[Federal Register Volume 80, Number 159 (Tuesday, August 18, 2015)]
[Rules and Regulations]
[Pages 50103-50187]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2015-19600]



[[Page 50103]]

Vol. 80

Tuesday,

No. 159

August 18, 2015

Part III





Securities and Exchange Commission





-----------------------------------------------------------------------





17 CFR Parts 229, 240, and 249





Pay Ratio Disclosure; Final Rule

Federal Register / Vol. 80 , No. 159 / Tuesday, August 18, 2015 / 
Rules and Regulations

[[Page 50104]]


-----------------------------------------------------------------------

SECURITIES AND EXCHANGE COMMISSION

17 CFR Parts 229, 240, and 249

[Release Nos. 33-9877; 34-75610; File No. S7-07-13]
RIN 3235-AL47


Pay Ratio Disclosure

AGENCY: Securities and Exchange Commission.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY:  We are adopting amendments to Item 402 of Regulation S-K to 
implement Section 953(b) of the Dodd-Frank Wall Street Reform and 
Consumer Protection Act. Section 953(b) directs the Commission to amend 
Item 402 of Regulation S-K to require disclosure of the median of the 
annual total compensation of all employees of a registrant (excluding 
the chief executive officer), the annual total compensation of that 
registrant's chief executive officer, and the ratio of the median of 
the annual total compensation of all employees to the annual total 
compensation of the chief executive officer. The disclosure is required 
in any annual report, proxy or information statement, or registration 
statement that requires executive compensation disclosure pursuant to 
Item 402 of Regulation S-K. The disclosure requirement does not apply 
to emerging growth companies, smaller reporting companies, or foreign 
private issuers.

DATES: Effective Date: October 19, 2015.
    Compliance Date: Registrants must comply with the final rule for 
the first fiscal year beginning on or after January 1, 2017.

FOR FURTHER INFORMATION CONTACT: John Fieldsend, Special Counsel in the 
Office of Rulemaking, at (202) 551-3430, in the Division of Corporation 
Finance; 100 F Street NE., Washington, DC 20549.

SUPPLEMENTARY INFORMATION: We are adopting amendments to Item 402 \1\ 
of Regulation S-K \2\ and a conforming amendment to Form 8-K \3\ under 
the Securities Exchange Act of 1934 (the ``Exchange Act'').\4\
---------------------------------------------------------------------------

    \1\ 17 CFR 229.402.
    \2\ 17 CFR 229.10 et seq.
    \3\ 17 CFR 249.308.
    \4\ 15 U.S.C. 78a et seq.
---------------------------------------------------------------------------

Table of Contents

I. Background
    A. Section 953(b) of the Dodd-Frank Act
    B. Summary of the Proposed Rule
    C. Summary of the General Comments on the Proposed Rule
    D. Summary of Changes in the Final Rule
    1. Non-U.S. Employee Exemptions and Additional Permitted 
Disclosure
    2. Employees of Consolidated Subsidiaries
    3. Employed on Any Date Within Three Months of the Last 
Completed Fiscal Year
    4. Identifying the Median Employee Once Every Three Years
    5. Initial Compliance Date
    6. Transition Period for New Registrants
    7. Additional Transition Periods
II. Discussion
    A. Scope of Final Rule
    1. Pay Ratio Disclosure Requirements Under New Paragraph (u) to
    Item 402 of Regulation S-K
    a. Proposed Rule
    b. Comments on the Proposed Rule
    c. Final Rule
    2. Pay Ratio Disclosure in Filings That Require Item 402 of
    Regulation S-K Information
    a. Proposed Rule
    b. Comments on the Proposed Rule
    c. Final Rule
    3. Excluded Registrants--Smaller Reporting Companies, Foreign 
Private Issuers, MJDS Filers, and Emerging Growth Companies
    a. Proposed Rule
    b. Comments on the Proposed Rule
    c. Final Rule
    B. Requirements of Final Rule
    1. ``All Employees'' Covered Under the Rule
    a. Types of Employees
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    b. Employed on Any Date Within Three Months of the Last 
Completed Fiscal Year
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    c. Employees Located Outside the United States
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    (a) Non-U.S. Employees Generally
    (b) Foreign Data Privacy Law Exemption
    (c) De Minimis Exemption
    (d) Cost-of-Living Adjustment
    d. Employees of Consolidated Subsidiaries
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    e. Any PEO Compensation in the Last Full Fiscal Year
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    f. Additional Information is Permissible
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    g. Annualizing Permanent Employees is Permissible, but Other 
Compensation Adjustments are Prohibited
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    2. Identifying the Median Employee and Calculating Annual Total 
Compensation
    a. Identifying the Median Employee
    i. Once Every Three Years
    (a) Proposed Rule
    (b) Comments on the Proposed Rule
    (c) Final Rule
    ii. Using Annual Total Compensation, Another Consistently 
Applied Compensation Measure, Statistical Sampling, Reasonable 
Estimates, or Other Reasonable Methods
    (a) Proposed Rule
    (b) Comments on the Proposed Rule
    (i) Flexibility
    (ii) Statistical Sampling
    (iii) Consistently Applied Compensation Measures
    (iv) The ``Median'' Employee
    (c) Final Rule
    (i) Flexibility
    (ii) Statistical Sampling
    (iii) Consistently Applied Compensation Measures
    (iv) The ``Median'' Employee
    b. Calculating Annual Total Compensation
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    3. Disclosure of Methodology, Assumptions, and Estimates
    a. Proposed Rule
    b. Comments on the Proposed Rule
    c. Final Rule
    4. Meaning of ``Annual''
    a. Proposed Rule
    b. Comments on the Proposed Rule
    c. Final Rule
    5. ``Filed'' Not ``Furnished''
    a. Proposed Rule
    b. Comments on the Proposed Rule
    c. Final Rule
    6. Timing of Disclosure
    a. Updating Pay Ratio Disclosure for the Last Completed Fiscal 
Year
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    b. Omitting Salary or Bonus Information for the PEO in Reliance 
on
    Instruction 1 to Item 402(c)(2)(iii) and (iv), and Technical 
Amendment to Item 5.02(f) of Form 8-K
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    c. Initial Compliance Date
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    d. Transition Period for New Registrants
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
    e. Additional Transition Periods
    i. Proposed Rule
    ii. Comments on the Proposed Rule
    iii. Final Rule
III. Economic Analysis
    A. Background
    B. Baseline
    C. Economic Effects From Mandated Disclosure Requirements
    1. Benefits
    2. Costs
    a. General
    b. Compliance Cost Estimates in Comment Letters
    i. Center on Executive Compensation Survey
    ii. Chamber of Commerce Survey

[[Page 50105]]

    iii. Other Specific Comments
    c. Quantification of Compliance Costs
    d. Indirect Costs
    3. Other Economic Effects
    D. Economic Effects From Exercise of Discretion
    1. General
    2. Implementation Choices and Alternatives
    a. Filings Subject to Pay Ratio Disclosure Requirements
    b. Registrants Subject to the Pay Ratio Disclosure Requirements
    c. Employees Included in the Determination of the Median
    i. Types of Employees
    ii. Workers Not Employed by the Registrant
    (i.e., ``Leased'' Workers'')
    iii. Employees of Consolidated Subsidiaries
    iv. Employees Located Outside of the United States
    v. Foreign Data Privacy Law Exemption
    vi. De Minimis Exemption
    vii. Calculation Date
    d. Adjustments to the Compensation of Employees
    e. Frequency of Identifying the Median Employee
    f. Method of Identifying the Median Employee
    i. Consistently Applied Compensation Measure
    ii. Statistical Sampling
    g. Disclosure of Methodology, Assumptions, and Estimates
    h. Determination of Total Compensation
    i. Defining ``Annual''
    j. Updating the Pay Ratio Disclosure for the Last Completed
    Fiscal Year
    k. Status of Disclosure as ``Filed''
    l. Compliance Date
    m. Transition Periods
IV. Paperwork Reduction Act
    A. Background
    B. Summary of Information Collections
    C. Summary of Comment Letters and Revisions to Proposal
    D. Revisions to PRA Reporting and Cost Burden Estimates
    1. Estimated Internal Burden Hours
    2. Estimated Cost Burdens
    3. Estimated Cost and Hour Burdens for Each Collection of 
Information
    a. Regulation S-K
    b. Form 10-K
    c. Form 8-K
    d. Proxy Statements on Schedule 14A
    e. Information Statements on Schedule 14C
    f. Form S-1
    g. Form S-4
    h. Form S-11
    i. Form N-2
    j. Form 10
    E. Summary to Changes to Annual Compliance Burden in Collection 
of Information
V. Final Regulatory Flexibility Act Certification
VI. Statutory Authority

I. Background

A. Section 953(b) of the Dodd-Frank Act

    Section 953(b)(1) of the Dodd-Frank Wall Street Reform and Consumer 
Protection Act (the ``Dodd-Frank Act'') \5\ directs us to amend Item 
402 of Regulation S-K (``Item 402'') \6\ to require each registrant, 
other than an emerging growth company, as that term is defined in 
Section 3(a) of the Exchange Act, to disclose in any filing of the 
registrant described in Item 10(a) of Regulation S-K (or any successor 
thereto): \7\ (A) The median of the annual total compensation of all 
employees of the registrant, except the chief executive officer 
(``CEO'') (or any equivalent position) of the registrant; (B) the 
annual total compensation of the CEO (or any equivalent position) of 
the registrant; and (C) the ratio of the median of the total 
compensation of all employees of the registrant to the annual total 
compensation of the CEO of the registrant. Section 953(b)(2) specifies 
that, for purposes of Section 953(b), ``total compensation'' of an 
employee of a registrant shall be determined in accordance with Item 
402(c)(2)(x) of Regulation S-K as in effect on the day before the date 
of enactment of the Dodd-Frank Act. As discussed in detail below, we 
are adopting amendments to Item 402 to implement Section 953(b).\8\ We 
refer to this disclosure of the median of the annual total compensation 
of all employees of the registrant, the annual total compensation of 
the principal executive officer (``PEO'') of the registrant,\9\ and the 
ratio of the two amounts as ``pay ratio'' disclosure.
---------------------------------------------------------------------------

    \5\ Public Law 111-203, sec. 953(b), 124 Stat. 1376, 1904 
(2010), as amended by Public Law 112-106, sec. 102(a)(3), 126 Stat. 
306, 309 (2012). Section 102(a)(3) of the JOBS Act amended Section 
953(b) of the Dodd-Frank Act to provide an exemption for registrants 
that are emerging growth companies as that term is defined in 
Section 3(a) of the Exchange Act.
    \6\ 17 CFR 229.402. As discussed in greater detail below, 
consistent with Section 953(b), the final rule requires a registrant 
to provide the pay ratio disclosure in any filing described in Item 
10(a) of Regulation S-K that calls for executive compensation 
disclosure under Item 402, including annual reports on Form 10-K, 
registration statements under the Securities Act and Exchange Act, 
and proxy and information statements, to the same extent that these 
forms require compliance with Item 402. Therefore, any company that 
provides such a filing is subject to the final rule. Section 953(b) 
refers to any such company as an ``issuer.'' In this release, to be 
consistent with other releases, we generally refer to such a company 
as a ``registrant.'' For the purposes of this release, unless 
otherwise expressly specified, these terms are used interchangeably.
    \7\ 17 CFR 229.10(a).
    \8\ On September 18, 2013, we proposed amendments to implement 
Section 953(b). See Pay Ratio Disclosure, Release No. 33-9452 (Sept. 
18, 2013) [78 FR 60560] (``Proposing Release'').
    \9\ The term ``CEO'' in the executive compensation rules was 
replaced by the term ``PEO'' as part of the 2006 amendments to Item 
402 in order to maintain consistency with the nomenclature used in 
Item 5.02 of Form 8-K. See Executive Compensation and Related Person 
Disclosure, Release No. 33-8732A, n. 326 (Aug. 29, 2006) [71 FR 
53158] (``2006 Adopting Release''). Consistent with the language of 
current Item 402, both the proposed rule and the final rule use the 
term ``PEO'' in lieu of ``CEO.''
---------------------------------------------------------------------------

    Congress did not expressly state the specific objectives or 
intended benefits of Section 953(b), and the legislative history of the 
Dodd-Frank Act also does not expressly state the Congressional purpose 
underlying Section 953(b).\10\ As discussed below, based on our 
analysis of the statute and comments received, we believe Section 
953(b) was intended to provide shareholders with a company-specific 
metric that can assist in their evaluation of a registrant's executive 
compensation practices. Accordingly, we have sought to tailor the final 
rule to meet that purpose while avoiding unnecessary costs.
---------------------------------------------------------------------------

    \10\ See letters from National Association of Manufacturers 
(Jul. 6, 2015) (``NAM II'') (stating that it concurs with our 
conclusion in the Proposing Release that neither the statute nor the 
related legislative history directly states the objectives or 
intended benefits of the provision) and WorldatWork (Jul. 6, 2015) 
(``WorldatWork II'').
---------------------------------------------------------------------------

    In informing our understanding of the Congressional purpose of 
Section 953(b), we have considered the surrounding provisions of the 
Dodd-Frank Act \11\ as well as the comments that we received during 
this rulemaking. Subtitle E of Title IX of the Dodd-Frank Act, headed 
--``Accountability and Executive Compensation'' is, as explained in the 
Conference Report for the legislation, ``designed to address 
shareholder rights and executive compensation practices.'' \12\ Its 
provisions, including Section 953(b), address various aspects of 
executive compensation with a focus on encouraging shareholder 
engagement in executive compensation matters by, among other things, 
increasing the transparency of compensation. In Section 951, for 
example, Congress required companies to provide for periodic 
shareholder votes on executive compensation. In implementing Congress's 
directive, we noted that a key function of the disclosures required 
incident to the new voting requirement was to ``provide shareholders 
and investors with timely information'' that was potentially useful to 
them ``as they

[[Page 50106]]

consider voting and investment decisions.'' \13\ Section 952 requires, 
in turn, that both compensation committee members of registrants and 
their advisors be independent. We noted that the rules implementing 
Section 952 could serve an informational purpose that benefits 
``investors to the extent they enable compensation committees to make 
better informed decisions regarding the amount or form of executive 
compensation.'' \14\ Further, as we noted in the release proposing 
implementation of Section 953(a), that section is intended to provide 
shareholders with metrics that will help them assess executive 
compensation relative to the registrant's performance.\15\ The Section 
953(a) information is intended, among other things, to assist 
shareholders when exercising their say-on-pay voting rights under 
Section 951.
---------------------------------------------------------------------------

    \11\ See, e.g., Medtronic, Inc. v. Lohr, 518 U.S. 470, 496 
(1996) (``Congress' intent, of course, primarily is discerned from 
the language of the . . . statute and the `statutory framework' 
surrounding it. Also relevant, however, is the `structure and 
purpose of the statute as a whole,' as revealed not only in the 
text, but through . . . reasoned understanding of the way in which 
congress intended the statute and its surrounding regulatory scheme 
to affect business, consumers, and law.'') and Maine Public 
Utilities Comm'n v. FERC, 545 F.3d 278, 282 (D.C. Cir. 2006) (``This 
court applies the traditional tools of statutory interpretation in 
determining congressional intent, looking to the text, structure, 
purpose, and legislative history of a statute.'').
    \12\ Dodd-Frank Act, H.R. Rep. 111-517, at 872 (2010) (Conf. 
Rep.).
    \13\ Shareholder Approval of Executive Compensation and Golden 
Parachute Compensation, Release No. 33-9178 (Jan. 25, 2011) [76 FR 
6010, 6037 (Feb. 2, 2011)].
    \14\ Listing Standards for Compensation Committees, Release No. 
33-9330 (June 20, 2012) [77 FR 38422, 38447 (Jun. 27, 2012)].
    \15\ Pay Versus Performance, Release No. 34-74835 (Apr. 29, 
2015) [80 FR 26329 (May 7, 2015)].
---------------------------------------------------------------------------

    We believe that Section 953(b) should be interpreted consonant with 
Subtitle E's general purpose of further facilitating shareholder 
engagement with executive compensation. Thus, we believe that Congress 
intended Section 953(b) to supplement the executive compensation 
information available to shareholders. Particularly, Section 953(b) 
provides new data points that shareholders may find relevant and useful 
when exercising their voting rights under Section 951.\16\ Several 
commenters stated affirmatively that they would find the new data 
points, including pay ratio disclosure, relevant and useful when making 
voting decisions.\17\ Some commenters in the pre-proposing period \18\ 
suggested specifically that shareholders of public companies could use 
the pay ratio information, together with pay-versus-performance 
disclosure, to help inform their say-on-pay votes, which could also be 
a tool for shareholders to hold companies accountable for their CEO 
compensation.\19\ A significant consideration for us in fashioning a 
final rule implementing Section 953(b), then, is the extent to which 
elements of the final rule further Congress's apparent goal of giving 
shareholders additional executive compensation information to enhance 
the shareholder engagement envisioned by Section 951.\20\
---------------------------------------------------------------------------

    \16\ We note that the say-on-pay votes extend to certain other 
senior officers at the registrant beyond the PEO while the pay ratio 
disclosure is solely focused on a comparison of the PEO's 
compensation to the median employee's compensation. However, we do 
not think that this diminishes the overall utility of the pay ratio 
disclosures to say-on-pay votes. The PEO will typically be the 
highest compensated officer at a registrant and, to the extent 
shareholders rely on the pay ratio disclosure to determine whether 
the PEO's compensation is appropriate or not, it also may inform 
shareholders' relative assessment of the compensation of the other 
senior officers whose compensation is subject to say-on-pay votes.
    \17\ See, e.g., letters from AFL-CIO (Dec. 2, 2013) (``AFL-CIO 
I''); American Federation of State County and Municipal Employees 
(Nov. 27, 2013) (``AFSCME''); Amalgamated Bank (Nov. 25, 2013) 
(``Amalgamated''); Bricklayers & Trowel Trades International Pension 
Fund (Nov. 27, 2013) (``Bricklayers International''); California 
State Teachers' Retirement System (Dec. 2, 2013) (``CalSTRS''); 
Calvert Investments (Dec. 5, 2013) (``Calvert''); Chevy Chase Trust 
(Nov. 25, 2013) (``Chevy Chase Trust''); Corporate Governance (Nov. 
25, 2013) (``CorpGov.net''); Form Letter C; Form Letter D; Form 
Letter E; Form Letter F; Laborers' International Union of North 
America (Oct. 10, 2013) (``LIUNA''); Local Authority Pension Fund 
Forum (Nov. 10, 2013) (``LAPFF''); New York State Comptroller (Nov. 
27, 2013) (``NY State Comptroller''); Pax World Management LLC (Oct. 
10, 2013) (``Pax World Funds''); Public Citizen (Nov. 26, 2013) 
(``Public Citizen I''); The Forum for Sustainable and Responsible 
Investment (Dec. 2, 2013) (``US SIF''); Trillium Asset Management, 
LLC (Dec. 2, 2013) (``Trillium I''); Trillium Asset Management, LLC 
(Jul. 31, 2014) (``Trillium II''); and UAW Retirement Medical 
Benefits Trust (Nov. 21, 2013) (``UAW Trust'').
    \18\ To facilitate public input on rulemaking required by the 
Dodd-Frank Act, we provided a series of email links, organized by 
topic, on our Web site at http://www.sec.gov/spotlight/regreformcomments.shtml, so that the public could provide comments 
before we proposed a rule. The comments relating to Section 953(b) 
are located at http://www.sec.gov/comments/df-title-ix/executive-compensation/executive-compensation.shtml (``Pre-Proposing Release 
Web site''). Comments that we received after we published the 
Proposing Release are located at http://www.sec.gov/comments/s7-07-13/s70713.shtml. Our references to comment letters refer to the 
comments on the proposal unless otherwise specified.
    \19\ See, e.g., letters from CtW Investment Group (Jun. 20, 
2011) (``CtW Investment Group pre-proposal letter) and Steven Towns 
(Feb. 6, 2012) (``S. Towns pre-proposal letter).
    \20\ We note that some commenters contended that the pay ratio 
disclosure is intended to publicly ``shame'' registrants concerning 
the size of the disparity between their CEO's compensation and their 
typical worker's compensation. See, e.g., letters from The Honorable 
Carolos Cardozo Campbell, Former Assistant Secretary of Commerce for 
Economic Development (Nov. 27, 2013) (``Former Assistant Secretary 
Campbell''); Center on Executive Compensation (Jul. 6, 2015) (``COEC 
III''); Hyster-Yale Materials Handling, Inc. (Dec. 2, 2013) 
(``Hyster-Yale''); NACCO Industries, Inc. (Dec. 2, 2013) 
(``NACCO''); and U.S. Chamber of Commerce (Dec. 2, 2013) (``Chamber 
I''). As discussed above, we have reached a different conclusion 
based on principles of statutory construction and have taken no such 
objective into account in framing the rule. In crafting the final 
rule, we have sought to carefully tailor the pay ratio disclosure 
requirement so that it provides shareholders with a company-specific 
metric that is relevant and useful to their say-on-pay voting.
---------------------------------------------------------------------------

    Consistent with this understanding of the Congressional purpose of 
Section 953(b), we believe the final pay ratio rule should be designed 
to allow shareholders to better understand and assess a particular 
registrant's compensation practices and pay ratio disclosures rather 
than to facilitate a comparison of this information from one registrant 
to another.\21\ As we noted in the Proposing Release, we do not believe 
that precise conformity or comparability of the pay ratio across 
companies is necessarily achievable given the variety of factors that 
could cause the ratio to differ. Consequently, we believe the primary 
benefit of the pay ratio disclosure is to provide shareholders with a 
company-specific metric that they can use to evaluate the PEO's 
compensation within the context of their company.
---------------------------------------------------------------------------

    \21\ See, e.g., letters from CalSTRS, LAPFF, RPMI Railpen 
Investments (Dec. 2, 2013) (``RPMI''), Rep. Keith Ellison et al. 
(Dec. 2, 2013) (``Rep. Ellison et al. I''), Rep. Keith Ellison et 
al. (Mar. 17, 2015) (``Rep. Ellison et al. II''), and Sen. Robert 
Menendez et al. (Dec. 16, 2014) (``Sen. Menendez et al. II'').
---------------------------------------------------------------------------

    On the other hand, some commenters asserted that the pay ratio 
disclosure would not provide meaningful or material information to 
shareholders in making voting or investment decisions.\22\ In support 
of this contention, some of these commenters cited studies 
demonstrating that shareholders are not interested in this 
information,\23\ some commenters cited shareholder votes indicating a 
high level of support for executive pay and little support for 
shareholder proposals advocating for pay ratio disclosure,\24\

[[Page 50107]]

and some commenters contended that pay ratio disclosure would confuse 
shareholders because they would rely on it without fully considering a 
company's detailed narrative disclosures.\25\ Notwithstanding the 
disagreement among commenters on the value of the pay ratio disclosure, 
in adopting the final rule we have sought to implement Congress's 
apparent determination that the pay ratio disclosure would be useful to 
shareholders.
---------------------------------------------------------------------------

    \22\ See, e.g., letters from James J. Angel, Ph.D., CFA, 
Visiting Associate Professor, The Wharton School, University of 
Pennsylvania (Nov. 22, 2013) (``Prof. Angel''); Avery Dennison 
Corporation (Nov. 26, 2013) (``Avery Dennison''); Bill Barrett 
Corporation (Dec. 2, 2013) (``Bill Barrett Corp.''); Business 
Roundtable (Dec. 2, 2013) (``Business Roundtable I''); Center on 
Executive Compensation (Dec. 2, 2013) (``COEC I''); Center on 
Executive Compensation (Sep. 26, 2014) (``COEC II''); COEC III; 
Hyster-Yale; Mercer, Inc. (Dec. 2, 2013) (``Mercer I''); NACCO; and 
Pearl Meyer and Partners (Dec. 2, 2013) (``PM&P'').
    \23\ See, e.g., letters from Chamber I (citing the Center for 
Audit Quality's 7th annual Main Street Investor Survey, which ranked 
CEO compensation last on the list of factors used by shareholders to 
make investment decisions, with only 16% saying it was essential to 
their decision); COEC I (acknowledging that, while some literature 
focuses on pay disparities among employees with comparable jobs, the 
study frequently cited for the impact of disparities on 
collaboration, ``Pay Disparities Within Top Management Groups: 
Evidence of Harmful Effects on Performance of High-Technology 
Firms'' by Phyllis Siegel and Donald C. Hambrick, concerns executive 
pay and pay disparities among top executives, it does not discuss 
pay disparities between the CEO and median employee), and 
International Bancshares Corporation (Nov. 25, 2013) (``IBC'') 
(citing a Wall Street Journal article that says only 10% of 
individuals polled believed pay ratio would have value to 
shareholders) .
    \24\ See, e.g., letters from Chamber I (indicating that the 
average support for management compensation in public companies was 
90.1%, and 97.6% of companies received majority shareholder support 
for executive compensation) and COEC I (noting that, since 2010, 
there have been 14 shareholder proposals advocating that companies 
provide pay ratio disclosure, and those proposals averaged 93% 
opposition from shareholders, with none receiving at least 10% 
support from shareholders).
    \25\ See, e.g., letters from Bill Barrett Corp., COEC I, 
National Investor Relations Institute (Oct. 17, 2013) (``NIRI''), 
and Semtech Corporation (Nov. 27, 2013) (``Semtech''). Two of these 
commenters suggested specifically that we should undertake an 
education effort to help shareholders understand the limits of pay 
ratio disclosure and remind them that they can find other 
information, such as an executive summary of the Compensation 
Discussion and Analysis section of a company's proxy statement, 
which can provide a more complete understanding of corporate pay 
practices. See letters from NIRI and Semtech.
---------------------------------------------------------------------------

    We also recognize that many commenters raised significant concerns 
about the costs of providing the required pay ratio disclosure. In 
implementing the statutory requirements, we have exercised our 
exemptive authority and provided flexibility in a manner that we expect 
will reduce costs and burdens for registrants, while preserving what we 
perceive to be the purpose and intended benefits of the disclosure 
required by Section 953(b).\26\ The significant cost estimates of the 
pay ratio disclosure submitted by some commenters support our view that 
some accommodations are appropriate.\27\ The final rule, therefore, 
both maintains the flexibility and accommodations from the proposal 
(such as permitting the use of statistical sampling and a consistent 
compensation measure to identify the median employee and reasonable 
estimates to calculate total compensation) and provides additional 
flexibility as follows: The final rule takes a flexible approach to the 
methodology a registrant can use to identify its median employee and 
calculate the median employee's annual total compensation; provides a 
de minimis exemption for non-U.S. employees and an exemption for 
registrants where, despite reasonable efforts to obtain or process the 
information necessary for compliance with the final rule, they are 
unable to do so without violating a foreign jurisdiction's laws or 
regulations governing data privacy; permits cost-of-living adjustments 
for the compensation of employees in jurisdictions other than the 
jurisdiction in which the PEO resides so that the compensation is 
adjusted to the cost of living in the jurisdiction in which the PEO 
resides; gives registrants the ability to make the median employee 
determination only once every three years and to choose as a 
determination date any date within the last three months of a 
registrant's fiscal year; and provides transition periods for new 
registrants, registrants engaging in business combinations or 
acquisitions, and registrants that cease to be smaller reporting 
companies or emerging growth companies.
---------------------------------------------------------------------------

    \26\ We are mindful of the principle that ``no legislation 
pursues its purposes at all costs,'' Rodriguez v. United States, 480 
U.S. 522, 525-26 (1987), and we believe that the accommodations that 
we have included within the final rule reflect an appropriate 
balance.
    \27\ The potential costs arising from the requirements of 
Section 953(b), as well as the potential costs relating to the final 
rule, are discussed in detail below in Section III of this release 
under the heading ``Economic Analysis''.
---------------------------------------------------------------------------

    Overall, we think the final rule will provide investors with 
information Congress intended them to have to assess the compensation 
and accountability of a company's PEO while seeking to limit the costs 
and practical difficulties of providing the disclosure.
    Finally, we recognize the possibility that, based on the specific 
facts and circumstances of a registrant's work force and corporate 
operations, the pay ratio disclosure may warrant additional disclosures 
from a registrant to ensure that, in the registrant's view, the pay 
ratio disclosure is a meaningful data point for investors when making 
their say-on-pay votes. While Congress appears to have believed that 
the pay ratio disclosure would be a useful data point, we recognize 
that its relative usefulness--taken alone without accompanying 
disclosures to provide potentially important context--may vary 
considerably. Rather than prescribe a one-size-fits-all catalogue of 
additional disclosures that registrants should provide to put the pay 
ratio disclosure in context, we believe it is the better course to 
provide registrants the flexibility to provide additional disclosures 
that they believe will assist investors' understanding of the meaning 
of pay ratio disclosure when making say-on-pay votes. In this way, we 
believe we can best fulfill Congress's directive in Section 953(b) 
while avoiding unnecessary costs and complexities that might result 
from mandating additional disclosures.

B. Summary of the Proposed Rule

    In September 2013, we proposed a new rule to implement Section 
953(b) of the Dodd-Frank Act.\28\ The proposal's goal was to implement 
the statutory directive, while minimizing costs. In response to public 
comments we received prior to the proposal about the significant 
potential costs of complying with this requirement, the proposed rule 
would allow registrants flexibility in developing the disclosure 
required by the statute. We recognized that a one-size-fits-all 
approach would not be appropriate, given the wide range of affected 
registrants and the disparate burdens that would be imposed on them 
based on such factors as their business types and the complexity of 
their payroll systems. We therefore proposed to implement Section 
953(b) in a manner that we believed would lower the cost of compliance 
while remaining consistent with the requirements of Section 953(b).
---------------------------------------------------------------------------

    \28\ See Proposing Release.
---------------------------------------------------------------------------

    The proposed rule would require companies to disclose the median of 
the annual total compensation of all its employees except the PEO, the 
annual total compensation of its PEO, and the ratio of the two amounts. 
The proposed rule would not have specified a single calculation 
methodology for identifying the median employee. Instead, it would 
permit registrants to select a methodology for identifying the median 
employee that was appropriate to the size and structure of their 
business and the way they compensate employees. Under the proposal, 
registrants could have chosen to identify the median employee by 
analyzing their full employee population or by using statistical 
sampling or another reasonable method. Also, to identify the median, 
registrants could have used ``total compensation,'' as defined in our 
existing rules, namely Item 402(c)(2)(x), or any consistently applied 
compensation measure, such as information derived from tax and/or 
payroll records. The proposed rule would not prescribe a particular 
methodology or specific computation parameters.
    Once the median employee was identified, the proposed rule would 
require the registrant to calculate the annual total compensation for 
that median employee in accordance with the definition of ``total 
compensation'' set forth in Item 402(c)(2)(x), which requires companies 
to provide extensive compensation information for the PEO and other 
named executive officers. ``Total compensation'' under Item 
402(c)(2)(x) is not ordinarily calculated for all employees. The 
proposed rule, therefore, would permit registrants to use reasonable 
estimates in calculating

[[Page 50108]]

any element of total compensation and in calculating the annual total 
compensation of the median employee. Also, the proposed rule would 
define ``annual total compensation'' to mean total compensation for the 
last completed fiscal year, which would be consistent with our existing 
executive compensation disclosure requirements.
    Under the proposal, if a registrant used a compensation measure 
other than annual total compensation to identify the median employee, 
it would be required to disclose the compensation measure it used. 
Also, the registrant would be required to briefly describe and 
consistently apply any methodology it used to identify the median and 
any material assumptions, adjustments, or estimates used to identify 
the median employee or determine total compensation or any elements of 
total compensation for that employee or the PEO, and the registrant 
would need to clearly identify any amounts it estimated. Finally, 
registrants would be permitted, but not required, to supplement their 
disclosure with a narrative discussion or additional ratios if they 
chose to do so.
    Section 953(b) does not define the term ``employee.'' The proposed 
rule would define that term, for purposes of pay ratio disclosure, to 
include any individual employed by the registrant or any of its 
subsidiaries as of the last day of the registrant's last completed 
fiscal year. The proposed definition would encompass any full-time, 
part-time, seasonal, or temporary employees of the registrant or any of 
its subsidiaries, including any non-U.S. employee. Also, a registrant 
would be permitted, but not required, to annualize the total 
compensation for a permanent employee who was employed at year-end but 
did not work for the entire year. In contrast, full-time equivalent 
adjustments for part-time employees, annualizing adjustments for 
temporary and seasonal employees, and cost-of-living adjustments for 
non-U.S. employees would not be permitted.
    Also, under the proposal, registrants would be required to provide 
the proposed pay ratio disclosure in registration statements, proxy and 
information statements, and annual reports required to include 
executive compensation information as set forth under Item 402. 
Registrants, however, would not be required to provide their pay ratio 
information in reports that did not include Item 402 executive 
compensation information, such as current and quarterly reports. 
Additionally, registrants would not be required to update their annual 
pay ratio disclosure until they filed their annual report on Form 10-K 
for their last completed fiscal year or, if later, their definitive 
proxy or information statement for their next annual meeting of 
shareholders (or written consents in lieu of such a meeting). 
Registrants, however, would still be required to file their pay ratio 
information no later than 120 days after the end of the last fiscal 
year as provided in General Instruction G(3) of Form 10-K.\29\
---------------------------------------------------------------------------

    \29\ 17 CFR 249.310.
---------------------------------------------------------------------------

    The proposal would provide a transition period for newly public 
companies. For these companies, initial compliance would be required 
with respect to compensation for the first fiscal year commencing on or 
after the date the company became subject to the reporting 
requirements. Also, as provided by the Jumpstart Our Business Startups 
Act (``JOBS Act''),\30\ the proposed rule would not apply to emerging 
growth companies. Finally, the proposed rule would not apply to smaller 
reporting companies or foreign private issuers.
---------------------------------------------------------------------------

    \30\ 15 U.S.C. 78c(a).
---------------------------------------------------------------------------

C. Summary of the General Comments on the Proposed Rule

    In the Proposing Release, we requested comment on many aspects of 
the proposed rule, including whether the proposed rule would address 
sufficiently the practical difficulties of data collection, whether 
other alternative approaches consistent with Section 953(b) could 
provide the potential benefits of pay ratio information at a lower 
cost, and whether the proposed flexible approach would appropriately 
implement Section 953(b). We received a large volume of comment letters 
from a variety of stakeholders. We received more than 287,400 comment 
letters, including over 1,540 individual letters that reflected a wide 
range of views concerning the proposed rule and the potential costs and 
benefits associated with its requirements. We received comments that 
addressed the proposed rule as a whole (including commenters that 
supported or opposed the rule in its entirety) as well as comments 
directed toward particular requirements of the rule, such as its 
application to foreign, part-time, temporary, and seasonal employees. 
In this section, we summarize the general comments on the proposal as a 
whole. Comments on particular provisions of the proposed rule are 
addressed as part of the discussion of each specific provision of the 
rule in Section II below.
    Of the over 287,400 total comment letters we received, over 285,900 
were form letters regarding the proposed rule. There were 12 types of 
form letters,\31\ and these letters either supported our proposed rule 
or supported the idea of adopting a rule based on Section 953(b) 
without specifically referencing the proposal.
---------------------------------------------------------------------------

    \31\ We have received 285,936 form letters. Form Letter A 
(19,965 letters) supporting a strong rule generally; Form Letter B 
(12,942 letters) supporting a strong rule generally; Form Letter C 
(20 letters) supporting the Proposed Rule; Form Letter D (5,428 
letters) supporting the Proposed Rule; Form Letter E (1,688 letters) 
supporting the Proposed Rule; Form Letter F (1,167 letters) 
supporting the Proposed Rule; Form Letter G (15,304 letters) 
supporting the Proposed Rule; Form Letter H (70,338 letters) 
supporting a rule generally; Form Letter I (36,299 letters) 
supporting the adoption of a rule; Form Letter J (75,333 letters) 
supporting a strong rule; Form Letter K (15,247 letters) supporting 
a strong rule generally; and Form Letter L (32,275 letters) 
supporting a rule.
---------------------------------------------------------------------------

    For example, one form letter asserted that the pay ratio disclosure 
is material to investors because high pay disparities can impair 
employee morale and productivity and have negative consequences on a 
company's overall performance and because investors will have a 
``valuable additional'' measure for evaluating executive compensation, 
including when making say-on-pay voting decisions.\32\ Also, the letter 
supported the proposed rule's inclusion of all employees, including 
non-U.S. and part-time employees, and its flexibility in identifying 
the median employee. Other form letters also indicated that the pay 
ratio disclosure is material to investors,\33\ including some that 
noted that the disclosure would aid them in making voting 
decisions.\34\
---------------------------------------------------------------------------

    \32\ See Form Letter C.
    \33\ See letters from Form Letter A, Form Letter B, Form Letter 
D, Form Letter E, Form Letter F, and Form Letter G.
    \34\ See letters from Form Letter D (``Pay ratio disclosure will 
help investors evaluate CEO pay levels when voting on executive 
compensation matters. The ratio of the CEO-to-worker pay is a 
valuable metric for investors, because it places CEO pay levels into 
a broader perspective.''), Form Letter E (``Pay ratio disclosure 
will help investors evaluate CEO pay levels when voting on executive 
compensation matters. The ratio of the CEO-to-worker pay is a 
valuable tool for investors in evaluating and voting on CEO pay; 
scrutinizing the performance of Boards of Directors; and, 
identifying possible investment risks.''), Form Letter F (``A pay 
ratio disclosure will help investors better evaluate CEO pay levels 
when voting on executive compensation matters. Compensation experts 
have found that there is a correlation between high CEO pay and poor 
performance. By mandating disclosure of the ratio of CEO to worker 
pay, inequities will be become more transparent.'').
---------------------------------------------------------------------------

    Additionally, the vast majority of the over 1,500 unique comment 
letters were from individual commenters who, like those submitting the 
form letters, supported the proposed rule or supported the idea of 
adopting a rule based on Section 953(b) without specifically 
referencing the proposal. Most of these individuals supported the

[[Page 50109]]

proposed rule or the pay ratio disclosure because they believed it 
would:
     Inform shareholders about executive compensation matters, 
especially with regard to say-on-pay voting; \35\
---------------------------------------------------------------------------

    \35\ See, e.g., letters from D.A. Alexander (Nov. 22, 2013) 
(``Alexander''), Jean M. Blair (Sep. 25, 2013) (``J. Blair''), Cathy 
Clemens (Sep. 25, 2013) (``Clemens''), Beth Finchler (Sep. 27, 2013) 
(``Finchler''), Amy Hevron (``Sep. 28, 2013) (``Hevron''), Emanuel 
Jacobowitz (Sep. 24, 2013) (``Jacobowitz''), Rachel LaBruyere (Oct. 
17, 2013) (``LaBruyere''), Gabrielle Loperfido (Sep. 25, 2013) 
(``Loperfido''), Carol Nix (Sep. 24, 2013) (``Nix''), Bonnie 
Overcott (Jan. 4, 2014) (``Overcott''), Lynn Reilly (Sep. 22, 2013) 
(``Reilly''), Kendall Simmons (Sep. 24, 2013) (``Simmons''), Dory 
Storms (Sep. 24, 2013) (``Storms''), Amy Sullivan-Greiner (Sep. 24, 
2013) (``Sullivan-Greiner''), Jackie Tortora (Oct. 21, 2013) 
(``Tortora''), and Bryan Taylor (Nov. 21, 2013) (``Taylor'').
---------------------------------------------------------------------------

     demonstrate a company's focus on its long-term health as 
opposed to short-term gains that benefit its executives at the expense 
of its shareholders; \36\
---------------------------------------------------------------------------

    \36\ See, e.g., letters from Anonymous (Dec. 1, 2013) 
(``Anonymous''), Eric C. Gade (Sep. 18, 2013) (``Gade''), Linda 
Kranen (Oct. 20, 2013) (``Kranen''), Alyce Lomax (Dec. 2, 2013) 
(``Lomax''), Holly Schroeder (Nov. 5, 2013) (``Schroeder''), Erika 
Skornia-Olsen (Nov. 5, 2013) (``Skornia-Olsen''), and Calvin Vu 
(Oct. 26, 2013) (``Vu'').
---------------------------------------------------------------------------

     discourage the pay practices that led to the 2008 
financial crisis; \37\
---------------------------------------------------------------------------

    \37\ See, e.g., letters from Lisbeth Caccese (Sep. 25, 2013) 
(``Caccese''), Hope Carr (Sep. 24, 2013) (``Carr''), Sarah McKee 
(Sep. 24, 2013) (``McKee''), Thomas Motes (Sep. 24, 2013) 
(``Motes''), Karl David Reinhardt (Sep. 24, 2013) (``Reinhardt''), 
Cynthia Sommer (Sep. 24, 2013) (``Sommer''), Cody Spann (Sep. 24, 
2013) (``Spann''), Kathy Van Dame (Sep. 24, 2013) (``Van Dame''), 
Marietta Whittlesey (Sep. 24, 2013) (``Whittlesey''), Susan Williams 
(Sep. 24, 2013) (``S. Williams''), Mary M. Williamson (Sep. 24, 
2013) (``M. Williamson''), and Robin Wittrock (Sep. 24, 2013) 
(``Wittrock'').
---------------------------------------------------------------------------

     reduce the inequitable wealth distribution in the U.S.; 
\38\ and
---------------------------------------------------------------------------

    \38\ See, e.g., letters from Paul Cohen (Oct. 3, 2013) 
(``Cohen''), Alan Harris (Sep. 19, 2013) (``Harris''), Liz A. King 
(Sep. 23, 2013) (``L. King''), Ben Leet (Oct. 27, 2013) (``Leet''), 
Laurie H. Norton (Sep. 18, 2013) (``Norton''), Debbie Notkin (Sep. 
24, 2013) (``Notkin''), and Desmonde Printz (Sep. 19, 2013) 
(``Printz'').
---------------------------------------------------------------------------

     highlight potential problems in a company due to the 
negative impact of a high pay ratio on employee morale and 
productivity.\39\
---------------------------------------------------------------------------

    \39\ See, e.g., letters from Daniel Grossman (Nov. 8, 2013) 
(``Grossman''), Peter Linton (Nov. 17, 2013) (``Linton''), Michael 
R.K. Mudd (Nov. 3, 2013) (``Mudd''), Vivian Rosati (Nov. 6, 2013) 
(``Rosati''), and Walden Asset Management (Nov. 15, 2013) 
(``Walden'').
---------------------------------------------------------------------------

    As discussed in greater detail below, many commenters supported the 
proposed rule's overall flexibility.\40\ Some commenters asserted that 
the permitted flexibility would lessen the costs and burdens of the 
proposed rule without reducing the rule's benefits,\41\ be consistent 
with the directives of Section 953(b),\42\ and have minimal effect on 
the pay ratio disclosure.\43\ Other commenters, however, opposed the 
proposed rule because they believed it provided too much flexibility, 
which they asserted would allow registrants to manipulate the ratio in 
their favor,\44\ decrease the ratio's utility (especially for comparing 
the ratios of different companies),\45\ and still lead to high 
costs.\46\ One commenter suggested that the final rule consist of 
little more than ``a simple restatement'' of Section 953(b), doing 
``little more than rearranging a few commas and adding a word or two to 
the statutory language.'' \47\
---------------------------------------------------------------------------

    \40\ See, e.g., letters from Prof. Angel; Aon Hewitt (Dec. 3, 
2013) (``Aon Hewitt''); B[acirc]tirente, Canada et al. (Nov. 28, 
2013) (``B[acirc]tirente et al.''); Best Buy Co., Inc., et al. (Dec. 
17, 2013) (``Best Buy et al.''), CalSTRS; Marcia K. Campbell, 
Trustee, Illinois Teachers Retirement System (Nov. 21, 2013) 
(``Trustee Campbell''); Capital Strategies Consulting, Inc. (Dec. 2, 
2013) (``Capital Strategies''); Center for Effective Government 
(Nov. 26, 2013) (``CEG''); Cummins Inc. (Dec. 2, 2013) (``Cummins 
Inc.''); CUPE Employees' Pension Plan (Nov. 21, 2013) (``CUPE''); 
Davis Polk & Wardwell LLP (Dec. 4, 2013) (``Davis Polk,''); Ernst & 
Young LLP (Dec. 2, 2013) (``E&Y''); First Affirmative Financial 
Network (Nov. 12, 2013) (``First Affirmative''); Fonds de 
solidarit[eacute] FTQ (Nov. 27, 2013) (``FS FTQ''); Freeport-McMoRan 
Copper & Gold Inc. (Dec. 2, 2013) (``Freeport-McMoRan''); Interfaith 
Center on Corporate Responsibility (Oct. 17, 2013) (``ICCR''); 
International Brotherhood of Teamsters (Oct. 9, 2013) 
(``Teamsters''); International Union of Bricklayers and Allied 
Craftworkers, Administrative District Council 1 of Illinois (Nov. 
15, 2013) (``IL Bricklayers and Craftworkers Union''); International 
Union of Bricklayers and Allied Craftworkers, Local 5 New York (Nov. 
15, 2013) (``NY Bricklayers and Craftworkers Union''); Intel 
Corporation (Nov. 27, 2013) (``Intel''); Johnson and Johnson (Dec. 
4, 2013) (``Johnson & Johnson''); Marco Consulting Group (Nov. 15, 
2013) (``Marco Consulting''); McMorgan & Company LLC (Dec. 2, 2013) 
(``McMorgan & Co.''); Mercer, Inc. (Dec. 2, 2013) (``Mercer I''); 
Meridian Compensation Partners, LLC (Dec. 2, 2013) (``Meridian''); 
Microsoft Corporation (Dec. 4, 2013) (``Microsoft''); Nathan 
Cummings Foundation (Nov. 21, 2013) (``Cummings Foundation''); 
Network for Sustainable Financial Markets (Dec. 2, 2013) (``NSFM''); 
New York City Bar Association (Dec. 5, 2013) (``NYC Bar''); Novara 
Tesija, PLLC (Nov. 6, 2013) (``Novara Tesija''); Organizational 
Capital Partners (Nov. 24, 2013) (``OCP''); Oxfam America (Oct. 16, 
2013) (``Oxfam''); PGGM (Dec. 2, 2013) (``PGGM''); Public Citizen I; 
Public School Teachers' Pension and Retirement Fund of Chicago (Nov. 
25, 2013) (``Chicago Teachers Fund''); Quintave (Dec. 1, 2013) 
(``Quintave''); Rep. Ellison et al. II, SEIU Master Trust (Dec. 2, 
2013) (``SEIU''); Sen. Menendez et al. II, Socially Responsive 
Financial Advisors/First Affirmative Financial Network (Oct. 11, 
2013) (``Socially Responsive Financial Advisors''); Society of 
Corporate Secretaries and Governance Professionals (Dec. 16, 2013) 
(``Corporate Secretaries''); Trillium I, UAW Trust; Vectren 
Corporation (Dec. 2, 2013) (``Vectren Corp.''); Washington State 
Investment Board (Nov. 26, 2013) (``WA State Investment Board''); 
and WorldatWork (Dec. 2, 2013) (``WorldatWork I'').
    \41\ See, e.g., letters from American Bar Association (Mar. 7, 
2014 (``ABA''), AFSCME, Chris Barnard (Nov. 6, 2013) (``Barnard''), 
Bricklayers International, Council of Institutional Investors (Nov. 
6, 2013) (``CII''), Kenneth Fowler (Nov. 3, 2013) (``Fowler''), 
LIUNA, Walter Mirczak (Oct. 21, 2013) (``Mirczak''), PNC Financial 
Services Group, Inc. (Dec. 2, 2013) (``PNC Financial Services''), 
Sen. Menendez et al. II, US SIF, Vivient Consulting LLC (Dec. 2, 
2013) (``Vivient''), and Walden.
    \42\ See, e.g., letters from Domini Social Investments LLC (Nov. 
27, 2013) (``Domini'') and PM&P.
    \43\ See, e.g., letters from Capital Strategies and WorldatWork 
I.
    \44\ See, e.g., letters from Sherry Bupp (Oct. 4, 2013) 
(``Bupp''), Chris Corayer (Oct. 8, 2013) (``Corayer''), Russell J. 
Fedewa (Oct. 4, 2013) (``Fedewa''), Eleanor J. Fox (Oct. 3, 2013) 
(``Fox''), Gary G. Friend II (Oct. 3, 2013) (``Friend''), Charles 
Grotzke (Oct. 3, 2013) (``Grotzke''), Bruce Hlodnicki (Sep. 19, 
2013) (``Hlodnicki''), Karla Kizzort (Oct. 4, 2013) (``Kizzort''), 
Christine Maly (Sep. 25, 2013) (``Maly''), B. A. Petricoin (Oct. 3, 
2013) (``Petricoin''), and Jasmine Van Pelt (Oct. 4, 2013) (``Van 
Pelt'').
    \45\ See, e.g., letters from Amundi Asset Management (Nov. 28, 
2013) (``Amundi''); British Columbia Investment Management 
Corporation (Dec. 2, 2013) (``BCIMC''); Paul Ciatto (Sep. 26, 2013) 
(``Ciatto''); Paul Glenn (Oct. 3, 2013) (``Glenn''); IBC; Karl T. 
Muth, Lecturer in Economics and Public Policy, Northwestern 
University (Sep. 24, 2013) (``Prof. Muth''); and NIRI.
    \46\ See, e.g., letters from ABA (stating that registrants will 
still incur significant costs even with the ability to select a 
methodology) and Financial Services Roundtable (Dec. 2, 2013) 
(``FSR'').
    \47\ See letter from Public Citizen (Jul. 6, 2015) (``Public 
Citizen II'').
---------------------------------------------------------------------------

    A number of commenters were critical of the proposed rule or 
particular aspects of it, as discussed in greater detail below. Some 
commenters stated specifically that they opposed the proposed rule or 
Section 953(b)'s requirement that we adopt any pay ratio rule. Some of 
these commenters asserted that the rule would not provide shareholders 
with material information.\48\ Other commenters noted that the pay 
ratio disclosure would not allow for meaningful comparisons among 
registrants.\49\ One commenter asserted that Section 953(b) and the 
proposed rule violate the First Amendment.\50\
---------------------------------------------------------------------------

    \48\ See, e.g., letters from American Benefits Council (Jan. 9, 
2014) (``American Benefits Council''); Mark Appleby (Oct. 10, 2013) 
(``Appleby''); Avery Dennison; Sean Bearly (Nov. 7, 2013) 
(``Bearly''); Joe Beltran (Nov. 21, 2013) (``Beltran''); Renato 
Berzolla (Nov. 21, 2013) (``Berzolla''); Former Assistant Secretary 
Campbell; Jonnie Dodge (Nov. 7, 2013) (``Dodge''); FSR; Hyster-Yale; 
IBC (supporting Congressional efforts to repeal of Section 953(b)); 
Jim Meyer (Oct. 18, 2013) (``Meyer''); NACCO; National Association 
of Manufacturers (Dec. 2, 2013) (``NAM I'') (supporting 
Congressional efforts to repeal of Section 953(b)); NAM II (same); 
National Retail Federation (Nov. 26, 2013) (``NRF'') (expressing 
``concern'' with the Proposed Rule); Elaine St. Miller (Nov. 21, 
2013) (``St. Miller''); Towers Watson (Dec. 2, 2013) (``Towers 
Watson'') (discussing ``reservations'' about a pay ratio rule); and 
WorldatWork I.
    \49\ See, e.g., letters from ABA (stating that ``the required 
disclosure will have little utility to investors other than to 
enable them to see the ratio of principal executive to employee 
compensation for a specific registrant from year to year'') and COEC 
I.
    \50\ See letter from COEC I. We do not believe that the pay 
ratio disclosure that Congress has mandated is inconsistent with the 
First Amendment. We believe that, in passing Section 953(b), 
Congress determined that the disclosure advances an important 
government interest, and we have carefully tailored the disclosure 
through this rulemaking to further that interest. Moreover, 
consistent with Congress's apparent purpose, commenters have stated 
that the pay ratio disclosure would be useful to shareholders in 
making say-on-pay votes. Accordingly, we believe the disclosure fits 
comfortably within the class of securities law disclosures that have 
been deemed to be consistent with the First Amendment. See Dun & 
Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 758, n.5 
(1985) (citing Ohralik v. Ohio State Bar Ass'n, 436 U.S. 447, 456 
(1978)). See also SEC v. Wall Street Publ'g Inst., 851 F.2d 365, 373 
(D.C. Cir. 1988).

---------------------------------------------------------------------------

[[Page 50110]]

    A few commenters stated that we should not adopt a final rule until 
we demonstrate that the rule is consistent with our mission and fully 
explain the benefits and costs of the rule.\51\ In this regard, one of 
these commenters criticized us for not making a statement about our 
precise goals or objectives for the rule, especially when Congress 
failed to hold hearings on Section 953(b).\52\ The commenter also 
stated that, without this statement and further explanations as to why 
we rejected less costly options, commenters cannot be fully informed 
and provide constructive comments. Several commenters argued that the 
proposed rule would be very costly to implement though many of these 
did not provide specific cost estimates.\53\ A majority of these 
commenters indicated that navigating their payroll systems and creating 
a single database of all their employees' compensation would be the 
most costly aspect of the proposed rule--especially with respect to 
non-U.S. employees.\54\ Commenters also mentioned other activities that 
would contribute to the costs, including data privacy compliance, 
foreign exchange calculations, data testing, establishing corporate 
guidelines, obtaining legal services, auditing results, public 
relations tasks, and litigation risk.\55\ As discussed below, some 
commenters provided specific cost and burden estimates about the 
proposed rule to demonstrate generally that it would impose high costs 
and burdens on registrants.\56\ In addition, some commenters argued 
that the pay ratio disclosure would impose a burden on competition or 
would cause competitive disadvantages for particular types of 
companies.\57\ Our analysis of the costs of the final rule, as well as 
an assessment of its impact on competition, is contained in the 
Economic Analysis section below.
---------------------------------------------------------------------------

    \51\ See, e.g., letters from Chamber I and Technical 
Compensation Advisors (Dec. 2, 2013) (``TCA.'').
    \52\ See letter from Chamber I.
    \53\ See, e.g., letters from Aon Hewitt; Business Roundtable I; 
Chesapeake Utilities Corporation (Dec. 17, 2013) (``Chesapeake 
Utilities''); Garmin, Ltd. (Nov. 11, 2013) (``Garmin''); IBC; KBR, 
Inc. (Nov. 26, 2013) (``KBR''); McGuireWoods LLP (Nov. 7, 2013) 
(``McGuireWoods''); National Association of Corporate Dealers (Dec. 
1, 2013) (``NACD''); NAM I; NAM II; NIRI; PNC Financial Services; 
and Semtech.
    \54\ See, e.g., letters from Aon Hewitt: Avery Dennison; 
Business Roundtable I; Chamber I; COEC I; Corporate Secretaries; 
Eaton (Dec. 2, 2013) (``Eaton''); FEI Company (Oct. 16, 2013) 
(``FEI''); FuelCell Energy, Inc. (Nov. 8, 2013) (``FuelCell 
Energy''); IBC; KBR; NACCO; NAM I; NAM II; and NIRI.
    \55\ See, e.g., letters from Aon Hewitt, Corporate Secretaries, 
and McGuireWoods.
    \56\ See, e.g., letters from ASA; Avery Dennison; COEC I; COEC 
II; Corporate Secretaries; Dover Corporation (Nov. 26, 2013) 
(``Dover Corp.''); Eaton; Exxon Mobil Corporation (Dec. 2, 2013) 
(``ExxonMobil''); FEI; FuelCell Energy; General Mills, Inc. (Dec. 2, 
2013) (``General Mills''); Hyster-Yale; Intel; NACCO; NRF; and U.S. 
Chamber of Commerce (May 22, 2014) (``Chamber II'').
    \57\ See, e.g., letters from ABA; Prof. Angel; Former Assistant 
Secretary Campbell; Chamber I; COEC I; Corporate Secretaries; IBC; 
NAM I; NAM II; and Korok Ray, Assistant Professor of Accounting, The 
George Washington University (Dec. 2, 2013) (``Prof. Ray'').
---------------------------------------------------------------------------

    A number of commenters recommended that we take additional 
preliminary steps before adopting a final rule. Some commenters 
requested that we extend the proposed rule's comment period.\58\ 
Another commenter suggested that we re-solicit comments after 
publishing the concerns it expressed.\59\ A few commenters advocated 
that we involve stakeholders in the rulemaking process by holding a 
roundtable, engaging in negotiated rulemaking, and/or conducting pilot 
programs.\60\ One of these commenters also recommended submitting the 
proposed rule to the Office of Information and Regulatory Affairs for 
an enhanced regulatory review.\61\ A few commenters suggested that we 
defer adopting a final rule under Section 953(b) until we adopt other 
rules required under the Dodd-Frank Act, particularly the pay-versus-
performance rule mandated by Section 953(a).\62\
---------------------------------------------------------------------------

    \58\ See, e.g., letters from American Insurance Association, et 
al. (Oct. 9, 2013) (``AIA et al.''); Bill Barrett Corp.; COEC I; 
IBC; NIRI; and WorldatWork I.
    \59\ See letter from Chamber I.
    \60\ See, e.g., letters from Chamber I, Chamber II, IBC, and 
NIRI.
    \61\ See letter from Chamber II.
    \62\ See, e.g., letters from Chamber I and WorldatWork I.
---------------------------------------------------------------------------

    As discussed above, members of the public interested in making 
their views known were invited to submit comment letters in advance of 
the official comment period for the proposed rule. In addition, we have 
continued to review and consider all comment letters submitted during 
and after the end of the comment period. Also, as discussed further in 
the Economic Analysis section below, we have considered and analyzed 
the numerous comments received regarding the costs and complexities of 
the mandated disclosure and have taken them into account in the final 
rule. Finally, we added to this rulemaking's public comment file 
additional analyses by the Commission's Division of Economic and Risk 
Analysis staff on the potential effects of excluding different 
percentages of employees from the pay ratio calculation, and commenters 
were expressly invited to comment on the analyses.
    This robust and public debate has informed us in developing our 
final rule. Overall, we believe interested parties have had sufficient 
opportunity to review the proposed rule, as well as the comment letters 
submitted, and to provide views on the proposal and on the other 
comment letters and data to inform our consideration of the final rule. 
Accordingly, we do not believe it is necessary to solicit additional 
public input before adopting the final rule.

D. Summary of Changes in the Final Rule

    The final rule we are adopting generally is consistent with the 
proposed rule. After considering all of the comments received on the 
proposal, however, and in particular, after considering specific 
suggestions from commenters on alternatives that could help to mitigate 
compliance costs and practical difficulties associated with the 
proposed rule, we are adopting a number of revisions in the final rule. 
We believe these revisions generally will preserve Congress's intent to 
require the disclosure of information that reflects the ratio of the 
PEO's compensation to the median employee's compensation while helping 
to minimize the expected costs and unintended consequences of the 
required disclosure. We summarize some of these changes here and 
discuss them in greater detail in Section II, below.
1. Non-U.S. Employee Exemptions and Additional Permitted Disclosure
    We proposed that an ``employee'' would include any U.S. and non-
U.S. employee of a registrant. We acknowledged in the Proposing Release 
that the inclusion of non-U.S. employees would raise compliance costs 
for multinational companies, would introduce cross-border compliance 
issues, could raise additional comparability concerns, and could have 
an adverse impact on competition. We indicated, however, that the 
inclusion of non-U.S. employees in the calculation of the median is 
consistent with the ``all employees'' language of the statute.

[[Page 50111]]

    The final rule defines the term ``employee'' to include U.S. 
employees and employees located in a jurisdiction outside the United 
States (``non-U.S. employees'') of a registrant, as proposed. We 
continue to believe that this is most consistent with the statutory 
language of Section 953(b) and with the purpose of providing a company-
specific metric that shareholders can use to evaluate a registrant's 
executive compensation. Including both U.S. and non-U.S. employees will 
result in pay ratio disclosure that reflects the actual composition of 
the registrant's workforce. Even assuming the statutory language could 
be viewed as ambiguous on this issue, we also believe that this 
approach is most consistent with the general nature of our disclosure 
regime, which does not limit registrants' disclosure obligations only 
to factors, events, or circumstances that exist in or take place within 
the United States. For example, a registrant must disclose the PEO's 
compensation whether or not the PEO actually works within the United 
States.
    To help address concerns about compliance costs, and consistent 
with the commenters' suggestions, the final rule provides two tailored 
exemptions from the definition of ``employee,'' which otherwise 
includes all of a registrant's U.S. and non-U.S. employees in the 
median employee determination. First, the final rule provides an 
exemption for circumstances in which foreign data privacy laws or 
regulations make registrants unable to comply with the final rule. 
Second, the final rule permits registrants to exempt non-U.S. employees 
where these employees account for 5% or less of the registrant's total 
U.S. and non-U.S. employees, with certain limitations.
    The Proposing Release acknowledged that data privacy laws or 
regulations in various foreign jurisdictions could affect a 
registrant's ability to gather the necessary data to identify its 
median employee. We did not propose any accommodation to address this 
concern, however, because we believed the flexibility of the proposed 
rule would permit registrants to manage any potential costs arising 
from these laws. In response to significant concerns expressed by a 
number of commenters over cross-border compliance issues that may arise 
from the pay ratio disclosure requirement, and consistent with 
commenters' suggestions, the final rule permits registrants to exclude 
from their determination of the median employee an employee who is 
employed in a foreign jurisdiction in which the laws or regulations 
governing data privacy are such that, despite its reasonable efforts to 
obtain or process the information necessary for compliance with the 
final rule, the registrant is unable to do so without violating such 
data privacy laws or regulations.
    The registrant's reasonable efforts must include using or seeking 
an exemption or other relief under any governing data privacy laws or 
regulations. If a registrant excludes any non-U.S. employees in a 
particular jurisdiction under this exemption, it must exclude all non-
U.S. employees in that jurisdiction, list the excluded jurisdictions, 
identify the specific data privacy law or regulation, explain how 
complying with the final rule violates such data privacy law or 
regulation (including the efforts made by the registrant to use or seek 
an exemption or other relief under such law or regulation), and provide 
the approximate number of employees exempted from each jurisdiction 
based on this exemption. In addition, the registrant must obtain a 
legal opinion from counsel that opines on the inability of the 
registrant to obtain or process the information necessary for 
compliance with the final rule without violating that jurisdiction's 
data privacy laws or regulations, including the registrant's inability 
to obtain an exemption or other relief under any governing laws or 
provisions.
    In addition to the data privacy exemption for non-U.S. employees, 
the final rule includes a de minimis exemption for non-U.S. employees. 
Under the final rule, if a registrant's non-U.S. employees account for 
5% or less of its total employees, it may exclude all of those 
employees when making its pay ratio calculations. In this circumstance, 
however, if the registrant chooses to exclude any non-U.S. employees, 
it must exclude all of them. If a registrant's non-U.S. employees 
exceed 5% of the registrant's total U.S. and non-U.S. employees, it may 
exclude up to 5% of its total employees who are non-U.S. employees. If 
a registrant excludes any non-U.S. employees in a particular 
jurisdiction, it must exclude all non-U.S. employees in that 
jurisdiction. The registrant must also disclose the jurisdictions from 
which its non-U.S. employees are being excluded, the approximate number 
of employees excluded from each jurisdiction under the de minimis 
exemption, the total number of its U.S. and non-U.S. employees 
irrespective of any exemption (de minimis or data privacy), and the 
total number of its U.S. and non-U.S. employees used for its de minimis 
calculation.
    In calculating the number of non-U.S. employees that may be 
excluded under the de minimis exemption, a registrant must count any 
non-U.S. employee exempted under the data privacy exemption against the 
availability. A registrant may exclude any non-U.S. employee that meets 
the data privacy exemption, even if the number of excluded employees 
exceeds 5% of the registrant's total employees. If, however, the number 
of employees excluded under the data privacy exemption equals or 
exceeds 5% of the registrant's total employees, the de minimis 
exemption will not be available. Additionally, if the number of 
employees excluded under the data privacy exemption is less than 5% of 
the registrant's total employees, the registrant may use the de minimis 
exemption to exclude no more than the number of non-U.S. employees 
that, combined with the data privacy exemption, equals 5% of the 
registrant's total employees.
    Finally, the final rule permits registrants to make cost-of-living 
adjustments for the compensation of employees in jurisdictions other 
than the jurisdiction in which the PEO resides to identify the median 
and calculate annual total compensation. In identifying the median 
employee, whether using annual total compensation or any other 
compensation measure that is consistently applied to all employees 
included in the calculation, the registrant may, but is not required 
to, make cost-of-living adjustments for the compensation of employees 
in jurisdictions other than the jurisdiction in which the PEO resides 
so that the compensation is adjusted to the cost of living in the 
jurisdiction in which the PEO resides. If the registrant uses a cost-
of-living adjustment to identify the median employee, and the median 
employee identified is an employee who does not reside in the same 
jurisdiction as the PEO, the registrant must use the same cost-of-
living adjustment in calculating the median employee's annual total 
compensation and disclose the country in which the median employee is 
located. The registrant is also required to briefly describe the cost-
of-living adjustments it used to identify the median employee and 
briefly describe the cost-of-living adjustments it used to calculate 
the median employee's annual total compensation, including the measure 
used as the basis for the cost-of-living

[[Page 50112]]

adjustment.\63\ To provide context for the Item 402(u)(1)(iii) 
disclosure, a registrant electing to present the pay ratio in this 
manner must also disclose the median employee's annual total 
compensation and pay ratio without the cost-of-living adjustments. To 
calculate this pay ratio, the registrant will need to identify the 
median employee without using any cost-of-living adjustments.
---------------------------------------------------------------------------

    \63\ For example, registrants may use cost-of-living adjustments 
based on purchasing power parity (``PPP'') conversion factors. A PPP 
conversion factor is the ratio of PPP exchange rate to the nominal 
exchange rate. For example, conversion factors for the US dollar are 
available at http://data.worldbank.org/indicator/PA.NUS.PPPC.RF. 
This ratio provides the number of units of a country's currency 
required to buy the same amount of goods and services in the 
domestic market as a U.S. dollar would buy in the United States.
---------------------------------------------------------------------------

    In the Proposing Release, we noted that some comments received 
prior to the proposal requested that the rule allow registrants to 
present separate pay ratios covering U.S. and non-U.S. employees to 
mitigate concerns that the comparison of the PEO to non-U.S. employees 
could substantially affect the pay ratio disclosure. The proposal did 
not prohibit such disclosure but did not expressly state it was 
permitted. For clarification, therefore, the final rule states that 
registrants are permitted, but not required, to provide additional pay 
ratios as long as any additional pay ratios are not misleading and are 
not presented with greater prominence than the required ratio.
 2. Employees of Consolidated Subsidiaries
    We proposed requiring a registrant's pay ratio disclosure to 
include the employees of any of its subsidiaries (including officers 
other than the PEO), in addition to its direct employees, in its pay 
ratio disclosure. Unlike the proposed rule, however, the final rule 
defines ``employee'' to include only the employees of the registrant's 
consolidated subsidiaries. As discussed in greater detail below, 
defining a ``subsidiary'' based on whether a registrant consolidates a 
company in its financial statements will likely decrease the costs and 
burdens on a registrant without significantly affecting the pay ratio 
because most registrants consolidate based on their ownership of over 
50% of the outstanding voting shares of their subsidiaries and guidance 
is readily available on when consolidation is appropriate.
3. Employed on Any Date Within Three Months of the Last Completed 
Fiscal Year
    The proposed rule would define ``employee'' as an individual 
employed as of the last day of the registrant's last completed fiscal 
year because this calculation date would be consistent with the one 
used for the determination of the three most highly compensated 
executive officers under existing Item 402(a)(3)(iii). In the Proposing 
Release, we also noted our preliminary view that a bright line 
calculation date for determining who is an employee would ease 
compliance for registrants by eliminating the need to monitor changes 
in workforce composition during the year. Further, we assumed the 
potential benefits of the pay ratio disclosure would not be 
significantly diminished by covering only individuals employed at year-
end, although we acknowledged that this approach could be costlier to 
registrants with seasonal or temporary employees who are employed at 
year end as opposed to other times during the year.
    Taking into consideration concerns raised by commenters about the 
desire for flexibility in choosing the calculation date, the final rule 
permits registrants to use any date within three months prior to the 
last day of their last completed fiscal year to identify the median 
employee. If in subsequent years the registrant changes the date it 
uses to identify the median employee, it must disclose this change and 
provide a brief explanation about the reason or reasons for the change. 
This provision provides consistency for individual registrants from 
year to year while also providing registrants with flexibility to 
choose the determination date. To provide additional transparency about 
how the pay ratio disclosure has been calculated, the final rule 
requires registrants to disclose the date used to identify the median 
employee.
4. Identifying the Median Employee Once Every Three Years
    The proposed rule would require registrants to identify the median 
employee every year. To help minimize compliance costs, we are revising 
the rule, as suggested by commenters, to allow registrants to identify 
the median employee every three years unless there has been a change in 
its employee population or employee compensation arrangements that the 
registrant reasonably believes would result in a significant change in 
the pay ratio disclosure. However, the registrant must still calculate 
the identified median employee's annual total compensation and use that 
figure in calculating its pay ratio every year. If there have been no 
changes that the registrant reasonably believes would significantly 
affect its pay ratio disclosure, the registrant must disclose that it 
is using the same median employee in its pay ratio calculation and 
describe briefly the basis for its reasonable belief. For example, the 
registrant could disclose that there has been no change in its employee 
population or employee compensation arrangements that it believes would 
significantly impact the pay ratio disclosure. If there has been such a 
change, the registrant must re-identify the median employee for that 
fiscal year.
    Under the final rule's approach, the registrant will identify its 
median employee for year one and then be permitted to use that employee 
or one who is similarly compensated (if, for example, the median 
employee is no longer in the same position or is no longer employed by 
the registrant) in the following two years for calculating the median 
employee's annual total compensation and the registrant's pay ratio. 
The registrant must calculate the median employee's annual total 
compensation in year one and then re-calculate the annual total 
compensation for that employee in year two and again in year three. If 
the median employee identified in year one is no longer in the same 
position or no longer employed by the registrant on the median employee 
determination date in year two or three, the final rule permits the 
registrant to replace its median employee with an employee in a 
similarly compensated position.
5. Initial Compliance Date
    We proposed that a registrant's first reporting period would begin 
in its first fiscal year commencing on or after the effective date of 
the final rule. Therefore, under the proposed rule, the registrant's 
initial pay ratio disclosure would be included in its first annual 
report on Form 10-K or proxy or information statement for its annual 
meeting of shareholders following the end of such year. Unlike the 
proposal, the compliance date set forth in this adopting release 
provides that the registrant's first reporting period for the pay ratio 
disclosure is its first full fiscal year beginning on or after January 
1, 2017 (instead of on or after the effective date of the final rule).
6. Transition Period for New Registrants
    The proposed rule would not have required pay ratio disclosure by 
new registrants subject to the rule in a registration statement on Form 
S-1 or

[[Page 50113]]

Form S-11 for an initial public offering or registration statement on 
Form 10. Consistent with the revised transition period for existing 
registrants, the final rule provides that the first pay ratio reporting 
period begins for new registrants with their first fiscal year 
commencing on or after January 1, 2017 that is after the date they 
first become subject to the requirements of Section 13(a) or 15(d) of 
the Exchange Act. In this way, new registrants will not become subject 
to the final rule sooner than existing registrants. Such registrants 
are also permitted to omit their initial pay ratio disclosure from 
their registration statements (or any other filing) made before their 
first annual report or proxy or information statement following the end 
of that reporting period, but not later than 120 days after the end of 
the fiscal year.
7. Additional Transition Periods
    We did not propose a transition period for registrants that cease 
to be smaller reporting companies or emerging growth companies, nor did 
we provide any special rules for registrants that engage in business 
combinations and/or acquisitions. We did, however, request comment on 
whether there should be transition periods in these situations and, if 
so, the appropriate length of time for any such transition period. One 
commenter requested that we include such a transition period.\64\
---------------------------------------------------------------------------

    \64\ See letter from ABA (recommending a transition period for 
issuers that cease to be smaller reporting companies in which such 
registrants would not be required to provide their pay ratio 
disclosure until the first full fiscal year commencing on or after 
the first anniversary of the end of the fiscal year in which the 
issuer is no longer a smaller reporting company).
---------------------------------------------------------------------------

    The final rule provides that registrants that cease to be smaller 
reporting companies or emerging growth companies are not required to 
provide pay ratio disclosure until they file a report for the first 
fiscal year commencing on or after they cease to be a smaller reporting 
company or emerging growth company. The final rule also permits 
registrants that engage in business combinations and/or acquisitions to 
omit the employees of a newly-acquired entity from their pay ratio 
calculation for the fiscal year in which the business combination or 
acquisition occurs. In these cases, a registrant does not have to 
include these individual employees in its median employee calculation 
until the first full fiscal year following the acquisition. Registrants 
that exclude employees as a result of a business combination must 
disclose the relevant acquired business and the approximate number of 
employees that are excluded from the pay ratio calculation.

II. Discussion

A. Scope of Final Rule

1. Pay Ratio Disclosure Requirements Under New Paragraph (u) to Item 
402 of Regulation S-K
a. Proposed Rule
    We proposed new paragraph (u) of Item 402 to require disclosure of: 
(A) the median of the annual total compensation of all employees of the 
registrant, (except the registrant's PEO); (B) the annual total 
compensation of the registrant's PEO; and (C) the ratio of the amount 
in (A) to the amount in (B), presented as a ratio in which the amount 
in (A) equaled one, or, alternatively, expressed narratively in terms 
of the multiple that the amount in (B) bears to the amount in (A).
    Although Section 953(b) calls for a ratio showing the median of the 
annual total compensation of all employees to the PEO's annual total 
compensation, it does not specify how the ratio should be expressed. To 
promote consistent presentation and address the potential for 
confusion, therefore, the proposed rule specified that registrants must 
express the ratio as one in which the median of the annual total 
compensation of all employees is equal to one.
b. Comments on the Proposed Rule
    No commenters objected to use of the term PEO or including the pay 
ratio disclosure requirements in new paragraph (u) to Item 402, and 
commenters discussing other aspects of the proposal did so on the 
assumption that we would include the pay ratio disclosure requirements 
in Item 402(u). Several commenters agreed that the pay ratio should 
show the PEO's compensation divided by the median employee's 
compensation because it would be easier to understand.\65\
---------------------------------------------------------------------------

    \65\ See, e.g., letters from FEI and Dr. Sue Ravenscroft, 
Professor of Accounting, Iowa State University (Sep. 20, 2013) 
(``Prof. Ravenscroft I'').
---------------------------------------------------------------------------

c. Final Rule
    After considering the comments, we are adopting the final rule as 
proposed. The final rule adds new paragraph (u) to Item 402 and 
requires disclosure of: (A) The median of the annual total compensation 
of all employees of the registrant (except the registrant's PEO); (B) 
the annual total compensation of the registrant's PEO; and (C) the 
ratio of the amount in (B) to the amount in (A), presented as a ratio 
in which the amount in (A) equals one, or, alternatively, expressed 
narratively in terms of the multiple that the amount in (B) bears to 
the amount in (A).\66\
---------------------------------------------------------------------------

    \66\ We did not propose to require that the pay ratio disclosure 
be provided in interactive data format, and are not adopting such a 
requirement for this disclosure. To the extent that we consider more 
generally the tagging of disclosures in XBRL format in our rules, we 
may consider revisiting the format in which the pay ratio disclosure 
is provided.
---------------------------------------------------------------------------

    Consistent with the proposal, the final rule also requires 
registrants to disclose the ratio such that the PEO's annual total 
compensation is always compared to the median employee's annual total 
compensation. Registrants may not present the median employee's annual 
total compensation as a percentage of the PEO's compensation. We 
believe expressing the ratio as ``a factor rather than a fraction'' 
makes the ratio easier to understand because allowing the inverse may 
be confusing.\67\ In other words, the ratio must always show how much 
larger or smaller the PEO's annual total compensation is as compared to 
the median employee's annual total compensation. We believe that 
requiring registrants to present the ratio in this manner will make it 
easier for shareholders to comprehend and allow them to use it in 
making voting decisions on executive compensation.\68\
---------------------------------------------------------------------------

    \67\ See letter from FEI. See also letter from Prof. Ravenscroft 
I (stating that ``the ratio of CEO compensation to median worker 
would be easier for most people to grasp than the ratio of median 
worker compensation to CEO compensation, a switch that I would see 
as not changing the intent of the law'').
    \68\ In the rare cases in which the PEO's yearly compensation is 
nominal (or is otherwise less than the median employee's 
compensation), the resulting ratio will be a number smaller than 
one. Despite this anomalous result, we believe that in the vast 
majority of cases setting the median compensation equal to one will 
result in a ratio that is easier to understand than the inverse. See 
generally Illinois Commerce Comm'n v. I.C.C., 776 F.2d 355, 359 
(D.C. Cir. 1985) (``General rules need not work perfectly in all 
their applications''). We also note that registrants are permitted 
to provide additional narrative discussion in cases where they feel 
the disclosed pay ratio may be confusing or incomplete without 
further explanation.
---------------------------------------------------------------------------

    The final rule permits registrants to choose one of two options to 
express the ratio. Registrants may disclose the pay ratio with the 
median of the annual total compensation of all employees equal to one 
and the PEO's compensation as the number compared to one.\69\ For

[[Page 50114]]

example, if a registrant's median annual total compensation for 
employees is $50,000 and the annual total compensation of the PEO is 
$2,500,000, the PEO's compensation is 50 times larger than the median 
employee's compensation. The registrant may describe the pay ratio as 
50 to 1 or 50:1. Alternatively, registrants may disclose the pay ratio 
narratively by stating how many times higher (or lower) the PEO's 
annual total compensation is than that of the median employee. For 
example, the registrant may state that ``the PEO's annual total 
compensation is 50 times that of the median of the annual total 
compensation of all employees.''
---------------------------------------------------------------------------

    \69\ We note that some commenters recommended that the final 
rule include a safe harbor or simplified reporting method such that 
a registrant may stipulate that its pay ratio exceeds 300-to-1. See 
letters from Hyster-Yale and NACCO. Under this suggestion, 
registrants would be permitted to forgo calculating and disclosing 
their company-specific pay ratio and would instead be permitted to 
simply disclose that the ratio exceeds the stipulated 300-to-1 
statistic. We have not adopted the suggestion to allow a registrant 
to disclose that its pay ratio exceeds a stipulated statistic, such 
as 300-to-1, because we do not believe that it would be consistent 
with Congress's apparent intent to provide a useful, relevant, 
company-specific pay ratio disclosure for investors to utilize when 
undertaking their say-on-pay votes.
---------------------------------------------------------------------------

2. Pay Ratio Disclosure in Filings That Require Item 402 of Regulation 
S-K Information
a. Proposed Rule
    The proposed rule required registrants to include their pay ratio 
disclosure in any filing described in Item 10(a) that requires 
executive compensation disclosure under Item 402, including annual 
reports on Form 10-K,\70\ Securities Act and Exchange Act registration 
statements, and proxy and information statements, to the same extent 
that these forms require compliance with Item 402. Section 953(b) does 
not direct us to amend any of our forms to add the pay ratio disclosure 
requirements to filings that do not already require disclosure of Item 
402 information, and we did not propose to do so. Additionally, we 
proposed not to require registrants to update their pay ratio 
disclosure for the most recently completed fiscal year until they file 
their annual reports on Form 10-K, or, if later, their proxy or 
information statements for their next annual meeting of shareholders 
(or written consents in lieu of such a meeting) but, in any event, not 
later than 120 days after the end of such fiscal year.
---------------------------------------------------------------------------

    \70\ 17 CFR 249.310.
---------------------------------------------------------------------------

b. Comments on the Proposed Rule
    All of the commenters discussing the issue agreed that we should 
limit pay ratio disclosure to the filings described in Item 10(a) that 
require executive compensation disclosure under Item 402, as 
proposed.\71\ One commenter suggested, however, that the final rule 
should allow registrants to include their pay ratio disclosure in other 
filings if they choose to do so.\72\
---------------------------------------------------------------------------

    \71\ See, e.g., letters from ABA, AFL-CIO I, California Public 
Employees' Retirement System (Dec. 2, 2013) (``CalPERS''), Calvert, 
Capital Strategies, CII, Connecticut State Treasurer (Dec. 2, 2013) 
(``CT State Treasurer''), Davis Polk, Intel, Johnson & Johnson, 
McGuireWoods, NIRI, NRF, Pax World Funds, PM&P, Prof. Ray, and 
WorldatWork I.
    \72\ See letter from Capital Strategies.
---------------------------------------------------------------------------

c. Final Rule
    We are adopting the final rule as proposed. It requires registrants 
to include their pay ratio disclosure in any filing described in Item 
10(a) that calls for executive compensation disclosure under Item 402, 
including annual reports on Form 10-K, registration statements under 
the Securities Act and Exchange Act, and proxy and information 
statements to the same extent that these forms require compliance with 
Item 402, consistent with the statutory directive. Registrants must 
follow the instructions in each form to determine whether Item 402 
information is required, including any instructions that allow for the 
omission of Item 402 information.\73\ The final rule does not require 
registrants to include the pay ratio disclosure in any filings that are 
not required to include Item 402 information, but registrants can 
voluntarily include non-mandated information in any of their filings if 
they choose to do so and the information is not misleading in the 
context of that filing. Further, registrants do not need to update 
their pay ratio disclosure for their most recently completed fiscal 
year until they file their annual report on Form 10-K, or, if later, 
their proxy or information statement for their next annual meeting of 
shareholders (or written consents in lieu of such a meeting) but, in 
any event, not later than 120 days after the end of such fiscal year.
---------------------------------------------------------------------------

    \73\ See, e.g., General Instructions I(2)(c) and J(1)(m) to Form 
10-K containing special provisions for the omission of Item 402 
information by wholly-owned subsidiaries and asset-backed 
registrants.
---------------------------------------------------------------------------

    We do not read Section 953(b) to require pay ratio disclosure in 
filings that do not contain other executive compensation information. 
In our view, the most meaningful way to present pay ratio disclosure is 
in context with other executive compensation disclosure, such as the 
Summary Compensation Table required by Item 402(c) and the Compensation 
Discussion and Analysis required by Item 402(b), rather than provided 
on a stand-alone basis. In this manner, the pay ratio information will 
be presented in the same context as other information that shareholders 
can use in making their voting decisions on executive compensation. 
Finally, although we understand the primary purpose of the pay ratio 
disclosure to be to inform shareholder's say-on-pay votes under Section 
951, we acknowledge that some commenters indicated the disclosure could 
be useful to investors in making investment decisions. For that reason, 
and in light of the statutory language of Section 953(b), the final 
rule retains the requirement to include this disclosure in registration 
statements under the Securities Act.\74\
---------------------------------------------------------------------------

    \74\ Although the final rule requires registrants to include the 
pay ratio disclosure in registration statements under the Securities 
Act, as discussed below, the final rule permits new registrants to 
delay compliance so that the pay ratio requirement is not required 
in a registration statement on Form S-1 or Form S-11 for an initial 
public offering or registration statement on Form 10.
---------------------------------------------------------------------------

3. Excluded Registrants--Smaller Reporting Companies, Foreign Private 
Issuers, MJDS Filers, and Emerging Growth Companies
a. Proposed Rule
    In the Proposing Release, we noted that the reference to ``each 
issuer'' in Section 953(b) could be read to apply to all registrants, 
including smaller reporting companies,\75\ foreign private issuers,\76\ 
U.S.-Canadian Multijurisdictional Disclosure System filers,\77\ and 
emerging growth companies. Because Section 953(b) refers specifically 
to the definition of total compensation in Item 402(c)(2)(x), and is 
silent on whether pay ratio disclosure should be required for 
registrants not previously subject to Item 402(c) requirements, 
however, we proposed to limit the pay ratio disclosure requirement to 
registrants required to provide Item 402(c) disclosure. As a result, 
the proposed rule stated that smaller reporting companies, foreign 
private issuers, and MJDS filers did not have to provide pay ratio 
disclosure in any of their filings.
---------------------------------------------------------------------------

    \75\ A ``smaller reporting company'' is an issuer that had a 
public float of less than $75 million as of the last business day of 
its most recently completed second fiscal quarter or had annual 
revenues of less than $50 million during the most recently completed 
fiscal year for which audited financial statements are available. 17 
CFR 229.10(f)(1).
    \76\ A ``foreign private issuer'' is any foreign issuer other 
than a foreign government, except for a registrant that, as of the 
last business day of its most recent fiscal year, has more than 50% 
of its outstanding voting securities held of record by United States 
residents and any of the following: a majority of its officers and 
directors are citizens or residents of the United States, more than 
50% of its assets are located in the United States, or its business 
is principally administered in the United States. 17 CFR 240.3b-
4(c).
    \77\ A U.S.-Canadian Multijurisdictional Disclosure System 
(``MJDS'') filer is a registrant that files reports and registration 
statements with us in accordance with the requirements of the MJDS.
---------------------------------------------------------------------------

    Also, the JOBS Act,\78\ which was passed by Congress subsequent to 
the

[[Page 50115]]

Dodd-Frank Act but prior to publication of the Proposing Release, 
specifically excluded registrants that qualify as emerging growth 
companies, as that term is defined in Section 3(a) of the Exchange 
Act,\79\ from the requirements of Section 953(b). To give effect to the 
statutory exemption, we proposed an instruction to Item 402(u) 
providing that a registrant that is an emerging growth company is not 
required to comply with Item 402(u).
---------------------------------------------------------------------------

    \78\ See Section 102(a)(3) of the JOBS Act.
    \79\ An ``emerging growth company'' is an issuer that had total 
annual gross revenues of less than $1 billion during its most 
recently completed fiscal year, has not reached the fifth 
anniversary of the date of the first sale of its common equity 
securities pursuant to an effective registration statement under the 
Securities Act, had not issued $1 billion in non-convertible debt 
during the previous 3-year period, or is deemed to be a ``large 
accelerated filer.'' 15 U.S.C. 78c(a).
---------------------------------------------------------------------------

b. Comments on the Proposed Rule
    Most commenters concurred with the proposed rule's exclusion of 
emerging growth companies, consistent with the JOBS Act.\80\ One 
commenter noted specifically that, while it ``believes all registrants 
accessing U.S. capital markets should be subject to comparable 
financial regulation,'' including smaller reporting companies, foreign 
private issuers, and MJDS filers, the pay ratio information ``is best 
viewed in the context of other compensation disclosures and the pay 
ratio disclosure should be limited to those registrants required to 
provide a summary compensation disclosure.'' \81\
---------------------------------------------------------------------------

    \80\ See, e.g., letters from CalPERS; Davis Polk; Hay Group, 
Inc. (Dec. 2, 2013) (``Hay Group''); NY State Comptroller; PM&P; and 
US SIF. Some commenters, however, disagreed or were uncomfortable 
with this exclusion. See letters from CII and Andrew Kushner (Nov. 
12, 2013) (``Kushner'').
    \81\ See letter from CalPERS.
---------------------------------------------------------------------------

    Additionally, most of the commenters who addressed the issue agreed 
that we should exclude smaller reporting companies from the pay ratio 
requirements.\82\ One commenter reasoned that, by excluding emerging 
growth companies, Congress demonstrated its intent to relieve this 
category of registrants from the costs and burdens of compliance, and 
because both emerging growth companies and smaller reporting companies 
are subject to scaled executive compensation disclosure, it would be 
consistent with Congressional intent to exclude smaller reporting 
companies.\83\ Another commenter recommended that the final rule exempt 
any business with revenues of less than $500 million or a market 
capitalization of less than $1 billion.\84\
---------------------------------------------------------------------------

    \82\ See, e.g., letters from ABA, Prof. Angel, CalPERS, Capital 
Strategies, Davis Polk, Hay Group, NIRI, NY State Comptroller, PM&P, 
Vivient, and WorldatWork I.
    \83\ See letter from ABA.
    \84\ See letter from Hay Group.
---------------------------------------------------------------------------

    By contrast, a few commenters asserted that the final rule should 
include smaller reporting companies.\85\ One of these commenters 
asserted that Congress did not expressly exclude smaller reporting 
companies because the phrase ``each issuer'' in Section 953(b)(1) 
signals its intent that there should be no exemption for any particular 
registrant and that excluding certain registrants from the disclosure 
would defeat the purpose and policy of Section 953(b).\86\
---------------------------------------------------------------------------

    \85\ See, e.g., letters from CII; Ashley Ray, University of 
Idaho College of Law (Nov. 13, 2013) (``Ray''); and US SIF.
    \86\ See letter from Ray.
---------------------------------------------------------------------------

    Finally, some commenters agreed that the proposed rule should 
exclude foreign private issuers and MJDS filers,\87\ while a few other 
commenters disagreed with excluding them.\88\ One commenter noted that, 
in ``view of the Commission's long-standing rules allowing foreign 
private issuers and MJDS filers to provide information about their 
executive compensation programs based on the applicable disclosure 
requirements of their home jurisdiction, we would find it anomalous to 
single out this specific Item 402-based disclosure requirement for 
mandatory application to these registrants without regard to important 
policy considerations that have led the Commission for decades to 
permit disclosure in this area based on home-country law.'' \89\
---------------------------------------------------------------------------

    \87\ See, e.g., letters from ABA, CalPERS, Capital Strategies, 
Davis Polk, Hay Group, and PM&P.
    \88\ See, e.g., letters from CII, Ray, and US SIF.
    \89\ See letter from ABA.
---------------------------------------------------------------------------

c. Final Rule
    After considering the comments, we are adopting the final rule as 
proposed. The final rule, therefore, does not require pay ratio 
disclosure by smaller reporting companies, foreign private issuers, 
MJDS filers, and emerging growth companies.\90\
---------------------------------------------------------------------------

    \90\ Registered investment companies will also not be required 
to provide Item 402(u) disclosure. Business development companies 
are a category of closed-end investment company that are not 
registered under the Investment Company Act [15 U.S.C. 80a-2(a)(48) 
and 80a-53-64]. Business development companies will be treated in 
the same manner as issuers other than registered investment 
companies and therefore will be subject to the pay ratio disclosure 
requirement.
---------------------------------------------------------------------------

    As stated above, Congress explicitly excluded emerging growth 
companies from the pay ratio disclosure requirement. Regarding smaller 
reporting companies, Section 953(b)(2) requires total compensation to 
be calculated in accordance with Item 402(c)(2)(x). Smaller reporting 
companies, however, are permitted to follow the scaled disclosure 
requirements set forth in Items 402(m)-(r),\91\ and therefore are not 
required to calculate compensation in accordance with Item 
402(c)(2)(x). Also, the requirement set forth in Item 402(n) for 
disclosure of Summary Compensation Table information, which includes 
disclosure of ``total compensation,'' does not require smaller 
reporting companies to include the same types of compensation required 
to be included in total compensation for other registrants under Item 
402(c)(2).\92\ Congress's express reference to Item 402(c)(2)(x) to 
calculate total compensation (without mentioning Item 402(n)(2)(x)) is 
consistent with the exclusion of smaller reporting companies from the 
pay ratio disclosure requirement.
---------------------------------------------------------------------------

    \91\ See Item 402(l).
    \92\ See Item 402(n)(2)(viii) (indicating that smaller reporting 
companies are not required to include the aggregate change in the 
actuarial present value of pension benefits that is required for 
companies subject to Item 402(c)(2)(viii)).
---------------------------------------------------------------------------

    Requiring smaller reporting companies to provide the pay ratio 
disclosure would compel them to collect data and calculate compensation 
for the PEO in ways that they otherwise are not required to do. Nothing 
in the statute indicates that was Congress's intent, and no commenters 
indicated that they believed there was such an intent. To clarify 
further that smaller reporting companies are excluded from the final 
rule, we are making a technical amendment to paragraph (l) of Item 402 
to add Item 402(u), as proposed, to the list of items that are not 
required for smaller reporting companies.\93\
---------------------------------------------------------------------------

    \93\ Smaller reporting companies are permitted to choose whether 
they want to comply with either the scaled disclosure requirements 
or the larger company disclosure requirements on an ``a la carte'' 
basis. As we discussed in the scaled disclosure adopting release, 
the staff evaluates compliance by smaller reporting companies with 
only the Regulation S-K requirements applicable to smaller reporting 
companies, even if the company chooses to comply with the larger 
company requirements. See Smaller Reporting Company Regulatory 
Relief and Simplification, Release No. 33-8876 (Dec. 19, 2007) [73 
FR 934], at 941.
---------------------------------------------------------------------------

    The final rule similarly does not apply to foreign private issuers 
and MJDS filers, which we believe is consistent with excluding 
registrants that are not currently required to provide Summary 
Compensation Table disclosure pursuant to Item 402(c). Foreign private 
issuers file annual reports and registration statements on Form 20-F 
\94\ and MJDS filers file annual reports and registration

[[Page 50116]]

statements on Form 40-F.\95\ Neither of these forms requires Item 402 
disclosure. As with smaller reporting companies, requiring foreign 
private issuers and MJDS filers to provide the pay ratio disclosure 
would require these registrants to collect data and calculate 
compensation for the PEO in ways they otherwise would not be required 
to do. The final rule, therefore, does not apply to foreign private 
issuers or MJDS filers.
---------------------------------------------------------------------------

    \94\ 17 CFR 249.220f.
    \95\ 17 CFR 249.240f.
---------------------------------------------------------------------------

    Finally, for the same reasons, the final rule, consistent with the 
proposal, does not change existing Item 402(a)(1) with respect to 
foreign private issuers. Item 402(a)(1) states that a ``foreign private 
issuer will be deemed to comply with Item 402 if it provides the 
information required by Items 6.B and 6.E.2 of Form 20-F, with more 
detailed information provided if otherwise made publicly available or 
required to be disclosed by the registrant's home jurisdiction or a 
market in which its securities are listed or traded.'' Foreign private 
issuers that file annual reports on Form 10-K, therefore, are still 
able to satisfy Item 402 requirements by following Items 6.B and 6.E.2 
of Form 20-F and are not required to provide the pay ratio disclosure 
mandated by Section 953(b).

B. Requirements of Final Rule

1. ``All Employees'' Covered Under the Rule
    The final rule defines ``employee'' to include a registrant's U.S. 
and non-U.S. employees, as well as its part-time, seasonal, and 
temporary employees, as proposed. We believe that the ``all employees 
of the issuer'' language in Section 953(b) is best implemented by 
including rather than excluding broad categories of employees. Further, 
even assuming there was any ambiguity in the statutory language, we 
believe that a more inclusive approach better serves Section 953(b)'s 
purpose of providing shareholders with additional information about a 
registrant's compensation practices that can be used in making voting 
decisions on executive compensation because it results in a pay ratio 
that is more reflective of the actual composition of the registrant's 
workforce.\96\ As discussed in greater detail below, however, in 
response to particular issues and concerns raised by comments, we have 
provided two tailored exemptions from the general requirement to 
include all employees. In particular, for the reasons discussed below, 
we have provided an exemption for employees in foreign jurisdictions in 
which it is not possible for a registrant to obtain or process 
information necessary to comply with the rule without violating the 
data privacy laws or regulations of that jurisdiction and a de minimis 
exemption for non-U.S. employees.
---------------------------------------------------------------------------

    \96\ As discussed below, commenters have made competing 
arguments that the pay ratio disclosure would be ``distorted'' in 
one way or another if the final rule included (or excluded) various 
categories of employees. See, e.g., letter from Prof. Ray (asserting 
that the pay ratio will be distorted depending on whether the final 
rule includes non-U.S. employees). We appreciate that a particular 
registrant's pay ratio information may vary--perhaps, in some cases, 
significantly--depending on whether the final rule includes or 
excludes these employee categories (e.g., non-U.S. employees; part-
time, temporary and seasonal employees; and leased workers). We do 
not, however, view the choices made on these issues as involving a 
``distortion'' of the pay ratio disclosure that Congress directed 
should be provided to investors. As noted elsewhere in this release, 
we have been guided by our general view that Section 953(b) reflects 
Congress's intention that the pay ratio disclosure should be broadly 
inclusive of all types of a registrant's employees; thus, absent a 
reason that takes into account the statutory objective, we have 
declined to make choices in the final rule that would exclude broad 
categories of employees from the process of identifying the median 
employee. At the same time, however, in an effort to mitigate the 
potential costs and burdens of the final rule, we have built in some 
flexibility and provided several other accommodations to elements of 
the final rule, where we have concluded that these measures would 
not result in any undue impact on the required pay ratio disclosure.
---------------------------------------------------------------------------

a. Types of Employees
i. Proposed Rule
    The proposed rule included in the definition of ``employee'' all of 
a registrant's full-time, part-time, seasonal, and temporary workers, 
including officers other than the PEO. In the Proposing Release, we 
reasoned that these individuals should be included in the rule because 
Section 953(b)(1)(A) expressly requires disclosure of the median of the 
annual total compensation of ``all employees,'' which would encompass 
full-time, part-time, seasonal, and temporary workers. Also, we 
proposed to include all of a registrant's officers other than the PEO 
in the definition of ``employee.''
    Workers not employed by a registrant (or its subsidiaries), 
however, such as independent contractors, ``leased'' workers, or other 
workers who are employed by a third party, were not covered by our 
proposed definition of ``employee.'' As an example, we noted that, if a 
registrant pays a fee to another company (such as a management company 
or an employee leasing agency) that supplies workers to the registrant, 
and those workers receive compensation from that other company, these 
workers should not be considered employees of the registrant for 
purposes of the disclosures required by Section 953(b) of the Dodd-
Frank Act.
ii. Comments on the Proposed Rule
    A number of commenters supported the proposed rule's requirement 
that registrants include their part-time, temporary, and seasonal 
employees in addition to their full-time employees in their median 
employee determination.\97\ Some commenters asserted that the reference 
in Section 953(b) to ``all employees'' demonstrates Congress's intent 
was not to limit the pay ratio to only full-time employees.\98\ Some 
commenters contended that including temporary employees would cost 
registrants very little because they routinely develop that information 
for their own internal use.\99\ Commenters supporting the proposed rule 
also contended that, if part-time, temporary, and seasonal employees 
were excluded from the pay ratio, the disclosure would be incomplete, 
inaccurate, and/or misleading.\100\ One commenter suggested that the 
exclusion of part-time, seasonal, and temporary employees would not 
reduce the regulatory burdens on registrants because registrants with 
such employees would have substantial flexibility in identifying the 
median employee.\101\
---------------------------------------------------------------------------

    \97\ See, e.g., letters from AFSCME, CalPERS, Calvert, Chicago 
Teachers Fund, Cummings Foundation, CUPE, Domini, Susan A. Estep 
(Nov. 15, 2013) (``Estep''), Frank Gould (Oct. 4, 2013) (``Gould''), 
ICCR, IL Bricklayers and Craftworkers Union, Marco Consulting, 
McMorgan & Co., Novara Tesija, NY State Comptroller, Oxfam, Pax 
World Funds, Sen. Robert Menendez et al. (Nov. 26, 2013) (``Sen. 
Menendez et al. I''), Sen. Menendez et al. II, Socially Responsive 
Financial Advisors, Teamsters, Trillium I, Trustee Campbell, US SIF, 
and Walden.
    \98\ See letters from AFL-CIO (Jul. 6, 2015) (``AFL-CIO II''), 
Institute for Policy Studies (Oct. 30, 2013) (``IPS''), and Sen. 
Menendez et al. II.
    \99\ See, e.g., letters from Virginia Fischer (Oct. 3, 2013) 
(``Fischer'') and Public Citizen II.
    \100\ See, e.g., letters from AFL-CIO I, AFL-CIO II, Americans 
for Financial Reform (Dec. 2, 2013) (``AFR''), B[acirc]tirente et 
al., Bricklayers International, CII, CT State Treasurer, FS FTQ, 
Public Citizen I, Public Citizen II, and John Theodore (Nov. 5, 
2013) (``Theodore'').
    \101\ See letter from AFL-CIO II.
---------------------------------------------------------------------------

    Other commenters contended that the final rule should include only 
full-time employees.\102\ Many of these commenters claimed that 
applying the rule to part-time, temporary, and seasonal employees would 
make the pay ratio disclosure less meaningful because the compensation 
of these different types of employees are not comparable to each other 
or to the PEO's compensation.\103\ Some commenters

[[Page 50117]]

also asserted that including employees other than full-time employees 
would be burdensome and increase costs.\104\ One commenter noted that, 
according to a survey it conducted of companies with more than 10,000 
employees, ``86 percent of the average employer's employees are full-
time, with the median employer having 95 percent of its workforce as 
full-time employees.'' This commenter asserted that the incremental 
information that would be obtained from including part-time or seasonal 
employees does not justify the effort to collect it and could provide a 
distorted picture of the employee's annual income.\105\
---------------------------------------------------------------------------

    \102\ See, e.g., letters from Best Buy et al., Brian Foley & 
Company, Inc. (Dec. 2, 2013) (``Brian Foley & Co.''), Chamber I, 
FuelCell Energy, Garmin, Hyster-Yale, Mercer I, NACCO, PM&P, 
Semtech, Steven Hall and Partners (Dec. 2, 2013) (``SH&P''), Vectren 
Corp., WorldatWork I, and WorldatWork II.
    \103\ See, e.g., letters from American Apparel & Footwear 
Association (Dec. 2, 2013) (``AAFA I''), American Apparel & Footwear 
Association (Dec. 23, 2013) (``AAFA II''), American Benefits 
Council, COEC I, Corporate Secretaries, Davis Polk, Dover Corp., 
ExxonMobil, FSR, General Mills, Hay Group, IBC, KBR, Meridian, NACD, 
NIRI, NRF, NYC Bar, Retail Industry Leaders Association (Dec. 2, 
2013) (``RILA''), and WorldatWork I.
    \104\ See, e.g., letters from American Benefits Council, COEC I, 
Corporate Secretaries, Davis Polk, and General Mills.
    \105\ See letter from COEC I.
---------------------------------------------------------------------------

    Another commenter asserted that the final rule should exclude part-
time, seasonal, and temporary employees unless a majority of a 
registrant's employees work on a part-time, temporary, and/or seasonal 
basis.\106\ A few commenters recommended that the rule should allow 
registrants to exclude any employee who was not employed for at least 
four months during the calendar year.\107\ Other commenters indicated 
that, if the rule is not limited to full-time employees, registrants 
should be able to annualize or make full-time equivalent adjustments to 
the compensation of part-time, seasonal, and/or temporary 
employees.\108\ One commenter suggested, in the alternative, that if 
the rule includes part-time, seasonal, and temporary employees, it 
should only include such employees who have been employed during the 
90-day period ending on the last day of the registrant's fiscal 
year.\109\ Finally, another commenter suggested, in the alternative, 
that the rule should require a ``primary disclosure'' that compares 
only full-time employees to the PEO's compensation.\110\
---------------------------------------------------------------------------

    \106\ See letter from ABA.
    \107\ See, e.g., letters from Hyster-Yale and NACCO (stating 
that such a provision could contain a consistency requirement to 
prevent companies from selectively choosing which employees to 
include or exclude).
    \108\ See, e.g., letters from AAFA II, American Benefits 
Council, Brian Foley & Co., Corporate Secretaries, Hay Group, KBR, 
NIRI, and NYC Bar.
    \109\ See letter from Meridian.
    \110\ See letter from NRF.
---------------------------------------------------------------------------

    Some commenters noted that neither Section 953(b) nor its 
legislative history states explicitly that Congress intended for the 
``all employees'' term to include part-time, seasonal, and temporary 
employees. These commenters contended we could thus interpret the ``all 
employees'' language to exclude such employees from the final 
rule.\111\ Some commenters believed that the minimal effect these 
employees would have on the pay ratio would not justify the high costs 
required to include those employees in determining the pay ratio.\112\ 
Conversely, commenters in support of including part-time, seasonal, and 
temporary employees in the pay ratio contended that failing to do so 
would distort registrants' pay ratios because many of their employees 
would not be included in the median calculation.\113\
---------------------------------------------------------------------------

    \111\ See, e.g., letter from ABA.
    \112\ See, e.g., letters from COEC I (stating that two-thirds of 
respondents to a survey it conducted indicated that limiting the 
application of the rule to full-time employees would reduce their 
costs, that the ``average savings for these respondents would be 
approximately 20 percent,'' and that the burden imposed by including 
``global, full-time, part-time and seasonal employees'' is not 
offset by other benefits) and General Mills (``We would expect 
moderate cost savings from limiting the analysis to full-time 
employees, versus covering our entire workforce, but the savings 
could be significant for registrants in other industries. . . 
Conversely, there has been little or no evidence to suggest that the 
benefits of the Proposed Rule would be diminished as a result of 
limiting its scope to full-time employees.'').
    \113\ See, e.g., letters from AFL-CIO I, AFR, B[acirc]tirente et 
al., Bricklayers International, CII, CT State Treasurer, FS FTQ, 
Public Citizen I, and Theodore.
---------------------------------------------------------------------------

    Several commenters agreed that a registrant's pay ratio should 
exclude ``leased'' workers.\114\ One of these commenters noted that 
including ``leased'' workers would add significant costs and distort 
the pay ratio.\115\ Also, according to this commenter, such workers are 
not ``statutory'' employees and the third parties employing these 
workers may be unwilling to provide the information. A few commenters 
recommended that the final rule require a registrant to include such 
workers in its pay ratio.\116\ One commenter asserted that a registrant 
should be required to clearly describe its reliance on ``leased'' 
workers if they comprise more than 40% of its workforce.\117\
---------------------------------------------------------------------------

    \114\ See, e.g., letters from ABA; American Staffing Association 
(Nov. 21, 2013) (``ASA''); CalPERS; CII; Corporate Secretaries; 
Emergent BioSolutions, Inc. (Nov. 27, 2013) (``Emergent''); 
ExxonMobil; Hyster-Yale; Intel; McGuireWoods, NACCO; and WorldatWork 
I.
    \115\ See letter from ABA.
    \116\ See, e.g., letters from Demos (Nov. 22, 2013) (``Demos 
I''), Fischer (referring to ``independent contractors''), and 
Vectren Corp. (arguing that the rule should include full-time, U.S.-
based contractors if they make up a significant portion of the 
registrant's workforce).
    \117\ See letter from LAPFF.
---------------------------------------------------------------------------

iii. Final Rule
    After considering the public comments, we have concluded that the 
final rule's definition of ``employee'' should include the full-time, 
part-time, seasonal, and temporary employees employed by the registrant 
or any of its consolidated subsidiaries. Because this definition refers 
to workers ``employed by the registrant,'' workers who provide services 
to the registrant or its consolidated subsidiaries as independent 
contractors or ``leased'' workers are excluded from the definition as 
long as they are employed, and their compensation is determined, by an 
unaffiliated third party. The final rule includes in the definition of 
``employee'' all of a registrant's officers other than the PEO, as 
proposed. Section 953(b)(1)(A) expressly directs disclosure of the 
median of the annual total compensation of ``all employees of the 
issuer, except the chief executive officer (or any other equivalent 
position) of the issuer.''
    We believe this statutory language indicates that Congress intended 
the final rule to include all types of a registrant's employees, 
including part-time, seasonal, and temporary workers, and we do not 
think it is appropriate to provide a wholesale exemption for those 
broad categories of employees that are not employed full-time.\118\ Any 
such exemption would risk producing pay ratio disclosure that is 
significantly different than the pay ratio disclosure that Congress 
expressly directed us to require when it said ``all employees.'' 
Further, as noted above, we have generally limited our use of 
discretionary or exemptive authority to those items that would not have 
an appreciable effect on the information that Congress intended that 
shareholders have when they make their say-on-pay votes. To the extent 
there is any statutory ambiguity, we would still elect this inclusive 
approach because we believe that it is more reflective of the actual 
composition of the registrant's workforce and thus furthers the purpose 
of providing shareholders with useful information about a registrant's 
overall compensation practices. While we are sensitive to

[[Page 50118]]

concerns raised by some commenters that inclusion of these broad 
categories of employees means that compliance with the final rule will 
be more costly than if we adopted a broad exemption, we note that the 
final rule provides other types of flexibility and accommodations 
designed to reduce compliance costs while remaining faithful to our 
understanding of the statutory directive and purpose of Section 953(b).
---------------------------------------------------------------------------

    \118\ See, e.g., County of Oakland v. Federal Housing Finance 
Agency, 716 F.3d 935, 940 (6th Cir. 2013) (explaining that ``a 
straightforward reading of the statute leads to the unremarkable 
conclusion that when Congress said `all taxation,' it meant all 
taxation'') (emphasis in original); Marie O. v. Edgar, 131 F.3d 610, 
620 (7th Cir. 1997) (interpreting the phrase ``all infants and 
toddlers'' and explaining that ``[a]ll is unambiguous; it means 
every eligible child''); cf. GEICO v. Fetisoff, 958 F.2d 1137, 1142 
(D.C. Cir. 1992) (describing ``the word `all''' as ``one of the 
least ambiguous words in the English language'').
---------------------------------------------------------------------------

    A registrant can supplement its pay ratio disclosure or provide 
additional pay ratios for its shareholders to consider if it wants to 
explain the effect of including part-time, seasonal and temporary 
employees on its pay ratio disclosure. While we do not believe a 
purpose of the rule is to facilitate comparisons among registrants, the 
opportunity to supplement the pay ratio disclosure and to provide 
additional pay ratios should help mitigate some concerns that 
shareholders may draw unwarranted conclusions from comparing one 
registrant's disclosed ratio to the ratio of others. In addition, our 
change to the proposed rule to allow a registrant some flexibility in 
selecting the date for identifying the median employee may enable 
registrants that employ temporary or seasonal employees only during a 
very limited period at the end of their fiscal year to choose a date 
that allows them to exclude these employees.
    One commenter pointed out that Item 402 does not contain a 
reference to hourly or overtime compensation and contended, therefore, 
that the rule should not apply to non-salaried employees who receive 
``wages plus overtime,'' rather than salary.\119\ We believe that the 
``all employees'' language is not limited to salaried employees. 
Moreover, we are concerned that a contrary reading would be arbitrary 
and would eliminate an entire category of employees from the pay ratio 
disclosure, potentially depriving shareholders of a more complete 
understanding of the median employee's compensation when making their 
say-on-pay votes. Thus, we believe that it is appropriate to include 
these employees as part of a registrant's pay ratio disclosure to 
reflect the manner in which the registrant establishes its workforce.
---------------------------------------------------------------------------

    \119\ See letter from Chamber I.
---------------------------------------------------------------------------

    The final rule excludes from the definition of ``employee'' those 
workers who are employed, and whose compensation is determined, by an 
unaffiliated third party but who provide services to the registrant or 
its consolidated subsidiaries as independent contractors or ``leased'' 
workers. Although it is unclear whether Congress intended to include 
these workers as ``employees of the issuer,'' or even considered the 
issue, we believe, as a matter of policy, these workers should not be 
included as ``employees of the issuer.''
    While one commenter stated that ``leased employees and other 
workers employed by a third party are not ``statutory'' employees of a 
registrant,'' \120\ some definitions of ``employee'' may include 
workers who are not employed directly by the registrant or its 
consolidated subsidiaries, such as independent contractors or 
``leased'' or ``borrowed'' workers, if they are employed by a third 
party.\121\ We note that the statute specifies employees ``of the 
issuer,'' and in light of this, to the extent there is any ambiguity on 
this point, we believe that the better reading of the statute is to 
exclude from the final rule's definition of ``employee'' workers who 
are not employed by the registrant or its consolidated subsidiaries, 
such as independent contractors or ``leased'' workers, if they are 
employed, and their compensation is determined, by an unaffiliated 
third party.
---------------------------------------------------------------------------

    \120\ See, e.g., letter from ABA.
    \121\ See, e.g., Black's Law Dictionary 602 (9th ed. 2009) 
(defining ``employee'' as a ``person who works in the service of 
another person (the employer) under an express or implied contract 
of hire, under which the employer has the right to control the 
details of work performance,'' and defines a ``borrowed employee'' 
as an ``employee whose services are, with the employee's consent, 
lent to another employer who temporarily assumes control over the 
employee's work'').
---------------------------------------------------------------------------

    We believe excluding such workers is appropriate because 
registrants generally do not control the level of compensation that 
these workers are paid. Instead, the registrant provides a payment for 
their services to an unaffiliated third party, which determines the 
compensation for the employees. As one commenter noted, there can be no 
assurance, therefore, that the registrant even has access to the 
workers' compensation information, which could make it difficult or 
impossible to obtain the information.\122\
---------------------------------------------------------------------------

    \122\ See letter from ABA.
---------------------------------------------------------------------------

    We do not believe it is appropriate for registrants to voluntarily 
include workers employed by third parties in their required pay ratio 
disclosure ``if such persons make up a significant portion of the 
workforce,'' as one commenter suggested, even if doing so may add to 
the ``flexibility'' of the final rule.\123\ For the reasons described 
above, we have not included these workers within the definition of 
employee, and we are concerned that allowing registrants the option to 
elect to include them in their required disclosures would introduce the 
potential for registrants to manipulate the pay ratio disclosure. 
Registrants, however, may discuss their reliance on ``leased'' workers, 
as suggested by another commenter, in their narrative disclosure.\124\ 
Also, they may provide additional ratios that factor in those workers, 
as long as any additional ratios are not misleading and are not more 
prominently displayed than the required ratio.
---------------------------------------------------------------------------

    \123\ See letter from Vectren Corp.
    \124\ See letter from LAPFF.
---------------------------------------------------------------------------

b. Employed on Any Date Within Three Months of the Last Completed 
Fiscal Year
i. Proposed Rule
    The proposed rule would define ``employee'' as an individual 
employed as of the last day of the registrant's last completed fiscal 
year. We proposed this calculation date for determining who is an 
employee because it is consistent with the one used for the 
determination of PEO and principal financial officer and the other 
three most highly compensated executive officers under Item 
402(a)(3)(iii). In the Proposing Release, we noted that the composition 
of a company's workforce typically changes throughout the fiscal year, 
and in some industries and businesses, it can change constantly. 
Although Section 953(b) requires the median calculation to cover ``all 
employees,'' it does not prescribe a particular calculation date for 
the determination of who should be treated as an employee for that 
purpose.
    We reasoned in the Proposing Release that a single date for 
determining who is an employee would ease compliance for registrants by 
eliminating the need to monitor changing workforce composition during 
the year, while providing a recent snapshot of the registrant's entire 
workforce. Also, we indicated that a requirement to track which 
employees have been continuously employed for the entire annual period 
could increase costs for registrants, and suggested that the most 
appropriate calculation date would be one that is consistent with the 
calculation date for determining the named executive officers under 
current Item 402 requirements.
    In proposing this approach, we assumed that the potential benefits 
of the disclosure mandated by Section 953(b) would not be significantly 
diminished by covering only individuals employed on a specific date

[[Page 50119]]

at calendar year-end, rather than covering every individual who was 
employed at any time during the year. Although this approach could help 
limit compliance costs for registrants, we acknowledged that it could 
have other costs. For example, this approach would not capture seasonal 
or temporary employees who are not employed at year end, with the 
result that a registrant with a significant number of such workers 
might identify a median employee from a pool that does not fully 
reflect the workforce that it requires to run its business. This 
approach might also cause the disclosure to be costlier for, and 
thereby have an anti-competitive impact on, registrants whose temporary 
or seasonal workers are employed at calendar year-end as opposed to 
other times during the year because registrants with temporary or 
seasonal employees at calendar year-end would have to include them in 
their median calculations but other registrants with temporary or 
seasonal employees at other times of the year would not have to do so. 
Finally, we noted that it would be possible, but unlikely, that 
registrants could try to structure their employment arrangements to 
reduce the number of lower paid employees employed on the determination 
date.
ii. Comments on the Proposed Rule
    Most commenters that discussed the issue agreed that registrants 
should be permitted to identify the median employee based on the 
composition of their workforce on a particular day of the year as 
opposed to the workforce employed throughout the year.\125\ Only a few 
commenters, however, supported using the last day of the fiscal year 
calculation date.\126\ It seems that these commenters supported this 
provision more to limit the calculation to a particular day of the 
year, thereby limiting the need to monitor a changing workforce during 
the year, than because they believed the appropriate date should be the 
last day of the registrant's last fiscal year.\127\ A number of 
commenters contended that the final rule should allow registrants the 
flexibility to choose a calculation date within the registrant's last 
fiscal year.\128\ As some of these commenters noted, requiring 
registrants to use the last day of the fiscal year could adversely 
affect retailers,\129\ may not allow enough time for registrants to 
collect and report on their pay ratio information,\130\ and could make 
it difficult for registrants that are not calendar year-end companies 
to use information derived from its tax and/or payroll records to 
calculate the ratio.\131\ Some commenters suggested that, if the final 
rule permits registrants to choose a determination date other than a 
registrant's fiscal year end, it should also require registrants to be 
consistent from year to year and/or briefly explain the reasons for not 
using the fiscal year-end date.\132\
---------------------------------------------------------------------------

    \125\ See letters from ABA, Capital Strategies, FEI, Hyster-
Yale, Johnson & Johnson, and NACCO.
    \126\ See, e.g., letters from CII and Vectren.
    \127\ Id.
    \128\ See, e.g., letters from ABA, Best Buy et al., Capital 
Strategies, COEC I, COEC II, Corporate Secretaries, Davis Polk, FEI, 
Hyster-Yale, Johnson & Johnson, Microsoft, NACCO, NAM I, PM&P, RILA, 
SH&P, and WorldatWork I. See also letter from Brianne H. McCoy, 
University of Idaho College of Law (Nov. 28, 2013) (``McCoy'') 
(stating that a registrant should be required to choose a date that 
``corresponds with the month in which the registrant had its highest 
gross operating revenues from the previous year'').
    \129\ See letter from PM&P.
    \130\ See, e.g., letters from COEC I, Corporate Secretaries, 
Davis Polk, Microsoft, and NAM I.
    \131\ See, e.g., letters COEC I, Microsoft, NAM I, and NAM II.
    \132\ See, e.g., letters from Capital Strategies, Microsoft, and 
SH&P.
---------------------------------------------------------------------------

    One commenter suggested that the rule could allow registrants to 
choose a calculation date within a designated time window, ``such as a 
date occurring during the 90-day period preceding the fiscal year 
end.'' \133\ Regarding the concern that allowing registrants to select 
the calculation date would create the potential for manipulation of the 
pay ratio, this commenter stated that the concern ``is unwarranted, 
particularly if the choice is restricted to a limited time period (such 
as the last fiscal quarter), since in general the employee population 
of a registrant would not vary significantly over such a period.'' 
\134\
---------------------------------------------------------------------------

    \133\ See letter from Davis Polk.
    \134\ Id.
---------------------------------------------------------------------------

    Another commenter proposed that registrants should be required to 
calculate the median annual compensation of all employees employed at 
any time over the preceding 365 days to ensure accurate disclosure for 
registrants that employ a high number of seasonal employees.\135\
---------------------------------------------------------------------------

    \135\ See letter from AFL-CIO II.
---------------------------------------------------------------------------

    One commenter recommended that the final rule permit registrants to 
use different determination dates for different segments of their 
workforce based on tax, payroll, and/or other established recordkeeping 
systems, accompanied by a brief statement of the basis for the 
different disclosure dates because a number of companies maintain their 
human resource/payroll systems for U.S. employees on a calendar-year 
basis, but do so for their foreign employees on a fiscal-year 
basis.\136\ The commenter also noted that using the end of the second 
or third fiscal quarter as a determination date would not be feasible 
because most payroll systems are set up to collect information on 
fiscal year-end or calendar year-end bases.
---------------------------------------------------------------------------

    \136\ See letter from ABA.
---------------------------------------------------------------------------

    Some commenters responded to the Proposing Release's request for 
comment on whether the rule's definition of ``employee'' would cause a 
registrant to change its corporate structure. Most of the commenters 
that responded said that the definition would not cause registrants to 
alter their corporate structure or employment arrangements,\137\ 
although one commenter disagreed.\138\
---------------------------------------------------------------------------

    \137\ See, e.g., letters from ABA, Capital Strategies, and 
WorldatWork I.
    \138\ See letter from Prof. Ray.
---------------------------------------------------------------------------

iii. Final Rule
    After considering commenters' desire for flexibility in choosing 
the median employee determination date, we are revising the final rule 
from the proposal. Unlike the proposed rule, which would define 
``employee'' as an individual employed as of the last day of the 
registrant's last completed fiscal year, the final rule defines 
``employee'' as an individual employed on any date of the registrant's 
choosing within the last three months of the registrant's last 
completed fiscal year. The final rule also requires registrants to 
disclose the date used to identify the median employee.
    Some commenters recommended that the final rule require registrants 
to explain the reason they selected a determination date other than the 
last day of the fiscal year.\139\ We do not believe such an explanation 
would lead to useful disclosure, as registrants would likely state that 
they chose any date other than the end of the fiscal year to provide 
more time to take the steps necessary to identify the median employee. 
If, however, a registrant changes the determination date from the prior 
year, we believe it should disclose the reason for the change. Under 
the final rule, therefore, if a registrant changes the date it uses to 
identify the median employee, the registrant must disclose the change 
and provide a brief explanation about the reason or reasons for the 
change.
---------------------------------------------------------------------------

    \139\ See, e.g., letters from Microsoft and SH&P.
---------------------------------------------------------------------------

    We note that allowing registrants to choose a determination date 
within a defined window, rather than be required to use the last day of 
the fiscal year, is a change from the proposal and differs from the 
approach in determining the

[[Page 50120]]

three most highly compensated executive officers under existing Item 
402(a)(3)(iii).\140\ We believe permitting registrants to choose a date 
within the last three months of their last completed fiscal year, as 
suggested by a commenter, is appropriate because it will provide 
registrants with some flexibility and could permit them additional time 
to identify their median employee in advance of their fiscal year 
end.\141\ At the same time, establishing a particular date certain will 
provide some consistency from year to year. We also note that this 
change may help to avoid some of the unintended consequences identified 
by commenters, such as anti-competitive effects on retailers with a 
significant number of employees at year end or inefficient changes in 
corporate structure made simply to avoid employing workers on the last 
day of the fiscal year. Finally, as discussed in the Proposing Release, 
we continue to believe that requiring the determination of the employee 
to be made as of a specific date, rather than over the course of the 
year, will ease compliance for registrants by eliminating the need to 
monitor changes in their workforce composition throughout the year.
---------------------------------------------------------------------------

    \140\ 17 CFR 230.402(a)(3)(iii).
    \141\ See letter from Davis Polk (noting that a ``90-day period 
preceding the fiscal year end'' would permit ``a registrant [to] 
begin the task of identifying its median employee in advance of its 
fiscal year end, which is the most costly and time-consuming part of 
the pay ratio calculation'').
---------------------------------------------------------------------------

c. Employees Located Outside the United States
i. Proposed Rule
    We proposed a definition of ``employee'' that would include any 
U.S. and non-U.S. employee of a registrant. In the Proposing Release, 
we acknowledged that the inclusion of non-U.S. employees raises 
compliance costs for multinational companies, introduces cross-border 
compliance issues, and could raise concerns about the impact of non-
U.S. pay structures on the comparability of the data to companies 
without off-shore operations. We also recognized that differences in 
relative compliance costs could have an adverse impact on competition. 
We weighed these considerations and proposed that the disclosure 
requirements would nonetheless cover all employees without exemptions 
for specific categories of employees, including non-U.S. employees.
    Additionally, we were cognizant that data privacy laws in various 
jurisdictions could have an impact on gathering and verifying the data 
needed to identify the median of the annual total compensation of all 
employees. Commenters in the pre-proposal period expressed concern 
that, in some cases, data privacy laws of foreign countries could 
prohibit a registrant's collection and transfer of personally 
identifiable compensation data that would be needed to identify the 
median employee. We also noted that some data privacy laws may make the 
collection or transfer of the underlying data more burdensome, but do 
not actually prohibit transfer of compensation data.
    For example, we indicated that multinational companies based in the 
United States might need to ensure compliance with data privacy 
regulations when taking certain actions to comply with the proposal, 
such as transmitting personally identifiable human resources data 
(``personal data'') of European Union (``E.U.'') employees onto global 
human resource information system networks in the United States; 
sending personal data in hard copy from the E.U. to the United States; 
or making personal data ``onward transfers'' to third-party payroll, 
pension, and benefits processors outside of the E.U.\142\ In some E.U. 
countries, employee consent is required, while in other countries 
consent may not be sufficient. We noted that other jurisdictions, such 
as Peru, Argentina, Canada, and Japan also have data privacy laws that 
could be implicated by the gathering of data for purposes of the 
proposed pay ratio disclosure.
---------------------------------------------------------------------------

    \142\ The E.U. Directive 95/46/EC, 1995 O.J. L 281 (European 
Union Directive on the Protection of Individuals with Regard to the 
Processing of Personal Data and on the Free Movement of Such Data) 
sets forth the regulatory framework governing the transfer of 
personal data from an E.U. Member State to a non-E.U. country.
---------------------------------------------------------------------------

    Although we did not propose any specific accommodation to address 
this concern, we stated our belief that the flexibility afforded to all 
registrants under the proposed rule could permit registrants to manage 
any potential costs arising from applicable data privacy laws. For 
example, the proposed rule would permit registrants in this situation 
to estimate the compensation of affected employees. We requested 
comment on whether the proposed flexibility afforded to registrants in 
selecting a method to identify the median, such as the use of 
statistical sampling or other reasonable estimation techniques and the 
use of consistently applied compensation measures to identify the 
median employee, could enable registrants to better manage any 
potential costs and burdens arising from local data privacy regulations 
or if there are other alternatives that would be consistent with 
Section 953(b).
ii. Comments on the Proposed Rule
    Many commenters agreed that non-U.S. employees should be included 
in a registrant's pay ratio disclosure.\143\ Some commenters contended 
that failing to include non-U.S. employees would cause the pay ratio 
disclosure to be incomplete, less informative, and misleading.\144\ 
Some commenters stressed that Congress intended to include non-U.S. 
employees because Section 953(b) refers specifically to ``all 
employees.'' \145\ One commenter suggested that the exclusion of non-
U.S. employees would not reduce the regulatory burdens on registrants 
because registrants with such employees would have substantial 
flexibility in identifying the median employee.\146\
---------------------------------------------------------------------------

    \143\ See, e.g., letters from AFL-CIO I, AFL-CIO II, AFSCME, 
Bricklayers International, CalPERS, Calvert, Chicago Teachers Fund, 
Corayer, CT State Treasurer, Cummings Foundation, CUPE, Estep, 
Fedewa, First Affirmative, Form Letter C, Gould, Douglas J. Matteson 
(Oct. 5, 2013) (``Matteson''), ICCR, IL Bricklayers and Craftworkers 
Union, Marco Consulting, McMorgan & Co., Novara Tesija, NY 
Bricklayers and Craftworkers Union, NY State Comptroller, Oxfam, Pax 
World Funds, Charmaine M. Phillips (Oct. 16, 2013) (``C. 
Phillips''), Public Citizen I, Public Citizen II, Douglas C. Rand 
(Sep. 26, 2013) (``Rand''), Sen. Menendez et al. I, Sen. Menendez et 
al. II, Socially Responsive Financial Advisors, Stephen Spofford 
(Sep. 24, 2013) (``S. Spofford''), Teamsters, Trillium I, Trustee 
Campbell, US SIF, and Walden.
    \144\ See, e.g., letters from AFR, B[acirc]tirente et al., CII, 
Domini, and FS FTQ.
    \145\ See, e.g., letters from AFL-CIO II, IPS, and Sen. Menendez 
et al. II.
    \146\ See letter from AFL-CIO II.
---------------------------------------------------------------------------

    Many commenters, however, disagreed with the proposed rule and 
contended that the final rule should include only the registrant's U.S. 
employees because including non-U.S. employees would be very costly 
and/or distort the pay ratio.\147\ A number of commenters asserted that 
the costs to registrants of including non-U.S. employees would outweigh 
any benefits of the disclosure to shareholders,\148\ and offered a 
variety of different estimates of how greatly the inclusion of non-U.S. 
employees would affect costs.\149\
---------------------------------------------------------------------------

    \147\ See, e.g., letters from Appleby, Avery Dennison, Chamber 
I, Chesapeake Utilities, Dover Corp., FuelCell Energy, Garmin, Hay 
Group, Meridian, NACD, PM&P, SH&P, Vectren Corp., Vivient, 
WorldatWork I, and WorldatWork II.
    \148\ See, e.g., letters from ABA, American Benefits Council, 
Business Roundtable I, Chamber I, Corporate Secretaries, ExxonMobil, 
Frederic W. Cook & Co., Inc. (Nov. 29, 2013) (``Frederic W. Cook & 
Co.''), RILA, Semtech, and Society for Human Resource Management 
(Dec. 30, 2013) (``SHRM'').
    \149\ See, e.g., letters from American Benefits Council (stating 
that costs could be 20 to 30 times higher if non-U.S. employees are 
included and also that some of its ``member companies have estimated 
that the annual cost to make the median employee determination on a 
worldwide basis could be hundreds of thousands of dollars''), 
Business Roundtable I (stating that costs could decrease by over 50% 
in some cases if non-U.S. employees are excluded), ExxonMobil, and 
FEI (stating that costs would decrease by 90% if the final rule 
excluded non-U.S. employees).

---------------------------------------------------------------------------

[[Page 50121]]

    Additionally, commenters noted that companies with international 
operations almost always have multiple payroll systems and databases 
for their employees' compensation that are difficult, if not 
impossible, to reconcile,\150\ with some of the commenters providing 
the following examples:
---------------------------------------------------------------------------

    \150\ See, e.g., letters from American Benefits Council, Aon 
Hewitt, BCIMC, Business Roundtable I, COEC I, COEC II, COEC III, 
Cummins Inc., Eaton, ExxonMobil, Freeport-McMoRan, MVC Associates 
International (Nov. 28, 2013) (``MVC Associates''), NIRI, Semtech, 
SHRM, and Tesoro Corporation (Nov. 21, 2013) (``Tesoro Corp'').
---------------------------------------------------------------------------

     One commenter indicated that it has 15 payroll systems 
that are not integrated, and those systems would have to be manually 
reconciled with ``substantial costs'' and ``extensive staff hours;'' 
\151\
---------------------------------------------------------------------------

    \151\ See letter from Freeport-McMoRan.
---------------------------------------------------------------------------

     one commenter stated that it has 30 payroll systems that 
do not interface; \152\
---------------------------------------------------------------------------

    \152\ See letter from Cummins Inc.
---------------------------------------------------------------------------

     one commenter cited a Human Resource Policy Association 
survey concluding that 84% of respondents could not easily calculate 
worldwide enterprise cash compensation for all their employees; \153\
---------------------------------------------------------------------------

    \153\ See letter from MVC Associates.
---------------------------------------------------------------------------

     one commenter cited its own survey finding that a 
registrant on average maintains 46 different payroll systems in 34 
different countries; \154\ and
---------------------------------------------------------------------------

    \154\ See letter from COEC I.
---------------------------------------------------------------------------

     one commenter stated that it does not have a single 
payroll system.\155\
---------------------------------------------------------------------------

    \155\ See letter from Tesoro Corp.
---------------------------------------------------------------------------

    Several commenters stated that many countries have data privacy and 
other laws that prevent registrants from transferring payroll data 
outside that country's borders (even if the transfer would be within 
the same company), which would make compiling the information necessary 
for the pay ratio problematic or even illegal.\156\ One of these 
commenters recommended that, if the final rule did not exclude all non-
U.S. employees, it should permit registrants to exclude from their 
methodology for identifying the median employee the data for any 
employees in a jurisdiction where such collection, analysis, and 
transmission would violate a registrant's existing data privacy 
obligations.\157\ A few commenters stated that the proposed rule's 
flexibility afforded to all registrants could permit registrants to 
manage any potential costs arising from applicable data privacy 
laws.\158\
---------------------------------------------------------------------------

    \156\ See, e.g., letters from AAFA II, ABA, American Benefits 
Council, BCIMC, Business Roundtable I, Chamber I, COEC I, COEC II, 
Corporate Secretaries, Cummins Inc., Eaton, ExxonMobil, Freeport-
McMoRan, Hyster-Yale, IBC, NACCO, NAM I, NAM II, NIRI, RILA, 
Semtech, SHRM, and WorldatWork I.
    \157\ See letter from ABA.
    \158\ See, e.g., letters from Capital Strategies (stating that 
data privacy laws would not affect registrants because U.S. firms 
must know their employee expenses and can use sampling techniques to 
negate any data privacy effects) and WorldatWork I (stating that, 
while data privacy laws will have a negative effect on some 
registrants, others may be able to gather the required information 
under existing waivers granted to them by the EU, and some 
registrants can estimate total compensation of their employees in 
countries with data privacy laws by placing employees in ``bands'' 
of similar compensation and benefits levels, and estimating total 
compensation using those bands).
---------------------------------------------------------------------------

    One commenter contended that any data privacy concerns can be 
addressed easily by anonymizing payroll data sets or conducting 
statistical sampling.\159\ If, however, these privacy safeguards proved 
insufficient, the commenter recommended that registrants be permitted 
to exclude employees in such countries only if they disclose the number 
of employees in the excluded countries and obtain and file as an 
exhibit to the periodic report in which the pay ratio disclosure 
appears a legal opinion by qualified outside counsel demonstrating that 
the privacy safeguards are inadequate under local law.
---------------------------------------------------------------------------

    \159\ See letter from AFL-CIO II.
---------------------------------------------------------------------------

    Additionally, a number of commenters contended that including non-
U.S. employees in the final rule would distort the pay ratio because of 
(1) differences in local pay practices,\160\ with some of these comment 
letters stating that, unlike many U.S. companies, companies in other 
countries include as ``compensation'' transportation, food, housing, 
wedding, birth, education, and phone expenses, as well as profit-
sharing arrangements and government provided benefits; \161\ (2) the 
exchange rates of foreign currencies; \162\ and (3) cost-of-living 
differences among countries.\163\
---------------------------------------------------------------------------

    \160\ See, e.g., letters from AAFA I, ABA, Aon Hewitt, American 
Benefits Council, Business Roundtable I, COEC I, Cummins Inc., 
Eaton, ExxonMobil, FEI, Freeport-McMoRan, Hyster-Yale, IBC, KBR, 
NACCO, NYC Bar, Prof. Ray, RILA, Semtech, and SHRM.
    \161\ See, e.g., letters from Eaton, Freeport-McMoRan, and SHRM.
    \162\ See, e.g., letters from American Benefits Council, BCIMC, 
Business Roundtable I, Cummins Inc., ExxonMobil, FEI, Freeport-
McMoRan, Prof. Ray, IBC, NAM I, NAM II, NYC Bar, Semtech, and SHRM.
    \163\ See, e.g., letters from AAFA I, American Benefits Council, 
BCIMC, COEC I, Corporate Secretaries, Cummins Inc., ExxonMobil FEI, 
Freeport-McMoRan, IBC, NAM I, NAM II, NYC Bar, and RILA.
---------------------------------------------------------------------------

    A few commenters argued that excluding non-U.S. employees was 
supported by statutory construction. They contended that, despite 
Section 953(b)'s reference to ``all employees,'' this statutory 
reference does not require the final rule to include non-U.S. 
employees.\164\ Some of these commenters asserted that excluding non-
U.S. employees would be consistent with Section 953(b) because the 
statute is silent as to whether ``all employees'' means non-U.S. 
employees.\165\ Others insisted that interpreting ``all employees'' to 
exclude non-U.S. employees would be appropriate because of a 
presumption against the extraterritoriality of U.S. laws.\166\
---------------------------------------------------------------------------

    \164\ See, e.g., letters from COEC I, ExxonMobil, Frederic W. 
Cook & Co., Freeport-McMoRan, NIRI, and NYC Bar.
    \165\ See, e.g., letters from Freeport-McMoRan, and NIRI.
    \166\ See, e.g., letters from COEC I and ExxonMobil.
---------------------------------------------------------------------------

    Some commenters advocated for a de minimis exemption for non-U.S. 
employees \167\ because, as one of these commenters stated, excluding a 
small number of employees is unlikely to affect ``in a material way'' 
the pay ratio and the nominal differences in ratios would be outweighed 
by the cost savings to registrants.\168\ Commenters provided a number 
of suggestions for a de minimis exemption. One commenter suggested 
that, if non-U.S. employees make up less than 20% of all a registrant's 
employees, the registrant should be permitted to exclude all non-U.S. 
employees.\169\ Another commenter stated that registrants should be 
permitted to exclude non-U.S. employees in any foreign country that 
comprises less than 5% of the registrant's aggregate global 
workforce.\170\
---------------------------------------------------------------------------

    \167\ See, e.g., letters from American Benefits Council (arguing 
in the alternative to excluding non-U.S. employees), Corporate 
Secretaries, ExxonMobil, Financial Services Institute (Dec. 2, 2013) 
(``FSI''), FSR, NACCO (arguing in the alternative to excluding non-
U.S. employees), NYC Bar (arguing in the alternative to excluding 
non-U.S. employees), and PNC Financial Services.
    \168\ See letter from PNC Financial Services.
    \169\ See letter from NACCO.
    \170\ See letter from American Benefits Council.
---------------------------------------------------------------------------

    A different commenter recommended that a registrant be permitted to 
exclude non-U.S. employees if they account for less than 5% of the 
registrant's total workforce because they would represent a de minimis 
number of employees.\171\ Also, according to that commenter, if the 
registrant's foreign employees account for more than 5% of all 
employees, this commenter recommended that the registrant should be 
permitted to exclude employees in any single foreign jurisdiction if 
they comprise less than 2% of total

[[Page 50122]]

employees, with an aggregate cap of 5%, and that the registrant be 
allowed to choose which country's employees to exclude.\172\ Another 
commenter noted that a registrant should be permitted to exclude non-
U.S. employees in a foreign country if the number of employees in that 
country is less than 1% of the registrant's total workforce.\173\
---------------------------------------------------------------------------

    \171\ See letter from FSR.
    \172\ See id.
    \173\ See letter from ExxonMobil.
---------------------------------------------------------------------------

    One commenter recommended that a registrant be able to exclude non-
U.S. employees if its CEO is based in the United States and more than 
50% of the registrant's employees also are based in the United 
States.\174\ Conversely, a different commenter contended that we should 
use our discretion to limit the scope of the term ``all employees'' to 
only U.S. employees if ``most of the registrant's employees'' and its 
principal executive officer ``work primarily outside the United 
States.'' \175\ Alternatively, the commenter suggested that, if the 
final rule includes non-U.S. employees, the rule should not include as 
``compensation'' the value of pension accruals, employer matching 
contributions for 401(k)s, and non-cash benefits. The commenter also 
advocated that, if non-U.S. employees are included, the final rule 
should exclude from the median identification employees who work in any 
E.U. jurisdiction that has strict data privacy rules and employs fewer 
than 50 people.
---------------------------------------------------------------------------

    \174\ See letter from NYC Bar.
    \175\ See letter from ABA. See also letter from NACCO 
(suggesting that, if non-U.S. employees make up more than 80% of a 
registrant's employees, the registrant should be exempt from the 
rule entirely).
---------------------------------------------------------------------------

    Another commenter contended that the final rule should include a 
principles-based exclusion that would permit companies the flexibility 
to exclude substantial percentages of employees if their compensation 
data is difficult to obtain and the impact would not be 
significant.\176\ The commenter cited the two memoranda analyzing the 
potential effects of excluding different percentages of employees on 
the pay ratio calculation that the staff from the Division of Economic 
and Risk Analysis (``DERA'').\177\ According to the commenter, these 
memoranda concluded that excluding a ``large share'' of employees from 
the pay ratio calculation would ``not have a significant impact'' on 
the ratio. As an example, the commenter noted that the June 30 
Memorandum states that excluding 40% of a registrant's employee 
population could reduce the pay ratio by 10.77% or increase it by 
12.08%. The commenter asserted these amounts would be ``negligible,'' 
and therefore registrants should be permitted to exclude these 
employees from their pay ratio calculations under the principles-based 
exclusionary approach advocated by the commenter. Additionally, the 
commenter noted that the average respondent to a survey it conducted 
indicated that 40% of its employees are located overseas and excluding 
these employees would decrease the compliance costs of the rule by 47%.
---------------------------------------------------------------------------

    \176\ See letter from COEC III.
    \177\ See Memorandum of the Division of Economic and Risk 
Analysis regarding the potential effect on pay ratio disclosure of 
exclusion of different percentages of employees at a range of 
thresholds (Jun. 4, 2015) (``June 4 Memorandum'') available at 
http://www.sec.gov/comments/s7-07-13/s70713-1556.pdf and Memorandum 
from the Division of Economic and Risk Analysis regarding an 
extension of the analysis of the potential effect on pay ratio 
disclosure of exclusion of different percentages of employees at a 
range of thresholds (Jun. 30, 2015) (``June 30 Memorandum'') 
available at http://www.sec.gov/comments/s7-07-13/s70713-1559.pdf. 
When the June 4 Memorandum was placed in the public comment file, we 
provided a press release announcing it was issued and an electronic 
alert through our RSS feed. The press release expressly advised that 
additional staff analyses might be placed in the comment file.
---------------------------------------------------------------------------

iii. Final Rule
(a) Non-U.S. Employees Generally
    After considering the comments, we have determined to include in 
the final rule's definition of ``employee'' a registrant's U.S. and 
non-U.S. employees, as proposed. We believe that the inclusion of non-
U.S. employees is the policy choice that most closely captures what 
Congress directed us to do in stating that the ratio should reflect 
``all employees.'' As noted above, we believe that the use of the word 
``all'' provides a general direction in favor of inclusion rather than 
exclusion of broad categories of employees. Although we recognize that 
our reading may impose more costs on registrants than if we excluded 
non-U.S. employees, given that Congress is undoubtedly aware that many 
U.S. registrants operate globally, we think Congress's use of ``all 
employees'' without any territorial limitation is a strong indication 
that Congress did not want us to categorically exclude non-U.S. 
employees. Further supporting our conclusion is the fact that, 
historically, the disclosures that are required of registrants under 
the securities laws apply to events, assets, conduct, or persons 
irrespective of whether those are located in the United States or 
abroad, and there is no indication that Congress intended to depart 
from this historical approach. However, as discussed below, we are 
adopting several tailored exemptions to address specific concerns 
raised by commenters.
    With respect to the assertion by some commenters that the rule 
should exclude non-U.S. employees because there is a presumption 
against the extraterritoriality of U.S. laws,\178\ we do not believe 
that including foreign employees within the ``employee'' definition 
constitutes an extraterritorial application of the statute. Generally, 
whether a particular application of a statute is ``extraterritorial'' 
turns on an analysis of whether the conduct that is the Congressional 
focus of the statute occurs here or abroad.\179\ The Congressional 
focus of Section 953(b) is disclosure by registrants that avail 
themselves of the U.S. public markets and thereby submit to U.S. 
law.\180\ As discussed above, Section 953(b)(1) directs us to amend 
Item 402, ``to require each issuer to disclose'' the pay ratio 
information ``in any filing'' described in Item 10(a). Companies are 
only required to provide a ``filing'' with us if they offer and/or sell 
securities in the United States and become subject to the our 
registration and filing requirements. Therefore, the final rule affects 
only registrants that come under the umbrella of United States laws.
---------------------------------------------------------------------------

    \178\ See, e.g., letters from COEC I and ExxonMobil.
    \179\ See Morrison v. Nat'l Australia Bank Ltd., 561 U.S. 247 
(2010).
    \180\ In this regard, we note that it is important that 
shareholders have access to information about both the domestic and 
foreign operations of a registrant given that a shareholder's 
investment in a U.S. registrant is typically exposed to the risks 
of, and the returns generated by, the global operations of the 
registrant.
---------------------------------------------------------------------------

    Our exemptive authority under Section 36 of the Exchange Act and 
Section 28 of the Securities Act would allow us to exempt registrants 
from including non-U.S. employees in the median employee determination 
required by Section 953(b).\181\ However,

[[Page 50123]]

after careful consideration, we decline to exercise our discretion to 
grant a wholesale exemption for non-U.S. employees. As we understand 
Section 953(b), Congress thought that it was important for shareholders 
to have pay ratio disclosure that reflects ``all employees'' of a 
registrant when making their say-on-pay votes. We do not believe it 
would be appropriate to second-guess Congress by granting a wholesale 
exemption for non-U.S. employees in a manner that we believe could 
fundamentally change the pay ratio information that Congress directed 
be provided to shareholders.\182\
---------------------------------------------------------------------------

    \181\ Section 36(a) of the Exchange Act and Section 28 of the 
Securities Act afford us general exemptive authority to 
conditionally or unconditionally exempt any person, security, or 
transaction, or any class or classes of persons, securities, or 
transactions, from any provision or provisions of this title (i.e., 
the Exchange Act and the Securities Act, respectively) or of any 
rule or regulation thereunder, to the extent that such exemption is 
necessary or appropriate in the public interest, and is consistent 
with the protection of investors. Although Section 953(b) was not 
expressly incorporated into either Act, our view is that, to the 
extent that the statutory criteria for invoking exemptive authority 
under these sections are met, the exemptive authorities afforded by 
Section 36(a) and Section 28 are available here. We construe Section 
953(b) as a Congressional directive to us to rely on our Exchange 
Act and Securities Act rulemaking authorities to amend Sec.  229.402 
of title 17, Code of Federal Regulations to require the pay ratio 
disclosure. The pay ratio amendments that we are adopting, 
therefore, are rules or regulations under the Exchange Act and the 
Securities Act and, thus, fall within the express terms of Section 
36(a) and Section 28.
    \182\ Our use of exemptive authority where foreign data privacy 
laws are involved (as discussed below) may result in different pay 
ratio disclosure in certain instances, but we believe the use of 
exemptive authority is appropriate because it reflects a carefully 
tailored accommodation necessary to address a situation (i.e., 
foreign data privacy laws) that Congress may not have contemplated.
---------------------------------------------------------------------------

    While the final rule does not exclude non-U.S. employees, in 
response to concerns that the inclusion of non-U.S. employees could 
raise compliance costs for multinational companies, introduce cross-
border compliance issues, and have an adverse impact on competition, we 
are exercising our exemptive authority to provide two tailored 
exemptions that we believe will alleviate some of these concerns: (1) 
an exemption that applies when a foreign jurisdiction's data privacy 
laws or regulations are such that, despite its reasonable efforts to 
obtain or process information necessary to comply with the rule, a 
registrant is unable to do so without violating those laws or 
regulations, and (2) a de minimis exemption. These exemptions are 
discussed in detail below.
(b) Foreign Data Privacy Law Exemption
    The first instance in which we believe it is appropriate to provide 
an exemption to the general requirement that non-U.S. employees be 
included in the pay ratio disclosure is when a jurisdiction's data 
privacy laws or regulations are such that, despite a registrant's 
reasonable efforts to obtain or process information necessary to comply 
with the rule, it is unable to do so without violating those laws or 
regulations. A number of commenters noted that many countries have data 
privacy and other laws that prevent registrants from transferring 
payroll data outside that country's borders (even if the transfer would 
be within the same company), which would make compiling the information 
necessary for the pay ratio disclosure illegal.\183\ For example, 
commenters noted that the E.U. prohibits the transfer of personal data 
to a third country that does not ensure an adequate level of privacy 
protections (the United States is considered not to ensure adequate 
privacy protections) and that China, Japan, Mexico, Canada, Peru, 
Australia, Russia, Switzerland, Argentina, and Singapore have adopted 
or are considering similar rules.\184\
---------------------------------------------------------------------------

    \183\ See, e.g., letters from AAFA II, ABA, American Benefits 
Council, BCIMC, Business Roundtable I, Chamber I, COEC I, Corporate 
Secretaries, Cummins Inc., Eaton, ExxonMobil, Freeport-McMoRan, 
Hyster-Yale, IBC, NACCO, NAM I, NAM II, NIRI, RILA, Semtech, SHRM, 
and WorldatWork I.
    \184\ See, e.g., letters from ABA, American Benefits Council, 
Business Roundtable I, COEC I, Cummins Inc., NAM I, NAM II, RILA, 
and WorldatWork I.
---------------------------------------------------------------------------

    One of these commenters acknowledged, however, that ``it would be 
reasonable to expect that registrants which employ workers abroad 
already have an understanding of their obligations under the data 
privacy laws of each jurisdiction in which they operate, and have 
undertaken to comply with those laws,'' but the commenter was concerned 
that existing actions taken by registrants to comply with those laws 
may not be sufficiently flexible to facilitate compliance with the 
rule.\185\ In this regard, the commenter noted that, under the E.U. 
Directive, data may be exempt from the dictates of the Directive if it 
is truly anonymous such that the data cannot be attributed to any 
identifiable person. It would seem, therefore, that a registrant 
employing hundreds or thousands of employees in an E.U. jurisdiction 
could collect compensation data for purposes of complying with Section 
953(b) in a way that would preserve employee anonymity while a 
registrant that employs only a handful of employees in an E.U. 
jurisdiction may not be able to collect the data in such a manner.\186\ 
The commenter therefore recommended that, if the final rule did not 
exclude all non-U.S. employees, it should permit registrants to exclude 
from the definition of ``employee'' any employee in a jurisdiction 
where such collection, analysis, and transmission would violate a 
registrant's existing data privacy obligations.\187\
---------------------------------------------------------------------------

    \185\ See letter from ABA.
    \186\ In such situations, we recognize that the de minimis 
exemption may be available.
    \187\ See letter from ABA (recommending that we consider 
permitting registrants to exclude from the identification of the 
median employee those employees who work in an E.U. jurisdiction 
that maintains strict data privacy laws and in which the registrant 
employs fewer than 50 employees).
---------------------------------------------------------------------------

    After considering the comments received, we are persuaded that a 
tailored exemption from the definition of ``employee'' is appropriate 
where a foreign country's data privacy laws or regulations are such 
that a registrant is not able to comply with the rule without violating 
those laws or regulations in spite of its reasonable efforts to obtain 
or process the necessary information.\188\
---------------------------------------------------------------------------

    \188\ As required by Section 28 of the Securities Act and 
Section 36 of the Exchange Act, we find that the exemption here is 
consistent with investor protection and is necessary or appropriate 
in the public interest. We make these findings based on the fact 
that, without the exemption, registrants operating in countries with 
applicable privacy laws could be forced into the difficult situation 
of violating either that country's laws or U.S. law, and we believe 
that because of the limited and tailored nature of the exemption, it 
will not materially impact the pay ratio disclosure.
---------------------------------------------------------------------------

    Although, as noted above, we believe the inclusion of non-U.S. 
employees is consistent with the Congressional directive and is 
important for providing pay ratio information that reflects a 
registrant's overall employment practices, we do not have any 
indication that Congress intended that a registrant should have to 
choose between complying with our disclosure rules and violating the 
laws of a foreign jurisdiction. We believe that, on balance, providing 
an accommodation in such situations would not substantially affect the 
utility of the Section 953(b) disclosures for shareholder say-on-pay 
votes.
    To prevent any potential manipulation, the rule requires the 
registrant to exercise reasonable efforts to obtain or process the 
information necessary for compliance with the final rule. As part of 
its reasonable efforts, the registrant must seek an exemption or other 
relief under the applicable jurisdiction's governing data privacy laws 
or regulations and use the exemption if granted.
    If a registrant excludes any non-U.S. employees in a particular 
jurisdiction under the data privacy exemption, it must exclude all non-
U.S. employees in that jurisdiction. Additionally, the registrant must 
list the excluded jurisdictions, identify the specific data privacy law 
or regulation, explain how complying with the final rule violates the 
law or regulation (including the efforts made by the registrant to use 
or seek an exemption or other relief under such law or regulation), and 
provide the approximate number of employees exempted from each 
jurisdiction based on this exemption.
    Also, the registrant must obtain a legal opinion that opines on the 
inability of the registrant to obtain or process the information 
necessary for compliance with the final rule without violating that 
jurisdiction's laws or regulations governing data privacy, including 
the registrant's inability to obtain an exemption or other relief under 
any

[[Page 50124]]

governing laws or regulations. The legal opinion must be filed as an 
exhibit with the filing in which the pay ratio disclosure is included. 
For filings other than proxy or information statements, the legal 
opinion must be filed as an exhibit under Exhibit 99.\189\
---------------------------------------------------------------------------

    \189\ See Item 601(b)(99) of Regulation S-K [17 CFR 
229.601(b)(99)]. The exhibit must be filed with proxy or information 
statements pursuant to Item 25 of Schedule 14A.
---------------------------------------------------------------------------

(c) De Minimis Exemption
    The second instance in which we believe it is appropriate to 
provide an exemption from the general requirement to include non-U.S. 
employees in identifying the median employee is when a de minimis 
number of a registrant's employees work outside the United States. The 
de minimis exemption is a change from the proposed rule. Under this 
exemption, registrants whose non-U.S. employees make up 5% or less of 
their total U.S. and non-U.S. employees may exclude all of them when 
identifying their median employee. If such a registrant chooses to 
exclude any non-U.S. employees under this exemption, it must exclude 
all of them. A registrant with more than 5% non-U.S. employees may also 
exclude non-U.S. employees up to the 5% threshold; provided that, if 
such a registrant excludes any non-U.S. employees in a particular 
foreign jurisdiction, it must exclude all the employees in that 
jurisdiction. The registrant may not pick and choose which employees to 
exclude in any one jurisdiction.
    We believe a de minimis exemption provides flexibility in a manner 
that will not meaningfully alter the pay ratio disclosure. We are 
persuaded that a de minimis exemption is appropriate after considering 
the potential cost savings to registrants and the small effect it would 
have on the pay ratio, as discussed below. The final rule establishes 
the de minimis threshold at 5%. The commenters that suggested specific 
de minimis thresholds did not provide reasons why the particular 
thresholds they suggested were suitable, but several of these 
commenters suggested a threshold of 5%.\190\ We believe the 5% 
threshold will both limit the exemption to an amount that is, in fact, 
de minimis and help address the payroll or other data challenges that 
may arise for registrants with a small percentage of non-U.S. 
employees.
---------------------------------------------------------------------------

    \190\ See letters from PNC Financial Services (suggesting 5% as 
the de minimis threshold) and FSR (recommending a registrant be 
permitted to exclude non-U.S. employees if they account for less 
than 5% of the registrant's employees, but, if the registrant's 
foreign employees account for more than 5% of all employees, the 
registrant may exclude employees in any single foreign jurisdiction 
if they comprise less than 2% of total employees, with an aggregate 
cap of 5%).
---------------------------------------------------------------------------

    Although commenters did not provide data about the effect on the 
pay ratio of potential de minimis thresholds, staff in DERA performed 
an analysis of the potential effect on pay ratio disclosure of 
excluding different percentages of employees at a range of thresholds 
and posted to the comment file two memoranda containing the 
analysis.\191\ As discussed in further detail both in the memoranda and 
in the Economic Analysis section below, under such analysis and based 
on the assumptions set forth in the analysis, the exclusion of 5% of 
employees may cause the pay ratio calculation to decrease by up to 3.4% 
or to increase by up to 3.5%. We believe such analysis confirms that 
the effect of the 5% threshold on the pay ratio disclosure will be de 
minimis. We note that one commenter observed that under one of the 
scenarios analyzed in the June 30 Memorandum, excluding 40% of a 
registrant's employees may cause the pay ratio calculation to decrease 
by up to 10.77% or to increase by up to 12.08%.\192\ The commenter 
called this impact ``not significant'' and ``negligible'' and, based on 
this impact, contended that the final rule should permit registrants to 
exclude up to 40% of their non-U.S. employees. We are not adopting this 
suggestion. We believe that the exclusion of 40% of employees would be 
a fundamentally different type of exclusion than the one we adopting 
here--that is, a de minimis exclusion designed to allow companies to 
exclude employees in jurisdictions where there are only a limited 
number of employees and where the costs of including such employees may 
be disproportionately greater than the incremental information they 
would add to the disclosure. The exclusion of 40% of employees is not a 
de minimis amount of employees, and, in contrast to the assertions of 
the commenter, we believe that a decrease in the pay ratio calculation 
of up to 10.77% or increase by up to 12.08% is significant and more 
than negligible, and we do not believe that it is de minimis. 
Additionally, as the commenter acknowledges, under other scenarios in 
the memorandum, exclusion of up to 40% of a registrant's employees 
could have an even greater effect on the pay ratio (e.g., by causing it 
to decrease by up to 25.06% or to increase by up to 33.43%). In 
contrast, the 5% de minimis threshold both results in the exclusion of 
a de minimis number of employees and has a de minimis effect on the pay 
ratio, regardless of which scenario is considered. Accordingly, the 
final rule's de minimis threshold for non-U.S. employees does not 
exceed 5%.
---------------------------------------------------------------------------

    \191\ See the June 4 Memorandum and the June 30 Memorandum. See 
also, Section III.D.2.c.vi below.
    \192\ See letter from COEC III.
---------------------------------------------------------------------------

    As one commenter warned, there is a possibility for intentional 
manipulation in identifying the median employee when a de minimis 
exemption is provided and a registrant is permitted to choose which 
jurisdictions to exclude.\193\ To provide safeguards against any 
potential manipulation, the final rule requires that, if a registrant 
with 5% or fewer non-U.S. employees chooses to exclude those employees 
from the calculation of its median employee, it may not pick and choose 
which of the 5% to exclude and must exclude all of its non-U.S. 
employees. Similarly, if a registrant with more than 5% non-U.S. 
employees excludes any employees in any jurisdiction, it must exclude 
all the employees in that jurisdiction. In this regard, we recognize 
that this requirement could prevent some registrants with more than 5% 
non-U.S. employees from excluding any of its foreign employees. A 
purpose of the de minimis exemption is to provide relief from the need 
to determine how to integrate payroll systems and compensation 
arrangements in jurisdictions where the number of employees may not 
justify the effort, and we believe that setting the threshold at 5% 
establishes an appropriate measure of relief.
---------------------------------------------------------------------------

    \193\ See letter from FSR.
---------------------------------------------------------------------------

    The final rule also requires a registrant using the de minimis 
exemption to provide certain disclosures. If the registrant excludes 
any non-U.S. employees under the de minimis exemption, it must disclose 
the jurisdiction or jurisdictions from which employees are being 
excluded, the approximate number of employees excluded from each 
jurisdiction under the de minimis exemption, the total number of its 
U.S. and non-U.S. employees irrespective of any exemption (data privacy 
or de minimis) and the total number of its U.S. and non-U.S. employees 
used for its de minimis calculation.
    In calculating the number of non-U.S. employees that may be 
excluded under the de minimis exemption, a registrant must count any 
non-U.S. employee exempted under the data privacy exemption against the 
availability. A registrant may exclude any non-U.S. employee that meets 
the data privacy exemption, even if the number of excluded employees 
exceeds 5% of the registrant's total employees. If, however,

[[Page 50125]]

the number of employees excluded under the data privacy exemption 
equals or exceeds 5% of the registrant's total employees, the 
registrant may not use the de minimis exemption to exclude additional 
non-U.S. employees.
    For example, a registrant has non-U.S. employees located in two 
foreign jurisdictions. One of the jurisdictions has 10% of the 
registrant's total employees who are non-U.S. employees and has data 
privacy laws that, despite its reasonable efforts to obtain or process 
the information necessary for compliance with the final rule, the 
registrant is unable to do so without violating those data privacy laws 
or regulations. The other jurisdiction has an additional 5% of the 
registrant's total employees who are non-U.S. employees and has no such 
data privacy laws or regulations. The registrant may exclude all the 
non-U.S. employees in the first jurisdiction, which has 10% of the 
registrant's total employees. In that situation, however, the 
registrant may not exclude the non-U.S. employees in the second 
jurisdiction, which has the additional 5% of the total employees, even 
though the 5% would otherwise constitute a de minimis amount of non-
U.S. employees, because the registrant is already excluding over 5% of 
its employees under the data privacy exemption.\194\
---------------------------------------------------------------------------

    \194\ We do not envision a scenario in which a registrant can 
forgo the data privacy exemption in favor of the de minimis 
exemption in the above or similar situations. The data privacy 
exemption is permitted only for circumstances in which a foreign 
jurisdiction's laws or regulations governing data privacy are such 
that a registrant is unable to comply with the final rule without 
violating the that jurisdiction's laws or regulations. If a 
registrant is in a position to forgo the data privacy exemption, it 
would not be considered eligible for the exemption.
---------------------------------------------------------------------------

    Moreover, if the number of non-U.S. employees excluded under the 
data privacy exemption is less than 5% of the registrant's total 
employees, the registrant may use the de minimis exemption to exclude 
no more than the number of non-U.S. employees that, combined with the 
data privacy exemption, equals 5% of the registrant's total employees.
    For example, a registrant has non-U.S. employees located in two 
foreign jurisdictions. One of the jurisdictions has 2.5% of the 
registrant's total employees who are non-U.S. employees and has data 
privacy laws that, despite its reasonable efforts to obtain or process 
the information necessary for compliance with the final rule, the 
registrant is unable to do so without violating those data privacy laws 
or regulations. The other jurisdiction has an additional 2.5% of the 
registrant's total employees who are non-U.S. employees and has no such 
data privacy laws or regulations. The registrant may exclude the 2.5% 
of total employees who are non-U.S. employees in the first jurisdiction 
under the data privacy exemption. The registrant may also exclude the 
additional 2.5% of the registrant's total employees who are non-U.S. 
employees from the second jurisdiction because the total number of 
exempted non-U.S. employees under both the data privacy and the de 
minimis exemptions equal only 5% of the registrant's total employees.
    Alternatively, in the above example, if the number of non-U.S. 
employees in the second jurisdiction was 3% of the registrant's total 
employees, the registrant could not exclude the non-U.S. employees in 
that jurisdiction because the registrant's number of excluded non-U.S. 
employees in both jurisdictions would be over 5% of its total 
employees.
(d) Cost-of-Living Adjustment
    In the Proposing Release, we requested comment on whether we should 
permit cost-of-living adjustments for employees in different countries. 
A number of commenters who addressed this issue contended that 
unadjusted cost of living differences between countries would cause the 
inclusion of non-U.S. employees to render the pay ratio disclosure 
misleading.\195\ Accordingly, some of these commenters suggested that 
the pay ratio disclosure could be more meaningful for some registrants 
if the final rule permitted cost-of-living adjustments.\196\ However, 
other commenters objected to permitting cost-of living-
adjustments,\197\ asserting that the compensation of a non-U.S. 
employee ``is more directly comparable to the total compensation of a 
registrant's [PEO] without a cost-of-living adjustment,'' \198\ and 
that permitting cost-of-living adjustments could add a level of 
subjectivity to the pay ratio disclosure.\199\
---------------------------------------------------------------------------

    \195\ See, e.g., letters from AAFA I, American Benefits Council, 
BCIMC, Corporate Secretaries, Cummins Inc., ExxonMobil, NAM I, NAM 
II, and SH&P.
    \196\ See letter from ExxonMobil (``Example 2: Differences in 
cost of living. Two employees hold similar jobs in two different 
countries, C and D. The employee in Country C receives total annual 
compensation of $100,000 and the employee in Country D receives 
total annual compensation of $75,000. The cost of living in Country 
D is approximately 50% of the cost of living in Country C. In real 
economic terms, the employee in Country D enjoys significantly 
higher pay than the employee in Country C. However, under the rules 
as proposed the appearance is reversed.'') (Emphasis in original.). 
See also, letters from NAM I, NAM II, and SH&P.
    \197\ See letters from ABA, Prof. Ray, and WorldatWork I.
    \198\ See letter from ABA (``We believe that the Commission 
should not allow registrants to make cost-of-living adjustments for 
non-U.S.-based employees (should the agency determine to include 
them), other than the annualization and full-time equivalent 
adjustments discussed above in our responses to Questions 23 and 24. 
We believe the total compensation of such full-time employees is 
more directly comparable to the total compensation of a registrant's 
principal executive officer without a cost-of-living adjustment than 
with it.'').
    \199\ See letter from Prof. Ray.
---------------------------------------------------------------------------

    We acknowledge that differences in the underlying economic 
conditions of the countries in which registrants operate likely have an 
effect on the compensation paid to employees in those jurisdictions. As 
a result, requiring registrants to determine their median employee and 
calculate the pay ratio without permitting them to adjust for these 
different underlying economic conditions could result in what some 
would consider a statistic that does not appropriately reflect the 
value of the compensation paid to individuals in those countries. The 
final rule, therefore, allows registrants the option to make cost-of-
living adjustments to the compensation of their employees in 
jurisdictions other than the jurisdiction in which the PEO resides when 
identifying the median employee (whether using annual total 
compensation or any other consistently applied compensation measure), 
provided that the adjustment is applied to all such employees included 
in the calculation.\200\ If the registrant chooses this option, the 
compensation of such employees will have to be adjusted to the cost of 
living in the jurisdiction in which the PEO resides. Further, if the 
registrant uses a cost-of-living adjustment to identify the median 
employee, and the median employee identified is an employee in a 
jurisdiction other than the one in which the PEO resides, the 
registrant must use the same cost-of-living adjustment in calculating 
the median employee's annual total compensation and disclose the median 
employee's jurisdiction. If a registrant does not make cost-of-living-
adjustments to its employees when identifying the median employee, the 
registrant is not permitted to make cost-of-living adjustments to the 
median employee's annual total compensation if the median employee is 
an employee in

[[Page 50126]]

a jurisdiction other than the jurisdiction in which the PEO resides.
---------------------------------------------------------------------------

    \200\ We are utilizing our authority under Section 28 of the 
Securities Act and Section 36 of the Exchange Act to permit 
registrants to elect to identify the median employee by first 
adjusting the compensation of employees in jurisdictions other than 
the jurisdiction in which the PEO resides to the cost of living in 
the CEO's jurisdiction of residence. Moreover, for the reasons 
described above, we believe that this conditional exemption is 
consistent with the public interest and investor protection.
---------------------------------------------------------------------------

    In the Proposing Release, we said that we preliminarily believed 
that certain adjustments, including cost-of-living adjustments, ``could 
distort an understanding of the registrant's compensation practices.'' 
Based on our fuller understanding of the Congressional purpose 
underlying the pay ratio disclosure and the comments received on the 
proposal, however, we are persuaded that allowing registrants the 
option of a cost-of-living adjustment for employees in jurisdictions 
other than the jurisdiction in which the PEO resides could be useful to 
investors as they make their say-on-pay votes. Put simply, a cost-of-
living adjustment could provide a more meaningful comparison of the 
PEO's compensation to the actual value of the median employee's 
compensation by effectively filtering out that part of the difference 
in compensation that results from differences in the cost of living 
between the PEO's place of residence (typically, the United States) and 
the median employee's jurisdiction. For some shareholders making their 
say-on-pay votes, we believe that what may matter is the value of 
compensation received by the median employee, rather than the dollar 
amount of the compensation paid. Although we are not mandating that 
registrants adjust for these cost-of-living considerations, we believe 
that it is appropriate to give them the option to make such adjustments 
where they determine that doing so would provide more useful 
information to their shareholders as they vote on executive 
compensation.\201\
---------------------------------------------------------------------------

    \201\ Although the final rule gives registrants the option to 
make cost-of-living adjustments for employees in jurisdictions other 
than the jurisdiction in which the PEO resides, we are not giving 
registrants the option of adjusting part-time or seasonal employees' 
compensation as though they were full time employees. For U.S. part-
time or seasonal employees, the unadjusted compensation reflects the 
actual relative value of the compensation received by that employee, 
unlike non-U.S. employees who may be working in countries with a 
significantly lower cost of living. Moreover, adjusting the 
compensation of part-time or seasonal employees to what they would 
have received if they had been full time employees would cause the 
median to not be reasonably representative of the registrant's 
actual employment arrangements for its workforce during the period.
---------------------------------------------------------------------------

    We recognize that providing registrants the flexibility to make 
cost-of-living adjustments could add a level of subjectivity to the pay 
ratio disclosure, make compliance with the rule more burdensome, or 
permit registrants to alter the reported ratio to achieve a particular 
objective with the ratio disclosure.\202\ Registrants with a 
significant number of employees in countries with higher cost-of-living 
than the jurisdiction in which the PEO resides may be unlikely to 
adjust those compensation figures downward, while registrants with a 
sizable work force in countries with a lower cost-of-living may be 
likely to adjust the compensation figures upward.
---------------------------------------------------------------------------

    \202\ See letters from Prof. Ray (stating that permitting cost-
of-living adjustments ``will only make the pay ratio more 
subjective,'' because they ``add subjectivity'' to the disclosure) 
and WorldatWork I (``Cost-of-living adjustments and full-time 
compensation adjustments would make compliance more burdensome by 
requiring more context in the explanation of how the ratio was 
calculated.'').
---------------------------------------------------------------------------

    We believe, however, that the final rule mitigates these concerns 
by requiring registrants to briefly describe any cost-of-living 
adjustments they used to identify the median employee or to calculate 
annual total compensation, including the measure used as the basis for 
the cost-of-living adjustment, and disclose the country in which the 
median employee is located. Additionally, the final rule requires that 
any registrant electing to present the pay ratio using a cost-of-living 
adjustment must also disclose the median employee's annual total 
compensation and pay ratio without the cost-of-living adjustments. To 
calculate this pay ratio, the registrant will need to identify the 
median employee without using any cost-of-living adjustments. In this 
way, shareholders would have pay ratio information both in terms of the 
value of compensation received by the employee and in terms of the 
compensation paid by the registrant.\203\
---------------------------------------------------------------------------

    \203\ We believe that requiring this disclosure of the 
unadjusted pay ratio for those registrants who choose to include a 
cost-of-living adjustment will help to mitigate the concerns noted 
in the Proposing Release about the impact that a cost-of-living 
adjustment could have on an understanding of a registrant's 
compensation practices.
---------------------------------------------------------------------------

    For registrants who choose to present the pay ratio using a cost-
of-living adjustment, the pay ratio required by Item 402(u)(1)(iii) 
will be the cost-of-living adjusted pay ratio. Disclosure of the 
unadjusted pay ratio will be available to provide context for the 
registrant's required pay ratio. Because the cost-of-living adjustment 
will be optional for registrants, we assume they will choose to avail 
themselves of this option only to the extent they believe the benefits 
of doing so will justify any additional costs to make the adjustment.
d. Employees of Consolidated Subsidiaries
i. Proposed Rule
    The proposed rule would cover employees of both a registrant and 
its subsidiaries, which is similar to the approach taken for other Item 
402 information.\204\ In the context of Item 402, a subsidiary of a 
registrant is an affiliate controlled by the registrant directly or 
indirectly through one or more intermediaries, as set forth in the 
definition of ``subsidiary'' under both Securities Act Rule 405 and 
Exchange Act Rule 12b-2. Therefore, the proposal would cover an 
employee if he or she was employed by the registrant or a subsidiary of 
the registrant as defined in Rule 405 and Rule 12b-2.
---------------------------------------------------------------------------

    \204\ See Item 402(a)(2) and Instruction 2 to Item 402(a)(3).
---------------------------------------------------------------------------

ii. Comments on the Proposed Rule
    The majority of commenters that discussed this issue recommended 
that the rule only require parent registrants to incorporate into their 
pay ratio disclosure the employees of their consolidated 
subsidiaries.\205\ One of the commenters claimed that there would be a 
91% increase in compliance costs if the final rule included minority-
owned subsidiaries and joint ventures because registrants would 
otherwise be required to ``engage in an extensive information gathering 
process'' without ``access to the payroll and human resources 
information needed for the pay ratio from subsidiaries or other 
entities with a more tenuous connection, such as joint ventures.'' 
\206\ Another commenter claimed that registrants do not exercise much 
influence on the compensation policies and practices of entities in 
which they have only a minority or nominal interest.\207\ Only one 
commenter asserted that the final rule should exclude compensation 
information from all subsidiaries and be limited to only the 
compensation of employees directly employed by the registrant.\208\
---------------------------------------------------------------------------

    \205\ See, e.g., letters from ABA (limiting the requirement only 
to employees of the registrant's wholly-owned or majority-owned 
subsidiaries with consolidated financial statements, but not 
subsidiaries that are portfolio companies of business development 
companies), Best Buy et al., Business Roundtable I, COEC I, COEC II, 
Corporate Secretaries, CT State Treasurer, Davis Polk, Eaton, 
ExxonMobil, Mercer I, Meridian, NACCO, NAM I, and NAM II.
    \206\ See letter from COEC I.
    \207\ See letter from ABA.
    \208\ See letter from WorldatWork I.
---------------------------------------------------------------------------

    Some commenters suggested that the final rule require registrants 
to incorporate only employees of their wholly-owned subsidiaries and 
not employees of their joint ventures.\209\ Other commenters stated 
that the final rule should allow a subsidiary to exclude its pay ratio 
disclosure in its filings if the subsidiary's employees are 
incorporated into its parent registrant's

[[Page 50127]]

pay ratio disclosure.\210\ One commenter asserted that the final rule 
should require a registrant to include the compensation information of 
its subsidiary only if the issuer has ``control'' over the subsidiary 
(as ``control'' is defined in our rules).\211\ Another commenter 
maintained that the final rule should require a registrant to include 
its subsidiary only if the registrant has ``actual control'' over the 
compensation decisions made at the subsidiary level.\212\ Finally, a 
few commenters contended that the final rule should require registrants 
to include the employees of their subsidiaries in their pay ratio 
generally without specifying the types of subsidiaries.\213\
---------------------------------------------------------------------------

    \209\ See, e.g., letters from AAFA I, Business Roundtable I, 
Corporate Secretaries, NACCO, NAM I, NAM II, and PM&P.
    \210\ See, e.g., letters from FSI and FSR.
    \211\ See letter from Capital Strategies.
    \212\ See letter from Cummins Inc.
    \213\ See, e.g., letters from AFL-CIO I, Domini, and US SIF.
---------------------------------------------------------------------------

iii. Final Rule
    After considering these comments, we are revising the final rule 
from the proposal. Unlike the proposed rule, the final rule defines 
``employee'' to include only the employees of the registrant and its 
consolidated subsidiaries rather than employees of subsidiaries that 
were affiliates it controlled directly or indirectly through one or 
more intermediaries, as set forth in the definition of ``subsidiary'' 
under both Securities Act Rule 405 and Exchange Act Rule 12b-2. This 
change should reduce costs and burdens for registrants, while 
maintaining the benefits of the pay ratio rule, as discussed below.
    Rule 12b-2 and Rule 405 define a ``subsidiary'' as ``an affiliate 
controlled by [an entity] directly, or indirectly through one or more 
intermediaries,'' while an ``affiliate'' is defined as ``a person that 
directly, or indirectly through one or more intermediaries, controls or 
is controlled by, or is under common control with, the person 
specified.'' \214\ The term ``control'' (including the terms 
``controlling,'' ``controlled by'' and ``under common control with'') 
means the possession, direct or indirect, of the power to direct or 
cause the direction of the management and policies of a person, whether 
through the ownership of voting securities, by contract, or 
otherwise.\215\ One commenter described this definition of 
``subsidiary'' as ``very expansive'' because it includes not just 
affiliates controlled directly by the registrant, but also those 
controlled indirectly by the registrant through one or more 
intermediaries or that are under common control with the 
registrant.\216\ In the Section 16 context, we have noted that many 
practitioners believe that individuals or entities holding as little as 
10% or more of the voting equity securities of a registrant may likely 
be considered an affiliate or control person.\217\ Further, whether an 
affiliate is controlled by an entity is based on the facts and 
circumstances of each situation, so whether a company should be 
considered a ``subsidiary'' of a registrant is not always clear.\218\ 
Therefore, depending on the specific facts and circumstances of the 
situation, if the rule used the term ``subsidiary'' as defined under 
Rules 12b-2 and 405, a registrant could potentially be required to 
include the employees of a company in which it holds as little as a 10% 
ownership stake. As commenters noted, obtaining compensation and 
payroll data from unconsolidated entities could be costly, burdensome, 
or potentially impossible.\219\ Additionally, one commenter suggested 
that, because compensation disclosure is designed to facilitate the 
comparison of the PEO's pay to the performance of the company based on 
its consolidated financial statements, the pay ratio should relate to 
the same consolidated financial performance of the company and not to 
non-consolidated entities and other factors if the purpose of the rule 
is to enhance compensation disclosure.\220\
---------------------------------------------------------------------------

    \214\ 17 CFR 240.12b-2 and 17 CFR 230.405.
    \215\ Id.
    \216\ See letter from ABA.
    \217\ See Revision of Rule 144, Rule 145 and Form 144, Release 
No. 33-7391, Section III.B (Jun. 27, 1995) (``Under the proposal [to 
make revisions to Rule 144, 145, and Form 144], the same criteria 
used to determine those persons that are not `insiders' under 
Exchange Act Section 16 would be used for Rule 144. Many 
practitioners already used Section 16 criteria as a guide. The 
Commission believes it is likely that most persons who are not 
officers, directors, or 10% holders are not in a `control' 
position.'').
    \218\ See, e.g., letters from ABA, Best Buy et al., Business 
Roundtable I, COEC I, and Davis Polk.
    \219\ See, e.g., letters from ABA, Best Buy et al., Business 
Roundtable I, COEC I, Corporate Secretaries, Davis Polk, Eaton, 
ExxonMobil, NACCO, and NAM I.
    \220\ See letter from NAM II.
---------------------------------------------------------------------------

    In contrast, defining a ``subsidiary'' based on whether a 
registrant consolidates a company in its financial statements likely 
will decrease the costs and burdens on a registrant compared with the 
proposal because most registrants consolidate based on their ownership 
of over 50% of the outstanding voting shares of their subsidiaries and 
more guidance is readily available on when consolidating subsidiaries 
is appropriate than when an entity should be considered a 
``subsidiary'' based on the concept of control. For example, the United 
States generally accepted accounting principles (``U.S. GAAP'') 
traditionally has required a company holding the ``controlling 
financial interest'' of another company to consolidate that 
company.\221\ The usual condition for consolidation is a controlling 
financial interest through majority ownership of over 50% of the 
outstanding voting shares.\222\ Determining whether a company is a 
``subsidiary'' under the consolidated financial statement method, as 
opposed to the using the definition of ``subsidiary'' under Rules 405 
and 12b-2, generally will provide a higher quantitative threshold and 
thus a smaller pool of employees to include in the median employee 
determination, which should help to reduce costs associated with making 
such a determination. Overall, the standard for consolidation and the 
definition of ``control'' under Rules 405 and 12b-2 are both driven by 
very similar concepts of control. Use of a consolidated subsidiary 
standard will typically exclude employees of entities where a company 
holds between a 10% to 50% voting interest in such entity.
---------------------------------------------------------------------------

    \221\ See Volume I, Section 1160.1.12 of PWC 2010 Accounting and 
Reporting Manual.
    \222\ See ASC 810-10-15-8 (``The usual condition for a 
controlling financial interest is ownership of a majority voting 
interest, and, therefore, as a general rule ownership by one 
reporting entity, directly or indirectly, of more than 50 percent of 
the outstanding voting shares of another entity is a condition 
pointing toward consolidation. The power to control may also exist 
with a lesser percentage of ownership, for example, by contract, 
lease, agreement with other stockholders, or by court decree.'').
---------------------------------------------------------------------------

    Although this change from the proposal generally will result in a 
smaller pool of employees being used for the median employee 
determination, we do not believe it will undermine the usefulness of 
the required disclosures or conflict with the purposes of Section 
953(b). As one commenter indicated, ``registrants do not exercise much, 
if any, influence on the compensation policies and practices of 
entities in which they have only a minority or nominal interest (unless 
the employees of such entities provide services directly to the 
registrant).'' \223\ According to the commenter, limiting the final 
rule to employees of a registrant's consolidated subsidiaries, 
therefore, ``would not deprive investors of useful information or 
important insights into a registrant's compensation structure.'' We 
believe that requiring registrants to consider only their employees and 
the employees of their consolidated subsidiaries in identifying their 
median employee should not limit the usefulness of the pay ratio 
disclosure as a data point for shareholders to use in making their

[[Page 50128]]

voting decisions on executive compensation under Section 951 of the 
Dodd-Frank Act in any significant way.
---------------------------------------------------------------------------

    \223\ See letter from ABA.
---------------------------------------------------------------------------

e. Any PEO Compensation in the Last Full Fiscal Year
i. Proposed Rule
    The Proposing Release did not discuss the compensation information 
that would be required if one or more of a registrant's PEOs served 
only part of a fiscal year.
ii. Comments on the Proposed Rule
    Some commenters suggested that a registrant should be required to 
disclose the compensation information only for the PEO holding the 
position at the end of the last completed fiscal year.\224\ One 
commenter suggested an exception where the PEO has served for most of a 
fiscal year but departs before the last day, in which case the 
commenter recommended that we require compensation information 
disclosure for the person who served as PEO for the majority of the 
fiscal year.\225\ Another commenter suggested the use of the PEO as of 
the last day of the most recently completed fiscal year and noted that 
the issue of that person's annual total compensation representing less 
than a full year's compensation ``could be addressed by requiring 
registrants to annualize the annual total compensation for the chief 
executive officer serving at fiscal year end.'' \226\
---------------------------------------------------------------------------

    \224\ See, e.g., letters from ABA, Corporate Secretaries, Davis 
Polk, McGuireWoods, and PM&P.
    \225\ See letter from ABA.
    \226\ See letter from Davis Polk.
---------------------------------------------------------------------------

iii. Final Rule
    The final rule allows a registrant a choice of two options in 
calculating the annual total compensation for its PEO in situations in 
which the registrant replaces its PEO with another PEO during its 
fiscal year. In these situations, the registrant must disclose which 
option it chose and how it calculated its PEO's annual total 
compensation. First, a registrant may take the total compensation 
calculated pursuant to Item 402(c)(2)(x), and reflected in the Summary 
Compensation Table, provided to each person who served as PEO during 
the year and combine those figures. This figure would constitute the 
registrant's annual total PEO compensation.
    Alternatively, a registrant may look to the PEO serving in that 
position on the date it selects to identify the median employee and 
annualize that PEO's compensation. For example, if the registrant 
chooses October 15 as the date to determine its median employee, the 
registrant would calculate the compensation of the person serving as 
PEO on that date and annualize that PEO's compensation. If the person 
was PEO for six months and received $100,000 of total compensation, the 
registrant would use $200,000 as the annual total compensation of its 
PEO. This approach is consistent with annualizing the total 
compensation of permanent employees, discussed below, which is 
permitted under the final rule. It is also similar to the approach 
suggested by one commenter.\227\
---------------------------------------------------------------------------

    \227\ See id.
---------------------------------------------------------------------------

    We are not adopting the approach advocated by some commenters of 
disclosing compensation information for the PEO holding that position 
at the end of the fiscal year. Section 953(b) requires the disclosure 
of the ``annual total compensation of the chief executive officer (or 
any equivalent position) of the issuer,'' and we think the better 
interpretation of that language is that it is intended to capture the 
annual compensation paid for that position (regardless of the 
individual who holds the position). Also, we believe that allowing the 
disclosure of only partial year compensation would fundamentally alter 
the pay ratio disclosure and would not capture the ratio of PEO 
compensation to median employee compensation.
f. Additional Information Is Permissible
i. Proposed Rule
    In the Proposing Release, we noted that we received some comments 
prior to the proposal suggesting that the rule should allow registrants 
to present separate pay ratios covering U.S. and non-U.S. employees to 
mitigate concerns that the comparison of the PEO to non-U.S. employees 
could distort the disclosure.\228\
---------------------------------------------------------------------------

    \228\ See, e.g., letters from AFL-CIO (Dec. 13, 2010) (``AFL-CIO 
pre-proposal letter''), Americans for Financial Reform (Mar. 23, 
2011) (``AFR pre-proposal letter''), Walden Asset Management (Apr. 
29, 2011) (``Walden pre-proposal letter''), and Social Investment 
Forum (Apr. 21, 2011) (``SIF pre-proposal letter'').
---------------------------------------------------------------------------

    The Proposing Release stated that we did not believe that it was 
necessary to include instructions in the new rule expressly permitting 
registrants to add disclosure to accompany the pay ratio. We indicated, 
that, as with other mandated disclosure under our rules, registrants 
would be permitted to supplement their required disclosure with a 
narrative discussion or additional ratios if they chose to do so. We 
indicated also that, as with other disclosure under our rules,\229\ any 
additional ratios should not be presented with greater prominence than 
the required pay ratio. We requested comment on whether the final rule 
should permit or require registrants to include two separate pay ratios 
covering U.S. employees and non-U.S. employees.
---------------------------------------------------------------------------

    \229\ See, e.g., Exchange Act Rule 12b-20 and Commission 
Guidance Regarding Management's Discussion and Analysis or Financial 
Condition and Results of Operations, Release No. 33-8350 (Dec. 19, 
2003) [68 FR 75056], at 75060.
---------------------------------------------------------------------------

ii. Comments on the Proposed Rule
    Some commenters opposed any requirement for two separate 
ratios.\230\ Other commenters also opposed a requirement for two 
separate ratios, but indicated that the final rule should permit 
registrants to provide separate ratios if they chose to do so.\231\ One 
commenter acknowledged that the Proposing Release permitted additional 
ratios but suggested that we should indicate that any additional ratios 
should not be misleading and not presented with greater prominence than 
the required ratio.\232\
---------------------------------------------------------------------------

    \230\ See, e.g., letters from American Benefits Council, Aon 
Hewitt, Frederic W. Cook & Co., and WorldatWork I.
    \231\ See, e.g., letters from ABA, Capital Strategies, CII, COEC 
I, CT State Treasurer, Cummings Foundation, Emergent, Hyster-Yale, 
Johnson & Johnson, NACCO, Comptroller of the City of New York (Nov. 
27, 2013) (``NYC Comptroller''), US SIF, and Vivient.
    \232\ See letter from E&Y.
---------------------------------------------------------------------------

iii. Final Rule
    After considering the comments, we have added an instruction to the 
final rule stating expressly that registrants may present additional 
ratios or other information to supplement the required ratio, but are 
not required to do so. The instruction states also that, if a 
registrant includes any additional ratios, the ratios must be clearly 
identified, not misleading, and not presented with greater prominence 
than the required ratio. Additional pay ratios are not limited to any 
particular information, such as pay ratios covering U.S. and non-U.S. 
employees.
g. Annualizing Permanent Employees is Permissible, but Other 
Compensation Adjustments are Prohibited
i. Proposed Rule
    The proposed rule included an instruction permitting, but not 
requiring, registrants to annualize the total compensation for 
permanent ``employees'' who did not work for the entire year, such as 
new hires, employees on leave under the Family and Medical Leave Act of 
1993,\233\ employees called for active military duty, or employees who 
took an unpaid

[[Page 50129]]

leave of absence during the period for another reason. We did not 
propose to require registrants to perform this type of adjustment, 
however, because we did not believe that the costs of requiring 
companies to make this extra calculation would be justified. The 
proposed rule applied to individuals who were employed on the last day 
of the fiscal year because it referred specifically to an ``employee.''
---------------------------------------------------------------------------

    \233\ 29 U.S.C. 2601 et seq.
---------------------------------------------------------------------------

    In addition, the proposed instruction would prohibit a registrant 
from annualizing some eligible employees and not others, and the 
instruction prohibited adjustments that would cause the ratio to not 
reflect the actual composition of the workforce, such as annualizing 
the compensation of seasonal or temporary employees. A registrant could 
annualize the compensation for a permanent part-time employee who had 
only worked a portion of the year (such as an employee who is 
permanently employed for three days a week and who took an unpaid leave 
of absence under the Family and Medical Leave Act for a portion of the 
year). In such a case, we explained that the adjustment should reflect 
compensation for the employee's part-time schedule over the entire 
year, but should not adjust the part-time schedule to a full-time 
equivalent schedule.
    Although we proposed to permit the annualizing adjustments 
described above, the proposed rule would not have permitted certain 
other adjustments or assumptions, such as full-time equivalent 
adjustments for part-time employees or annualizing adjustments for 
temporary or seasonal employees. We believed such adjustments would 
cause the median to not be reasonably representative of the 
registrant's actual employment and compensation arrangements for its 
workforce during the period and could diminish the potential usefulness 
of the disclosure.
ii. Comments on the Proposed Rule
    Some commenters concurred that registrants should be permitted to 
annualize the total compensation for all permanent employees (which 
would exclude temporary and seasonal positions) that were employed by 
the registrant for less than the full fiscal year.\234\ Some commenters 
also agreed that registrants should not be permitted to adjust the 
salaries of part-time, temporary, and seasonal employees to the 
equivalent of full-time status contending that it would be misleading 
to do so.\235\
---------------------------------------------------------------------------

    \234\ See, e.g., letters from ABA, Intel, NACCO, PM&P, SH&P, and 
WorldatWork I.
    \235\ See, e.g., letters from Domini, Capital Strategies, and CT 
State Treasurer (stating that substantial reliance on non-full-time 
employees is important for shareholders' understanding of a 
company's workforce and pay structure).
---------------------------------------------------------------------------

    Many commenters, however, contended that the final rule should 
permit registrants to provide full-time equivalent adjustments for the 
salaries of part-time, temporary, and seasonal employees.\236\ Some of 
these commenters asserted that not doing so would distort the pay 
ratio.\237\ One of these commenters asserted that permitting full-time 
equivalent adjustments would not undermine disclosure or make it less 
accurate because any potential concern about shareholders' 
understanding of the pay ratio would be mitigated by requiring 
registrants to disclose their full-time equivalency and the approximate 
number of part-time, seasonal, and temporary employees for which it 
made the calculation.\238\ Another commenter indicated that, if the 
final rule requires disclosure of any adjustment calculations, the 
disclosure should be limited to brief statements.\239\
---------------------------------------------------------------------------

    \236\ See, e.g., letters from ABA, ASA, Business Roundtable I, 
Chesapeake Utilities, COEC I, COEC II, Dover Corp., ExxonMobil, 
Fedewa, FEI, Frederic W. Cook & Co., Johnson & Johnson, LAPFF, 
McCoy, PM&P, SH&P, SHRM, TCA, and Vectren Corp.
    \237\ See, e.g., letters from ABA, ASA, Chesapeake Utilities, 
COEC I, and COEC II.
    \238\ See letter from ABA.
    \239\ See letter from Corporate Secretaries.
---------------------------------------------------------------------------

iii. Final Rule
    After considering the comments, we are adopting the final rule as 
proposed. Annualization and full-time equivalent adjustments are 
separate concepts. Annualization involves taking the compensation of an 
employee who worked for only part of the registrant's fiscal year and 
projecting that compensation as if the employee worked the full fiscal 
year at the schedule that the employee worked for the portion of the 
year the employee worked. Annualization is allowed under the rule for 
full-time and part-time employees who did not work for the registrant's 
full fiscal year for some reason, such as they were employees who were 
newly hired, on leave under the Family and Medical Leave Act of 1993, 
called for active military duty, or took an unpaid leave of absence 
during the period. Annualization is only allowed for permanent 
employees; it is not allowed under the final rule for seasonal or 
temporary employees. A full-time equivalent adjustment involves taking 
the compensation of a part-time employee and projecting what the 
employee would have made if the employee were employed on a full-time 
basis. Full-time equivalent adjustments are prohibited under the final 
rule under all circumstances.
    We are taking this approach in the final rule because we believe it 
most accurately captures the workforce and compensation practices that 
the registrant has chosen to employ. The limited ability to annualize a 
permanent full-time or part-time employee reflects the fact that these 
employees are a permanent part of the registrant's workforce despite 
having only worked for part of that particular year. In contrast, a 
temporary or seasonal employee is not a permanent part of the 
registrant's workforce. Full-time equivalent adjustments of these 
employees' compensation would reflect a different workforce composition 
and compensation structure than used by the registrant. To the extent a 
registrant believes that not making full-time equivalent adjustments 
for temporary or seasonal employees might not provide shareholders a 
complete understanding of the registrant's compensation practices as 
they exercise their say-on-pay votes, the registrant is permitted under 
the final rule to provide additional disclosure.
2. Identifying the Median Employee and Calculating Annual Total 
Compensation
a. Identifying the Median Employee
i. Once Every Three Years
(a) Proposed Rule
    The proposed rule required registrants to disclose the ``median of 
the annual total compensation of all employees of the registrant.'' 
\240\ The proposed rule defined ``annual total compensation'' to mean 
``total compensation for the registrant's last completed fiscal year,'' 
\241\ and ``employee'' to mean ``an individual employed by the 
registrant or any of its subsidiaries as of the last day of the 
registrant's last completed fiscal year.'' \242\ Therefore, the 
proposed rule suggested that registrants would need to undertake the 
full process of identifying anew the median employee for each completed 
fiscal year.
---------------------------------------------------------------------------

    \240\ See proposed Item 402(u)(1)(i) of Regulation S-K.
    \241\ See proposed Item 402(u)(2)(ii) of Regulation S-K.
    \242\ See proposed Item 402(u)(3) of Regulation S-K.
---------------------------------------------------------------------------

(b) Comments on the Proposed Rule
    A few commenters suggested that the final rule should allow 
registrants to undertake the full process of identifying the median 
employee periodically (such as once every three years) and use the

[[Page 50130]]

compensation of that same or a similarly situated employee during the 
intervening two years as long as no material or substantial changes to 
the workforce had occurred.\243\ These commenters indicated that, to 
remain consistent with Section 953(b), a registrant should be required 
to undertake the process to re-identify the median employee for any 
year in which it has experienced a change in its employee population 
which the registrant reasonably believes would result in a significant 
change in the pay ratio disclosure. One of these commenters suggested 
that the registrant should exercise ``requisite diligence'' annually to 
determine whether such a material change had occurred.\244\
---------------------------------------------------------------------------

    \243\ See, e.g., letters from ABA, COEC I, and COEC II.
    \244\ See letter from ABA.
---------------------------------------------------------------------------

(c) Final Rule
    The final rule allows a registrant to identify the median employee 
whose compensation will be used for the annual total compensation 
calculation once every three years unless there has been a change in 
its employee population or employee compensation arrangements that it 
reasonably believes would result in a significant change in the pay 
ratio disclosure. If there have been no changes that the registrant 
reasonably believes would significantly affect its pay ratio 
disclosure, the registrant must disclose that it is using the same 
median employee in its pay ratio calculation and describe briefly the 
basis for its reasonable belief. For example, the registrant could 
disclose that there has been no change in its employee population or 
employee compensation arrangements that it believes would significantly 
affect the pay ratio disclosure. If the registrant is using the same 
median employee, it must calculate that median employee's annual total 
compensation each year and use that figure to update its pay ratio 
disclosure each year.
    For example, the registrant is required to identify the median 
employee and calculate that median employee's annual total compensation 
in year one. In years two and three, however, the registrant may use 
that same median employee (or an employee whose compensation is 
substantially similar to the original median employee based on the 
compensation measure used to select that median employee, as discussed 
below) to re-calculate the annual total compensation for that employee 
without re-identifying the median employee as would otherwise be 
required under the final rule if it satisfies the above conditions.
    We believe this approach is appropriate because, as commenters 
noted, it is consistent with the requirements of Section 953(b) to 
provide annual pay ratio disclosure, while at the same time reducing 
registrants' costs and burdens of re-calculating the median employee 
more than once every three years unless there has been a change in the 
registrant's employee population or employee compensation arrangements 
that it reasonably believes would result in a significant change in the 
pay ratio disclosure.\245\ Also, we note that the final rule permits 
reasonable estimates in identifying the median employee. Permitting 
registrants to identify the median employee once every three years, 
absent a change in the registrant's employee population or employee 
compensation arrangements that it reasonably believes would result in a 
significant change in its pay ratio disclosure, is another way of 
allowing an estimate of the median employee in a situation where it is 
unlikely to result in a significant change to the pay ratio disclosure. 
Therefore, this provision will help to minimize burdens and costs while 
not significantly affecting registrants' pay ratios. We do not believe 
that allowing this flexibility will limit the usefulness of the pay 
ratio for shareholders as a data point in making their voting decisions 
on executive compensation under Section 951 of the Dodd-Frank Act.
---------------------------------------------------------------------------

    \245\ See, e.g., letters from ABA and COEC I.
---------------------------------------------------------------------------

    Further, we note that allowing a registrant in appropriate 
circumstances to go up to three years without engaging in the full 
process to re-identify the median employee is consistent with the outer 
limit established by Section 951 for a registrant's say-on-pay vote. In 
our view, just as registrants must have a new say-on-pay vote three 
years after the previous vote, we think it is reasonable and 
appropriate that they engage in the process of re-identifying the 
median employee no less frequently than every three years.
    Also, there may be situations in which there has been no change in 
a registrant's employee population or employee compensation 
arrangements but it is no longer appropriate for the registrant to use 
the median employee identified in year one as the median employee in 
years two or three because of a change in that employee's 
circumstances. In such a situation, the registrant may use another 
employee whose compensation is substantially similar to the original 
median employee based on the compensation measure used to select the 
median employee in year one. For example, if the median employee 
identified in year one is no longer employed by the registrant in years 
two or three or that employee's compensation significantly changed in 
years two and three (for example, a promotion that significantly 
increased his or her compensation), the registrant is permitted to 
identify its median employee in each of the following two years from 
among employees that had similar compensation to the median employee in 
year one. If no other employee has similar compensation, however, the 
registrant must re-identify the median employee as required under the 
final rule.
ii. Using Annual Total Compensation, Another Consistently Applied 
Compensation Measure, Statistical Sampling, Reasonable Estimates, or 
Other Reasonable Methods
(a) Proposed Rule
    In the Proposing Release, we noted that Congress specifically chose 
the ``median'' as the point of comparison for Section 953(b), rather 
than the average or some other measure. Therefore, the proposed rule 
required the median. Section 953(b) does not prescribe a methodology 
that must be used to identify the median. To allow the greatest degree 
of flexibility while remaining consistent with the statutory provision, 
the proposed rule did not specify any required calculation 
methodologies for identifying the median. Instead, we provided 
instructions and guidance designed to allow registrants to choose from 
several alternative methods to identify the median, so that they would 
be able to use the method that worked best for their own facts and 
circumstances.
    For instance, registrants would be able to provide the proposed 
disclosure using total compensation for each employee under Item 
402(c)(2)(x), statistical sampling, reasonable estimates, or the use of 
any consistently applied compensation measures to identify the median. 
Once the registrant identified the median employee based on the 
selected compensation measure applied to each employee in the sample, 
the registrant would calculate that employee's annual total 
compensation in accordance with Item 402(c)(2)(x) and disclose that 
amount as part of the pay ratio disclosure. The proposal did not 
prescribe what a reasonable estimate would entail because that would 
necessarily depend on the registrant's particular facts and 
circumstances. In addition, the proposed rule did not

[[Page 50131]]

prescribe specific estimation techniques or confidence levels for 
identifying the median employee because we believed that companies 
would be in the best position to determine what is reasonable in light 
of their own employee population and access to compensation data. We 
proposed to require registrants to briefly describe and consistently 
use the methodology and any material assumptions, adjustments, or 
estimates applied to identify the median employee, and for any 
estimated amounts to be clearly identified as such.
    We proposed this flexible approach because we believed that the 
most appropriate and cost effective methodology would necessarily 
depend on a registrant's particular facts and circumstances, including, 
among others, such variables as: The size and nature of the 
registrant's workforce; the complexity of its organization; the 
stratification of pay levels across the workforce; the types of 
compensation paid to employees; the extent that different currencies 
are involved; the number of tax and accounting regimes involved; and 
the number of payroll systems the registrant has and the degree of 
difficulty involved in integrating payroll systems to readily compile 
total compensation information for all employees. We believed that 
these likely are the same factors that would cause substantial 
variation in the costs of compliance.
    In the Proposing Release, we stated our belief that the proposed 
rule's flexibility would enable registrants to manage compliance costs 
more effectively. We also stated that, by allowing registrants to 
better manage costs, a flexible approach could mitigate, to some 
extent, any potential negative effects on competition arising from the 
mandated requirements. We recognized, however, that a flexible approach 
could increase uncertainty for registrants that would prefer more 
specificity on how to comply with the proposed rule, particularly for 
those registrants that do not use statistical analysis in the ordinary 
course of managing their businesses.
    In the Proposing Release, we offered guidance on two permissible 
methodologies under the proposal: (1) Statistical sampling and (2) use 
of a consistently applied compensation measure. The variance of 
underlying compensation distributions (that is, how widely employee 
compensation is spread out or distributed around the mean) could 
appreciably affect the sample size needed for reasonable statistical 
sampling. We conducted an analysis about sample size that we described 
in the Proposing Release. Our analysis used mean and median wage 
estimates from the U.S. Department of Labor's Bureau of Labor 
Statistics (``BLS'') at the 4-digit North American Industry 
Classification System (``NAICS'') industry level (290 industries) and 
assumed a lognormal wage distribution, a 95% confidence interval with 
0.5% margin of error. The analysis focused on the registrants that have 
a single business or geographical unit. The analysis also assumed that 
the sampling method would be a true random sampling because it would 
not be biased by region, occupation, rank, or other factor. In our 
analysis, the appropriate sample size for the registrants with a single 
business or geographical unit varied between 81 and 1,065 across 
industries, with the average estimated sample size close to 560.
    We acknowledged, however, that variation in the types of employees 
at a registrant across business units and geographical regions would 
add complexity to the sampling procedure. While we generally agreed 
that a relatively small sample size would be appropriate in some 
situations, a reasonable determination of sample size would ultimately 
depend on the underlying distribution of compensation data. We noted 
that reasonable estimates of the median for registrants with multiple 
business lines or geographical units could be arrived at through more 
than one statistical sampling approach. All approaches, however, would 
require drawing observations from each business or geographical unit 
with a reasonable assumption on each unit's compensation distribution 
and inferring the registrant's overall median based on the observations 
drawn. Certain cases may not easily generate confidence intervals 
around the estimates or prescribe the appropriate minimum sample size. 
As a result, compliance costs would vary across registrants according 
to the characteristics of their compensation distributions. 
Nevertheless, we concluded that permitting registrants to use 
statistical sampling could lead to a reduction in compliance costs as 
compared with other methods of identifying the median.
    Additionally, we noted that the identification of a median employee 
would not necessarily require a determination of exact compensation 
amounts for every employee included in the sample. A registrant could, 
rather than calculating exact compensation, identify the employees in 
the sample that have extremely low or extremely high pay that would 
fall completely on either end of the pay spectrum. Since identifying 
the median involves finding the employee in the middle, it might not be 
necessary to determine the exact compensation amounts for every 
employee paid more or less than the employee in the middle.
    In addition to statistical sampling, the Proposing Release also 
highlighted the use of a consistently applied compensation measure. We 
recognized concerns about expected compliance costs arising from the 
complexity of the ``total compensation'' calculation under Item 
402(c)(2)(x) and, in particular, the determination of total 
compensation in accordance with Item 402(c)(2)(x) for employees when 
identifying the median. To address these concerns, the proposed rule 
would allow companies to use total direct compensation (such as annual 
salary, hourly wages, and any other performance-based pay) or cash 
compensation to first identify a median employee and then calculate 
that median employee's annual total compensation in accordance with 
Item 402(c)(2)(x). We noted that this approach would provide a workable 
identification of the median for many registrants, and we expected that 
the costs of compliance would be reduced if registrants were permitted 
to identify the median using a less complex, more readily available 
figure, such as salary and wages, rather than total compensation as 
determined in accordance with Item 402(c)(2)(x).
    This approach could also reduce costs for registrants that are 
unable to use statistical sampling techniques, as registrants would be 
permitted to use a consistently-applied compensation measure to 
identify the median employee regardless of whether they use statistical 
sampling. Further, because of concerns that using cash compensation 
could be just as burdensome to calculate for registrants with multiple 
payroll systems in various countries, we did not propose to require 
companies to use a specific compensation measure like cash compensation 
or total direct compensation when they were identifying the median 
employee. Instead, we believed that registrants would be in the best 
position to select a compensation measure that was appropriate to their 
own facts and circumstances and that a consistently applied 
compensation measure would result in a reasonable estimate of a median 
employee at a substantially reduced cost. Therefore, the proposed rule 
permitted a registrant to identify a median employee based on any 
consistently applied compensation measure, such as information derived 
from its tax and/or payroll records, as long as the registrant briefly 
disclosed the measure that it used.

[[Page 50132]]

    We also recognized that the annual period used for payroll and/or 
tax recordkeeping could sometimes differ from the registrant's fiscal 
year. For purposes of calculating the annual total compensation amounts 
when using a consistently applied compensation measure, the proposed 
rule permitted registrants to use the same annual period that was used 
in the payroll and/or tax records from which the compensation amounts 
were derived. We did not propose to define or limit what would qualify 
as payroll and/or tax records. We noted, however, that the proposed 
accommodation was intended to be construed broadly enough to allow 
registrants to use information that they already tracked and compiled 
for payroll and/or tax purposes. We thought that permitting registrants 
to use compensation information in the form that it was maintained in 
their own books and records would reduce compliance costs without 
appreciably affecting the quality of the disclosure.
    Additionally, although our proposed flexible approach could reduce 
comparability of the pay ratio disclosure across registrants, we stated 
our belief that precise conformity or comparability of the ratio across 
companies was not necessary and indicated that a possible benefit of 
the pay ratio disclosure would be providing a company-specific metric 
that shareholders could use to evaluate the PEO's compensation within 
the context of his or her own company. Accordingly, we did not believe 
that improving the comparability of the disclosure across companies by 
mandating a specific method for identifying the median would be 
justified in light of the costs that would be imposed on registrants by 
a more prescriptive rule.
    Also, even assuming the benefits of comparability across 
registrants as a desirable goal, we did not believe that mandating a 
particular methodology would necessarily improve comparability because 
of the numerous other factors that could also cause the ratios to be 
less meaningful for company-to-company comparison, such as differences 
in industry and business type; variations in the way companies organize 
their workforces to accomplish similar tasks; differences in the 
geographical distribution of employees (domestic or international, as 
well as in high- or low-cost areas); degree of vertical integration; 
reliance on contract and outsourced workers; and ownership structure. 
We also note that some commenters asserted that disclosing the pay 
ratio could potentially increase the likelihood that a registrant's 
competitors could infer proprietary or sensitive information about the 
registrant's business, which could cause a competitive disadvantage for 
registrants.\246\
---------------------------------------------------------------------------

    \246\ See, e.g., letters from Huan Lou (Dec. 2, 2013) (``Lou'') 
and Prof. Ray.
---------------------------------------------------------------------------

    Finally, we recognized that allowing registrants to select a 
methodology for identifying the median, including identifying the 
median employee based on any consistently applied compensation measure 
and allowing the use of reasonable estimates, rather than prescribing a 
methodology or set of methodologies, could potentially permit a 
registrant to alter the reported ratio to achieve a particular 
objective with the pay ratio disclosure, thereby potentially reducing 
the usefulness of the information. But, as we explained, we believed 
that requiring the use of a consistently applied compensation measure 
should lessen this concern.
(b) Comments on the Proposed Rule
(i) Flexibility
    A large number of commenters indicated that they supported the 
flexibility permitted in the proposed rule generally, or more 
specifically supported the flexibility of the proposed rule in 
permitting registrants to choose a methodology for calculating the 
median.\247\ Some commenters contended that the flexibility would 
lessen the costs and burdens of the proposed rule without reducing the 
rule's benefits.\248\ A few commenters suggested that the proposed 
rule's flexibility would not undermine the rule and would be consistent 
with the directives of Section 953(b).\249\ Other commenters indicated 
that, although the proposed rule's flexibility might alter pay ratio 
disclosures, any distortion would be minimal.\250\ One commenter 
claimed that the proposed rule was not flexible enough because 
Instruction 2(iii) to Item 402(u) would permit a registrant to identify 
the median employee only using ``compensation'' measures, whereas the 
commenter advocated for use of a ``job-level'' measure.\251\ Despite 
the permitted flexibility, another commenter remained concerned about 
the complexity of finding the median employee.\252\
---------------------------------------------------------------------------

    \247\ See, e.g., letters from Prof. Angel, Aon Hewitt, 
B[acirc]tirente et al., Best Buy et al., CalSTRS, Capital 
Strategies, CEG, Chicago Teachers Fund, Corporate Secretaries, 
Cummings Foundation, Cummins Inc., CUPE, Davis Polk, E&Y, First 
Affirmative, Freeport-McMoRan, FS FTQ, ICCR, IL Bricklayers and 
Craftworkers Union, Intel, Johnson & Johnson, Marco Consulting, 
McMorgan & Co., Mercer I, Meridian, Microsoft, Novara Tesija, NSFM, 
NY Bricklayers and Craftworkers Union, NYC Bar, OCP, Oxfam, PGGM, 
Public Citizen I, Quintave, Rep. Ellison et al. II, SEIU, Sen. 
Menendez et al. II, Socially Responsive Financial Advisors, 
Teamsters, Trillium I, Trustee Campbell, UAW Trust, Vectren Corp., 
WA State Investment Board, and WorldatWork I.
    \248\ See, e.g., letters from ABA, AFSCME, Barnard, Bricklayers 
International, CII, LIUNA, Fowler, Mirczak, PNC Financial Services, 
Sen. Menendez et al. II, US SIF, Vivient, and Walden.
    \249\ See, e.g., letters from Domini and PM&P.
    \250\ See, e.g., letters from Capital Strategies and WorldatWork 
I.
    \251\ See letter from Towers Watson. According to the commenter, 
in a job-level approach, a global company assigns all jobs in its 
organization to one of X number of job levels (determined by how the 
particular job role contributes to the organization based on tasks, 
skills, expertise, leadership, functional strategy and business 
strategy). The company would readily be able to determine the job or 
job family that would fall at the median of skill levels within the 
company based on a ratio that compares the total number of employees 
at each job level to the total employee population. The job or job 
family in the median of skill levels would be where the median-level 
employee resides and, once it is determined that the median employee 
resides at a particular level, the company would then be able to 
apply a statistical sampling approach to that job level, taking into 
account compensation earned in different locations or countries, to 
further reduce the number of payroll files that will need to be 
examined. The commenter indicated that more details of this approach 
can be found at: http://www.towerswatson.com/en/Services/Tools/job-leveling-global-grading-and-career-map.
    \252\ See letter from Vectren Corp.
---------------------------------------------------------------------------

    One commenter contended that a registrant should be permitted to 
develop its own methodology for identifying their median employee to 
mitigate costs \253\ if the rule required the registrant to accurately 
describe its methodology, perform consistent calculations each year, 
disclose when and how it chose to deviate from the prior year's 
methodology, select the methodology in good faith, and make reasonable 
assumptions, adjustments, and estimates. Some commenters recommended 
that, once a registrant chooses its methodology for identifying the 
median employee, the registrant should be required to use that 
methodology going forward.\254\ Other commenters supported the proposed 
rule's flexibility in allowing registrants to choose a methodology for 
identifying the median employee, provided that registrants are required 
to disclose the methodologies they used.\255\ One commenter suggested 
that the final rule provide some type of check generally on a 
registrant's methodology.\256\
---------------------------------------------------------------------------

    \253\ See letter from ABA.
    \254\ See, e.g., letters from PGGM and RPMI.
    \255\ See, e.g., letters from Australian Council of 
Superannuation Investors (Dec. 2, 2013) (``ACSI'') and IPS.
    \256\ See letter from Mirczak.
---------------------------------------------------------------------------

    Some commenters were opposed to the proposed rule's flexibility. 
Most of these commenters were individuals who claimed that the proposed 
rule's flexibility would allow registrants to reduce the ratio in 
inappropriate

[[Page 50133]]

ways.\257\ Other commenters contended that the permitted flexibility 
would decrease the ratio's utility, especially for comparing the ratios 
of different companies.\258\ A few commenters maintained that the 
proposed rule would still lead to high costs, even taking into account 
its flexibility.\259\
---------------------------------------------------------------------------

    \257\ See, e.g., letters from Bupp, Corayer, Fedewa, Fox, 
Friend, Grotzke, Hlodnicki, Kizzort, Maly, Petricoin, and Van Pelt.
    \258\ See, e.g., letters from Amundi, BCIMC, Ciatto, Glenn, IBC, 
Prof. Muth, and NIRI.
    \259\ See, e.g., letters from ABA (stating that registrants will 
still incur significant costs even with the ability to select a 
methodology) and FSR.
---------------------------------------------------------------------------

    Some commenters recommended that the final rule include a safe 
harbor for identifying the median employee to minimize burdens, provide 
greater comparability, and limit liability.\260\ One commenter 
encouraged us to retain the proposed rule's flexibility in allowing a 
registrant to remove some percentage of the distribution from both the 
high and low ends of an employee sample because this would not distort 
the median employee determination while reducing costs.\261\
---------------------------------------------------------------------------

    \260\ See, e.g., letters from NSFM, OCP, PM&P, Quintave, and 
SHRM.
    \261\ See letter from E&Y.
---------------------------------------------------------------------------

    Only a few commenters commented specifically on using reasonable 
estimates for identifying the median employee. One commenter declared 
that the final rule should not permit any estimates at all.\262\ Other 
commenters contended that the final rule should permit reasonable 
estimates.\263\ One of these commenters noted that permitting 
reasonable estimates would mitigate the rule's costs while not 
``materially'' impacting its usefulness.\264\ Also, some commenters 
recommended that the final rule not specify any requirements, guidance, 
or safe harbors regarding the estimates.\265\ Several commenters 
asserted that the final rule should not provide guidance regarding 
assumptions about error rates or confidence levels.\266\ One of these 
commenters expressed concern that, if the final rule provided such 
guidance, the assumptions would become de facto requirements.\267\
---------------------------------------------------------------------------

    \262\ See letter from Dennis T. (Nov. 19, 2013) (``Dennis T'').
    \263\ See, e.g., letters from ABA, Capital Strategies, Davis 
Polk, Hyster-Yale, Johnson & Johnson, NACCO, and WorldatWork I.
    \264\ See letter from ABA.
    \265\ See, e.g., letters from ABA (noting that further guidance 
is not needed because there is already sufficient deterrence for 
unreasonable estimates with the principles-based disclosure 
framework and anti-fraud provisions), Capital Strategies, Hyster-
Yale, Johnson & Johnson, and NACCO.
    \266\ See, e.g., letters from Prof. Angel, Hyster-Yale, and 
NACCO.
    \267\ See letter from Prof. Angel.
---------------------------------------------------------------------------

    A number of commenters provided suggestions on the methodologies a 
registrant should be permitted to use in identifying the median 
employee. A few commenters stated that the final rule should provide 
more explicit guidance on what ``other reasonable methods'' for 
identifying the median employee are available.\268\ One of these 
commenters suggested allowing the following methods: (1) Specific safe 
harbor assumptions about the statistical distribution of compensation 
within the company and its business units; (2) formulaic, numerical, 
and other computational approaches to estimate the median compensation; 
and (3) disclosure of a reasonable range of outcomes rather than 
requiring the ``right'' outcome.\269\
---------------------------------------------------------------------------

    \268\ See, e.g., letters from E&Y and TCA.
    \269\ See letter from TCA.
---------------------------------------------------------------------------

    Additionally, other commenters provided specific suggestions for 
methods that we should consider ``reasonable'' in the final rule. One 
commenter requested that the final rule allow registrants to identify 
the median employee based on rates of pay or compensation schedules 
applicable to classes of employees instead of pay actually earned over 
the course of the year.\270\ Another commenter suggested that, where a 
registrant has an even number of employees and, therefore, is unable to 
select one median employee, the registrant should be permitted to 
select and disclose an average of the compensation of the two employees 
nearest the median.\271\ Finally, one commenter indicated that a 
registrant's methodology should be based on a ``good faith compliance'' 
standard that is akin to that type of standard in our proposed 
crowdfunding rules.\272\
---------------------------------------------------------------------------

    \270\ See letter from RILA (``In order to reduce the cost of 
using payroll or other data and annualizing information for 
employees such as new hires who are employed for less than an entire 
fiscal or other year, we would propose that registrants be permitted 
as an alternative to determine the median employee based upon 
employee rate of pay on the measurement date. For some issuers, 
identifying the median employee based upon rates of pay, rather than 
pay earned over the course of a year, will reduce the burden and 
minimize the skewing effects on the ratio of a large number of part-
time, temporary and seasonal employees. Provided that the method is 
disclosed, the final rule should give issuers flexibility in this 
regard.'').
    \271\ See letter from SH&P.
    \272\ See letter from FSR. See also Crowdfunding, Release No. 
33-9470 (Oct. 23, 2013) [78 FR 66427].
---------------------------------------------------------------------------

(ii) Statistical Sampling
    We received many comments on using statistical sampling for 
identifying the median employee, with a majority of these commenters 
supporting the use of statistical sampling, as permitted in the 
proposed rule. Many of these commenters suggested that allowing the use 
of statistical sampling would reduce the costs for registrants, without 
specifically quantifying a reduction.\273\ Some commenters that 
supported statistical sampling recommended that the final rule not 
specify requirements for statistical sampling (such as appropriate 
sample size, confidence levels, or other requirements).\274\ One of 
these commenters contended that specifying requirements for statistical 
sampling would unduly constrain registrants from developing the most 
appropriate methodology and would be inconsistent with 
flexibility.\275\
---------------------------------------------------------------------------

    \273\ See, e.g., letters from AFL-CIO I, AFSCME, Prof. Angel, 
Bricklayers International, Calvert, Chesapeake Utilities, Chicago 
Teachers Fund, CUPE, Davis Polk, First Affirmative, FS FTQ, FSI, 
ICCR, IL Bricklayers and Craftworkers Union, LIUNA, Marco 
Consulting, McMorgan & Co., Meridian, NACD, Novara Tesija, NRF, NY 
Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, Rep. 
Ellison et al. I; Socially Responsive Financial Advisors, TCA, 
Teamsters, Trillium I, Trustee Campbell, US SIF, and WA State 
Investment Board.
    \274\ See, e.g., letters from ABA, IBC, Johnson & Johnson, and 
WorldatWork I.
    \275\ See letter from ABA.
---------------------------------------------------------------------------

    Another commenter asserted that the statute permits statistical 
sampling because Section 953(b) does not prescribe a particular way to 
identify the median employee.\276\ Some commenters stated that 
registrants would likely use statistical sampling in identifying their 
median employee,\277\ and that statistical sampling is feasible.\278\ A 
few commenters agreed with the proposal that registrants should be 
allowed to identify the median employee by using statistical sampling 
based on a definition of compensation other than ``annual total 
compensation'' under Item 402.\279\ Some commenters indicated that the 
final rule should include a safe harbor for statistical sampling.\280\
---------------------------------------------------------------------------

    \276\ See letter from Michael Ohlrogge, Stanford Law School and 
Stanford Department of Management Science and Engineering (Sep. 25, 
2013) (``Ohlrogge I'').
    \277\ See, e.g., letters from Capital Strategies and Mercer I.
    \278\ See, e.g., letters from Ohlrogge I, Michael Ohlrogge, 
Stanford Law School and Stanford Department of Management Science 
and Engineering (Jul. 5, 2015) (``Ohlrogge II''), and Mike Petty 
(Oct. 21, 2013) (``M. Petty'').
    \279\ See, e.g., letters from Emergent and McGuireWoods.
    \280\ See, e.g., letters from American Benefits Council; E&Y; 
FSR; Dr. Sandy J. Miles, Professor of Human Resource Management, 
Murray State University (Nov. 29, 2013) (``Prof. Miles''); Strus and 
Associates Inc. (Nov. 30, 2013) (``Strus and Assoc.''); and TCA.
---------------------------------------------------------------------------

    Although the majority of commenters that discussed statistical 
sampling supported its use in identifying the median employee, one 
commenter stated specifically that the final rule should discourage the 
use of statistical sampling in favor of information

[[Page 50134]]

derived from tax and/or payroll records to determine actual employee 
pay rather than an estimated amount.\281\ Other commenters, while not 
necessarily opposed to statistical sampling, contended that it would 
not mitigate costs of collecting and assembling employee compensation 
data, which, in their view, is the most expensive part of the 
rule.\282\
---------------------------------------------------------------------------

    \281\ See letter from LAPFF.
    \282\ See, e.g., letters from Business Roundtable I, Chamber I, 
COEC I, COEC II, ExxonMobil, Freeport-McMoRan, FuelCell Energy, 
Garmin, Johnson & Johnson, Microsoft, NAM I, NAM II, NIRI, and 
WorldatWork I.
---------------------------------------------------------------------------

(iii) Consistently Applied Compensation Measures
    Most commenters that discussed using consistently applied 
compensation measures, such as information derived from tax and/or 
payroll records, to identify the median employee agreed with the 
proposed rule's approach.\283\ Generally, these commenters contended 
that permitting the use of such measures would reduce costs while not 
impairing the pay ratio's usefulness. For example, one commenter noted 
that, while using a consistently applied compensation measure may 
exclude benefits, perquisites, and other allowances, it would still 
capture salary, incentive cash earned, and stock awards.\284\ 
Therefore, the measure would include ``the substantial majority of 
compensation and [would] not lead to distortion of the median.'' 
Another commenter asserted that the final rule should be flexible 
enough to encompass different approaches across jurisdictions.\285\
---------------------------------------------------------------------------

    \283\ See, e.g., letters from ABA, AFL-CIO I, Calvert, 
Chesapeake Utilities, Chicago Teachers Fund, CII, First Affirmative, 
FS FTQ, ICCR, IL Bricklayers and Craftworkers Union, Johnson & 
Johnson, Marco Consulting, Meridian, Microsoft, Novara Tesija, NY 
Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, SH&P, 
WA State Investment Board, and WorldatWork I.
    \284\ See letter from Microsoft.
    \285\ See letter from ABA.
---------------------------------------------------------------------------

    One commenter disagreed with permitting the use of consistently 
applied compensation measures on the basis that it would alter the pay 
ratio because not all compensation would be included.\286\ Some 
commenters noted that using consistently applied compensation measures 
would not reduce the costs for registrants with non-U.S. 
employees.\287\ Therefore, several of these commenters recommended that 
the final rule should allow registrants to use one or more comparable 
consistently applied compensation measures and not be limited to a 
single consistently applied compensation measure for employees in 
different international jurisdictions to reduce registrants' costs and 
burdens.\288\
---------------------------------------------------------------------------

    \286\ See letter from Tumeh.
    \287\ See, e.g., letters from Corporate Secretaries, Davis Polk, 
and FSR.
    \288\ See, e.g., letters from Corporate Secretaries (``For 
example, a company may be able to use W-2 or payroll information to 
relatively efficiently and accurately quantify annual cash 
compensation for US employees, but this data could be impracticable 
to replicate for certain non-US employees due to privacy concerns, 
comparability of compensation schemes, tax systems, or otherwise. . 
. . In light of this, we propose that the Commission consider 
including in the final rules that for employees based in non-US 
jurisdictions, registrants be permitted to use (i) the same 
compensation measure used in the US or (ii) another reasonably 
comparable compensation measure for any non-US jurisdiction where 
the same compensation measure is not used, unless a different 
application is required or compensation cannot be estimated in a 
particular jurisdiction for a particular reason (i.e., data privacy 
laws).'') and Davis Polk (``For example, a registrant using cash 
compensation as its consistently applied compensation measure for 
purposes of determining its median employee may intend to include 
only base salary and bonus amounts for its U.S. employees, but may 
find it appropriate to also include other benefits commonly 
considered to be part of base compensation for its non-U.S. 
employees, such as meal stipends or automobile allowances. We 
believe that it would be appropriate for registrants to use these 
reasonably comparable measures, even if not the exact same measure, 
across different employee populations in order to provide for more 
meaningful comparisons across the varying compensation structures in 
international jurisdictions and to reduce unnecessary burdens on 
registrants.'').
---------------------------------------------------------------------------

(iv) The ``Median'' Employee
    A number of commenters recommended that the rule use compensation 
of an employee, or employees, other than the median employee as is 
required in Section 953(b). A few of these commenters suggested that, 
instead of using their median employee in their pay ratio, registrants 
should be permitted to use their ``average'' employee.\289\ The 
commenters contended that using ``average'' instead of ``median'' would 
reduce costs, would be better understood by the public, and could be 
calculated easily using tax records.\290\ One of these commenters 
acknowledged that the statutory language of Section 953(b) uses the 
word ``median'' rather than ``average,'' but pointed out that the 
Proposing Release quotes from two letters submitted by members of 
Congress, including two co-sponsors of Section 953(b), in which the 
members refer to the ``average'' and ``typical'' employee instead of 
the ``median'' employee when discussing the statute.\291\ Also, the 
commenter noted that we stated in the Proposing Release that ``Section 
953(b) does not expressly set forth a methodology that must be used to 
identify the median, nor does it mandate that [we] must do so in [our] 
rules.'' \292\
---------------------------------------------------------------------------

    \289\ See, e.g., letters from Hyster-Yale, Mercer I, NACCO, and 
Powers.
    \290\ See, e.g., letters from Hyster-Yale and NACCO. But see 
letter from Suzanne Laatsch (Nov. 29, 2013) (``Laatsch'') (asserting 
that using the median employee, as opposed to using the average 
employee, would actually reduce costs).
    \291\ See letter from NACCO.
    \292\ See letter from NACCO.
---------------------------------------------------------------------------

    Other commenters recommended that the final rule allow registrants 
to use existing BLS data to calculate their pay ratios, which is based 
on average employee compensation figures.\293\ One of these commenters 
contended that allowing registrants to use existing data sources, 
including BLS data, would be faithful to the intent of Section 953(b) 
and not be ``materially'' different than using the median, even though 
it would be less precise.\294\
---------------------------------------------------------------------------

    \293\ See, e.g., letters from Brian Foley & Co., NACD, and NIRI.
    \294\ See letter from NIRI. But see letter from IPS (disagreeing 
specifically with the NIRI letter and asserting that the final rule 
should not permit registrants to use BLS data).
---------------------------------------------------------------------------

    A few commenters contended that registrants should be permitted to 
use the compensation of a range of employees as the median compensation 
instead of the compensation of the median employee.\295\ One of these 
commenters noted that it would be unlikely for a registrant, using a 
convenient and cost-effective measure, to determine a single employee 
that is the median.\296\ According to this commenter, it would be more 
likely that the registrant's calculations would yield a group of 
employees, any of whom who could serve as the median employee. 
Therefore, once the range of employees is determined, registrants 
should be permitted to use any reasonable method to determine which 
employee to use as the median employee.
---------------------------------------------------------------------------

    \295\ See, e.g., letters from Capital Strategies, Hyster-Yale, 
Mercer I, and NACCO.
    \296\ See letter from NYC Bar.
---------------------------------------------------------------------------

    Similarly, one commenter recommended that the final rule permit 
registrants, after identifying the median employee using whatever 
methodology they select, to use another employee as the median employee 
if that employee is within a 1% variance of the median and the original 
employee has anomalous compensation characteristics that would create 
the risk of a distorted pay ratio.\297\ The commenter recognized that 
Section 953(b) refers to the ratio of the ``median'' employee's annual 
compensation to the compensation of the PEO but contended that some 
deviation from that precise statutory language should be acceptable if 
it furthers the statute's intent to show the

[[Page 50135]]

ratio of the compensation of the typical or representative employee to 
that of the PEO.
---------------------------------------------------------------------------

    \297\ See letter from PNC Financial Services.
---------------------------------------------------------------------------

    Another commenter advocated using one of two alternative approaches 
based on whether the registrant has international employees that would 
segregate employees with similar positions into different groups. 
Registrants could identify in which group the median resides based on a 
ratio that compares the total number of employees at each job level to 
the total employee population and use sampling or another technique to 
identify the median of that group, which the registrant would use as 
its median for pay ratio purposes.\298\
---------------------------------------------------------------------------

    \298\ See letter from Tower Watson. See also letter from Mercer 
I (advocating a somewhat similar approach using multiple statistical 
samples).
---------------------------------------------------------------------------

(c) Final Rule
    We are adopting the final rule as proposed. Consistent with the 
proposal, the final rule does not specify any required methodology for 
registrants to use in identifying the median employee. Instead, the 
final rule permits registrants the flexibility to choose a method to 
identify the median employee based on their own facts and 
circumstances. To identify the median employee, registrants may use a 
methodology that uses reasonable estimates. The median employee may be 
identified using annual total compensation or any other compensation 
measure that is consistently applied to all employees included in the 
calculation, such as information derived from tax and/or payroll 
records. Also, in determining the employees from which the median is 
identified, a registrant is permitted to use its employee population or 
statistical sampling and/or other reasonable methods. In any event, the 
final rule requires a registrant to briefly describe the methodology it 
used to identify the median employee and any material assumptions, 
adjustments (including any cost-of-living adjustments), or estimates it 
used to identify the median employee or to determine total compensation 
or any elements of total compensation, which shall be consistently 
applied. The registrant also must clearly identify any estimates used.
(i) Flexibility
    As we noted in the Proposing Release, we believe that allowing 
registrants the flexibility to choose a method that works best for 
their particular facts and circumstances will help them comply with the 
rule in a relatively cost-efficient manner while still fulfilling the 
purpose of Section 953(b). We recognize that a flexible approach could 
increase uncertainty for registrants that prefer more specificity on 
how to comply with the final rule, particularly for those registrants 
that do not use statistical analysis in the ordinary course of managing 
their businesses. We believe that any negative effects caused by any 
uncertainty would be offset by the positive effects of permitting 
flexibility. Also, the final rule establishes certain methodologies and 
permissible uses of estimates, such that registrants may use reasonable 
estimates both in the methodology used to identify the median employee 
and in calculating annual total compensation or any elements of total 
compensation for employees other than the PEO; use their employee 
population, statistical sampling, or another reasonable methods in 
determining the employees from which the median employee is identified, 
and identify the median employee using annual total compensation or any 
other compensation measure that is consistently applied to all 
employees included in the calculation.
    We believe also that any uncertainty provided by the final rule's 
flexibility is offset by other benefits. Particularly, as we noted in 
the Proposing Release, the final rule's flexibility allows registrants 
to provide the required disclosure in a relatively cost-efficient 
manner based on the registrant's own facts and circumstances using 
total compensation for each employee under Item 402(c)(2)(x), 
statistical sampling, reasonable estimates, or the use of any 
consistently applied compensation measures to identify the median. A 
large number of commenters supported this flexibility because it would 
reduce the rule's costs without significantly diminishing its 
benefits.\299\ In this regard, we do not believe permitting this 
flexibility will limit the usefulness of the pay ratio for shareholders 
as a data point in making their voting decisions on executive 
compensation under Section 951 of the Dodd-Frank Act.
---------------------------------------------------------------------------

    \299\ See, e.g., letters from ABA, AFSCME, Barnard, Bricklayers 
International, Capital Strategies, CII, Domini, LIUNA, Fowler, 
Mirczak, PM&P, PNC Financial Services, US SIF, Vivient, Walden, and 
WorldatWork I.
---------------------------------------------------------------------------

    In determining their methodology, registrants may consider, among 
other factors, such variables as: The size and nature of the workforce; 
the complexity of the organization; the stratification of pay levels 
across the workforce; the types of compensation the employees receive; 
the extent that different currencies are involved; the number of tax 
and accounting regimes involved; and the number of payroll systems the 
registrant has and the degree of difficulty involved in integrating 
payroll systems to readily compile total compensation information for 
all employees. These likely are the same factors that could cause 
substantial variation in the costs of compliance, but the final rule's 
flexibility should help registrants reduce these costs.
    As part of the flexibility permitted by the final rule, registrants 
may use a methodology that uses reasonable estimates in identifying the 
median employee and calculating the annual total compensation or any 
elements of compensation for employees other than the PEO, as proposed. 
Most commenters on this issue agreed that the final rule should permit 
reasonable estimates to mitigate the rule's costs.\300\ Some of these 
commenters suggested that the final rule should not include further 
requirements or guidance for making reasonable estimates, including 
safe harbors, assumptions for error rates, or confidence levels.\301\ 
We are not persuaded that further guidance is necessary.
---------------------------------------------------------------------------

    \300\ See, e.g., letters from ABA, Capital Strategies, Davis 
Polk, Hyster-Yale, Johnson & Johnson, NACCO, and WorldatWork I.
    \301\ See, e.g., letters from ABA, Capital Strategies, Hyster-
Yale, Johnson & Johnson, and NACCO. But see letters from NSFM, OCP, 
PM&P, Quintave, and SHRM (recommending that the final rule include a 
safe harbor for identifying the median employee).
---------------------------------------------------------------------------

    Further, as proposed, in determining the employees from which the 
median is identified, the final rule permits registrants to use ``other 
reasonable methods'' in addition to using its employee population or 
statistical sampling. Some commenters provided suggestions on the 
``other reasonable methods'' a registrant should be permitted to use in 
identifying the median employee.\302\ We are not specifying the ``other 
reasonable methods'' that may be appropriate because we seek to allow 
each registrant

[[Page 50136]]

the flexibility to determine the method that best suits its own facts 
and circumstances, which may include some of the suggestions made by 
these commenters.
---------------------------------------------------------------------------

    \302\ See, e.g., letters from E&Y (``In the adopting release, we 
believe the Commission should identify additional methods that, if 
used, would satisfy the requirements for determining which employees 
should be included in the analysis.''), FSR (indicating that an 
issuer's methodology should be based on a ``good faith compliance'' 
standard that is akin to our proposed crowdfunding rules), RILA 
(requesting that the final rule allow issuers to identify the median 
employee based on rates of pay instead of pay earned over the course 
of the year), TCA (suggesting that final rule allow the following 
methods: (1) Specific safe harbor assumptions about the statistical 
distribution of compensation within the company and its business 
units; (2) formulaic, numerical, and other computational approaches 
to estimate the median compensation; and (3) disclosure of a 
reasonable range of outcomes rather than requiring the ``right'' 
outcome), and Towers Watson (recommending that the final rule permit 
registrants to ``employ a methodology using salary grades or job 
levels, as appropriate, to reasonable identify the median'' 
(emphasis in original)).
---------------------------------------------------------------------------

(ii) Statistical Sampling
    Consistent with the proposal, the final rule permits registrants to 
use statistical sampling in determining the employees from which the 
median is identified. We believe permitting statistical sampling is 
appropriate under Section 953(b). Statistical sampling is based on 
statistical theory to make sampling more efficient. For example, with 
large populations, results accurate enough to be useful can be obtained 
from samples that represent only a small fraction of the 
population.\303\ We note that one commenter recommended that we 
discourage statistical sampling in favor of using the information 
derived from tax and/or payroll records to determine actual employee 
pay rather than allowing registrants to use an estimation.\304\ We 
believe, however, that statistical sampling can be a reliable means of 
identifying the median employee.
---------------------------------------------------------------------------

    \303\ See generally Cochran, W. G. (1977), Sampling Techniques, 
3rd Edition, New York: Wiley. Different statistical sampling methods 
have been developed and have been well established in literature and 
practice since 1786. See Laplace, P.S. (1786). ``Sur Les Naissances, 
Les Mariages Et Les Morts,'' In Histoire de L'Academie Royale des 
Sciences, 1783, Paris, 693-702; Stephan, F. F. (1948). History of 
the uses of modern sampling procedures. Journal of the American 
Statistical Association 43:12-37; Godambe, V. P. (1954). A unified 
theory of sampling from finite populations. Journal of the Royal 
Statistical Society 17(2):269-278; Sukhatme, P. V. (1966). Major 
developments in sampling theory and practice. Research papers in 
statistics. F.N. David Ed. John Wiley & Sons. 367-409pp; Lohr, S.L. 
(2009), Sampling: Design and Analysis, 2nd Edition, Cengage 
Learning; and Rao, J.N.K. (2011). Impact of frequentist and Bayesian 
methods on survey sampling practice: a selective appraisal. 
Statistical Science 26(2): 240-256pp.
    \304\ See letter from LAPFF.
---------------------------------------------------------------------------

    Some commenters indicated that permitting statistical sampling in 
the final rule would not mitigate costs because the final rule would 
still require registrants to collect and assemble employee compensation 
data, which the commenters view as the most expensive part of the 
rule.\305\ We believe, however, that permitting registrants to use 
statistical sampling could lead to a reduction in compliance costs as 
compared with other methods of identifying the median employee without 
significantly affecting the pay ratio because registrants are not 
required to calculate the total compensation for each of their 
employees. Therefore, although the final rule still requires 
registrants to collect and assemble employee compensation data, the 
availability of statistical sampling may allow them to assemble far 
less employee compensation data than if the final rule prohibited such 
sampling. We note that a number of other commenters indicated that 
permitting statistical sampling would reduce costs.\306\ Also, because 
of the reliability of the result achieved through appropriately 
conducted statistical sampling, we do not believe the use of sampling 
will limit the usefulness of the pay ratio for shareholders as a data 
point in making their voting decisions on executive compensation under 
Section 951 of the Dodd-Frank Act.
---------------------------------------------------------------------------

    \305\ See, e.g., letters from Business Roundtable I, Chamber I, 
COEC I, ExxonMobil, Freeport-McMoRan, FuelCell Energy, Garmin, 
Johnson & Johnson, Microsoft, NAM I, NAM II, NIRI, and WorldatWork 
I.
    \306\ See, e.g., letters from AFL-CIO I, AFSCME, Bricklayers 
International, Calvert, Chesapeake Utilities, Chicago Teachers Fund, 
CUPE, Davis Polk, Dennis T, First Affirmative, FS FTQ, FSI, ICCR, IL 
Bricklayers and Craftworkers Union, LIUNA, Marco Consulting, 
McMorgan & Co., Meridian, NACD, Novara Tesija, NRF, NY Bricklayers 
and Craftworkers Union, Oxfam, Public Citizen I, Rep. Ellison et al. 
I, Socially Responsive Financial Advisors, TCA, Teamsters, Trillium 
I, Trustee Campbell, Prof. Angel, US SIF, and WA State Investment 
Board.
---------------------------------------------------------------------------

    The final rule does not provide specific parameters for statistical 
sampling, including the appropriate sample size. We agree with 
commenters that specifying requirements for statistical sampling, 
including appropriate sample sizes, confidence levels, or other 
requirements would unduly constrain registrants from developing the 
most appropriate methodology.\307\ Instead, we believe registrants must 
make their own determinations on what is appropriate based on their own 
facts and circumstances.
---------------------------------------------------------------------------

    \307\ See, e.g., letters from ABA, IBC, Johnson & Johnson, and 
WorldatWork I.
---------------------------------------------------------------------------

    We are, however, providing some guidance for registrants when using 
statistical sampling. In this regard, based on the analysis we 
described in the Proposing Release, we believe that a relatively small 
sample size may be appropriate in certain situations.\308\ A reasonable 
determination of sample size ultimately depends on the underlying 
distribution of compensation data. Further, we believe that reasonable 
estimates of the median for registrants with multiple business lines or 
geographical units may be determined using more than one statistical 
sampling approach. Additionally, all statistical sampling approaches 
should draw observations from each business or geographical unit with a 
reasonable assumption on each unit's compensation distribution and 
infer the registrant's overall median based on the observations drawn.
---------------------------------------------------------------------------

    \308\ In that analysis, we determined that the appropriate 
sample size for the registrants with a single business or 
geographical unit varied between 81 and 1,065 employees across 
industries, with the average estimated sample size close to 506.
---------------------------------------------------------------------------

    Moreover, as we noted in the Proposing Release, the identification 
of a median employee does not necessarily require a determination of 
exact compensation amounts for every employee included in the sample. 
For example, rather than calculating exact compensation, a registrant 
could identify the employees in its sample that have extremely low or 
extremely high pay that would, therefore, fall on either end of the 
compensation spectrum. Since identifying the median involves finding 
the employee in the middle, it may not be necessary to determine the 
exact compensation amounts for every employee paid more or less than 
that employee in the middle. Instead, just noting that the employees 
are above or below the median may be sufficient for finding the 
employee in the middle of the compensation spectrum.
(iii) Consistently Applied Compensation Measures
    As proposed, the final rule permits registrants to use a 
consistently applied compensation measure, such as information derived 
from tax and/or payroll records, in determining the employees from 
which the median is identified as long as the registrant discloses the 
compensation measure used. Due to concerns about expected compliance 
costs arising from the complexity of using the ``total compensation'' 
calculation under Item 402(c)(2)(x) when identifying the median, we 
sought a reasonable alternative to identifying the median employee that 
is easier to calculate.
    As we noted in the Proposing Release, this approach provides a 
workable identification of the median employee for many registrants, 
and we expect it will reduce the costs of compliance. Most commenters 
discussing this issue agreed with this position and supported 
permitting registrants to use consistently applied compensation 
measures, such as information derived from tax and/or payroll records, 
to identify the median employee because it would reduce costs while not 
significantly affecting a registrant's pay ratio.\309\ As one commenter 
noted, while

[[Page 50137]]

a consistently applied compensation measure may exclude benefits, 
perquisites, and other allowances, it will still capture salary, 
incentive cash earned, and stock awards, which will encompass ``the 
substantial majority of compensation and [should] not lead to 
distortion of the median.'' \310\ In light of these comments, we do not 
believe this provision will hinder shareholders in using the pay ratio 
as a potentially useful data point in making their voting decisions on 
executive compensation under Section 951 of the Dodd-Frank Act.
---------------------------------------------------------------------------

    \309\ See, e.g., letters from ABA, AFL-CIO I, Calvert, 
Chesapeake Utilities, Chicago Teachers Fund, CII, First Affirmative, 
FS FTQ, ICCR, IL Bricklayers and Craftworkers Union, Johnson & 
Johnson, Marco Consulting, Meridian, Microsoft, Novara Tesija, NY 
Bricklayers and Craftworkers Union, Oxfam, Public Citizen I, SH&P, 
WA State Investment Board, and WorldatWork I.
    \310\ See letter from Microsoft.
---------------------------------------------------------------------------

    One commenter asserted that the final rule should be flexible 
enough to permit a registrant to use an internally consistent measure 
that is not necessarily internally identical.\311\ According to the 
commenter, the consistently applied compensation measure could be a 
measure based on the registrant's individual organizational structure 
that is reasonably designed to identify the median employee. As an 
example, the commenter stated that if a registrant selects to use 
``taxable wages'' as the consistently applied compensation measure, it 
is possible that the measure may be defined differently across multiple 
jurisdictions and be calculated over differing time periods. In such a 
situation, the commenter suggested that, although the registrant should 
attempt to use the non-United States equivalent to a Form W-2 for 
purposes of conducting its analysis, the final rule should provide 
sufficient flexibility to permit the use of other reasonable data 
sources for collecting the comparable information relating to non-U.S. 
employees as long as such measures are consistently applied within each 
subject jurisdiction.
---------------------------------------------------------------------------

    \311\ See letter from ABA.
---------------------------------------------------------------------------

    We agree. We note that a consistently applied compensation measure, 
such as ``taxable wages'' or ``cash compensation,'' \312\ may be 
defined differently across jurisdictions and may include different 
annual periods. For purposes of calculating the annual total 
compensation amounts when using a consistently applied compensation 
measure, the final rule permits registrants to use a measure that is 
defined differently across jurisdictions and may include different 
annual periods as long as within each jurisdiction, the measure is 
consistently applied. A registrant, however, would not be permitted to 
use an entirely different type of measure across jurisdictions that 
would not be consistently applied. The final rule does not require 
registrants to use any specific compensation measure when identifying 
the median employee. We continue to believe that registrants are in the 
best position to select a compensation measure that is appropriate to 
their own facts and circumstances. Therefore, consistent with the 
proposal, the final rule permits registrants to identify a median 
employee based on any consistently applied compensation measure, such 
as information derived from tax and/or payroll records, as long as the 
registrant briefly discloses the measure that it used. After the median 
employee is identified, registrants must calculate that median 
employee's annual total compensation in accordance with Item 
402(c)(2)(x).
---------------------------------------------------------------------------

    \312\ See, e.g., letters from Corporate Secretaries and Davis 
Polk.
---------------------------------------------------------------------------

(iv) The ``Median'' Employee
    Some commenters recommended that the final rule permit registrants 
to calculate the ratio using a figure other than the median, such as 
the employee earnings estimates available through the BLS,\313\ which 
may reduce costs for registrants and promote comparability across 
companies. As we stated in the Proposing Release, although such an 
approach would greatly reduce the compliance burden for registrants, we 
do not believe it is consistent with Section 953(b).
---------------------------------------------------------------------------

    \313\ See, e.g., letters from Brian Foley & Co., NACD, and NIRI. 
Some commenters recommended that the final rule permit registrants 
to use the average salary of its employees instead of the median. 
See, e.g., letters from Hyster-Yale, and NACCO, which they suggest 
also may reduce costs for registrants and promote comparability 
across companies. As with using the BLS data, however, we do not 
believe this approach is consistent with Section 953(b).
---------------------------------------------------------------------------

    Consistent with the proposal, the final rule requires that 
registrants provide their pay ratio disclosure using the compensation 
of their median employee. Section 953(b)(1)(A) states specifically that 
registrants must disclose ``the median of the annual total compensation 
of all employees of the issuer, except the chief executive officer.'' 
We believe, therefore, that Congress intended that the final rule 
should require registrants to use their median employee in their pay 
ratio determination.
    Although we considered commenters' arguments that ``median,'' as 
used in Section 953(b), can be interpreted to mean a measure other than 
``median,'' such as ``average'' or ``mean,'' we believe the better 
reading of the statutory language is that it means the statistical 
median of all employee compensation. ``Median'' has a specific meaning 
in statistics and probability theory,\314\ and there is no reason to 
believe that, when Congress chose to use this term in Section 953(b), 
it intended that it not have its established meaning. In the context of 
the disclosure required by Section 953(b), the alternative measures 
suggested by commenters could produce a significantly different number 
than the ``median,'' and we do not believe that would be appropriate in 
light of the statutory language and our understanding of its purpose. 
Further, we believe the median may provide a more useful or relevant 
data point for shareholders making their say-on-pay votes than would a 
mathematical average because the use of median can decrease the 
significance of outliers.
---------------------------------------------------------------------------

    \314\ The dictionary defines ``median'' as ``the middle number 
in a sequence, or the average of the two middle numbers when the 
sequence has an even number of numbers.'' Random House Webster's 
Dictionary, 411 (2d ed. 1996).
---------------------------------------------------------------------------

    Similarly, we note that some commenters suggested that a registrant 
should be permitted to use the compensation of a range of employees as 
the median compensation instead of the compensation of the exact median 
employee.\315\ We disagree. No matter what method a registrant chooses 
to identify its median employee, it must identify an actual employee 
and determine that employee's annual total compensation to use in the 
pay ratio disclosure. We note, however, that the final rule does not 
require a registrant to disclose any personally identifiable 
information about that employee other than his or her compensation. A 
registrant may choose to generally identify the employee's position to 
put the employee's compensation in context, but the registrant is not 
required to provide this information and should not do so if providing 
the information could identify any specific individual.
---------------------------------------------------------------------------

    \315\ See, e.g., letters from Capital Strategies, Hyster-Yale, 
Mercer I, NACCO, and NYC Bar.
---------------------------------------------------------------------------

    Another commenter recommended that a registrant, after identifying 
the median employee, should be able to select another employee as the 
median if that employee is within a 1% variance of the median and the 
original employee has anomalous compensation characteristics that would 
result in a pay ratio that did not accurately reflect the relationship 
between the compensation practices for a typical employee and the 
compensation of the

[[Page 50138]]

CEO.\316\ Given the significant flexibility that the rule provides 
registrants in identifying the median employee, we believe it would be 
appropriate to substitute another substantially similarly situated 
employee in these circumstances. Thus, when calculating the total 
compensation for that median employee in accordance with Item 
402(c)(2)(x), if the registrant reasonably determines that there are 
anomalous compensation characteristics of that employee's compensation 
that would have a significant higher or lower impact on the pay ratio, 
we will not object if the registrant substitutes another employee with 
substantially similar compensation to the original median employee 
based on the compensation measure used to select the median employee. 
The registrant must, however, disclose this fact as part of its brief 
description of the methodology it used to identify the median employee.
---------------------------------------------------------------------------

    \316\ See letter from PNC Financial Services (noting that if the 
median employee is chosen by a metric other than total Item 402(c) 
compensation, when that median employee's compensation is then 
calculated using Item 402(c), it is possible that such employee may 
have pay elements or be missing pay elements that would make his or 
her compensation anomalous when compared with others at the same 
overall compensation level, and providing some examples such as if 
the employee does not participate in certain benefit programs that 
are reflected in Item 402(c) compensation but not in W-2 
compensation).
---------------------------------------------------------------------------

b. Calculating Annual Total Compensation
i. Proposed Rule
    The proposed rule would define ``total compensation'' by reference 
to Item 402(c)(2)(x). In the Proposing Release, we stated that, because 
of the complexity of the requirements of Item 402(c)(2)(x), registrants 
typically compile information required by Item 402(c) manually for the 
named executive officers, which takes significant time and resources. 
Given the specificity of the definition used in Section 953(b), the 
proposed rule incorporated the Item 402(c)(2)(x) definition of ``total 
compensation'' for purposes of disclosing the median of the annual 
total compensation of employees and the pay ratio. Because the total 
compensation calculation using Item 402(c)(2)(x) would only be required 
for one additional employee (the median employee), we did not propose 
to simplify the total compensation definition that is required to be 
used to disclose the median employee compensation and the ratio.
    Additionally, the proposed rule permitted the use of reasonable 
estimates in determining any elements of total compensation of 
employees other than the PEO under Item 402(c)(2)(x). For registrants 
using estimates, an instruction to the proposed rule would require them 
to disclose and consistently apply any material estimate used to 
identify the median employee or to determine that employee's total 
compensation or any elements of total compensation, and to clearly 
identify any estimates used. In using an estimate for annual total 
compensation (or for a particular element of total compensation), a 
registrant would be required to have a reasonable basis to conclude 
that the estimate approximates the actual amount of compensation under 
Item 402(c)(2)(x) (or for a particular element of compensation under 
Item 402(c)(2)(iv)-(ix)) awarded to, earned by, or paid to the 
employee. We did not specify what a reasonable basis would entail 
because we believed that would necessarily depend on the registrant's 
particular facts and circumstances.
    Because the requirements of Item 402(c)(2)(x) were promulgated to 
address executive officer compensation, rather than compensation for 
all employees, we considered some interpretive questions that 
registrants could face in applying the requirements of Item 
402(c)(2)(x) to employees who are not executive officers and proposed 
ways to address those questions. Those included questions concerning: 
applying the definition of ``total compensation'' to an employee who is 
not an executive officer and valuation issues for certain elements of 
total compensation, including for non-U.S. employees.
ii. Comments on the Proposed Rule
    One commenter agreed with the Proposed Rule's requirement that 
``total compensation'' be calculated using the requirements of Item 
402(c)(2)(x).\317\ Also, this commenter and a few others stated 
specifically that they supported the proposed rule's flexibility in 
permitting the use of reasonable estimates to calculate the annual 
total compensation or any elements of total compensation for employees 
other than the PEO.\318\ One commenter indicated that it expected its 
clients to use reasonable estimates for calculating total 
compensation.\319\ Some commenters requested additional guidance as to 
what estimates would be considered reasonable,\320\ but other 
commenters said that no additional guidance is required.\321\
---------------------------------------------------------------------------

    \317\ See letter from COEC II.
    \318\ See, e.g., letters from ABA, Prof. Angel, COEC I, COEC II, 
Davis Polk, and Vectren.
    \319\ See letter from Mercer I.
    \320\ See, e.g., letters from Aon Hewitt (requesting also a safe 
harbor in the final rule) and PM&P.
    \321\ See, e.g., letters from ABA and COEC II.
---------------------------------------------------------------------------

    One commenter stated that it was ``deeply troubled'' that the 
proposed rule would require registrants to calculate the median 
employee's total compensation using Item 402(c)(2)(x) because no 
registrant uses this measure to calculate a non-executive officer's 
compensation.\322\ According to the commenter, a significant difficulty 
with using Item 402(c)(2)(x) is the inclusion of pension benefits 
because the actuarial value of pensions vacillate dramatically from 
year to year, which would significantly impact total compensation. The 
commenter recommended excluding pension value from the total 
compensation calculation. If the final rule does not exclude pensions, 
the commenter suggested that government-related pensions for non-U.S. 
employees should be excluded as they are for named executive officers 
under Item 402(c)(2)(x).
---------------------------------------------------------------------------

    \322\ See letter from PM&P.
---------------------------------------------------------------------------

    Other commenters also contended that the final rule should not 
require calculating total compensation using Item 402(c)(2)(x) because 
certain compensation measures are excluded from that item requirement, 
such as benefits, non-discriminatory plans, perquisites, and personal 
benefits that aggregate less than $10,000, which would cause the median 
employee's total compensation to be understated.\323\ Another 
commenter, while not necessarily advocating against using Item 
402(c)(2)(x) to calculate total compensation, noted that it would be 
difficult for registrants to use that item requirement because it would 
require them to include as compensation bonuses, stock compensation, 
pensions, and other benefits for the median employee that are usually 
not factored into annual salary amounts in company records.\324\
---------------------------------------------------------------------------

    \323\ See, e.g., letters from Hyster-Yale, NACCO, and NACD.
    \324\ See letter from IBC.
---------------------------------------------------------------------------

    Some commenters suggested that the final rule prescribe a 
methodology for calculating total compensation.\325\ One commenter 
requested specifically that we include guidance about excluding 
government-mandated pension plans.\326\ Another commenter suggested 
that registrants should be permitted (but not required) to include 
benefits, non-discriminatory plans, perquisites, and personal benefits 
that aggregate less than $10,000 in total compensation because Item 
402(c)(2)(x) merely

[[Page 50139]]

permits and does not require exclusion of such items for executive 
officers.\327\ Other commenters contended that the calculation of total 
compensation in the final rule should include all compensation of a 
registrant's PEO and median employee, such as benefits, perks, bonuses, 
stock options, and other forms of compensation; \328\ be limited to 
cash and stock-based compensation; \329\ or include only compensation 
on the W-2 Form of the PEO and median employee.\330\
---------------------------------------------------------------------------

    \325\ See, e.g., letters from Friend and William Preston (Oct. 
5, 2013) (``Preston'').
    \326\ See letter from ABA.
    \327\ Id.
    \328\ See, e.g., letters from Corayer, Fedewa, Gould, LAPFF, 
Matteson, and Anne C. Somers (Oct. 24, 2013) (``Somers'').
    \329\ See, e.g., letters from Brian Foley & Co., Dover Corp., 
and Semtech.
    \330\ See, e.g., letters from Capital Strategies, FEI, Hyster-
Yale, and NACCO.
---------------------------------------------------------------------------

    One commenter recommended that total compensation either include 
the average change in defined benefit pension values or exclude defined 
benefit pension values entirely.\331\ Another commenter indicated that 
total compensation should include welfare and retirement benefits 
included to union members.\332\ One commenter indicated that total 
compensation should be calculated based upon the same method used when 
identifying the three most highly compensated executive officers in 
accordance with Instruction 1 of Item 402(a)(3), which excludes the 
value of the aggregate change in actuarial present value of defined 
pension benefits under Item 402(c)(2)(viii).\333\
---------------------------------------------------------------------------

    \331\ See letter from Mercer I.
    \332\ See letter from Vectren Corp.
    \333\ See letter from SH&P.
---------------------------------------------------------------------------

    Finally, one commenter suggested that, if the final rule allows 
registrants to identify the median employee only every three years, 
registrants should similarly be allowed to calculate total compensation 
either every year or only when a new median is determined unless there 
has been a material change in annual total compensation.\334\
---------------------------------------------------------------------------

    \334\ See letter from ABA.
---------------------------------------------------------------------------

iii. Final Rule
    The definition of ``total compensation'' in the final rule is 
unchanged from the proposal. The final rule requires that ``total 
compensation'' for both the median employee and PEO be calculated using 
the requirements of Item 402(c)(2)(x). As with the proposed rule, the 
final rule provides registrants with flexibility in identifying the 
median employee, and does not require registrants to identify the 
median employee by calculating the total compensation for each 
employee. Because the total compensation calculation using Item 
402(c)(2)(x) is only required for one additional employee (the median 
employee), we do not believe it necessary in the final rule to simplify 
the total compensation definition that is required to be used in that 
calculation.
    THE FINAL RULE PERMITS REGISTRANTS TO USE REASONABLE ESTIMATES IN 
CALCULATING THE ANNUAL TOTAL COMPENSATION OF THEIR MEDIAN EMPLOYEE, 
INCLUDING ANY ELEMENTS OF THE TOTAL COMPENSATION, UNDER ITEM 
402(C)(2)(X). A FEW COMMENTERS SUPPORTED SUCH FLEXIBILITY.\335\ WE 
BELIEVE, AS WE NOTED IN THE PROPOSING RELEASE, THAT THE USE OF 
REASONABLE ESTIMATES DOES NOT DIMINISH THE POTENTIAL USEFULNESS OF THE 
PAY RATIO DISCLOSURE, IS CONSISTENT WITH SECTION 953(B), AND WILL 
RESULT IN LOWER COMPLIANCE COSTS ON REGISTRANTS.
---------------------------------------------------------------------------

    \335\ See, e.g., letters from ABA, Capital Strategies, COEC I, 
Davis Polk, Johnson & Johnson, Mercer I, Vectren, Corp., and 
WorldatWork I.
---------------------------------------------------------------------------

    Under the final rule, registrants must clearly identify any 
estimates used. Additionally, registrants must have a reasonable basis 
to conclude that their estimates approximate the actual amounts of Item 
402(c)(2)(x) compensation, or a particular element of compensation 
under Item 402(c)(2)(iv)-(ix), that are awarded to, earned by, or paid 
to the median employee. Some commenters requested that we provide 
additional guidance as to what estimates would be considered 
reasonable.\336\ Consistent with the Proposing Release, we are not 
prescribing what a reasonable basis would entail in the final rule 
because we believe that will necessarily depend on the registrant's 
particular facts and circumstances.
---------------------------------------------------------------------------

    \336\ See, e.g., letters from Aon Hewitt (requesting also a safe 
harbor in the final rule) and PM&P.
---------------------------------------------------------------------------

    As we discussed in the Proposing Release, our rules for 
compensation disclosure focus on the compensation of executive officers 
and directors rather than compensation for all employees. Some 
commenters urged us not to require registrants to calculate total 
compensation using Item 402(c)(2)(x).\337\ We believe, however, that it 
is appropriate to require ``total compensation'' to be calculated using 
Item 402(c)(2)(x) for both the median employee and PEO. Using different 
measures of total compensation for the median employee and the PEO 
would be inconsistent with the statutory language and alter the pay 
ratio. Due to the concerns about calculating median employee 
compensation using requirements meant only for executive offices, 
however, we are reiterating the discussion that we included in the 
Proposing Release to help registrants understand our views on how the 
final rule should be applied.
---------------------------------------------------------------------------

    \337\ See, e.g., letters from Hyster-Yale, IBC, NACCO, NACD, and 
PM&P.
---------------------------------------------------------------------------

    In calculating the annual total compensation of employees in 
accordance with Item 402(c)(2)(x), applicable references to ``named 
executive officer'' in Item 402 and the related instructions are deemed 
in the final rule to refer instead to ``employee,'' as proposed. Also, 
the final rule clarifies that, for non-salaried employees, references 
to ``base salary'' and ``salary'' in Item 402 are deemed to refer 
instead, as applicable, to ``wages plus overtime.'' We are adopting 
this provision to help registrants calculate the total compensation for 
a median employee that happens to be non-salaried.
    Additionally, registrants may use reasonable estimates, as 
described above, in determining an amount that reasonably approximates 
the aggregate change in actuarial present value of an employee's 
defined pension benefit for purposes of Item 402(c)(2)(viii), as we 
stated in the Proposing Release. For example, in the case of pension 
benefits provided to union members in connection with a multi-employer 
defined benefit pension plan, the participating employers typically do 
not have access to information (or may not have access in the timeframe 
needed to compile pay ratio disclosure) from the plan administrator 
that would be needed to calculate the aggregate change in actuarial 
present value of the accumulated benefit of a particular individual 
under the plan.\338\ In such circumstances, we believe it would be 
appropriate for a registrant to use reasonable estimates in determining 
an amount that reasonably approximates the aggregate change in 
actuarial present value of an employee's defined pension benefit for 
purposes of Item 402(c)(2)(viii).
---------------------------------------------------------------------------

    \338\ Section 101(k) and related regulations under the Employee 
Retirement Income Security Act of 1974, as amended [21 U.S.C. 
1021(k)], govern the requirements for plan administrators to provide 
actuarial reports relating to the plan. Under the rules, a plan 
administrator has thirty days to respond to a request for an 
actuarial report, and it is not required to provide access to any 
reports that have not been its possession for more than thirty days. 
In addition, the rules prohibit the disclosure of reports that 
include information that the plan administrator reasonably 
determines to be personally identifiable information regarding a 
plan participant, beneficiary or contributing employer. See 29 CFR 
2520.101-6.
---------------------------------------------------------------------------

    The instructions to Item 402(c)(2)(ix) permit the exclusion of 
personal benefits as long as the total value for the

[[Page 50140]]

employee is less than $10,000, based on the basis of the aggregate 
incremental cost to the registrant.\339\ In calculating any such 
amounts for purposes of calculating the annual total compensation of 
employees other than the PEO, a registrant may use reasonable estimates 
in the manner described above, as proposed. In light of concerns about 
the difficulty and complexity in the valuation of government-mandated 
pension plans, we acknowledged in the proposing release that some 
registrants might need clarity as to how to treat government-mandated 
pension plans for purposes of calculating an employee's total 
compensation and, specifically, for purposes of determining the 
aggregate change in actuarial present value of defined pension benefits 
under Item 402(c)(2)(viii). Item 402(c)(2)(viii) applies to a defined 
benefit plan, which, as explained in the 2006 Adopting Release, is a 
retirement plan in which the company pays the executive specified 
amounts at retirement that are not tied to the investment performance 
of the contributions that fund the plan.\340\ In contrast, under many 
government-mandated pension plans, the employee ultimately receives the 
pension benefit payment from the government, not the employer, and the 
purpose of the mandated pension benefit is not to provide compensation 
to the employee for services performed for the employer.\341\ 
Notwithstanding any amounts that an employer may be obligated to pay 
(typically as a tax) to the government in respect of an employee or 
amounts the employee may be obligated to have withheld from wages and 
paid to the government, where a pension benefit is being provided to 
the employee from the government and not by the registrant, a 
government-mandated defined benefit pension plan should not be 
considered a ``defined benefit plan'' for purposes of Item 
402(c)(2)(viii) and any accrued pension benefit under such a plan 
should not be considered compensation for purposes of Item 
402(c)(2)(x).
---------------------------------------------------------------------------

    \339\ See Instruction 4 to Item 402(c)(2)(ix). This instruction 
would apply to perquisites and personal benefits. Accordingly, 
perquisites provided to executive officers who are included in the 
identification of the median would be treated as set forth in 
Instruction 4. For this purpose, however, benefits that were 
provided to all employees or all salaried employees would not have 
been considered ``perquisites.''
    \340\ See 2006 Adopting Release at 53175. This definition serves 
to distinguish defined benefit pension plans from defined 
contribution plans, in which the amount payable at retirement is 
tied to the performance of the contributions that fund the plan.
    \341\ Although Item 402(a)(2) includes compensation transactions 
between a registrant and a third party where the purpose of the 
transaction is to furnish compensation to the employee, we generally 
would not consider a government-mandated pension plan to be such a 
transaction.
---------------------------------------------------------------------------

    Some commenters expressed concern over including pension benefits 
in calculating total compensation.\342\ One commenter indicated that it 
would be difficult to use Item 402(c)(2)(x) in calculating ``total 
compensation'' due to the inclusion of pension benefits because the 
actuarial value of pensions vacillate dramatically from year to year, 
which could significantly impact total compensation.\343\ Another 
commenter requested specifically that we include guidance about 
excluding government-mandated pension plans.\344\ In response to these 
comments, we are clarifying that, in calculating ``total 
compensation,'' registrants may exclude government-related pension 
benefits for non-U.S. employees, just as Social Security benefits are 
excluded under ``total compensation'' for U.S. employees. Ultimately, 
as to both concerns, we believe that the ability to use reasonable 
estimates should help with calculating total compensation for the 
median employee who has pension benefits.
---------------------------------------------------------------------------

    \342\ See, e.g., letters from ABA and PM&P.
    \343\ See letter from PM&P.
    \344\ See letter from ABA.
---------------------------------------------------------------------------

    We acknowledge that the application of the definition of total 
compensation under Item 402(c)(2)(x) to employees who are not executive 
officers could understate the overall compensation paid to such 
employees. Item 402 captures all of the various compensation components 
received by a named executive officer, excluding certain limited items 
like benefits under non-discriminatory plans and perquisites and 
personal benefits that aggregate less than $10,000. By excluding 
certain benefit plans and perquisites that do not exceed the $10,000 
threshold, however, the rules may understate the median employee's 
actual total compensation. To address this, the final rule permits 
registrants, at their discretion, to include personal benefits that 
aggregate less than $10,000 and compensation under non-discriminatory 
benefit plans in calculating the annual total compensation of the 
median employee. To be consistent, however, the PEO's total 
compensation used in the related pay ratio disclosure must also reflect 
the same approach to these items used for the median employee. The 
registrant must also explain any difference between the PEO total 
compensation used in the pay ratio disclosure and the total 
compensation amounts reflected in the Summary Compensation Table, if 
material.
    One commenter suggested that, if the final rule allows registrants 
to identify the median employee only every three years, registrants 
should similarly be required to calculate total compensation only when 
a new median is determined or when there is a material change to the 
annual total compensation figure.\345\ The final rule allows 
registrants to identify the median employee every three years, but 
requires total compensation for that employee to be calculated each 
year. The primary reason for our decision to permit registrants to 
calculate the median employee every three years is to reduce the costs 
and burdens to registrants. Based on the comments we received, we 
understand that much of the cost associated with the proposed rule 
arises from the task of identifying the median employee.\346\ Several 
commenters stated that navigating a registrant's payroll systems and 
creating a single database of all of its employees' compensation, 
especially non-U.S. employees' compensation, would be the most costly 
aspect of the proposed rule.\347\ Other activities mentioned by 
commenters that would contribute to the costs of the proposed rule 
included data privacy compliance, foreign exchange calculations, data 
testing, establishing corporate guidelines, obtaining legal services, 
auditing results, public relations tasks, and litigation risk.\348\ 
Many of these activities also must be undertaken in identifying the 
median employee. Once the median employee has been identified, however, 
it does not appear that calculating the annual total compensation of 
that one additional employee is a source of significant additional 
cost. Additionally, since the PEO's compensation will be updated 
annually, we believe that it is appropriate to have a consistent 
reflection of that year's compensation both for the PEO and the median 
employee. Therefore, the final rule requires registrants to calculate 
the median employee's annual total compensation every year.
---------------------------------------------------------------------------

    \345\ See letter from ABA.
    \346\ One commenter stated explicitly that the primary cost 
associated with the proposed rule would be in identifying the median 
employee. See letter from McGuireWoods.
    \347\ See, e.g., letters from Aon Hewitt, Avery Dennison, 
Business Roundtable I, Chamber I, COEC I, Corporate Secretaries, 
Eaton, FEI, FuelCell Energy, IBC, KBR, NACCO, NAM I, NAM II, and 
NIRI.
    \348\ See, e.g., letters from Aon Hewitt, Corporate Secretaries, 
and McGuireWoods.
---------------------------------------------------------------------------

    Finally, Section 953(b)(2) states that ``total compensation'' shall 
be determined in accordance with Item 402(c)(2)(x) ``as in effect on 
the day before the date of enactment of this

[[Page 50141]]

Act.'' One commenter suggested that this statement does not preclude 
any amendment of Regulation S-K subsequent to the passage of the Dodd-
Frank Act that would alter the definition of ``total compensation'' in 
Item 402 in effect on the day before the date of enactment of the 
Act.\349\ In the Proposing Release, we noted that Section 953(b) refers 
to Item 402(c)(2)(x) in effect on the day before enactment of the Dodd-
Frank Act, or July 20, 2010. We also indicated that, because no 
substantive amendments have been made to Item 402(c) since that date, 
the proposed rule would refer to Item 402(c)(2)(x) without reference to 
the rules in effect on July 20, 2010. We further stated that we expect 
to address the impact on the proposed rule of any future amendments to 
Item 402(c)(2)(x) if and when such future amendments are considered. No 
substantive amendments have been made to Item 402(c) since July 20, 
2010. We continue, therefore, to take the approach articulated in the 
Proposing Release on this issue.
---------------------------------------------------------------------------

    \349\ See letter from Chamber I.
---------------------------------------------------------------------------

3. Disclosure of Methodology, Assumptions, and Estimates
a. Proposed Rule
    The proposed rule required registrants to briefly describe and 
consistently apply any methodology used to identify the median and any 
material assumptions, adjustments, or estimates used to identify the 
median or to determine total compensation or any elements of total 
compensation. The proposed rule also provided that, if a registrant 
changes methodology, material assumptions, adjustments, or estimates 
from those used in its pay ratio disclosure for the prior fiscal year, 
and if the effects of any such change are material, the registrant must 
briefly describe the change and the reasons for the change, and provide 
an estimate of the impact of the change on the median and the ratio. 
The proposed rule would not require registrants to provide technical 
analyses or formulas (such as statistical formulas, confidence levels 
or the steps used in data analysis).
b. Comments on the Proposed Rule
    Many commenters indicated that the rule should require registrants 
to provide narrative information about the methodology and material 
assumptions, adjustments, or estimates they used in identifying the 
median or calculating annual total compensation for employees.\350\ 
Only a few commenters asserted that the rule should not require 
registrants to provide narrative information.\351\ One of these 
commenters recommended also that we clarify the nature of the 
information that we expect registrants to disclose without imposing 
restrictions on methodologies; state expressly that disclosure is only 
required if a registrant used material assumptions, adjustments, or 
estimates; and state that no negative statement is required if a 
registrant did not use material assumptions, adjustments, or 
estimates.\352\
---------------------------------------------------------------------------

    \350\ See, e.g., letters from AFL-CIO I, Barnard, Business 
Roundtable I, CalPERS, CalSTRS, Calvert, CII, COEC I, COEC II, CT 
State Treasurer, Domini, Hermes Equity Ownership Services (Nov. 18, 
2013) (``Hermes''), Alex Kasner (Nov. 12, 2013) (``Kasner''), LAPFF, 
Meridian, Microsoft, Somers, Wesley Sze (Nov. 13, 2013) (``Sze''), 
UAW Trust, US SIF, and WorldatWork I.
    \351\ See, e.g., letters from Prof. Angel, Capital Strategies, 
and Prof. Ray.
    \352\ See letter from ABA.
---------------------------------------------------------------------------

    One commenter expressed concern that the disclosure would not be 
brief because of the registrant's need to use many estimates and 
assumptions for the median, especially for non-U.S. employees.\353\ 
Another commenter cited survey data in which 66% of respondents 
anticipated that they would feel compelled to provide more than a brief 
narrative to explain how they determined the pay ratio.\354\ Some 
commenters did not support requiring any additional narrative 
disclosure beyond what was already in the proposed rule.\355\ Some of 
these commenters, as well as a large number of other commenters, 
asserted that the final rule should permit registrants to provide 
additional narrative disclosures if they chose to do so.\356\
---------------------------------------------------------------------------

    \353\ See letter from Business Roundtable I.
    \354\ See letter from COEC I.
    \355\ See, e.g., letters from ABA, COEC I, Hyster-Yale, Intel, 
Meridian, NACCO, NYC Comptroller, Vivient, and WorldatWork I.
    \356\ See, e.g., letters from AFL-CIO I, AFSCME, B[acirc]tirente 
et al., CBIS, CT State Treasurer, E&Y, First Affirmative, CUPE, FS 
FTQ, Hyster-Yale, ICCR, IL Bricklayers and Craftworkers Union, 
Intel, Marco Consulting, McGuireWoods, McMorgan & Co., NYC 
Comptroller, NACCO, Novara Tesija, NY Bricklayers and Craftworkers 
Union, Oxfam, Public Citizen I, Rep. Ellison et al. I, Socially 
Responsive Financial Advisors, Teamsters, Trillium I, Trustee 
Campbell, Vivient, and Walden.
---------------------------------------------------------------------------

    Some commenters noted that the final rule should require 
registrants to disclose material changes to their assumptions, 
adjustments, or estimates from previous years, as proposed.\357\ One 
commenter suggested that the final rule allow a good-faith compliance 
period of two years in which a registrant can change its initial 
methodology without having to specifically explain and quantify the 
change.\358\
---------------------------------------------------------------------------

    \357\ See, e.g., letters from ABA, CEG, CT Treasuer, Domini, 
Kasner, McGuireWoods, and WorldatWork I.
    \358\ See letter from Frederic W. Cook & Co.
---------------------------------------------------------------------------

    Finally, a few commenters requested that the final rule require 
some additional metrics, such as upper and lower quartiles, mean, and 
standard deviation,\359\ and one commenter suggested that companies 
voluntarily disclose both the ratio between average employee pay and 
average executive pay and the ratio of pay between the top and bottom 
10% of earners within the company.\360\ Other commenters, however, 
stated specifically that the final rule should not require the 
disclosure of any additional metrics.\361\
---------------------------------------------------------------------------

    \359\ See, e.g., letters from Prof. Ray and Rebecca Vogel (Nov. 
13, 2013) (``R. Vogel'').
    \360\ See letter from LAPFF.
    \361\ See, e.g., letters from Capital Strategies, Johnson & 
Johnson, and WorldatWork I.
---------------------------------------------------------------------------

c. Final Rule
    The final rule, consistent with the proposal, requires registrants 
to briefly describe and consistently apply any methodology used to 
identify the median and any material assumptions, adjustments 
(including any cost-of-living adjustments), or estimates used to 
identify the median or to determine total compensation or any elements 
of total compensation. The final rule also requires a registrant to 
clearly identify any estimates used. For example, when statistical 
sampling is used, registrants must describe the size of both the sample 
and the estimated whole population, any material assumptions used in 
determining the sample size and the sampling method (or methods) is 
used. Additionally, although the required descriptions must provide 
sufficient information for readers to evaluate the appropriateness of 
the methodologies used, registrants are not required to include any 
technical analyses, formulas, confidence levels, or the steps used in 
data analysis.\362\ Although one commenter suggested that the final 
rule allow a good-faith compliance period of two years in which an 
issuer may change its initial methodology without having to 
specifically explain and quantify the change,\363\ the final rule 
requires registrants to disclose any change in methodology, significant 
assumption, adjustment, or estimate from the prior year if the effects 
of any such change are significant. Registrants must also disclose if 
they changed from using the cost-of-living adjustment to not using that 
adjustment and if they changed from not using the cost-of-living

[[Page 50142]]

adjustment to using it. We believe that it is important for 
shareholders to understand changes to a registrant's methodology so 
that they may make informed voting decisions on executive compensation 
under Section 951 of the Dodd-Frank Act.
---------------------------------------------------------------------------

    \362\ A registrant, however, must include the measure used as 
the basis for any cost-of-living adjustments when briefly describing 
the cost-of-living adjustments it used to identify the median 
employee and calculate the median employee's annual total 
compensation.
    \363\ See letter from Frederic W. Cook & Co.
---------------------------------------------------------------------------

    We believe that requiring registrants to include the brief overview 
will make it easier for shareholders to understand the pay ratio 
disclosure for that company and better evaluate its utility in 
assessing the compensation and accountability of a registrant's 
executives, including in making their voting decisions on executive 
compensation under Section 951. We do not believe that requiring 
registrants to provide additional metrics, such as a more detailed or 
technical analysis will help shareholders in this manner. We note that 
other of our rules require similar disclosures, particularly where 
registrants are given the flexibility to choose a methodology, such as 
the valuation method for determining the present value of accrued 
pension benefits in Item 402(h)(2) or the description of models, 
assumptions, and parameters in Item 305 of Regulation S-K (quantitative 
and qualitative disclosures about market risk). Several commenters 
agreed that the rule should require registrants to provide this 
narrative description, as proposed, and no additional information.\364\ 
Further, a number of these commenters indicated that the final rule 
should clarify that the narrative should be brief.\365\ Consistent with 
these comments, the final rule specifically states that registrants 
must ``briefly'' describe this information.
---------------------------------------------------------------------------

    \364\ See, e.g., letters from ABA, AFL-CIO I, Barnard, Business 
Roundtable I, CalPERS, CalSTRS, Calvert, CII, COEC I, COEC II, CT 
State Treasurer, Domini, Hermes, Hyster-Yale, Kasner, Intel, LAPFF, 
Meridian, Microsoft, NACCO, Comptroller of the City of New York 
(Nov. 27, 2013) (``New York City Comptroller''), Somers, UAW Trust, 
US SIF, Vivient, and WorldatWork I.
    \365\ See, e.g., letters from ABA, AFL-CIO I, Business 
Roundtable I, COEC I, COEC II, Domini, Meridian, Microsoft, and 
WorldatWork I.
---------------------------------------------------------------------------

    We note that some commenters contended that the final rule should 
require additional metrics,\366\ whereas other commenters stated 
specifically that the final rule should not require the disclosure of 
any additional metrics.\367\ As we discussed in the Proposing Release, 
we are sensitive to the costs of the mandated disclosure, and we 
believe that narrative disclosure in addition to what is already 
required about the ratio would not, for many registrants, provide 
useful information for shareholders. For example, disclosures about 
employment policies, use of part-time employees, use of seasonal 
employee workers, and outsourcing and off-shoring strategies are not 
required under the final rule. The final rule does, however, allow 
registrants the flexibility to provide those additional disclosures 
that they believe will assist shareholders' understanding of the 
meaning of the pay ratio disclosure for their particular circumstances. 
We believe this approach is preferable to imposing a requirement on all 
registrants to provide additional metrics that may not be useful in 
many cases.
---------------------------------------------------------------------------

    \366\ See, e.g., letters from Prof. Ray, LAPFF, and R. Vogel.
    \367\ See, e.g., letters from Capital Strategies, Johnson & 
Johnson, and WorldatWork I.
---------------------------------------------------------------------------

4. Meaning of ``Annual''
a. Proposed Rule
    The proposed rule would define ``annual total compensation'' to 
mean total compensation for the last completed fiscal year, consistent 
with the time period used for the other Item 402 disclosure 
requirements. This provision was intended to address concerns about the 
need to update the pay ratio disclosure throughout the year and to make 
clear that the disclosure does not need to be updated more than once a 
year. Although we considered other ``annual'' periods that may have 
reduced compliance costs for registrants by giving them the ability to 
use information in the form that it is currently compiled for other 
purposes, we believed it was appropriate for the time period for the 
pay ratio disclosure to be the same as the time period used for the 
PEO's compensation. Registrants, therefore, would be required to 
calculate the total compensation for the median employee for their last 
completed fiscal year.
    For purposes of identifying the median employee, however, we 
proposed allowing registrants to use compensation amounts derived from 
the information derived from tax and/or payroll records for the annual 
period used in those records. We believed that permitting companies to 
identify the median employee using compensation information in the form 
that it is maintained in their own books and records would reduce 
compliance costs. Registrants using the information derived from tax 
and/or payroll records to identify the median employee would still be 
required to calculate the Item 402(c)(2)(x) total compensation for that 
median employee for the last completed fiscal year, rather than the 
annual period used in the payroll and/or tax records.
b. Comments on the Proposed Rule
    Some commenters agreed that the final rule should require the pay 
ratio to be calculated for the last completed fiscal year, as proposed, 
rather than some other annual period.\368\ Other commenters, however, 
contended that the final rule should provide another annual period. 
Some of these commenters recommended that the final rule permit 
registrants to use the year prior to the registrant's last completed 
fiscal year for identifying the median employee, annual total 
compensation, or both to give the registrants more time to identify the 
median employee and calculate his or her total compensation.\369\ Using 
the employee population from the year before, one commenter stated, 
would not have a material impact on the ratio and the added value of 
more contemporaneous information would likely be negligible.\370\ 
Another commenter suggested using this time period for registrants with 
non-U.S. employees.\371\
---------------------------------------------------------------------------

    \368\ See, e.g., letters from ABA, CalPERs and UAW Trust.
    \369\ See, e.g., letters from Aon Hewitt, Business Roundtable I, 
Corporate Secretaries, and Eaton.
    \370\ See letter from Aon Hewitt.
    \371\ See letter from FSR.
---------------------------------------------------------------------------

    Some commenters urged us to adopt a final rule that permitted use 
of the time periods used for payroll and/or tax records when 
calculating compensation to identify the median employee and the pay 
ratio for that employee.\372\ These commenters indicated that the 
periods could be different across jurisdictions. In this regard, 
another commenter noted that there is no need to have exact overlap of 
the time periods because the pay ratio ``won't change all that 
much.''\373\
---------------------------------------------------------------------------

    \372\ See, e.g., letters from Hyster-Yale and NACCO.
    \373\ See letter from Prof. Angel.
---------------------------------------------------------------------------

    Other commenters asserted that registrants should be given 
flexibility to choose any annual period in identifying the median 
employee and/or that employee's total compensation.\374\ One of these 
commenters stated, however, that the annual period must substantially 
relate to the fiscal year for which the pay ratio disclosure is being 
provided, regardless of whether the last day of such annual period 
falls before or after the end of the registrant's fiscal year for the 
purposes of identifying the median employee.\375\ As an example, this 
commenter stated that, if the registrant has a fiscal year ending on 
November 30, the registrant should be permitted to identify the median 
employee based on a compensation

[[Page 50143]]

measure calculated from January 1 through December 31 of that year, as 
long as such records substantially relate to the fiscal year for which 
pay ratio disclosure is being provided.
---------------------------------------------------------------------------

    \374\ See, e.g., letters from Davis Polk and WorldatWork I.
    \375\ See letter from Davis Polk.
---------------------------------------------------------------------------

c. Final Rule
    The final rule defines ``annual total compensation'' to mean 
``total compensation'' for the registrant's last completed fiscal year, 
as proposed. Although there were other ``annual'' periods suggested by 
commenters, such as the year prior to the registrant's last completed 
fiscal year \376\ or the time periods used for the information derived 
from tax and/or payroll records,\377\ we believe the registrant's last 
completed fiscal year is more appropriate. Using the registrant's last 
completed fiscal year is consistent with the time period used for the 
other Item 402 disclosure requirements. Registrants are required, 
therefore, to disclose the ``total compensation,'' using Item 
402(c)(2)(x), for their median employee and PEO based on the 
compensation they provided for these individuals in the last completed 
fiscal year. We believe that making the time period for the pay ratio 
disclosure consistent with other Item 402 disclosures will better 
enable shareholders to use it in conjunction with the other Item 402 
disclosures to assess the compensation and accountability of a 
registrant's executives. For this same reason, we are not permitting 
registrants to select any annual period or the year prior to the last 
completed fiscal year to calculate total compensation.
---------------------------------------------------------------------------

    \376\ See, e.g., letters from Aon Hewitt, Business Roundtable I, 
Corporate Secretaries, Eaton, and FSR.
    \377\ See, e.g., letters from Hyster-Yale and NACCO.
---------------------------------------------------------------------------

    As discussed above, registrants may use compensation amounts 
derived from the information derived from their tax and/or payroll 
records for the same annual period used in those records to identify 
their median employee because we believe this reduces compliance costs. 
Registrants using the information derived from tax and/or payroll 
records to identify the median employee are still required to calculate 
the Item 402(c)(2)(x) total compensation for that median employee for 
the registrant's last completed fiscal year, rather than the annual 
period used in the payroll and/or tax records because identifying the 
median is a separate process from calculating total compensation.
5. ``Filed'' Not ``Furnished''
a. Proposed Rule
    Under the proposal, the pay ratio disclosure would be considered 
``filed'' for purposes of the Securities Act and Exchange Act, which is 
the same as for other Item 402 information.
b. Comments on the Proposed Rule
    Only one commenter stated explicitly that the pay ratio disclosure 
should be ``filed,'' as proposed.\378\ This commenter agreed that the 
information should be filed because Section 953(b) refers to 
``filings.'' Further, the commenter stressed that any concerns 
registrants may have about the pay ratio information being ``filed'' 
are mitigated by the proposed rule's flexibility.
---------------------------------------------------------------------------

    \378\ See letter from US SIF.
---------------------------------------------------------------------------

    Commenters that opposed the proposed rule generally indicated that 
the pay ratio disclosure should be ``furnished'' rather than ``filed.'' 
\379\ The commenters contending that the pay ratio information should 
be ``furnished'' argued that, in making the calculations for 
identifying the median employee and total compensation, registrants 
will have to review a large amount of data and make a significant 
number of estimates, assumptions, and judgment calls, which will 
necessarily lead to imprecision.\380\ Some noted that this imprecision 
will subject a registrant to potential liability and litigation,\381\ 
make it difficult to validate the information sufficiently for 
Sarbanes-Oxley Act \382\ certification purposes,\383\ and/or not permit 
the information to be audited (or greatly increase the costs of the 
audits).\384\
---------------------------------------------------------------------------

    \379\ See, e.g., letters from AAFA II, ABA, American Benefits 
Council, Aon Hewitt, Best Buy et al., Bill Barrett Corp., Business 
Roundtable I, Chamber I, Chesapeake Utilities, COEC I, COEC II, 
Corporate Secretaries, Eaton, Freeport-McMoRan, General Mills, 
Intel, Mercer I, NAM I, NIRI, NRF, PM&P, RILA, SHRM, and Vectren 
Corp.
    \380\ See, e.g., letters from AAFA II, ABA, Business Roundtable 
I, Chesapeake Utilities, COEC I, COEC II, Corporate Secretaries, 
Eaton, General Mills, NRF, RILA, and Vectren Corp.
    \381\ See, e.g., letters from ABA, American Benefits Council, 
Aon Hewitt, Bill Barrett Corp., Chamber I, General Mills, Mercer I, 
and PM&P.
    \382\ See Pub. L. 107-204, 116 Stat. 745 (2002).
    \383\ See, e.g., letters from ABA, Best Buy et al., Corporate 
Secretaries, Freeport-McMoRan, Intel, NAM I, NAM II, and SHRM.
    \384\ See, e.g., letters from COEC I, COEC II, Corporate 
Secretaries, and NIRI.
---------------------------------------------------------------------------

    Some commenters asserted that use of the word ``filing'' in Section 
953(b) does not demonstrate a Congressional desire that the disclosure 
be ``filed'' rather than ``furnished.'' Some of these commenters 
pointed out that the statutory language refers to information to be 
included in ``filings'' rather than requiring the information to be 
``filed.'' \385\ Also, the commenters noted that there is some 
information in our ``filings'' that is ``furnished,'' such as Items 
2.02 and 7.01 of Form 8-K, the glossy annual reports to shareholders, 
the audit committee reports (Item 407(d)), the stock performance graphs 
(Item 2.01(e)), the compensation committee reports (Item 407(e)(5)), 
and that executive compensation information in ``filings'' was 
``furnished'' until 2006.\386\
---------------------------------------------------------------------------

    \385\ See, e.g., letters from ABA, COEC I, COEC II, Corporate 
Secretaries, PNC Financial Services, and RILA.
    \386\ See, e.g., letters from ABA, Business Roundtable I, COEC 
I, General Mills, and Mercer I.
---------------------------------------------------------------------------

    One commenter recommended that, in the event that the pay ratio 
disclosure must be ``filed,'' we consider providing a ``safe harbor'' 
excluding the disclosure from the portion of a registrant's filings 
that must be certified pursuant to Exchange Act Rules 13a-14 and 15d-14 
and are also subject to Section 906 of the Sarbanes-Oxley Act of 
2002.\387\ Additionally, another commenter recommended that, at least 
initially, we make the pay ratio disclosure an addendum to documents 
required under Regulation S-K and have that addendum deemed 
``furnished.'' \388\ This commenter indicated that this approach could 
minimize some of the rule's costs and burdens.
---------------------------------------------------------------------------

    \387\ See letter from ABA.
    \388\ See letter from Chamber I.
---------------------------------------------------------------------------

c. Final Rule
    The final rule treats the pay ratio disclosure, as with other Item 
402 information, as ``filed'' for purposes of the Securities Act and 
Exchange Act, and, therefore, subject to potential liabilities under 
those statutes, including Exchange Act Section 18 liability.\389\ 
Information required to be disclosed by registrants pursuant to the 
federal securities laws generally is filed with us and subject to the 
liabilities thereunder, unless a specific exception applies. Although 
we recognize that identifying the median employee and calculating total 
compensation may require registrants to review a large amount of data 
and make a significant number of estimates, assumptions, and judgment 
calls, we do not believe this fact alone justifies exempting this 
information from being ``filed.'' Many of the disclosures required by 
our rules require complex calculations and estimates. Moreover, the 
fact that registrants will be required to provide disclosure about how 
they have arrived at their pay ratio calculations, and in particular 
the required disclosure about the assumptions and methodologies 
underlying the calculations, will permit registrants to clearly explain 
to shareholders where potential imprecisions may be introduced into the 
reported statistic. We also note that all

[[Page 50144]]

other Item 402 information is considered ``filed'' rather than 
``furnished'' and that the disclosure called for by Item 402(u)--
information pertaining to the registrant's operations and workforce 
composition--differs from the types of information we typically permit 
to be ``furnished'' rather than ``filed.'' \390\ For similar reasons, 
we do not believe these disclosures should be exempted from the 
certification requirements of Exchange Act Rules 13a-14 and 15d-14 or 
Section 906 of the Sarbanes-Oxley Act of 2002 or be provided as an 
addendum to filings referenced in Regulation S-K.
---------------------------------------------------------------------------

    \389\ 15 U.S.C. 78r.
    \390\ Some examples of information that are ``furnished'' 
include: the Results of Operations and Financial Condition 
information (Item 2.02 of Form 8-K); Regulation FD disclosures (Item 
7.01 of Form 8-K); the Stock Performance Graph (Item 201(e) of 
Regulation S-K); the Audit Committee Report (Item 407(d) of 
Regulation S-K); the Compensation Committee Report (Item 407(e)(5) 
of Regulation S-K); and the annual reports to shareholders (Rules 
14a-3(b) and 14c-3(a) under the Exchange Act and General Instruction 
G(2) to Form 10-K).
---------------------------------------------------------------------------

    In addition, we note that Section 18 of the Exchange Act does not 
create strict liability for ``filed'' information. Rather, it states 
that a person shall not be liable for misleading statements in a filed 
document if it can establish that it acted in good faith and had no 
knowledge that the statement was false or misleading.\391\ A plaintiff 
asserting a claim under Section 18 would need to meet the elements of 
the statute to establish a claim, including purchasing or selling a 
security at a price that was affected by the false or misleading 
statement in reliance on the misstatement, and damages caused by that 
reliance. Finally, regardless of whether the information is ``filed'' 
or ``furnished,'' registrants that fail to comply with the final rule 
could also be violating Exchange Act Sections 13(a) and 15(d), as 
applicable, and would also be subject to potential liability under 
Exchange Act Section 10(b) \392\ and Rule 10b-5,\393\ promulgated 
thereunder, for any false or misleading material statements in the 
information disclosed pursuant to the rule.
---------------------------------------------------------------------------

    \391\ Exchange Act Section 18(a) provides that any person who 
shall make or cause to be made any statement in any application, 
report, or document filed pursuant to this title or any rule or 
regulation thereunder or any undertaking contained in a registration 
statement as provided in subsection (d) of section 15 of this title, 
which statement was at the time and in the light of the 
circumstances under which it was made false or misleading with 
respect to any material fact, shall be liable to any person (not 
knowing that such statement was false or misleading) who, in 
reliance upon such statement shall have purchased or sold a security 
at a price which was affected by such statement, for damages caused 
by such reliance, unless the person sued shall prove that he acted 
in good faith and had no knowledge that such statement was false or 
misleading. A person seeking to enforce such liability may sue at 
law or in equity in any court of competent jurisdiction. In any such 
suit the court may, in its discretion, require an undertaking for 
the payment of the costs of such suit, and assess reasonable costs, 
including reasonable attorneys' fees, against either party litigant.
    \392\ 15 U.S.C. 78j.
    \393\ 17 CFR 240.10b-5
---------------------------------------------------------------------------

6. Timing of Disclosure
a. Updating Pay Ratio Disclosure for the Last Completed Fiscal Year
i. Proposed Rule
    The proposed rule would not have required the pay ratio for the 
registrant's last completed fiscal year to be disclosed until the 
filing of its annual report on Form 10-K for that fiscal year or, if 
later, the filing of a definitive proxy or information statement 
relating to its next annual meeting of shareholders (or written 
consents in lieu of such a meeting) following the end of such fiscal 
year. The proposed rule would require pay ratio information to be 
filed, in any event, not later than 120 days after the end of such 
fiscal year as provided in General Instruction G(3) of Form 10-K. Also, 
in any filing a registrant made after the end of its last completed 
fiscal year and before the filing of such Form 10-K or proxy or 
information statement, as applicable, a registrant that was subject to 
the proposed rule for the fiscal year prior to the last completed 
fiscal year would be permitted to include or incorporate by reference 
the pay ratio disclosure information for that prior fiscal year. We 
proposed this provision because, as discussed above, the proposed rule 
would require annual total compensation amounts used in the ratio to be 
calculated for the registrant's last completed fiscal year. In 
addition, pay ratio disclosure would be required in any filing by the 
registrant that required Item 402 disclosure.
    Although the annual update of the pay ratio was not required to be 
disclosed until the filing of an annual report for the last completed 
fiscal year, or if later, the filing of a definitive proxy statement or 
information statement relating to the registrant's annual meeting of 
shareholders, this provision would not have altered the requirements 
for Item 402 disclosure under Item 8 of Schedule 14A in other proxy or 
information statement filings.
ii. Comments on the Proposed Rule
    Some commenters generally agreed with the proposed rule's 
requirement that the pay ratio disclosure be updated no earlier than 
the filing of a registrant's annual report on Form 10-K or, if later, 
the filing of a proxy or information statement for the registrant's 
annual meeting of shareholders (or written consents in lieu of such a 
meeting), and in any event no later than 120 days after the end of its 
fiscal year.\394\ One of these commenters agreed with the proposal that 
all registrants, should be allowed to file their pay ratio disclosure 
no later than 120 days after fiscal year end by either an amended Form 
10-K or Form 8-K because permitting the later filing would allow 
registrants not subject to the proxy rules to have the same amount of 
time to file their pay ratio disclosure as filers that are subject to 
the proxy rules.\395\ One commenter indicated that, although timing of 
the disclosure is ``not important,'' the information should be required 
in the registrant's annual report on Form 10-K and permitted in other 
filings.\396\ Another commenter contended that the information should 
at least be available in a registrant's proxy statement for the annual 
meeting so that shareholders may use the information for voting.\397\
---------------------------------------------------------------------------

    \394\ See, e.g., letters from ABA, CalPERS, Calvert, CII, E&Y, 
Trillium II, UAW Trust, US SIF, and Vectren Corp.
    \395\ See letter from ABA.
    \396\ See letter from Capital Strategies.
    \397\ See letter from UAW Trust.
---------------------------------------------------------------------------

    One commenter stated that it would not object to the proposed delay 
because it would not diminish the usefulness of the disclosure to 
investors.\398\ The commenter, however, noted that the proposed delay 
still might not provide registrants enough time after the end of the 
fiscal year for all registrants to calculate and disclose the pay ratio 
in their annual proxy statement. Therefore, the commenter stated that 
it ``generally would not object to the rules providing for some 
additional accommodation to the extent that it does not significantly 
diminish the usefulness of the disclosure to investors.''\399\
---------------------------------------------------------------------------

    \398\ See letter from CII.
    \399\ Id.
---------------------------------------------------------------------------

    Other commenters disagreed with the proposed rule's requirement 
that registrants disclose their pay ratio information on Form 10-K, the 
proxy or information statement, or 120 days after the end of its fiscal 
year. Mainly, these commenters believed the requirement would not 
provide sufficient time for registrants to identify the median 
employee, calculate total compensation and the pay ratio, and file 
their information.\400\ Most of these commenters recommended that the 
final rule permit disclosure on Form 8-K at

[[Page 50145]]

some other time of the year, including when the information is able to 
be calculated,\401\ ``within some extended period (such as 180 days 
after fiscal year end, as is the case for Form 11-Ks and other 
reports),''\402\ any time during the first five months after fiscal 
year-end,\403\ before the end of the registrant's second quarter,\404\ 
and 14 days before the annual meeting of shareholders.\405\
---------------------------------------------------------------------------

    \400\ See, e.g., letters from American Benefits Council, Brian 
Foley & Co., Chesapeake Utilities, Frederic W. Cook & Co., Hyster-
Yale, Mercer I, NACCO, and PM&P.
    \401\ See, e.g., letters from Chesapeake Utilities, Mercer I, 
and PM&P.
    \402\ See letter from American Benefits Council.
    \403\ See letter from Brian Foley & Co.
    \404\ See letter from Frederic W. Cook & Co.
    \405\ See, e.g., letters from Hyster-Yale and NACCO.
---------------------------------------------------------------------------

iii. Final Rule
    The final rule does not require registrants to provide the pay 
ratio disclosure information for the registrant's last completed fiscal 
year until it files its annual report on Form 10-K for that year or, if 
later, it files the definitive proxy or information statement relating 
to its next annual meeting of shareholders (or written consents in lieu 
of such a meeting). In any event, the final rule requires registrants 
to file their pay ratio information not later than 120 days after the 
end of such fiscal year in a manner similar to General Instruction G(3) 
of Form 10-K.\406\ This requirement is consistent with the proposal. 
Also, consistent with the proposed rule, a registration statement that 
incorporates by reference a Form 10-K (or amended Form 10-K) containing 
all Part III information other than updated pay ratio information could 
be declared effective before the registrant's definitive proxy or 
information statement containing updated pay ratio information is filed 
in accordance with General Instruction G(3).
---------------------------------------------------------------------------

    \406\ General Instruction G(3) of Form 10-K permits registrants 
to incorporate by reference Part III of its Form 10-K, which 
includes the Item 402 information, from their definitive proxy or 
information statements filed in connection with the registrant's 
annual meeting if such definitive proxy or information statements 
are filed within 120 days after the end of the fiscal year covered 
by the Form 10-K. If a definitive proxy or information statement is 
not filed within this 120-day period, Items comprising the Part III 
information must be filed as part of the Form 10-K, or as an 
amendment to the Form l0-K, not later than the end of the 120-day 
period.
---------------------------------------------------------------------------

    Additionally, although the annual update is not required to be 
disclosed until the filing of an annual report for the last completed 
fiscal year, or if later, the filing of a definitive proxy statement or 
information statement relating to the registrant's annual meeting of 
shareholders, as discussed in the Proposing Release, this provision 
does not alter the requirements for Item 402 disclosure under Item 8 of 
Schedule 14A in other proxy or information statement filings. For 
example, if a registrant filed a proxy statement (other than the 
definitive proxy statement for its annual meeting) that required Item 
402 information pursuant to Item 8 of Schedule 14A, the registrant 
would be required to include or incorporate by reference pay ratio 
disclosure for the most recent period that had been filed in its Form 
10-K or definitive proxy statement for its annual meeting.
    We continue to believe this provision is appropriate for the 
reasons discussed in the Proposing Release. Without it, a registrant 
would be required to include its pay ratio disclosure in a filing (such 
as a registration statement) filed after the end of the prior fiscal 
year, but before it was able to compile its executive compensation 
information for that fiscal year, which is usually included in a 
registrant's proxy statement relating to its annual meeting of 
shareholders following the end of the fiscal year, which could raise 
additional incremental costs for registrants that elect to provide 
executive compensation disclosure in their annual proxy statement 
rather than their annual report and for registrants that are conducting 
registered offerings at the beginning of their fiscal year.
    We note that a number of commenters agreed with our approach. In 
response to other comments stating that our approach will not provide 
registrants sufficient time to identify the median employee, calculate 
total compensation and the pay ratio, and file their information, we 
note that the final rule retains the significant flexibility afforded 
to registrants in the proposal and includes several additional 
accommodations intended to reduce the burdens of producing the required 
disclosure. We believe these provisions will make it feasible for 
registrants to file their pay ratio disclosure within the timeframes 
set forth in the final rule. We do not believe it would be appropriate 
to extend the deadline by which the pay ratio disclosure should be 
updated in light of its relevance to shareholders in making their 
voting decisions under Section 951 of the Dodd-Frank Act. If 
registrants were not required to provide the pay ratio disclosure when 
they file their annual report on Form 10-K or, if later, the definitive 
proxy or information statement for their next annual meeting of 
shareholders (or written consent in lieu of such a meeting), this could 
result in the disclosure not being presented together with other 
relevant executive compensation information to which it relates and not 
being available to inform shareholders as they exercise their say-on-
pay voting rights, which we understand to be the disclosure's primary 
purpose. For all of these reasons, we believe the timing requirements 
in the final rule are reasonable and appropriate.
b. Omitting Salary or Bonus Information for the PEO in Reliance on 
Instruction 1 to Item 402(c)(2)(iii) and (iv), and Technical Amendment 
to Item 5.02(f) of Form 8-K
i. Proposed Rule
    In cases where a registrant is relying on Instruction 1 to Items 
402(c)(2)(iii) and (iv) of Regulation S-K to omit salary or bonus of 
the PEO that is not calculable until a later date, the proposed rule 
would permit registrants to omit pay ratio disclosure until those 
elements of the PEO's total compensation are determined. The proposed 
rule would also have required registrants relying on that instruction 
to provide their pay ratio disclosure in the same Form 8-K filing in 
which the PEO's salary or bonus is disclosed. We proposed a conforming 
amendment to Item 5.02(f) of Form 8-K to reflect the addition of this 
pay ratio disclosure requirement. Although a filing is triggered under 
Item 5.02(f) when the omitted salary or bonus becomes calculable in 
whole or in part, under the proposed amendment to Form 8-K, the pay 
ratio information would be required only when the salary or bonus 
became calculable in whole, which would avoid the need for multiple 
updates to the pay ratio disclosure until the final total compensation 
amount for the PEO is known.
ii. Comments on the Proposed Rule
    Only a few commenters discussed this proposed instruction and they 
generally agreed with the proposed rule.\407\ One commenter contended 
that, if the registrant is relying on Instruction 1 to Item 
402(c)(2)(iii) and (iv), the final rule should require neither an 
estimate of compensation nor additional supplemental disclosure prior 
to the Item 5.02 8-K because it would further dilute the utility of the 
pay ratio information for shareholders.\408\ One commenter suggested 
that delaying the pay ratio information under Instruction 1 to Item 
402(c)(2)(iii) and (iv) would diminish the information's usefulness, 
but did not object to the proposed instruction because it would only 
affect a small number of registrants.\409\
---------------------------------------------------------------------------

    \407\ See, e.g., letters from ABA, Capital Strategies, and CII.
    \408\ See letter from ABA.
    \409\ See letter from CII.

---------------------------------------------------------------------------

[[Page 50146]]

iii. Final Rule
    As proposed, the final rule permits registrants to omit pay ratio 
disclosure until the salary or bonus of their PEO's total compensation 
is determined in cases in which the registrant is relying on 
Instruction 1 to Items 402(c)(2)(iii) and (iv) of Regulation S-K \410\ 
to omit the salary or bonus of the PEO that is not calculable until a 
later date. Commenters on this provision generally agreed with our 
approach. The final rule also includes a conforming amendment to Item 
5.02(f) of Form 8-K to reflect the addition of this pay ratio 
disclosure requirement. However, although a filing is triggered under 
Item 5.02(f) when the PEO's omitted salary or bonus becomes calculable 
in whole or in part, under the conforming amendment to Form 8-K, the 
pay ratio information is required only when the salary or bonus become 
calculable in whole, which avoids the need for multiple updates to the 
pay ratio disclosure until the final total compensation amount for the 
PEO is known.
---------------------------------------------------------------------------

    \410\ Instruction 1 to Items 402(c)(2)(iii) and (iv) of 
Regulation S-K, under our existing executive compensation disclosure 
rules, permits registrants to omit disclosure in the Summary 
Compensation Table of the salary or bonus of a named executive 
officer if it is not calculable as of the latest practicable date. 
In that circumstance, the registrant must include a footnote 
disclosing that fact and providing the date that the amount is 
expected to be determined, and the amount must be disclosed at that 
time by filing a Form 8-K. Item 5.02(f) of Form 8-K sets forth the 
requirements for the filing of information that was omitted from 
Item 402 disclosure in accordance with Instruction 1 to Items 
402(c)(2)(iii) and (iv), including the requirement to include a new 
total compensation figure for the named executive officer.
---------------------------------------------------------------------------

    The final rule includes an instruction that provides that a 
registrant relying on Instruction 1 to Items 402(c)(2)(iii) and (iv) 
with respect to the salary or bonus of the PEO would be required to 
disclose that the pay ratio disclosure is not calculable until the PEO 
salary or bonus, as applicable, is determined and disclose the date 
that the PEO's actual total compensation is expected to be determined. 
The instruction also requires the registrant to include its pay ratio 
disclosure in the filing on Form 8-K that includes the omitted salary 
or bonus information as contemplated by Instruction 1 to Items 
402(c)(2)(iii) and (iv).
    We believe, as stated in the Proposing Release, that the potential 
benefits of the complete and up-to-date pay ratio disclosure could be 
diminished if the pay ratio were to be calculated using less than the 
entire amount of the PEO's total compensation for the period and that 
these potential benefits could justify the potential costs to 
shareholders of a delay in the timing of the disclosure. For example, 
in some cases, the amount of compensation that is omitted under 
Instruction 1 to Items 402(c)(2)(iii) and (iv) could be significant, 
and, therefore, the pay ratio would be lower if presented using that 
incomplete compensation amount. Similarly, we believe that the 
potential benefits of the complete and up-to-date pay ratio disclosure 
could be diminished if the registrant used the prior year's pay ratio 
information to calculate an approximate pay ratio for the current year, 
especially if there is a significant change to the PEO's compensation 
from the prior year. Also, based on the number of registrants that have 
historically relied on Instruction 1 to Items 402(c)(2)(iii) and 
(iv),\411\ we do not expect that the instruction will impact a 
significant number of registrants each year.
---------------------------------------------------------------------------

    \411\ For example, based on a review of EDGAR filings for 
calendar years 2012 and 2013, we estimate that approximately 11 
Forms 8-K are filed pursuant to Item 5.02(f) annually and 
approximately 90% of these relate to disclosure of PEO compensation.
---------------------------------------------------------------------------

c. Initial Compliance Date
i. Proposed Rule
    We proposed to require a registrant to comply with proposed Item 
402(u) with respect to compensation for the registrant's first fiscal 
year commencing on or after the effective date of the rule. We also 
proposed to permit a registrant to omit this initial pay ratio 
disclosure until the filing of its annual report on Form 10-K for that 
fiscal year or, if later, the filing of a proxy or information 
statement for its next annual meeting of shareholders (or written 
consents in lieu of a meeting) following the end of such year. In any 
event, the information would be required to be filed not later than 120 
days after the end of such fiscal year. We recognized in the Proposing 
Release that a transition period would likely be needed by large, 
multinational registrants and any registrants that did not have a 
centralized, consolidated payroll, benefits, and pension system that 
captures the information necessary to identify the median. We expected 
that it would take registrants one full reporting cycle to implement 
and test any necessary systems.
ii. Comments on the Proposed Rule
    One commenter disagreed with the initial transition period in the 
proposed rule on the grounds that further delays in having access to 
the pay ratio disclosure are not in the best interests of 
shareholders.\412\ Other commenters contended that the transition 
period in the Proposing Release would disadvantage registrants with 
fiscal years that end on or close to the effective date of the final 
rule and suggested that the transition period be extended until:
---------------------------------------------------------------------------

    \412\ See letter from IPS.
---------------------------------------------------------------------------

     a registrant's first fiscal year commencing on or after 
six months following the effective date of the final rule; \413\
---------------------------------------------------------------------------

    \413\ See, e.g., letters from American Benefits Council, COEC I, 
Frederic W. Cook & Co., and Microsoft.
---------------------------------------------------------------------------

     a registrant's first fiscal year commencing one year after 
the effective date of the final rule; \414\
---------------------------------------------------------------------------

    \414\ See, e.g., letters from AAFA II and NRF.
---------------------------------------------------------------------------

     a registrant's first fiscal year commencing after the 
second anniversary of the effective date of the final rule (or, 
alternatively, a registrant's first fiscal year commencing on or after 
December 15 of the year in which the rule becomes effective); \415\
---------------------------------------------------------------------------

    \415\ See letter from ABA.
---------------------------------------------------------------------------

     a registrant's first fiscal year commencing on or after 
the first January 1 after the effective date of the final rule; \416\
---------------------------------------------------------------------------

    \416\ See, e.g., letters from Best Buy et al., Corporate 
Secretaries, General Mills, Meridian, PM&P, and SH&P.
---------------------------------------------------------------------------

     a registrant's 2016 fiscal year, if the final rule is 
adopted in 2014; \417\
---------------------------------------------------------------------------

    \417\ See, e.g., letters from Chesapeake Utilities, Intel, and 
Mercer I.
---------------------------------------------------------------------------

     a registrant's 2017 fiscal year; \418\
---------------------------------------------------------------------------

    \418\ See, e.g., letters from Hay Group, Hyster-Yale, and NACCO.
---------------------------------------------------------------------------

     one year after the Proposing Release's compliance date 
(i.e., one year after a registrant's first fiscal year commencing on or 
after the effective date of the final rule); \419\
---------------------------------------------------------------------------

    \419\ See letter from RILA.
---------------------------------------------------------------------------

     two full years after the effective date of the final rule; 
\420\ and
---------------------------------------------------------------------------

    \420\ See, e.g., letters from Business Roundtable I, Eaton, and 
SHRM.
---------------------------------------------------------------------------

     three years after the effective date of the final 
rule.\421\
---------------------------------------------------------------------------

    \421\ See letter from FSI.
---------------------------------------------------------------------------

    One commenter recommended delaying compliance with ``the most 
onerous parts of this rule,'' \422\ and a further commenter requested 
that we phase in various requirements of the rule.\423\ Neither of 
these commenters, however, was more specific as to which parts of the 
rule to delay or phase-in. One commenter suggested that we include a 
three-year sunset provision in the final rule.\424\ Other commenters 
suggested various transition periods for companies with non-U.S. 
employees. Some commenters requested that, if the final rule includes 
non-U.S. employees,

[[Page 50147]]

we permit registrants to exclude non-U.S. employees from the pay ratio 
for an additional two years.\425\ A few commenters recommended that we 
extend the transition period for multinational registrants to permit 
them to begin to comply with the final rules with respect to 
compensation for their first full fiscal year commencing on or after 
the second anniversary of the effective date of the final rules 
(assuming foreign employees are not excluded from the ``median 
employee'' determination).\426\ One commenter urging a transition 
period for non-U.S. employees stated that ``a staged implementation 
would allow companies to design methodologies for pay ratio compliance 
during the first year and test them on an employee population where 
data collection is more manageable.'' \427\
---------------------------------------------------------------------------

    \422\ See letter from Semtech.
    \423\ See letter from Chamber I.
    \424\ See letter from COEC I.
    \425\ See, e.g., letters from FSR (suggesting that the final 
rule permit registrants with non-U.S. employees at least two full 
fiscal years to comply, and providing two suggestions for doing so: 
(1) providing a ``transition period during which the registrant may 
report its pay ratio disclosure solely on the basis of the data 
available in respect of its employees based in the United States;'' 
or (2) providing a transition period for any registrant with more 
than a de minimis non-U.S. workforce), FuelCell Energy (requesting 
that we ``provide an additional two years before companies must 
include overseas workers in their pay ratio calculations''), Garmin 
(same), NIRI (same), and Semtech (same).
    \426\ See letter from ABA and American Benefits Council.
    \427\ See letter from COEC I.
---------------------------------------------------------------------------

iii. Final Rule
    The final rule provides that registrants' first reporting period is 
their first full fiscal year beginning on or after January 1, 2017, 
instead of on or after the effective date of the rule, as proposed. For 
example, the reporting period for a company with a fiscal year that 
ends on December 31 will begin on January 1, 2017.\428\ We believe a 
transition period is appropriate because, as we noted in the Proposing 
Release, certain registrants may need additional time to implement 
systems to compile the disclosure and verify its accuracy.
---------------------------------------------------------------------------

    \428\ Approximately 70% of registrants have fiscal years that 
begin on January 1. We determined this figure based on the number of 
current reporting companies. There are 8,529 total registrants, and 
5,799 of these registrants have a fiscal year end of December 31, 
which is approximately 68% (5,799/8,529=.67991).
---------------------------------------------------------------------------

    We are changing our approach from the proposal because a number of 
commenters contended that the proposed transition period would be 
burdensome to registrants, and would particularly disadvantage 
registrants with fiscal years that end on or close to the effective 
date of the final rule.\429\ Additionally, a number of these commenters 
indicated at least another additional year would be required for 
registrants to establish systems to comply with the final rule.\430\ 
One commenter claimed, in particular, that registrants would need ``an 
initial year to establish and test the systems that may be necessary to 
collect and analyze the data required to identify their median employee 
and develop the necessary disclosure controls and procedures, and then 
a second year involving a full reporting cycle to actually put their 
selected system into operation.'' \431\
---------------------------------------------------------------------------

    \429\ See, e.g., letters from AAFA II, ABA, American Benefits 
Council, Best Buy et al., Business Roundtable I, Chamber I, 
Chesapeake Utilities, COEC I, Corporate Secretaries, Eaton, Frederic 
W. Cook & Co., FSI, General Mills, Hay Group, Hyster-Yale, Intel, 
Mercer I, Meridian, Microsoft, NACCO, NRF, PM&P, RILA, Semtech, 
SH&P, and SHRM.
    \430\ See, e.g., letters from ABA, Business Roundtable I, 
Chesapeake Utilities, Eaton, FSI, Hay Group, Hyster-Yale, Intel, 
Mercer I, NACCO, and SHRM.
    \431\ See letter from ABA.
---------------------------------------------------------------------------

    We are not providing an additional transition period or staggered 
compliance for registrants with non-U.S. employees, as requested by 
some commenters. We believe that the final rule provides sufficient 
time for all registrants, including multinationals and those with non-
U.S. employees, to identify the median employee and calculate annual 
total compensation for that employee and the PEO. Additionally, we note 
that the de minimis and foreign privacy law exemptions to the 
definition of ``employee'' in the final rule may help reduce the burden 
on such registrants in preparing the necessary disclosure.
d. Transition Period for New Registrants
i. Proposed Rule
    The proposed rule would permit new registrants to delay compliance 
so that the pay ratio disclosure would not be required in a 
registration statement on Form S-1 \432\ or Form S-11 \433\ for an 
initial public offering or registration statement on Form 10.\434\ Such 
registrants would be required to comply with proposed Item 402(u) with 
respect to compensation for the first fiscal year commencing on or 
after the date the registrant became subject to the requirements of 
Section 13(a) or Section 15(d) of the Exchange Act, and the registrant 
could omit this initial pay ratio disclosure from its filings until the 
filing of its Form 10-K for such fiscal year or, if later, the filing 
of a proxy or information statement for its next annual meeting of 
shareholders (or written consents in lieu of a meeting) following the 
end of such fiscal year. Similar to the proposed instructions for 
updating pay ratio disclosure, these proposed instructions also would 
require that this initial pay ratio disclosure be filed, in any event, 
as provided in connection with General Instruction G(3) of Form 10-K 
not later than 120 days after the end of such fiscal year.
---------------------------------------------------------------------------

    \432\ 17 CFR 239.11.
    \433\ 17 CFR 239.18.
    \434\ 17 CFR 249.210.
---------------------------------------------------------------------------

ii. Comments on the Proposed Rule
    All the commenters that discussed the topic agreed that new 
registrants should not be required to provide pay ratio disclosure in 
their initial registration statements on Form S-1, Form S-11, or Form 
10.\435\ A few commenters also agreed that a new public company should 
not have to provide any pay ratio disclosure until it completes its 
first full fiscal year as a public company.\436\
---------------------------------------------------------------------------

    \435\ See, e.g., letters from ABA, CalPERS, CII, Corporate 
Secretaries, Hyster-Yale, NACCO and PM&P.
    \436\ See, e.g., letters from ABA and Hyster-Yale.
---------------------------------------------------------------------------

iii. Final Rule
    Similar to the transition period for existing registrants, the 
final rule provides that a new registrant's first pay ratio disclosure 
must follow its first full fiscal year beginning after the registrant 
has (i) been subject to the requirements of Sections 13(a) or 15(d) of 
the Exchange Act for a period of at least twelve calendar months 
beginning on or after January 1, 2017 and (ii) filed at least one 
annual report pursuant to Sections 13(a) or 15(d) of the Exchange Act 
that does not contain the pay ratio disclosure.\437\
---------------------------------------------------------------------------

    \437\ For example, company with a fiscal year ending on December 
31 that completes its initial public offering on March 1, 2017 will 
not be required to include any pay ratio information in its 
registration statement on Form S-1. The registrant will be first 
required to include pay ratio disclosure in its Form 10-K for its 
2018 fiscal year or its definitive proxy or information statement 
for its 2019 annual meeting of shareholders, but no later than 120 
days following the end of its 2018 fiscal year. The registrant's pay 
ratio disclosure will be required for its 2018 fiscal year because 
it filed its registration statement after January 1, 2017 (March 1, 
2017), it will have been subject to the requirements of Sections 
13(a) or 15(d) of the Exchange Act for a period of at least twelve 
calendar months (March 1, 2017 to March 1, 2018), and it will have 
filed at least one annual report pursuant to Sections 13(a) or 15(d) 
of the Exchange Act (for fiscal year 2017).
---------------------------------------------------------------------------

    This change aligns the transition for new registrants with the 
change we made to the initial transition period for existing 
registrants. As discussed above, the final rule provides that a 
registrant's first reporting period is its first full fiscal year 
beginning on or after January 1, 2017, instead of on or after the 
effective date of the rule, as proposed. Also, this change is 
consistent with

[[Page 50148]]

some commenters' recommendation that new registrants not be required to 
provide any pay ratio disclosure until they complete their first full 
fiscal year as a public company.\438\
---------------------------------------------------------------------------

    \438\ See, e.g., letters from ABA and Hyster-Yale.
---------------------------------------------------------------------------

    Additionally, as proposed, the final rule does not require the pay 
ratio to be disclosed in a registration statement on Form S-1 or Form 
S-11 for an initial public offering or an initial registration 
statement on Form 10. Also, new registrants are permitted to omit their 
pay ratio disclosure from their filings until after the later of (i) 
their first full fiscal year beginning on the date they first become 
subject to the requirements of Section 13(a) or 15(d) of the Exchange 
Act and (ii) January 1, 2017. All commenters that discussed the topic 
agreed that new registrants should not be required to provide pay ratio 
disclosure in their initial registration statements on Form S-1, Form 
S-11, or Form 10.\439\
---------------------------------------------------------------------------

    \439\ See, e.g., letters from ABA, CalPERS, CII, Corporate 
Secretaries, Hyster-Yale, NACCO and PM&P.
---------------------------------------------------------------------------

    We noted in the Proposing Release that shareholders might benefit 
from pay ratio disclosure in connection with an initial public offering 
or Exchange Act registration. Even so, we continue to believe that it 
is appropriate to give companies time to develop any needed systems to 
compile the disclosure and verify its accuracy. This is particularly so 
since we believe the primary purpose of the pay ratio disclosure is to 
provide a useful data point for shareholders in making their voting 
decisions on executive compensation, including their say-on-pay votes, 
which is unlikely to occur for those registrants until at least a year 
after the initial public offering has occurred. The transition period 
for new registrants is similar to the time frame provided for other 
registrants to comply with pay ratio disclosure requirements following 
the effective date of the final rule.
    As we stated in the Proposing Release, we are sensitive to the 
impact that a rule could have on capital formation. Section 953(b), as 
amended by the JOBS Act, distinguished between certain newly public 
companies and all other registrants by providing an exemption for 
emerging growth companies. We note further that, without a transition 
period, the incremental time needed to compile pay ratio disclosure 
could cause companies that are not emerging growth companies to delay 
an initial public offering, which could have a negative impact on 
capital formation. In this regard, we assume that companies that are no 
longer eligible for emerging growth company status are likely to be 
businesses with more extensive operations or a greater number of 
employees, which could increase the initial efforts needed to comply 
with the proposed requirements. We continue to believe that providing a 
transition period for these newly public companies could mitigate this 
potential impact on capital formation and will not significantly affect 
shareholders' ability to assess the compensation and accountability of 
a registrant's executives.
e. Additional Transition Periods
i. Proposed Rule
    We did not propose a transition period for registrants that cease 
to be smaller reporting companies or emerging growth companies or 
engage in business combinations and/or acquisitions. We did, however, 
request comment on whether there should be such transition periods and 
the appropriate length of time for any such transition period.
ii. Comments on the Proposed Rule
    One commenter recommended a transition period for registrants that 
cease to be smaller reporting companies.\440\ The commenter recommended 
that those registrants not be required to provide their pay ratio 
disclosure until the first full fiscal year commencing on or after the 
first anniversary of the end of the fiscal year in which the registrant 
is no longer a smaller reporting company.
---------------------------------------------------------------------------

    \440\ See letter from ABA.
---------------------------------------------------------------------------

    Several commenters supported generally a transition period for 
registrants that engage in a business combination and/or an acquisition 
to delay including any new employees acquired in the transaction in the 
acquirer's pay ratio.\441\ Other commenters suggested specific 
transition periods for such registrants, including: six months after 
the end of the fiscal year in which the transaction closes; \442\ six 
or more months after the transaction; \443\ one full fiscal year 
following the transaction; \444\ the fiscal year beginning 18 months 
after closing of the transaction; \445\ and three years after the 
transaction.\446\
---------------------------------------------------------------------------

    \441\ See, e.g., letters from Eaton, Hyster-Yale, NAM I, NAM II, 
and PM&P.
    \442\ See letter from Brian Foley & Co.
    \443\ See id. (providing an alternative to initial 
recommendation of six months after the end of the fiscal year in 
which the transaction closes).
    \444\ See, e.g., letters from ABA, Business Roundtable I, 
Chesapeake Utilities, Frederic W. Cook & Co., NACCO, and NYC Bar.
    \445\ See letter from Microsoft.
    \446\ See letter from Cummins Inc.
---------------------------------------------------------------------------

iii. Final Rule
    In response to comments, the final rule provides that a registrant 
that ceases to be a smaller reporting company or an emerging growth 
company will not be required to provide pay ratio disclosure until 
after the first full fiscal year after exiting such status and not for 
any fiscal year commencing before January 1, 2017. For example, if a 
calendar year-end smaller reporting company registrant's public float 
exceeds $75 million as of the end of its second fiscal quarter in 2017, 
the registrant will cease to be a smaller reporting company as of the 
beginning of its fiscal year starting on January 1, 2018.\447\ The 
registrant, therefore, must include its pay ratio disclosure in its 
Form 10-K for 2018 or a proxy or information statement for its 2019 
annual meeting of shareholders (or written consents in lieu of a 
meeting) following the end of the 2018 fiscal year, but not later than 
120 days after the end of the 2018 fiscal year. We believe that this 
approach is appropriate because, as commenters noted, smaller reporting 
companies ``will encounter the same challenges in preparing to comply 
with the pay ratio disclosure requirement as registrants generally and, 
therefore, will need time to determine how they will collect the data 
necessary to identify their median employee and prepare the necessary 
disclosure,'' but, ``as relatively small entities, these registrants 
are not likely to need as much time as `regular' (larger) registrants 
to transition to compliance with the pay ratio disclosure 
requirement.'' \448\ This new transition period is consistent with the 
one comment we received on this issue, and we believe it will provide 
sufficient time for these registrants to prepare their disclosure.
---------------------------------------------------------------------------

    \447\ See 17 CFR 229.10(f)(1).
    \448\ See letter from ABA.
---------------------------------------------------------------------------

    Similarly, in 2017, if a calendar year-end emerging growth company 
had total annual gross revenues of $1 billion or more, exceeded the $1 
billion threshold in non-convertible debt for the previous 3-year 
period, has reached the fifth anniversary of the date of the first sale 
of its common equity securities pursuant to an effective registration 
statement under the Securities Act, had not issued $1 billion in non-
convertible debt during the previous 3-year period, or is deemed to be 
a ``large accelerated filer,'' the registrant will cease to be an 
emerging growth company at the beginning of its fiscal year starting on 
January 1, 2018.\449\ The registrant,

[[Page 50149]]

therefore, will be required to include pay ratio disclosure in its Form 
10-K for 2018 or a proxy or information statement for its 2019 annual 
meeting of shareholders (or written consents in lieu of a meeting) 
following the end of the 2018 fiscal year, but not later than 120 days 
after the end of the 2018 fiscal year.
---------------------------------------------------------------------------

    \449\ 17 CFR 229.10(a).
---------------------------------------------------------------------------

    We have decided to adopt this provision because it is consistent 
with a commenter's similar recommendation for a transition period when 
a registrant ceases to be a smaller reporting company. The reasoning 
for the approach for both types of registrants is similar in that 
emerging growth companies will need time to determine how they will 
collect the data necessary to identify their median employee and 
prepare the necessary disclosure.
    The final rule also permits a registrant that engages in a business 
combination and/or an acquisition to omit the employees of a newly-
acquired entity from their pay ratio calculation for the fiscal year in 
which the business combination or acquisition becomes effective. For 
example, for a calendar year-end registrant that engages in a business 
combination and/or acquisition in 2017, the registrant's first period 
for which it will have to include the newly-acquired employees in the 
pay ratio disclosure would be fiscal year 2018, with the disclosure 
included in its Form 10-K for 2018 or a proxy or information statement 
for their next annual meeting of shareholders (or written consents in 
lieu of a meeting) following the end of the 2018 fiscal year, but not 
later than 120 days after the end of such fiscal year.
    A number of commenters recommended a transition period for such 
registrants. Suggestions for the length of the transition period ranged 
from six months \450\ to three years.\451\ The transition period being 
adopted is generally consistent with commenters' suggestions that the 
disclosure not be required until one full fiscal year after the 
transaction. We also believe that the final rule's approach will allow 
a registrant sufficient time to incorporate the payroll, compensation, 
and/or recordkeeping structures of the newly-acquired entity into the 
registrant's pay ratio disclosure framework.
---------------------------------------------------------------------------

    \450\ See letter from Brian Foley & Co. (suggesting that the 
final rule allow registrants conducting a merger/acquisition 
transaction to delay reporting ``until at least 6 months after the 
end of the fiscal year in which the M&A transaction closes, or 6 or 
more months after the closing date'').
    \451\ See letter from Cummins Inc.
---------------------------------------------------------------------------

    Finally, under the provision in the final rule for triennial 
calculations of median employee compensation discussed above, in the 
year of the acquisition or business combination, the registrant need 
not evaluate whether the acquisition or business combination would 
result in a substantial change to its pay ratio disclosure that would 
necessitate the re-identification of the median employee. Rather, 
consistent with the one year transition for incorporating the new 
employees in the pay ratio disclosure, the first time the registrant 
must evaluate whether the business combination or acquisition would 
result in a substantial change to its pay ratio disclosure that would 
necessitate a re-identification of the median employee is in the fiscal 
year following the acquisition or business combination. We believe this 
will provide registrants sufficient time to integrate the new business 
or acquisition. Nevertheless, those registrants must identify the 
acquired business excluded and the approximate number of employees for 
the fiscal year in which the business combination or an acquisition 
becomes effective to provide transparency about what the pay ratio 
disclosure does and does not include.

III. Economic Analysis

A. Background

    We have performed an economic analysis of the main economic effects 
that may result from the final rule, relative to the baseline discussed 
below. Section 2(b) of the Securities Act and Section 3(f) of the 
Exchange Act require us, when engaging in rulemaking that requires us 
to consider or determine whether an action is necessary or appropriate 
in the public interest, also to consider whether the action will 
promote efficiency, competition, and capital formation.\452\ Further, 
Section 23(a)(2) of the Exchange Act requires us, when adopting rules 
under the Exchange Act, to consider the impact that any new rule will 
have on competition and to not adopt any rule that would impose a 
burden on competition that is not necessary or appropriate in 
furtherance of the purposes of the Exchange Act.\453\
---------------------------------------------------------------------------

    \452\ See 15 U.S.C. 77b(b) and15 U.S.C. 78c(f).
    \453\ See 15 U.S.C. 78w(a)(2).
---------------------------------------------------------------------------

    As discussed above, Section 953(b) of the Dodd-Frank Act directs us 
to amend Item 402 of Regulation S-K to add a pay ratio disclosure 
requirement. Section 953(b) imposes a new requirement on registrants to 
disclose the median of the annual total compensation of all employees 
and the ratio of that median to the annual total compensation of the 
CEO. In doing so, Section 953(b) requires registrants to determine 
total compensation in accordance with Item 402(c)(2)(x). Our rules for 
compensation disclosure generally focus on compensation matters that 
relate to executive officers and directors. While registrants subject 
to Item 402 are required to provide extensive information about the 
compensation of the PEO and other named executive officers identified 
pursuant to Item 402(a), current disclosure rules do not require 
registrants to disclose compensation information for other employees in 
their filings with us.
    As directed by Congress, we proposed amendments to Item 402 to 
require the disclosure of the annual total compensation of a 
registrant's PEO, the median annual total compensation of all employees 
of that registrant (excluding the PEO), and the ratio of the median 
annual total compensation of all employees to the annual total 
compensation of the PEO. We considered the statutory language and 
exercised our discretion to develop a proposal designed to lower 
compliance costs while remaining consistent with the mandate of Section 
953(b). In particular, among other things, we proposed a rule that 
would permit registrants to use reasonable estimates to identify the 
median employee, including by using statistical sampling, and a 
consistently applied compensation measure (such as payroll or tax 
records). The proposed rule would also allow the use of reasonable 
estimates in calculating the annual total compensation or any elements 
of total compensation for employees. The proposed flexible approach was 
aimed at decreasing compliance costs while taking into consideration a 
registrant's particular facts and circumstances. We received thousands 
of comment letters in response to the proposal.
    To satisfy the statutory mandate of Section 953(b), we are adopting 
amendments to Item 402 substantially as proposed, with modifications 
intended to address some of the concerns raised by commenters and 
provide further flexibility in the determination of the pay ratio. We 
believe the primary benefit that Congress intended with pay ratio 
disclosure is to provide shareholders with a company-specific metric 
that they can use to inform their voting decisions regarding executive 
compensation under Section 951 of the Dodd-Frank Act. Several 
commenters stated affirmatively that they would find the new data 
points, including pay ratio disclosure, relevant and useful when

[[Page 50150]]

making voting decisions.\454\ As discussed above, while neither the 
statute nor the related legislative history directly states the 
objectives or intended benefits of the provision, we believe, based on 
our analysis of the statute and comments received, that Section 953(b) 
was intended to provide shareholders with a company-specific metric 
that can assist in their evaluation of a registrant's executive 
compensation practices.
---------------------------------------------------------------------------

    \454\ See, e.g., letters from AFL-CIO I, AFSCME, Amalgamated, 
Bricklayers International, CalSTRS, Calvert, Chevy Chase Trust, 
CorpGov.net, Form Letter C, Form Letter D, Form Letter E, Form 
Letter F, LIUNA, LAPFF, NY State Comptroller, Pax World Funds, 
Public Citizen I, Rep. Ellison et al. I, Rep. Ellison et al. II, 
Trillium I, Trillium II, UAW Trust, and US SIF.
---------------------------------------------------------------------------

    We are sensitive to the costs and benefits that stem from the final 
rule. Some of the costs and benefits stem directly from the statutory 
mandate in Section 953(b), while others are affected by the discretion 
we exercise in implementing that mandate. Our economic analysis of the 
final rule addresses both the costs and benefits that stem directly 
from the mandate of Section 953(b) and those arising from the policy 
choices made using our discretion, recognizing that it may be difficult 
to separate the discretionary aspects of the rule from those elements 
required by statute.
    In the economic analysis that follows, we first examine the current 
regulatory and economic landscape to form a baseline for our analysis. 
We then analyze the likely economic effects--including benefits and 
costs and impact on efficiency, competition, and capital formation--
arising from the new mandatory disclosure requirement prescribed by the 
Dodd-Frank Act and from the choices we have made in exercising our 
discretion, relative to the baseline discussed below.

B. Baseline

    To assess the economic impact of the final rule, we are using as 
our baseline the current state of the market without a requirement for 
registrants to disclose pay ratio information. At present, registrants 
that are required to comply with Item 402(c) of Regulation S-K provide 
disclosure of their PEO's compensation as Section 953(b) requires. 
Other registrants, such as emerging growth companies, smaller reporting 
companies, foreign private issuers, and MJDS filers, are not required 
to comply with Item 402(c).\455\ We do not expect that many 
registrants, if any, currently maintain payroll and information systems 
that track total compensation as determined pursuant to Item 402 for 
all their employees, or make that information publicly available. Some 
registrants have reported ratios of CEO compensation to employee 
pay.\456\ We note, however, that the voluntarily reported information 
is different from the elements covered by Item 402(c)(2)(x), and 
therefore is not identical to what would be required under the final 
rule.
---------------------------------------------------------------------------

    \455\ These registrants are required to provide disclosure of 
executive compensation, but the disclosure requirements for these 
registrants do not fall under Item 402(c)(2)(x).
    \456\ For example, Noble Energy Inc. voluntarily disclosed that 
``our Chairman and CEO's total annual direct compensation was 
approximately 85 times that of the median annual total direct 
compensation of all of our other employees'' in their 2014 proxy 
filing and disclosed a pay ratio in their 2015 proxy filing. MBIA 
Inc. provided voluntary disclosure about average and median salary 
and bonus for all employees other than the NEOs, and compared them 
to the CEO's salary and bonus in its 2011 and 2012 proxy filings. 
NorthWestern Corp. disclosed in its 2011 proxy filing that its CEO 
compensation (salary, annual incentive and long-term incentive) was 
``18 times the median pay of all our employees'' and disclosed a pay 
ratio of 19-24 in its 2012-2015 proxy filings. Whole Foods Market 
Inc. disclosed in its filings a ``salary cap'' on executive cash 
compensation based on a multiple of the employee ``average annual 
wage''. Other examples of registrants that disclosed a pay ratio or 
median employee pay in its proxy filings include Advanced 
Environmental Recycling Technologies Inc., First Real Estate 
Investment Trust of New Jersey, Inter Parfums Inc., Itex 
Corporation, and Penn Virginia Corp. See also Simpson Thacher survey 
of pay ratio disclosures, available at http://www.compensationstandards.com/member/memos/firms/Simpson/03_15_ratio.pdf. We note that the pay ratio in these voluntary 
disclosures may differ from the pay ratio required to be disclosed 
in the final rule. In addition, registrants that currently disclose 
pay ratio are not necessarily the same registrants subject to the 
final rule.
---------------------------------------------------------------------------

    Currently, shareholders cannot calculate registrant-specific median 
employee compensation or the ratio of the PEO compensation to median 
employee compensation because there are no existing publicly available 
sources for this data. In the absence of such data, researchers have 
approximated the ratio using other available data, such as average 
employee pay.\457\ Statistics on the median earnings of U.S. workers in 
various ``industries'' are publicly available from the BLS,\458\ 
enabling shareholders to approximate the ratio using the industry 
median employee compensation and the information about PEO compensation 
for those registrants subject to Item 402(c).\459\ The distribution of 
the ratios of CEO to industry median employee compensation for a sample 
of large reporting companies is reported by NAICS industry sectors in 
the figure below for fiscal year 2013.\460\ Using this data, it is 
possible, for example, to determine that, for the median manufacturing 
firm with available data, CEO pay was approximately 105 times industry 
median employee pay. The 25th-75th percentile range for manufacturing 
firms was approximately 51-195.
---------------------------------------------------------------------------

    \457\ In a working paper entitled ``The CEO-Employee Pay 
Ratio,'' Dr. Steven Crawford finds that during the 1995-2012 period 
the ratio of CEO compensation to the average employee pay at U.S. 
commercial banks was on average 16.6 (with a median of 8.4). See 
``The CEO-Employee Pay Ratio'' available at http://ssrn.com/abstract=2529112.
    \458\ See BLS Occupational Employment Statistics available at 
http://www.bls.gov/oes/.
    \459\ See letter from Prof. Angel (``CEO compensation is already 
disclosed for public, but not private companies. The only new 
information is the pay of the median employee. However, there is 
already pretty good information about median compensation in various 
industries. For example, a few seconds of Googling leads to http://www.bls.gov/oes/current/oessrci.htm.'').
    \460\ The ratios in the figure are calculated for each 
registrant with executive compensation data from the Standard and 
Poor's Compustat Executive Compensation database which tracks 
compensation for the companies currently or previously in the S&P 
1500 index and industry median employee wage information at each 3-
digit NAICS level from the BLS as of May 2014. The data in the 
Compustat Executive Compensation database is for fiscal year 2013, 
which is the most recent fiscal year with complete coverage at the 
time of this analysis. The distribution of the registrant-level 
ratios within each NAICS industry sector (2-digit) is represented 
using horizontal box plots that show the minimum and maximum, and 
25th, 50th (median), and 75th percentiles.

---------------------------------------------------------------------------

[[Page 50151]]

[GRAPHIC] [TIFF OMITTED] TR18AU15.003

    We caution that any pay ratio estimate that can be made with 
currently available information would be different from the ratio 
required under the final rule. The above example uses the BLS median 
wage information of U.S. workers within the same 3-digit NAICS 
industries, while the final rule mandates registrants to use 
registrant-specific information about median employee compensation for 
``all employees,'' including employees in workplaces outside the U.S., 
subject to certain exemptions. Also, the example is based on only wages 
and does not consider other forms of compensation for employees other 
than PEOs because the BLS does not report those components for detailed 
industry definitions. In contrast, the final rule requires registrants 
to present the ratio using ``total compensation,'' which includes all 
forms of compensation in Item 402(c)(2)(x). Thus, while existing public 
data may permit shareholders to estimate median pay ratios across 
industry sectors, it does not allow for the particularized, registrant-
specific assessment that, in our view, Section 953(b) was intended to 
facilitate.
    To assess the economic effects of the final rule, we consider its 
impact on shareholders, registrants subject to the pay ratio 
disclosure, and all registrants' employees, including executive 
officers. We estimate that the final rule applies to approximately 
3,571 registrants.\461\
---------------------------------------------------------------------------

    \461\ Based on a review of EDGAR filings in calendar year 2014, 
we estimate that there were approximately 7,619 annual reports on 
Form 10-K filed in that year with available Xtensible Business 
Reporting Language (XBRL) data tags (available at http://www.sec.gov/dera/data/financial-statement-data-sets.html). From this 
number we subtracted the annual reports filed by approximately 678 
emerging growth companies (EGCs), 2,958 smaller reporting companies 
(SRCs), and 412 ABS issuers. These ABS issuers typically omit 
executive compensation disclosures in accordance with General 
Instruction J to Form 10-K. To the extent that the number of EGCs is 
growing each year, we might be underestimating the number of 
registered EGCs because we look only at registrants that file an 
annual report on Form 10-K. Registrants can fall into multiple 
categories among emerging growth companies, smaller reporting 
companies, and foreign private issuers. For example, 371 smaller 
reporting companies self-identified also as emerging growth 
companies. Therefore, we did not include these 371 registrants in 
the 2,958 smaller reporting companies that we subtracted from the 
7,619 registrants that file annual reports on Form 10-K because they 
were already included as emerging growth companies. Foreign private 
issuers and MJDS filers that file annual reports on Form 20-F and 
Form 40-F, respectively, are not required to provide Item 402 
information. They are therefore not included in the 3,571 affected 
registrants estimated above.
---------------------------------------------------------------------------

    Important potential determinants of the economic effects of the pay 
ratio disclosure requirements on the affected registrants are the 
differences in size, nature, and location of the workforce; complexity 
of the organization; and the degree of integration of payroll systems 
that are likely to exist among these registrants. In particular, the 
number of business and/or geographic segments within a particular 
registrant can significantly affect the compliance costs associated 
with the final rule. The registrants that operate in different 
geographic and business segments will likely have a less homogeneous 
workforce and are also less likely to maintain a single centralized 
payroll system.\462\ The average number of geographic and business 
segments and employees per each segment disclosed by some of the 
potentially affected

[[Page 50152]]

registrants in the calendar year 2014 are reported in the table 
below.\463\
---------------------------------------------------------------------------

    \462\ See letter from Chamber II (reporting that 39 companies 
that conduct operations in more than 50 countries with an average of 
90 different ``employee data systems'' worldwide would have average 
estimated labor costs of $311,800 for first year compliance. In 
contrast, 37 companies that operate in fewer than 10 countries with 
an average of 4.2 employee data systems would have average estimated 
labor costs of $67,200 for first year compliance (according to the 
25 firms that provided this data)).
    \463\ The corporate segments data used in the table come from 
the Standard and Poor's Compustat Segments database for companies 
with a business or geographic segment listed under ``segment type''. 
Segment information is self-reported by companies. As such, it is 
not based on standardized definitions of lines-of-business and 
geographic areas. The database provides some geographic segment 
information for approximately 63% of the potentially affected 
registrants and some business segment information for approximately 
68% of the potentially affected registrants.

                           Table 1--Characteristics of Registrants With Segments Data
----------------------------------------------------------------------------------------------------------------
                                                                                                     Number of
                                      Average           Min           Median            Max         registrants
----------------------------------------------------------------------------------------------------------------
Total Assets ($ millions).......          13,250               1           1,840       3,248,176           2,681
Number of Employees per                   15,453               0           2,978         537,000           1,575
 Registrant.....................
Number of Geographic Segments...            3.11               1               2              31           2,263
Geographic Segment Assets ($              11,892             3.1           1,512       3,248,176           1,037
 millions)......................
Number of Employees per                    8,742               0           1,522       1,100,000           1,210
 Geographic Segment.............
Number of Business Segments.....            2.45               1               1              11           2,444
Business Segment Assets ($                 4,614               1             811         812,044           2,171
 millions)......................
Number of Employees per Business           7,849               0           1,022         420,000           1,429
 Segment........................
----------------------------------------------------------------------------------------------------------------

    Table 1 shows that, in 2014, potentially affected registrants had 
an average of three geographic segments and two business segments. 
Also, the average number of employees was approximately 8,700 per 
geographic segment and 7,800 per business segment. We do not have 
complete information on how the registrants maintain their payroll 
systems across multiple geographic and business segments, but we 
believe that, because it is probable that registrants with multiple 
geographic and business segments will have multiple payroll systems and 
therefore lack easily accessible employee-level data on compensation, 
the number of such segments serves as an indication of the complexity 
and costs of trying to comply with the final rule (whether by sampling 
at each segment and aggregating the samples across the segments or by 
aggregating the payroll observations and sampling from the aggregated 
pool). The estimated costs associated with compliance for registrants 
with multiple geographic and business segments employing multiple 
payroll systems are discussed below.
    One commenter asserted that the pay ratio disclosure may affect PEO 
compensation.\464\ If the pay ratio disclosure were to significantly 
affect PEO compensation, the rule may have adverse effects on 
registrants' ability to attract and retain PEOs focused on such 
compensation. We note that there may be other factors affecting the 
ability of a registrant to attract and retain executive talent, such as 
the general structure and conditions of the labor market for 
executives. However, we do not have enough information to assess the 
effect of the new rule on PEO compensation or on the level of 
competition in the labor market for PEOs.
---------------------------------------------------------------------------

    \464\ See letter from Lou (``In spite of the fact that the pay 
ratio does not necessarily lead to CEO pay cuts, some companies may 
decrease executive pay if the number is too embarrassing. But this 
productivity-unrelated deduction can artificially depress the U.S. 
CEO market and IPOs.'').
---------------------------------------------------------------------------

    Relative to the baseline discussed above, the economic analysis 
that follows focuses initially on the likely economic effects--
including benefits and costs and impact on efficiency, competition, and 
capital formation--arising from the new mandatory disclosure 
requirement prescribed by the Dodd-Frank Act, and it then focuses on 
those that arise from the choices we have made in exercising our 
discretion.\465\
---------------------------------------------------------------------------

    \465\ Although we have divided our discussion of the economic 
effects of the rule between mandatory and discretionary features, we 
do not mean to imply that Congress unambiguously compelled us to 
adopt all of the items discussed under the mandatory requirements 
discussion. Specifically, we recognize that we retain exemptive 
authority and interpretive authority over many aspects of the rule, 
including many of those that we discuss within the mandatory 
requirements section. Generally speaking, we have chosen to classify 
items as mandatory because, in our view, these particular items 
appear to us to be consistent with the Congressional intent 
underlying Section 953(b).
---------------------------------------------------------------------------

C. Economic Effects From Mandated Disclosure Requirements

1. Benefits
    The following discussion is mainly intended to address benefits of 
the mandated disclosure to shareholders and shareholders of the 
registrants that are subject to the disclosure requirements mandated by 
Section 953(b).
    Although Congress neither expressly identified in Section 953(b) a 
specific market failure intended to be addressed by the new disclosure 
requirement nor expressly stated the specific objectives and intended 
benefits of Section 953(b), we nonetheless believe that the context in 
which the provision appears provides useful evidence of Congress' 
purpose in enacting the provision. As discussed above, we believe that 
Congress intended Section 953(b) to enhance the executive compensation 
information available to shareholders. Particularly, Section 953(b) 
provides new data points that shareholders may find relevant and useful 
when exercising their voting rights under Section 951. We believe, 
therefore, that Section 953(b) should be interpreted consonant with 
Subtitle E's general purpose of further facilitating shareholder 
engagement in executive compensation decisions. A significant 
consideration for us in fashioning a final rule implementing Section 
953(b), then, is the extent to which elements of the final rule further 
Congress' apparent goal of giving shareholders additional executive 
compensation information to use as part of the shareholder engagement 
envisioned by Section 951.
    Moreover, as discussed earlier, a number of commenters stated that 
they would find the pay ratio disclosure relevant when making voting 
decisions.\466\ We acknowledge the views of these commenters and regard 
the informational benefit of facilitating shareholder engagement in 
executive compensation decisions as potentially a significant new 
benefit to shareholders when they exercise their say-on-pay voting 
rights. We note that registrants have not historically been required to 
provide shareholders with access to information that would allow them 
to

[[Page 50153]]

assess the level of a PEO's compensation as it compares to employees at 
the same registrant and, as a result, shareholders generally have not 
been provided such information.
---------------------------------------------------------------------------

    \466\ See, e.g., letters from AFL-CIO I, AFSCME, Amalgamated, 
Bricklayers International, CalSTRS, Calvert, Chevy Chase Trust, 
CorpGov.net, Form Letter C, Form Letter D, Form Letter E, Form 
Letter F, LAPFF, LIUNA, NY State Comptroller, Pax World Funds, 
Public Citizen I, Rep. Ellison et al. I, Rep. Ellison et al. II, 
Trillium I, Trillium II, UAW Trust, and US SIF.
---------------------------------------------------------------------------

    While we believe that the pay ratio disclosure may provide an 
informational benefit to shareholders in their say-on-pay voting, we 
are unable to quantify this benefit. This is so for a number of 
reasons. First, the primary benefit that results from the pay ratio 
disclosure is not directly tied to an immediate economic transaction, 
such as the purchase or sale of a security, which makes it difficult 
for us to quantify in monetary terms the likely benefit to shareholders 
of this information. Second, the pay ratio disclosure is but one data 
point among many considerations that shareholders might find relevant 
when exercising their say-on-pay votes, which also makes it difficult 
for us to quantify the precise benefit that shareholders may 
experience. Third, even in situations where the pay ratio may be a 
significant or dispositive consideration for shareholders, because the 
say-on-pay vote is advisory and not binding, it is difficult for us to 
link the disclosure with certainty to a potential change in PEO 
compensation and even more speculative for us to link the disclosure to 
an economic outcome at a registrant. Further, we note that no commenter 
provided us with data that would allow us to quantify the potential 
benefits nor did any commenter suggest a source of data or a 
methodology that we could look to in quantifying the rule's potential 
benefits.
    We also think it is important to observe that, despite our 
inability to quantify the benefits, Congress has directed us to 
promulgate this disclosure rule. Thus, we believe it reasonable to rely 
on Congress's determination that the rule will produce benefits for 
shareholders and that its costs (which we discuss further below) are 
necessary and appropriate in furthering shareholders' ability to 
meaningfully exercise their say-on-pay voting rights. Because Congress 
expressly directed us to undertake this rulemaking, we do not believe 
it would be appropriate to second-guess its apparent conclusion that 
the benefits from this rule justify its adoption. In any event, as 
noted above, we concur with Congress's judgment that the pay ratio 
disclosure could be beneficial for shareholders.
    Commenters also suggested that pay ratio disclosure can be a 
valuable tool in evaluating PEO compensation practices in general.\467\ 
Among other uses of the pay ratio information, some commenters 
suggested that comparing the total compensation of the median employee 
and PEO would assist investors in their ability to evaluate the PEO's 
compensation in the context of the registrant's overall business,\468\ 
and could provide insight into the effectiveness of board oversight and 
sound board governance.\469\ Other commenters noted that they 
incorporate social and governance issues, like pay equity, as part of 
their investment decisions.\470\
---------------------------------------------------------------------------

    \467\ See, e.g., letters from Alexander, Bricklayers 
International, CalSTRS, Calvert, Chicago Teachers Fund, First 
Affirmative, Grossman, Grosvenor Capital, ICCR, IL Bricklayers and 
Craftworkers Union, Cummings Foundation, LaBruyere, Linton, Marco 
Consulting, NEI Investments (Dec. 2, 2013) (``NEI Investments I''), 
NEI Investments (Feb. 28, 2014) (``NEI Investments II''), Novara 
Tesija, NY Bricklayers and Craftworkers Union, NY State Comptroller, 
Oxfam, Pax World Funds, Public Citizen I, RPMI, Taylor, Tortora, US 
SIF, and Walden.
    \468\ See, e.g., letter from UAW Retiree Medical Benefits Trust 
(Apr. 15, 2011) (``UAW Trust pre-proposal'') (``Disclosure of 
internal pay equity, whether the ratio between median employee wages 
and those of the CEO or the ratio between compensation awarded to 
the CEO and to other top executives, will ultimately help investors 
evaluate executive pay practices by better contextualizing the 
information provided to the shareholders through the proxy statement 
and other corporate filings.'').
    \469\ See, e.g., letters from Allied Value, LLC (``Allied 
Value''); Kranen; Lynne L. Dallas, Professor of Law, University of 
San Diego (Dec. 1, 2013) (``Prof. Dallas''); UAW Trust pre-proposal 
(noting ``we view Section 953(b) as an essential tool that will 
increase corporate board accountability to investors''); and WA 
State Investment Board.
    \470\ See, e.g., letters from AFR, CT State Treasurer, Cummings 
Foundation, and Form Letter B (``I support Dodd-Frank rule 953(b), 
which strikes me as being all about the intersection of pay equity 
and investor value.'').
---------------------------------------------------------------------------

    As noted above, we recognize that there are significant limitations 
to using the pay ratio information for comparative purposes in light of 
the various factors that could affect employee compensation at a 
particular registrant and the flexibility we are providing. We believe 
that the informational benefit to shareholders from the final rule is 
in providing information about a particular registrant's executive 
compensation.
    In addition to its utility in assessing PEO compensation practices, 
commenters identified a number of ancillary benefits that may arise 
from the required pay ratio disclosure. Some commenters suggested that 
the new disclosure could offset an upward bias in executive 
compensation resulting from the practice of benchmarking executive pay 
solely against the compensation of other executives to the extent that 
current benchmarking practices are inefficient.\471\ Other commenters 
suggested that a comparison of PEO compensation to employee 
compensation could be used by shareholders to approximate employee 
morale and/or productivity \472\ or analyzed as a measure of a 
particular registrant's approach to managing human capital.\473\ One 
commenter cited his own research showing that large pay disparities 
within a corporation contribute to an unethical culture within the 
corporation.\474\ Finally, some commenters asserted that the 
registrant-specific information about the median employee compensation 
may be used to address a broader public policy concern relating to 
income inequality and income mobility, which they suggest is 
exacerbated by increasingly high levels of PEO compensation relative to 
other workers.\475\
---------------------------------------------------------------------------

    \471\ See, e.g., letters from AFR, Cummings Foundation, LIUNA, 
PGGM, and RPMI.
    \472\ See, e.g., letters from AFL-CIO I, AFR, AFSCME, Alexander, 
Allied Value, B[acirc]tirente et al., CalSTRS, Calvert, Cummings 
Foundation, CUPE, Demos I, Domini, Professors Charles M. Elson and 
Craig K. Ferrere (Dec. 2, 2013) (``Profs. Elson and Ferrere''), 
Estep, Form Letter A, FS FTQ, Daniel Greenwood (Oct. 26, 2013) 
(``Greenwood''), Grossman, IPS, LaBruyere, Linton, McMorgan Co., 
Mudd, NY State Comptroller, Overcott, Pax World Funds, Public 
Citizen I, Rosati, SEIU, Socially Responsive Financial Advisors, 
Somers, Taylor, Teamsters, Tortora, Trillium I, Trillium II, Trustee 
Campbell, US SIF, and Walden.
    \473\ See, e.g., letters from Bricklayers International, Demos 
(May 30, 2014) (``Demos II''), FS FTQ, MVC Associates, Quintave, 
OCP, Rep. Ellison et al. I, and UAW Trust.
    \474\ See letter from Prof. Dallas.
    \475\ See letters from Rep. Ellison et al. I; Rep. Ellison et 
al. II; and Dr. Sue Ravenscroft, Professor of Accounting, Iowa State 
University (Jun. 18, 2014) (``Prof. Ravenscroft II'').
---------------------------------------------------------------------------

    With respect to employee morale and productivity, commenters did 
not specify what effect a pay ratio disclosure would have on these 
conditions relative to other environment-specific and registrant-
specific factors. In particular, the pay ratio disclosure may be 
significantly dependent on how a registrant structures its business. 
For example, one registrant might outsource the labor-related 
(manufacturing) aspects of its business to a third party to focus on 
product innovation, while another registrant competing in the same 
industry might choose to retain the labor aspect of its business. To 
the extent that product innovation requires higher pay than 
manufacturing, the outsourcing company will have a lower pay ratio for 
the same PEO pay. Therefore, the potential value of this disclosure for 
assessing issues related to employee morale, productivity, and 
investment in human capital may be

[[Page 50154]]

diminished by the variation in business structures.
    Some commenters suggested that the pay ratio disclosure would 
promote capital formation.\476\ The main rationale given for this 
effect is that the new disclosure would help shareholders understand 
the assets of the firms they invest in or that it will let shareholders 
choose registrants that invest in their workforce.\477\ On the other 
hand, some commenters asserted that the rule would discourage capital 
formation because it would discourage firms from accessing the U.S. 
capital markets.\478\ We note that the final rule does not apply to 
emerging growth companies, which conduct the bulk of initial public 
offerings.\479\ While the pay ratio disclosure could be costly, it is 
not clear whether it would significantly affect a registrant's ongoing 
capital raising activity.
---------------------------------------------------------------------------

    \476\ See, e.g., letters from Alexander, Capital Strategies, 
Change to Win (Nov. 25, 2013) (``Change to Win''), CUPE, Greenwood, 
Grossman, LaBruyere, Mudd, Overcott, Prof. Ravenscroft I, Taylor, 
and Tortora.
    \477\ See, e.g., letters from Alexander, Greenwood, Grossman, 
LaBruyere, Mudd, Overcott, Taylor, and Tortora.
    \478\ See, e.g., letters from Prof. Angel, COEC I, Lou, and 
NIRI.
    \479\ M. Dambra, L. Field and M. Gustafson, The JOBS Act and IPO 
Volume: Evidence that Disclosure Costs Affect the IPO Decision, J. 
of Financial Economics. (2014) (documenting that 88% of the U.S. 
IPOs filed in the 4/1/12-3/31/14 period are eligible for the EGC 
revenue threshold).
---------------------------------------------------------------------------

    Overall, while certain shareholders may use the pay ratio for their 
investment decisions, it is unclear whether the final rule would impact 
the capital formation of U.S. capital markets in a significant way. As 
discussed above, shareholders may be able to approximate the industry 
level pay ratio using industry employee compensation data from BLS and 
the information about PEO compensation for registrants subject to Item 
402(c). In this regard, adding the pay ratio statistic to the mix of 
reported financial and operational data may not change the investment 
decision of investors who access this data. On the other hand, the pay 
ratio disclosure is company-specific, which adds information not 
otherwise available to investors.
    In contrast to commenters supporting the required disclosure, some 
commenters stated that the disclosure mandated by Section 953(b) would 
not have any benefit, or would not have benefits sufficient to justify 
the compliance costs, which many of those commenters anticipate would 
be substantial.\480\ Some of the commenters questioned the materiality 
of pay ratio information to an investment decision \481\ This view was 
also asserted by the minority in the Senate report accompanying the 
legislation.\482\
---------------------------------------------------------------------------

    \480\ See, e.g., letters from AAFA I, American Benefits Council, 
ASA, Brian Foley & Co. Chamber I, Chamber II, NIRI, Tesoro Corp., 
WorldatWork I, and WorldatWork II.
    \481\ See, e.g., American Benefits Council, ExxonMobil, Prof. 
Muth, and RILA. Another commenter noted that pay ratio disclosure 
that includes all employees and is based on a large, global, full- 
and part-time pool of employees will not be meaningful without 
substantial explanation. See letter from WorldatWork II.
    \482\ See S. Rep. No. 111-176 (2010) (``Although provisions like 
this appeal to popular notions that chief executive officer salaries 
are too high, they do not provide material information to investors 
who are trying to make a reasoned assessment of how executive 
compensation levels are set. Existing SEC disclosures already do 
this.'').
---------------------------------------------------------------------------

    While we acknowledge these concerns about the usefulness or 
materiality of the mandated disclosure, we note that other commenters 
asserted that certain shareholders incorporate social and governance 
issues, like pay equity, as part of their decision making.\483\ These 
shareholders may realize non-economic benefits associated with their 
decision making based on this type of information. These commenters, 
however, did not quantify the extent to which shareholders would value 
pay ratio information or would incorporate the disclosure required by 
Section 953(b) into their investment or voting decision, if at all. 
Academic research suggests that the stock market does not fully 
incorporate employee satisfaction into stock prices.\484\ As mentioned 
above, because company-specific pay ratio information is not currently 
reported, it is not possible to assess the usefulness to shareholders 
of this information as required by Section 953(b) relative to the 
usefulness of publicly available statistics of median compensation, or 
the usefulness of any other company-specific metric of employee 
compensation or satisfaction.
---------------------------------------------------------------------------

    \483\ See, e.g., letters from AFR and Form Letter B.
    \484\ Existing research has studied whether there is a 
correlation between information about employee satisfaction and 
long-term equity returns in an effort to understand how the market 
values a public company's intangible assets. This research uses 
different information than what is provided in the pay ratio 
disclosure. This research was based on the equity prices of 
companies that were identified on Fortune Magazine's list of the 
``100 Best Companies to Work For in America.'' See A. Edmans, Does 
the stock market fully value intangibles? Employee satisfaction and 
equity prices, J. of Financial Economics 101, 621-640 (2011) 
(finding evidence implying that the market fails to incorporate 
intangible assets, like employee satisfaction, fully into stock 
valuations until the intangible subsequently manifests in tangibles, 
such as earnings, that are valued by the market, and finding 
evidence suggesting that the non-incorporation of intangibles into 
stock prices is not simply due to the lack of salient information 
about them).
---------------------------------------------------------------------------

    Some commenters were particularly concerned that the comparisons of 
pay ratios across registrants may be inappropriate to the extent that 
registrants employ workers in different countries that have unique 
compensation practices,\485\ use different methodologies to calculate 
the median employee,\486\ employ workers with different skill 
levels,\487\ and have different corporate structures.\488\ As noted 
above, we believe that the purpose of the pay ratio disclosure is to 
provide shareholders of a registrant with new data points that they may 
find relevant and useful when exercising their voting rights under 
Section 951. As we noted in the Proposing Release, we believe that a 
variety of factors can potentially limit the comparability of the pay 
ratio across registrants. We also acknowledge that the final rule we 
are adopting allows for significant flexibility in determining the pay 
ratio to address concerns raised by a number of commenters about the 
potential costs of the pay ratio disclosure. One result of allowing for 
this flexibility, however, is that the comparability of the pay ratio 
from registrant to registrant may be further diminished. We recognize 
this consequence but believe it is justified in light of the cost 
savings that such flexibility will provide and because we do not regard 
precise comparability as the primary objective of the final rule.
---------------------------------------------------------------------------

    \485\ See, e.g., letters from Avery Dennison, BCIMC, and COEC I.
    \486\ See letter from BCIMC.
    \487\ See letter from COEC I.
    \488\ See letters from Prof. Angel, COEC I, and Tesoro Corp.
---------------------------------------------------------------------------

2. Costs
a. General
    The following discussion is mainly intended to address costs to 
registrants that are subject to the pay ratio disclosure. The analysis 
of costs focuses on direct compliance costs on registrants.
    As discussed above, the final rule permits registrants to choose 
from several options to identify the median employee. First, 
registrants can choose to use Item 402(c)(2)(x) to calculate the annual 
total compensation for each employee and then identify the median 
employee. Second, registrants can choose a statistical method that is 
appropriate to the size and structure of their own businesses and the 
way in which they compensate employees to identify the median employee, 
and then use Item 402(c)(2)(x) to calculate the median employee's 
compensation. Third, registrants can use a consistently applied 
compensation measure, whether with respect to the entire employee 
population or in conjunction with

[[Page 50155]]

statistical sampling, to identify the median employee, and then 
calculate and disclose that median employee's total compensation in 
accordance with Item 402(c)(2)(x).
    In addition to providing flexibility in identifying the median 
employee's compensation, the final rule allows flexibility in several 
other respects. Registrants may:
     Use reasonable estimates when applying Item 402(c)(2)(x) 
to calculate the annual total compensation for employees other than the 
PEO, including when disclosing the annual total compensation of the 
median employee identified using a consistently applied compensation 
measure;
     identify the median employee every three years to the 
extent that there is no significant change in the registrant's employee 
population or employee compensation arrangements;
     consistently choose any date within the last three months 
of a registrant's fiscal year to identify the median employee.
    Moreover, the final rule allows flexibility for registrants with 
non-U.S. employees by providing (1) a foreign data privacy law 
exemption reducing the burden on registrants that operate in certain 
foreign jurisdictions, and (2) a de minimis exemption that may reduce 
the number of payroll systems that need to be used to identify the 
median employee and will allow registrants some flexibility in 
addressing payroll matters that may result from having employees in 
multiple jurisdictions. Finally, the final rule also provides that 
registrants, when determining the compensation of the median employee 
for purposes of identifying the median employee and making the pay 
ratio disclosure, may elect to adjust the compensation of their 
employees in jurisdictions other than the jurisdiction in which the PEO 
resides to reflect the cost of living in the PEO's country of 
residence, but they must also provide disclosure of the registrant's 
pay ratio calculated without the cost-of-living adjustment. Overall, we 
believe that the flexible approach allowed by the final rule is 
consistent with Section 953(b) and, in certain circumstances discussed 
below, may reduce costs compared to other methods of implementing the 
statute.
b. Compliance Cost Estimates in Comment Letters
    In the pre-proposing period, we received estimates of the costs of 
compliance for certain registrants from some commenters.\489\ These 
estimates varied significantly and were based on the commenters' 
initial reading and interpretation of the statute and not on the 
proposed rule, which would allow for flexibility not accounted for in 
the pre-proposal letters. For example, prior to the proposal, one 
commenter estimated the cost of compliance with Section 953(b) would be 
$250,000 to $500,000 annually, and revised its cost estimate downward 
to $15,000 annually after the proposed rule was released ``primarily 
due to the ability afforded by the proposed rule for registrants to use 
a consistently applied compensation measure, such as payroll records or 
W-2 reportable wages and the equivalents for non-U.S. employees, to 
identify the median employee.'' \490\
---------------------------------------------------------------------------

    \489\ For example, one pre-proposal comment letter from an 
industry group reported that a member company estimated that it 
would require approximately $7.6 million and 26 weeks to prepare the 
pay ratio disclosure and that a separate member company estimated 
that it would cost approximately $2 million annually to determine 
the actuarial value of employee pension benefits. See American 
Benefits Council et al., (Jan. 19, 2012) (``American Benefits 
Council et al. pre-proposal letter'').
    \490\ See letter from Intel.
---------------------------------------------------------------------------

    In the Proposing Release, we did not estimate the costs of the 
calculation and disclosure of a registrant's pay ratio because we did 
not have enough data for such estimation. In response to the Proposing 
Release, a number of commenters evaluated our estimates of the 
compliance costs represented by the estimated Paperwork Reduction Act 
(``PRA'') burdens imposed by the proposed rule. Most commenters 
generally indicated that those PRA burdens underestimated the 
compliance costs associated with the disclosure requirement, and some 
provided more specific cost estimates. For example, one commenter noted 
that our PRA estimate of an average of 340 hours of internal company 
time in year one to comply with the proposed rule significantly 
understates the time that many companies would need to comply 
(especially those with non-U.S. employees).\491\ Below, we discuss the 
specific comments that we consider to be the most useful to estimate 
the compliance costs of the pay ratio disclosure. We note that, in 
providing specific comments, commenters did not typically distinguish 
between costs derived from the statutory mandate and costs derived from 
the exercise of our discretion. Furthermore, they typically did not 
distinguish between internal costs in burden hours versus external 
professional costs in dollar amounts for PRA purposes.
---------------------------------------------------------------------------

    \491\ See letter from NIRI.
---------------------------------------------------------------------------

    Two commenters provided survey studies with several relevant 
estimates of the compliance costs associated with the proposed rule, as 
well as characteristics of the types of registrants that would be 
affected. As discussed in greater detail below, the aggregate initial 
external compliance cost estimates provided by these commenters range 
between $187 million \492\ and approximately $711 million.\493\ These 
estimates are based on responses to the surveys discussed below and may 
not be representative of all registrants affected by the final rule.
---------------------------------------------------------------------------

    \492\ See letter from COEC I. We note that this estimate 
represents only the cost of outside professionals and does not take 
into account internal company hours as an additional cost of 
compliance with the rule. This commenter estimated that compliance 
would require at least 801,000 hours of in-house personnel time (at 
least 255,000 additional company hours above the 546,000 company 
hours we estimated) in addition to a cost of $187 million.
    \493\ See letter from Chamber II. This commenter estimated an 
annual internal compliance burden of 3.6 million hours in addition 
to an annual cost of $710.9 million.
---------------------------------------------------------------------------

i. Center on Executive Compensation Survey
    One commenter provided the results of a joint survey it conducted 
among its members.\494\ The results are based on the responses from 128 
public companies out of 1,270 surveyed.\495\ Most of the respondents 
are large registrants, with average revenue of approximately $28 
billion; 59% of the registrants had revenues greater than $10 billion. 
Nearly 80% of respondents had 10,000 or more employees, most of them 
employed full-time. In addition, nine out of ten respondents had 
foreign operations with employees located outside the United States. On 
average, respondents operated in 34 countries, and about two-fifths of 
their employees worked in foreign countries. The average number of 
separate employee data systems that respondents had worldwide was 46.
---------------------------------------------------------------------------

    \494\ This survey was jointly conducted by the COEC, Human 
Resource Policy Association, and Corporate Secretaries. See letters 
from COEC I and Corporate Secretaries. See also letter from Business 
Roundtable I (referencing the results from the COEC I survey).
    \495\ See letter from COEC I. We note that the letter from 
Corporate Secretaries refers to results from the same survey, but 
for the 127 respondents who are also members of Corporate 
Secretaries.
---------------------------------------------------------------------------

    In its letter, the commenter questioned our estimate, for PRA 
purposes, of $400 per hour for outside professional costs and the 
estimated PRA hour burden. More than half of the survey respondents 
indicated that the average hourly fee for their company's

[[Page 50156]]

external securities compliance counsel is above $700. The respondents 
also indicated that, on average, 72% of the estimated initial 
compliance costs are expected to be incurred in subsequent years. Based 
on these survey results, the commenter asserted that compliance with 
the rule would require at least 255,000 additional company hours and an 
additional $114.1 million in costs across all affected registrants for 
outside professional services above the our PRA burden estimates in the 
Proposing Release ($72.77 million). Using these updated estimates, the 
commenter arrived at a total initial compliance cost estimate of at 
least $186.9 million.\496\ We note that, although labeled ``total 
compliance costs'' by the commenter, that estimate of compliance costs 
includes only the cost of outside professionals, and thus is only part 
of the expected total compliance costs. The estimate does not take into 
account internal company hours as an additional cost of compliance with 
the rule. Additionally, the commenter assumed that all affected 
registrants will bear the same compliance costs, which may bias its 
total cost estimate because compliance costs are likely to vary between 
registrants with and without foreign operations, or between small and 
large registrants.
---------------------------------------------------------------------------

    \496\ The commenter reports the additional cost of outside 
professional services of $114 million in excess of the cost of 
outside professional services estimated by us in the Proposing 
Release. The total cost estimate reported as $186.9 million can be 
obtained by adding $114 million and $72.77 million, with the 
difference likely due to rounding. See letter from COEC I.
---------------------------------------------------------------------------

    The survey provided several estimates of how compliance costs might 
change if there were certain changes in the rule. For instance, the 
commenter's letter argued that the final rule should apply only to a 
registrant's consolidated subsidiaries, noting that its survey 
indicated that, if the final rule were to include employees of all 
minority-owned subsidiaries and joint ventures, a registrant's 
compliance costs would increase by an average of 91%, with a mid-range 
of 20%. The letter from the other commenter that jointly conducted the 
survey also presented information about the inclusion of all minority-
owned subsidiaries and joint ventures, but its letter presented the 
survey data in a different format. It presented the average and median 
anticipated increases categorized based on the company's annual 
revenue.\497\ According to this comment letter, for registrants with 
annual revenue of over $30 billion, the median increase in cost would 
be approximately 35% if employees in minority-owned subsidiaries and 
joint ventures were included. For registrants with annual revenue 
between $5 billion and $30 billion, that median increase would be 
approximately 20%, while for registrants with annual revenue below $5 
billion, the median increase in compliance cost would be 10%. These 
numbers, however, appear to reflect an increase in the compliance cost 
if the coverage of subsidiaries and joint ventures were to be increased 
from the suggested coverage under the Proposing Release to complete 
(100%) coverage of subsidiaries and joint ventures. Thus, they do not 
directly correspond to the changes we made in this release, including 
the change that we made to include only employees of a registrant's 
consolidated subsidiaries, as suggested by the commenter.
---------------------------------------------------------------------------

    \497\ See letter from Corporate Secretaries.
---------------------------------------------------------------------------

    If a registrant were permitted to calculate the pay ratio based on 
full-time, permanent employees only, then according to the survey 
responses, the compliance cost would decrease by a mid-range of 10% or 
an average of approximately 11%.\498\ Another commenter suggested that 
the median decrease in compliance costs would be approximately 10% or 
an average of 12% if a registrant were permitted to calculate the pay 
ratio based on full-time, permanent employees only.\499\ Requiring only 
U.S. employees to be used when estimating the pay ratio would decrease 
costs on average by 40%, while the mid-range decrease would be 
approximately 50%.\500\ In contrast, if the rule did not contain the 
flexibility allowed under the proposal and instead total compensation 
as calculated in the Summary Compensation Table was required to be used 
to identify the median employee, 99% of the respondents said that their 
cost would increase and 49.1% said that the cost would increase by over 
100%. Another commenter, relying on information from the same survey, 
suggested that the average cost increase would be 4,689% and the median 
cost increase would be approximately 175% if total compensation as 
calculated in the Summary Compensation Table was required to be used to 
identify median employee pay.\501\
---------------------------------------------------------------------------

    \498\ See letter from COEC I. While the survey data attached to 
the comment letter showed that the average anticipated cost 
reduction would be 11%, the text of the comment letter discussing 
the survey also stated that for the subset of respondents that 
anticipated that limiting the rule to full-time employees would 
lower costs, ``the average savings would be approximately 20 
percent.''
    \499\ See letter from Corporate Secretaries.
    \500\ See letter from COEC I (mentioning in the text of the 
comment letter that the expected decrease is ``47% on average for 
firms with non-U.S. employees''). See also letter from Corporate 
Secretaries.
    \501\ See letter from Corporate Secretaries.
---------------------------------------------------------------------------

ii. Chamber of Commerce Survey
    A different commenter also provided estimates of compliance costs 
of the proposed rule based on survey results.\502\ This commenter's 
survey is a version of the COEC survey that included only 118 
respondents, approximately ``3.1% of all covered businesses.'' The 
commenter did not elaborate on how its version of the survey is 
different from the COEC survey, other than including fewer 
respondents.\503\ The commenter's letter provides no information on the 
survey's respondent size characteristics to provide context with 
respect to the respondents' potential organizational complexity and 
associated challenges in complying with the proposed rule. Based on the 
survey, the commenter concludes that the average labor cost per company 
of complying with the proposed rule would be approximately $185,600 for 
the initial year. That commenter also estimated, but did not monetize, 
an annual compliance time of 3.6 million hours.\504\ The survey results 
also show a wide divergence in cost estimates across survey 
respondents, with 42 respondents estimating the value of the time 
necessary to comply with the proposal to be at least $100,000, while 13 
respondents estimated this value to be less than $10,000. On average, 
respondents estimated 952 hours needed to comply with the proposed 
rule. Respondents that conduct operations in foreign countries will 
have higher compliance costs according to the survey results. Thirty-
nine respondents that conduct operations in more than 50 countries 
indicated an average labor cost of $311,800 to comply with the proposed 
rule. These respondents also reported an average of 90 different 
employee data systems worldwide. On the other hand, for 37 respondents 
that operate in fewer than 10 countries, the average compliance cost 
was estimated to be $67,200. Based on the survey results, the commenter 
asserted that the total external compliance costs for the private 
sector could be approximately $711 million and that total cost could 
increase to $1.1 billion (in addition to the internal compliance time) 
if every

[[Page 50157]]

affected registrant has an average cost of $311,800.
---------------------------------------------------------------------------

    \502\ See letter from Chamber II.
    \503\ The letter from Chamber II does not specify how many 
companies were surveyed but the letter indicates that the Chamber 
represents over ``3 million businesses and organizations of every 
size, sector and region.''
    \504\ See letter from Chamber II.
---------------------------------------------------------------------------

iii. Other Specific Comments
    In addition to the two surveys, several other commenters provided 
the following cost estimates based on the proposed rule. In these 
estimates, the commenters did not distinguish between the costs arising 
from the mandated disclosure and the costs arising from the exercise of 
our discretion. The estimates for the proposal were as follows:
     $500,000 to $1 million to automate a large global 
registrant's processes; \505\
---------------------------------------------------------------------------

    \505\ See letter from KBR.
---------------------------------------------------------------------------

     between $1 million and $1.5 million for at least 10 to 15 
internal staff members and two or three external advisors, with 100 to 
150 hours of internal work and 20 to 40 hours in external consulting 
time; \506\
---------------------------------------------------------------------------

    \506\ See letter from Avery Dennison.
---------------------------------------------------------------------------

     annual compliance cost of $15,000 for an issuer with 
global operations; \507\
---------------------------------------------------------------------------

    \507\ See letter from Intel.
---------------------------------------------------------------------------

     a ``likely to be conservative'' estimate of $100,000 per 
company, based on what mid-sized retail corporations informed the 
commenter.\508\
---------------------------------------------------------------------------

    \508\ See letter from NRF. We note that this commenter did not 
indicate the number of employees of such mid-sized retail 
corporations.
---------------------------------------------------------------------------

     approximately $250,000 for 1,000 internal hours initially 
and $100,000 per year for 500 hours annually thereafter (but, according 
to the commenter, the workload would perhaps drop by 90% if the final 
rule includes only employees employed by the U.S. parent organization 
and all U.S.-based subsidiaries in addition to other changes 
recommended by the commenter); \509\
---------------------------------------------------------------------------

    \509\ See letter from FEI.
---------------------------------------------------------------------------

     ``thousands of dollars to hire a dedicated resource and 
overhaul our payroll and human resource information system in order to 
prepare our first pay ratio disclosures under this rule;'' \510\
---------------------------------------------------------------------------

    \510\ See letter from FuelCell Energy.
---------------------------------------------------------------------------

     between $50,000 and $100,000 to test the commenter's 
current payroll system for a quote on identifying the median employee, 
with an ``actual cost'' that ``is indeterminable'' but that the 
commenter believes could cost over $500,000; \511\
---------------------------------------------------------------------------

    \511\ See letter from NACCO.
---------------------------------------------------------------------------

     $500,000 to $1 million for 50 internal employees and 
outside advisors; \512\
---------------------------------------------------------------------------

    \512\ See letter from General Mills.
---------------------------------------------------------------------------

     3,000 work hours in the initial year and 850 work hours 
annually thereafter (but, according to the commenter, these costs would 
be reduced by 90% if the final rule excluded non-U.S. employees); \513\
---------------------------------------------------------------------------

    \513\ See letter from ExxonMobil.
---------------------------------------------------------------------------

     over $1.6 million not including modifications of payroll 
or accounting systems; \514\
---------------------------------------------------------------------------

    \514\ See letter from Eaton.
---------------------------------------------------------------------------

     actual cost is indeterminable, but believed to exceed 
$500,000 due to substantial non-U.S. employee base; \515\
---------------------------------------------------------------------------

    \515\ See letter from Hyster-Yale.
---------------------------------------------------------------------------

     cost for many registrants would likely to be in the 
millions of dollars; \516\ and
---------------------------------------------------------------------------

    \516\ See letter from ASA.
---------------------------------------------------------------------------

     annual cost to collect required data would exceed $2 
million.\517\
---------------------------------------------------------------------------

    \517\ See letter from Dover Corp.
---------------------------------------------------------------------------

    The overall cost range provided by individual commenters for 
initial compliance by a large registrant was between $15,000 per year 
to $2 million.\518\ We note that all of these comments concerned the 
proposed rule, rather than the final rule. As discussed below, the 
final rule allows for further flexibility, which we believe will reduce 
the cost of compliance.
---------------------------------------------------------------------------

    \518\ See letters from Dover Corp. and Intel.
---------------------------------------------------------------------------

    These estimates provide a significant number of data points on the 
anticipated compliance costs that we use in our quantification of the 
estimated compliance costs of the final rule below. However, we caution 
that these estimates do not necessarily represent an accurate 
indication of the expected costs because they use different 
methodologies and assumptions in arriving at these numbers, some of 
which might change with the different requirements under the final 
rule. Moreover, only a few commenters discussed the complexity of their 
payroll systems; \519\ the degree to which the estimated costs reflect 
internal personnel costs \520\ or the costs of outside service 
providers \521\ and outside professionals; \522\ and the precise 
assumptions used in deriving the estimates, all of which may be 
relevant for assessing the estimates provided.\523\ Also, although most 
of these estimates do not precisely distinguish between initial and 
ongoing costs, we expect that, for many registrants, the overall 
compliance burden will diminish after systems are in place to gather 
and verify the underlying data.\524\ Some commenters noted that the 
costs are impossible to determine before they start the process.\525\ 
The provided cost estimates were also given prior to additional cost-
reducing measures adopted in the final rule in response to comments, 
including comments about costs. These measures include: The ability to 
calculate the median employee once every three years, the exemption 
from the definition of ``employee'' of a de minimis percentage of non-
U.S. employees, the requirement to consider only registrant's own 
employees and those of their consolidated subsidiaries when identifying 
the median employee, and the exemption for employees in foreign 
jurisdictions in which it is not possible for a registrant to obtain or 
process information necessary to comply with the rule without violating 
the data privacy laws or regulations of that jurisdiction. We expect 
that these changes will further reduce the costs of compliance.
---------------------------------------------------------------------------

    \519\ See letter from Hyster-Yale (commenting that ``of the 
4,967 people who received Form W-2s (or similar documents) for 2012, 
2,375 (48%) resided outside the U.S. and were paid on twenty 
different payroll systems (none of which are integrated with our 
U.S. system or any other system)''). This commenter, however, did 
not comment on the expected cost of integrating these payroll 
systems.
    \520\ See letter from Avery Dennison (estimating that they will 
use 10-15 internal staff members with 100-150 hours of internal 
work), FEI (estimating 1,000 hours of internal time to develop the 
database and methodology to derive the data and 500 hours to 
maintain this database in the following years), and ExxonMobil 
(estimating 3,000 work hours by internal personnel in the first year 
and 850 work hours per year thereafter).
    \521\ See letter from ExxonMobil (stating ``are not able to 
provide a specific estimate, but expect that a significant work 
effort would also be required on the part of each of our 60 third-
party vendors'').
    \522\ See letters from Avery Dennison (estimating that they will 
use two or three external advisors (e.g., legal counsel, human 
resources consultants) and 20-40 hours of external consulting time,) 
and Hyster-Yale (stating that it will need to hire an outside 
consulting firm to assist with the process).
    \523\ See letter from General Mills (providing a list of steps 
it will take). However, the commenter did not provide detailed cost 
estimates for these steps.
    \524\ A few commenters addressed this point, and the estimates 
they provided were very different. See, e.g., letters from 
ExxonMobil (expecting a 72% drop), FEI (estimating a 60% drop in 
compliance costs from the first year to the next year), and General 
Mills (indicating that most of the costs for the first year would 
also apply to subsequent years of compliance). But see letter from 
Business Roundtable I (asserting that 42% of respondents to a survey 
it conducted indicated that they would have to update their 
methodology every year).
    \525\ See, e.g., letter from NACCO (stating that consulting 
companies were unable to provide quotes about the cost to assist 
with statistical sampling until they are able to test their various 
payroll systems). This commenter also stated that the actual cost is 
indeterminable but could exceed $500,000, as it will depend on (i) 
the availability and accuracy of employee data, (ii) the scope of 
the final rule and (iii) whether registrants choose to disclose the 
minimum required disclosure or if they decide to provide various 
alternative disclosures that would provide shareholders with more 
context.
---------------------------------------------------------------------------

    In contrast to these estimates, a significant number of the 
commenters, generally the same commenters that perceived the benefits 
of the rule, asserted that the rule would not impose high costs and 
burdens. The majority of these commenters indicated that the

[[Page 50158]]

permitted flexibility in complying with the proposed rule would reduce 
costs,\526\ and that the registrants already have the data necessary to 
make their pay ratio calculations.\527\ While we agree that the 
permitted flexibility should lower costs for many registrants, we 
recognize that registrants who operate in various geographic and 
business segments may need to reconcile their systems to compile and 
provide the required information at a potentially significant 
cost.\528\
---------------------------------------------------------------------------

    \526\ See e.g., letters from AFR, Bricklayers International, 
CalPERS, Calvert, Capital Strategies, Chicago Teachers Fund, 
Cummings Foundation, First Affirmative, ICCR, IL Bricklayers and 
Craftworkers Union, LIUNA, Marco Consulting, Novara Tesija, NY 
Bricklayers and Craftworkers Union, Oxfam, Pax World Funds, Prof. 
Ravenscroft I, Public Citizen I, Rep. Ellison et al. I (referring 
specifically to sampling), Sen. Menendez et al. I, Sen. Menendez et 
al. II, and US SIF.
    \527\ See, e.g., letters from AFR, Allied Value, CalPERS, LAPFF, 
LIUNA, Pax World Funds, and Theodore.
    \528\ See, e.g., letters from American Benefits Council, Aon 
Hewitt, Avery Dennison, BCIMC, Business Roundtable I, Business 
Roundtable (Jul. 21, 2015) (``Business Roundtable II''), Chamber I, 
COEC I, Corporate Secretaries, Cummins Inc., Eaton, ExxonMobil, FEI, 
Freeport-McMoRan, FuelCell Energy, IBC, KBR, MVC Associates, NACCO, 
NAM I, NAM II, NIRI, Semtech, SHRM, and Tesoro.
---------------------------------------------------------------------------

c. Quantification of Compliance Costs
    While our overarching consideration of the costs of the rule takes 
into account the information provided by a broad range of commenters, 
the most useful frameworks for considering costs were provided by 
commenters that provided data on company-wide potential costs. Other 
commenters provided certain valuable insights into how our rule would 
be implemented, but were either not as transparent in their analytical 
frameworks or not easily generalizable in terms of aggregating the 
costs across multiple registrants.
    Some commenters indicated that the most important driver for the 
compliance costs of the required disclosure is the presence of foreign 
operations and the complexity of dealing with, or the need to create 
compatible employee data and payroll systems that track the 
compensation of employees of such foreign operations.\529\ Underscoring 
the importance of foreign operations for the costs of compliance with 
the rule, one commenter's survey results indicate that the compliance 
costs for registrants would decrease by a median of 50% or on average 
by 40% if only U.S. employees were included in the calculation of the 
median compensation.\530\ In addition, most of the survey participants 
appear to have had at least some international operations. The 
participants in the commenter's survey on average operate in 34 
countries (including the United States) \531\ and have about 38% of all 
employees and 44% of their full-time employees outside of the United 
States. The survey reflected predominantly larger registrants and 
therefore may not reflect the characteristics of a large number of 
registrants subject to Section 953(b) requirements. Another commenter 
suggested that the cost of implementing the final rule would be 20-30 
times higher if foreign employees are included in the calculation of 
the median compensation.\532\ Based on these comments and survey 
results, we acknowledge that there may be significant differences in 
the potential costs of the rule between registrants with foreign 
operations and registrants without foreign operations.
---------------------------------------------------------------------------

    \529\ See id.
    \530\ See letter from COEC I. The survey was conducted jointly 
by COEC, Human Resource Policy Association, and Corporate 
Secretaries. See supra note 494.
    \531\ See id. The Chamber II survey participants were similarly 
international in operation. Over 68% of the participants in the 
Chamber II survey operate in more than 10 countries.
    \532\ See letter from American Benefits Council.
---------------------------------------------------------------------------

    Commenters have pointed out other potential cost drivers, such as 
including part-time, seasonal, and temporary employees in the 
calculation of the median. However, the effect of these other factors 
seems to be less significant. For example, one commenter's survey 
results suggest that the compliance costs for registrants would 
decrease by a median of 10% and on average by 11% if only full-time 
permanent employees are included in the determination of the median 
compensation, compared to an expected median reduction of 50% or 
average reduction of 40% if no foreign employees are included in the 
determination.\533\
---------------------------------------------------------------------------

    \533\ See letters from COEC I.
---------------------------------------------------------------------------

    Some of the compliance costs outlined above may be ameliorated by 
the de minimis exemption in the final rule that allows for some 
flexibility in identifying the median employee for registrants that 
have employees in multiple countries. The final rule also provides a 
foreign data privacy law exemption that should help to reduce the 
burden on registrants that operate in certain foreign jurisdictions. 
For some registrants with small foreign operations, the de minimis 
exemption and the foreign data privacy law exemption might greatly 
reduce the importance of foreign operations as a driver of compliance 
costs.
    Several commenters subject to the proposed rule provided compliance 
costs estimates specific to their particular situation. Other 
commenters provided cost estimates for what appear to be anonymous but 
real companies.\534\ For all estimates received, we have used data 
available from Standard and Poor's Compustat,\535\ to obtain or confirm 
information about the number of employees at the registrant. Because, 
as noted above, certain registrants were not specifically identified, 
we could not confirm or obtain their number of employees or use them to 
construct a reliable cost-per-employee estimate, which is a key factor 
in our analysis below. Accordingly, we decided to focus our analysis on 
those estimates where the registrant was identified, although we have 
noted below what the impact on our estimates would be if we were to 
include the estimates from the unidentified registrants whose 
employment data we cannot verify.
---------------------------------------------------------------------------

    \534\ One commenter mentioned that ``Company B,'' a U.S. 
multinational manufacturer with approximately 130,000 employees in 
about 275 locations worldwide, including 30,000 employees in the 
United States and 100,000 overseas expects that the cost to build 
the global human resources information system needed to comply with 
the proposed rule would exceed $18 million. See letters from NAM I 
and NAM II. Another commenter mentioned that one survey respondent 
with over 50,000 employees across 69 countries described the costs 
of data gathering as over $10 million. See letter from Corporate 
Secretaries. The commenter only notes that the survey respondent has 
over a certain number of employees.
    \535\ The Standard and Poor's Compustat Database is a 
comprehensive database of company financials routinely used by us, 
academics, and practitioners in empirical assessments involving 
public companies.
---------------------------------------------------------------------------

    To estimate the potential compliance costs of the final rule, we 
analyze the detailed information on compliance costs provided in 
comment letters from 10 registrants \536\ and provide an assessment, 
below, of how their estimates might relate more broadly to other 
affected registrants. The reported expected costs vary in nature, but 
the common element among the commenters is the cost associated with 
modifying current payroll or accounting systems to compile the 
information necessary for the identification of the median employee and 
the calculation of the employee's compensation. In quantifying these 
and other reported potential costs, when registrants presented a range 
of cost estimates, we use the mid-range value (i.e., if a registrant 
indicated that its costs would range from $0.5 million to $1 million, 
we use $0.75 million in the analysis).\537\

[[Page 50159]]

When registrants indicated that they expect to incur at least a certain 
dollar amount of costs, we use that value in our estimation, and, as 
such, this represents a lower bound because we have no way of 
estimating the top of their range (i.e., if a registrant indicated that 
its costs would be at least $0.5 million, we use $0.5 million in the 
analysis).
---------------------------------------------------------------------------

    \536\ See letters from Avery Dennison, Dover Corp., Eaton, 
ExxonMobil, General Mills, Intel, NACCO, FEI, Hyster-Yale, and KBR.
    \537\ Since commenters provided no information on how likely it 
is that the costs would be close to the lower or upper bound of the 
range, we believe that using the midpoint of the range would provide 
a reasonable cost estimate.
---------------------------------------------------------------------------

    One commenter provided costs in terms of the number of hours rather 
than in dollar value terms.\538\ In this case, we converted the hours 
into a dollar amount using the hourly rates reported in one commenter's 
survey,\539\ according to which the median respondent expected to incur 
the cost of $700 per hour to comply with the proposed rule. This rate 
is higher than our estimate of $400 per hour for professional costs in 
the PRA section of the Proposing Release, and we believe it may 
overestimate the costs of compliance with the final rule. We continue 
to believe that $400 per hour is the appropriate rate to use for PRA 
purposes as this reflects the average cost for outside professional 
services for all registrants, including smaller and mid-sized 
registrants. Nevertheless, we have used the $700 per hour rate here to 
be conservative in our estimates and because the commenter in question 
is a large registrant with globally diversified operations. The 
estimate of $700 per hour for outside compliance counsel may be 
justified for a certain type of registrant with certain size and 
characteristics but may not be representative of all registrants. We 
also note that we have monetized the commenter's entire hourly estimate 
using the $700 per hour rate. This likely overstates the actual dollar 
value of these costs, as we expect that some of the compliance burden 
of the final rule will be carried internally by the registrant (e.g., 
using the registrant's existing workforce) and at rates significantly 
lower than $700 per hour.
---------------------------------------------------------------------------

    \538\ See letter from ExxonMobil.
    \539\ See letter from COEC I.
---------------------------------------------------------------------------

    The 10 registrants that quantified expected registrant-wide 
compliance costs tend to be large registrants (in terms of both assets 
and revenues) and have globally diverse operations, and as such, may 
not be representative of the costs incurred by smaller registrants or 
registrants that do not have foreign operations. Thus, we restrict the 
use of the information provided by them to estimate the expected 
compliance costs for only registrants subject to the requirements that 
we identify as having foreign operations. Since we believe that the 
compliance costs will be generally proportionate to the size of the 
registrant's work force, we calculate the cost per employee using the 
number of employees reported for fiscal year 2013.\540\
---------------------------------------------------------------------------

    \540\ This assumption implies that all compliance costs are 
variable. It is possible that some compliance costs have a fixed 
component (i.e., payroll system configuration costs and training), 
and thus compliance costs per employee may be higher for small and 
mid-size firms with globally diversified operations than for 
globally diversified large firms, which may result in adverse 
effects on competition to the extent that the former are not SRCs 
and are not covered by the de minimis or foreign data privacy 
exemptions. Data on the number of employees and revenues is taken 
from Standard and Poor's Compustat.
---------------------------------------------------------------------------

    For the 10 registrants with compliance cost estimates, we estimate 
the ratio of ``Compliance cost estimates'' to ``Number of employees.'' 
We then take the median ratio and use it to estimate the expected 
initial compliance costs for registrants with U.S.-based operations and 
registrants with foreign operations. We use the median instead of the 
average to diminish the influence of outliers. The individual estimates 
and the average and median are presented in the table below. We 
estimate that the average cost-per-employee for these registrants would 
be $50.70 and the median cost-per-employee would be $38.04. The average 
cost per registrant is approximately $971,500, while the median is 
$750,000.\541\ We note that the 10 registrants in this analysis are 
larger than the average and median registrant subject to the final 
rule. To adjust for differences in registrant size, we make the 
assumption that the compliance costs of the rule will be proportionate 
to the size of the registrant's work force, which enables us to use the 
cost-per-employee ratio to estimate the potential compliance costs of 
affected registrants.
---------------------------------------------------------------------------

    \541\ Our estimate of the cost per registrant aggregates the 
cost of outside professional services and monetized internal burden 
hours. One commenter referred to an average cost of $311,800 per 
registrant to comply with the pay ratio provision. The commenter 
also estimated that the median registrant would require 1,825 
internal burden hours to comply with the rule. These estimates are 
based on companies with operations in more than 50 countries that 
averaged 90 different employee data systems. See letter from Chamber 
II.

               Table 2--Total Initial Compliance Costs per Employee Based on Commenters' Estimates
----------------------------------------------------------------------------------------------------------------
                                                       Total
                                                    compliance      Revenue ($       Number of    Estimated cost
            Commenter                   CIK       cost estimates     millions)       employees     per employee
                                                     (dollars)                      (thousands)      (dollars)
----------------------------------------------------------------------------------------------------------------
Avery Dennison..................      0000008818       1,250,000         6,140.0              26           48.08
Dover Corporation...............      0000029905       2,000,000         8,729.8              37           54.05
Eaton Corporation...............      0000031277       1,600,000        22,046.0             102           15.69
ExxonMobil Corp.................      0000034088       2,100,000       390,247.0              75           28.00
General Mills...................      0000040704         750,000        17,909.6              43           17.44
Intel...........................      0000050863          15,000        52,708.0           107.6            0.14
NACCO...........................      0000789933         500,000           932.7             4.1          121.95
FEI Company.....................      0000914329         250,000           927.5            2.61           95.79
Hyster-Yale Materials Handling..      0001173514         500,000         2,666.3             5.1           98.04
KBR.............................      0001357615         750,000         7,214.0              27           27.78
                                 -------------------------------------------------------------------------------
Average.........................  ..............  ..............  ..............  ..............           50.70
Median \542\....................  ..............  ..............  ..............  ..............           38.04
----------------------------------------------------------------------------------------------------------------

    We estimate the aggregate costs for the 3,571 registrants to whom 
the rule will apply, including both registrants with U.S.-based 
operations and registrants with foreign operations, using data from 
Standard and Poor's Compustat. We

[[Page 50160]]

classify registrants as having foreign operations if they report having 
at least one geographical segment outside the United States. If they 
report only U.S.-based segments, then we classify them as having U.S.-
based operations only. We note several challenges in calculating these 
estimates. First, 13 of both the registrants with foreign operations 
and the registrants with U.S.-based operations only do not report the 
total number of employees. For these registrants, we impute the total 
number of employees by using the median number of employees for the 
registrants with foreign operations and the registrants with U.S.-based 
operations, respectively. Second, about 37% of affected registrants do 
not report geographic segment data, so we cannot classify them as 
registrants with U.S.-based operations or registrants with foreign 
operations.\543\ We compare the total assets and total revenues of 
these to the group of registrants with U.S.-based operations or 
registrants with foreign operations. These registrants tend to more 
closely resemble the registrants with U.S.-based operations, so we 
classify them as such. We recognize that this may underestimate the 
compliance cost for registrants with non-U.S. operations that choose 
not to report geographic segments.
---------------------------------------------------------------------------

    \542\ If we were to include the data for the anonymous 
registrants in the NAM I/NAM II letters and the Corporate 
Secretaries letter, the median would be $51.07. We note that the 
data for the anonymous registrant in the Corporate Secretaries 
letter did not specify a single number of employees, indicating 
instead that registrant had over 50,000 employees, thus making it 
difficult to compute a cost per employee ratio without additional 
assumptions. We assume that the number of employees of that 
anonymous registrant was 50,000 and its compliance cost was $10 
million.
    \543\ Compustat, like most other databases, has incomplete 
coverage of small companies, and reporting companies that do not 
trade on national exchanges.
---------------------------------------------------------------------------

    We estimate the initial compliance costs for the registrants with 
foreign operations by first aggregating the number of employees across 
all such registrants and then multiplying that number by the median 
estimated cost per employee, calculated as $38.04 in Table 2 above.
    To estimate the expected costs for registrants with U.S.-based 
operations only, we rely on one commenter's survey results that 
indicate that the median decrease in registrants' compliance costs 
would be 50% if only U.S. employees are included in the determination 
of the median employee compensation.\544\ Thus, we assume that the 
expected compliance costs for registrants with U.S.-based operations 
only will be 50% lower than the expected compliance costs for 
registrants with foreign operations. Accordingly, we estimate the 
compliance costs for registrants with only U.S.-based operations by 
multiplying the total number of employees across such registrants by 
the (median estimated costs per employee * (1-0.5)), which is $19.02 
per employee.\545\
---------------------------------------------------------------------------

    \544\ See letters from COEC I.
    \545\ $38.04*(1-0.5) = $19.02
---------------------------------------------------------------------------

    Lastly, to estimate the total compliance costs for registrants that 
do not report geographic segment data, which we reclassify as 
registrants with U.S.-based operations only, we first determine whether 
we have information on their number of employees.\546\ For the 973 
registrants that we reclassify as registrants with U.S.-based 
operations and for which we have information as to their number of 
employees, we aggregate the total number of employees for those 
registrants and multiply it by $19.02. For the other 335 registrants 
that we reclassify as registrants with U.S.-based operations, we do not 
have information on the number of employees. As mentioned above, these 
registrants are similar to registrants with U.S.-based operations with 
respect to total assets and total revenues. Thus, to estimate the 
compliance cost for those registrants, we estimate the median cost per 
registrant with U.S.-based operations only and multiply it by the 
number of such registrants (335). To estimate the median cost per 
registrant with U.S.-based operations only, we first multiplied $19.02 
per employee by the number of employees of each of the 793 registrants 
with U.S.-based operations only. We then found the median of those 793 
amounts, which was $27,008.40. Multiplying this number by the total 
number of registrants with missing employee data (335), we reached a 
total cost estimate of $9,047,814. Consistent with other estimates in 
our analysis, we used median costs rather than average costs to reduce 
the significance of outliers. We believe that our approach to the 
estimate is more appropriate because, by using median numbers, we 
reduce the significance of outliers, but we acknowledge that had we 
instead estimated based on average numbers, a significantly higher cost 
estimate for this group would result. Our cost estimates are presented 
in Table 3 below.
---------------------------------------------------------------------------

    \546\ For about a third of registrants with missing segment 
data, Compustat reports no employee data.

                    Table 3--Total Initial Compliance Cost Estimates for Affected Registrants
----------------------------------------------------------------------------------------------------------------
              Type of registrant                      Estimates                       Calculation
----------------------------------------------------------------------------------------------------------------
Registrants with foreign operations:
    Registrants affected......................                 1,470
    Total number of employees.................            27,595,305
    Cost/employee ratio.......................                $38.04
    Total cost................................        $1,049,725,402  27,595,305*$38.04
    Average cost per registrant...............              $714,099  $1,049,725,402/1,470
Registrants with U.S.-based operations only:
    Registrants affected......................                   793
    Total number of employees.................             6,522,626
    Cost/employee ratio.......................                $19.02  $38.04*(1-0.5)
    Total cost................................          $124,060,347  6,522,626*$19.02
    Average cost per registrant...............              $156,444  $124,060,347/793
    Median cost per registrant................            $27,008.40  This number represents the median of
                                                                       (number of employees * $19.02) across the
                                                                       793 U.S.-based registrants
Registrants with missing data, reclassified as
 registrants with U.S.-based operations only:
    Registrants affected......................                 1,308
    Registrants with available employee data..                   973
    Total number of employees for the 973                  6,932,754
     registrants.
    Cost/employee ratio.......................                $19.02  $38.04*(1-0.5)
    Total cost for the 973 registrants........          $131,860,981  6,932,754*$19.02
    Registrants with no employee data.........                   335  1,308-973
    Total cost for the 335 registrants........            $9,047,814  335 * $27,008.4

[[Page 50161]]

 
    Total cost................................          $140,908,795  131,860,981 + 9,047,814
        Total.................................   $1,314,694,544\547\  $1,049,725,402 + $124,060,347 +
                                                                       $140,908,795
----------------------------------------------------------------------------------------------------------------

     
---------------------------------------------------------------------------

    \547\ If we had included the cost estimates of the anonymous 
registrants provided by two commenters in calculating the median 
cost per employee ratio, as referenced above, the median cost per 
employee would be $51.07, and the total compliance cost for all 
3,571 registrants would have been $1,765 million. See letters from 
NAM I, NAM II (providing the same estimate as in the NAM I letter), 
and Corporate Secretaries.
---------------------------------------------------------------------------

    Based on our calculations, the average initial cost of compliance 
for a registrant with foreign operations is expected to be 
approximately $714,099 and for a registrant with U.S.-based operations 
only is expected to be approximately $156,444. The total initial cost 
of compliance for all 3,571 registrants affected by the Section 953(b) 
requirements is expected to be approximately $1,315 million, which 
includes both internal company costs as well as the costs of outside 
professionals.\548\ Thus, we estimate the initial cost of compliance 
for the average registrant to be approximately $368,159.\549\
---------------------------------------------------------------------------

    \548\ If average cost per registrant with U.S.-based operations 
($156,444) were used in lieu of median for the 335 registrants with 
missing employee data ($27,008.40), the total cost would instead be 
estimated as $1,049,725,402 + $124,060,347 + $131,860,981 + 
(335*$156,444) = $1,358 million.
    \549\ $1,314,694,544/3,571 = $368,159.
---------------------------------------------------------------------------

    It is important to note that this estimate does not reflect the de 
minimis and foreign data privacy exemptions, or the change to include 
only employees of consolidated subsidiaries, which would lead to some 
cost reductions for some registrants and which we are not able to fully 
quantify. Our cost estimate is higher than the survey estimate of $187 
million for the cost of outside professionals provided by one 
commenter, although we note that the commenter also estimated, but did 
not monetize, an internal company burden of 800,870 hours.\550\ 
Similarly, our estimated cost is higher than the other commenter's 
survey estimate of over $710 million, although that commenter also 
estimated, but did not monetize, an annual compliance time of 3.6 
million hours.\551\
---------------------------------------------------------------------------

    \550\ See letter from COEC I.
    \551\ See letter from Chamber II.
---------------------------------------------------------------------------

    Next, we estimate the ongoing compliance costs. Unlike in the case 
of the initial compliance costs, we received very few specific 
estimates of the ongoing compliance costs from commenters. One 
commenter suggested that ongoing annual costs would be approximately 
28% of the initial compliance costs.\552\ Another commenter reported 
expected ongoing compliance costs of 40%.\553\ One commenter's survey 
results suggest that the ongoing costs are expected to be about 80% 
(mid-range) or 72% (average) of the initial compliance costs.\554\ We 
note that some compliance costs of the final rule, such as burden hours 
and professional costs associated with making the disclosure, may 
remain consistent from year to year. Other compliance costs, however, 
will largely be upfront fixed costs, such as those associated with the 
modification of payroll or accounting systems to allow a registrant to 
compile the information and costs associated with developing the 
methodology needed to identify the median employee and calculate his or 
her pay.\555\ Given these upfront fixed costs, it is likely that that 
part of the initial compliance costs would decline after the first 
year.
---------------------------------------------------------------------------

    \552\ See letter from ExxonMobil. The commenter stated that it 
expects 3,000 work hours for initial compliance costs and 850 work 
hours for ongoing compliance costs. This suggests the ongoing 
compliance cost is approximately 28.3% (850/3000) of the initial 
compliance costs.
    \553\ See letter from FEI. According to this commenter, the 
initial compliance costs would be approximately $250,000, while the 
ongoing compliance costs would be approximately $100,000, suggesting 
that the latter is 40% of the former ($100,000/$250,000).
    \554\ See letter from COEC I.
    \555\ But see letter from Business Roundtable I (asserting that 
42% of respondents to a survey indicated that they would have to 
update their methodology every year). We note that ongoing costs of 
compliance may represent a higher percentage of the initial costs of 
compliance for these respondents.
---------------------------------------------------------------------------

    The specific estimates provided by commenters (28% to 72%) \556\ 
yield a range of ongoing compliance cost estimates of between $368 
million and $947 million per year, with the median of the estimates 
provided by these commenters (40%) yielding an ongoing compliance cost 
of approximately $526 million per year.\557\ We note, however, that the 
Proposing Release did not provide registrants with the flexibility to 
identify the median employee every three years. We assume these three 
estimates are based on the commenters' reading of the Proposing 
Release, and hence include the requirement that the median employee be 
identified every year.
---------------------------------------------------------------------------

    \556\ Other commenters made more general assertions about 
ongoing compliance costs, but because they did not provide specific 
cost estimates, we did not include them in this calculation. See, 
e.g., letters from General Mills (asserting that ``most of the costs 
would also apply to subsequent years of compliance'') and Business 
Roundtable I.
    \557\ $1,315 million * 0.28 = $368 million; $1,315 million * 
0.72 = $947 million; $1,315 million * 0.4 = $526 million.
---------------------------------------------------------------------------

d. Indirect Costs
    Registrants covered by the final rule also could be affected by 
indirect costs.\558\ They could be at a competitive disadvantage to 
registrants (including private companies, foreign private issuers, 
smaller reporting companies and emerging growth companies) that are 
outside the scope of the final rule. Some commenters suggested that the 
proposed rule would cause competitive disadvantages for public 
companies,\559\ U.S. companies, especially those with overseas 
employees,\560\ issuers with subsidiaries,\561\ and issuers with low-
wage workers.\562\ In addition, we understand from commenters that some 
registrants covered by the final rule would likely incur higher costs 
of compliance based on size, business type, and level of integration of 
payroll and benefits systems--such as large, multinational registrants 
that do not maintain integrated employee compensation information on a 
global basis.\563\ Therefore, the competitive impact of compliance with 
the disclosure requirements prescribed by Section 953(b) could 
disproportionately fall on U.S. registrants with large workforces and 
global operations, although the de minimis exemption and the foreign 
data privacy law exemption in the final rule would likely reduce some 
of these compliance costs and the competitive effects of the final 
rule. While we expect that the incremental

[[Page 50162]]

impact of the fixed components of compliance costs will have a lower 
impact on larger registrants than on smaller registrants, as discussed 
in the previous section, the overall compliance costs will likely 
lessen after systems are in place to gather and verify the underlying 
data, reducing the competitive effects of the final rule over the 
years.\564\
---------------------------------------------------------------------------

    \558\ We cannot precisely quantify the indirect costs because 
they depend on the registrant's business structure and competitive 
environment.
    \559\ See, e.g., letters from ABA, Prof. Angel, Former Assistant 
Secretary Campbell, Chamber I, COEC I, Corporate Secretaries, NAM I, 
NAM II.
    \560\ See, e.g., letters from ABA, Chamber I, COEC I, and 
Corporate Secretaries.
    \561\ See letter from COEC I.
    \562\ See, e.g., letters from IBC and Prof. Ray.
    \563\ See supra note 528.
    \564\ But see letters from Business Roundtable I (noting that 
42% of respondents to a survey ``indicated that they expect having 
to update their pay ratio methodology every year because of changes 
to their business organization or structure'') and COEC I (arguing 
that the burden hours and professional costs may decrease after the 
first year but not by the estimated 50% in the Proposing Release).
---------------------------------------------------------------------------

    Registrants subject to the final rule could also face a competitive 
disadvantage if their competitors are able to infer proprietary or 
sensitive information from the disclosure of the median of the annual 
total compensation of all employees.\565\ For example, it could be 
possible in some instances for a competitor to infer sensitive 
information about a registrant's cost structure based on information 
about median levels of employee compensation, especially for small 
registrants operating in a single industry. While the final rule does 
not apply to smaller reporting companies and emerging growth companies, 
we still estimate that at least 131 affected registrants operated in a 
single business segment and had book value of assets less than $100 
million in 2014.\566\ As we noted above, a registrant subject to 
Section 953(b) could potentially be at a competitive disadvantage when 
hiring or retaining a PEO if there is pressure to limit PEO wages based 
on the pay ratio disclosure while non-covered registrants are not 
subject to the same pressure. However, there may be other factors 
affecting the ability of a registrant to attract and retain executive 
talent.
---------------------------------------------------------------------------

    \565\ See letter from Prof. Ray. But see letter from Capital 
Strategies (stating, on the contrary, that competitors will not be 
able to decipher any proprietary or sensitive information from the 
pay ratio disclosure).
    \566\ This estimate is based on data from the Standard and 
Poor's Compustat Segments database as of the end of December 2014.
---------------------------------------------------------------------------

    One of the commenters indicated that 55% of the respondents in a 
survey of members it conducted anticipated indirect costs (i.e., 
adverse impact on sales, brand damage, increased public relations costs 
etc.)\567\ Although we acknowledge the possibility of these indirect 
costs, we cannot quantify them and lack sufficient data to analyze 
them.
---------------------------------------------------------------------------

    \567\ See letter from Corporate Secretaries.
---------------------------------------------------------------------------

3. Other Economic Effects
    Several commenters indicated that the pay ratio rule would promote 
economic efficiency.\568\ In contrast, one commenter argued the rule 
would inhibit economic efficiency without providing specific 
details.\569\ As noted above, the pay ratio disclosure is not well 
suited to compare pay practices across registrants, and thus, it is 
unclear whether the final rule would affect economic efficiency. Some 
commenters suggested that registrants may decide to alter their pay 
structure or workforce structure in ways that are different from their 
efficient labor market decisions.\570\ For example, one commenter 
suggested that registrants may decide to shift their labor force to 
workers employed by a third party or reduce their foreign 
operations.\571\ Also, the same commenter asserted that registrants may 
change the relative wages of employees in a way that increases the 
median employee pay while reducing the pay of employees below the 
median employee in the pay distribution.\572\ Another commenter 
asserted that pressure for a registrant to maintain a low pay ratio 
could also curtail the expansion of business operations into lower cost 
geographies.\573\ We expect that such changes, if they were to occur, 
would move the registrant away from efficient business practices and 
could result in inefficient outcomes. Other commenters suggested, 
however, that workforce restructuring in response to the pay ratio 
disclosure was not likely.\574\ While we believe that registrants are 
unlikely to make critical labor decisions solely to impact the pay 
ratio disclosure, we cannot assess the prevalence of such effects at 
this time because these commenters did not quantify or otherwise 
provide data relevant to the expected changes in business practices.
---------------------------------------------------------------------------

    \568\ See, e.g., letters from Capital Strategies (asserting that 
``The proposed rules would promote market efficiency as they would 
steer investors to the best value CEO, that is, the one producing 
the most returns for the least pay, when compared to peers.''), 
Change to Win (asserting that CEO pay increases have been driven by 
rent seeking behavior and ``if in fact the disclosures provided by 
Section 953(b) do induce more investors to insist on limiting 
executive pay, this will result in increased, rather than reduced, 
economic efficiency''), and Form Letter D (``Disclosure of the pay 
ratios will help the capital markets better allocate capital to 
those companies that invest in their workforces.'').
    \569\ See letter from NIRI.
    \570\ See, e.g., letters from CEG (stating that the rule can be 
``gamed by outsourcing lower wage jobs''), NIRI (stating that the 
rule will adversely impact ``U.S. states and cities with lower labor 
costs''), and Prof. Ray (arguing that employers will change their 
corporate structure or employment arrangements as a response to the 
pay ratio rule).
    \571\ See letter from Lou (``[S]ome companies will be 
incentivized to outsource poorly paid jobs to increase the median 
payment number. The flip side is the increase of administrative cost 
and losses of profits generated by the to-be-outsourced department. 
Or companies will reduce workforce in their foreign subsidiaries 
where the labor cost is relatively low. But obviously this reduction 
sacrifices the low labor cost advantage. Therefore the proposal does 
not motivate CEOs to maximize companies' interests.'').
    \572\ Id.
    \573\ See letter from Prof. Ray (providing the example of a firm 
that might prefer to hire in the U.S. rather than in India because a 
strong exchange rate of the U.S. dollar against the Indian rupee 
will make the Indian wages appear low and can lead to high pay ratio 
if the firm hires employees in India).
    \574\ See letters from ABA (asserting that registrants would not 
incur the related costs to alter their organizational structure or 
workforce in order to improve their pay ratio disclosure), Capital 
Strategies (asserting that the definition of ``all employees'' would 
not cause registrants to alter their corporate structure or 
employment arrangements), and WorldatWork I (same).
---------------------------------------------------------------------------

D. Economic Effects From Exercise of Discretion

1. General
    In this section, we discuss the choices we have made in 
implementing the statutory requirements and the associated economic 
effects, including the likely benefits and costs and the likely impact 
on efficiency, competition, and capital formation. In addition to the 
statutory benefits and costs described above, we believe that the use 
of our discretion in implementing the statute could result in benefits 
and costs to registrants and users of the pay ratio disclosure.
    In general, the final rule implementing Section 953(b) is designed 
to comply with the statutory mandate. In light of the significant 
potential costs that commenters attribute to the requirements of 
Section 953(b),\575\ we believe that it is appropriate for the final 
rule to permit registrants certain flexibility in calculating the pay 
ratio, which we believe should help lower the costs of compliance 
generally while still providing the information directed by Section 
953(b). In addition, the final rule generally seeks to implement 
Section 953(b) without imposing additional requirements that are not 
mandated by the Dodd-Frank Act. In this respect, the final rule 
reflects our consideration of the relative costs and benefits of a more 
flexible approach as opposed to a more prescriptive approach.
---------------------------------------------------------------------------

    \575\ See, e.g., letters from Aon Hewitt, Avery Dennison, 
Business Roundtable I, Chamber II, Chesapeake, COEC I, COEC II, COEC 
III, Corporate Secretaries, Eaton, FEI, FuelCell Energy, Garmin, 
General Mills, Hyster-Yale, IBC, KBR, NACCO, NACD, NAM I, NAM II, 
NIRI, PNC Financial Services, and Semtech.
---------------------------------------------------------------------------

    In evaluating alternatives, we considered whether to adopt a rule 
that would be prescriptive enough that the

[[Page 50163]]

resulting pay ratio disclosure would be more directly comparable across 
registrants. As noted above, we believe that comparability of the ratio 
across registrants has significant limits, due to the variety of 
factors that could influence the ratio. We believe that providing a 
flexible approach would not significantly diminish the potential 
benefits of the mandated disclosure and would achieve the purposes that 
Congress intended at a significantly reduced burden for registrants. In 
this respect, we note that some commenters indicated that the expected 
benefits of pay ratio disclosure derive from its ability to facilitate 
a company-specific assessment, by providing a metric by which a PEO's 
compensation can be evaluated within the context of that particular 
company.\576\ We also acknowledge that some commenters that support the 
pay ratio disclosure suggested that it could be used to compare 
compensation practices between registrants and/or for the same 
registrant over time.\577\ We note, however, that using the ratios to 
compare compensation practices between registrants, and for a 
registrant over time (e.g., in the case of business acquisitions or 
significant structural or business model changes), without taking into 
account inherent differences in business models between registrants and 
for a registrant over time, which may not be readily available 
information, could potentially lead to unwarranted conclusions.
---------------------------------------------------------------------------

    \576\ See, e.g., letters from B[acirc]tirente et al., 
Bricklayers International, CalSTRS, Calvert, Domini, FS FTQ, Pax 
World Funds, Walden, and WA State Investment Board.
    \577\ See, e.g., letters from B[acirc]tirente et al., 
Bricklayers International, CalPERS, CalSTRS, Calvert, Domini, 
EnTrust Capital (Nov. 20, 2013) (``EnTrust''), FS FTQ, LIUNA, Pax 
World Funds, Public Citizen I, RPMI, Walden, and WA State Investment 
Board.
---------------------------------------------------------------------------

2. Implementation Choices and Alternatives
a. Filings Subject to the Pay Ratio Disclosure Requirements
    Commenters suggested that the final rule should apply only to those 
filings for which the applicable form requires Item 402 
disclosure.\578\ Some commenters stated that requiring pay ratio 
disclosure in every filing would be unnecessary and even 
confusing.\579\ The final rule follows the proposed approach to require 
pay ratio disclosure in filings described in Item 10(a) of Regulation 
S-K that require executive compensation disclosure under Item 402 of 
Regulation S-K. We believe that requiring pay ratio disclosure in 
filings that do not contain other executive compensation information 
would not present this information in the most meaningful context. Some 
commenters asserted that the pay ratio disclosure would provide another 
metric to evaluate executive compensation disclosure,\580\ and we 
believe that the intended purpose of the disclosure is to provide new 
data points that shareholders can use when exercising their new voting 
rights under Section 951. We believe that the pay ratio disclosure 
would be presented in a more meaningful context if it were accompanied 
by other Item 402 information, such as the Summary Compensation Table 
required by Item 402(c) and the Compensation Discussion and Analysis 
required by Item 402(b). Therefore, we believe that this choice 
preserves the intended benefits of Section 953(b) while reducing 
reporting costs relative to a requirement to include pay ratio 
disclosure in every filing, including in filings that do not require 
other Item 402 information.
---------------------------------------------------------------------------

    \578\ See letters from ABA, AFL-CIO I, CalPERS, Calvert, Capital 
Strategies, CII, CT State Treasurer, Davis Polk, Intel, Johnson & 
Johnson, McGuireWoods, NIRI, NRF, Pax World Funds, PM&P, Prof. Ray, 
and WorldatWork.
    \579\ See, e.g., letters from ABA (``In our view, it is neither 
reasonable nor sensible to mandate the inclusion of pay ratio 
disclosure in filings where no other executive compensation 
disclosure required by Item 402 will appear to provide meaningful 
context.'') and PM&P (``Providing such information in multiple 
filings (e.g., registration statements, annual reports or other 
filings) throughout the year is unnecessary and would dilute the 
usefulness, if any, of the disclosure.'').
    \580\ See supra note 34.
---------------------------------------------------------------------------

b. Registrants Subject to the Pay Ratio Disclosure Requirements
    We recognize that the reference to ``each issuer'' in Section 
953(b) could be interpreted to apply to all registrants. However, as a 
result of the specific reference in Section 953(b) to the definition of 
``total compensation'' contained in Item 402(c)(2)(x) and the absence 
of Congressional direction to apply this requirement to registrants not 
previously subject to Item 402(c) requirements, the final rule does not 
apply to registrants that are not subject to Item 402(c) requirements. 
Thus, smaller reporting companies, foreign private issuers, and MJDS 
filers are excluded. In addition, Congress exempted emerging growth 
companies from the requirement in the JOBS Act.
    We considered a number of alternative approaches. We considered 
whether a broader reading of the statute was warranted in the context 
of SRCs as suggested by some commenters.\581\ However, most commenters 
agreed with the approach to exclude SRCs from the requirements.\582\ 
Commenters either argued that these registrants are not currently 
required to provide a Summary Compensation Table under Item 402(c) and 
therefore should not be required to comply with the pay ratio rule 
\583\ or cited high costs of compliance.\584\ Requiring SRCs to provide 
the pay ratio disclosure consistent with the requirement for other 
registrants would require them to collect data and calculate 
compensation for the PEO in a manner they otherwise would not do, and 
there would be some incremental costs in doing so. However, these 
incremental costs may be limited to the extent that smaller reporting 
companies are less likely to have defined benefit and actuarial pension 
plans.\585\ In contrast, the costs of complying with the other 
requirements prescribed by Section 953(b)--namely, identifying the 
median employee and calculating annual total compensation for that 
employee--are more extensive. We can estimate those costs using the 
approach and estimates we made in the ``Quantification of Compliance 
Costs'' section above. We identify 2,958 registrants as SRCs that are 
not EGCs as of the end of fiscal year 2014. Of these, 494 have data on 
segments and number of employees in Compustat. Following the approach 
in the ``Quantification of Compliance Costs'' section above, we use 
information on international geographic segments reported in Compustat 
to identify registrants with and without international operations. Of 
the 2,958 SRCs, 212 have foreign operations and 282 have U.S.-based 
operations only; the rest does not have

[[Page 50164]]

segment data in Compustat.\586\ We next obtain the total number of 
employees for SRCs with foreign operations and SRCs with U.S.-based 
operations only from Compustat. We apply the respective ``compliance 
cost per employee'' ratios for registrants with foreign and U.S.-based 
operations only estimated in section ``Quantification of Compliance 
Costs'' above to estimate the average and total compliance costs for 
SRCs with foreign operations and SRCs with U.S.-based operations only. 
We reclassify the remaining 2,464 registrants as SRCs with U.S.-based 
operations only because our analysis suggests that registrants that are 
not covered by Compustat are usually smaller companies without foreign 
operations. Consistent with our methodology for Table 3, to estimate 
the total compliance costs for SRCs that do not report segment data, 
which we reclassify as SRCs with U.S.-based operations only, we 
estimate their total number of employees and multiply it by $19.02. For 
those SRCs we reclassify as SRCs with U.S.-based operations only that 
do not have employee data available, we estimate the median cost per 
SRC with U.S.-based operations only and multiply it by the number of 
such registrants. The table below presents our estimates of the 
compliance costs of SRCs with foreign operations and SRCs with U.S.-
based operations only, as well as total compliance costs for SRCs.
---------------------------------------------------------------------------

    \581\ See, e.g., letters from CII (indicating that two of three 
CII members that commented on the proposed rule ``were `not 
comfortable' with the proposed exemption from the pay ratio 
disclosure requirements for emerging growth companies, smaller 
reporting companies, foreign private issuers, and MJDS filers''), 
Ray (``To support Congress's demand for greater pay-related 
disclosures, I strongly suggest for the Commission to expand the 
disclosure requirement to ensure that all smaller reporting 
companies disclose their pay ratio.''), and US SIF (``While we 
understand the SEC's reasons for several exemptions from the 
proposed rule for emerging growth companies, smaller companies and 
foreign private issuers, we are, nonetheless, uncomfortable with 
these proposed exemptions.'').
    \582\ See, e.g., letters from ABA, Prof. Angel, CalPERS, Capital 
Strategies, Davis Polk, Hay Group, NIRI, NY State Comptroller, PM&P, 
Vivient, and WorldatWork.
    \583\ See, e.g., letters from CalPERS and WorldatWork (``Smaller 
companies would face a double compliance burden if asked to publish 
summary compensation tables and calculate the pay ratio.'').
    \584\ See, e.g., letters from Prof. Angel and Vivient.
    \585\ See supra note 92.
    \586\ For approximately 60% of SRCs with missing segment data, 
Compustat reports no employee data.

                            Table 4--Total Initial Compliance Cost Estimates for SRCs
----------------------------------------------------------------------------------------------------------------
              Type of registrant                      Estimates                       Calculation
----------------------------------------------------------------------------------------------------------------
Registrants with foreign operations:
    Registrants affected......................                   212
    Total number of employees.................                60,280
    Cost/employee ratio.......................                $38.04
    Total cost................................            $2,293,051  60,280*$38.04
    Average cost per registrant...............               $10,816  2,293,051/212
Registrants with U.S.-based operations only:
    Registrants affected......................                   282
    Total number of employees.................                89,625
    Cost/employee ratio.......................                $19.02  $38.04*(1-0.5)
    Total cost................................            $1,704,668  89,625 *$19.02
    Average cost per registrant...............                $6,045  1,704,668/282
    Median cost per registrant................                $1,103  This number represents the median of
                                                                       (number of employees * $19.02) across the
                                                                       282 U.S.-based registrants
Registrants with missing data, reclassified as
 U.S.-based operations only:
    Registrants affected......................                 2,464
    Registrants with available employee data..                 1,150
    Total number of employees for the 1,150                  245,752
     registrants.
    Cost/employee ratio.......................                $19.02  $38.04*(1-0.5)
    Total cost for the 1,150 registrants......            $4,674,203  245,752*$19.02
    Registrants with no employee data.........                 1,314  2,464-1,150
    Total cost for the 1,314 registrants......            $1,449,342  1,314 * $1,103
    Total cost................................            $6,123,545  4,674,203+1,449,341
        Total.................................           $10,121,264  2,293,051 +1,704,668
                                                                      +6,123,545
----------------------------------------------------------------------------------------------------------------

    Our decision not to require SRCs to comply with the pay ratio 
disclosure requirements prescribed by Section 953(b) would save the 
average SRC with foreign operations approximately $10,816, and the 
average SRC with U.S.-based operations only approximately $6,045. We 
note that these cost savings are the savings from not having to 
identify the median employee and calculate the total compensation for 
that employee. Those cost savings from exercise of our discretion with 
respect to SRCs total approximately $10 million.\587\ We expect there 
also would be cost savings from not having to calculate PEO 
compensation pursuant to Item 402, but we are unable to quantify those 
savings.
---------------------------------------------------------------------------

    \587\ If average cost per SRC with U.S.-based operations were 
used in lieu of median for registrants with missing data, the total 
cost savings would instead be estimated as $2,293,051 + $1,704,668 + 
$4,674,203 + 1,314*$6,045 = $16.62 million.
---------------------------------------------------------------------------

    To the extent that these costs have a fixed component that does not 
depend on the registrant's size of operations, the compliance burden 
for small registrants may be disproportionately large.\588\ Moreover, 
small companies are more likely to operate in a single geographic or 
business segment, making the disclosure of the median employee pay more 
likely to reveal sensitive or proprietary information that can put 
these registrants at a competitive disadvantage.
---------------------------------------------------------------------------

    \588\ See letter from Vivient (``From a cost standpoint, 
requiring smaller reporting companies to disclose the pay ratio 
would create a significant cost and administrative burden. Given the 
size of their employee population and administrative budget, very 
few smaller reporting companies employ full-time senior level human 
resource professionals who could be assigned the responsibility of 
complying with the proposed ruling.'').
---------------------------------------------------------------------------

    We also considered expanding the coverage of the final rule to 
registrants, such as foreign private issuers and MJDS filers, which are 
not currently required to provide Item 402 disclosure. Most commenters 
agreed with the exclusion of foreign private issuers and MJDS 
filers,\589\ but other commenters expressed concerns about excluding 
them.\590\ Although quantifying the costs to these registrants of 
calculating PEO compensation under Item 402(c)(2)(x) or of complying 
with the requirements

[[Page 50165]]

prescribed by Section 953(b) is not currently feasible because of lack 
of data, we assume that these costs (and in particular the costs of 
computing the median employee) could be significant. In particular, 
these costs may be higher for foreign private issuers and MJDS than for 
SRCs because these registrants are not currently required to provide 
any Item 402 disclosure. Based on a review of EDGAR filings for 
calendar year 2014, we estimate that there are approximately 677 
foreign private issuers filing on Form 20-F and 143 MJDS filers filing 
on form 40-F that will benefit from the exclusion from the pay ratio 
disclosure requirements.
---------------------------------------------------------------------------

    \589\ See letters from ABA (stating that Section 953(b) does not 
require expanding the scope of Item 402 of Regulation S-K to apply 
to registrants not currently required to comply with the Item 402 
disclosure, such as foreign private issuers and MJDS filers, so the 
Commission should not expand its rules to do so), Capital 
Strategies, Davis Polk, Hay Group, and PM&P.
    \590\ See letters from CalPERS, CII, and US SIF.
---------------------------------------------------------------------------

c. Employees Included in the Determination of the Median
    Section 953(b) requires disclosure of the median of the annual 
total compensation of ``all employees of the issuer.'' Consistent with 
that mandate, the final rule includes in the definition of ``employee'' 
or ``employee of the registrant'' any U.S. and non-U.S. full-time, 
part-time, seasonal, or temporary worker (including officers other than 
the PEO) employed by the registrant or any of its subsidiaries as of 
the last day of the registrant's last completed fiscal year. 
Additionally, as set forth above, we are excluding from the 
determination of the median employee any workers who are not employed 
by the registrant or its consolidated subsidiaries, such as independent 
contractors and ``leased'' workers who are employed by, and whose 
compensation is determined by, an unaffiliated third party.
i. Types of Employees
    Commenters were generally split on whether the rule should include 
part-time, seasonal, or temporary employees. A number of commenters 
agreed with the proposed requirements to include part-time, temporary, 
and seasonal workers because of Section 953(b)'s reference to ``all 
employees'' and believed that excluding these employees would distort 
the pay ratio by rendering it incomplete or misleading.\591\ Other 
commenters suggested that the final rule should exclude part-time, 
seasonal, or temporary employees. These commenters asserted that 
compensation for these employees is not comparable to full-time 
employees, so their inclusion would distort the pay ratio.\592\ Some 
commenters believed that the final rule should exclude part-time, 
seasonal, or temporary employees because the potential benefits from 
including them would not justify the high costs.\593\ Another commenter 
recommended that the final rule should exclude part-time, seasonal, and 
temporary employees unless a majority of a registrant's employees work 
on a part-time, temporary, and/or seasonal basis.\594\
---------------------------------------------------------------------------

    \591\ See letters from AFL-CIO I, AFR, B[acirc]tirente et al., 
Bricklayers International, CII, CT State Treasurer, FS FTQ, Public 
Citizen I, and Theodore.
    \592\ See letters from AAFA I, AAFA II, American Benefits 
Council, COEC I, Corporate Secretaries, and Davis Polk.
    \593\ See letters from COEC I (``Nearly all firms, 122 of 128, 
have part-time and/or seasonal employees. With this in mind, two-
thirds of survey respondents to the survey indicated that limiting 
the application of the proposed pay ratio rules to full-time 
employees only would reduce their costs. The average savings for 
these respondents would be approximately 20 percent.''), General 
Mills (which estimated that approximately 13% of their employees are 
part-time, seasonal or temporary), and WorldatWork II (recommending 
exclusion of part-time, seasonal, and temporary employees from the 
pay ratio).
    According to one of the commenters, the average percentage of 
full-time employees reported by survey respondents was 86% (mid-
range was 95%). See letter from COEC I. Another commenter estimates 
the average (median) percentage of full-time employees to be 87% 
(95%). See letter from Corporate Secretaries. Based on BLS data, 
approximately 81% of workers were employed full time as of 2014. See 
BLS Labor Force Statistics from the Current Population Survey, 
available at http://www.bls.gov/cps/cpsaat12.htm. We note that BLS 
data incorporates workers at a broader range of firms, including 
privately held firms and small firms, which may not be 
representative of the composition of the workforce at the 
registrants subject to the final rule.
    \594\ See letter from ABA.
---------------------------------------------------------------------------

    The final rule requires registrants to include part-time, 
temporary, and seasonal employees when identifying the median employee. 
We could have chosen an alternative approach, namely to allow 
registrants to base their pay ratio disclosure on full-time employees 
only. This approach would have led to a lower cost of compliance. 
According to one survey, such flexibility would have generated median 
savings of approximately 10% in compliance costs.\595\ Applying this 
estimate to our compliance cost estimate of $1,315 million, had we 
chosen this alternative the total compliance costs would have been 
approximately $1,183.5 million, or savings of approximately $131.5 
million. According to another commenter, excluding part-time employees 
could reduce costs of compliance by 20%, which would raise the estimate 
of potential cost savings to approximately $263 million.\596\ Another 
commenter noted that more than 30 percent of respondents to its survey 
believe that limiting the median employee calculation to full-time 
employees would yield cost savings of more than 10 percent and an 
additional 18 percent of respondents believe that limiting the median 
employee calculation to full-time employees will yield cost savings of 
more than 20 percent.\597\ Despite these potential cost savings, we are 
not adopting this alternative. Several commenters argued that excluding 
part-time, temporary, and seasonal workers from the definition of 
``employee'' would make the disclosure incomplete and/or not 
representative of the registrant's actual workforce.\598\ We agree and 
have concluded, as discussed more fully in section B.1.a. above, that 
this approach could significantly undermine the intent of the rule to 
allow a company-specific assessment of a registrant's compensation 
practices with respect to ``all employees''.
---------------------------------------------------------------------------

    \595\ See letter from Corporate Secretaries.
    \596\ See letter from COEC III.
    \597\ See letter from WorldatWork I.
    \598\ See, e.g., letters from AFL-CIO I, AFR, B[acirc]tirente et 
al., Bricklayers International, CII, CT State Treasurer, FS FTQ, and 
Public Citizen I.
---------------------------------------------------------------------------

ii. Workers Not Employed by the Registrant (i.e., Leased Workers)
    The final rule, as proposed, excludes independent contractors and 
``leased'' workers who are employed by, and whose compensation is 
determined by, an unaffiliated third party. Commenters generally 
supported this approach.\599\ As discussed, we believe excluding such 
workers is appropriate because registrants generally do not control the 
level of compensation that these workers are paid.
---------------------------------------------------------------------------

    \599\ See supra note 114.
---------------------------------------------------------------------------

    We recognize that it is possible that a registrant could alter its 
corporate structure or its employment arrangements to reduce the number 
of employees covered by the final rule and, therefore, reduce its costs 
of compliance or alter its pay ratio disclosure to achieve a particular 
objective. For example, a registrant could choose to use only 
independent contractors or ``leased'' workers instead of hiring 
employees. A registrant could also choose to outsource some aspects of 
its business to achieve similar objectives. Although one commenter 
asserted that registrants would change their corporate structure or 
employment arrangements based on the definition of ``employee,'' \600\ 
other commenters questioned the likelihood of this behavior.\601\ We 
cannot quantify the expected prevalence of this behavior. However, 
given the inherent complexities involved in altering a registrant's 
corporate structures or employment arrangements, we do not expect that 
many registrants would undertake such changes merely for the

[[Page 50166]]

purposes of lowering compliance costs or achieving a particular pay 
ratio.
---------------------------------------------------------------------------

    \600\ See letter from Prof. Ray.
    \601\ See, e.g., letters from ABA, Capital Strategies, and 
WorldatWork I.
---------------------------------------------------------------------------

iii. Employees of Consolidated Subsidiaries
    As discussed above, of the commenters that discussed whether to 
include employees of a subsidiary, the majority recommended that the 
final rule require registrants to include only employees of certain 
types of subsidiaries, in particular consolidated or wholly-owned 
subsidiaries.\602\ One of those commenters claimed that there would be 
a median increase in compliance costs of approximately 20% if the final 
rule included employees of all minority-owned subsidiaries and joint 
ventures.\603\ Another commenter argued that limiting the final rule to 
consolidated subsidiaries would reduce costs and burdens and is 
consistent with Rule 405 of the Securities Act and Rule 12b-2 of the 
Exchange Act.\604\
---------------------------------------------------------------------------

    \602\ See letters from ABA, Best Buy et al., Business Roundtable 
I, COEC I, Corporate Secretaries, CT State Treasurer, Davis Polk, 
Eaton, ExxonMobil, General Mills, Mercer I, Meridian, NACCO, NAM I, 
and NAM II.
    \603\ See letter from Corporate Secretaries.
    \604\ See letter from ABA.
---------------------------------------------------------------------------

    The final rule defines ``employee'' to include only the employees 
of the registrant and its consolidated subsidiaries rather than 
employees of subsidiaries that were affiliates it controlled directly 
or indirectly through one or more intermediaries, as set forth in the 
definition of ``subsidiary'' under both Securities Act Rule 405 and 
Exchange Act Rule 12b-2. This change will affect registrants that have 
unconsolidated subsidiaries with a significant number of workers.
    We believe that excluding employees of unconsolidated subsidiaries 
may provide a better representation of the compensation practices of 
the registrant itself since the compensation provided by unconsolidated 
subsidiaries may be beyond the control of the registrant covered by 
Section 953(b).
    Based on our analysis of Compustat firms for calendar year 2014, 
excluding firms identified as emerging growth companies, smaller 
reporting companies, foreign private issuers, and MJDS filers, 
approximately 23% of firms reported positive equity investments in 
unconsolidated subsidiaries,\605\ with the median investment of 
approximately 1.7% of the book value of total assets. The majority of 
the firms were in the financial, regulated utilities, and oil and gas 
industries. We lack information on the number of employees at 
unconsolidated subsidiaries to quantify the potential magnitude of 
their effect on the pay ratio disclosure.
---------------------------------------------------------------------------

    \605\ Our analysis excludes investments in unconsolidated 
subsidiaries accounted for by the cost method as those are not 
identified separately by filers in the data available to us.
---------------------------------------------------------------------------

    We also reviewed Exhibit 21 to the annual reports on Form 10-K, 
which contains subsidiary information, for a sample of 24 firms that 
submitted individual unique comment letters pertaining to the proposal. 
The median number of subsidiaries in that set of firms was 
approximately 31. Only three registrants explicitly identified the 
number of consolidated subsidiaries, with the median being 
approximately 36. The registrants we studied may not be random as firms 
with more subsidiaries may be more likely to submit a comment letter. 
The information in the exhibits indicates that some firms have complex 
organizational structures but it does not allow us to systematically 
differentiate between consolidated and unconsolidated subsidiaries.
    The final rule allows registrants to exclude employees from 
unconsolidated subsidiaries when identifying the median employee. This 
change from the proposed rule could lead to significant cost 
savings.\606\ First, limiting the definition in this way will result in 
a smaller pool of employees from which to identify the median employee, 
thereby helping to reduce compliance costs associated with this step. 
Second, registrants are more likely to maintain integrated systems with 
their consolidated subsidiaries because these subsidiaries have to 
consolidate their financial statements with those of the registrant, 
which should make it easier to collect and analyze the relevant 
data.\607\ Finally, as the consolidated subsidiary standard is commonly 
applied in other disclosures, there may be less cost for registrants to 
identify subsidiaries relevant for the disclosure. In summary, the 
final rule could provide a potential competitive advantage to 
registrants with a significant percentage of the workforce at 
unconsolidated subsidiaries over registrants with consolidated 
subsidiaries due to lower compliance costs associated with having fewer 
workers covered by the rule.
---------------------------------------------------------------------------

    \606\ See, e.g., letters from ABA, COEC I (claiming that there 
would be a 91% increase on average in costs if registrants were 
required to include all minority-owned subsidiaries and joint 
ventures in the definition of ``employee''), and Corporate 
Secretaries (survey reporting a median increase in costs of 20% if 
the definition of ``employee'' included all minority-owned 
subsidiaries and joint ventures of the registrants).
    \607\ See letter from COEC I. The letter indicates that in the 
majority of cases, a registrant's access to the information 
necessary to calculate the pay ratio will only extend to wholly-
owned subsidiaries that consolidate their financial statements with 
those of the registrant.
---------------------------------------------------------------------------

    We have attempted to quantify the expected decrease in compliance 
costs from the revised definition of subsidiaries of the registrant, 
but did not obtain estimates on what these costs would be. One 
commenter's survey results suggested that compliance costs would 
increase by approximately 20% (median) compared to the proposal if the 
final rule required registrants to include employees of all minority-
owned subsidiaries and joint ventures.\608\ However, the effect of 
allowing registrants to exclude employees of unconsolidated 
subsidiaries on compliance costs relative to the proposed rule is not 
clear from this estimate. In light of this uncertainty and because we 
do not have other data available on the effects on the compliance cost 
estimate of the exclusion of employees of unconsolidated subsidiaries 
versus the exclusion of ``minority-owned subsidiaries and joint 
ventures,'' we are not reducing our initial cost estimates to account 
for the change to only include employees of consolidated subsidiaries. 
We acknowledge that our estimates may therefore overstate the 
compliance cost for companies with unconsolidated subsidiaries. 
Although we are unable to estimate the magnitude of this cost savings, 
as noted above, and consistent with the views of commenters, we believe 
that it could be significant for some companies.
---------------------------------------------------------------------------

    \608\ See letter from Corporate Secretaries. See also letter 
from COEC I.
---------------------------------------------------------------------------

iv. Employees Located Outside the United States
    As discussed above, a number of commenters asserted that non-U.S. 
employees should be included in the final rule.\609\ Some who supported 
this view argued that the failure to include foreign workers would 
substantially affect the pay ratio disclosure.\610\ On the other hand, 
many commenters indicated that including those employees would lead to 
significantly higher costs and suggested that the final rule allow 
registrants to use only their U.S. employees when identifying the 
median

[[Page 50167]]

employee.\611\ One commenter indicated that costs would be 20 to 30 
times higher, or hundreds of thousands of dollars higher if non-U.S. 
employees are included.\612\ Other commenters asserted that costs would 
decrease by over 50% \613\ or 90% \614\ if non-U.S. employees are 
excluded. Another commenter indicated that nearly half of respondents 
to its survey expected a U.S.-employee -only ratio to reduce compliance 
costs by more than 20 percent, while 29 percent of respondents expected 
it would reduce compliance costs by more than 40 percent.\615\
---------------------------------------------------------------------------

    \609\ See, e.g., letters from AFL-CIO I, AFSCME, Bricklayers 
International, CalPERS, Calvert, Chicago Teachers Fund, Corayer, CT 
State Treasurer, Cummings Foundation, CUPE, Estep, Fedewa, First 
Affirmative, Gould, ICCR, IL Bricklayers and Craftworkers Union, 
Marco Consulting, Matteson, McMorgan Co., Sen. Menendez et al. I, 
Novara Tesija, NY Bricklayers and Craftworkers Union, NY State 
Comptroller, Oxfam, Pax World Funds, C. Phillips, Public Citizen I, 
Rand, S. Spofford, Socially Responsive Financial Advisors, 
Teamsters, Trillium I, Trustee Campbell, US SIF, and Walden.
    \610\ See, e.g., letters from AFR, B[acirc]tirente et al., CII, 
Domini, and FS FTQ.
    \611\ See, e.g., letters from ABA, American Benefits Council, 
Business Roundtable I, Business Roundtable II, Chamber I, Corporate 
Secretaries, ExxonMobil, Frederick W. Cook & Co., RILA, Semtech, and 
SHRM.
    \612\ See letter from American Benefits Council.
    \613\ See letter from Business Roundtable I.
    \614\ See, e.g., letters from ExxonMobil and FEI.
    \615\ See letter from WorldatWork I.
---------------------------------------------------------------------------

    Comment letters addressing costs associated with including non-U.S. 
employees often noted that multinational registrants have multiple 
payroll systems and databases for their employees' compensation that 
are difficult if not impossible to reconcile.\616\ One commenter 
indicated that it has 15 payroll systems that are not integrated, and 
those payroll systems would have to be manually reconciled with 
``substantial costs'' and ``extensive staff hours.'' \617\ Another 
commenter stated that it used 30 payroll systems that are not 
connected.\618\ Yet another commenter cited a Human Resource Policy 
Association survey indicating that 84% of respondents could not easily 
calculate worldwide enterprise cash compensation for all their 
employees.\619\
---------------------------------------------------------------------------

    \616\ See, e.g., letters from American Benefits Council, Aon 
Hewitt, BCIMC, Business Roundtable I, COEC I, Cummins, Eaton, 
ExxonMobil, Freeport-McMoRan, MVC Associates, NIRI, Semtech, SHRM, 
and Tesoro.
    \617\ See letter from Freeport-McMoRan. Similarly, a different 
commenter stated that it does not have a single payroll system that 
can easily analyze the type of data required for the calculation. 
See letter from Tesoro.
    \618\ See letter from Cummins Inc.
    \619\ See letter from MVC Associates (citing to COEC I).
---------------------------------------------------------------------------

    The final rule does not allow registrants to exclude their non-U.S. 
employees when identifying the median employee, other than the limited 
exceptions described below. One commenter \620\ estimated that the 
average company in a survey it conducted had 40% of its workforce 
located outside the United States and recommended a principles-based 
approach that would permit registrants to exclude up to and exceeding 
40% of their employee population. This commenter estimated that 
permitting registrants to exclude non-U.S. employees would reduce 
compliance costs by 47%. Another commenter estimated that the median 
decrease in the compliance costs for registrants with foreign 
operations would be approximately 50% if the final rule excluded non-
U.S. employees.\621\ Given our estimate of aggregate initial compliance 
costs for registrants with foreign operations of approximately $1,050 
million, had we instead excluded non-U.S. employees, the survey results 
suggest such registrants' initial compliance costs would instead be 
$525 million.
---------------------------------------------------------------------------

    \620\ See letter from COEC III.
    \621\ See letter from Corporate Secretaries.
---------------------------------------------------------------------------

v. Foreign Data Privacy Law Exemption
    The final rule also provides a foreign data privacy exemption that 
gives registrants the ability to exclude from their median employee 
computation non-U.S. employees in jurisdictions in which data privacy 
laws or regulations prohibit the use or transfer of the necessary 
information required to comply with the final rule. According to one 
commenter's survey, 45.8% of respondents anticipated ``being prohibited 
or limited by non-U.S. data privacy laws'' in their efforts ``to access 
information necessary to collect data to identify the median employee 
or make the pay ratio calculation.'' \622\ The foreign data privacy 
exemption may lower the costs of calculating the pay ratio for 
registrants with employees in such jurisdictions, although we do not 
have data from which to estimate the magnitude of the cost savings. We 
recognize that it may also affect the median employee compensation 
determination. For example, based on the latest available Bureau of 
Economic Analysis (BEA) data for 2012, U.S. multinational companies 
were estimated to have approximately 11.5% of their employees at 
foreign affiliates in the EU, 3.2% in Canada, 4.2% in China, 3.7% in 
Mexico, 1.4% in Japan, and 2.6% combined in Switzerland, Australia, 
Argentina, Russia, and Singapore. Lower estimates are obtained when 
only majority-owned foreign affiliates are considered. Each of these 
jurisdictions was identified by commenters as having laws that may 
prohibit or restrict the transfer of information necessary to make the 
pay ratio calculation.\623\ To the extent that data privacy 
restrictions may be present both in high-income and low-income 
jurisdictions, the direction of the effect of the exemption on the pay 
ratio is ambiguous and may vary from registrant to registrant. To the 
extent that registrants with non-U.S. workers in jurisdictions with 
data privacy laws would have experienced a significantly higher cost of 
calculating the pay ratio than registrants with the same percentage of 
non-U.S. workers but in jurisdictions without such laws, this change 
from the proposed rule mitigates the potential adverse competitive 
effects of the pay ratio disclosure requirement on registrants with 
non-U.S. workers in jurisdictions with data privacy laws. Consistent 
with a commenter's suggestion,\624\ the final rule requires registrants 
to obtain a legal opinion from counsel in that jurisdiction on the 
inability of the registrant to obtain or process the information 
necessary for compliance with the final rule without violating that 
jurisdictions' laws or regulations governing data privacy, including 
the registrant's inability to obtain an exemption or other relief under 
any governing laws or regulations. The legal opinion must be

[[Page 50168]]

filed as an exhibit with the filing in which the pay ratio disclosure 
is included.
---------------------------------------------------------------------------

    \622\ See letter from COEC I.
    \623\ See, e.g., letters from ABA (``The best known of these 
data privacy regimes is the European Union's Directive 95/46/EC of 
the European Parliament and the Council of 24 October 1995 on the 
Protection of Individuals with Regard to the Processing of Personal 
Data and the Free Movement of Such Data, 1995 O.J. L 281 . . . . 
Other jurisdictions, including Argentina, Canada, Japan, and 
Switzerland, have also adopted strong data privacy laws.''), 
American Benefits Council (``Among other things, the EU data privacy 
regime prohibits `the transfer of personal data to a third country 
which does not ensure an adequate level of protection.' We also 
understand that many other countries, including China, Japan, 
Mexico, Canada, Peru, and Singapore, have or are in the process of 
implementing similar data privacy rules.''), Business Roundtable I 
(``Finally, countries outside of the EU, including Japan and 
Singapore, either already have developed domestic data privacy 
regimes similar to the EU Directive or are in the process of doing 
so.''), Business Roundtable II, COEC I (stating that, ``in addition 
to the 27 jurisdictions which have implemented the EU Directive, our 
survey respondents noted that there are several other countries 
which have restrictive data privacy laws including China, Japan and 
Mexico''), Corporate Secretaries (indicating that ``there are 27 
countries in the EU that have implemented the EU Privacy Law. . 
.other countries such as Japan and Singapore have developed or are 
in the process of developing domestic data privacy regulations''), 
NAM I (``More specifically, compliance with the data protection laws 
of each European Union member country, as well as data protection 
laws of Australia, will be a significant obstacle to collection of 
necessary information. . . . Indeed, a Manufacturer operating in 
Russia found that, according to that nation's data privacy laws, the 
company will need to get the personal sign-off from every Russian 
employee to share the data with the corporate headquarters.''), and 
WorldatWork I (``Aside from the EU's laws, there are other countries 
that have confidentiality laws which may impact this information 
gathering, such as Argentina's confidentiality laws concerning 
equity awards.'').
    \624\ See letter from AFL-CIO II.
---------------------------------------------------------------------------

    The exemption could potentially provide a competitive advantage to 
registrants with a significant overall percentage of the workforce 
located in jurisdictions with data privacy laws over other registrants 
due to lower compliance costs associated with having fewer workers 
covered by the rule. However, the limited and tailored nature of the 
exemption, as well as the reduction or elimination of the de minimis 
exemption for registrants that exclude employees under the foreign data 
privacy exemption, as discussed below, mitigates this possibility.
vi. De Minimis Exemption
    While we define the term ``employee'' to include any U.S. and non-
U.S. employee of a registrant, the final rule provides for a de minimis 
exemption for employees in foreign countries, up to 5% of a 
registrant's workforce, under certain conditions.\625\ This type of 
exemption was suggested by several commenters,\626\ and it should 
provide cost savings to eligible registrants. The suggested de minimis 
amount varied significantly across commenters.\627\
---------------------------------------------------------------------------

    \625\ See Section II.B.1.c.iii.
    \626\ See letters from American Benefits Council, ExxonMobil, 
FSI, FSR, NYC Bar, and PNC Financial Services.
    \627\ See, e.g., letters from American Benefits Council 
(suggesting that a registrant be permitted to exclude non-U.S. 
employees in any foreign jurisdiction that comprises less than 5% of 
the issuer's aggregate global workforce), ExxonMobil (indicating 
that a registrant be permitted to exclude non-U.S. employees in a 
foreign jurisdiction if the number of employees in that jurisdiction 
is less than 1% of the issuer's total workforce), FSR (recommending 
that non-U.S. employees be excluded if they account for less than 5% 
of the registrant's total workforce or employees in any single 
foreign jurisdiction if they comprise less than 2% of total 
employees with an aggregate cap of 5% (if the registrant's non-U.S. 
employees account for more than 5% of all employees)), and NACCO 
(suggesting that a registrant be permitted to exclude non-U.S. 
employees if they make up less than 20% of the employee population).
---------------------------------------------------------------------------

    In the June 4 Memorandum and June 30 Memorandum, staff attempted to 
quantify the effects of the exemption. However, because staff lacked 
more specific information about potentially affected registrants, 
including comprehensive data on the intra-company distribution of 
compensation of these categories of employees at companies that may be 
subject to the rule, the analyses necessarily relied on certain 
assumptions. Commenters also did not provide this data.
    The projections in the two staff memoranda were based on evidence 
obtained from other studies, aggregate statistics, and other 
assumptions that may result in over- or underestimating the magnitude 
of the effect on the pay ratio calculation. The memoranda made the 
following assumptions: companies have excluded the percent of employees 
equal to the specified percentage threshold; the distribution of pay is 
described by a lognormal distribution \628\ (with various estimates of 
the standard deviation of the log of pay \629\ that broadly incorporate 
the ranges of estimates from the studies cited in the June 4 
Memorandum, 0.25, 0.35, 0.45, and 0.55 \630\); and the level of PEO pay 
is independent of the exclusion threshold. The estimates of the effect 
on the pay ratio calculation of excluding different percentages of 
employees were sensitive to the above assumptions.\631\
---------------------------------------------------------------------------

    \628\ A commenter noted that a lognormal distribution may be 
inadequate for actual firms. See letter from Public Citizen II. As 
we noted in the Proposing Release, registrants that have multiple 
business or geographical segments may not necessarily have a 
lognormal distribution of pay. However, a distributional assumption 
is necessary for the analysis because staff could not observe the 
actual distribution of wages within the affected firms in the data 
available to them. This assumption is motivated by the positive 
skewness in dollar wages and the distribution of log of wages 
approximating normal distribution. See, e.g., Blundell, R., Reed, 
H., Stoker, T., 2003, Interpreting aggregate wage growth: The role 
of labor market participation, American Economic Review, Vol. 93(4), 
pp. 1114-1131; Measuring the distribution of wages in the United 
States from 1996 through 2010 using the Occupational Employment 
Survey, BLS Monthly Labor Review, May 2014, http://www.bls.gov/opub/mlr/2014/article/measuring-the-distribution-of-wages-in-the-united-states-from-1996-through-2010-using-the-occupational-employment-survey-1.htm). The assumptions used in this analysis are for 
illustration purposes only. As we note, while the lognormal 
assumption may be appropriate for some registrants, it may not be 
appropriate for all registrants. While one commenter suggested that 
the final rule permit a registrant to determine the median employee 
based on an assumption that compensation is lognormally distributed 
within a company or segment, we believe that registrants can and 
thus should determine for themselves whether the use of a lognormal 
assumption is appropriate given their own compensation 
distributions. See letter from TCA. The final rule does not specify 
any required methodology for registrants to use in identifying the 
median employee and permits registrants the flexibility to choose a 
method to identify the median employee based on their own facts and 
circumstances so long as a registrant's methodology uses reasonable 
estimates. Indeed, more generally, we believe that it is appropriate 
for registrants to make their own determinations about whether a 
particular methodological assumption constitutes a reasonable 
estimate for their particular firms.
    \629\ As we noted in the Proposing Release, each registrant 
would have a company-specific compensation variance, which is 
impossible to be generally assumed.
    \630\ Two commenters noted that these assumptions may understate 
the variability of employee pay within actual publicly-traded 
companies, particularly companies with part-time, seasonal or 
temporary employees. See letters from AFL-CIO II and Public Citizen 
II. As discussed in the June 4 memorandum, the above standard 
deviation assumptions could understate intra-firm wage variation in 
employee pay, which would in turn potentially understate the effects 
of the exclusion on the pay ratio. However, the staff lacked data on 
the actual intra-firm distribution of wages for registrants affected 
by the final rule to perform additional analysis.
    \631\ The letter from COEC III cites additional research on wage 
dispersion within and between firms not cited in the June 4 
Memorandum. See Jae Song, David J. Price, Fatih Guvenen, Nicholas 
Bloom, and Till von Wachter, Firming Up Inequality, NBER Working 
Paper 21199, May 2015, available at http://www.nber.org/papers/w21199 (``Song et al. (2015)''). The estimates of within-firm wage 
variation in this paper are in a format from which the staff cannot 
directly infer the standard deviation estimates required for its 
analysis, and the staff lacks access to the source data to compute 
these standard deviation estimates. The paper concludes that within-
firm wage inequality changed little over time, which is broadly 
similar to the conclusion in Barth et al. (2014) based on a 
different dataset, cited in the June 4 Memorandum. The staff 
analysis utilizes a range of standard deviation assumptions to 
illustrate the effects of various potential levels of within-firm 
wage variability.
---------------------------------------------------------------------------

    The June 4 Memorandum, under the assumptions above evaluated the 
effects on the pay ratio calculation of excluding different percentages 
(between 1% and 20%) of pay observations from a lognormal distribution 
for each set of assumptions about intra-company standard deviation of 
the log of pay ([sigma]) and for each of the two scenarios below 
concerning excluded pay observations: Scenario I (all excluded 
observations are below the median for the underlying distribution of 
pay); and Scenario II (all excluded observations are above the median 
for the underlying distribution of pay). Under these scenarios, for a 
given standard deviation level, the effect on the pay ratio is larger 
in magnitude when a larger percentage of employees are excluded. For 
example, the exclusion of 5% of employees may cause the pay ratio to 
decrease by up to 3.4% in Scenario I or to increase by up to 3.5% in 
Scenario II (an aggregate range of 6.9%). Under a 20% threshold, the 
pay ratio may decrease by up to 13% or increase by up to 15% (an 
aggregate range of 28%), depending on the scenario considered.
    The June 30 Memorandum extended the analysis contained in the June 
4 Memorandum by showing, under the same assumptions, the potential 
effects of excluding percentages greater than 20% and up to 95%. As 
expected, under the same assumptions, excluding a broader range of 
exclusion thresholds (between 20% and 95%) yielded a larger magnitude 
of the effect on the pay ratio for Scenarios I and II. In addition, the 
June 30 Memorandum included different intermediate scenarios between 
Scenarios I and II, with some observations excluded from above the 
median and some from below the median of the underlying

[[Page 50169]]

distribution.\632\ As expected, under the same assumptions, including 
these intermediate scenarios yielded effects within the range 
delineated by Scenarios I and II.\633\
---------------------------------------------------------------------------

    \632\ Specifically, the June 30 Memorandum included the 
following three scenarios: Scenario I(a) (75% of the excluded 
observations are below the median, with the remaining 25% of the 
excluded observations above the median); Scenario I(b) (50% of the 
excluded observations are below the median, with the remaining 50% 
of the excluded observations above the median); and Scenario I(c) 
(75% of the excluded observations are above the median, with the 
remaining 25% of the excluded observations below the median). See 
June 30 Memorandum. One commenter noted that non-random exclusion of 
low-paid employees would cause the pay ratio estimate to decline 
more than random exclusion. See letter from Public Citizen II. 
Consistent with the commenter, as the staff analysis demonstrates, 
exclusion scenarios in which over half of the excluded workers are 
paid below the true median cause the pay ratio estimate to decline. 
As noted in the June 4 Memorandum, non-U.S. employees of U.S. 
multinational firms outside the United States on average receive 
lower compensation than employees located inside the United States. 
See June 4 Memorandum. However, for some firms with employees 
outside the United States in highly skilled occupations or firms 
with employees in jurisdictions with high labor costs, some 
employees outside the United States may receive higher compensation 
than U.S. employees.
    \633\ We note that, if observations are equally excluded from 
either side of the median, the estimated effect is zero regardless 
of the percentage of employees excluded from the distribution.
---------------------------------------------------------------------------

    As the memoranda indicate, under the assumptions considered, 
excluding 5% of employees yields an effect on the pay ratio in the 
range between -3.4% and 3.5%.\634\
---------------------------------------------------------------------------

    \634\ As mentioned above, we lack information about the actual 
intra-firm distribution of pay for affected registrants but, even if 
the true distribution is lognormal, we do not have information to 
know where the effect may fall within the range.
---------------------------------------------------------------------------

    We further recognize that the estimates of the effects of the de 
minimis exemption on the pay ratio are sensitive to the assumptions 
made and may understate or overstate the actual magnitude of the effect 
if any of the above assumptions, for instance, the assumptions about 
lognormal distribution or magnitude of intra-firm variation in wages, 
do not hold. If the affected registrant's true intra-firm distribution 
of the log of employee pay is not normal, depending on the shape of the 
true distribution, the actual effects on the median may significantly 
differ from the estimated effects reported in the memoranda. 
Importantly, if the true intra-firm distribution of pay at an affected 
registrant deviates from the lognormal assumption, estimates of the 
effects under these scenarios may correspondingly decrease in accuracy 
as the percentage of the excluded observations increases.
    We note that the de minimis exemption may not affect some 
registrants because not all registrants will be eligible to use it or 
choose to use it to exclude up to 5% of the total workforce. We also 
note that, in some instances, this exemption may result in the 
exclusion of employees from jurisdictions with low pay, which may 
increase the difficulty of interpreting the pay ratio. The requirements 
to disclose the jurisdiction(s) and the approximate number of employees 
from each jurisdiction being excluded should mitigate this concern.
    We have considered several reasonable alternatives to the final 
rule's 5% de minimis exemption. One alternative would be to apply a 
different de minimis threshold. A lower de minimis percentage may 
increase registrants' costs of calculating the ratio for workers in 
non-U.S. jurisdictions. Lowering the de minimis threshold below 5% 
would not meaningfully reduce the impact on the pay ratio under the 
assumptions in our analysis. A higher de minimis threshold could yield 
potentially larger savings in the costs of calculating the ratio for 
registrants with workers in non-U.S. jurisdictions. However, as seen in 
Table 1 in the June 4 Memorandum, such a threshold would have a 
potentially larger effect on the pay ratio than the 5% threshold. 
Specifically, as discussed above, under the assumptions made and 
depending on the scenario considered, the exclusion of 10% of employees 
may decrease the pay ratio by up to 6.7% or increase it by up to 7.2%; 
the exclusion of 15% of employees may decrease the pay ratio by up to 
9.9% or increase it by up to 11%; the exclusion of 20% of employees may 
decrease the pay ratio by up to 13% or increase it by up to 15%. Each 
of these alternatives thus could result in an impact on pay ratio that 
is greater than the impact under the de minimis threshold.\635\
---------------------------------------------------------------------------

    \635\ See Section II.B.1.c.iii(c) for the definition of the de 
minimis threshold at 5%.
---------------------------------------------------------------------------

    Under alternative de minimis exemptions suggested by commenters, 
registrants would be permitted to exclude non-U.S. employees in every 
foreign country that comprises less than 1% \636\ or less than 5% \637\ 
of the registrant's aggregate global workforce. These alternative 
definitions of the de minimis exemption can reduce calculation costs 
for U.S. multinational registrants with a high level of international 
diversification in their workforce. However, they may potentially 
result in the exclusion of a large percentage of employees at 
registrants with a large percentage of non-U.S. employees diversified 
across countries, which may affect the pay ratio considerably, as we 
indicate in the discussion above. These alternatives may also offer a 
larger relative competitive advantage to internationally diversified 
U.S. registrants compared to U.S. registrants with the same total 
percentage of non-U.S. employees concentrated in fewer countries and 
thus ineligible for the exemption under these alternatives.
---------------------------------------------------------------------------

    \636\ See letter from ExxonMobil.
    \637\ See letter from American Benefits Council.
---------------------------------------------------------------------------

    A different alternative exemption proposed by a commenter would 
permit registrants to exclude all employees in any single foreign 
jurisdiction if they comprise less than 2% of total employees, with an 
aggregate cap of 5% (if the registrant's foreign employees account for 
more than 5% of all employees).\638\ The exemption in our final rule is 
defined more broadly than this alternative definition and enables 
savings in calculation costs for a potentially larger fraction of 
registrants with non-U.S. workers.
---------------------------------------------------------------------------

    \638\ See letter from FSR.
---------------------------------------------------------------------------

    Registrants that are eligible for the de minimis exemption and 
choose to use it may have lower compliance costs than registrants that 
do not use the exemption.\639\ By excluding foreign workers, the 
exemption makes eligible registrants with foreign operations more 
similar to registrants with U.S.-based operations only and reduces the 
number of employees considered in the identification of the median 
employee, with the effect of reducing compliance costs for the eligible 
registrants. Relying on some reasonable assumptions, we are able to 
quantify some of the cost savings from the de minimis exemption. First, 
we assume that all registrants with foreign operations will use the 
exemption and eliminate 5% of their workforce. We also assume that the 
savings in compliance costs are directly proportionate to the number of 
employees excluded and that the compliance cost per excluded foreign 
employee is equal to the estimate for firms with some foreign 
employees, $38.04. Using these assumptions and our estimates of the 
number of registrants with foreign operations and the total number of 
employees for these registrants from Table 3, the total savings from 
the use of the de minimis

[[Page 50170]]

exemption would be approximately $52.5 million.\640\
---------------------------------------------------------------------------

    \639\ Some commenters noted that the staff analysis did not 
incorporate a cost-benefit analysis. See letters from COEC III and 
WorldatWork II. Potential cost savings from the de minimis exemption 
are discussed in this section. See Section III.D.2.c.iv, above, for 
a discussion of potential cost savings from the exclusion of all 
non-U.S. employees. Commenters did not provide more detailed 
estimates of the cost savings from excluding some but not all non-
U.S. employees.
    \640\ The savings are calculated as 5% * (27,595,305 * $38.04) 
for firms with some foreign employees. Under similar assumptions, if 
companies with foreign operations were permitted to exclude 10%, 
20%, 30%, and 40% of their employees, respectively, potential cost 
savings could amount to that percentage multiplied by (27,595,305 * 
$38.04) or $105 million, $210 million, $315 million, and $420 
million, respectively. Actual cost savings may differ if there are 
fixed costs associated with the inclusion of non-U.S. employees in 
the pay ratio calculation. Another commenter estimated from a survey 
that the mid-range percentage of non-U.S. employees was 40% and that 
the mid-range cost savings from excluding all non-U.S. employees 
would be 50%. See letter from COEC.
---------------------------------------------------------------------------

    We note that actual cost savings incrementally attributed to the de 
minimis exemption are likely to be lower. Some registrants may have 
less than 5% of the total workforce outside the United States. Other 
registrants may have more than 5% of the total workforce outside the 
United States but less than 5% of the total workforce in aggregate 
across foreign jurisdictions in which it can exclude employees under 
the de minimis exemption (with no more than 5% of the total workforce 
in each such jurisdiction). The incremental cost savings from the de 
minimis exemption are further reduced or potentially eliminated for 
registrants that exclude some employees under the foreign data privacy 
exemption as such registrants may be ineligible for the de minimis 
exemption or eligible for the exemption but unable to exclude employees 
at the maximum level under the de minimis exemption due to the 
concurrent use of the foreign data privacy exemption, concentrated 
nature of their non-U.S. workforce, or a combination of the two 
factors.\641\ Other registrants with non-U.S. workers may elect not to 
use the exemption. We also note that the actual cost savings could vary 
significantly depending on whether cost savings increase uniformly with 
the percent of employees excluded and whether registrants can exclude 
any employees outside the United States under the foreign data privacy 
law exemption.
---------------------------------------------------------------------------

    \641\ The final rule does not limit a registrant's ability to 
rely on the foreign data privacy exemption, provided the conditions 
of the exemption are met.
---------------------------------------------------------------------------

    To the extent that registrants with non-U.S. workers experience a 
higher cost of calculating the pay ratio than registrants with U.S. 
workers only and that the de minimis exemption reduces such costs, it 
reduces the potential adverse competitive effects of the pay ratio 
disclosure requirement on registrants with non-U.S. workers eligible 
for the exemption. Registrants with non-U.S. workers concentrated in 
fewer foreign jurisdictions are more likely to exceed the 5% threshold 
for any single foreign jurisdiction and thus be ineligible for the 
exemption than registrants with the same total percentage of non-U.S. 
workers diversified across more foreign jurisdictions. If the inability 
to use the de minimis exemption increases the costs of calculating the 
pay ratio significantly (for instance, when the overall percentage of 
non-U.S. workers is higher than but relatively close to 5%), 
registrants with geographically concentrated non-U.S. workers may be at 
a relative competitive disadvantage.
vii. Calculation Date
    The final rule permits registrants to choose any date within three 
months of the end of registrant's fiscal year to identify the median 
employee for that year and requires registrants to disclose the date 
used. Compared with prescribing a given date, such as the last day of 
the completed fiscal year, as proposed, this approach may reduce 
compliance costs by providing flexibility to registrants that may not 
have enough time to collect and report on their pay ratio information 
at year-end. Commenters suggested that allowing registrants to select 
the date would allow them to pick a date that does not coincide with 
other required reporting or that better utilizes the internal resources 
of the registrants.\642\ This approach also might reduce compliance 
costs for registrants that use many employees at the end of the 
calendar year by permitting those registrants to choose a date on which 
those seasonal or temporary employees are not employed.\643\ 
Registrants in industries with more fluctuations in employment within 
the year due to seasonality may realize larger benefits from this 
approach.\644\ Hence, it is possible that registrants could choose a 
date and structure their employment arrangements around that date to 
reduce the number of workers employed on the calculation date or to 
alter the reported ratio to achieve a particular objective with the pay 
ratio disclosure. One commenter specifically addressed this issue and 
noted its belief that this concern ``is unwarranted, particularly if 
the choice is restricted to a limited time period (such as the last 
fiscal quarter), since in general the employee population of a 
registrant would not vary significantly over such a period.'' \645\ 
Another commenter suggested that not allowing such an adjustment will 
produce ``artificially low'' median employee pay for registrants that 
have many temporary and seasonal workers at year-end.\646\ Yet another 
commenter recommended allowing flexibility with respect to the 
measurement date but requiring that the calculation include annual 
compensation of all employees employed at any time over the preceding 
365 days.\647\ However, we note that such an alternative would increase 
the cost for registrants with significant fluctuations in the number of 
employees within the year relative to the final rule. Based on the 
comments we received, we believe that the rule as adopted will reduce 
compliance costs compared to the proposed rule. However, we did not 
receive data that would allow us to quantify the cost reduction.
---------------------------------------------------------------------------

    \642\ See, e.g., letters from Chamber I, COEC I, COEC II, and 
Microsoft.
    \643\ See, e.g., letters from American Benefits Council, NACCO, 
and PM&P.
    \644\ For example, analysis by staff in the Division of Economic 
and Risk Analysis of BLS employment statistics by industry for 2014 
revealed that the following sectors appear to have the largest 
employment fluctuations due to seasonality, as proxied by the 
average magnitude of the percentage difference between seasonally 
adjusted and not seasonally adjusted employment: arts, entertainment 
and recreation, educational services, and construction, as well as 
administrative and support services, accommodation and food services 
(when data for all months is used) and retail trade and warehousing 
and storage (when data for the final three months is used, to 
account for the fiscal year end being December for the majority of 
registrants). See http://www.bls.gov/data/#employment. Agricultural 
employment is excluded from the above calculations. It is also 
likely to be subject to significant fluctuations within the year due 
to seasonality. We note that BLS data may not be representative of 
the employment fluctuations within the year for the registrants 
subject to the final rule and that the value of the flexibility to 
select the calculation date will vary across registrants.
    \645\ See letter from Davis Polk.
    \646\ See letter from PM&P.
    \647\ See letter from AFL-CIO II.
---------------------------------------------------------------------------

d. Adjustments to the Compensation of Employees
    The final rule includes an instruction that permits a registrant to 
annualize the compensation for all permanent employees (full-time or 
part-time) employed on the calculation date who did not work for the 
registrant for the full fiscal year. The final rule does not permit 
annualization for employees in temporary or seasonal positions. The 
final rule also does not permit the use of full-time-equivalent 
adjustments for any of a registrant's employees in the required pay 
ratio disclosure, although such adjustments are permitted to derive an 
additional ratio if the registrant chooses. We believe that our 
approach provides appropriate accommodations to registrants to 
represent the annual composition of

[[Page 50171]]

their workforce without significantly diminishing the potential 
usefulness of the disclosure mandated by Section 953(b).
    We believe that by permitting annualization adjustments for 
permanent employees but not seasonal or temporary ones, our approach 
more closely captures the composition of a registrant's workforce and 
compensation practices. For these annualizing adjustments to have any 
significant impact on the reported pay ratio, both the fraction of 
permanent new hires to all employees of the registrant and their 
annualized compensation would have to be relatively large. We also note 
that some commenters were supportive of allowing annualizing 
adjustments.\648\ One commenter suggested that this adjustment will 
make the ratio more representative of the registrant's labor 
arrangements.\649\ This procedure is purely optional and registrants do 
not need to annualize compensation, such as if they believe that the 
additional cost of the adjustment does not warrant the perceived 
benefit.
---------------------------------------------------------------------------

    \648\ See, e.g., letters from ABA, Corporate Secretaries, Intel, 
NACCO, PM&P, SH&P, and WorldatWork I.
    \649\ See letter from Corporate Secretaries (``We believe that 
allowing annualizing adjustments effectuates the Dodd-Frank Act 
requirement that each registrant disclose the median of the annual 
total compensation of its employees. Such adjustments would result 
in a calculation that is closer to a fair and reasonable 
representation of the registrant's actual compensation practices and 
labor costs. The adjustment also would help eliminate the potential 
distorting effects of mid-year hires, including those that result 
from a merger or acquisition'').
---------------------------------------------------------------------------

    By permitting registrants to annualize compensation for these 
employees, the comparability of disclosure across registrants could be 
reduced compared to an alternative of either requiring or prohibiting 
such annualization. As noted above, however, we believe that precise 
comparability of disclosure from registrant to registrant could be 
difficult to achieve due to the variety of factors that could cause the 
ratio to differ. Accordingly, we do not believe that the costs 
associated with promoting precise comparability would be justified.
    Another alternative would have been to permit a registrant to 
annualize the compensation for all temporary and seasonal employees who 
were employed for less than the full fiscal year and to use that 
annualized compensation in the mandated pay ratio disclosure. Some 
commenters supported this alternative and indicated that not allowing 
an annualizing adjustment for these employees would distort the pay 
ratio.\650\ Some commenters urged us to permit the use of full-time 
equivalent adjustments for part-time employees, temporary, and seasonal 
employees.\651\ The final rule does not permit the mandated pay ratio 
disclosure to include either the use of annualized adjustments for 
seasonal or temporary employees or the use of full-time-equivalent 
adjustments for part-time employees. We believe that such adjustments 
would reflect a different workforce composition and compensation 
structure than that utilized by the registrant. To the extent that a 
registrant relies primarily on part-time, temporary, or seasonal 
workers, computing a ratio based on their annualized compensation of 
temporary and seasonal workers or the full-time-equivalent of a part-
time worker, unlike annualizing adjustments for permanent employees, 
could have a significant impact on the ratio.
---------------------------------------------------------------------------

    \650\ See, e.g., letters from AAFA II, American Benefits 
Council, Brian Foley & Co., Corporate Secretaries, Hay Group, KBR, 
NIRI, and NYC Bar.
    \651\ See supra note 108.
---------------------------------------------------------------------------

    Although we are not permitting full-time-equivalent adjustments or 
annualization adjustments for seasonal and temporary employees to be 
made for purposes of calculating the annual total compensation in the 
mandated pay ratio disclosure, the final rule does permit registrants 
to provide additional disclosure. For example, registrants can report 
additional ratios, including ratios that reflect one or more of those 
adjustments, if they choose, provided that any additional ratio is 
clearly identified, not misleading, and not presented with greater 
prominence than the required ratio.
    The final rule permits but does not require registrants to adjust 
compensation to the cost of living in the PEO's jurisdiction of 
residence. While some commenters stated that international differences 
in the cost of living can distort the reported pay ratio,\652\ one 
commenter suggested that compensation is more directly comparable to 
the total compensation of a registrant's PEO without a cost-of-living 
adjustment.\653\ Moreover, as some commenters suggested,\654\ the cost-
of-living adjustment may alleviate concerns about pay comparability 
between U.S. employees and those non-U.S. employees that are located in 
foreign countries where the purchasing power of the average foreign pay 
in dollar terms deviates significantly from the purchasing power of the 
average domestic pay in dollar terms.
---------------------------------------------------------------------------

    \652\ See, e.g., letters from AAFA I, American Benefits Council, 
Corporate Secretaries, and ExxonMobil.
    \653\ See letter from ABA.
    \654\ See letters from NAM I, NAM II, and SH&P.
---------------------------------------------------------------------------

    Providing the option to use a cost-of-living adjustment is not 
expected to increase the compliance cost for registrants. Country-level 
cost-of-living data is widely available. The incremental cost of 
identifying the median employee based on pay without the cost-of-living 
adjustment and calculating the pay ratio without the cost-of-living 
adjustment is expected to be small once pay data for all employees or 
for samples of employees from individual countries have been obtained. 
Thus, registrants that believe the cost-of-living adjusted ratio to be 
more meaningful given the structure of their workforce may benefit from 
the option to present the pay ratio with the cost-of-living adjustment 
as the ratio required by Item 402(u)(1)(iii).
    The cost-of-living adjustment of the compensation of a registrant's 
employees may have an effect on the determination of the median 
employee and on the calculation of the pay ratio for registrants with 
employees in countries whose cost of living differs from the cost of 
living in the PEO's country of residence. We are limited in our ability 
to quantify the impact of this adjustment on the pay ratio calculation 
by our lack of data on the intra-firm distribution of pay of employees 
outside the PEO's country of residence for the affected registrants and 
by limited data available to us on the distribution of employees by 
country at the individual registrant level. As noted elsewhere, because 
we lack data regarding intra-firm distributions, we cannot predict the 
effects of a cost-of-living adjustment on those distributions, as the 
adjustment may, in some cases, have an effect on the combined employee 
pay distribution at the individual registrant level by potentially 
changing the median employee within the same country or by locating the 
median employee in a different country. We therefore analyze 
qualitatively the main factors that may contribute to more significant 
effects of the cost-of-living adjustment on the determination of the 
median employee and on the calculation of the pay ratio.
    The cost-of-living adjustment option could affect the pay ratio 
calculation for registrants with some employees located outside the 
PEO's country of residence that elect to use this option. The effect of 
the cost-of-living adjustment could be potentially larger for 
registrants with a larger percentage of employees outside the PEO's 
country of residence and for registrants with employees in countries 
with a cost of living that differs significantly from the PEO's country 
of residence.

[[Page 50172]]

    According to the results of one commenter's survey, the average 
(median) respondent had 62% (60%) of its employees located in the 
United States.\655\ According to another commenter's survey, 
approximately 55% of respondents reported having the majority of their 
employees located in the United States.\656\ As set forth in Table 3 
above, out of the registrants potentially subject to the final rule for 
which we have data on the presence of geographic segments,\657\ 
approximately a third are estimated to have U.S.-based operations only 
and two-thirds are estimated to have some non-U.S. operations. Based on 
aggregate BEA data on U.S. multinational companies for 2012,\658\ 
employees at foreign affiliates and at majority-owned foreign 
affiliates of U.S. multinational companies comprised approximately 38% 
and 33-34%, respectively, of the total employment of U.S. multinational 
companies. We note that the respondents in the surveys cited above, 
companies that report geographic segments, and companies in the BEA 
sample may not be representative of the full set of registrants subject 
to the final rule, and the PEO's country of residence may be different 
than the United States.
---------------------------------------------------------------------------

    \655\ See letter from Corporate Secretaries.
    \656\ See letter from COEC I.
    \657\ Compustat segments database reports data on the presence 
of geographic segments for approximately 63% of registrants 
potentially subject to the final rule.
    \658\ See U.S. BEA data on Direct Investment & Multinational 
Enterprises (MNEs) available at http://bea.gov/iTable/index_MNC.cfm.
---------------------------------------------------------------------------

    Based on aggregate BEA data on the distribution of employees of 
U.S. multinational companies by country for 2012,\659\ the majority of 
employees at foreign affiliates of U.S. multinational companies were 
estimated to be located in the following regions: Europe (34%, 
including 30% in the EU); Asia and Pacific (34%); and Latin America 
(20%). Looking at individual jurisdictions, the most employees of 
foreign affiliates of U.S. multinational companies were estimated to be 
located in the following ten countries: China (11.2%); the United 
Kingdom (10.3%); Mexico (9.8%); Canada (8.4%); India (6.9%); Germany 
(4.9%); Brazil (4.6%); Japan (3.7%); France (3.5%); and Australia 
(2.5%), which in aggregate accounted for 65.8% of foreign employees of 
U.S. multinational firms.\660\
---------------------------------------------------------------------------

    \659\ Id.
    \660\ Id.
---------------------------------------------------------------------------

    However, these aggregate statistics on the location of employees by 
country do not capture the distribution of employees by countries at 
individual registrants. While these aggregate statistics may offer an 
average perspective across all firms with a non-U.S. workforce in the 
BEA sample, they do not enable us to draw strong conclusions about the 
ultimate effects on the pay ratio for those registrants that are 
subject to final rule and decide to opt for the cost-of-living 
adjustment. We believe that registrants anticipating an increase in the 
pay ratio after the adjustment may be less likely to opt for the cost-
of-living adjustment. Based on 2014 data from the International 
Monetary Fund (``IMF''),\661\ we note that the E.U. on aggregate had a 
cost of living similar to the U.S. level. Based on 2014 data on PPP 
conversion factors from the World Bank,\662\ of the above locations, 
the United Kingdom, Canada, Germany, Japan, France, and Australia had a 
cost of living similar to or above the U.S. level. Based on the same 
measure, of the above locations, India, China, Mexico, and Brazil had a 
cost of living below the U.S. level.\663\ As we noted above, the actual 
effects of cost-of-living adjustment on the pay ratio calculation will 
depend on the countries where employees are located, the actual 
distribution of employee pay, and the specific cost-of-living measure 
used.
---------------------------------------------------------------------------

    \661\ We lack PPP estimates for the E.U. in aggregate. An 
indirect proxy is the PPP conversion factor implied by the E.U. 
gross domestic product (``GDP'') reported in dollar terms and PPP-
adjusted dollar terms. See IMF World Economic Outlook data by 
country groups, available at http://www.imf.org/external/pubs/ft/weo/2015/01/weodata/weoselagr.aspx.
    \662\ PPP conversion factor is the ratio of the PPP exchange 
rate to the nominal dollar exchange rate, available at http://data.worldbank.org/indicator/PA.NUS.PPPC.RF. This ratio makes it 
possible to compare the cost of the bundle of goods that make up GDP 
across countries. It measures how many dollars are needed to buy a 
dollar's worth of goods in the country as compared to the United 
States.
    \663\ Id. PPP conversion factors calculated based on the 2011 
International Comparison Program round and 2014 market exchange 
rates for the countries above are: China (0.6); the United Kingdom 
(1.2); Mexico (0.6); Canada (1.1); India (0.3); Germany (1.0); 
Brazil (0.7); Japan (1.0); France (1.1); and Australia (1.4).
---------------------------------------------------------------------------

    Below we illustrate the potential effect of the cost-of-living 
adjustment on the pay ratio for a hypothetical registrant.\664\ In this 
example, we make the following assumptions: We assume that the 
registrant has 70% of employees in the country of residence of the PEO 
and 30% of employees in another country; the pay of the registrant's 
employees in the PEO's country, expressed in the currency of the PEO's 
country, is lognormally distributed (with mean log of pay of 10.5 and 
standard deviation of log of pay of 0.5); the pay of the registrant's 
employees in the other country, expressed in the currency of the PEO's 
country, is lognormally distributed (with mean log of pay of 9 and 
standard deviation of log of pay of 0.5); and the cost of living is two 
times higher in the country of residence of the PEO than in the other 
country. In this hypothetical example, a cost-of-living adjustment 
would cause the pay ratio to decrease by approximately 6.4%. If the 
cost of living were three times higher in the country of residence of 
the PEO than in the other country, holding other assumptions unchanged, 
a cost-of-living adjustment would cause the pay ratio to decrease by 
approximately 14.9%.
---------------------------------------------------------------------------

    \664\ This is intended only as a hypothetical example for 
illustration purposes and not as an indication of the effects at an 
actual or representative registrant. The effects are highly 
sensitive to the assumptions described. The estimates are obtained 
numerically.
---------------------------------------------------------------------------

    Some commenters \665\ suggested that a cost-of-living adjustment 
could introduce an element of subjectivity into the pay ratio 
calculation or permit registrants to alter the reported ratio to 
achieve a particular objective with the ratio disclosure. In the final 
rule, the requirements to apply a consistent methodology, to disclose 
the use of the adjustment, and to provide disclosure of the 
registrant's pay ratio calculated without the cost-of-living adjustment 
should address these concerns.
---------------------------------------------------------------------------

    \665\ See letters from Prof. Ray and WorldatWork I.
---------------------------------------------------------------------------

e. Frequency of Identifying the Median Employee
    Unlike the proposed rule, which required registrants to identify 
the median employee every year, the final rule allows registrants to 
identify the median employee once every three years unless there has 
been a change in the registrant's employee population or employee 
compensation arrangements that it reasonably believes would result in a 
significant change in the pay ratio disclosure. Registrants must still 
provide annual disclosure of their pay ratio by recalculating the 
previously identified median employee's annual total compensation each 
year.
    Under this approach, a registrant may identify its median employee 
for year one and then use that employee or one who has substantially 
similar compensation as its median employee in the following two years 
for calculating the employee's annual total compensation and the 
registrant's pay ratio. A couple of commenters suggested this approach, 
noting that it would still result in a registrant providing a pay ratio 
disclosure on an annual basis while reducing the burden and costs 
required to identify the median employee annually when there have not 
been any interim changes in the

[[Page 50173]]

registrant's workforce or compensation structure.\666\
---------------------------------------------------------------------------

    \666\ See, e.g., letters from ABA, COEC I, and COEC II.
---------------------------------------------------------------------------

    Choosing this approach is likely to result in lower ongoing 
compliance costs for affected registrants. We expect that some 
registrants will identify the median employee every three years, while 
others may identify it every year or every two years. Thus, depending 
on how frequently registrants would have to identify the median 
employee, and on whether the identification of the median employee is 
the main ongoing cost of compliance, with the change in the final 
release, the ongoing compliance costs could range approximately from 
$123 million to $947 million per year.\667\
---------------------------------------------------------------------------

    \667\ Ongoing costs without this adjustment are estimated to be 
between $368 million and $947 million. See Section III.C.2.c. 
Ongoing costs with the adjustment are estimated as follows. The $123 
million is estimated as $368 million divided by three, based on the 
lowest of the estimates of ongoing compliance costs and the 
assumptions that all registrants identify the median employee every 
three years and the identification of the median employee accounts 
for the entirety of the ongoing cost. This estimate represents the 
aggregate annual cost in this scenario averaged over three years. 
The $947 million is based on the highest of the estimates of ongoing 
compliance costs and the assumption that all registrants incur the 
full ongoing cost every year.
---------------------------------------------------------------------------

f. Method of Identifying the Median Employee
    In order to allow the greatest degree of flexibility while 
maintaining consistency with the statutory provision, the final rule 
does not specify a particular methodology for identifying the median. 
Instead, it allows registrants a choice of multiple methods, including 
several with significant flexibility.
    We are adopting this flexible approach because we believe that the 
appropriate and most cost-effective methodology for identifying the 
median employee necessarily depends on a registrant's particular facts 
and circumstances, including, among others, such variables as size and 
nature of the workforce, complexity of the organization, the 
stratification of pay levels across the workforce, the types of 
compensation the employees receive, the extent that different 
currencies are involved, the number of tax and accounting regimes 
involved, the number of payroll systems the registrant has, and the 
degree of difficulty involved in integrating payroll systems to readily 
compile total compensation information for all employees. We believe 
that these are likely the same factors that would cause substantial 
variation in the costs of compliance. By not prescribing specific 
methodologies that must be used, the final rule allows registrants to 
choose a method to identify the median employee that is appropriate to 
the size, structure, and compensation practices of their own 
businesses, including permitting a registrant to identify the median 
employee using any consistently applied compensation measure.
    In addition, the final rule's flexibility could enable registrants 
to manage compliance costs more effectively than a more prescriptive 
approach would allow.\668\ We also believe that, by allowing 
registrants to minimize direct compliance costs, a flexible approach 
could mitigate, to some extent, any potential negative effects of the 
mandated requirements on competition. We recognize, however, that a 
flexible approach could increase uncertainty for registrants that 
prefer more specificity on how to comply with the final rule, 
particularly for registrants that do not use statistical analyses in 
the ordinary course of managing their businesses. In light of this 
potential uncertainty, the final rule establishes certain parameters on 
the use of this flexibility, such as by specifying that the use of 
statistical sampling or other reasonable estimates in identifying the 
median is permitted, as is identifying the median employee based on any 
consistently applied compensation measure.
---------------------------------------------------------------------------

    \668\ For example, one commenter estimated that using all 
elements of compensation rather than the sum of salary earned, 
incentive cash earned, and stock awards granted to identify the 
median employee in a worldwide workforce would increase the initial 
expense and ongoing workload by a factor of more than five times. 
See letter from Microsoft.
---------------------------------------------------------------------------

    We believe that a flexible approach would not significantly 
diminish the potential benefits of the disclosure mandated by Section 
953(b). As discussed above, we believe that the intended purpose of the 
pay ratio disclosure is to provide shareholders with a company-specific 
metric to evaluate the PEO's compensation, rather than a benchmark for 
compensation arrangements across registrants. Also as discussed above, 
we are not persuaded that mandating a particular methodology will 
necessarily improve the comparability of pay ratio disclosure across 
registrants because of the numerous other factors that could also cause 
the ratios to be less meaningful for registrant-to-registrant 
comparison. Even if such comparability could be marginally enhanced by 
mandating a specific method for identifying the median, we do not 
believe this marginal improvement in comparability would be justified 
in light of the costs that would be imposed on registrants by a more 
prescriptive rule. We also note that some commenters expressed the view 
that greater comparability across registrants could increase the 
likelihood that a registrant's competitors could infer proprietary or 
sensitive information about the registrant's business. This in turn 
could increase the indirect costs to registrants of the adopted 
requirements, such as competitive harms in labor markets discussed in 
the previous section or general costs arising from the mandated 
disclosure requirement.
    Finally, we recognize that allowing registrants to select a 
methodology to identify the median, rather than prescribing a 
methodology or set of methodologies, could reduce the benefits for 
shareholders if that flexibility results in a pay ratio statistic that 
is less useful than a more precisely and consistently calculated ratio. 
In particular, some commenters claimed that permitting flexibility in 
the rule would allow registrants to manipulate the ratio in their 
favor.\669\ While we acknowledge that the flexibility we are providing 
creates some risk that registrants will attempt to use this flexibility 
to produce a more favorable pay ratio, we think that this risk is 
mitigated by the disclosures we are requiring with respect to the 
methodologies and assumptions used to identify the median employee, as 
discussed below. The final rule specifically discusses two particular 
permitted methods of identifying the median employee--using a 
consistently applied compensation measure and using statistical 
sampling. For all of these reasons, we believe the benefits of the 
final rule's flexibility outweigh those of a more prescriptive 
approach.
---------------------------------------------------------------------------

    \669\ See, e.g., letters from Bupp, Corayer, Fedewa, Fox, 
Friend, Grotzke, Hlodnicki, Kizzort, Maly, Petricoin, and Van Pelt.
---------------------------------------------------------------------------

i. Consistently Applied Compensation Measure
    We proposed to allow registrants to use any consistently applied 
compensation measure, such as amounts derived from the registrant's 
payroll or tax records, to identify the median employee and then 
calculate that median employee's annual total compensation in 
accordance with Item 402(c)(2)(x). We are adopting this approach as 
proposed.
    Allowing registrants this flexibility is likely to reduce 
registrants' compliance costs significantly, compared to the 
alternative of requiring registrants to calculate total compensation in 
accordance with Item 402(c)(2)(x) for all employees, or for a 
statistically valid sample, and then identify the median.

[[Page 50174]]

This view was shared by commenters.\670\ Registrants that choose this 
approach will be able to identify a median employee from employee 
compensation data that they may already track or record or that may be 
less expensive for them to acquire than obtaining and computing all of 
the Item 402(c)(2)(x) compensation information for each employee. Using 
one commenter's survey results,\671\ we can estimate the potential 
savings resulting from our exercise of discretion. According to the 
survey, which provided both average and median cost increase estimates, 
costs would increase on average by 4,689% if registrants were required 
to calculate total compensation in accordance with Item 402(c)(2)(x) 
for all employees. The median increase in compliance costs would be 
175%.\672\ Applying this median percentage to our initial compliance 
cost estimate of $1,315 million, this alternative would increase the 
total compliance costs to approximately $3,616 million. Thus, we 
estimate that allowing registrants flexibility in identifying the 
median employee could result in a total savings of approximately $2,301 
million.
---------------------------------------------------------------------------

    \670\ See, e.g., letters from CalPERS (``By offering companies a 
number of alternatives, companies will be able to determine which 
methodology works best for their company and/or tailor it for their 
special circumstances. Moreover, this flexibility will allow 
companies to select a methodology that is most cost effective for 
them. Finally, since companies will already have the dataset 
necessary for this calculation (in order to prepare their financial 
statements and tax returns), we do not envisage costs will be a 
barrier to compliance.'') and COEC I (``The value of this 
flexibility appears to be significant in terms of cost. Indeed, 
survey respondents were asked how their compliance costs would be 
affected in the event they were required to calculate median 
employee compensation using the same method employed in the `Summary 
Compensation Table.' This method is currently used to calculate 
total primary executive officer compensation and is equivalent to 
the first approach offered by the SEC. 99 percent of respondents 
answered that their costs would increase if they were forced to 
calculate median employee compensation using the Summary 
Compensation Table approach. Including all responses, the median 
increase is reported as 100 percent. The data strongly indicate that 
adhering to the Summary Compensation Table approach would lead to 
additional significant increases in compliance costs relative to the 
Proposed Rule'').
    \671\ See letter from Corporate Secretaries.
    \672\ Another commenter, based on survey results, argued that 
the average increase in costs would be 4,592%, but did not provide a 
median estimate. See letter from COEC I.
---------------------------------------------------------------------------

    We acknowledge, however, that some registrants will still incur 
costs if they have to combine or sample from separately maintained 
payroll systems across segments and/or geographic locations.
ii. Statistical Sampling
    The final rule, as proposed, also allows registrants to use 
statistical sampling in their determination of the median employee. The 
size of the reduction in compliance costs that can be achieved by using 
statistical sampling or other reasonable estimates in identifying the 
median employee ultimately depends on a registrant's particular facts 
and circumstances. Below we provide an illustration of how various 
registrants' characteristics might affect the sampling size. We note 
that these numbers are intended to provide examples and should not be 
treated as recommendations about the appropriate sampling size. For 
example, in the following figure and tables, we show that the variance 
of underlying wage distributions can materially affect the appropriate 
sample size for statistical sampling.\673\ Industries characterized by 
the BLS as having low wage variances, such as electric power 
generation, coal mining, and metal ore mining, have estimated minimum 
appropriate sample sizes for an accurate median estimate of less than 
135 employees. In contrast, industries characterized by high wage 
variances, such as offices of physicians, health and personal care 
stores, and spectator sports, have estimated minimum appropriate sample 
sizes of more than 1,263 employees. Figure 2 shows the distribution of 
estimated minimum appropriate sample sizes for registrants operating in 
each of the 290 4-digit NAICS industries tracked by the BLS.\674\
---------------------------------------------------------------------------

    \673\ The analysis uses average and median wage estimates from 
the BLS at the 4-digit NAICS industry level (290 industries) and 
assumes a lognormal wage distribution. We use a 95% confidence 
interval with relative 0.5% margin of error in the estimate of the 
average of the logarithm of wage. We estimate the median wage by 
taking the exponential of the sample average of the logarithm of 
wage, which is the sample geometric average. This median estimator 
is the maximum likelihood estimator (``MLE'') of the population 
median for lognormal distribution, and it is an unbiased estimator 
when the sample size is large. The 95% confidence interval for the 
population wage median can be obtained by taking the exponential of 
the endpoints of the 95% confidence interval for the sample average 
of logarithm of wage (E.L. Crow and K. Simizu, Lognormal 
Distributions: Theory and Applications 29, (Marcel-Dekker: New York, 
1988), R. Serfling, Efficient and Robust Fitting of Lognormal 
Distributions, North American Actuarial Journal 6, 95-109 (2002), G. 
Casella & R. L. Berger, Statistical Inference 320 (Duxbury, 2nd ed. 
2002), T. B. Parkin, and J. A. Robinson, Statistical Evaluation of 
Median Estimators for Lognormally Distributed Variable, Soil Science 
Society of America Journal 57, 317-323 (1993). The lognormal wage 
distribution assumption is supported by the following studies: F. 
Clementi, and M. Gallegati, Pareto's Law of Income Distribution: 
Evidence for Germany, the United Kingdom, and the United States. 
Econophysics of Wealth Distributions, New Economic Window. 3-14 
(2005), and J. L[oacute]pez and L. Serv[eacute]n, A Normal 
Relationship? Poverty, Growth and Inequality. World Bank Policy 
Research Working Paper 3814 (2006). See also M. Pinkovskiy and X. 
Sala-i-Martin, Parametric Estimations of the World Distribution of 
Income, NBER working paper 15433, (2009). It is common in practice 
to control the relative margin of error (instead of absolute margin 
of error) to determine the sample size. Accordingly, the sample size 
depends on the coefficient of variation (CV) of the underlying 
distribution. The square of the CV, also known as ``relative 
variance,'' is often more stable and easier to guess in advance than 
variance (W. G. Cochran, Sampling Techniques 77, (New York: Wiley, 
3rd ed. 1977); S. L. Lohr, Sampling: Design and Analysis 46-47, 
(Cengage Learning, 2nd ed. 2009). This analysis also assumes that 
when the sampling is implemented, the sampling method would be a 
true random sampling (i.e., it would not be biased by region, 
occupation, rank, or other factor).
    \674\ Our analysis excludes the Motor Vehicle Manufacturing and 
Postal Service industries from our sample because, for these 
industries, mean wage was lower than median wage in dollar terms, 
appearing to contradict our lognormal distributional assumption.

                 Table 6--The Industries With the Smallest and Largest Appropriate Sample Sizes
----------------------------------------------------------------------------------------------------------------
                                                                                                    Example of
                            Industry                               Average wage     Median wage     registrant
                                                                        ($)             ($)         sample size
----------------------------------------------------------------------------------------------------------------
                            10 Industries With Smallest Variance in Wage Distribution
----------------------------------------------------------------------------------------------------------------
Electric Power Generation, Transmission and Distribution........          72,800          70,380              84
Coal Mining.....................................................          55,740          53,310             116
Metal Ore Mining................................................          58,000          55,060             135
Software Publishers.............................................          96,730          90,390             160
Wired Telecommunications Carriers...............................          65,580          61,510             162
Natural Gas Distribution........................................          74,270          69,350             170
Rail Transportation.............................................          59,990          56,120             172
Computer and Peripheral Equipment Manufacturing.................          94,850          88,160             174

[[Page 50175]]

 
Interurban and Rural Bus Transportation.........................          35,950          33,690             184
School and Employee Bus Transportation..........................          33,440          31,200             199
----------------------------------------------------------------------------------------------------------------
                            10 Industries With Largest Variance in Wage Distribution
----------------------------------------------------------------------------------------------------------------
Offices of Physicians...........................................          72,040          40,510           1,572
Health and Personal Care Stores.................................          41,890          27,060           1,290
Spectator Sports................................................          42,540          27,680           1,263
Agents and Managers for Artists, Athletes, Entertainers, and              71,960          45,100           1,251
 Other Public Figures...........................................
Motion Picture and Video Industries.............................          56,540          36,420           1,226
Home Health Care Services.......................................          37,780          25,100           1,225
Cut and Sew Apparel Manufacturing...............................          34,800          23,580           1,180
Amusement Parks and Arcades.....................................          29,580          20,800           1,095
Apparel, Piece Goods, and Notions Merchant Wholesalers..........          52,350          35,800           1,063
Other Professional, Scientific, and Technical Services..........          48,140          33,150           1,059
----------------------------------------------------------------------------------------------------------------

                                                                                                  [GRAPHIC] [TIFF OMITTED] TR18AU15.004
                                                                                                  
    Because these estimated minimum appropriate sample sizes are based 
on wage distributions measured by the BLS in standardized industries, 
they may not correspond to the appropriate minimum sample size at 
registrants with an employee base that does not correspond precisely to 
one of these industries. Even for registrants whose operations are 
wholly within one of these standardized industries, their appropriate 
sample size may also be different to the extent that their distribution 
of employee wages is different than that of the industry. In these 
instances, a registrant's appropriate sample size could be higher or 
lower than that estimated for its industry.
    In 2014, of the nearly 3,571 registrants that we believe will be 
subject to the final rule, we estimate that approximately 68% and 63% 
report business and geographic segments, respectively. Approximately 
50% and 65% of the potentially affected registrants that self-report 
business and geographic segments, respectively, report a single segment 
of that type.\675\ Of the registrants that self-report a single 
business segment for which we have industry classifications that match 
the BLS data, Table 7 shows estimated minimum appropriate sample sizes 
assuming that each registrant's wage

[[Page 50176]]

distribution is similar to the BLS-measured industry distribution.
---------------------------------------------------------------------------

    \675\ This estimate is based on data from Standard and Poor's 
Compustat Segment database. We note that the segment information is 
self-reported by the companies, so it is not based on standardized 
definitions of geographic areas such as states, countries, or 
regions.

     Table 7--Number of Registrants According to Sample Size Ranges
------------------------------------------------------------------------
                                                             Number of
                  Sample size(n) ranges                     registrants
------------------------------------------------------------------------
n<100...................................................              30
100<=n<250..............................................             113
250<=n<500..............................................             260
500<=n<750..............................................             664
750<=n<1000.............................................             127
n>=1000.................................................              20
                                                         ---------------
    Total...............................................           1,214
------------------------------------------------------------------------

    The example in Table 7 is simplified by the assumptions that 
registrants have employees in a single industry and that employee pay 
is described by a lognormal distribution with parameters based on 
aggregate statistics for that industry. We recognize that statistical 
sampling may be more complicated for registrants with different types 
of pay distributions or multiple business and geographic segments, each 
of which may have different parameters of the distribution. While one 
commenter suggested simple random sampling could be used for these 
registrants,\676\ other approaches, such as stratified cluster 
sampling,\677\ may yield more efficient estimates in some instances. We 
also recognize that the implementation of statistical sampling for 
registrants with multiple payroll systems may require additional steps. 
While we believe that statistical sampling can produce reasonable 
estimates of the median for these types of registrants, we lack 
information on intra-firm employee pay distributions and registrants' 
costs of sampling to estimate the proportion of registrants for which 
specific sampling approaches may most efficiently produce reasonable 
estimates of median pay.
---------------------------------------------------------------------------

    \676\ See letter from Ohlrogge II.
    \677\ See, e.g., S. Gross. Median estimation in sample surveys. 
In Proceedings of the Section on Survey Research Methods. American 
Statistical Association, 181-184. (1980).
---------------------------------------------------------------------------

    While some commenters argued that statistical sampling will not 
lead to significant reductions in compliance costs,\678\ the majority 
of commenters supported using statistical sampling for calculating the 
median employee, implying that this approach can reduce costs for some 
registrants.\679\ In light of the comments received, we continue to 
believe that permitting registrants to use statistical sampling will 
lead to an overall reduction in compliance costs as compared to not 
permitting this method of identifying the median.
---------------------------------------------------------------------------

    \678\ See, e.g., letters from Business Roundtable I, Chamber I, 
COEC I, COEC II, ExxonMobil, Freeport-McMoRan, FuelCell Energy, 
Garmin, Johnson & Johnson, Microsoft, NAM I, NAM II, NIRI, and 
WorldatWork I.
---------------------------------------------------------------------------

g. Disclosure of Methodology, Assumptions, and Estimates
    The final rule requires registrants to briefly describe and 
consistently apply any methodology used to identify the median employee 
and disclose any material assumptions, adjustments, or estimates used 
to identify the median or to determine total compensation or any 
elements of total compensation. Registrants also must clearly identify 
any estimates used. Registrants' disclosure of the methodology and 
material assumptions, adjustments, and estimates used must be designed 
to provide information for a reader to be able to evaluate the 
appropriateness of the methodologies used.
    This disclosure is intended to aid shareholders and investors in 
their use of the pay ratio disclosure and alert them to any material 
changes in the methodology that might change the reported pay ratio. 
Many commenters indicated that the rule should require registrants to 
provide this narrative information,\680\ but a number of these 
commenters indicated that the rule should clarify that the narrative be 
brief.\681\
---------------------------------------------------------------------------

    \680\ See, e.g., letters from AFL-CIO I, Barnard, Business 
Roundtable I, CalPERS, CalSTRS, Calvert, CII, COEC I, COEC II, CT 
State Treasurer, Domini, Hermes, Kasner, LAPFF, Meridian, Microsoft, 
Somers, Sze, UAW Trust, US SIF, and WorldatWork I.
    \681\ See, e.g., letters from ABA, AFL-CIO I, Business 
Roundtable I, COEC I, COEC II, Domini, Meridian, Microsoft, and 
WorldatWork I.
---------------------------------------------------------------------------

    Alternatively, we could have required registrants to provide a 
detailed description of every operational step and methodological 
assumption used in identifying the median employee. Because we are 
concerned that disclosure about methodology, assumptions, adjustments, 
and estimates could become dense and overly technical,\682\ which we 
believe would limit its usefulness, the final rule asks for a brief 
overview and makes clear that it is not necessary to provide technical 
analyses or formulas. We do not believe that a detailed, technical 
discussion (such as statistical formulas, confidence levels, or the 
steps used in data analysis) would appreciably enhance shareholders' 
understanding of how the pay ratio was calculated. We recognize, as 
commenters noted, that registrants will incur some costs in developing 
and reviewing the appropriate language to describe the approach taken. 
However, we expect that the costs of this disclosure will be marginal, 
as these additional disclosures are intended to simply describe what 
has already been done or assumed in the calculations, and therefore 
will not require additional analysis by registrants.
---------------------------------------------------------------------------

    \682\ See, e.g., letter from Business Roundtable I (commenting 
that the disclosure could not be brief because of the issuer's need 
to use many estimates and assumptions).
---------------------------------------------------------------------------

h. Determination of Total Compensation
    As mandated by Section 953(b), the final rule defines ``total 
compensation'' by reference to Item 402(c)(2)(x). We received comments 
supporting the use of estimates in calculating the annual total 
compensation or any elements of total compensation for employees other 
than the PEO.\683\ As proposed, the final rule permits registrants to 
use reasonable estimates to determine elements of ``total 
compensation.''
---------------------------------------------------------------------------

    \683\ See, e.g., letters from ABA, COEC I, COEC II, Davis Polk, 
Prof. Angel, and Vectren Corp.
---------------------------------------------------------------------------

    We acknowledge that, to the extent that the use of estimates causes 
the disclosure to present a less precise measure of the ``total 
compensation'' of the registrant's median employee than if we 
prohibited the use of estimates, it could diminish the potential 
usefulness of the disclosure. However, commenters did not suggest that 
allowing for the use of reasonable estimates in determining the ``total 
compensation'' would diminish the potential usefulness of the 
disclosure and we likewise believe it is not likely to have such an 
effect.\684\
---------------------------------------------------------------------------

    \684\ But see letter from Dennis T (``No estimates. [N]o 
sampling. We demand the actual data.'').
---------------------------------------------------------------------------

i. Defining ``Annual''
    As proposed, the final rule defines ``annual total compensation'' 
to mean total compensation for the last completed fiscal year, 
consistent with the time period used for the other Item 402 disclosure 
requirements.
    Some commenters agreed that the pay ratio disclosure should be 
calculated based on data from the last completed fiscal year.\685\ 
Other commenters, however, recommended that the rule permit registrants 
to use another period, such as the fiscal year preceding the 
registrant's last completed fiscal year.\686\ Commenters also asked 
that registrants be permitted to choose the period.\687\ We understand 
that these suggestions are intended to reduce compliance costs for 
registrants by giving registrants extra time to comply with the rule or 
the ability to use information in the form

[[Page 50177]]

that would best suit their particular facts and circumstances. We 
believe, however, that it is appropriate for the time period used for 
the pay ratio disclosure to be the same as the time period used for the 
registrant's other executive compensation disclosures, although the 
flexibility in identifying the median employee could help to address 
the concerns raised by these commenters. In particular, using the same 
time period as for other executive compensation disclosure will avoid 
any possible confusion for shareholders using this disclosure.
---------------------------------------------------------------------------

    \685\ See, e.g., letters from ABA, CalPERs, and UAW Trust.
    \686\ See, e.g., letters from Aon Hewitt, Business Roundtable I, 
Corporate Secretaries, and Eaton.
    \687\ See, e.g., letters from Davis Polk and WorldatWork I.
---------------------------------------------------------------------------

j. Updating the Pay Ratio Disclosure for the Last Completed Fiscal Year
    The final rule includes instructions to clarify the timing for 
updating pay ratio disclosure after the end of a registrant's fiscal 
year. Without this provision, a registrant could be required to include 
pay ratio disclosure in a filing after the end of the fiscal year, but 
before it has compiled the executive compensation information for that 
fiscal year for inclusion in its proxy statement relating to its annual 
meeting of shareholders. This could impose additional costs on 
registrants that elect to provide executive compensation disclosure in 
their annual proxy statement rather than in their annual report and for 
registrants that are conducting registered offerings at the beginning 
of their fiscal year.
    To address this concern, we considered the recommendation of 
commenters that pay ratio disclosure not be required to be updated for 
the most recently completed fiscal year until the registrant files its 
proxy statement for its annual meeting of shareholders. The final rule 
generally follows this recommended approach and also provides a similar 
accommodation for registrants that do not file annual proxy 
statements.\688\ It also aligns the final rule to the filing deadlines 
for providing Item 402 disclosure in annual reports and proxy and 
information statements. We believe that such an approach will reduce 
costs to registrants without diminishing the potential usefulness of 
the disclosure.
---------------------------------------------------------------------------

    \688\ Based on a review of EDGAR filings in calendar year 2013, 
approximately 250 registrants that would be subject to the final 
rule do not file proxy or information statements in connection with 
annual meetings of shareholders, including 15D filers (other than 
SRCs and ABS issuers) and registrants that are not corporate 
entities required to hold annual meetings of shareholders.
---------------------------------------------------------------------------

    We also believe that this approach could reduce costs for 
registrants in connection with filings made or required to be made 
before the filing of the proxy or information statement for the annual 
meeting of shareholders (or written consents in lieu of such a meeting) 
that would typically contain the registrant's other Item 402 disclosure 
covering the most recently completed fiscal year. In addition, under 
the final rule, updating the pay ratio disclosure is not an additional 
impediment for a registrant that requests effectiveness of a 
registration statement after the end of its fiscal year and before the 
filing of the proxy statement for its annual meeting of shareholders. 
In this regard, this approach could alleviate some of the final rule's 
potential impact on capital formation.
k. Status of Disclosure as ``Filed''
    Under the final rule, the pay ratio disclosure will be considered 
``filed'' for purposes of the Securities Act and Exchange Act, like 
other Item 402 information. A number of commenters recommended that the 
pay ratio disclosure be ``furnished'' rather than ``filed'' \689\ 
because registrants will have to review a large amount of data and make 
a significant number of estimates, assumptions, and judgment calls, 
which will necessarily lead to imprecision.\690\ This, in turn, could 
subject registrants to potential liability and litigation,\691\ make it 
difficult to validate the information sufficiently for Sarbanes-Oxley 
Act certification purposes,\692\ and/or not permit the information to 
be audited (or greatly increase the costs of the audits).\693\ We 
recognize that some registrants could have more difficulty in gathering 
and verifying the information than others. We believe, however, that 
the flexibility afforded to registrants in connection with identifying 
the median employee could reduce some of the difficulties of compiling 
the required information because registrants will be able to tailor the 
methodology to reflect their own facts and circumstances. In addition, 
we believe that the final rule's transition periods, which are 
discussed below, could mitigate some concerns about compiling and 
verifying the information because they are designed to give registrants 
sufficient time to develop and implement compliance procedures.
---------------------------------------------------------------------------

    \689\ See, e.g., letters from AAFA II, ABA, American Benefits 
Council, Aon Hewitt, Best Buy et al., Bill Barrett Corp., Business 
Roundtable I, Business Roundtable II, Chamber I, Chesapeake 
Utilities, COEC I, COEC II, Corporate Secretaries, Eaton, Freeport-
McMoRan, General Mills, Intel, Mercer I, NAM I, NAM II, NIRI, NRF, 
PM&P, RILA, SHRM, and Vectren Corp.
    \690\ See, e.g., letters from AAFA II, ABA, Business Roundtable 
I, Chesapeake Utilities, COEC I, COEC II, Corporate Secretaries, 
Eaton, General Mills, NRF, RILA, and Vectren Corp.
    \691\ See, e.g., letters from ABA, American Benefits Council, 
Aon Hewitt, Bill Barrett Corp., Chamber I, General Mills, Mercer I, 
and PM&P.
    \692\ See, e.g., letters from ABA, Best Buy et al., Corporate 
Secretaries, Freeport-McMoRan, Intel, NAM I, and SHRM.
    \693\ See, e.g., letters from COEC I, COEC II, Corporate 
Secretaries, and NIRI.
---------------------------------------------------------------------------

    Requiring registrants to ``file'' their pay ratio information may 
make the final rule more costly for registrants than the alternative of 
allowing them to ``furnish'' such information. Treating the pay ratio 
disclosure as ``filed'' will mean that registrants could potentially be 
subject to litigation under Section 18 of the Exchange Act, although, 
as mentioned earlier, Section 18 does not create strict liability for 
misstatements in ``filed'' information and requires that a plaintiff 
establish that it relied on the misleading information in purchasing or 
selling a security and suffered damages caused by that reliance. On the 
other hand, under the final rule, this potential liability for 
misleading pay ratio disclosure could make registrants more accountable 
for the disclosure than if we instead permitted the disclosure to be 
`furnished' and result overall in fewer inaccuracies in the required 
pay ratio disclosure. To the extent that registrants perceive there to 
be a greater likelihood of private litigation under the final rule than 
if they were permitted to ``furnish'' the information, registrants may 
decide to apply a more costly process to identify the median employee, 
or retain additional counsel, thus increasing compliance costs.
l. Compliance Date
    Section 953(b) does not specify a date when registrants must begin 
to comply with the final rule. In a change from the Proposing Release, 
the final rule requires that a registrant must begin to comply with 
Item 402(u) with respect to compensation for the registrant's first 
full fiscal year commencing on or after January 1, 2017.
    As discussed above, the change from the proposal provides calendar 
year-end filers and registrants with fiscal years beginning January 1, 
2015 until the day before the final rule's effectiveness one additional 
year to provide their pay ratio disclosure relative to the proposal. 
The final rule also changes the compliance schedule for registrants 
with fiscal years starting on or after effectiveness through December 
30, 2015. These registrants will receive two additional years to 
provide their pay ratio disclosure relative to the proposal. Assuming a 
hypothetical effective date of November 1, 2015, we estimate that this 
change will lead to a one-time cost deferral of approximately $147 
million for two years and to savings of approximately between $27.3 
million and $212 million for 223 registrants subject to the final rule 
that have fiscal

[[Page 50178]]

years that end from October 31, 2015, through December 30, 2015 \694\ 
and a one-time cost deferral of approximately $1,169 million for one 
year and to savings of approximately between $109 million and $842 
million for the remaining 3,348 registrants subject to the final 
rule.\695\ For this estimation, we identified the fiscal year ends for 
all affected registrants. We found that 223 registrants have fiscal 
years that end between October 31 and December 30, of which 130 are 
multinational registrants and 93 are registrants with U.S.-only 
segments. We found that of the remaining 3,348 registrants estimated to 
be subject to the final rule, 1,340 are multinational registrants and 
2,008 are registrants with U.S.-only segments.\696\
---------------------------------------------------------------------------

    \694\ For these registrants, the initial compliance cost is 
assumed to be deferred for two years. There are 130 multinational 
registrants and 93 registrants with U.S.-only segments. The annual 
cost for the 130 multinational registrants is equal to the product 
of their cumulative number of employees (3,354,869) and the cost per 
employee ($38.04), or a total of approximately $128 million. The 
annual cost for the 93 registrants with U.S.-only segments is equal 
to the product of their cumulative number of employees (1,018,780) 
and the cost per employee ($19.02), or a total of approximately $19 
million. Thus, the total initial compliance cost for the 223 
registrants that is deferred for two years is approximately $147 
million. These registrants also will not have to incur two years of 
ongoing compliance costs. Assuming the ongoing cost as a percentage 
of the initial cost is between one-third of 28% (9.3%) and 72% of 
the initial cost, the cost savings are estimated to be between ($147 
million*9.3%)*2 = $27.3 million and ($147 million*72%)*2 = $212 
million. See Sections III.C.2.c and III.D.2.e.
    \695\ For these registrants, the initial compliance cost is 
assumed to be deferred for one year. Of the 3,348 registrants, 1,340 
are multinational registrants with 24,240,445 employees in total and 
a total annual cost of $922 million; 1,691 are registrants with 
U.S.-only segments with 12,530,180 employees in total and a total 
annual cost of $238 million; and 317 are registrants with U.S.-only 
segments with missing employee data and a total annual cost of $9 
million. Thus, the total cost for the 3,348 registrants is 
approximately $1,169 million over one year. These registrants also 
will not have to incur one year of ongoing compliance costs. 
Assuming the ongoing cost as a percentage of the initial cost is 
between one-third of 28% (9.3%) and 72% of the initial cost, the 
cost savings are estimated to be between ($1,169 million*9.3%) = 
$109 million and ($1,169 million*72%) = $842 million. See Sections 
III.C.2.c and III.D.2.e.
    \696\ For registrants without segment or employee data, we 
followed the same approach as in Table 3.
---------------------------------------------------------------------------

m. Transition Periods
    The final rule also includes a transition period for new 
registrants because we are sensitive to the impact that the rule could 
have on capital formation. We note that the requirements of Section 
953(b), as amended by the JOBS Act, distinguish between certain newly 
public companies and all other registrants by providing an exemption 
for emerging growth companies. We also note that the incremental time 
needed to compile pay ratio disclosure could cause registrants that are 
not emerging growth companies to delay an initial public offering, 
which could have a negative impact on capital formation.\697\ In this 
regard, we expect that registrants that are not emerging growth 
companies are likely to be businesses with more extensive operations 
and a greater number of employees than many emerging growth companies, 
which could increase the initial efforts needed to comply with the 
final rule. We believe that providing a transition period for these 
newly public companies could mitigate this potential impact on capital 
formation.
---------------------------------------------------------------------------

    \697\ See, e.g., letters from ABA (``Although we doubt that a 
company considering an initial public offering of its securities 
would decide to forego such a transaction simply because of the pay 
ratio disclosure obligation, in some situations, the time and costs 
associated with Item 402(u) compliance could certainly weigh in the 
timing of the offering.''), Lou (``The competitive disadvantages 
raise the costs of raising capital through public trading markets 
and thereafter discourage companies to go public. The additional 
monetary cost obviously makes an initial public offering (`IPO') [a] 
less attractive mean[s] to raise capital though the negative impact 
might not constitute a fatal factor that would kill the IPO. However 
a CEO would feel reluctant to list the company because of the 
threatening embarrassment of pay ratio disclosure. Therefore the pay 
ratio exerts a negative effect on IPOs.''), and PM&P (``We agree 
with the proposed transition period that new registrants should not 
be required to include pay ratio disclosure in their initial 
registration statements, and that to do so could significantly delay 
the IPO.'').
---------------------------------------------------------------------------

    To address these concerns, the final rule also includes 
instructions that would permit new registrants to delay compliance, so 
that pay ratio disclosure would not be required in a registration 
statement on Form S-1 or S-11 for an initial public offering or a 
registration statement on Form 10. Instead, such a registrant would be 
required to first comply with Item 402(u) with respect to compensation 
for the first fiscal year commencing on or after the year in which the 
registrant first becomes subject to the requirements of Section 13(a) 
or Section 15(d) of the Exchange Act, but no earlier than the year 
commencing January 1, 2017. Additionally, in response to comments, the 
final rule provides that a registrant that ceases to be a smaller 
reporting company or an emerging growth company will not be required to 
provide pay ratio disclosure until the first fiscal year after exiting 
such status, but no earlier than the year commencing January 1, 2017. 
This change from the proposed rule allows registrants exiting smaller 
reporting company or an emerging growth company status to delay their 
initial compliance by a year, and will give them additional time to 
decide how they will identify their median employee and prepare the 
necessary disclosure. Further, the final rule permits registrants that 
engage in business combinations and/or acquisitions to not include in 
the median employee determination employees of a newly-acquired entity 
for the fiscal year in which the business combination or acquisition 
occurs. We believe that the exercise of discretion used in allowing 
these additional transitional periods will result in cost savings for 
the affected registrants and will further mitigate any effects of the 
rule on capital formation.

IV. Paperwork Reduction Act

A. Background

    Certain provisions of the final amendments contain ``collection of 
information'' requirements within the meaning of the Paperwork 
Reduction Act of 1995 (the ``PRA'').\698\ We published a notice 
requesting comment on the collection of information requirements in the 
Proposing Release for the rule amendments, and we submitted these 
collections of information requirements to the Office of Management and 
Budget (``OMB'') for review in accordance with the PRA.\699\ The titles 
for the collections of information are:
---------------------------------------------------------------------------

    \698\ 44 U.S.C. 3501 et seq.
    \699\ 44 U.S.C. 3507(d) and 5 CFR 1320.11.
---------------------------------------------------------------------------

     ``Regulation S-K'' (OMB Control No. 3235-0071);
     ``Form 10-K'' (OMB Control No. 3235-0063);
     ``Regulation 14A and Schedule 14A'' (OMB Control No. 3235-
0059);
     ``Regulation 14C and Schedule 14C'' (OMB Control No. 3235-
0057);
     ``Form 8-K'' (OMB Control No. 3235-0060);
     ``Form S-1'' (OMB Control No. 3235-0065);
     ``Form S-4'' (OMB Control No. 3235-0324);
     ``Form S-11'' (OMB Control No. 3235-0067);
     ``Form 10'' (OMB Control No. 3235-0064); and
     ``Form N-2'' (OMB Control No. 3235-0026).
    These regulations, schedules and forms were adopted under the 
Securities Act and the Exchange Act, and in the case of Form N-2,\700\ 
the Investment Company Act of 1940.\701\ The regulations, forms and 
schedules set forth the disclosure requirements for periodic reports, 
registration statements,

[[Page 50179]]

and proxy and information statements filed by companies to help 
investors make informed investment and voting decisions. The hours and 
costs associated with preparing, filing and sending the form or 
schedule constitute reporting and cost burdens imposed by each 
collection of information. An agency may not conduct or sponsor, and a 
person is not required to respond to, a collection of information 
unless it displays a currently valid OMB control number.
---------------------------------------------------------------------------

    \700\ 17 CFR 239.14 and 274.11a-1.
    \701\ 15 U.S.C. 80a-1 et seq.
---------------------------------------------------------------------------

    Our amendments to the forms and regulations are intended to satisfy 
the requirements of Section 953(b) of the Dodd-Frank Act, which directs 
the Commission to amend Item 402 of Regulation S-K to add the pay ratio 
disclosure requirements specified by that provision. Compliance with 
the final rule will be mandatory for affected registrants. Responses to 
the information collections will not be kept confidential, and there 
will be no mandatory retention period for the information disclosed.

B. Summary of Information Collections

    In order to satisfy the legislative mandate in Section 953(b), we 
are adopting new paragraph (u) to Item 402 of Regulation S-K. This new 
paragraph (u) will require registrants to disclose:
     the median of the annual total compensation of all 
employees of the registrant (excluding the PEO);
     the annual total compensation of the registrant's PEO; and
     the ratio between these two amounts.
    For this purpose, Section 953(b) specifies that total compensation 
is to be determined in accordance with Item 402(c)(2)(x). Item 402 
already requires registrants to disclose the annual total compensation 
of the PEO in accordance with Item 402(c)(2)(x).\702\ The median of the 
annual total compensation of all employees and the ratio are new, 
incremental disclosure burdens and will require affected registrants to 
collect compensation information for employees that is not currently 
required to be disclosed.
---------------------------------------------------------------------------

    \702\ As of the date of this release, the requirements for the 
calculation of total compensation under Item 402(c)(2)(x) are the 
same as those in effect on July 20, 2010. Therefore, for purposes of 
this PRA analysis, we have assumed that registrants would not need 
to recalculate the annual total compensation for the PEO in 
connection with the pay ratio disclosure.
---------------------------------------------------------------------------

    The additional disclosure under new paragraph (u) of Item 402 will 
be required in any annual report, proxy or information statement, or 
registration statement that requires executive compensation disclosure 
pursuant to Item 402 of Regulation S-K.\703\ In addition, the 
requirements will allow certain registrants to omit the disclosure 
otherwise required by Item 402(u) from filings made during a specified 
transition period.
---------------------------------------------------------------------------

    \703\ Consistent with the scope of Section 953(b), the new 
requirements will not apply to the annual reports and proxy and 
information statements of emerging growth companies, smaller 
reporting companies or foreign private issuers. In addition, 
consistent with the instructions J and I of Form 10-K, the new 
requirements will not apply to the annual reports of issuers of 
asset-backed securities or to wholly-owned subsidiary registrants.
---------------------------------------------------------------------------

    Finally, in order to conform the amendments to current rules for 
the disclosure of PEO compensation when certain elements of 
compensation are not yet known, we are adopting a conforming amendment 
to Item 5.02 of Form 8-K. New paragraph (1) of Item 5.02(f) will also 
require registrants that are disclosing PEO total compensation in 
accordance with Item 5.02 of Form 8-K to provide in that filing the 
updated pay ratio disclosure required by Item 402(u). Because Item 5.02 
of Form 8-K provides a delayed method of filing information that would 
otherwise be required in the registrant's proxy or information 
statement or annual report, the PRA analysis assumes that the burden 
and cost of compliance with new Item 402(u) would be associated 
primarily with those forms and schedules rather than Form 8-K.

C. Summary of Comment Letters and Revisions to Proposals

    In the Proposing Release, we requested comment on the PRA analysis. 
We received letters from two commenters that directly addressed the PRA 
estimates,\704\ as well as a number of other comment letters and 
submissions that discussed the costs and burdens to issuers that would 
have an effect on the PRA analysis.\705\ A detailed discussion of these 
comments is included in the Section III above.\706\
---------------------------------------------------------------------------

    \704\ See letters from COEC I, COEC II, and Chamber II.
    \705\ See letters from Avery Dennison, ExxonMobil, FEI and KBR.
    \706\ See Section III.C.2.b.
---------------------------------------------------------------------------

    One of the two commenters analyzed data from a survey of 118 
companies to conclude that it would take registrants an average of 952 
hours per year to comply with the pay ratio disclosure requirement at 
an average labor cost of $185,600, which we assume refers to external 
costs only.\707\ The other commenter disagreed with the our assumption 
in the Proposing Release that ongoing compliance costs in the second 
and third year would significantly decrease from the initial compliance 
costs in the first year.\708\ Based on the proposed rule, which did not 
include several accommodations adopted in the final rule, that 
commenter estimated that ongoing compliance costs would be 80% (mid-
range) or 72% (average) of the initial compliance costs for each 
successive year. This commenter also cited a survey completed by 128 
public companies, the majority of which were large companies with 
assets well over $2 billion, to assert that the average cost of outside 
securities compliance counsel is $700 per hour, rather than the $400 
per hour used in the PRA estimates.\709\
---------------------------------------------------------------------------

    \707\ See letter from Chamber II. This commenter estimated that 
the annual cost of compliance would be $710.9 million and an annual 
compliance time of 3.6 million hours. The commenter also stated that 
we may have underestimated costs by more than 870% and 
underestimated compliance time by 560%. From this information, we 
infer that the average labor cost of $185,600 refers to external 
costs, as multiplying the number of registrants estimated to be 
subject to the proposed rule (3,830) by the average labor cost 
estimated by this commenter ($185,600) equals $710,848,000.
    \708\ See letters from COEC I and COEC II.
    \709\ As noted in our Economic Analysis, we continue to believe 
that $400 per hour is the appropriate rate to use for PRA purposes. 
This is the rate we typically estimate for outside legal services 
used in connection with public company reporting and is intended to 
represent an average to cover all registrants of varying sizes. In 
addition, some commenters indicated that they would retain external 
advisors such as payroll specialists, human resource consultants, 
and compensation consultants. See letters from Avery Dennison 
(stating that it expects to retain two to three external advisors, 
including legal advisers and HR consultants) and General Mills 
(indicating that it expects to hire outside advisors, such as 
compensation consultants and payroll specialists). Generally, we 
expect the hourly fees for such external advisors to be much lower 
than those of legal counsel.
---------------------------------------------------------------------------

    Several companies submitted estimates of burdens and costs without 
commenting on the our estimates. One company estimated that compliance 
with the proposed rule would require 100-150 hours of work by internal 
staff and 20 to 40 hours of external consulting time at a total 
``internal cost'' of $1 million to $1.5 million.\710\ Another company 
estimated that it would take over 1,000 internal burden hours to 
develop the database and methodology to derive the pay ratio 
information, and that ongoing burden hours would be approximately 50% 
(500 hours) of the initial compliance burden hours.\711\ It also 
asserted that this represents an approximate cost of over $250,000 on 
an initial basis and $100,000 on an ongoing

[[Page 50180]]

basis.\712\ Another large corporate commenter asserted that it would 
require up to approximately 3,000 internal burden hours to comply with 
the proposed rule in the initial year of compliance, and that ongoing 
compliance burdens after the initial compliance year would be 
approximately 28% of the initial burden hours (850 hours per year 
thereafter).\713\ Another global issuer estimated that it may take 
between $500,000 and $1 million to establish and automate the process 
to comply with the proposed rule.\714\
---------------------------------------------------------------------------

    \710\ See letter from Avery Dennison. Although this commenter 
used the phrase ``internal cost'' of compliance, we assume that this 
cost includes more than internal staff time. Otherwise, the internal 
cost of compliance could range from approximately $6,667 to $15,000 
per hour ($1 million divided by 150 hours or $1.5 million divided by 
100 hours).
    \711\ See letter from FEI.
    \712\ Id.
    \713\ See letter from ExxonMobil.
    \714\ See letter from KBR.
---------------------------------------------------------------------------

    We are adopting the final rule as proposed with modifications that 
may help mitigate compliance costs and burdens. First, we provide two 
tailored exemptions for non-U.S. employees from the definition of 
``employee'': an exemption for circumstances in which a foreign 
jurisdiction's laws or regulations governing privacy are such that, 
despite its reasonable efforts to obtain or process the information 
necessary for compliance with the final rule, the registrant is unable 
to do so without violating such data privacy laws or regulations and a 
de minimis exemption. Second, the final rule defines ``employee'' to 
include only the employees of the registrant's consolidated 
subsidiaries, instead of all subsidiaries as proposed. Third, to 
provide consistency and flexibility, the final rule permits registrants 
to use any date within three months of the last day of their last 
completed fiscal year to identify the median employee. Fourth, the 
final rule allows registrants to identify the median employee every 
three years, instead of every year, if there has been no change in 
their employee population or employee compensation arrangements that 
they reasonably believe would result in a significant change in their 
pay ratio disclosure.

D. Revisions to PRA Reporting and Cost Burden Estimates

    For purposes of the PRA, in the Proposing Release, we estimated 
that the total annual increase in the paperwork burden for all affected 
companies to prepare the disclosure that would be required under the 
adopted amendments would be approximately 545,792 hours of company 
personnel time and a cost of approximately $72,772,200 for the services 
of outside professionals. As discussed in more detail below, we are 
revising our PRA burden and cost estimates to reflect the responses of 
commenters, as well as the modifications we have made to the final rule 
to reduce compliance burdens.
    For purposes of the PRA for the final rule, we estimate the total 
annual increase in the paperwork burden for all affected companies to 
comply with the collection of information requirements in our final 
rule is approximately 2,367,573 hours of company personnel time and 
approximately $315,390,720 for the services of outside 
professionals.\715\ These estimates include the time and the cost of 
implementing data gathering systems and disclosure controls and 
procedures, compiling necessary data, preparing and reviewing 
disclosure, filing documents and retaining records.
---------------------------------------------------------------------------

    \715\ We describe how we derived the three-year average hour and 
cost burdens per response below. The portion of the burden carried 
by outside professionals is reflected as a cost, while the portion 
of the burden carried by the company internally is reflected in 
hours. For administrative convenience, the presentation of the 
totals related to the paperwork burden hours have been rounded to 
the nearest whole number and the cost totals have been rounded to 
the nearest dollar.
---------------------------------------------------------------------------

    In deriving these estimates, we have assumed that:
     Registrants subject to the final rule would satisfy the 
new requirements by either including the information directly in annual 
reports on Form 10-K or incorporating the information by reference from 
a proxy statement on Schedule 14A or information statement on Schedule 
14C. Our estimates assume that substantially all of the burden relating 
to the new disclosure requirements would be associated with Form 10-K;
     For registrants that would be permitted to provide their 
pay ratio disclosure in a filing made in accordance with Item 5.02 of 
Form 8-K, rather than in Form 10-K, the burden relating to the new 
disclosure requirements would be associated primarily with Form 10-K 
rather than Form 8-K; \716\ and
---------------------------------------------------------------------------

    \716\ Our PRA estimates for Form 8-K include an estimated one 
hour burden to account for the inclusion of the new pay ratio 
disclosure.
---------------------------------------------------------------------------

     100% of new registrants would use the transition 
provisions allowing them to omit the required disclosure from their 
initial registration statements and, for follow-on offerings by these 
registrants, the burden relating to the new disclosure requirements 
would be associated primarily with Form 10-K rather than Forms S-1, S-
11 or N-2 as applicable (because registrants would incorporate the 
disclosure from Form 10-K).
    We understand from commenters that the burdens and costs of 
compliance will likely vary among individual companies based on a 
number of factors, including the size and complexity of their 
organizations, the nature of their operations and workforce, the 
location of their operations, and, significantly, the extent that their 
existing payroll systems collect the information necessary to identify 
the median of the annual total compensation of their employees. Because 
the final rule provides additional flexibility in identifying the 
median and the annual total compensation of employees, the actual 
burden could be lower if the methodology used is able to reduce the 
effort needed to collect the data or if the registrant is able to use 
information that it collects for other purposes.\717\ We believe that 
the actual burdens will likely vary significantly among individual 
companies based on these factors. Our estimates in this PRA analysis 
reflect average burdens, and, therefore, some companies may experience 
costs in excess of our estimates and some companies may experience 
costs that are lower than our estimates.\718\
---------------------------------------------------------------------------

    \717\ See Section II of this release for a discussion of the 
requirements.
    \718\ As in our Economic Analysis, we estimated the PRA costs 
and burdens to reflect a broad range of registrants.
---------------------------------------------------------------------------

1. Estimated Internal Burden Hours
    Commenters estimated that registrants would spend anywhere from 100 
burden hours \719\ to 3000 burden hours to prepare and review the pay 
ratio disclosure.\720\ One commenter estimated that affected companies 
would on average spend 952 hours per year to comply with the new 
disclosure requirement.\721\ This estimate was based on a survey of a 
range of companies, some with operations in more than 50 countries and 
others with operations in fewer than 10 countries, and did not take 
into account the modifications that were made to the rule to reduce 
compliance costs.
---------------------------------------------------------------------------

    \719\ See letter from Avery Dennison.
    \720\ See letter from ExxonMobil.
    \721\ See letter from Chamber II.
---------------------------------------------------------------------------

    In our analysis of the economic costs and benefits of the rule, we 
estimated that the total initial compliance costs would be 
$1,314,694,544 or approximately $368,159 per registrant.\722\ Our 
estimate did not break down the costs between internal burden hours and 
external costs, which is how the burdens and costs are described for 
PRA purposes. As discussed later in our analysis of the estimated cost 
and hour burdens for each collection of

[[Page 50181]]

information,\723\ we believe that substantially all of the burden 
relating to the new disclosure requirements will be associated with 
Form 10-K. For Exchange Act reports on Form 10-K, we estimate that 75% 
of the burden of preparation is carried by the company internally and 
that 25% of the burden of preparation is carried by outside 
professionals. Using that formula, we estimate that the average 
registrant will spend 1,105 internal burden hours preparing and 
reviewing the disclosure for the initial year of compliance.\724\
---------------------------------------------------------------------------

    \722\ See discussion in Section III.C.2.c. $1,314,694,544/3,571 
= $368,159.
    \723\ See discussion in Section IV.D.3.
    \724\ We did not receive any estimates of the cost per hour 
related to preparation of disclosures by the company internally, but 
expect that such costs will be less than the cost of hiring outside 
professionals. For ease of analysis, we assume that internal hourly 
costs will be approximately half the cost of hiring outside 
professionals ($400/2 = $200). Assuming 75% of burden hours are 
carried internally and 25% are carried externally, the average 
compliance cost of $368,159 per registrant corresponds to $368,159/
((.75)$200 + (.25)$400) = 1,473 hours, of which 1,105 hours 
(1,473(.75)) are internal and 368 hours (1,473(.25)) are external.
---------------------------------------------------------------------------

    In the Proposing Release, we estimated that the internal burden 
hours would be greatest during the first year of compliance with the 
rule and would diminish in subsequent years. As discussed above, we 
received few estimates of ongoing compliance costs from commenters, and 
the estimates we received varied widely. Some commenters suggested that 
the internal burden hours and external professional costs would not 
decrease after the initial compliance year.\725\ As discussed earlier, 
we believe that part of the initial compliance costs would decline 
after the first year.\726\ Other commenters provided estimates of 
ongoing compliance costs that ranged from 28% to 80% of the initial 
compliance costs. For example, one commenter estimated that its burden 
would decrease by approximately 72%, from 3,000 internal hours during 
the first year to 850 internal hours in subsequent years, approximately 
28% of the initial burden.\727\ Another commenter suggested, based on 
results from a survey that it conducted, that ongoing costs could be 
about 80% (median) or 72% (average) of the initial compliance 
costs.\728\ Yet another commenter estimated its initial burden at 1,000 
internal hours and ongoing burden at 50% of the initial compliance 
costs (500 internal hours).\729\ This commenter also estimated initial 
compliance costs at $250,000 with ongoing compliance costs of $100,000, 
or 40% of the initial compliance costs.
---------------------------------------------------------------------------

    \725\ See letters from Chamber II and Intel.
    \726\ See discussion in Section III.C.2.c.
    \727\ See letter from ExxonMobil.
    \728\ See letter from COEC I.
    \729\ See letter from FEI.
---------------------------------------------------------------------------

    Because of the limited number of ongoing cost estimates and their 
wide dispersion, for the purposes of the PRA we assume that ongoing 
compliance burdens and costs will be approximately 40% (the median of 
the estimates) of the initial compliance burdens and costs. Thus, we 
used one commenter's estimated ongoing burden of 40% of the initial 
burden for our estimated three-year average burden. We utilize an 
estimated burden of 1,105 hours in the initial year and 442 \730\ hours 
in the two years thereafter, for a three-year average burden of 663 
hours.\731\
---------------------------------------------------------------------------

    \730\ 1,105 x 40% = 442 burden hours.
    \731\ (1,105 + 442 + 442)/3 = 663 burden hours.
---------------------------------------------------------------------------

 2. Estimated Cost Burdens
    Commenters provided a wide range of estimated external cost 
burdens. Commenters provided estimates of $185,600 per year on average 
for each company,\732\ $250,000 on an initial basis and $100,000 on an 
ongoing basis,\733\ and 20-40 hours of external consulting time.\734\ 
Another commenter estimated that it may cost between $500,000 and $1 
million for each company to establish and automate the process to 
comply with the proposed rule, although it is not clear whether this 
includes both internal and external costs and burdens.\735\ As 
discussed above, we estimate that total compliance burdens for the 
initial year of compliance will be $1,314,694,544 or $368,159 per 
registrant. Assuming that 25% of the burden of preparing the disclosure 
is carried by outside professionals, we estimate that the average 
registrant will incur $147,200 in outside professional costs in the 
first year to comply with the disclosure requirement.\736\
---------------------------------------------------------------------------

    \732\ See letter from Chamber II.
    \733\ See letter from FEI.
    \734\ See letter from Avery Dennison.
    \735\ See letter from KBR.
    \736\ 368 x $400 = $147,200.
---------------------------------------------------------------------------

    As with the estimated internal burden hours, we assume that the 
compliance costs after the initial year will be reduced because a 
substantial portion of the costs will be related to establishing 
systems and processes to collect the payroll data in the initial year 
of compliance. Applying the same assumption used above that the ongoing 
compliance costs will be approximately 40% of the estimate for the 
initial compliance year, we estimate that ongoing compliance costs will 
be approximately $58,880 per year on average for each affected company 
\737\ so that the three-year average cost of compliance is 
$88,320.\738\
---------------------------------------------------------------------------

    \737\ $147,200x 40% = $58,880.
    \738\ ($147,200 + 58,880 + 58,880)/3 = $88,320.
---------------------------------------------------------------------------

3. Estimated Cost and Hour Burdens for Each Collection of Information
    For each collection of information, we estimate the following cost 
and hour burdens:
a. Regulation S-K
    While the adopted amendments would make revisions to Regulation S-
K, the collection of information requirements for that regulation are 
reflected in the burden hours estimated for the forms and schedules 
listed below. The rules in Regulation S-K do not impose any separate 
burden. Consistent with historical practice, we are retaining an 
estimate of one burden hour to Regulation S-K for administrative 
convenience.
b. Form 10-K
    Only Forms 10-K that are filed by registrants that are not smaller 
reporting companies or emerging growth companies will be required to 
include the pay ratio disclosure. For purposes of our PRA estimates, we 
have assumed that 100% of asset-backed securities issuers will omit 
Item 402 disclosure from Form 10-K pursuant to Instruction J of Form 
10-K and 100% of wholly-owned subsidiary registrants will omit Item 402 
disclosure from Form 10-K pursuant to Instruction I of Form 10-K, and, 
accordingly, these registrants will also not be subject to the new 
disclosure requirements. Based on a review of EDGAR filings in calendar 
year 2014, we estimate that of the approximately 7,619 annual reports 
filed in that year, approximately 3,571 annual reports are filed by 
registrants that would be subject to the new disclosure 
requirements.\739\ We estimate that the new disclosure requirements 
will add an average of 663 burden hours \740\ to the total burden hours 
required to produce each Form 10-K that is subject to the new 
requirements and approximately $88,320 for outside professionals.\741\
---------------------------------------------------------------------------

    \739\ Based on a review of EDGAR filings in 2014, approximately 
678 annual reports were filed by EGCs, 2,958 by SRCs, and 412 by ABS 
issuers. See Section III.D.2.b above.
    \740\ As we discuss below, we estimate that 10 of the Forms 8-K 
filed in a given year would require one additional hour for 
preparing the disclosure required by the amendments. Thus, 
substantially all of the internal burden of the pay ratio disclosure 
is allocated to Form 10-K: 663(3,571)-(110) = 2,367,563 or 
approximately 663 (2,367,563/3,571) per response. Burden hours are 
rounded to the next hour.
    \741\ As discussed below, we estimate that the requirement to 
provide updated pay ratio disclosure on Form 8-K will result in one 
additional burden hour for that form. We attribute the external 
costs of the required pay ratio disclosure proportionately between 
Form 10-K and Form 8-K based on the estimated internal burden hours 
for each form. [1/(663+1)] x $88,320 = $133 per Form 8-K response. 
The remaining costs have been attributed to Form 10-K: 
$88,320)3,571--$133(10) = $315,389,390 in aggregate or $88,320 
($315,389,390/3,571) per response. Costs are rounded up to the next 
dollar.

---------------------------------------------------------------------------

[[Page 50182]]

    We estimate that the preparation of annual reports currently 
results in a total annual compliance burden of 12,198,095 hours and an 
annual cost of outside professionals of $1,627,400,000. Under the final 
rule, we estimate that the total incremental cost of outside 
professionals for annual reports will be approximately $315,389,390 per 
year and the total incremental internal burden will be approximately 
2,367,563 hours per year.
c. Form 8-K
    As described in this release, the final rule will require a 
registrant that is filing its PEO total compensation on a delayed basis 
due to the unavailability of certain components of compensation on Form 
8-K (in accordance with Instruction 1 to Items 402(c)(2)(iii) and (iv) 
of Regulation S-K and Item 5.02(f) of Form 8-K) to provide the pay 
ratio disclosure at the same time. The final rule also includes a 
conforming amendment to Item 5.02 of Form 8-K that will require a 
registrant to include updated pay ratio disclosure in the Form 8-K that 
it files to disclose its PEO total compensation information.\742\ We 
estimate that the burden for adding the pay ratio disclosure to that 
Form 8-K filing will be one hour per registrant.\743\ We also estimate 
that the Form 8-K amendment will not result in additional Form 8-K 
filings because registrants who omit disclosure in reliance on 
Instruction 1 to Items 402(c)(2)(iii) and (iv) are already required to 
file a Form 8-K. The amendments will, however, add pay ratio disclosure 
requirements to that Form 8-K filing.
---------------------------------------------------------------------------

    \742\ See Section II.B.6.b.
    \743\ As noted above, we have assumed that the burden relating 
to the new pay ratio requirements would remain associated with the 
registrant's proxy or information statement or annual report, and, 
therefore, our PRA estimates for those forms reflect that burden.
---------------------------------------------------------------------------

    Based on a review of EDGAR filings for calendar years 2012 and 
2013, we estimate that on average approximately 11 Forms 8-K are filed 
pursuant to Item 5.02(f) annually and approximately 10 of these relate 
to disclosure of PEO compensation. As a result, we estimate that 10 of 
the Forms 8-K filed in a given year will require one additional hour 
for preparing the disclosure required by the amendments, in addition to 
the total burden hours required to produce each Form 8-K.
    We estimate that the preparation of current reports on Form 8-K 
currently results in a total annual compliance burden of 507,675 hours 
and an annual cost of outside professionals of $67,688,700. As result 
of the rule, we estimate that the incremental company burden will be 
approximately 10 hours per year and approximately $1,330 in the 
incremental cost of outside professionals for current reports on Form 
8-K.
d. Proxy Statements on Schedule 14A
    Only proxy statements on Schedule 14A that are required to include 
Item 402 information, and that are not filed by smaller reporting 
companies or emerging growth companies, will be required to include the 
new pay ratio disclosure. For purposes of our PRA estimates, consistent 
with past amendments to Item 402,\744\ we have assumed that all of the 
burden relating to the new disclosure requirements will be associated 
with Form 10-K, even if registrants include the new disclosure required 
in Form 10-K by incorporating that disclosure by reference from a proxy 
statement on Schedule 14A.
---------------------------------------------------------------------------

    \744\ We took a similar approach in connection with the rules 
for Summary Compensation Table disclosure required by the 2006 
amendments to Item 402. See 2006 Adopting Release, supra note 9.
---------------------------------------------------------------------------

e. Information Statements on Schedule 14C
    Only information statements on Schedule 14C that are required to 
include Item 402 information, and that are not filed by smaller 
reporting companies or emerging growth companies, are required to 
include the pay ratio disclosure. For purposes of our PRA estimates, 
consistent with past amendments to Item 402, we have assumed that all 
of the burden relating to the disclosure requirements will be 
associated with Form 10-K, even if registrants include the disclosure 
required in Form 10-K by incorporating that disclosure by reference 
from an information statement on Schedule 14C.
f. Form S-1
    Because we have assumed that all new registrants will take 
advantage of the transition period afforded to them under the final 
rule, so that all of the registration statements on Form S-1 that will 
be required to include the pay ratio disclosure will incorporate by 
reference the registrant's disclosure contained in its annual report, 
we have assumed that all of the burden relating to the new disclosure 
requirements will be associated with Form 10-K.
g. Form S-4
    We have assumed that registrants filing on Form S-4 for whom 
executive compensation information under Item 402 is required pursuant 
to Items 18 or 19 of Form S-4 will incorporate by reference the pay 
ratio disclosure contained in the registrant's annual report. Thus, we 
have assumed that all of the burden relating to the new disclosure 
requirements will be associated with Form 10-K.
h. Form S-11
    Because we have assumed that all new registrants will take 
advantage of the transition period afforded to them under the final 
rule, so that all of the registration statements on Form S-11 that will 
be required to include the pay ratio disclosure will incorporate by 
reference the registrant's pay ratio disclosure contained in its annual 
report, we have assumed that all of the burden relating to the new 
disclosure requirements will be associated with Form 10-K.
i. Form N-2
    Only Forms N-2 filed by business development companies (BDCs) will 
be subject to the new disclosure requirements. Furthermore, the final 
rule will apply only to BDCs internally managed such that they 
compensate their own employees. Rather, such employees are generally 
compensated by the BDC's investment adviser. Because we assume that all 
of the Forms N-2 that will be filed by internally managed BDCs will 
incorporate by reference the registrant's disclosure contained in its 
annual report, we have assumed that all of the burden relating to the 
new disclosure requirements would be associated with Form 10-K.
j. Form 10
    Because we have assumed that all new registrants would take 
advantage of the transition period afforded to them under the final 
rule, we estimate no annual incremental increase in the paperwork 
burden associated with Form 10 as a result of the new requirements.

E. Summary of Changes to Annual Compliance Burden in Collection of 
Information

    Tables 1 and 2 below illustrate the total annual compliance burden 
of the collection of information in hours and in cost under the final 
rule for annual reports on Form 10-K and current reports on Form 8-K 
under the Exchange Act. \745\ The burden estimates

[[Page 50183]]

were calculated by multiplying the estimated number of annual responses 
by the estimated average number of hours it would take a company to 
prepare and review the new disclosure.
---------------------------------------------------------------------------

    \745\ Figures in both tables have been rounded to the nearest 
whole number.
---------------------------------------------------------------------------

    As discussed above, there is no change to the estimated burden of 
the collection of information under Forms S-1, S-4, S-11 or N-2 or 
under Schedule 14A and 14C because we have assumed that the burden 
relating to the new disclosure requirements would be associated 
primarily with Form 10-K. In addition, there is no change to the 
estimated burden of the collection of information under Form 10 because 
we have assumed that all new registrants would take advantage of the 
transition period. There is no change to the estimated burden of the 
collection of information under Regulation S-K because the burdens that 
Regulation S-K imposes are reflected in our revised estimates for the 
forms.\746\
---------------------------------------------------------------------------

    \746\ Estimates in columns C and E of Table 1 are affected by 
rounding to the next burden hour. See Section III.D.1. and 2. above 
for an analysis of the derivation of the estimated incremental 
burden hours and costs.

                           Table 1--Incremental Paperwork Burden under the Final Rule
----------------------------------------------------------------------------------------------------------------
                                                                       Total                           Total
                                     Number of      Hour burden     incremental     Incremental     incremental
                                      annual       per response   company burden   professional    professional
                                   responses (A)        (B)         hours (C) =      costs (D)      costs (E) =
                                                                     (A) * (B)                       (A) * (D)
----------------------------------------------------------------------------------------------------------------
Form 10-K.......................           3,571             663       2,367,563      $88,319.63    $315,389,390
Form 8-K........................              10               1              10            $133          $1,330
                                 -------------------------------------------------------------------------------
    Total.......................  ..............  ..............       2,367,573  ..............    $315,390,720
----------------------------------------------------------------------------------------------------------------


                                                   Table 2--Calculation of Total PRA Burden Estimates
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                  Current                 Increase in
                                                   annual      Current       burden       Burden         Current         Increase in      Professional
                                                 responses      burden     hours (C)   hours (D) =     professional     professional   costs (G) = (E) +
                                                  (A)\747\    hours (B)      \748\      (B) + (C)       costs (E)         costs (F)           (F)
--------------------------------------------------------------------------------------------------------------------------------------------------------
Form 10-K.....................................        8,137   12,198,095    2,367,563   14,565,658     $1,627,400,000    $315,389,390     $1,942,789,390
Form 8-K......................................       74,911      507,665           10      507,675        $67,688,700          $1,330        $67,690,030
                                               ---------------------------------------------------------------------------------------------------------
    Total.....................................       83,048   12,705,760    2,367,573   15,073,333     $1,695,088,700    $315,390,720     $2,010,479,420
--------------------------------------------------------------------------------------------------------------------------------------------------------

V. Final Regulatory Flexibility Act Certification

    The Regulatory Flexibility Act of 1980 (``RFA'') \749\ requires us, 
in promulgating rules, to consider the impact of those rules on small 
entities. The Commission certified in the Proposing Release, pursuant 
to Section 605(b) \750\ of the RFA, that the proposed rule, if adopted, 
would not have a significant economic impact on a substantial number of 
small entities. We received no comments on this certification.
---------------------------------------------------------------------------

    \747\ For these forms, the number of current annual responses 
reflected in the table equals the three-year average of the number 
of forms filed with us and currently reported by us to OMB.
    \748\ The increase in burden hours reflected in the table is 
based on the aggregate incremental burden hours per form multiplied 
by the annual responses that will be required to include additional 
disclosure under the new rules as adopted. As explained in the 
discussion above, for purposes of determining the total increase in 
burden hours, we have reduced the current number of annual responses 
to reflect that the disclosure requirements will not apply to all 
forms filed. See Table 1 for estimates per response.
    \749\ 5 U.S.C. 601 et seq.
    \750\ 5 U.S.C. 605(b).
---------------------------------------------------------------------------

    The final rule amends Item 402 by adding paragraph (u) to implement 
Section 953(b) of the Dodd-Frank Act. Specifically, the final rule 
requires registrants, other than emerging growth companies, smaller 
reporting companies and foreign private issuers, to disclose the median 
of the annual total compensation of all employees of the registrant 
(excluding the PEO), the annual total compensation of the registrant's 
PEO, and the ratio between these two amounts. The disclosure is 
required in any filing described in Item 10(a) of Reg. S-K that 
requires executive compensation disclosure pursuant to Item 402.
    For purposes of the RFA, under our rules, an issuer, other than an 
investment company,\751\ is a ``small business'' or ``small 
organization'' if it has total assets of $5 million or less as of the 
end of its most recent fiscal year and is engaged or proposing to 
engage in an offering of securities which does not exceed $5 
million.\752\ We believe that the final rule will affect some small 
entities that are business development companies that have a class of 
securities registered under Section 12 of the Exchange Act. We estimate 
that there are approximately five of those business development 
companies that may be considered small entities.\753\ As discussed 
above, emerging growth companies and smaller reporting companies are 
excluded from the final rule. An ``emerging growth company'' is an 
issuer that had total annual gross revenues of less than $1 billion 
during its most recently completed fiscal year.\754\ A smaller 
reporting company is an issuer, other than certain classes of issuers 
(including investment companies), that had a public float of less than 
$75 million as of the end of its most recently completed second fiscal 
quarter, or in the case of an initial registration statement under the 
Securities Act or Exchange Act for the shares of its common equity, had 
a public float of less than $75 million as of a date within 30 days of 
the date of filing of the registration statement.\755\ To

[[Page 50184]]

the extent that a small entity is a registrant, we believe that there 
are few, if any, small entities that do not qualify as emerging growth 
companies or smaller reporting companies because it is unlikely that an 
entity with total assets of $5 million or less would have total annual 
gross revenues of $1 billion or more, or would have a public float of 
$75 million or more. Because emerging growth companies and smaller 
reporting companies are excluded from the new disclosure requirement, 
we believe that the final rule applies to few, if any, small entities, 
other than the five business development companies.
---------------------------------------------------------------------------

    \751\ For purposes of the RFA, an investment company is a 
``small business'' or ``small organization'' that, together with 
other investment companies in the same group of related investment 
companies, has net assets of $50 million or less as of the end of 
its most recent fiscal year. 17 CFR 270.0-10.
    \752\ See Securities Act Rule 157 [17 CFR 230.157] and Exchange 
Act Rule 0-10(a) [17 CFR 240.0-10(a)].
    \753\ We estimate that there are 13 business development 
companies that will be subject to the final rule, five of which may 
be considered small entities for purposes of the RFA.
    \754\ See Securities Act Section 2(a)(19) [15 U.S.C. 
77b(a)(19)].
    \755\ See Securities Act Rule 405 [17 CFR 230.405]. In the case 
of an issuer whose public float was zero, the issuer could qualify 
as a smaller reporting company if it had annual revenues of less 
than $50 million during the most recently completed fiscal year for 
which audited financial statements are available.
---------------------------------------------------------------------------

    For the above reasons, we again certify, pursuant to 5 U.S.C. 
605(b), that the final rule will not have a significant economic impact 
on a substantial number of small entities.

VI. Statutory Authority

    The amendments contained herein are being proposed pursuant to 
Sections 7, 10, 19(a), and 28 of the Securities Act, Sections 3(b), 12, 
13, 14, 15(d), 23(a), and 36(a) of the Exchange Act, Section 953(b) of 
the Dodd-Frank Act, as amended, and Section 102(a)(3) of the JOBS Act.

List of Subjects

17 CFR Part 229

    Reporting and recordkeeping requirements, Securities.

17 CFR Part 240

    Brokers, Confidential business information, Fraud, Reporting and 
recordkeeping requirements, Securities.

17 CFR Part 249

    Brokers, Reporting and recordkeeping requirements, Securities.

Text of the Amendments

    In accordance with the foregoing, title 17, chapter II of the Code 
of Federal Regulations, is amended as follows:

PART 229 --STANDARD INSTRUCTIONS FOR FILING FORMS UNDER SECURITIES 
ACT OF 1933, SECURITIES EXCHANGE ACT OF 1934 AND ENERGY POLICY AND 
CONSERVATION ACT OF 1975--REGULATION S-K

0
1. The authority citation for part 229 is revised to read as follows:

    Authority:  15 U.S.C. 77e, 77f, 77g, 77h, 77j, 77k, 77s, 77z-2, 
77z-3, 77aa(25), 77aa(26), 77ddd, 77eee, 77ggg, 77hhh, 77iii, 77jjj, 
77nnn, 77sss, 78c, 78i, 78j, 78j-3, 78l, 78m, 78n, 78n-1, 78o, 78u-
5, 78w, 78ll, 78 mm, 80a-8, 80a-9, 80a-20, 80a-29, 80a-30, 80a-
31(c), 80a-37, 80a-38(a), 80a-39, 80b-11 and 7201 et seq.; 18 U.S.C. 
1350; Sec. 953(b), Pub. L. 111-203, 124 Stat. 1904; and Sec. 
102(a)(3), Pub. L. 112-106, 126 Stat. 309.

0
2. Amend Sec.  229.402 by:
0
a. In paragraph (l), removing ``(k) and (s)'' and adding in its place 
``(k), (s), and (u)''; and
0
b. Adding paragraph (u) directly after the Instructions to Item 402(t).
    The addition reads as follows:


Sec.  229.402  (Item 402) Executive compensation.

* * * * *
    (u) Pay ratio disclosure--(1) Disclose. (i) The median of the 
annual total compensation of all employees of the registrant, except 
the PEO of the registrant;
    (ii) The annual total compensation of the PEO of the registrant; 
and
    (iii) The ratio of the amount in paragraph (u)(1)(i) of this Item 
to the amount in paragraph (u)(1)(ii) of this Item. For purposes of the 
ratio required by this paragraph (u)(1)(iii), the amount in paragraph 
(u)(1)(i) of this Item shall equal one, or, alternatively, the ratio 
may be expressed narratively as the multiple that the amount in 
paragraph (u)(1)(ii) of this Item bears to the amount in paragraph 
(u)(1)(i) of this Item.
    (2) For purposes of this paragraph (u):
    (i) Total compensation for the median of annual total compensation 
of all employees of the registrant and the PEO of the registrant shall 
be determined in accordance with paragraph (c)(2)(x) of this Item. In 
determining the total compensation, all references to ``named executive 
officer'' in this Item and the instructions thereto may be deemed to 
refer instead, as applicable, to ``employee'' and, for non-salaried 
employees, references to ``base salary'' and ``salary'' in this Item 
and the instructions thereto may be deemed to refer instead, as 
applicable, to ``wages plus overtime'';
    (ii) Annual total compensation means total compensation for the 
registrant's last completed fiscal year; and
    (iii) Registrant means the registrant and its consolidated 
subsidiaries.
    (3) For purposes of this paragraph (u), employee or employee of the 
registrant means an individual employed by the registrant or any of its 
consolidated subsidiaries, whether as a full-time, part-time, seasonal, 
or temporary worker, as of a date chosen by the registrant within the 
last three months of the registrant's last completed fiscal year. The 
definition of employee or employee of the registrant does not include 
those workers who are employed, and whose compensation is determined, 
by an unaffiliated third party but who provide services to the 
registrant or its consolidated subsidiaries as independent contractors 
or ``leased'' workers.
    (4) For purposes of this paragraph (u), an employee located in a 
jurisdiction outside the United States (a ``non-U.S. employee'') may be 
exempt from the definition of employee or employee of the registrant 
under either of the following conditions:
    (i) The employee is employed in a foreign jurisdiction in which the 
laws or regulations governing data privacy are such that, despite its 
reasonable efforts to obtain or process the information necessary for 
compliance with this paragraph (u), the registrant is unable to do so 
without violating such data privacy laws or regulations. The 
registrant's reasonable efforts shall include, at a minimum, using or 
seeking an exemption or other relief under any governing data privacy 
laws or regulations. If the registrant chooses to exclude any employees 
using this exemption, it shall list the excluded jurisdictions, 
identify the specific data privacy law or regulation, explain how 
complying with this paragraph (u) violates such data privacy law or 
regulation (including the efforts made by the registrant to use or seek 
an exemption or other relief under such law or regulation), and provide 
the approximate number of employees exempted from each jurisdiction 
based on this exemption. In addition, if a registrant excludes any non-
U.S. employees in a particular jurisdiction under this exemption, it 
must exclude all non-U.S. employees in that jurisdiction. Further, the 
registrant shall obtain a legal opinion from counsel that opines on the 
inability of the registrant to obtain or process the information 
necessary for compliance with this paragraph (u) without violating the 
jurisdiction's laws or regulations governing data privacy, including 
the registrant's inability to obtain an exemption or other relief under 
any governing laws or regulations. The registrant shall file the legal 
opinion as an exhibit to the filing in which the pay ratio disclosure 
is included.
    (ii) The registrant's non-U.S. employees account for 5% or less of 
the registrant's total employees. In that circumstance, if the 
registrant chooses to exclude any non-U.S. employees under this 
exemption, it must exclude all non-U.S. employees. Additionally, if a 
registrant's non-U.S. employees exceed 5% of the registrant's total 
U.S. and non-U.S. employees, it may exclude

[[Page 50185]]

up to 5% of its total employees who are non-U.S. employees; provided, 
however, if a registrant excludes any non-U.S. employees in a 
particular jurisdiction, it must exclude all non-U.S. employees in that 
jurisdiction. If more than 5% of a registrant's employees are located 
in any one non-U.S. jurisdiction, the registrant may not exclude any 
employees in that jurisdiction under this exemption.
    (A) In calculating the number of non-U.S. employees that may be 
excluded under this Item 402(u)(4)(ii) (``de minimis'' exemption), a 
registrant shall count against the total any non-U.S. employee exempted 
under the data privacy law exemption under Item 402(u)(4)(i) (``data 
privacy'' exemption). A registrant may exclude any non-U.S. employee 
from a jurisdiction that meets the data privacy exemption, even if the 
number of excluded employees exceeds 5% of the registrant's total 
employees. If, however, the number of employees excluded under the data 
privacy exemption equals or exceeds 5% of the registrant's total 
employees, the registrant may not use the de minimis exemption. 
Additionally, if the number of employees excluded under the data 
privacy exemption is less than 5% of the registrant's total employees, 
the registrant may use the de minimis exemption to exclude no more than 
the number of non-U.S. employees that, combined with the data privacy 
exemption, does not exceed 5% of the registrant's total employees.
    (B) If a registrant excludes non-U.S. employees under the de 
minimis exemption, it must disclose the jurisdiction or jurisdictions 
from which those employees are being excluded, the approximate number 
of employees excluded from each jurisdiction under the de minimis 
exemption, the total number of its U.S. and non-U.S. employees 
irrespective of any exemption (data privacy or de minimis), and the 
total number of its U.S. and non-U.S. employees used for its de minimis 
calculation.
    Instruction 1 to Item 402(u)--Disclosing the date chosen for 
identifying the median employee. A registrant shall disclose the date 
within the last three months of its last completed fiscal year that it 
selected pursuant to paragraph (u)(3) of this Item to identify its 
median employee. If the registrant changes the date it uses to identify 
the median employee from the prior year, the registrant shall disclose 
this change and provide a brief explanation about the reason or reasons 
for the change.
    Instruction 2 to Item 402(u)--Identifying the median employee. A 
registrant is required to identify its median employee only once every 
three years and calculate total compensation for that employee each 
year; provided that, during a registrant's last completed fiscal year 
there has been no change in its employee population or employee 
compensation arrangements that it reasonably believes would result in a 
significant change to its pay ratio disclosure. If there have been no 
changes that the registrant reasonably believes would significantly 
affect its pay ratio disclosure, the registrant shall disclose that it 
is using the same median employee in its pay ratio calculation and 
describe briefly the basis for its reasonable belief. For example, the 
registrant could disclose that there has been no change in its employee 
population or employee compensation arrangements that it believes would 
significantly impact the pay ratio disclosure. If there has been a 
change in the registrant's employee population or employee compensation 
arrangements that the registrant reasonably believes would result in a 
significant change in its pay ratio disclosure, the registrant shall 
re-identify the median employee for that fiscal year. If it is no 
longer appropriate for the registrant to use the median employee 
identified in year one as the median employee in years two or three 
because of a change in the original median employee's circumstances 
that the registrant reasonably believes would result in a significant 
change in its pay ratio disclosure, the registrant may use another 
employee whose compensation is substantially similar to the original 
median employee based on the compensation measure used to select the 
original median employee.
    Instruction 3 to Item 402(u)--Updating for the last completed 
fiscal year. Pay ratio information (i.e., the disclosure called for by 
paragraph (u)(1) of this Item) with respect to the registrant's last 
completed fiscal year is not required to be disclosed until the filing 
of its annual report on Form 10-K for that last completed fiscal year 
or, if later, the filing of a definitive proxy or information statement 
relating to its next annual meeting of shareholders (or written 
consents in lieu of such a meeting) following the end of such fiscal 
year; provided that, the required pay ratio information must, in any 
event, be filed as provided in General Instruction G(3) of Form 10-K 
(17 CFR 249.310) not later than 120 days after the end of such fiscal 
year.
    Instruction 4 to Item 402(u)--Methodology and use of estimates. 1. 
Registrants may use reasonable estimates both in the methodology used 
to identify the median employee and in calculating the annual total 
compensation or any elements of total compensation for employees other 
than the PEO.
    2. In determining the employees from which the median employee is 
identified, a registrant may use its employee population or statistical 
sampling and/or other reasonable methods.
    3. A registrant may identify the median employee using annual total 
compensation or any other compensation measure that is consistently 
applied to all employees included in the calculation, such as 
information derived from the registrant's tax and/or payroll records. 
In using a compensation measure other than annual total compensation to 
identify the median employee, if that measure is recorded on a basis 
other than the registrant's fiscal year (such as information derived 
from tax and/or payroll records), the registrant may use the same 
annual period that is used to derive those amounts. Where a 
compensation measure other than annual total compensation is used to 
identify the median employee, the registrant must disclose the 
compensation measure used.
    4. In identifying the median employee, whether using annual total 
compensation or any other compensation measure that is consistently 
applied to all employees included in the calculation, the registrant 
may make cost-of-living adjustments to the compensation of employees in 
jurisdictions other than the jurisdiction in which the PEO resides so 
that the compensation is adjusted to the cost of living in the 
jurisdiction in which the PEO resides. If the registrant uses a cost-
of-living adjustment to identify the median employee, and the median 
employee identified is an employee in a jurisdiction other than the 
jurisdiction in which the PEO resides, the registrant must use the same 
cost-of-living adjustment in calculating the median employee's annual 
total compensation and disclose the median employee's jurisdiction. The 
registrant also shall briefly describe the cost-of-living adjustments 
it used to identify the median employee and briefly describe the cost-
of-living adjustments it used to calculate the median employee's annual 
total compensation, including the measure used as the basis for the 
cost-of-living adjustment. A registrant electing to present the pay 
ratio in this manner also shall disclose the median employee's annual 
total compensation and pay ratio without the cost-of-living adjustment. 
To calculate this pay ratio, the registrant will need to identify the

[[Page 50186]]

median employee without using any cost-of-living adjustments.
    5. The registrant shall briefly describe the methodology it used to 
identify the median employee. It shall also briefly describe any 
material assumptions, adjustments (including any cost-of-living 
adjustments), or estimates it used to identify the median employee or 
to determine total compensation or any elements of total compensation, 
which shall be consistently applied. The registrant shall clearly 
identify any estimates used. The required descriptions should be a 
brief overview; it is not necessary for the registrant to provide 
technical analyses or formulas. If a registrant changes its methodology 
or its material assumptions, adjustments, or estimates from those used 
in its pay ratio disclosure for the prior fiscal year, and if the 
effects of any such change are significant, the registrant shall 
briefly describe the change and the reasons for the change. Registrants 
must also disclose if they changed from using the cost-of-living 
adjustment to not using that adjustment and if they changed from not 
using the cost-of-living adjustment to using it.
    6. Registrants may, at their discretion, include personal benefits 
that aggregate less than $10,000 and compensation under non-
discriminatory benefit plans in calculating the annual total 
compensation of the median employee as long as these items are also 
included in calculating the PEO's annual total compensation. The 
registrant shall also explain any difference between the PEO's annual 
total compensation used in the pay ratio disclosure and the total 
compensation amounts reflected in the Summary Compensation Table, if 
material.
    Instruction 5 to Item 402(u)--Permitted annualizing adjustments. A 
registrant may annualize the total compensation for all permanent 
employees (full-time or part-time) that were employed by the registrant 
for less than the full fiscal year (such as newly hired employees or 
permanent employees on an unpaid leave of absence during the period). A 
registrant may not annualize the total compensation for employees in 
temporary or seasonal positions. A registrant may not make a full-time 
equivalent adjustment for any employee.
    Instruction 6 to Item 402(u)--PEO compensation not available. A 
registrant that is relying on Instruction 1 to Item 402(c)(2)(iii) and 
(iv) in connection with the salary or bonus of the PEO for the last 
completed fiscal year, shall disclose that the pay ratio required by 
paragraph (u) of this Item is not calculable until the PEO salary or 
bonus, as applicable, is determined and shall disclose the date that 
the PEO's actual total compensation is expected to be determined. The 
disclosure required by paragraph (u) of this Item shall then be 
disclosed in the filing under Item 5.02(f) of Form 8-K (17 CFR 249.308) 
that discloses the PEO's salary or bonus in accordance with Instruction 
1 to Item 402(c)(2)(iii) and (iv).
    Instruction 7 to Item 402(u)--Transition periods for registrants. 
1. Upon becoming subject to the requirements of Section 13(a) or 15(d) 
of the Exchange Act (15 U.S.C. 78m or 78o(d)), a registrant shall 
comply with paragraph (u) of this Item with respect to compensation for 
the first fiscal year following the year in which it became subject to 
such requirements, but not for any fiscal year commencing before 
January 1, 2017. The registrant may omit the disclosure required by 
paragraph (u) of this Item from any filing until the filing of its 
annual report on Form 10-K (17 CFR 249.310) for such fiscal year or, if 
later, the filing of a proxy or information statement relating to its 
next annual meeting of shareholders (or written consents in lieu of 
such a meeting) following the end of such year; provided that, such 
disclosure shall, in any event, be filed as provided in General 
Instruction G(3) of Form 10-K not later than 120 days after the end of 
such fiscal year.
    2. A registrant may omit any employees that became its employees as 
the result of the business combination or acquisition of a business for 
the fiscal year in which the transaction becomes effective, but the 
registrant must disclose the approximate number of employees it is 
omitting. Those employees shall be included in the total employee count 
for the triennial calculations of the median employee in the year 
following the transaction for purposes of evaluating whether a 
significant change had occurred. The registrant shall identify the 
acquired business excluded for the fiscal year in which the business 
combination or acquisition becomes effective.
    3. A registrant shall comply with paragraph (u) of this Item with 
respect to compensation for the first fiscal year commencing on or 
after the date the registrant ceases to be a smaller reporting company, 
but not for any fiscal year commencing before January 1, 2017.
    Instruction 8 to Item 402(u)--Emerging growth companies. A 
registrant is not required to comply with paragraph (u) of this Item if 
it is an emerging growth company as defined in Section 2(a)(19) of the 
Securities Act (15 U.S.C. 77(b)(a)(19)) or Section 3(a)(80) of the 
Exchange Act (15 U.S.C. 78c(a)(80)). A registrant shall comply with 
paragraph (u) of this Item with respect to compensation for the first 
fiscal year commencing on or after the date the registrant ceases to be 
an emerging growth company, but not for any fiscal year commencing 
before January 1, 2017.
    Instruction 9 to Item 402(u)--Additional information. Registrants 
may present additional information, including additional ratios, to 
supplement the required ratio, but are not required to do so. Any 
additional information shall be clearly identified, not misleading, and 
not presented with greater prominence than the required ratio.
    Instruction 10 to Item 402(u)--Multiple PEOs during the year. A 
registrant with more than one non-concurrent PEO serving during its 
fiscal year may calculate the annual total compensation for its PEO in 
either of the following manners:
    1. The registrant may calculate the compensation provided to each 
person who served as PEO during the year for the time he or she served 
as PEO and combine those figures; or
    2. The registrant may look to the PEO serving in that position on 
the date it selects to identify the median employee and annualize that 
PEO's compensation.
    Regardless of the alternative selected, the registrant shall 
disclose which option it chose and how it calculated its PEO's annual 
total compensation.
    Instruction 11 to Item 402(u)--Employees' personally identifiable 
information. Registrants are not required to, and should not, disclose 
any personally identifiable information about that employee other than 
his or her compensation. Registrants may choose to generally identify 
an employee's position to put the employee's compensation in context, 
but registrants are not required to provide this information and should 
not do so if providing the information could identify any specific 
individual.
* * * * *

PART 240--GENERAL RULES AND REGULATIONS, SECURITIES EXCHANGE ACT OF 
1934

0
3. The authority citation for part 240 continues to read, in part, as 
follows:

    Authority: 15 U.S.C. 77c, 77d, 77g, 77j, 77s, 77z-2, 77z-3, 
77eee, 77ggg, 77nnn, 77sss, 77ttt, 78c, 78c-3, 78c-5, 78d, 78e, 78f, 
78g, 78i, 78j, 78j-1, 78k, 78k-1, 78m, 78n, 78n-1, 78o, 78o-4, 78o-
10, 78p, 78q, 78q-1, 78s, 78u-5, 78w, 78x, 78ll, 78mm, 80a-20, 80a-
23, 80a-29, 80a-37, 80b-3, 80b-4, 80b-

[[Page 50187]]

11, and 7210 et seq., and 8302; 7 U.S.C. 2(c)(2)(E); 12 U.S.C. 
5521(e)(3); 18 U.S.C. 1350; and Pub. L. 111-203, 939A, 124 Stat. 
1376 (2010), unless otherwise noted.
* * * * *

0
4. Amend Sec.  240.14a-101 by adding Item 25 at the end to read as 
follows:


Sec.  240.14a-101  Schedule 14A. Information required in proxy 
statement.

    SCHEDULE 14A INFORMATION
* * * * *
    Item 25. Exhibits. Provide the legal opinion required to be filed 
by Item 402(u)(4)(i) of Regulation S-K (17 CFR 229.402(u)) in an 
exhibit to this Schedule 14A.

PART 249 -- FORMS, SECURITIES EXCHANGE ACT OF 1934

0
5. The authority citation for part 249 is revised, in part, to read as 
follows:

    Authority:  15 U.S.C. 78a et seq. and 7201 et seq.; 12 U.S.C. 
5461 et seq.; 18 U.S.C. 1350; Sec. 953(b), Pub. L. 111-203, 124 
Stat. 1904; and Sec. 102(a)(3), Pub. L. 112-106, 126 Stat. 309, 
unless otherwise noted.
* * * * *

0
6. Form 8-K (referenced in Sec.  249.308) is amended by redesignating 
paragraph (f) as (f)(1) and adding paragraph (f)(2) to read as follows:
    Form 8-K
* * * * *
    Item 5.02 Departure of Directors or Certain Officers; Election of 
Directors; Appointment of Certain Officers; Compensatory Arrangements 
of Certain Officers.
* * * * *
    (f)(1) * * *
    (2) As specified in Instruction 6 to Item 402(u) of Regulation S-K 
(17 CFR 229.402(u)), disclosure under this Item 5.02(f) with respect to 
the salary or bonus of a principal executive officer shall include pay 
ratio disclosure pursuant to Item 402(u) of Regulation S-K calculated 
using the new total compensation figure for the principal executive 
officer. Pay ratio disclosure is not required under this Item 5.02(f) 
until the omitted salary or bonus amounts for such principal executive 
officer become calculable in whole.
* * * * *

    Note:  The text of Form 8-K does not, and this amendment will 
not, appear in the Code of Federal Regulations.


    By the Commission.
    Dated: August 5, 2015.
Brent J. Fields,
Secretary.
[FR Doc. 2015-19600 Filed 8-17-15; 8:45 am]
BILLING CODE 8011-01-P