oversight

Thrift Savings Plan: Adding a Socially Responsible Index Fund Presents Challenges

Published by the Government Accountability Office on 2012-06-26.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

             United States Government Accountability Office

GAO          Report to the Ranking Member, Subcommittee
             on Technology, Information Policy,
             Intergovernmental Relations and Procurement
             Reform, Committee on Oversight and
             Government Reform, House of Representatives

June 2012
             THRIFT SAVINGS
             PLAN
             Adding a Socially
             Responsible Index
             Fund Presents
             Challenges




GAO-12-664
                                                June 2012

                                                THRIFT SAVINGS PLAN
                                                Adding a Socially Responsible Index Fund Presents
                                                Challenges
Highlights of GAO-12-664, a report to Ranking
Member, Subcommittee on Technology,
Information Policy, Intergovernmental
Relations and Procurement Reform
Committee on Oversight and Government
Reform, House of Representatives
                                                s




Why GAO Did This Study                          What GAO Found
Socially responsible investment—                Officials at the Thrift Savings Plan (TSP) and the other public retirement plans
investment made on the basis of                 that had considered socially responsible investment (SRI) associated a number
environmental, social, religious, or            of common challenges with SRI adoption. While none of these plans were
corporate governance criteria— in               identical to TSP in scope or demographics, many plan officials shared similar
U.S.-based mutual funds exceeded                challenges and concerns with TSP. For example, they identified participant
$300 billion in value in 2010. TSP—a            demand, SRI screening criteria, and costs as the most common challenges.
$308 billion retirement plan with more          Officials at public retirement plans that had adopted SRI cited some short-term
than 4.5 million participants—currently         benefits of SRI, such as providing participants an opportunity to invest in
offers five distinct low-cost investment
                                                accordance with their values, but said that the long-term benefits were unknown.
options, and is authorized to offer a
service that enables direct participant         When compared to the past performance of the TSP stock portfolio, the addition
investment in mutual funds outside              of a hypothetical SRI index fund tracking the best-performing U.S.-based SRI
TSP. GAO was asked to consider the              stock index would not have both increased returns and lowered volatility in any
value of adding an SRI option to TSP.           allocation scenario that GAO tested. Specifically, over the last 20 years, if TSP
GAO examined: (1) What challenges               had included such an SRI index fund in its existing stock portfolio, it could have
might TSP face in adopting an SRI               resulted in (1) lower returns and lower volatility, (2) lower returns and higher
option? (2) How would the addition of           volatility, or (3) higher returns and higher volatility, based on GAO’s analysis of
an index fund tracking an SRI index             evenly distributed portfolio allocations. The managers of the SRI index explained
have affected past TSP stock portfolio
                                                the difference in the index’s performance over the last 20 years was a result of
performance? (3) How do the
                                                having different sector weightings than the overall market to align with the fund’s
performance and costs of SRI mutual
funds compare to those of non-SRI               SRI strategy. Moreover, the addition of this SRI fund would have resulted in
mutual funds?                                   overlap with the TSP stock portfolio, and would not have provided a substantial
                                                opportunity for additional portfolio diversification.
To analyze the challenges surrounding
SRI, GAO interviewed federal officials,         Effect of Adding an SRI Index Fund to a Portfolio of the TSP Stock Index Funds (1992 to 2011)
SRI experts, and representatives of
public retirement plans that had
considered SRI adoption. To examine
the impact of adding an SRI fund to the
existing TSP funds, GAO analyzed
monthly benchmark return data. To
examine mutual fund performance
trends and costs, GAO analyzed
historical summary data on US-based
mutual funds.
GAO provided a copy of this draft
report to the Board, the Department of
                                                Looking more broadly at SRI mutual funds—the most common form of SRI in the
Labor, and the Department of the
Treasury for review and comment.                United States—GAO found the comparative performance of SRI and non-SRI
None of the agencies provided formal            mutual funds to vary by asset class while costs were nearly the same. Regarding
comments on the report.                         performance, SRI bond mutual funds had better risk-adjusted performance than
                                                their non-SRI counterparts over the last 15 years, while SRI stock and balanced
                                                funds did not. However, after controlling for various factors such as fund size,
What GAO Recommends
                                                SRI stock mutual funds had better estimated performance as well. Regarding
This report contains no                         costs, in fiscal year 2010, the costs of SRI institutional grade mutual funds were
recommendations.                                similar to their non-SRI counterparts. Although TSP participants cannot currently
                                                invest in mutual funds through TSP, the Federal Retirement Thrift Investment
                                                Board (Board) is authorized to offer a mutual fund window if it determines that it
View GAO-12-664. For more information,
contact Charles Jeszeck at (202) 512-7215 or    is in the best interests of participants.
jeszeckc@gao.gov.
                                                                                             United States Government Accountability Office
Contents


Letter                                                                                      1
               Background                                                                   4
               Officials Identified Several Challenges to SRI Adoption and Said
                 the Long-term Economic Benefits of SRI Were Unknown                      13
               Adding an SRI Index Fund Would Not Have Improved Past TSP
                 Portfolio Performance in Most Allocation Scenarios                       19
               SRI Mutual Fund Performance Varied by Asset Class, but Had
                 Similar Costs to Non-SRI Mutual Funds                                    24
               Concluding Observations                                                    26
               Agency Comments                                                            27

Appendix I     Objectives, Scope, and Methodology                                         28



Appendix II    Variation of Annual Compound Rates of Return and Standard Deviation
               among Evenly Distributed Allocations of the Best Performing SRI Stock
               Index and Existing TSP Stock Index Funds, 1992 through 2011           36



Appendix III   Additional Information on the Performance of SRI and Non-SRI
               Mutual Funds by Asset Class                                                37



Appendix IV    GAO Contact and Staff Acknowledgment                                       42



Tables
               Table 1: Investment Requirements for Existing TSP Funds                      6
               Table 2: Regression Coefficients on Non-SRI Indicator: U.S. Stock
                        Funds, Institutional, Front-load, and No-load Share Classes       31
               Table 3: Regression Coefficients on Non-SRI Indicator: U.S.
                        Balanced Funds, Institutional, Front-load, and No-load
                        Share Classes                                                     33
               Table 4: Regression Coefficients on Non-SRI Indicator: U.S. Bond
                        Funds, Institutional, Front-load, and No-load Share Classes       34
               Table 5: Coefficient on SRI Indicator in Regression with Annual
                        Report Net Expense Ratio as Dependent Variable With and
                        Without Control Variables                                         35



               Page i                                           GAO-12-664 Thrift Savings Plan
          Table 6: Sharpe Ratio SRI and Non-SRI Bond Mutual Funds                  39
          Table 7: Sortino Ratio SRI and Non-SRI Bond Mutual Funds                 39
          Table 8: Sharpe Ratio SRI and Non-SRI Stock Mutual Funds                 40
          Table 9: Sortino Ratio SRI and Non-SRI Stock Mutual Funds                40
          Table 10: Sharpe Ratio SRI and Non-SRI Balanced Mutual Funds             41
          Table 11: Sortino Ratio SRI and Non-SRI Balanced Mutual Funds            41


Figures
          Figure 1: Cumulative Growth in the Number of U.S. Open-end SRI
                   Mutual Funds, 1952 through 2011                                  9
          Figure 2: Investment Screens Applied by SRI Mutual Funds                 12
          Figure 3: Number of Public Retirement Plans Identifying
                   Challenges Associated with SRI Adoption Similar to Ones
                   Identified by TSP Officials                                     14
          Figure 4: Effect of Adding a Hypothetical TSP SRI Index Fund on
                   Past TSP Stock Portfolio Performance, 1992 to 2011              20
          Figure 5: Annual and Cumulative Returns of the Indices Underlying
                   the C, S, I, and SRI Funds, 1992 to 2011                        22
          Figure 6: Comparison of Annual Net Expense Ratios of SRI and
                   Non-SRI Mutual Funds                                            26




          Page ii                                        GAO-12-664 Thrift Savings Plan
Abbreviations
Board        Federal Retirement Thrift Investment Board
C Fund       Common Stock Index Investment Fund
DB           defined benefit
DC           defined contribution
ESG          environmental, social, and corporate governance
ERISA        Employee Retirement Income Security Act of 1974
ETAC         Employee Thrift Advisory Council
FERS         Federal Employees’ Retirement System
FERSA        Federal Employees’ Retirement System Act of 1986
F Fund       Fixed Income Investment Fund
G Fund       Government Securities Investment Fund
I Fund       International Stock Index Investment Fund
S Fund       Small Capitalization Stock Index Investment Fund
SRI          socially responsible investment
TSP          Thrift Savings Plan

This is a work of the U.S. government and is not subject to copyright protection in the
United States. The published product may be reproduced and distributed in its entirety
without further permission from GAO. However, because this work may contain
copyrighted images or other material, permission from the copyright holder may be
necessary if you wish to reproduce this material separately.




Page iii                                                    GAO-12-664 Thrift Savings Plan
United States Government Accountability Office
Washington, DC 20548




                                   June 26, 2012

                                   The Honorable Gerald E. Connolly
                                   Ranking Member
                                   Subcommittee on Technology,
                                     Information Policy,
                                     Intergovernmental Relations
                                     and Procurement Reform
                                   Committee on Oversight
                                     and Government Reform
                                   House of Representatives

                                   Dear Mr. Connolly:

                                   Since its inception in 1986, the federal government’s Thrift Savings Plan
                                   (TSP) has become a central component of federal employees’ retirement
                                   savings. Intended to resemble 401(k) pension plans in the private sector,
                                   TSP allows employees to contribute a portion of their current
                                   compensation through payroll salary deductions and to allocate their
                                   contributions and any associated earnings among the available
                                   investment options. As of February 2012, TSP held about $308 billion in
                                   retirement assets and had more than 4.5 million participants, making it
                                   one of the largest retirement savings plans in the United States.

                                   The Federal Employees’ Retirement System Act of 1986 (FERSA)
                                   established TSP and the Federal Retirement Thrift Investment Board
                                   (Board), an independent agency in the executive branch, to administer it. 1
                                   The five-member, presidentially appointed Board must manage TSP
                                   prudently and solely in the interest of the participants and their
                                   beneficiaries and cannot exercise the voting rights associated with
                                   ownership of securities in TSP. FERSA also established three TSP
                                   investment funds: the Government Securities Investment Fund (G Fund),
                                   Common Stock Index Investment Fund (C Fund), and Fixed Income
                                   Investment Fund (F Fund). In 1996, Congress authorized TSP to broaden
                                   its offerings with the International Stock Index Investment Fund (I Fund)



                                   1
                                    5 U.S.C.§ 8472.




                                   Page 1                                           GAO-12-664 Thrift Savings Plan
and Small Capitalization Stock Index Investment Fund (S Fund). Since
August 2005, TSP participants also have had the option to choose one of
five lifecycle funds, which diversify participant accounts among the G, F,
C, S, and I Funds using investment mixes tailored for different retirement
time horizons.

Congress has proposed legislation to require TSP to offer additional
investment options in an effort to give participants and beneficiaries
greater investment choice. Several of these proposals have sought to
require TSP to include a responsible investment index fund that would
track a market index comprised of stocks that were prescreened for
corporate governance, environmental practices, workplace relations and
benefits, product safety and impact, international operations and human
rights, involvement with repressive regimes, and community relations. 2
Commonly referred to as socially responsible investment (SRI), this kind
of investment strategy seeks to generate long-term competitive financial
returns while realizing positive societal impact by investing in accordance
with one’s values or certain environmental, social, or corporate
governance criteria. Investment in SRI mutual funds—the most common
SRI vehicle—exceeded $300 billion in 2010. Congress asked us to
examine the value of adding an additional investment option to TSP in the
form of an SRI index fund. In response to this request, we examined the
following questions:

•   What challenges might TSP face in adopting an SRI option?

•   How would the addition of an index fund tracking an SRI index have
    affected past TSP stock fund portfolio performance?

•   How do the performance and costs of SRI mutual funds compare to
    non-SRI mutual funds?

To determine the challenges associated with SRI, we reviewed relevant
federal laws and regulations related to FERSA and TSP, and literature. We
also interviewed officials from TSP, the Employee Thrift Advisory Council
(ETAC), investment management or consultant firms, and a


2
 The Financial Industry Regulatory Authority defines index funds as passive funds that
seek to replicate the performance of a market index instead of outperforming it. Index
funds differ from mutual funds which are investment companies that pool money from
many investors and invest it based on specific investment goals.




Page 2                                                     GAO-12-664 Thrift Savings Plan
nonrepresentative sample of 15 public pension plans. To identify our
sample, we contacted plans that were signatories of the United Nations’
Principles for Responsible Investment and employed a snowball sampling
technique based on recommendations of interviewees. To assess how the
addition of an index fund tracking an SRI index would have affected past
TSP stock fund portfolio performance, we identified the best performing
U.S.-based SRI stock index based on 10-year risk-adjusted performance
data, and analyzed its potential impact on the TSP stock portfolio over the
past 20 years (1992 through 2011). 3 Because this analysis is strictly based
on past performance, this result does not guarantee or imply that the
addition of an SRI index would have the same effect on future TSP stock
fund portfolio performance. To determine how the performance and costs
of SRI mutual funds compared with those of non-SRI mutual funds, we
analyzed historical performance data for several time periods over the past
15 years (dating back from December 2011)—the longest time period for
which data were available—and cost data from fiscal year 2010. We
focused our analysis exclusively on three share classes of mutual funds—
institutional, front-load, and no-load—of U.S. domiciled open-end mutual
funds, which experts identified as the most common form of SRI funds. 4 To
better demonstrate the full range of performance and costs trends between
SRI and non-SRI mutual funds, we analyzed funds by subcategories such
as share class, and also ran regressions to control for factors such as fund
asset size and investment strategy.

We assessed the reliability of the quantitative data used in this
engagement by reviewing related documentation, interviewing
knowledgeable officials, reviewing related internal controls, comparing to
published data, and tracing a selection of data to source documentation.
Based on this evaluation, we determined these data were reliable for the
purposes of this report. Additional details regarding our methodology can
be found in appendix I.




3
 We selected the best performing SRI stock index from the five U.S.-based SRI stock
indices with at least a 10-year history that were active as of December 2011. (We
identified three other U.S.-based SRI stock indices that did not have a 10-year history.)
We did not run a similar analysis on an SRI bond index because no U.S-based SRI bond
indices exist.
4
 Morningstar, Inc., identified three share classes of mutual funds—institutional, front-load,
and no-load—as the most common share classes used by institutional investors. We refer
to these three share classes together as institutional grade mutual funds.




Page 3                                                       GAO-12-664 Thrift Savings Plan
                          We conducted this performance audit from July 2011 to June 2012 in
                          accordance with generally accepted government auditing standards.
                          Those standards require that we plan and perform the audit to obtain
                          sufficient, appropriate evidence to provide a reasonable basis for our
                          findings and conclusions based on our audit objectives. We believe that
                          the evidence obtained provides a reasonable basis for our findings and
                          conclusions based on our audit objectives.



Background
The Thrift Savings Plan   FERSA created TSP to provide options for retirement planning and
                          encourage personal retirement saving among the federal workforce. Most
                          federal workers are allowed to participate in TSP, which is available to
                          federal and postal employees, including members of Congress and
                          congressional employees and members of the uniformed services, and
                          members of the judicial branch. TSP is structured to allow eligible federal
                          employees to contribute a fixed percentage of their annual base pay or a
                          flat amount, subject to Internal Revenue Service limits, into an individual
                          tax-deferred account. Additionally, Federal Employees’ Retirement
                          System (FERS) participants are eligible for automatic 1-percent
                          contributions and limited matching contributions from the employing
                          federal agency. TSP provides federal (and in most cases, state) income
                          tax deferral on contributions and their related earnings, similar to those
                          offered by many private sector 401(k)-type pension plans. 5 As is typical in
                          defined contribution (DC) plans, TSP allows participants to manage their
                          accounts and conduct a variety of transactions similar to those available
                          to 401(k) participants, including reallocating contributions or account




                          5
                           FERSA created FERS. As part of FERS, TSP is also part of the current three-part
                          retirement system for federal employees, which includes Social Security benefits, the
                          basic defined benefit (DB) plan, and TSP. The Office of Personnel Management trains
                          retirement counselors about each part of the plan. Prior to FERS, most federal employees
                          were covered by the Civil Service Retirement System.




                          Page 4                                                    GAO-12-664 Thrift Savings Plan
balances, borrowing from the account, making withdrawals, or purchasing
annuities. 6

Administration of TSP falls under the purview of the Board, an agency
established by Congress under FERSA. The Board is composed of five
members appointed by the President, with the advice and consent of the
Senate. They are authorized to appoint the executive director who hires
additional personnel, and ETAC—a 15-member council that provides
advice to the Board and the executive director on the investment policies
and administration of TSP. 7 The Board is responsible for establishing
policies for the investment and management of TSP, as well as
administration of the plan. The executive director and Board staff are
responsible for implementing the Board’s policies and managing the day-
to-day operations of TSP, prescribing regulations to administer FERSA,
and other duties. The Board members and the executive director serve as
plan fiduciaries. 8 FERSA has other investment policy provisions, such as
who can exercise voting rights associated with the ownership of stocks
held by TSP. 9 For example, the Board and the executive director may not
exercise voting rights associated with the ownership of TSP securities.

The Board has less discretion than private sector plan sponsors in setting
investment policy because the investment options available to TSP
participants are outlined in federal law, whereas private sector plan



6
 Pension plans may be classified either as DB or as DC plans. DB plans promise to
provide, generally, a fixed level of monthly retirement income for life that is typically
determined by a formula based on particular factors specified by the plan, such as salary,
years of service, or age at retirement regardless of how the plan’s investments perform. In
contrast, benefits from DC plans will vary with the contributions to and the performance of
the investments in individual accounts, which may fluctuate in value.
7
5 U.S.C. § 8473.
8
 Similar to fiduciaries in private pension plan under the Employee Retirement Income
Security Act of 1974 (ERISA), TSP fiduciaries are the persons who have discretionary
control or authority over the management or administration of the plan, including
management of the plan’s assets. TSP fiduciaries are required to perform their
responsibilities in the interest of participants and beneficiaries for the exclusive purpose of
providing benefits to participants and their beneficiaries and defraying the reasonable
expenses of administering the TSP. Other fiduciary responsibilities include the duty to act
prudently and, to the extent permitted by law, to diversify the investment of the fund so as
to minimize the risk of large losses, unless under the circumstances it is clearly prudent
not to do so.
9
    5 U.S.C. § 8438(f).




Page 5                                                         GAO-12-664 Thrift Savings Plan
                                             sponsors have greater discretion in choosing which investment options to
                                             offer participants. 10 In addition, Congress must amend FERSA to approve
                                             a change in TSP investment options offered to participants. TSP’s
                                             authorizing statute specifies the number and types of funds available to
                                             participants, and requires that some of these funds track indexes, which
                                             are broad, diversified market indicators. The Board may select the
                                             particular indices for the funds to follow as well as review the investment
                                             options and suggest additional funds. The Board has developed
                                             investment policies for each TSP fund. These policies, which the Board
                                             reaffirms quarterly, provide the rationale for selecting the fund’s
                                             investments. Table 1 shows FERSA requirements and Board policies
                                             regarding each fund and its underlying index.

Table 1: Investment Requirements for Existing TSP Funds

Fund       FERSA investment requirements                                      Board investment policies
G Fund     Treasury securities specially issued to TSP with a                 Short-term securities that mature in 1 to 4 days
           maturity determined by the executive director that
           provide the generally higher interest rates of securities
           with a term of at least 4 years [5 U.S.C. § 8438(a)(4)].
F Fund     Fixed-income securities [5 U.S.C. § 8438(a)(3)].                   An index including bonds and asset-backed securities to track
                                                                              the Barclay’s Capital U.S. Aggregate Bond Index.
C Fund     A portfolio that tracks a broad index representing the             An index of stocks of large to medium-sized companies to track
           U.S. stock market [5 U.S.C. § 8438(a)(1)].                         the Standard & Poor’s 500 Index.
S Fund     A portfolio that tracks a broad index representing U.S.            An index of stocks in small and medium-sized companies not
           stocks not included in the C Fund [5 U.S.C. §                      represented in the Standard & Poor’s 500 Index to track the Dow
           8438(a)(3)(B)].                                                    Jones Total Stock Market Completion Index.
I Fund     A portfolio that tracks a broad index representing                 An index of the stock markets of the developed world outside of
           international stock markets outside of the United                  the United States and Canada to track the Morgan Stanley
           States [5 U.S.C. § 8438(a)(5)].                                    Capital International Europe, Australasia, and Far East Index.
                                             Source: GAO analysis of FERSA.



                                             Members of Congress have introduced bills calling for new investment
                                             options to be added to TSP. In the past four sessions of Congress, a
                                             number of bills have been proposed to add investment options to TSP,
                                             including a corporate responsibility stock index fund, a precious metals
                                             investment fund, a real estate stock index fund, and a terror-free
                                             international investment option. In addition, Congress passed the Federal
                                             Retirement Reform Act of 2009, which among other things, authorized


                                             10
                                               GAO, Federal Retirement Thrift Investment Board: Many Responsibilities and
                                             Investment Policies Set by Congress, GAO-07-611 (Washington, D.C.: June 21, 2007).




                                             Page 6                                                               GAO-12-664 Thrift Savings Plan
TSP to offer a service that would enable participants to invest in mutual
funds outside TSP, if the Board determined that such a mutual fund
window was in the best interests of participants. 11 The law stipulated that
the Board had to ensure that any expenses charged for use of the mutual
fund window would be borne solely by the participants who used it. The
Board has not implemented the mutual fund window. According to TSP
officials, both the Board and ETAC were similarly split on whether to
include a mutual fund window and the Board tabled the discussion of the
mutual fund window to address more immediate issues, such as adding a
Roth TSP option. 12 They also noted that while TSP has not moved
forward with adding a mutual fund window, it may at some future point. 13

FERSA also requires the Board to defray reasonable expenses of
administering TSP. 14 TSP’s administrative expenses include management
fees for each investment fund; the costs of operating and maintaining
TSP’s recordkeeping system; the cost of providing participant services;
and the printing and mailing of notices, statements, and publications. 15
These expenses are offset by (1) forfeitures of agency automatic (1
percent) contributions made to employees who participated in the FERS




11
  Pub.L.No. 111-31, Div.B, §104, 123 Stat.1852, 1854 (codified at 5 U.S.C.§ 8438(b)(1)).
12
  The Roth TSP, as authorized by the Thrift Savings Plan Enhancement Act of 2009
(enacted June 22, 2009), allows federal civilian employees and members of the uniformed
services to contribute after-tax dollars into the TSP for the first time. Both the contributions
and their earnings will be tax-free when withdrawn, as long as IRS requirements are met.
The Roth TSP option has been available to participants since May 2012.
13
   TSP’s 2008 participant survey found that 39 percent of respondents said TSP would be
a better program if it provided a self-directed mutual fund window, 24 percent said they
would transfer some of their TSP account balance, and 10 percent would be willing to pay
$100 annual fee to use the window. In a 2009 memorandum to the Board, the TSP
executive director recommended the addition of a mutual fund window as an appealing
feature to participants who sought more specialized or sophisticated TSP investments.
Moreover, the window would allow participants to invest in funds that better matched their
individualized risk tolerance or particular interests, such as SRI funds. According to the
Department of Labor, the Board would need to consider the application of FERSA’s
fiduciary standards to the design and operation of such a mutual fund window.
14
  5 U.S.C. § 8477(b)(1)(A)(ii).
15
  According to the Department of Labor, the ERISA fiduciary standards governing
investment decisions apply to an ERISA fiduciary’s selection of a “socially responsible”
mutual fund as a designated investment alternative under a private-sector pension plan.




Page 7                                                         GAO-12-664 Thrift Savings Plan
                       but left federal service before becoming vested, 16 (2) other forfeitures,
                       and (3) loan fees. Because these amounts are not sufficient to cover all of
                       TSP’s expenses, TSP participants share in the remainder of the costs.
                       The cost to participants to invest in TSP is measured as an expense ratio
                       of the total administrative expenses charged to a fund during a specific
                       time period, divided by that fund’s average balance for that specific time
                       period. In 2011, expenses charged to each TSP account were
                       approximately 25 cents per each $1,000 of investment, or 2.5 basis
                       points. TSP’s expense ratio typically falls below the average expense
                       ratio of other 401(k) type plans.


Socially Responsible   SRI—investment made on the basis of environmental, social, and
Investment             corporate governance (ESG) criteria—is a global phenomenon and is
                       growing in popularity in the United States. In 2006, the United Nations
                       issued Principles for Responsible Investment that maintained a belief that
                       ESG issues can affect the performance of investment portfolios and
                       therefore must be given appropriate consideration by investors if they are
                       to fulfill their fiduciary duty. 17 By supporting the principles, institutional
                       investors commit to better align investors with broader societal goals
                       while acting in the best long-term interests of their beneficiaries.
                       Specifically, signatories agreed to

                       •    incorporate ESG issues into investment analysis and decision-making
                            processes,

                       •    be active owners and incorporate ESG issues into their ownership
                            policies and practices,

                       •    seek appropriate disclosure on ESG issues by the entities in which
                            they invest,

                       •    promote acceptance and implementation of these principles within the
                            investment industry to affect the performance of investment,




                       16
                        Vesting means that the TSP provides that an employee’s right to a benefit is
                       nonforfeitable upon the attainment of the required period of service under the law.
                       17
                        For more information on these principles, see http://www.unpri.org, accessed May 24,
                       2012.




                       Page 8                                                      GAO-12-664 Thrift Savings Plan
                                       •   work together to enhance the effectiveness of the principles, and

                                       •   report on their activities and progress in implementing the principles.

                                       In 2012, there were more than 1,000 asset owners, investment
                                       managers, and professional service partners that had committed to these
                                       principles worldwide—136 of them in the United States, according to the
                                       United Nations’ website.

                                       According to a 2010 report by US SIF 18—a leading SRI advocacy group—
                                       U.S.-based, open-end mutual funds comprise the largest share of funds
                                       that incorporate ESG factors in the United States with holdings of more
                                       than $300 billion. Other investment vehicles that incorporate ESG factors
                                       include exchange traded funds, variable annuity products, and alternative
                                       investment funds. As shown in figure 1, the number of SRI mutual funds
                                       open to investors has grown steadily since 1990.

Figure 1: Cumulative Growth in the Number of U.S. Open-end SRI Mutual Funds, 1952 through 2011




                                       18
                                         US SIF was formerly known as the Social Investment Forum (SIF). US SIF is a U.S.
                                       membership association for professionals, firms, institutions, and organizations engaged
                                       in sustainable and responsible investing.




                                       Page 9                                                     GAO-12-664 Thrift Savings Plan
According to US SIF, there are three principal components of SRI:

•   Investment screening. Investment screening is a practice of
    evaluating investment portfolios or mutual funds based on ESG
    criteria. Screening may involve including only strong performers,
    avoiding poor performers, or otherwise incorporating ESG factors into
    the process of investment analysis and management. Some
    responsible investors may avoid investment in companies whose
    products and business practices are harmful to individuals,
    communities, or the environment. Others may apply SRI screens to
    invest in companies that are leaders in adopting clean technologies
    and exceptional social and governance practices.

•   Shareholder advocacy. Shareholder advocacy involves engaging
    companies directly on issues of concern through shareholder
    resolutions or proxy voting. Shareholder advocacy frequently involves
    the filing of shareholder resolutions to improve business practices.
    The resolutions are then voted on by all owners of a corporation. For
    example, shareholders can use resolutions to urge a company to
    improve its sustainability practices, such as by reducing its carbon
    emissions, or by improving its supply chain to ensure that labor laws
    are fully enforced. Similarly, proxy voting is the primary means by
    which shareholders are able to direct company management to act in
    a socially responsible manner. As partial owners of companies,
    shareholders have the right to weigh in on important issues through
    the process of proxy voting.

•   Community investing. Community investing directs capital from
    investors and lenders to communities that are underserved by
    traditional financial services institutions. Community investing
    provides access to credit, equity, capital, and basic banking products
    that these communities would otherwise lack.

According to experts, investors pursue SRI for a variety of reasons. Some
investors may invest solely in accordance with their values without
considering financial implications. Other investors may pursue SRI for
purely financial motives and seek out well managed companies that take
steps to manage their ESG risks in the belief that these companies will
outperform their competitors in the long run. SRI mutual funds offer a
broad range of investment screens, including negative screens, which
exclude industry sectors associated with certain products or processes;
restricted screens, which set a threshold for investment in companies
associated with certain products or processes; and positive screens,



Page 10                                          GAO-12-664 Thrift Savings Plan
which include industry sectors associated with certain products,
processes, or values, or invest in “best-in-class” companies that
demonstrate the strongest record of incorporating ESG criteria into their
business models. Figure 2 provides an overview of investment screens
applied by current SRI mutual funds in the United States.




Page 11                                          GAO-12-664 Thrift Savings Plan
Figure 2: Investment Screens Applied by SRI Mutual Funds




                                       Page 12             GAO-12-664 Thrift Savings Plan
Officials Identified
Several Challenges to
SRI Adoption and
Said the Long-term
Economic Benefits of
SRI Were Unknown

Challenges              Officials at TSP and the other public retirement plans that had considered
                        adding an SRI option associated a number of common challenges with
                        the implementation of SRI. While none of the plan officials that we
                        contacted had plans that were identical to TSP in terms of its federal
                        scope or participant demographics, many of them shared similar
                        challenges and concerns with TSP. 19 As shown in figure 3, participant
                        demand, SRI screening criteria, and costs were the most common
                        challenges identified by public retirement plans.




                        19
                          We included responses from officials from three public DB plans who, like TSP officials,
                        were responsible for managing plans with large asset holdings and large, diverse
                        participant groups. Although the fiduciary role of these officials differed significantly from
                        that of TSP officials in terms of their ability to pursue SRI, these officials identified other
                        challenges that were similar to the challenges that TSP officials identified.




                        Page 13                                                        GAO-12-664 Thrift Savings Plan
                     Figure 3: Number of Public Retirement Plans Identifying Challenges Associated
                     with SRI Adoption Similar to Ones Identified by TSP Officials




Participant Demand   TSP and most other public plan officials we contacted identified low
                     participant demand for SRI as a challenge to adopting SRI. TSP officials
                     told us that based on the results of their participant surveys and the
                     experiences of ETAC there was little demand for an SRI fund among TSP
                     participants. Specifically, they noted that the results of periodic participant
                     surveys have consistently indicated that there was no overwhelming
                     demand for any new investment options, including an SRI option. 20 In
                     addition, ETAC members told us that they were unaware of demand for
                     SRI among TSP participants. They said that they would respond if
                     demand ever presented itself. While consultants and fund managers that
                     we contacted reported a growing demand for SRI in the United States,
                     public plan officials that we spoke with generally reported low participant
                     interest in SRI adoption. Officials at several plans noted that continued
                     pressure and repeated demands from small vocal groups of participants
                     in support of SRI had been a principal driver in the plans’ decision to have
                     an SRI option. However, officials at several of these plans said that, while


                     20
                       TSP’s 2006-2007 participant survey found that none of the additional investment options
                     that participants were asked to rate, including a socially responsible fund, had a majority
                     of respondents who were willing to add the funds if there would be additional costs
                     associated with utilizing those funds.




                     Page 14                                                     GAO-12-664 Thrift Savings Plan
                            the SRI option did attract a small percentage of participants, overall
                            participation in the SRI funds ranged from less than 0.5 percent (in one
                            plan with 20 investment options) to about 10 percent (in a plan that
                            offered 9 investment options). 21

Selection of SRI criteria   TSP and most other public plan officials we contacted identified the
                            difficulty of finding broadly acceptable SRI criteria as a distinct challenge to
                            adopting SRI. According to TSP officials, different interpretations of what
                            social criteria to apply to an SRI fund could lead to the need to develop
                            multiple funds to satisfy participants. Officials also noted that it would be
                            hard to reach agreement on what values an SRI fund should endorse.
                            Moreover, officials at most of the other plans we contacted said that the
                            lack of a common definition of SRI and the selection of SRI criteria was
                            challenging. For example, officials at one plan noted that social issues were
                            difficult to incorporate into an investment approach because, while some
                            basic social issues, such as child labor, imprisonment, and forced slavery,
                            were generally acceptable screens, reaching broad consensus on other
                            issues, such as labor laws, workers’ rights, weapons, guns, and tobacco,
                            was more difficult. An official at another plan noted that it was incumbent
                            upon participants to tell them what social policy they wished to pursue.
                            Officials at the plans we contacted that considered selection of screening
                            criteria to be a challenge overcame this challenge by either using an off-the
                            shelf SRI fund, or relying on the expertise and experience of the SRI fund
                            managers, or educating participants on why fund managers selected the
                            investments they did.

Costs                       TSP and other public plan officials we contacted had varied opinions on
                            the degree to which the costs associated with the creation and
                            administration of an SRI index fund presented a challenge to adopting
                            SRI. According to TSP officials, the costs to create a new index fund
                            would be considerable. In addition, they said an SRI index fund would
                            cost more because it requires additional screening and monitoring. Under
                            TSP’s current cost structure, any costs associated with a new SRI index
                            fund would be borne by all participants whether or not they chose to



                            21
                              This range of participation rates in SRI funds is not a reflection on SRI as a viable
                            investment option. For example, plans with more than 10 investment options could have
                            10 percent of participants investing in each option, including the SRI option. In contrast,
                            plans like TSP with fewer investment options might consider a 10-percent SRI
                            participation rate low when compared to the participation rates in the plan’s other
                            investment options.




                            Page 15                                                      GAO-12-664 Thrift Savings Plan
                   invest in the SRI index fund. 22 Other public plan officials we contacted
                   had varied opinions on the degree to which these costs presented a
                   challenge. While there would be certain upfront costs associated with
                   adding an SRI fund, which could include member communication and
                   manager selection, officials at several plans said that adding a new fund
                   to its existing portfolio would not adversely affect administrative costs.
                   According to some investment managers we contacted, the key factors
                   affecting cost of any fund are (1) its asset size—the larger the asset base,
                   the better the economies of scale and the lower the overall cost ratio—
                   and (2) whether the fund’s investment strategy requires active
                   management or passively tracking of a market index.

Fiduciary Issues   While TSP and some public plan officials we contacted asserted that their
                   role as fiduciary was a challenge in that it precluded the adoption of SRI,
                   officials at other public plans with an SRI option said there were no
                   fiduciary concerns surrounding the inclusion of an SRI option in a DC plan.
                   According to a 1990 memorandum sent from the TSP executive director to
                   the Board, Congress considered and rejected the concept of social
                   investment when creating TSP. The memorandum noted that the strict
                   fiduciary provisions of the law, which require the Board to discharge its
                   responsibilities solely in the interest of participants, excluded the possibility
                   of social investing, and that any authorization outside the realm of interest
                   of all participants would be inconsistent with the notion of employee
                   ownership of TSP assets. Officials at all of the public plans that had not
                   implemented an SRI option considered fiduciary issues a challenge to
                   adopting SRI, while officials at other plans did not. For example, an official
                   at one plan with an SRI option stated that fiduciary duty was not a
                   challenge when adding an SRI fund to the investment options to a DC plan
                   because participants have individual account ownership, are free to choose
                   how they invest, and must assume responsibility for any risks associated
                   with the underlying investments. One plan official noted that the fiduciary
                   responsibility of a DC plan extends to exercising due diligence in the
                   selection of a fund manager, providing appropriate participant
                   communications about the fund, offering enough investment options, and
                   acting in the best interest of the majority of participants.


                   22
                     According to TSP regulations, the amount of accrued administrative expenses not
                   covered by forfeitures (agency contributions made to employees who leave federal service
                   before vesting) and not directly attributable to an individual fund (such as fund
                   management fees) are charged on a pro rata basis to all TSP funds, based on the
                   respective fund balances on the last business day of the prior month end.




                   Page 16                                                   GAO-12-664 Thrift Savings Plan
Concerns about Political   TSP and some public plan officials that we contacted asserted that
Interference               investment decisions made on any basis other than the economic welfare
                           of participants could present a challenge in that it would expose the plan
                           to potential political interference. In its 2006 investment option review, a
                           consulting firm hired by TSP eliminated SRI from consideration in part on
                           the grounds that identifying screening criteria that all could agree to would
                           be difficult to find and likely draw attention from opposing parties of
                           interest. 23 According to the 1990 memorandum from the TSP executive
                           director, the laws that established the current TSP funds prevent the
                           political manipulation of TSP funds, and officials told us that TSP has
                           taken steps in the past to avoid political interference. Officials at the
                           public plans we contacted had different views on the extent to which
                           political interference was a challenge. For example, officials at some
                           public plans that did not implement SRI identified political interference as
                           one of the reasons they chose not to do so. On the other hand, officials at
                           other public plans that had implemented SRI said that political
                           interference was not a challenge. For example, officials at one public plan
                           noted that the state’s legislative mandate to maximize returns and
                           improve levels of risk prevented political interference. An official at
                           another public plan that adopted SRI told us that although they had
                           anticipated political interference by state officials following their decision
                           to divest from alcohol or tobacco companies, it had not materialized.

SRI Performance            Officials at TSP and some other public plans identified SRI fund
                           performance as a challenge to adopting an SRI fund. According to TSP
                           officials, participants who allocated assets to an SRI fund instead of a
                           standard fund that included all relevant companies would narrow the
                           number of companies in which they were indirectly investing, thereby
                           limiting their exposure to the performance of the broader, more diversified
                           market. While officials at some public plans that we contacted believed
                           that SRI funds had lower performance than other funds, other officials
                           had mixed views on whether the performance of SRI funds was any more
                           challenging than the performance of non-SRI funds. Officials at several
                           plans, which had considered but not implemented SRI, cited SRI
                           performance as a reason for not incorporating SRI in their plans. Officials
                           at one plan said they would reconsider offering an SRI fund if it
                           demonstrated better long-term performance than non-SRI funds.


                           23
                             See “Investment Option Review for the Federal Thrift Savings Plan,” Ennis Knupp and
                           Associates, Oct. 2006. The report is accessible on the Board’s website:
                           http://www.frtib.gov/FOIA/index.html, accessed June 15, 2012.




                           Page 17                                                  GAO-12-664 Thrift Savings Plan
                     However, officials at several other plans that implemented SRI told us
                     that the SRI fund produced comparable and sometimes better returns
                     than other funds in their portfolio. Officials at one plan said the plan would
                     terminate its working relationship with an external fund manager if its SRI
                     investments did not perform as well as other funds.

Peer Practice        While TSP officials considered the lack of peer implementation of SRI as
                     a challenge to adopting an SRI fund, officials at other public plans we
                     contacted said that it was not a challenge. As part of its investment
                     options review in 2006, a consulting firm advised TSP that SRI funds
                     were not a common practice among TSP’s peers and identified this
                     criterion as a reason for eliminating SRI from further consideration.
                     According to TSP officials, the fact that similar plans had not adopted SRI
                     was a challenge in that TSP had no precedent to follow. We found a
                     number plans similar in asset size and membership to TSP that applied
                     SRI principles through investment screening. Officials at several plans we
                     contacted said that peer implementation of SRI did not factor in their
                     decision to incorporate SRI into their investment strategy.

Investment Overlap   Officials at most of the public plans we contacted had no restrictions
                     regarding investment overlap between funds and thus did not view such
                     overlap as a challenge to adopting SRI. According to TSP officials, the
                     Board is permitted to suggest legislation to address any gaps in
                     investment options as long as there is no evident overlap. In the past, for
                     example, TSP proposed legislation authorizing the addition of the S Fund
                     and the I Fund to provide participants with options for greater
                     diversification of investments in the small capitalization and international
                     markets. According to TSP officials, each of the current funds tracks
                     different companies in different segments of the overall financial market
                     without overlap, helping to reduce the risk of incurring large losses on a
                     broader portfolio. Officials at other public plans, which did not face the
                     same restrictions as TSP, said that overlap was not a consideration, and
                     that certain amount of overlap with existing investments was both
                     expected and accepted. Some officials noted that the purpose of SRI was
                     to select companies that met certain criteria and provided an alternative
                     investment choice.


Benefits             Officials at some of the nine public plans we contacted that offered an
                     SRI option cited some short-term benefits associated with SRI but said
                     that the long-term benefits were unknown. For example, officials at
                     several plans noted that the greatest benefit of having an SRI option is
                     giving participants broader investment choice and an opportunity to make


                     Page 18                                            GAO-12-664 Thrift Savings Plan
                        a statement in the way they invest. Officials at other plans said that
                        having an SRI option could serve as a recruiting tool for the plan in that it
                        encouraged more eligible employees to join the plan. Regarding the long-
                        term benefits of SRI, officials at two of the public plans stated that it was
                        still too early to judge the benefit of SRI. As one plan official noted,
                        responsible investment involves making investment decisions that are
                        important to the long-term value and profitability of a company over time.


                        When compared to the past performance of the TSP stock portfolio (the
Adding an SRI Index     C, S, and I Funds), the addition of a hypothetical SRI index fund tracking
Fund Would Not Have     the best-performing U.S.-based SRI stock index would not have both
                        increased returns and lowered volatility in any allocation scenario we
Improved Past TSP       tested. Specifically, over the last 20 years, if TSP had included such an
Portfolio Performance   SRI index fund (SRI Fund) in its existing stock portfolio, 24 it could have
in Most Allocation      resulted in (1) lower returns and lower volatility, (2) lower returns and
                        higher volatility, or (3) higher returns and higher volatility, based on our
Scenarios               analysis of evenly distributed portfolio allocations containing the SRI Fund
                        against the TSP stock portfolio alone (a C, S, and I Funds combination). 25
                        For example, as shown in figure 4, adding the SRI Fund to the existing
                        TSP stock funds (an SRI, C, S, and I Funds combination) would have
                        resulted in lower returns and lower volatility; substituting the SRI Fund for
                        the C Fund (an SRI, S, and I Funds combination) would have resulted in
                        lower returns and higher volatility; and substituting the SRI Fund for the I
                        Fund (an SRI, C, and S Funds combination) would have resulted in
                        higher returns and higher volatility. 26 Because this analysis is strictly
                        based on past performance, this result does not guarantee or imply that




                        24
                          TSP’s stock portfolio includes TSP’s three stock index funds—C fund, S fund, and I
                        fund. We did not include a similar analysis of bond funds because no U.S.-based SRI
                        bond indices exist.
                        25
                           For example, four-way combinations allocate 25 percent to each fund and three-way
                        combinations allocate 33 percent to each fund. Portfolios were rebalanced annually to
                        maintain even distributions. In this section, returns are measured by the compound rate of
                        return and volatility is measured by standard deviation of annual returns. Standard
                        deviation is a statistical measure of the range of a fund’s performance. When a fund has a
                        high standard deviation, its range of performance has been very wide, indicating that there
                        is a greater potential for volatility.
                        26
                          For more information on the annual compound rates of return and standard deviation of
                        these funds, see appendix II.




                        Page 19                                                     GAO-12-664 Thrift Savings Plan
                                        the addition of an SRI index would have the same effect on future TSP
                                        stock fund portfolio performance.

Figure 4: Effect of Adding a Hypothetical TSP SRI Index Fund on Past TSP Stock Portfolio Performance, 1992 to 2011




                                        Page 20                                                GAO-12-664 Thrift Savings Plan
Notes
1. We used a time-weighted basis for our analysis and rebalanced portfolios annually to maintain the
evenly distributed allocations. A time-weighted basis gives equal weight to each unit of time; thus, the
annual rate of return in 1992 gets just as much weight in the analysis as the annual rate of return in
2011. We used a time-weighted analysis in order to focus on investment performance itself, rather
than on the particular economic consequences in the time period under study. For a TSP participant,
the overall compound rate of return would be affected by interim cash flow into and out of the plan.
For example, for a TSP participant who made regular contributions to the plan during the 1992 to
2011 period, the overall rate of return would be more influenced by particular performance in the later
years, when more contributions are at stake, than in the earlier years.
2. In this figure, returns represent annualized compound return (i.e., geometric average annual
return), while volatility represents standard deviation of annual return.


Overall, portfolio performance is directly tied to the individual fund
performance, which varied by time period. A comparison of the underlying
indices of these four funds shows that, while the SRI Fund had higher
cumulative returns than the I Fund over the last 20 years, it had lower
cumulative returns than all three of the TSP funds over the last 10 years.
Figure 5 shows the funds’ annual and cumulative returns and highlights
their performance during market cycles.




Page 21                                                              GAO-12-664 Thrift Savings Plan
Figure 5: Annual and Cumulative Returns of the Indices Underlying the C, S, I, and SRI Funds, 1992 to 2011




                                         The managers of the SRI index explained the difference in the index’s
                                         performance over the last 20 years in comparison with the Standard &
                                         Poor’s 500 Index (the C Fund index) was a result of having different
                                         sector weightings than the overall market to align with the fund’s SRI
                                         strategy. For example, they told us that, in the late 1990s, the index was
                                         relatively overweighted in technology, consumer, and finance stocks and
                                         underweighted in energy and utilities, resulting in higher performance in
                                         the “dot com” boom of the late 1990s and lower performance in the 2001
                                         recession. Moreover, the SRI index also excludes companies involved in
                                         the production of military weapons, which may have contributed to lower
                                         returns over the past decade while the country has been at war.




                                         Page 22                                                GAO-12-664 Thrift Savings Plan
In addition to providing less return overall than the C Fund over the 20-
year period, the inclusion of this SRI Fund would have resulted in overlap
with the C Fund and not have provided a substantial opportunity for
additional portfolio diversification. 27 By law, holdings in TSP stock funds
may not overlap. Fifty-seven percent of the companies included in the
SRI Fund index, which includes large, mid, and small capitalization
stocks, overlap with companies included in the C Fund index. 28 In part, as
a result of this overlap, the SRI Fund and the C Fund are highly
correlated in their returns, and thus adding this SRI Fund would not
provide a substantial opportunity for additional portfolio diversification.
Portfolio diversification aims to reduce risk by investing in various
financial instruments and markets so that market events will not affect all
assets in the same way. Diversification opportunities exist if investments
have independent price movement, and therefore, independent returns.
The price movement between these two funds over the last 20 years was
1.94 percent independent, suggesting that the same external causes
affected their returns to nearly the same degree. 29 By contrast, over the
same time period, the independence in price movement between the S
Fund and C Fund was 17.27 percent, and between the I Fund and C
Fund was 42.19 percent.




27
  Investors may pursue SRI for reasons other than portfolio diversification, including
directing their retirement savings to socially responsible companies.
28
  Other SRI indices may have had more or less overlap with the C Fund index.
29
  Price movements between two funds can be between 0 and 100 percent independent.
The independence of price movements is measured by 100 percent independence of
price movements less the common variance. The common variance is a statistical
measure that measures independence of price movements, and is the square of the
correlation coefficient. The correlation coefficient of the two funds was 0.99 over this time
period. Correlation coefficient is a statistical measure that falls between -1.0 and +1.0. A
coefficient of +1.0 means that the two indices move in the same direction at the same time
in the same (relative) amounts.




Page 23                                                      GAO-12-664 Thrift Savings Plan
SRI Mutual Fund
Performance Varied
by Asset Class, but
Had Similar Costs to
Non-SRI Mutual
Funds

After Controlling for Fund   Looking more broadly at SRI mutual funds, the most common form of SRI
Size and Strategy, SRI       in the United States, we found that the comparative performance of SRI
Bond Mutual Funds and        and non-SRI mutual funds over the last 15 years varied by asset class.
                             While TSP participants cannot currently invest in mutual funds through
SRI Stock Mutual Funds       TSP, the Board is authorized to offer a mutual fund window if it
Outperformed Their Non-      determines that it is in the best interests of participants. Specifically, our
SRI Counterparts             analysis of institutional-grade mutual funds 30 over the last 5, 10, and 15
                             years (dating back from December 2011) found that SRI bond mutual
                             funds had better risk-adjusted performance than their non-SRI
                             counterparts. In contrast, SRI stock funds and SRI balanced funds—
                             which hold bonds and stocks—had worse risk-adjusted performance than
                             their non-SRI counterparts over these time periods. 31 Because this
                             analysis is strictly based on past performance, these results do not


                             30
                               Morningstar identified three share classes of mutual funds—institutional, front-load and
                             no-load—as the most common share classes used by institutional investors. We refer to
                             these three share classes together as institutional grade mutual funds. We did not analyze
                             other forms of SRI funds such as variable annuities and exchange traded funds. Hence,
                             the results of our analysis are not generalizable to non-institutional grade mutual funds or
                             to other forms of SRI. We included three asset classes of mutual funds in our analysis,
                             stocks, bonds, and balanced. Stock funds invest in stocks, bond funds invest in bonds,
                             and balanced funds invest in a combination of stocks and bonds. Looking at institutional
                             grade share classes of mutual funds, we identified: (1) 21 SRI bond mutual funds, 118 SRI
                             stock mutual funds, and 30 SRI balanced mutual funds, and (2) 2,403 non-SRI bond
                             mutual funds, 3,579 non-SRI stock mutual funds, and 1,366 non-SRI balanced mutual
                             funds active as of December 2011.
                             31
                               In this section, risk-adjusted returns are measured by Sharpe and Sortino ratios. Sharpe
                             ratio is calculated by using standard deviation and excess return to determine reward per
                             unit of risk. The higher the Sharpe ratio is, the better the fund’s historical risk-adjusted
                             performance. The Sortino ratio differentiates harmful volatility from volatility in general by
                             using a value for downside deviation. The Sortino ratio is the excess return over the risk-
                             free rate divided by the downside semi-variance. For additional details on performance by
                             asset class, see appendix III.




                             Page 24                                                       GAO-12-664 Thrift Savings Plan
                          guarantee or imply that these asset classes would perform similarly in the
                          future.

                          After controlling for fund size and investment strategies (other than SRI
                          approaches), we found that the performance gap between the SRI and
                          non-SRI mutual funds narrowed significantly for stock funds but not for
                          balanced funds. Moreover, our regression estimates showed that SRI
                          stock mutual funds performed better than their non-SRI stock
                          counterparts in the 5- and 15-year timeframes, after controlling for
                          differences in asset size, share class, and investment strategies. 32 (See
                          appendix I for additional information on the regression analyses.)


SRI and Non-SRI Mutual    In fiscal year 2010, the costs of SRI institutional grade mutual funds were
Funds Had Similar Costs   similar to their non-SRI counterparts. It is important to note that our cost
                          analysis included only the most recent year of data available (fiscal year
                          2010) for three share classes of institutional grade mutual funds, and it
                          did not look at all SRI product types such as variable annuities or
                          exchange traded funds, which may have had higher costs. 33 In addition,
                          fiscal year 2010 cost data are not indicative of past or future costs. As
                          shown in figure 7, there was considerable overlap in the costs associated
                          with these funds, as measured by their annual net expense ratio—the
                          actual percentage of assets deducted each fiscal year for fund
                          expenses. 34




                          32
                             We ran regressions to control for fund size, share class, and investment strategy other
                          than SRI approaches. The investment strategy control variables were whether a fund is
                          actively managed or passively tracks an index, the broad investment category of a fund
                          based on portfolio statistics and composition (e.g., natural resources, real estate, or
                          financial), the more narrowly defined institutional investment category of a fund based on
                          portfolio statistics and composition (e.g., materials, domestic energy, or technology), and
                          the market capitalization and type of stock (value, blend, and growth).
                          33
                            We ran cost regressions to control for fund size, share class, and investment strategy
                          other than SRI approaches.
                          34
                            According to Morningstar, the annual net expense ratio reflects the actual fees charged
                          during a particular fiscal year. The expense ratio is expressed as the percentage of assets
                          deducted each fiscal year for fund expenses.




                          Page 25                                                      GAO-12-664 Thrift Savings Plan
Figure 6: Comparison of Annual Net Expense Ratios of SRI and Non-SRI Mutual Funds




                                       a
                                       The non-SRI maximum annual net expense ratio was 7.93 percent


                                       While non-SRI mutual funds had a broader range of costs than SRI mutual
                                       funds, the vast majority of SRI and non-SRI funds reported expense ratios
                                       from 0.12 to 1.81 percent. On average, the reported expense ratios for SRI
                                       mutual funds were 0.2 percentage points higher than non-SRI mutual
                                       funds. 35 When asset size and investment strategy were taken into account,
                                       SRI mutual fund cost ratios were estimated to be only 0.06 percentage
                                       points higher than non-SRI mutual fund cost ratios. 36 For additional details
                                       on our regression analysis on cost ratios, see appendix I.


                                       Adoption of an SRI index fund would present challenges for TSP.
Concluding                             Currently, the law limits the types of funds that TSP can offer, prohibits
Observations                           overlap among existing funds, and charges TSP to keep its costs low.
                                       First, TSP would have difficulty finding an SRI index fund that did not
                                       overlap with the existing TSP funds, limiting opportunities for additional
                                       portfolio diversification. However, officials at other DC plans, which do not
                                       face the same restrictions as TSP, said that a certain amount of overlap
                                       with SRI and other investment options was acceptable and the purpose of
                                       SRI was to provide an alternative investment choice. Second, TSP would
                                       have difficulty selecting SRI screening criteria that all participants and the



                                       35
                                         In fiscal year 2010, SRI mutual funds had an average net-expense ratio of 1.08 percent
                                       and non-SRI mutual had an average net-expense ratio of 0.88 percent.
                                       36
                                         We ran cost regressions to control for fund size, share class, and investment strategy
                                       other than SRI approaches.




                                       Page 26                                                     GAO-12-664 Thrift Savings Plan
                  Congress would find acceptable. While challenging, a number of plans
                  have a long history of SRI in their plans. Finally, under TSP’s current
                  structure, the costs of adding a new fund would be distributed among all
                  participants regardless of whether they participated in that fund. We note
                  that the Board has the authority to open a mutual fund window for
                  participants to invest in mutual funds managed outside TSP. If the Board
                  decides to act on this authority and allow the mutual fund window,
                  participants seeking other forms of investment, including SRI, could invest
                  in mutual funds and would bear the costs associated with this investment.


                  We provided a copy of this draft report to the Federal Retirement Thrift
Agency Comments   Investment Board, the Department of Labor, and the Department of the
                  Treasury for review and comment. None of the agencies provided formal
                  comments. The Department of Labor provided technical comments, which
                  we incorporated in the report, as appropriate.


                  As agreed with your office, unless you publicly announce its contents
                  earlier, we plan no further distribution of this report until 30 days from its
                  issue date. At that time, we will send copies of this report to relevant
                  congressional committees and other interested parties. In addition, this
                  report will be available at no charge on GAO’s website at
                  http://www.gao.gov.

                  If you or your staff have any questions about this report, please contact
                  me at (202) 512-7215 or jeszeckc@gao.gov. Contact points for our
                  Offices of Congressional Relations and Public Affairs may be found on
                  the last page of this report. GAO staff making key contributions to this
                  report is listed in appendix IV.

                  Sincerely yours,




                  Charles A. Jeszeck
                  Director
                  Education, Workforce,
                    and Income Security Issues



                  Page 27                                              GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
              Appendix I: Objectives, Scope, and
              Methodology



Methodology

              To determine the challenges associated with socially responsible
              investment (SRI), we reviewed relevant federal laws, regulations, and
              literature. For example, we reviewed the Federal Employees’ Retirement
              System Act of 1986, the Federal Retirement Reform Act of 2009, and the
              2006 United Nations’ Principles for Responsible Investment. We also
              interviewed officials from the Thrift Savings Plan (TSP), the Employee Thrift
              Advisory Council, investment management and consultant firms, and 15
              selected public pension plans. Our nonrepresentative sample of pension
              plans included 9 domestic defined contribution (DC) and defined benefit
              (DB) plans that incorporated SRI, and 6 plans that considered but did not
              adopt an SRI component. To identify our sample, we contacted plans that
              were signatories of the United Nations’ Principles for Responsible
              Investment and employed a snowball sampling technique based on
              recommendations of interviewees. We analyzed interview responses of
              pension plan officials and other SRI experts on the challenges and benefits
              associated with SRI, and how these experiences might affect TSP.

              To determine how the addition of an SRI index fund to a TSP stock portfolio
              would have affected past TSP stock portfolio performance, we identified the
              best performing U.S.-based SRI index and assessed its potential impact on
              the TSP stock portfolio based on historical performance data of the three
              TSP stock fund underlying indices. 1 To identify the best performing U.S.-
              based SRI stock index (SRI Fund) we (1) identified all U.S.-based SRI stock
              indices with at least a 10-year history, and (2) selected the index with the
              best 10-year Sharpe ratio (dating back from December 2011). 2 To determine
              how the SRI Fund would have affected TSP stock portfolio performance from
              1992 to 2011, we analyzed monthly total return data for the SRI Fund and
              the underlying indices of the three TSP stock funds provided by Morningstar,
              Inc.—a leading independent financial market research firm. We used these
              data to analyze changes in annual returns and volatility, in a manner similar
              to past analysis conducted by TSP when considering whether to add funds


              1
               The Common Stock Index Investment Fund (C Fund) is indexed to the Standard & Poor’s
              500 Index. The Small Capitalization Stock Index Investment Fund (S Fund) is indexed to
              the Dow Jones U.S. Completion Total Stock Market Index. The International Stock Index
              Investment Fund (I Fund) is indexed to the Morgan Stanley Capital International Europe,
              Australasia, and Far East Index.
              2
               We selected the best performing SRI stock index from the five U.S.-based SRI stock
              indices with at least a 10-year history that were active as of December 2011. Sharpe ratio
              is calculated by using standard deviation and excess return to determine reward per unit
              of risk. As such, the higher the Sharpe ratio, the better the fund’s historical risk-adjusted
              performance.




              Page 28                                                       GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




to the TSP portfolio. An important element of any performance statistic is the
unit of time measurement. Our analysis measures returns on an annual
basis, and measures risk based on the variation in year-to-year returns.
Using a different unit of time, such as a month or even a multi-year period,
could give a different picture of the risk/reward tradeoff. We calculated the
compound rates of return and standard deviation3 based on annual rates of
return from 1992 to 2011 for an annually rebalanced, evenly distributed
portfolio of the three existing TSP stock fund indices (a distribution of 33
percent, 33 percent, and 33 percent). We then calculated the change in
compound rates of return and standard deviation of annual returns for the
following evenly distributed portfolios:

•   a four-way combination (25 percent, 25 percent, 25 percent, and 25
    percent) of the SRI Fund and the three TSP stock funds indices,

•   all of the three-way combinations (33 percent, 33 percent, and 33
    percent) of the SRI Fund with two of the TSP stock fund indices,

•   all two-way combinations (50 percent and 50 percent) of the SRI Fund
    with one of the TSP stock fund indices, and

•   the SRI Fund alone (100 percent).

Another decision in any performance assessment is whether to do the
analysis on a time-weighted or a dollar-weighted basis. A time-weighted
basis gives equal weight to each unit of time; thus, the annual rate of
return in 1992 gets just as much weight in the analysis as the annual rate
of return in 2011. A dollar-weighted basis gives greater weight to the
periods when more money is at stake. For example, for a TSP participant
who made regular contributions to the plan during the 1992 to 2011
period, the overall rate of return would be more influenced by particular
performance in the later years, when more contributions are at stake,
than in the earlier years. We used a time-weighted basis for our analysis,
in order to focus on investment performance itself, rather than on the
particular economic consequences in the time period under study.



3
 Specifically, we calculated annualized compound return (i.e., geometric average annual
return) and standard deviation of annual return. Standard deviation is a statistical measure
of the range of a fund’s performance. When a fund has a high standard deviation, its
range of performance has been very wide, indicating that there is a greater potential for
volatility.




Page 29                                                     GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




To further assess the performance of the SRI Fund, we compared annual
rates of return and compound cumulative rates of return for the three TSP
stock fund indices over various time periods. Specifically, we reviewed
performance over the 20-year period (1992 to 2011), the 10-year period
(2002 to 2011), and periods of market weakness. Because this analysis is
strictly based on past performance, this result does not guarantee or
imply that the addition of an SRI index would have the same effect on
future TSP stock fund portfolio performance. In addition, we analyzed the
overlap of holdings of the SRI Fund and the C Fund as of April 2012. To
analyze the diversification potential of the SRI Fund for the TSP stock
portfolio, we analyzed the correlation coefficient, common variance, and
independence of price movement between the SRI Fund and the C Fund
over the last 20 years. 4

To determine how the performance and cost of SRI mutual funds compare
with those of non-SRI mutual funds, we compared performance over the
past 15 years (1997 to 2011)— the longest time period for which data were
available—and costs as of fiscal year 2010 provided by Morningstar. To
identify the universe of SRI mutual funds, we included mutual funds from
Morningstar considered to be SRI mutual funds based on ethical screen
employed and data on SRI mutual funds maintained by US SIF. 5 To
analyze performance and cost of SRI and non-SRI mutual funds active as
of December 2011, we focused our analysis exclusively on three
institutional grade share classes—institutional, front-load and no-load—of
U.S. domiciled open-end mutual funds, which experts identified as the most
common form of SRI funds. We did not examine other forms of SRI, such
as exchange traded funds, hedge funds, or variable annuities. Because this
analysis is strictly based on past performance, these results do not
guarantee or imply that these asset classes would perform similarly in the
future. Performance statistics include measures of risk-adjusted returns
over 5-, 10-, and 15-year time periods dating back from December 2011.



4
 Correlation coefficient is a statistical measure that falls between -1.0 and +1.0. A
coefficient of +1.0 means that the two indices move in the same direction at the same time
in the same (relative) amounts. The independence of price movements is measured by
100 percent independence of price movements less the common variance. The common
variance is a statistical measure that measures independence of price movements, and is
the square of the correlation coefficient.
5
 US SIF was formerly the Social Investment Forum (SIF). US SIF is a U.S. membership
association for professionals, firms, institutions and organizations engaged in sustainable
and responsible investing.




Page 30                                                     GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




Some mutual funds had more recent inception dates, thus limiting the
number of funds in longer-term performance comparisons. Risk-adjusted
return statistics include the Sharpe and Sortino ratios. Cost measures
include fiscal year 2010 annual report net-expense ratios. 6

To investigate why SRI and non-SRI mutual funds differed in
performance, we ran regressions with and without controls for fund size,
share class, and investment strategy not inherently related to SRI. We
used the risk-adjusted performance measures, Sharpe and Sortino ratios,
as the dependent variables. Table 2 summarizes the results of 24
regressions for U.S. stock funds. The numbers in the columns labeled
“Outcome Variable 1” and “Outcome Variable 2” are coefficient estimates
on a flag indicating that a fund was non-SRI. 7

Table 2: Regression Coefficients on Non-SRI Indicator: U.S. Stock Funds,
Institutional, Front-load, and No-load Share Classes

                                                                                 Average
                                                                             percentage
                                                                               change in
                                         Outcome              Outcome            non-SRI
Time frame and control                  variable 1,          variable 2,      coefficient
variables                              Sharpe ratio        Sortino ratio    with controls
5-year, without control variables              0.014               0.017
    with fund size                             0.001              -0.001           -100%
    with strategy categories                   0.009               0.010              -37
    with all controls                         -0.004              -0.008             -139
10-year, without control                       0.077               0.110
variables
    with fund size                             0.054               0.077              -30
    with strategy categories                   0.044               0.061              -44
    with all controls                           0.02               0.027              -75
15-year, without control                       0.026               0.037
variables




6
 Annual report expense ratios reflect the actual fees charged during a particular fiscal
year. The expense ratio expresses the percentage of assets deducted each fiscal year for
fund expenses.
7
 We do not report standard errors of estimates here because our results are based on the
universe of mutual funds that fit our criteria, rather than a sample. Thus standard errors
are not applicable.




Page 31                                                    GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




                                                                                             Average
                                                                                         percentage
                                                                                           change in
                                                           Outcome         Outcome           non-SRI
 Time frame and control                                   variable 1,     variable 2,     coefficient
 variables                                               Sharpe ratio   Sortino ratio   with controls
    with fund size                                             0.002           0.002              -95
    with strategy categories                                   0.009           0.011              -68
    with all controls                                          -0.015         -0.024             -161


 Average change in coefficient across all time periods with inclusion of                          -75
 fund size categories
 Average change in coefficient across all time periods with inclusion of                          -50
 fund strategy control variables
 Average change in coefficient across all time periods with inclusion of                         -125
 full set of controls
Source: GAO analysis of data received from Morningstar, Inc.



As shown above, the Sharpe ratio served as the outcome variable for the
first set of regressions (second column of the table). The Sortino ratio
served as the outcome variable for the second set of regressions (third
column of the table). The last column shows the average impact on the
non-SRI flag coefficient of inclusion of the control variables for the row. For
each outcome variable, inclusion of the control variables generally reduced
the estimated performance premium of non-SRI funds versus SRI funds.
The fund size and strategy variable sets both had substantial impacts on
the estimated difference in SRI and non-SRI fund performance.

The rows of the table show regression results for the 5-, 10-, and 15-year
time frames. The rows indicating no control variables included only the
non-SRI flag. The rows indicating “with fund size” included fund size,
along with the SRI variable as explanatory variables. 8 The row indicating
“with strategy categories” include controls for whether a fund is actively
managed or passively tracks an index, the broad investment category of a
fund based on portfolio statistics and composition (e.g., natural resources,
real estate, or financial), the more narrowly defined institutional
investment category of a fund based on portfolio statistics and


8
 The size intervals were determined using an SPSS Decision Tree algorithm. The algorithm
found six size categories which differed significantly from each other in the ratio of SRI to
non-SRI funds. Breaking fund size into categories allowed including funds without fund size
data in the analysis, because the missing data were grouped into a category.




Page 32                                                                 GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




composition (e.g., materials, domestic energy, technology, or utilities),
and the market capitalization and type of stock (value, blend, and growth).
The row indicating “with all controls” provides results for regressions with
share class as an explanatory variable in addition to the fund size and
strategy variables. 9

Table 3 shows results from 24 regressions for balanced funds. The
methodology for these regressions was the same as that used for the
regressions in table 2. Accounting for covariates did not have a consistent
impact on the estimated difference in performance between SRI and non-
SRI funds for these funds, with the addition of control variables to the
regressions sometimes increasing the estimated difference in
performance and sometimes decreasing it.

Table 3: Regression Coefficients on Non-SRI Indicator: U.S. Balanced Funds,
Institutional, Front-load, and No-load Share Classes

                                                                             Average
                                                                          percentage
                                          Outcome            Outcome change in non-
Time frame and control                   variable 1         variable 2 SRI coefficient
variables                              Sharpe ratio       Sortino ratio with controls
5-year, without control variables              0.057              0.082
    with fund size                             0.040              0.060                -28%
    with strategy categories                   0.078              0.112                  37
    with all controls                          0.063               0.09                  10
10-year, without control variables             0.164              0.241
    with fund size                             0.130              0.191                 -21
    with strategy categories                   0.148              0.217                  -1
    with all controls                          0.125              0.185                 -23
15-year, without control variables             0.111              0.166
    with fund size                             0.056              0.083                 -50
    with strategy categories                   0.094              0.139                 -16
    with all controls                          0.063              0.093                 -43


Average change in coefficient across all time periods with inclusion of                 -33
fund size categories



9
 We did not include standard deviation as an explanatory variable in the regressions
because volatility is accounted for in the risk-adjusted performance measures.




Page 33                                                     GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




                                                                                            Average
                                                                                         percentage
                                                             Outcome        Outcome change in non-
 Time frame and control                                     variable 1     variable 2 SRI coefficient
 variables                                                Sharpe ratio   Sortino ratio with controls
 Average change in coefficient across all time periods with inclusion of                              4
 fund strategy control variables
 Average change in coefficient across all time periods with inclusion of                            -19
 full set of controls
Source: GAO analysis of data received from Morningstar, Inc.



Table 4 shows results for 24 regressions for bond funds. The
methodology for these regressions was the same as that used for the
regressions in table 2 except that one field (equity style box) was not
included as an explanatory variable because it was not populated for 90
percent of these funds. Accounting for covariates generally decreased the
estimated difference in performance between the SRI and non-SRI funds.

Table 4: Regression Coefficients on Non-SRI Indicator: U.S. Bond Funds,
Institutional, Front-load, and No-load Share Classes

                                                                                            Average
                                                                                         percentage
                                                             Outcome        Outcome change in non-
 Time frame and control                                     variable 1     variable 2 SRI coefficient
 variables                                                Sharpe ratio   Sortino ratio with controls
 5-year, without control variables                              -0.264         -0.452
  with fund size                                                -0.277         -0.486              -6%
  with strategy categories                                      -0.028          0.055              101
  with all controls                                             -0.028          0.037                99
 10-year, without control variables                             -0.154         -0.222
  with fund size                                                -0.159         -0.235                -5
  with strategy categories                                      -0.006          0.047              109
  with all controls                                             -0.005          0.045              109
 15-year, without control variables                             -0.252         -0.380
  with fund size                                                -0.250         -0.375                 1
  with strategy categories                                      -0.071         -0.085                75
  with all controls                                             -0.065         -0.072                78


 Average change in coefficient                                                                       -3
 across all time periods with
 inclusion of fund size variables




Page 34                                                                   GAO-12-664 Thrift Savings Plan
Appendix I: Objectives, Scope, and
Methodology




                                                                                                        Average
                                                                                                     percentage
                                                              Outcome                   Outcome change in non-
 Time frame and control                                      variable 1                variable 2 SRI coefficient
 variables                                                 Sharpe ratio              Sortino ratio with controls
 Average change in coefficient                                                                                        95
 across all time periods with
 inclusion of fund strategy control
 variables
 Average change in coefficient                                                                                        95
 across all time periods with
 inclusion of full set of controls
Source: GAO analysis of data received from Morningstar, Inc.



To investigate disparities between SRI and non-SRI mutual fund costs,
we ran regressions that controlled for fund size, share class, and
investment strategy. The coefficient for a flag indicating SRI status is
reported for four regressions in table 5. The regression reported in the
third column used fund size categories along with the SRI variable as
explanatory variables. The strategy variables used in the regressions
reported in the fourth and fifth columns were the same as those used for
the performance regressions reported above. Once the fund size and
investment strategy variables are taken into account, the estimated
difference in cost between SRI and non-SRI funds falls to 0.06 percent.

Table 5: Coefficient on SRI Indicator in Regression with Annual Report Net Expense
Ratio as Dependent Variable With and Without Control Variables

                                                                                   Control variables
                                                                                                           Strategy share
                                            No control             Fund size             Strategy         class, and fund
                                             variables            categories            variables         size categories
 Estimated percentage                                 0.20                  0.15                0.12                0.06
 point difference in cost
 between SRI and non-
 SRI mutual funds
Source: GAO analysis of mutual fund annual report net-expense ratio data received from Morningstar Inc.



We assessed the reliability of the quantitative data used in this engagement
provided by Morningstar by reviewing related documentation, interviewing
knowledgeable officials, reviewing related internal controls, comparing to
published data, and tracing a selection of data to source documentation.
Based on this evaluation, we determined these data were reliable for the
purposes of this report. We supplemented our quantitative analysis with
qualitative data obtained from our interviews.



Page 35                                                                                GAO-12-664 Thrift Savings Plan
Appendix II: Variation of Annual Compound Rates of
                  Appendix II: Variation of Annual Compound
                  Rates of Return and Standard Deviation among
                  Evenly Distributed Allocations of the Best
Return and Standard Deviation among Evenly Distributed
                  Performing SRI Stock Index and Existing TSP
                  Stock Index Funds, 1992 through 2011

Allocations of the Best Performing SRI Stock Index and
Existing TSP Stock Index Funds, 1992 through 2011

                                                                    Change in compound rate                            Change in standard
                                                                      of return from baseline                      deviation from baseline
                   Allocations                                         (percentage per year)                          (percentage overall)
                   Existing TSP Funds (C, S, I)                                                        0.00                          0.00
                   (baseline)
                   SRI ,C, S, I                                                                       -0.18                          -0.24
                   SRI, C, I                                                                          -0.75                          -0.57
                   SRI, C, S                                                                           0.39                          0.12
                    SRI, S, I                                                                         -0.43                          0.03
                    SRI, C                                                                            -0.20                          -0.06
                    SRI, S                                                                             0.33                          0.47
                    SRI, I                                                                            -1.42                          -0.33
                   SRI Fund only                                                                      -0.88                          0.21
                   C Fund only                                                                         0.46                          -0.24
                   S Fund only                                                                         1.38                          1.79
                   I Fund only                                                                        -2.41                          2.00
                  Source: GAO analysis of annual rates of total return based on monthly total return data from Morningstar, Inc.




                  Page 36                                                                                  GAO-12-664 Thrift Savings Plan
Appendix III: Additional Information on the
               Appendix III: Additional Information on the
               Performance of SRI and Non-SRI Mutual Funds
               by Asset Class


Performance of SRI and Non-SRI Mutual
Funds by Asset Class
               Our analysis of institutional-grade mutual funds over the last 5, 10, and 15
               years (dating back from December 2011) showed that SRI bond mutual
               funds had better risk-adjusted performance than their non-SRI
               counterparts, while SRI stock and SRI balanced funds did not.1
               Specifically, we reviewed measures of risk-adjusted performance
               including the Sharpe and Sortino ratios. Morningstar Inc. defines these
               ratios as follows:

                  Sharpe ratio. This statistic is a reward to variability ratio, which offers
                   a means of locating an optimal risky portfolio. The equation is:




                   rA= expected return on the series or portfolio

                   σ= the standard deviation

                   rF= the risk-free rate

                   And where standard deviation is:


                               1
                                   1




               1
                Morningstar identified three share classes of mutual funds—institutional, front-load and
               no-load—as the most common share classes used by institutional investors. We refer to
               these three share classes together as institutional grade mutual funds. We did not analyze
               other forms of SRI funds such as variable annuities and exchange traded funds. Hence,
               the results of our analysis are not generalizable to non-institutional grade mutual funds or
               to other forms of SRI. We included three asset classes of mutual funds in our analysis,
               stocks, bonds, and balanced. Stock funds invest in stocks, bond funds invest in bonds,
               and balanced funds invest in a combination of stocks and bonds. After controlling for fund
               size and investment strategies (other than SRI approaches), we found that the gap
               between the performance of SRI and non-SRI mutual funds narrowed significantly for
               stock funds but not for balanced funds.




               Page 37                                                     GAO-12-664 Thrift Savings Plan
Appendix III: Additional Information on the
Performance of SRI and Non-SRI Mutual Funds
by Asset Class




    Where:

    σ= the Greek letter commonly used to denote standard deviation

    rt= expected return on the series or portfolio

    ‫ݎ‬ഺ= the arithmetic mean of the return series r

    n= the number of periods

   Sortino ratio. The Sortino ratio is a risk adjusted return ratio that
    considers excess return over a designated target return and the risk of
    not achieving that target return. Excess return is defined as the series’
    return less the target return; risk is considered to be the semi-
    standard deviation below the target return. The Sortino ratio therefore
    tells you how well you are being compensated by a series for each
    unit of shortfall risk you are incurring.

     The formula for the Sortino ratio:

     Where:

     T= target return

     SD= the target semi-standard deviation of the return series in
     question over the period in question. This is the square root of the
     target semi-variance, with T as the target return

     തതതത
     ܴ  ௣ the arithmetic average return of the return series in question over
     the period in question




Page 38                                              GAO-12-664 Thrift Savings Plan
Appendix III: Additional Information on the
Performance of SRI and Non-SRI Mutual Funds
by Asset Class




The following tables provide additional information on risk-adjusted
performance of SRI and non-SRI mutual funds by asset class.

Bond Mutual Funds

Table 6: Sharpe Ratio SRI and Non-SRI Bond Mutual Funds

                                                                            Annualized time period
 Type of fund               Measure                               5 year                  10 year            15 year
 SRI
                            Mean                                    1.02                     0.82               0.77
                            Median                                  1.00                     0.87               0.75
                            Maximum                                 1.53                     1.20               0.84
                            Minimum                                 0.27                     0.38               0.73
                            Number                                     18                      17                  4
 Non-SRI
                            Mean                                    0.76                     0.66               0.52
                            Median                                  0.68                     0.64               0.48
                            Maximum                                 2.82                     2.05               1.84
                            Minimum                                -2.98                     -2.09             -2.71
                            Number                                 2,103                    1,799             1,449
Source: GAO analysis of mutual fund performance data received from Morningstar, Inc.




Table 7: Sortino Ratio SRI and Non-SRI Bond Mutual Funds

                                                                            Annualized time period
 Type of fund               Measure                              5 year                    10 year          15 year
 SRI
                            Mean                                    1.80                      1.27              1.19
                            Median                                  1.69                      1.32              1.19
                            Maximum                                 3.01                      2.15              1.28
                            Minimum                                 0.36                      0.51              1.09
                            Number                                     18                       17                 4
 Non-SRI
                            Mean                                    1.35                      1.05              0.81
                            Median                                  1.04                      0.96              0.71
                            Maximum                               14.80                       3.96              3.69
                            Minimum                                -2.48                     -2.60             -2.76
                            Number                                2,103                      1,799            1,449
Source: GAO analysis of mutual fund performance data received from Morningstar, Inc.




Page 39                                                                                GAO-12-664 Thrift Savings Plan
Appendix III: Additional Information on the
Performance of SRI and Non-SRI Mutual Funds
by Asset Class




Stock Mutual Funds

Table 8: Sharpe Ratio SRI and Non-SRI Stock Mutual Funds

                                                                            Annualized time period
 Type of fund               Measure                              5 year                    10 year          15 year
 SRI
                            Mean                                    0.05                      0.14              0.24
                            Median                                  0.06                      0.12              0.23
                            Maximum                                 0.33                      0.52              0.47
                            Minimum                                -0.49                     -0.05             -0.01
                            Number                                     86                       68                37
 Non-SRI
                            Mean                                    0.07                      0.22              0.27
                            Median                                  0.06                      0.20              0.26
                            Maximum                                 0.73                      0.68              0.72
                            Minimum                                -0.94                     -0.38             -0.13
                            Number                                3,165                      2,466            1,431
Source: GAO analysis of mutual fund performance data received from Morningstar, Inc.




Table 9: Sortino Ratio SRI and Non-SRI Stock Mutual Funds

                                                                            Annualized time period
 Type of fund               Measure                              5 year                    10 year          15 year
 SRI
                            Mean                                    0.08                      0.20              0.35
                            Median                                  0.09                      0.17              0.33
                            Maximum                                 0.50                      0.78              0.69
                            Minimum                                -0.57                     -0.06             -0.01
                            Number                                     86                       68                37
 Non-SRI
                            Mean                                    0.10                      0.31              0.39
                            Median                                  0.09                      0.28              0.37
                            Maximum                                 1.29                      1.04              1.12
                            Minimum                                -1.02                     -0.45             -0.16
                            Number                                3,165                      2,466            1,431
Source: GAO analysis of mutual fund performance data received from Morningstar, Inc.




Page 40                                                                                GAO-12-664 Thrift Savings Plan
Appendix III: Additional Information on the
Performance of SRI and Non-SRI Mutual Funds
by Asset Class




Balanced Mutual Funds

Table 10: Sharpe Ratio SRI and Non-SRI Balanced Mutual Funds

                                                                            Annualized time period
 Type of fund               Measure                              5 year                    10 year          15 year
 SRI
                            Mean                                    0.05                      0.13              0.21
                            Median                                  0.02                      0.15              0.20
                            Maximum                                 0.38                      0.23              0.36
                            Minimum                                -0.15                     -0.16             -0.02
                            Number                                     20                       13                 8
 Non-SRI
                            Mean                                    0.11                      0.29              0.32
                            Median                                  0.09                      0.27              0.30
                            Maximum                                 0.89                      0.95              0.78
                            Minimum                                -1.20                     -1.18             -0.93
                            Number                                   897                       454              318
Source: GAO analysis of mutual fund performance data received from Morningstar, Inc.




Table 11: Sortino Ratio SRI and Non-SRI Balanced Mutual Funds

                                                                            Annualized time period
 Type of fund               Measure                              5 year                    10 year          15 year
 SRI
                            Mean                                    0.07                      0.18              0.29
                            Median                                  0.02                      0.20              0.28
                            Maximum                                 0.57                      0.31              0.51
                            Minimum                                -0.19                     -0.20             -0.03
                            Number                                     20                       13                 8
 Non-SRI
                            Mean                                    0.16                      0.42              0.45
                            Median                                  0.12                      0.38              0.42
                            Maximum                                 1.57                      1.64              1.32
                            Minimum                                -1.26                     -1.16             -1.00
                            Number                                   897                       454              318
Source: GAO analysis of mutual fund performance data received from Morningstar, Inc.




Page 41                                                                                GAO-12-664 Thrift Savings Plan
Appendix IV: GAO Contact and Staff
                 Appendix IV: GAO Contact and Staff
                 Acknowledgment



Acknowledgment

                 Charles A. Jeszeck, (202) 512-7215 or jeszeckc@gao.gov
GAO Contact
                 In addition to the individual named above, Kimberley Granger, Assistant
Staff            Director; Jonathan S. McMurray, Analyst-in-Charge; James Bennett; and
Acknowledgment   Sarah Kaczmarek made significant contributions to this report. Kenneth
                 Stockbridge, Roger Thomas, and Jack Wang also made key
                 contributions.




(131099)
                 Page 42                                        GAO-12-664 Thrift Savings Plan
GAO’s Mission         The Government Accountability Office, the audit, evaluation, and
                      investigative arm of Congress, exists to support Congress in meeting its
                      constitutional responsibilities and to help improve the performance and
                      accountability of the federal government for the American people. GAO
                      examines the use of public funds; evaluates federal programs and
                      policies; and provides analyses, recommendations, and other assistance
                      to help Congress make informed oversight, policy, and funding decisions.
                      GAO’s commitment to good government is reflected in its core values of
                      accountability, integrity, and reliability.

                      The fastest and easiest way to obtain copies of GAO documents at no
Obtaining Copies of   cost is through GAO’s website (www.gao.gov). Each weekday afternoon,
GAO Reports and       GAO posts on its website newly released reports, testimony, and
                      correspondence. To have GAO e-mail you a list of newly posted products,
Testimony             go to www.gao.gov and select “E-mail Updates.”

Order by Phone        The price of each GAO publication reflects GAO’s actual cost of
                      production and distribution and depends on the number of pages in the
                      publication and whether the publication is printed in color or black and
                      white. Pricing and ordering information is posted on GAO’s website,
                      http://www.gao.gov/ordering.htm.
                      Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
                      TDD (202) 512-2537.
                      Orders may be paid for using American Express, Discover Card,
                      MasterCard, Visa, check, or money order. Call for additional information.
                      Connect with GAO on Facebook, Flickr, Twitter, and YouTube.
Connect with GAO      Subscribe to our RSS Feeds or E-mail Updates. Listen to our Podcasts.
                      Visit GAO on the web at www.gao.gov.
                      Contact:
To Report Fraud,
Waste, and Abuse in   Website: www.gao.gov/fraudnet/fraudnet.htm
                      E-mail: fraudnet@gao.gov
Federal Programs      Automated answering system: (800) 424-5454 or (202) 512-7470

                      Katherine Siggerud, Managing Director, siggerudk@gao.gov, (202) 512-
Congressional         4400, U.S. Government Accountability Office, 441 G Street NW, Room
Relations             7125, Washington, DC 20548

                      Chuck Young, Managing Director, youngc1@gao.gov, (202) 512-4800
Public Affairs        U.S. Government Accountability Office, 441 G Street NW, Room 7149
                      Washington, DC 20548




                        Please Print on Recycled Paper.