PII: S1057-0810(01)00066-X


Design considerations for large public sector defined
contribution plans<

Kevin W. SigRista,1, Stewart L. Brownb,*
aFlorida State Board Of Administration

bFlorida State University, Tallahassee, FL 32306, USA

Received 14 May 2000; received in revised form 20 October 2000; accepted 1 December 2000

Abstract

The paper identifies differences between private (401 (k)) plans, which have evolved under ERISA
and existing public plans, which have not. Examination of model legislation reveals that public plans
should largely conform to ERISA going forward and reflect best practices in the private sector.
Empirical analysis of equity mutual funds with $1.8 trillion in assets and institutional equity accounts
with $98 billion in assets demonstrates efficiencies in separately procured institutional investment,
administrative and educational services relative to retail investment products. The analysis points to
tension between the duties of trustees and the demands of participants requesting large numbers of
retail investment options. © 2001 Elsevier Science Inc. All rights reserved.

Jel classification:G23; G28; H55; H72; J26

1. Introduction

The public sector increasingly uses defined contribution (DC) pension plans, either as
supplement or in place of the more traditional defined benefit (DB) pension plans. Optional
DC programs have been in place for a number of years for public university faculty, but in
recent years a number of large public employee retirement systems have adopted or are in

< The views expressed in this paper are solely those of the authors and do not represent the positions of their
employers. The authors would like to express their appreciation for helpful comments from Jim Francis, Russ
Ivinjack and John Freeman.

* Corresponding author. Tel.:11-850-644-9657; fax:11-850-644-4225.
E-mail addresses:sigrist_kevin@fsba.state.fl.us (K.W. SigRist), sbrown@cob.fsu.edu (S.L. Brown).

Financial Services Review 9 (2000) 197–218

1057-0810/00/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S1057-0810(01)00066-X



the process of implementing DC arrangements (e.g., Michigan, Colorado, Washington,
Louisiana, Florida, Montana, Ohio, and South Carolina). Moreover, proposed Social Security
reforms that would introduce self-directed accounts are essentially federal DC accounts.
Thus, investment policy issues unique to public sector DC plans are important and timely.

There is little literature exploring the special circumstances of public sector DC plans and
thus no clear enunciation of best practices. However, there are apparent differences in the
organization and administration of public and private sector DC plans. The purpose of the
paper is to address the policy issues encountered in the design of large public sector DC
programs.

Many corporate DC plans, which have evolved under the Employee Retirement Income
Security Act (U.S. Department of Labor, 1974), utilize an institutional approach where
services are unbundled; that is, trustees contract separately for components of investment
advisory, educational and administrative services. Morgan Stanley (1999) reports that of 375
401(k) plans surveyed in 1999, about half of the programs purchased unbundled investment
products and administrative services and more than half of the programs offered some
institutional (nonretail) investment products to their participants. Another survey reported
that 80% of 401(k) plans with more than 10,000 participants were at least partially unbundled
in 1999 (BARRA RogersCasey/IOMA, 1999). This has the impact of minimizing plan
expenses and increasing participant benefits. Generally such plans are consistent with
Modern Portfolio Theory (MPT) and offer limited investment options. It is not unusual to
find index funds in such plans.

In contrast, public DC plans have evolved without the benefit of ERISA. Enabling laws
and statutes are unique to each state. Such plans often lack a trustee governance structure and
commonly utilize a bundled investment architecture where vendors provide investment
advisory, educational and administrative services under one umbrella fee. Vendors in this
sector include Variable Annuity and Mutual Fund companies and fees are often higher than
the fees of unbundled DC plans. It is common to find multiple competing vendors in the same
plan, negating some of the advantages of economies of scale and sometimes engendering
conflicts of interest among vendors (Jacobius, 2000). Public plans typically have dozens and
sometime hundreds of different investment options (KPMG, 1999; National Association of
Deferred Compensation Administrators (NADCA), 1998).

Trustees of new public DC pension programs, including potential Social Security reforms,
must ultimately choose between the two approaches or implement a hybrid program. The
salient issues are the number of investment options available to participants and the manner
in which those products and other services are procured: 1) name-branded retail investment
options which bundle together investment, administrative and educational services versus 2)
an unbundled procurement of private label institutional/wholesale investment options which
may be independent of administrative and educational services. Other issues remain impor-
tant, such as active versus passive management. However, actively and passively managed
portfolios coexist comfortably in most corporate DC plans and we view this issue as outside
the scope of the paper.

The issue of the number and diversity of investment options is a thorny thicket constantly
faced by public sector and corporate DC trustees. Choice is commonly equated with
freedom; the more the better. Thus, for plan participants, there is an intuitive appeal and

198 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



comfort that arises from the ability to choose among a broad array of branded, hence familiar,
investment products. Vendors reinforce the importance of product diversity and trustees can
be put in the uncomfortable position of defending design decisions with seemingly arcane
academic arguments.

However, even in a theoretically pure liberal democracy, freedom is constrained to protect
property rights. Similarly, trustees must critically evaluate the expected marginal costs and
benefits to the overall group of participants associated with providing a diverse choice among
duplicative investment products. In the main, corporate DC trustees have largely opted for
a cost-effective approach that limits the investment options to a handful (i.e., 10 or less
options) of relatively distinct products. Moreover, EnnisKnupp (1999) surveyed corporate
plan administrators and over two-thirds of 110 respondents replied that institutional accounts
are more effective than mutual funds.

We explore these issues in more depth in the body of the paper. It is organized as follows:
we identify the major types of public plans, identify unique pressures on the organizers/
trustees of such plans, discuss the evolving legal infrastructure underpinning public pension
plans and discuss model legislation: the 1997 Uniform Management of Public Employee
Retirement Systems Act (National Conference of Commissioners on Uniform State Laws,
“MPERS”). We then explore the implications of MPERS for the investment/administrative
design of a public DC plan and analyze a unique data set comparing the administrative and
investment advisory costs associated with institutional accounts and mutual funds. Finally,
we explore the implications of a hybrid approach where institutional accounts and mutual
funds coexist.

MPERS and MPT strongly suggest that trustees/fiduciaries implement an unbundled
institutional approach. Analysis reveals substantially greater economies of scale associated
with institutional than mutual fund accounts. Further analysis reveals that a hybrid structure
has serious shortcomings because it results in higher costs (lessened economies of scale) for
a subset of the participants: those who choose the institutional investment options.

The higher costs associated with retail accounts significantly reduces the future benefits of
participants. With an initial account value of $500 million, a 30-year horizon and annual
market appreciation of 10%, large-cap mutual fund fees would lead to an account value that
was $870 million lower than that produced by a large-cap institutional fund account. Further,
in hybrid structures, we show that for a 10-year horizon, participants that prefer the low-cost
institutional option over the higher cost name-brand options are forced to pay about 40%
more in cumulative fees because assets move into the branded options.

2. First principles: trusts, trustees and fiduciary duties

Garla (1998) provides a taxonomy of public-sector tax-qualified and nontax-qualified DC
plans. Section 403(b) and Section 457 Plans (designated by the Internal Revenue Code
Section) are by far the most common public-sector DC-type plans. Section 403(b) Plans
allow employees of state educational organizations and governmental entities to elect to have
their employer make tax-deferred contributions for them to purchase an annuity contract or
make contributions to a custodial account for investment in mutual fund shares. Section 457

199K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



Plans are deferred compensation arrangements, for employee contributions, that can be
established by a state or other governmental entity. In 1996, Section 457 was amended so that
as of January 1, 1999 all assets and income of 457 plans must be held in trust for the
exclusive benefit of participants and beneficiaries. Section 401(k) Plans are available to a
small number of public employers who had them adopted prior to May 6, 1986. The more
recent wave of new and large public sector DC plans have been established as Section 401(a)
Money Purchase Plans. In 1986, the Federal Thrift Savings Plan (a 401(a) plan), currently
with $85 billion in assets, was authorized for federal employees.

As a rule, public sector retirement systems are exempt from the Employment Retirement
Income Security Act (ERISA) of 1974 (Employment Retirement Income Security Act,
1974), which governs corporate 401(k) programs. Without the guiding principles of ERISA,
the laws regulating public sector DC plans have evolved separately and vary significantly
across states. Moreover, both DB and DC public sector retirement plans have often lagged
developments in modern financial theory and practice. Garla (1998) surveys governmental
retirement system law and reports the widespread legacy of statutory lists and other legal
guidelines that govern permissible investments.

Legal considerations are typically ignored in the investment policy literature because
ERISA has provided a legal framework. Absent ERISA, specific laws and statutes are
necessary to initiate a public sector program and govern responsibilities for the design and
administration of the program. Therefore, this section discusses the key legal, fiduciary and
governance foundations that are necessary to establish a sound investment policy in the
public sector context.

2.1. Interested parties and the political system

Besides ERISA, another difference between public sector plans and corporate plans is that
interest groups routinely attempt to use the political process to enhance their well being and
profitability—a phenomenon known generically as rent-seeking. Rent-seeking is not the
same as lobbying activity. The latter principally communicates information on political
positions to decision-makers and their staff. Such communications may include proprietary
information and analysis relevant to assessing the economic impact of potential decisions.
Rent-seeking is an exercise in redistributing economic resources by obtaining preferential
tax, regulatory or procurement policies.

In the context of public retirement systems, rent-seeking behavior can take several forms.
First, potential vendors may try to manipulate decision-makers through the statutory, bud-
getary or administrative law processes. Second, potential vendors have apparently estab-
lished economic and financial relationships with traditional lobbying organizations, such as
local government associations (e.g., Associations of Counties and Leagues of Cities), unions
and professional associations (Pinkston, 2000; Jacobius, 2000). Third, some lobbying orga-
nizations, such as local government associations, may have captive money management
operations. Fourth, other interest groups can be expected to lobby for economically targeted
investment policies. Finally, trustees of a retirement system may face rent-seeking activity
within the ranks of their professional staff, whose interests may be mis-aligned with those of
participants, particularly if hiring and firing of staff is restricted by civil service laws.

200 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



The existence of interest group pressures dictate that implementing statutes must incor-
porate good financial and governance policy to ensure that participants and taxpayers are
sheltered from the rent-seeking behavior of interest groups. Such behavior generally in-
creases the costs of DC plans and thus reduces the eventual benefits that accrue to partici-
pants.

It is speculative to predict whether, and under what circumstances, these types of political
efforts might cause public sector decision makers to take actions potentially adverse to the
interests of participants. For instance, trustees are generally accomplished professionals with
integrity and substance and possessing sufficient resources to help them recognize and
understand their duties to participants. At the same time, the history of public sector DB
plans is instructive. Political pressures have led to legal restrictions on investments (consti-
tutional and statutory), economically targeted investments, social investments and occasional
budget actions affecting contributions or plan assets (Garla, 1998; Garthwaite, 1999; Useem
& Hess, 1999).

2.2. The Uniform Management of Public Employee Retirement Systems Act

Importantly, there is movement toward standardizing the legal infrastructure governing
public sector plans and incorporating some of the key components of ERISA. The Uniform
Management of Public Employee Retirement Systems Act (“MPERS”) (National Conference
of Commissioners on Uniform State Laws, 1997) develops a uniform legal framework for the
administration and operation of state and local public sector DB and DC plans. MPERS sets
out much of the legal infrastructure that is necessary to facilitate an effective investment
policy and the balance of this section explores certain aspects of MPERS.

One of the guiding principles of MPERS (and ERISA) is the requirement that assets of
retirement systems be held in trust and that trustees have theexclusiveauthority to invest and
manage those assets. Further, MPERS establishes that trustees should be sufficiently inde-
pendent to effectively and efficiently perform their duties. Among the exclusive powers that
MPERS confers are:

The power to establish an administrative budget sufficient to perform the trustee’s duties
and as appropriate and reasonable draw upon assets of the retirement system to fund the
budget;
The power to obtain by contract the services necessary to exercise the trustees’ duties;
The power to procure and dispose of goods and property necessary to exercise the
trustees’ powers and perform the trustee’s duties.

Trustee independence is essential because it permits them to perform their duties in the
face of political pressure from “interested parties.” In the absence of independence, trustees
may be forced to decide between fulfilling their fiduciary obligation to participants or
following the suggestions of groups whose interests may not be aligned with those of the
participants. The surest protection for trustees’ independence is through a constitutional
amendment. A supermajority requirement for future changes to trustees’ powers, creates a
close approximation to constitutional protection.

Of course, an independent board of trustees raises the specter of unethical behavior that

201K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



harms participants and taxpayers (self-dealing, bloated expenses etc.). The antidote to such
behavior is to subject trustees to: stringent fiduciary duties and standards of care; a require-
ment to publish an investment policy statement consistent with those duties; a requirement
to regularly report on investment performance net of fees and relative to financial market
benchmarks; annual independent audits; and personal liability for any losses resulting from
the breach of duties, with limited ability to shift fiduciary liability onto others.

2.3. MPERS and the prudent expert rule

The one major shortcoming of MPERS is its incorporation of a prudent person standard
of care rather than the more stringent prudent expert standard embodied in ERISA and
common law. The prudent expert standard requires that the retirement program’s assets are
invested, on behalf of the participants, with the care, skill, and diligence that a prudent
investor acting in a like manner would undertake. Under this standard, if a fiduciary is not
an expert, they have an obligation to obtain expert advice.

The official MPERS commentary explains that the prudent expert standard is too exacting
given the diversity among existing public retirement systems. In other words, a lower
standard is a legacy of existing public sector plans that evolved without ERISA standards.
Invoking the lesser prudent man standard favors trustees at the expense of participants
because trustee’s actions are judged on the basis of how trustees in other public plans have
behaved. Such behavior may not be exemplary.

The prudent expert standard applies to corporate retirement plans under ERISA and
private trusts under the common law. There does not appear to be a compelling argument for
a weaker standard for public sector plan participants.

2.4. Requirements for trustees under MPERS

MPERS requires that trustees must discharge their duties for the sole interest and
exclusive purpose of providing benefits to plan participants and beneficiaries and defraying
reasonable expenses of administering the plan. Trustees must also discharge their duties by
impartially taking into account any differing interests of participants and beneficiaries, by
incurring only costs that are appropriate and reasonable and in accordance with a good faith
interpretation of the laws governing the program.

Fiduciaries responsible for the investment or management of retirement assets are further
required to: “consider a broad range of economic and financial circumstances in establishing
investment strategies, to diversify investments, to make a reasonable effort to verify facts
relevant to the investment program, and may consider the benefits created by an investment
in addition to investment return only if the trustee determines the investment was prudent
even without the collateral benefits.” Collateral benefits refer to social investing, economi-
cally targeted investing and so forth

MPERS confers an affirmative duty for fiduciaries to adopt a detailed investment policy
statement that incorporates the investment objectives, the desired rates of return and accept-
able levels of risk for each asset class, guidelines for the delegation of authority and
information on the types of reports to be used to evaluate investment performance. MPERS

202 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



requirements comport reasonably well with current corporate practice. Table 1 indicates the
subject areas contained in the investment policy statements of a sample of 250 corporate
401(k) DC programs (BARRA RogersCasey/IOMA, 1999). This source indicates that about
one-half of corporate 401(k) programs have formally adopted investment policy statements,
although Chambers (1999) notes that there is no requirement under ERISA to have an
investment policy statement.

DC programs are intended to be participant-directed where participants bear the risks of
their investment decisions. Therefore, MPERS effectively incorporates section 404(c) of
ERISA: if a participant exercises control over the assets in their account, no program
fiduciary is liable for any loss to a participant’s account which results from such participant’s
exercise of control.

To obtain this protection for program fiduciaries, participants must be given meaningful,
independent control over the assets in their account with the opportunity to:

Choose from a broad range of investment alternatives that allow diversification within and
among such alternatives;
Give investment instructions with a frequency that is appropriate in light of the market
volatility of the investment alternatives; and
Obtain sufficient information to make informed investment decisions.

The incorporation of ERISA 404(c) is an important legal foundation in the design of a
public sector DC plan. In order to obtain 404(c) protection, program fiduciaries must
establish a program with investment, administrative and educational features that meet
minimum standards. However, the fact that participants are in control of certain investment
decisions does not relieve trustees of the duty to provide an appropriate design and conduct
on-going monitoring of the program to ensure it operates as planned and remains competitive

Table 1
Areas covered in 401(k) investment policy statements, 1999

All plans Large plans

Monitoring investment options 80% 88%
Determining type of investment options 78% 76%
Setting 401(k) plan objectives 70% 68%
Selecting investment options 67% 74%
Determining number of investment options 61% 65%
Benchmarking investment options 61% 76%
Terminating investment options 60% 65%
Amending investment policy 48% 38%
Designating roles and responsibilities 46% 56%
Plan expenses 36% 41%
Enforcing investment policy 35% 26%
Administration guidelines/requirements 31% 18%
Communication guidelines/requirements 27% 21%
Other 4% 0%
Memo: Number with Statement/Total Sample 250/446 34/56

Source: Barra RogersCasey/IOMA Annual Defined Contribution Survey. Large Plans have at least 10,000
participants.

203K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



with other alternatives. Public sector trustees have a duty to establish and maintain “best
practices” in the design of a DC program. As we shall see, cost considerations are central to
those decisions.

3. Public DC program objectives

Having reviewed the legal and fiduciary framework within which trustees should operate,
it is relatively straightforward to identify a DC program’s objectives:

1. Offer a diversified mix of low-cost investment options that span the risk-return
spectrum and give participants the opportunity to accumulate portable retirement
benefits.

2. Offer investment options that avoid excessive risk, have a prudent degree of diversi-
fication relative to broad market indices and provide a long-term rate of return, net of
all expenses and fees, that achieves or exceeds the returns on comparable market
benchmark indices.

3. Offer participants meaningful, independent control over the assets in their account with
the opportunity to:
a) Obtain sufficient information about the plan and investment alternatives to make

informed investment decisions;
b) Direct contributions and account balances between approved investment options

with a frequency that is appropriate in light of the market volatility of the invest-
ment options;

c) Direct contributions and account balances between approved investment options
without the limitation of fees or charges; and

d) Remove accrued benefits from the plan without undue delay or penalties.
4. Offer participants cost-effective administrative and educational services that will help

the plan maintain compliance with U.S. Department of Labor section 404(c) regula-
tions and provide participants with impartial and balanced information about invest-
ment choices that will help facilitate their portfolio decisions.

Most of the objectives are direct offshoots of MPERS, the prudent expert standard of care
and the 404(c) requirements. Others such as portability are not controversial. However,
Objective 4 deserves detailed discussion.

3.1. Impartial and balanced educational services

Corporate DC educational services have significantly evolved and expanded under recent
federal regulations. Corporations were naturally averse to providing education to 401(k)
participants that might later be interpreted by courts as investment advice. Therefore, in
1996, the U.S. Department of Labor promulgated Interpretive Bulletin 96–1 (IB 96–1) to
encourage the provision of educational services under 404(c) of ERISA. IB 96–1 identified
broad safe harbors for education, including: general financial and investment information,
asset allocation models and interactive investment materials (e.g., worksheets and software).

204 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



The Department of Labor also officially stated their interest in the provision of impartial
and balanced educational services. For example, to comply with the IB 96–1 safe harbors,
asset allocation models and interactive investment materials must incorporate generally-
accepted investment theory and disclose all material assumptions. Similarly, if interactive
materials and allocation models identify specific investment products, the participant must be
alerted to the existence of other substitutes and be given sources of information on those
alternative products. In IB 96–1, the Department stated that their intent behind these
requirements was, “to address the concern that a service provider could effectively steer
participants to a specific investment alternative by only specifying one particular fund in
connection with an asset allocation model.”

The Department’s concerns are of central importance for public sector DC plans because
they are not subject to the prohibited transactions regulations of ERISA. Under such
regulations, investment product providers and other fiduciaries are generally prohibited from
actions that would benefit them at the expense of the participants; for example, advising a
participant to choose one of their investment offerings. Therefore, unless public sector
trustees take an affirmative action, in law or the program objectives, participants will not be
afforded one of the basic protections enjoyed by members of corporate 401(k) programs.
Unfortunately, MPERS does not invoke the prohibited transactions regulations of ERISA.

Public sector trustees have a strong incentive to arrange cost-effective educational services
to help facilitate investment decisions and improve retirement benefits. First, providing
impartial and balanced educational information appears to be broadly consistent with the
404(c) objective of ensuring that participants remain in control of their assets. Second,
provision of even low levels of impartial educational services along the lines of IB 96–1
should entail significant increases in retirement benefits. Studies of consumer investment
behavior indicate a sizeable group of consumers do not understand or follow basic invest-
ment principals related to asset allocation and cost management.1 Moreover, the rule of
thumb is that more than 90% of DC retirement benefits will be determined by an investor’s
asset allocation (Ibbotson & Kaplan, 2000) and every 100 basis points of excess cost lowers
a DC account balance by approximately 20% over 30 years.

Trustees are also in a position to utilize the program’s group purchasing power to acquire
a standardized package of educational services for substantially lower cost than can be
acquired by individuals at retail prices. Effectively, the long-term cost to taxpayers of a DC
program should be lower per dollar of ultimate retirement benefit with relatively low
educational expenditures. Of course, trustees have a duty to monitor the cost effectiveness of
their chosen package of educational services, since they have a duty to incur only costs that
are appropriate and reasonable.

4. Implications for general program design and procurement

The objectives imply that the selection of all program vendors must be guided by
“best-in-class” principles, unless there are significant perceived advantages to purchasing
products and services bundled. However, there are several compelling arguments for pur-
suing an unbundled architecture with separately negotiated and procured investment, admin-

205K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



istrative and education services. First, it enhances the trustees’ negotiating position, as the
purchasing agent for the participants. Second, it provides the opportunity for a full disclosure
and review of performance and all costs borne by participants and taxpayers. Importantly,
investment and administrative services are subject to different cost economics and there
appears to be significantly different price structures in the retail and wholesale/institutional
marketplaces (a point we empirically test below). For instance, a U.S. Department of Labor
(1998) report on 401(k) costs stated:

“Larger plans enjoy potentially significant economies of scale. In the case of investment
expenses, they have access to more providers offering a wide range of investment vehicles
at lower cost. Very large plans may be able to reduce investment expense even more through
fee-reduction negotiations with the providers or use lower-cost institutional accounts. In
other expense categories, the combination of flat (or nearly flat) fees regardless of plan size,
plus declining per-capita charges for basic administration fee, reduce per-participant admin-
istrative costs among larger plans.”

Finally, an independent education vendor is needed to provide participants with impartial and
balanced information about investment choices. An independent educator can sidestep the
inherent conflicts of interest that arise when an investment product vendor educates and
counsels on asset allocation, financial planning issues and investment fees/costs in the
context of their products. An independent education vendor is also the natural agent to
empower participants with regular, understandable and standardized disclosure of product
fees and investment option performance versus market index benchmarks.

5. Implications for investment program design and procurement

The number and types of investment options offered in a DC plan are critical. The
overriding objective should be to provide a range of options that allows participants to
choose a point as close as possible to the efficient frontier consistent with individual risk
preferences. Within the context of the program objectives, investment options should be
consistent with MPT, finance theory and best practices.

There are three interrelated issues influencing the number and types of investment options:
diversification, costs and active management. Assume for a moment that markets are
perfectly efficient and consider the issue of diversification. In a perfectly efficient market,
index funds are the investment vehicle of choice and investment management costs are
minimal. How many and what types of investment vehicles are necessary to “span the
risk/return spectrum?” At least three: US Stocks, US Bonds and Cash Equivalents. For many
years, those were essentially the options available to federal employees in the Federal Thrift
Savings Plan.

These three choices minimally meet the diversification requirement of 404(c). However,
we would argue this concise list should be expanded to include two categories: Foreign
Stocks and Inflation-Indexed Bonds. An asset type should be defined at the highest pragmatic
level of aggregation possible, where the next highest level of aggregation would combine
securities with materially different legal, financial and economic characteristics. Foreign

206 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



Stocks and Inflation-Indexed Bonds have materially different characteristics than the three
options.

Providing five index fund options representing the broadest representations of these asset
categories would minimally satisfy the program’s investment objectives. Among the set of
liquid, highly diversified and unleveraged portfolios composed of public market securities,
these five index options would qualify as low-cost and effectively span the lion’s share of the
risk and return spectrum. Importantly, these options span the short-term market risk dimen-
sion and the long-term income replacement risk dimension. The latter is defined by the
possibility that a participant’s account balance will prove to be insufficient to maintain a
reasonable post-retirement standard of living. Over time, an employee’s income tends to
grow with inflation, improving productivity and prevailing market wages for similar occu-
pations and levels of managerial responsibility. U.S. stocks have historically provided high
average annual real returns and this long-term real return has been largely unaffected by the
rate of inflation (Boudoukh & Richardson, 1993). The 1997 introduction of inflation-indexed
bonds (TIPS) in the U.S. created a new asset type with extremely attractive characteristics
for tax-deferred retirement accounts.

Ibbotson and Kaplan (2000) examined performance for balanced mutual funds and large
DB pension funds and found that 99% to 112% of the total return level was explained by the
fund’s long-term asset allocation (the authors subdivided U.S. stocks into large- and small-
cap sectors). Thus, five indexed portfolios representing broad asset types would allow plan
participants to articulate their risk preferences. The average contribution of active manage-
ment in the Ibbotson and Kaplan study was less than invigorating. However, the appropriate
use of active management strategies is outside the scope of this paper. For guidance, it is
useful to turn to best practices in the 401(k) arena where plan sponsors are subject to ERISA.

Table 2 provides summary data on 401(k) programs’ investment options (BARRA
RogersCasey/IOMA, 1999). Despite diversity in the aggregate, this same survey indicated
that the median number of options was between 8 and 9 in 1999 and the majority of program
administrators believed that the number of investment options became excessive around 11.
A representative list of 8 options that generally comports with these corporate practices and
recent academic research would be: money market option; total market nominal bond option;
inflation-indexed bond option; total market U.S. stock index option; U.S. small stock option;
U.S. value stock option; U.S. growth stock option; and foreign stock option. In addition, two
or three balanced funds, optimized to reside on the efficient frontier, would provide one-stop
shopping for participants reluctant to perform their own asset allocation analysis.

Waring et al. (2000) make a persuasive case that DC investment options should predom-
inantly be optimized balanced funds. In their model design, asset type options should
function as “specialty options.” In part, they argue that participants should not be put in an
environment that facilitates focusing on the short run game of trying to pick hot funds, to the
exclusion of focusing on more important long run considerations, such as asset allocation.
This model is not commonly used in corporate or public sector DC plans.

As indicated in the introduction, the relatively low numbers of options in corporate DC
plans stand in stark contrast to the high numbers of investment options that are often put
before participants in public sector plans. Moreover, public sector 403(b) and 457 DC plans
often utilize multiple full-service bundled providers to generate diversity of investment

207K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



options within the program; a practice that is virtually unheard of in 401(k) programs.
Clearly, duplicative administrative and educational services bloat costs and reduce overall
benefits. Moreover, bundled providers have incentives to favor their branded alternatives,
which creates an inherent conflict of interest in any educational services they might offer. In
the next section, we empirically identify performance factors that support an unbundled
procurement of private label institutional/wholesale investment options.

6. Comparative cost analysis

Table 3 provides cost data from 1996 that illustrates the basic attraction of institutional
separate accounts relative to mutual funds: costs are far lower. Institutional mutual funds are
special share classes with high initial minimum balances (e.g., $100,000) and limited
provision of administrative services to the client, as well as lower charges for marketing and
distribution. Institutional separate accounts typically operate wherein a custody bank safe-
keeps securities in a group account and a manager(s) is authorized to trade securities in
that account; individual and collective recordkeeping is performed by the custody agent
or a third party recordkeeper. Institutional commingled accounts operate in a similar
fashion as an institutional mutual fund, but pricing is generally at the level of separate
accounts.

Table 2
Investment options offered by corporate 401(k) programs in 1999

Percentage of plans
all plans

Offering option
large plans

Balanced Fund 77% 77%
International Equity 71% 70%
U.S. Large Cap Equity 70% 72%
U.S. Equity—Indexed 66% 75%
U.S. Bonds 63% 65%
U.S. Small Cap Equity 61% 51%
Stable Value (GICs/Synthetics) 60% 74%
Money Market 56% 42%
U.S. Mid Cap Equity 48% 40%
Company Stock 28% 56%
Global Equity 27% 21%
Lifestyle Options 19% 25%
Emerging Markets 15% 9%
International Bonds 8% 2%
Real Estate Investment Trusts 6% 5%
Self-directed Brokerage 9% 4%
Mutual Fund Window 7% 4%
Other 26% 30%
Memo: Respondents 448 57

Source:BARRA RogersCasey/IOMA Annual Defined Contribution Survey. Large Plans have at least 10,000
participants.

208 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



It is important to note that the mutual fund costs in Table 3 include investment and
administrative fees, but the institutional separate accounts include only investment fees.
However, institutional accounts with more than $25 million in assets should enjoy even
lower investment advisory fees because their fees for institutional separate and commingled
accounts (institutional accounts) are generally negotiated on a sliding fee basis according to
the level of assets under management. In contrast, retail mutual funds charge all owners of
each share class the same expenses, regardless of whether the owner has 100 shares or
2,000,000 shares.

The empirical analysis in this section compares investment advisory fee levels for
institutional accounts and mutual funds and tests whether either vehicle lowers their invest-
ment advisory fees as assets under management increase. We find that institutional accounts
have much lower investment advisory fees than retail mutual funds and provide significantly
greater fee concessions as assets under management grow. In contrast, mutual funds provide
fee concessions for administrative services. However, these services can be independently
procured at costs competitive with the mutual funds’ charges.

Our empirical analysis uses two data sets compiled by Freeman and Brown (2001) to
compare the costs of actively managed domestic equity mutual funds and externally managed
equity institutional investment products. Readers are referred to that paper for methodolog-
ical details.

Freeman and Brown sent inquiries to the 100 largest DB public pension funds listed in the
January 25, 1999 edition of Pensions and Investments. Data for 1999 was collected on 220
individual external actively managed domestic equity portfolios with a total of $97.5 billion
in assets. The average portfolio size was $443 million, with the range extending from $15
million to $4.8 billion.

Morningstar’s Principia Pro compilation for October 1999 was the chief source of mutual
fund data. After eliminating funds with zero assets and missing data, the sample consisted of
4,943 equity funds. Multiclass funds were aggregated into single funds where weighted
averages of various expense ratios were obtained, using subfund assets as weights. Invest-
ment advisory fee and administrative fee ratios were separately compiled, where adminis-

Table 3
Average mutual fund expense ratios and separate account management fees data from 1996 and expressed as
percent of assets

Fund categories Most common retail
mutual funds
in DC plans

Institutional
mutual funds

$25 million
separate account

Indexed U.S. Stocks 0.27% 0.35% 0.13%
Active Large Stocks 0.83% 0.91% 0.63%
Active Small Stocks 1.06% 1.01% 0.95%
Foreign Stocks 1.33% 1.15% 0.75%
Active U.S. Bonds NA 0.69% 0.37%

Source: Cerulli Associates data cited in U.S. Department of Labor publication: Study of 401(k) Plan Fees and
Expenses, April 13, 1998, Contract No. J-P-7-0046, Task Order

209K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



trative fees were defined to exclude marketing and distribution fees (e.g., 12b-1). Adminis-
trative fees include transfer agent fees, custodial services, accounting fees and director’s fees.
Screens were applied to generate a sample of mutual funds closely corresponding to
characteristics of portfolios of public pension funds. The final mutual fund sample consisted
of 1,343 funds representing a total market value of about $1.77 trillion. The average portfolio
size was $1.3 billion, with the range extending from $15 million to $92.2 billion. Within the
sample are 1205 retail mutual funds with a market value of $1.71 trillion and 138 institutional
mutual funds with a market value of $56 billion. Institutional mutual funds were defined as
single-class funds having a minimum initial balance of $100,000 or more.

Importantly, there are no significant differences between the cost economics of providing
investment advisory services to mutual funds and DB pension fund separate accounts.
Investment management firms commonly offer nearly identical investment products in both
the institutional and retail markets. Investment management firms also access the same
markets for inputs: human capital, information and technology. Thus, investment advisory
services are essentially a commodity and fungible between mutual funds and institutional
accounts.

However, there are significant administrative cost differences between mutual funds and
DB pension fund separate accounts. Mutual funds are valued daily, rather than monthly.
Also, retail mutual fund administrative fees incorporate accounting/record-keeping costs at
the individual account level. These accounts are much smaller than the typical DB institu-
tional account. Smaller account size entails higher average administrative fees. Therefore, we
separate administrative fees and investment advisory fees in the following analysis.

This paper expands on Freeman and Brown’s analysis in several ways. First, a more
efficient simultaneous equation model is used to test for different economies of scale for
mutual fund administrative fees and investment advisory fees. Second, cost equations utilize
dummy variables to capture different cost structures for large-cap mutual funds, midcap
mutual funds and small-cap mutual funds. Regression equations shown in 1 and 2 below
were run on the mutual fund data and the results are presented in Table 4. Expense ratios are
scaled in basis points and size is scaled in millions of dollars under management. Because
of the exhaustive classification of the mutual funds, no intercept is included in the equations.

1. Administrative Expense Ratio5 a(1)*(Natural Log of Size)1a(2)*(Large-Cap
Dummy)1a(3)*(Mid-Cap Dummy)1a(4)*(Small-Cap Dummy)

2. Investment Advisory Expense Ratio5 b(1)*(Natural Log of Size)1b(2)*(Large-Cap
Dummy)1b(3)*(Mid-Cap Dummy)1b(4)*(Small-Cap Dummy)

The regression equations are highly statistically significant, as are all of the individual
coefficients, except the size coefficient for institutional mutual funds. Significant economies
of scale are apparent for both administrative expenses and investment advisory expenses.
However, scale economies are roughly 2.5 times greater for administrative fees than invest-
ment advisory fees. This difference is statistically significant for the retail mutual funds and
all mutual funds. For example, for a $500 million investment in a large-cap retail mutual fund
portfolio, total administrative costs are about one-half the investment advisory fee.2

The gap in expense ratios between retail and institutional mutual funds is generally
consistent with the fees shown in Table 3; noting that this latter report aggregated admin-

210 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



istrative and investment advisory fees. The smaller, although still significant, economies of
scale present in the administrative services of institutional mutual funds is likely a result of
the fact that they are bundled with materially lower administrative services and fees (see Fig.
1). Since institutional mutual fund share classes often perform limited individual services,
they have to prepare fewer prospectuses, allocate fewer phone representatives, compile and
maintain limited account data and so forth

Regressions of the form described in 3 below were run on the pension fund data. Expense
ratios are scaled in basis points and size is scaled in millions of dollars under management.
Because of the lack of an exhaustive classification of the pension fund data, an intercept is
included in the equation.

3. Investment Advisory Expense Ratio5 c(1)1c(2)*(Natural Log of Size)1c(3)*(Large-
Cap Dummy)1c(4)*(Mid-Cap Dummy)1c(5)*(Small-Cap Dummy)

Table 5 contains the pension fund regression results and the comparable mutual fund
results from Table 4. The analysis indicates that at low asset levels, investment advisory fees
for the pension funds and institutional mutual funds are comparable, but are significantly

Table 4
Simultaneous equations estimates of active domestic equity mutual fund administrative and investment
advisory fees per $10,000 of assets under management

Coefficient estimates
(T-statistics)

All mutual
funds

Retail mutual
funds

Institutional mutual
funds

Administrative Expense Ratio
Natural Log of Size 26.4 26.9 23.6

(214.1) (214.3) (23.0)
Large-Cap Dummy 71.5 75.3 45.9

(28.4) (28.0) (7.1)
Mid-Cap Dummy 72.0 74.8 53.7

(29.5) (28.8) (8.4)
Small-Cap Dummy 71.8 75.2 48.8

(28.8) (28.8) (7.1)
R-Squared 18.7% 20.8% 9.6%
Investment Advisory Expense Ratio
Natural Log of Size 22.5 22.8 22.8

(26.8) (27.4) (21.7)
Large-Cap Dummy 79.1 82.3 64.9

(35.3) (35.3) (7.0)
Mid-Cap Dummy 88.5 89.7 91.2

(37.8) (37.1) (9.8)
Small-Cap Dummy 95.6 97.4 93.1

(46.1) (46.8) (10.0)
R-Squared 14.8% 15.0% 24.9%
Sample Size 1343 1205 138
Assets in Billions $1,769 $1,713 $56
Chi-Square Test on Difference of Natural Log size
Coefficients
Chi-Square/Probability 39.6/0.0% 39.3/0.0% 0.13/71.5%

211K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



below that of retail mutual funds for the large-cap category. Significant economies of scale
are apparent in the pension fund account investment advisory fees. The slope coefficients for
the pension fund regression is 2 times greater than the mutual fund regression reflecting that
pension fund advisory fees are twice as sensitive to assets under management than mutual
fund fees. At larger asset levels, institutional mutual funds have lower fee structures than

Table 5
Estimates of active domestic equity pension fund and mutual fund investment advisory fees per $10,000 of
assets under management

Coefficient estimates
(T-statistics)

Pension fund Retail mutual
funds

Institutional mutual
funds

Investment Advisory Expense Ratio
Natural Log of Size 26.1 22.8 22.8

(25.0) (27.4) (21.7)
Intercept 66.8 NA NA

(8.6)
Large-Cap Dummy 25.1 82.3 64.9

(22.1) (35.3) (7.0)
Mid-Cap Dummy 13.5 89.7 91.2

(2.6) (37.1) (9.8)
Small-Cap Dummy 24.8 97.4 93.1

(5.8) (46.8) (10.0)
R-Squared 50.8% 15.0% 24.9%
Sample Size 220 1205 138
Assets in Billions $98 $1,713 $56

Fig. 1. Domestic Equity Administrative Expense Ratios for Active Retail Mutual Funds and Active Institutional
Mutual Funds

212 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



retail mutual funds, but do not appear to be competitive with the pension fund institutional
accounts (see Fig. 2).

The economic impact of the differential fee structures is significant. For example, for a
$500 million investment, the coefficients in Table 5 predict an investment advisory fee of 65
basis points for the large-cap mutual fund versus 24 basis points for a large-cap pension fund
account. For an initial account value of $500 million, a 30-year horizon and annual market
appreciation at 10%, the large-cap mutual fund fees would lead to an account value that was
$870 million lower than that produced by the large-cap pension fund account.

It is clear why institutional accounts are considered a best practice arrangement in the
corporate 401(k) arena: they have investment advisory costs that are one-half to two-thirds
lower than retail mutual funds and are able to deliver additional cost savings as assets under
management grow. Similarly, mutual fund’s practice of passing through the benefits of
economies of scale in administrative services, but not in investment advisory services is an
important motivator for trustees to negotiate and procure unbundled investment products and
administrative services.

Freeman and Brown (2001) conclude that the chief reason for substantial investment
advisory fee level differences between equity DB pension fund portfolio managers and
equity mutual fund portfolio managers is that advisory fees in the pension field are deter-
mined in a marketplace where arm’s length bargaining occurs. They argue that mutual fund
shareholders as a rule do not benefit from arm’s length bargaining due to a dysfunctional
regulatory system, and case law that, to date, has unduly favored advisory firms at the
expense of their captive funds’ shareholders. Mutual fund boards are typically dominated by
employees of the investment advisory firm, even though the mutual fund is to be operated in
the interests of the mutual fund shareholders.

Similar to investment advisory services, administrative services are essentially a com-
modity and there is limited product differentiation. Unlike advisory services, however,

Fig. 2. Domestic Equity Investment Advisory Fees for Active Retail Mutual Funds, Active Institutional Mutual
Funds and Active DB Pension Accounts

213K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



competition appears to be more vibrant in the administrative services arena. It is common for
small- and midsized mutual funds to purchase administrative services from third parties;
rather than from their affiliated investment management companies. The natural economies
of scale in administrative lines of business are demanded by the mutual fund companies and
then, at least in part, passed on to mutual fund shareholders.

Under MPERS trustees have a duty to, “make a reasonable effort to verify facts relevant
to the investment program.” The basic industry cost information presented above strongly
argues that trustees procure separate bids for administrative services and investment advisory
services. Trustees and consultants should be able to obtain sufficient information to identify
a normal operating profit rate for the stand-alone administrative services, given the DC plan’s
size and other circumstances. Moreover, all else equal, a large DC plan with an average DC
account balance that exceeds average mutual fund account balances should be able to use an
independent administrative vendor to meet or beat the cost structure of retail mutual funds.
As noted above, flat or declining per-capita charges for basic administration fees, reduce
per-participant administrative costs among larger plans when expressed as a percentage of
assets (U.S. Department of Labor, 1998). Separate procurement of administrative services
eliminates the mutual fund as a middle-man and any associated profit margin built into its
procurement of services for shareholders.

7. The trade-off between participant choice and effectiveness

The existence of economies of scale in institutional investment products sheds a negative
light on the practice of offering multiple versions of the same type of investment option; for
example, two active large-cap options. The common rationale for offering multiple versions
of the same option is to afford participants choice; with the expectation that participants
might be able to exit a poorly performing fund and choose a substitute that performs better
going forward. However, trustees must critically evaluate the expected marginal costs and
benefits associated with providing choice on this dimension.

Even with a sophisticated participant base, it seems unlikely that the expected benefits of
expanding the choice of duplicative investment options would materialize. There is a major
gap between the information retail investors can obtain on retail products and the information
professional staff can obtain on institutional products. In the search and monitoring phases
for institutional pension fund managers, professional staff typically gathers and analyzes
detailed information on aggregate performance and dispersion across subaccounts, invest-
ment strategies, real-time holdings data and trading activity, portfolio management, trading
and research personnel, risk management processes, trading arrangements, trading effective-
ness, and so forth. In contrast, retail mutual fund investors are generally limited to historical
aggregate return information, semiannual holdings data and historical turnover statistics.
Unfortunately, recent studies (Carhart, 1997; Kahn & Rudd, 1995; Malkiel, 1995) show that
past performance has virtually no ability to predict future performance by domestic equity
managers and has mixed results for fixed income managers.

On the other hand, the expected marginal costs resulting from expanded choice are certain
and material. First, for a given asset base, expanding the set of options compromises a

214 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



program’s ability to maintain low costs. For instance, splitting a $1 billion large-cap active
institutional mandate into three institutional mandates causes costs to rise by 34%. Addi-
tionally, under an institutional product structure, there are cost externalities created by
facilitating choice among duplicative products. These externalities are a nontrivial matter,
given the duty of impartiality: trustees’ obligation to consider the impact of design decisions
by taking into account any differing interests of participants.

Table 6 provides an estimate of the externality created when two mutual funds are
introduced to compete with an identical institutional fund. We assume that after the expan-
sion of choice, each option gets one-third of the assets that had originally been invested in
the institutional option. This assumption is conservatively based on data from Fidelity
Institutional Retirement Services Company (1999) and the recent allocation of participant
balances within the State of Michigan DC plan. For a ten year horizon, participants that
prefer the low-cost generic institutional option over the higher cost name-brand options are
forced to pay about 40% more in fees because assets move into the branded options.

This analysis also points to the unavoidable tension between participant desires, especially
the demands of vocal subsets of participants requesting more investment options, and the
trustees’ duties to the overall group of participants. Using the cost analysis from above, it is
possible to demonstrate that fiduciary duty precludes trustees from maximizing participants’
choice of investment options; to do so would be in direct conflict with the trustees’ obligation
to discharge their duties for the sole interest and exclusive purpose of providing benefits.

Imagine that a poll of participants were conducted regarding their nominees for invest-
ment options, prior to drafting an investment policy statement or conducting an education
program. It seems reasonable to expect that their preferences would resemble the retail
mutual fund market, with a smattering of bank deposits and insurance products. Adopting an
investment policy statement that incorporated each of these nominated products would
maximize product choice from each individual’s perspective.

However, the high cost of such a “maximal product choice design” would substantially
erode terminal retirement benefits for participants in the aggregate, relative to benefits
provided through a set of low-cost investment options that represent the basic asset types.
Consolidating all assets in comparable retail investment options (e.g., large-cap domestic
equities) into a single comparable retail product would create significant economies of scale
in administration and advisory fees for participants. In turn, that retail product would be
higher cost than an equivalent institutional mutual fund share class; which would be higher
cost than an institutional separate account.

Finally, in terms of investment performance, there is no basis to expect that higher cost

Table 6
Allocation of 10-year cumulative costs for $1 billion active large cap domestic equity structures. All
institutional funds and institutional/mutual fund mix

(Cost in millions) All institutional 1 Institutional/2 mutual funds

Direct Cost to Mutual Fund Investors $17.56 $71.41
Direct Cost to Institutional Users $8.78 $8.78
Indirect Cost to Institutional Users $3.60
Total Costs $26.34 $83.79

215K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



investment products (i.e., retail vs. institutional) are more effective. In fact, large plan DB
and DC practice suggests quite the opposite. Domestic equity mutual funds are rarely used
within large DB plans and EnnisKnupp (1999) reveals the vast majority of corporate DC plan
administrators believe institutional accounts are more effective than mutual funds.

8. Conclusion

New public sector DC plans face important challenges in program design. However, there
is a clear legal, fiduciary and governance infrastructure that should be adopted. With such an
appropriate infrastructure, public sector plans should largely conform to ERISA going
forward and reflect best practices in the private sector.

The major challenge facing public sector trustees is to reconcile fiduciary duties with the
desires of individual plan participants. ERISA, MPT and best practices in the corporate arena
strongly suggest that the number of investment options should be limited and an institutional
structure should be utilized where all costs are identified separately and tightly controlled. In
contrast, participants and other interested parties can be expected to argue vociferously for
an expansive list of recognizable, branded investment options. In the context of proposed
Social Security reforms, the massive stakes for retail investment firms can be expected to
generate proportionally large rent-seeking activities.

Based on empirical analysis presented above, the two views appear difficult to reconcile.
The retail approach, although intuitively appealing and familiar will result in higher costs and
lower aggregate participant benefits in the long run. Ultimately, trustees must choose
between what is right and what is easy and popular. Giving trustees sufficient independence
and subjecting them to stringent fiduciary duties should ensure that they make the appropriate
design choices.

Notes

1. A 1999 national survey of defined contribution plans by John Hancock Financial
Services indicates that less than one-quarter of respondents consider themselves knowl-
edgeable investors. Forty-one percentage believe that money market funds include
equities and 49% believe they contain bonds. Most have no strategy or goal for
allocating their retirement investments. Most say they would pull money out of equities
or reduce future allocations to equities if the stock market experienced a significant
downturn (John Hancock Financial Services, 1999). Recent academic studies indicate
that individual 401(k) investors naively diversify among the available investment
options and individual investors tend to chase the hottest performing sectors, trading
excessively and generating poor results (Benartzi & Thaler, 2001; Barber & Odean,
1999; Benartzi, 2001). A 1996 survey conducted by the Securities and Exchange
Commission and the Comptroller of the Currency found that fewer than one in five
mutual fund investors could give any estimate of expenses for their largest mutual
fund, and that fewer than one in six fund investors understood that higher expenses can
lead to lower investment returns.

216 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



2. A full information likelihood technique was used to estimate the simultaneous equa-
tions. In addition, standard errors and covariances were estimated in a single equation
context using Newey-West HAC and White Heteroskedasticity-Consistent technique.
Neither technique materially affected the overall results. The general results were also
insensitive to excluding the smallest and largest 10% of the retail or institutional
mutual fund sample. Results were generally insensitive to approaches using Morning-
star’s equity style categories to segregate the sample or develop exogenous variables.
Regressions were also performed on subsets of the mutual fund universe using simple
one variable equations in order to determine whether scale economies might differ for
large-cap mutual funds, midcap mutual funds and small-cap mutual funds. Results for
administrative expenses for retail mutual funds are generally consistent with those from
the systems approach presented in Table 4. However, investment advisory expenses
appear to differ materially for the midcap mutual funds (no economies of scale) and
small-cap mutual funds (economies of scale comparable to administrative expense)
relative to results in Table 4. Nonetheless, at $500 million of assets under management
the small-cap mutual fund coefficients in Table 4 predict an investment advisory fee of
80 basis points versus the 76 basis points predicted by the single equation coefficients.
The predicted results for midcap funds are similarly close.

References

Barber, B. M., & Odean, T. (1999). The Courage of Misguided Convictions.Financial Analysts Journal, 55(6), 41–55.
Barra Rogers Casey/IOMA. (1999).1999 Annual Defined Contribution Survey.Darien, CT.
Benartzi, S. (2001). Excessive Extrapolation and the Allocation of 401(k) Accounts to Company Stock, Forth-

coming;The Journal of Finance.
Benartzi, S., & Thaler, R. H. (2001). Naı¨ve Diversification Strategies in Defined Contribution Savings Plans,

Forthcoming;American Economic Review.
Boudoukh, J., & Richardson, M. (1993). Stock returns and inflation: a long-horizon perspective.American

Economic Review, 83(5), 1346–1355.
Carhart, M. M. (1997). On persistence in mutual fund performance.The Journal of Finance, 52(1), 57–82.
Chambers, J. C. (1999). Investment policy statements for defined contribution plans.Journal of Pension Benefits,

7(1), 29–33.
EnnisKnupp. (1999).Survey of Corporate Plan Sponsors,Chicago, IL.
Employee Retirement Income Security Act (“ERISA”) of. (1974). sec. 2(a), Pub. L. No. 93–406,88 Stat. 829

(1974) (codified at 29 U.S.C. Ch. 18), 29 U.S.C. sec. 1001(a) (1994).
Fidelity Institutional Retirement Services Company. (1999).Building Futures: How American Companies Are

Helping Their Employees Retire, A Report on Corporate Defined Contribution Plans.Boston, MA.
Freeman, J., & Brown, S. L. (2001). Mutual Fund Advisory Fees: The Cost of Conflicts of Interest.Journal of

Corporation Law,forthcoming, 25 (3).
Garla, B. E. (1998)., Investments of Governmental Plans of States and their Political Subdivisions. 415 PLI/Tax

391,Practicing Law Institute.
Garthwaite, C. (1999). Playing Politics with Retirement Accounts.American Legislative Exchange Council Issue

Analysis.
Ibbotson, R. G., & Kaplan, P. D. (2000). Does Asset Allocation Policy Explain 40, 90, or 100 percent of

performance?Financial Analysts Journal,56(1), 26–33.
Jacobius, A. (2000). NEA 403(b) endorsement shifting to security benefit as Nationwide exits market.Pensions

and Investments, 28(21), 44–45.

217K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218



John Hancock Financial Services. (1999). Insight into Participant Investment Knowledge and Behavior, Sixth
Defined Contribution Plan Survey. Boston, MA.

Kahn, R. N., & Rudd, A. (1995). Does historical performance predict future performance.Financial Analysts
Journal, 51(6), 43–52.

KPMG Peat Marwick LLP. (1998).Retirement Benefits in the 1990s: 1998 Survey Data.New York, NY. KPMG.
Malkiel, B. G. (1995). Returns from investing in equity mutual funds 1971 to 1991.The Journal of Finance,

50(2), 549–572.
Morgan Stanley Dean Witter Investment Management Defined Contribution Services. (1999).1999 Survey of the

410(k) Market,New York, NY.
National Association of Deferred Compensation Administrators (NADCA). (1998).1997 Survey of 457 Plans.

National Association of Deferred Compensation Administrators. Lexington, KY.
National Conference of Commissioners on Uniform State Laws. (1997).Uniform Management of Public

Employee Retirement Systems Act.Chicago, IL.
Pinkston, W. (2000). New Retirement Program is Created for Georgia Counties.Wall Street Journal, Southeastern

Edition. September, 20.
Useem, M., & Hess, D. (1999). Governance and Investments of Public Pension Funds.Pension Research Council

Working Paper, 99–11.Wharton School. Philadelphia, PA.
U. S. Department of Labor, Pension and Welfare Benefits Administration. (1996). 29 CFR Part 2509 Interpretive

Bulletin 96–1;Participant Investment Education,Federal Register61 (113), 29586–29590. Washington, D.C.
U. S. Department of Labor, Pension Welfare Benefits Administration. (1998).Study of 401(k) Plan Fees and

Expenses, Contract No. J-P-7–0046, Task Order 1.Washington, D.C.
Waring, M. B., Harbert, L. D., & Seigel, L. B. (2000). Mind the Gap! Why DC Plans Underperform DB Plans,

and How to Fix Them,Investment Insights, Barclays Global Investors, 3 (1), San Francisco, CA.

218 K.W. SigRist, S.L. Brown / Financial Services Review 9 (2000) 197–218