Volume3 Number1 April 2000 Investment Insights · taking inappropriate amounts of risk, either too much or too little. These shortcomings can have a tremendous impact on a participant’s

Mind the Gap!Why DC Plans Underperform DB Plans,and How to Fix Them

Volume 3 • Number 1 • April 2000

Investment InsightsCreating, transforming and sharing knowledge

∫

Mind the Gap! Why DC Plans Underperform DB Plans, and How to Fix Them

M. Barton Waring

Head of Client Advisory Group

Managing Director

Barton Waring currently runs our Client Advisory Group. He has an extensive back-

ground in investment strategy and policy, asset allocation and quantitative asset

management issues from his work with large defined-benefit and defined-contribution

plans. Prior to joining BGI in 1995, Barton held senior positions at Morgan Stanley

Asset Management, Towers Perrin Asset Consulting and Ibbotson Associates

Consulting. He frequently writes and speaks about investment strategy issues.

Lee D. Harbert

Director of Defined Contribution Business Development

Managing Director

Lee Harbert joined BGI in December 1998 as head of business development and

relationship management for BGI’s defined contribution (DC) business channel,

coming from Fidelity Investments Tax-Exempt Services Company, where he was

senior vice president for client services. Prior to joining Fidelity, Lee was senior

vice president for global securities services at Bank of America. He holds a BA

in business administration from the University of Michigan at Ann Arbor.

Laurence B. Siegel

Director of Investment Policy Research

Ford Foundation

Guest co-author Laurence B. Siegel is director of investment policy research at the

Ford Foundation in New York, where he has worked since 1994. Previously, he was a

managing director of Ibbotson Associates, a Chicago-based investment consulting

and data firm he helped to establish in 1979. He has also worked at the Marmon

Group and the American Enterprise Institute. Larry is a contributing editor of Invest-ment Policy magazine and has published over 40 articles in professional journals and

magazines. He is also on the editorial board of the Journal of Portfolio Management and

the Journal of Investing. Larry received his BA in urban studies and his MBA in finance

from the University of Chicago.

1

Investment Insights April 2000

Workers of the world, con-gratulations! You have beenpromoted to chief investmentofficer for your portion of theretirement fund!We’re kidding, of course. (Well, maybewe’re not.) But the task facing partici-pants in defined-contribution (DC) pension plans closely resembles thatperformed by chief investment officers,who make the asset-allocation andfund-selection decisions for traditionaldefined-benefit (DB) pension plans,endowments, and foundations. Thereare three major differences: (1) leadersof institutional funds have the trainingand experience to do the needed work,while the vast majority of participantsdo not, (2) they don’t have to do it intheir personal time as participants mustdo, and (3) it’s not their money. In con-trast, workers who fail at the task ofmanaging their DC plan assets willhave to live with the consequences ofretirement incomes lower than theywould have had if they had investedtheir retirement fund in competentlymanaged, risk-controlled, strategicallyoriented investment disciplines.

Because nearly all DC plan participantsare poorly positioned to make invest-ment decisions, their investment returnsare low. According to a leading benefitsconsulting firm, the returns achievedby DC plan participants have laggedinstitutional investors’ returns by 2%annually.1 Lest anyone mistake this rateof underperformance for a small number,note that $100,000 invested at 10% forthirty years grows to $1,744,940, whilethe same amount invested at 8% forthirty years grows to only $1,006,266.The missing 2% compounds to nearlythree-quarters of a million missing dollarsfor a hypothetical investor with a 30-yeartime horizon, roughly the average timebetween mid-career and mid-retirementfor today’s long-lived individuals.2

Clearly, the “little guy” has gotten verylittle benefit from the last half century’smany advances in the art and scienceof portfolio management. It’s time tochange that.

Mind the Gap! Why DC Plans UnderperformDB Plans, and How to Fix Them


2

A simple two-step solution

with the best chance of success

First, since participants are unlikely tohave or to acquire the skills needed tobuild suitable portfolios, it should bedone for them, and done right, using thebest practices of today’s most sophisti-cated investors: the large institutionalfunds. Sponsors should offer prepack-aged, mean-variance-efficient mixes of funds (that is, funds offering the high-est expected return for a given level ofacceptable risk), which provide diversi-fied exposures to all major asset classes,at reasonable levels of fees. This canclose most, or all, of the gap, at least forthose well-informed participants smartenough to choose this wise option. Afterall, how is this really different from manyof the purchases we make today?

A second, but equally important task, isto get participants to invest in the pre-packaged mixes (since we can’t requirethem to do so, and some participantswill still choose the traditional types ofDC fund offerings). By refocusing par-ticipant communications on investmentstrategy, we can hope to move themaway from their current focus on theplethora of fund choices now beingoffered and toward the more importantdecision: the total portfolio solution.

What are the “best practices” of institu-tional investors? We describe them inmore detail below, but basically they

consist of getting the steps of the invest-ment process in the right order. Thevery first step for institutional investorsis to determine their investment strategy.This means using the teachings of finan-cial economics to estimate expectedreturns and risks for each asset class,calculate the efficient frontier, andspecify an asset-class mix—all top-level decisions. Only when the strategyis completely in hand, do they move tothe more tactical issues of fund selec-tion and implementation. In sum, bestpractices are about asset allocation orinvestment strategy, appropriateness of investment vehicles chosen, andmanagement of fees and costs.

The DC-plan investment process usuallygets this backwards, with the choice offunds given much more focus and prior-ity than the decision about investmentstrategy. We would describe today’s typ-ical DC plan as “fund-oriented,” ratherthan “strategy-oriented.” When a plan isfund-oriented, the investment strategyis an accidental by-product of fund selec-tion, and is almost inevitably flawed. Themany symptoms of the “performancedisease” of DC plans—low returns, highcosts, inattention to risk—can be tracedto this reversal of priorities.

3


1. Explaining the shortfall

Strategy reasons for

underperformance

Generally, DC participants have twoseparate, but related, problems in form-ing suitable portfolios. The first problemis that they are unable to get to the “effi-cient frontier” because they don’t haveaccess to the requisite tools to success-fully build diversified portfolios offeringthe highest expected return for eachgiven level of risk—as a real strategicapproach to investing would accomplish.The second problem is that they aretaking inappropriate amounts of risk,either too much or too little. Theseshortcomings can have a tremendousimpact on a participant’s nest egg overa typical 10-to 20-year investment horizon (Figure 3, page 11).

A. Investors’ portfolios are off

the efficient frontier

Despite attempts by leading providers ofinvestment funds and DC administrationservices to educate participants, themost significant obstacle to efficientinvesting by plan participants is lack ofinvestment knowledge. The techniquesand suggestions designed to help partic-ipants create sensible investment strate-gies have simply not been successful.In the modest amount of time the over-whelming number of participants arewilling to give to the problem, they areunlikely to learn the tools mastered byonly a relative handful of sophisticated

Underinvesting Participants often contribute less to the plan than is

needed to achieve their goals. Underinvesting is not a

source of low returns per dollar invested—it does not

contribute to the missing 2%—but it is worth a mention

because of the tremendous contribution of underin-

vesting to the broader problem of DC plans failing to

fulfill investor goals. Because one can accomplish only

so much by varying one’s investment aggressiveness,

no asset provides the return needed for many plan

participants to achieve the level of retirement income

they expect. The rest of any shortfall must be—can

only be—made up by additional contributions or by

accepting a lower level of retirement spending.

To the extent the actual or potential legal liability of

the defined-contribution sponsor is created by failure

to persuade investors to invest enough to meet retire-

ment needs, underinvesting is or should be a source of

concern to sponsors. Sponsors should be aware of the

responsibility to provide plan participants with sound

advice on how much to invest, what rate of return can

be reasonably expected, and what one’s end-of-career

wealth and income are likely to be. Many sponsors

engage in such an effort now, but the effort is general-

ly ineffectual, with participants’ investment rates said

by some observers to be bimodally distributed—one

large group invests the maximum permitted, while

another large group invests the minimum. While

employers should understand that personal circum-

stances differ and not everyone can invest what would

ideally be required, the employer should play a leader-

ship role in encouraging employees to maximize their

retirement-plan contributions.


4

investment advisors, who have had theadvantage of graduate education andyears of training. Even if participantscould overcome this obstacle, there areother obstacles to building strategiesthat provide the same efficiencies avail-able to the serious institutional investor.

Until recently, the historical choice setfor DC participants typically consistedof guaranteed investment contracts(GICs), a balanced fund, a growth fund,and, only very recently, an index fund(all US only). Such a choice set simplydoes not contain the asset-class buildingblocks that are required to constructmean-variance efficient portfolios. Noresponsible defined-benefit pensionplan sponsor would ever build a portfolioout of these “primitive” pieces.

While the list of the “best” asset-classbuilding blocks merits an in-depth dis-cussion of its own, a list representingthe minimum types of funds that can be considered reflective of institutionalbest practices, with likely benchmarksin parentheses, might be:

• Large-capitalization US equities (S&P 500 Index)

• Smaller-capitalization US equities(Russell 2000 or various extendedmarket indices)

• International equities (MSCI EAFEor ACWI Indices)

• Diversified domestic fixed income(Lehman Aggregate Index)

• Cash

…where each building block is itselffully diversified, and managed at a reasonable cost. For each asset class,institutional investors go to a great deal of trouble to align their managers’benchmarks to their own to avoid glar-ing “benchmark misfits,” and to keepthe level of active risk and the level offees reasonably low. In an environmentwhere fund choices predominate, theseimportant concerns are completely lost.

Even so, the recent trend has been forDC sponsors to enrich the choice set toinclude a greater variety of equity andbond funds, including international,high-yield, and growth and value styles,and sometimes even multiple managerswithin an asset class. This improves theinvestor’s situation, but fails to addressthe basic problem of poor asset alloca-tion, caused by the emphasis on fundselection rather than on investmentstrategy. Under the current system,investors typically don’t even know theyare making risk or asset-class choices.Because investors are allowed to focuson making fund choices, most partici-pants do not build portfolios that have a sensible investment policy, despiteexpensive efforts to educate them.

5


In addition, the longer lists of fundsnow often offered are dominated byactive funds with high fees. While insti-tutional investors have long relied onlow-residual-risk “core” investments ineach of their asset classes to reduceboth risk and expense, DC participantshave not been encouraged to mimicthis sensible design. So, even with thismore fully fleshed out choice set, thechoices would not be comfortable tomost sophisticated defined-benefit plansponsors if they were required to usethem for their DB plans.

Figure 1 shows two frontiers: the lowerone constructed from traditional DC-plan components (A), and the higherone constructed from the list we saidwas representative of institutional bestpractices (B). If investors are onlyoffered the portfolio choices used toconstruct frontier A (we are reluctantto call it an “efficient” frontier), regard-less of their individual skill, they can-not reach the true efficient frontier (B)and will underperform over time forthat reason. In Figure 2, which enu-merates the many sources of underper-formance in typical DC plans, we callthis underperformance the loss frominadequate fund opportunities. Evenworse, investors in traditional DC planscannot even reach frontier A since theywould have to hold an optimal combi-nation of the funds composing it, andthey do not have the investment strategy

EX

PE

CT

ED

RE

TU

RN

R ISK

Cash

GICs

S&P 500

Index Fund

Growth

Fund

Company

Stock

FRONTIER A

A. FRONTIER OF FUND CHOICES IN TRADITIONAL DC PLAN

EX

PE

CT

ED

RE

TU

RN

R ISK

Cash

Diversified Domestic

Fixed Income

S&P 500

Index Fund

Int,l Equity Ł

Fund

Small-Cap

US Equity Fund

FRONTIER A

B. FRONTIER OF FUND CHOICES IN WELL-MANAGED DB PLAN

FRONTIER B

Figure 1


6

knowledge or tools to do so; we call thisthe loss from inefficient use of existingfund opportunities.

B. Investors take the wrong amount

of risk

Many investors allocate their DC invest-ments in ways that are completelyunsuited to the achievement of theirinvestment goals. By far, the most com-mon misallocation is to take too littlemarket risk, with the bulk of one’sassets in cash GICs and other stable-value investments. Workers young andold are affected by this folly, as areskilled workers and executives in“sophisticated” companies that sell oruse advanced technology or financialservices. These investors hold little ornothing in stocks, and even their hold-ings of bonds are often quite limited.Because the return on cash and cash-likeinstruments is often many percentagepoints below that on more appropriatemixes that include stock and bond funds,investors who take too little market riskmay experience a shortfall far greaterthan 2% over a long time horizon (see“Underinvesting,” page 3).3

A less common problem is excessiverisk-taking. A small but visible group of investors seeks extraordinary returnsthrough a 100% equity allocation in theirDC plans, often using the plan’s fund-switching mechanism to pursue the hotfund of the week. If the plan is heavilyfunded with company stock and that

stock has been a strong performer, theseparticipants may refuse to diversify,confusing future with past performance(the board of directors is often happilycomplicit in building this very undiver-sified exposure). If a DC plan has a self-directed brokerage option that allowsparticipants to invest in individualstocks, these investors often hold morespeculative positions such as biotechand Internet startups. Because undiver-

COLOR KEY:

Effect of macro fund structure

Effects of behavior of poorly educated participants

Effects of active management

Fees and transaction costs

SOURCES OF UNDERPERFORMANCE IN TYPICAL DEFINED-Ł

CONTRIBUTION PLANS RELATIVE TO WELL-MANAGED Ł

DEFINED-BENEFIT PLANS

Loss from inadequate fund opportunities

(frontier A below frontier B)

Loss from inefficient use of existing fund opportunities

(investor does not even reach frontier A)

Loss from insufficient risk-taking

(even if investor could reach frontier, chooses wrong point on it)

Loss in utility from taking active risk with no payoff

(even if active managers average to index performance with index fee)

Underperformance of active management

(compared to index or to well-chosen active funds)

Fees and hidden administrative costs

(compared to institutionally negociated fee structure)

Transaction costs: commissions, spreads, and market impact

(compared to index or to low-turnover active strategy)

Figure 2

7


sified portfolios rarely perform well inthe long run, at least on a risk-adjustedbasis, investors in this category are apart of the underperformance problem.

If one looks at aggregate asset-allocationdata for DC plans, the mix is not bad—about 70% in stocks (including companystock) and 30% in fixed-income, stable-value, and cash instruments.4 But if onedisaggregates the data and looks atindividuals’ allocations, extreme posi-tions such as those described above areshockingly common. The future wealthdivide between good investors, on theone hand, and poor investors (whetherin cash or risky stocks), on the otherhand, could become a serious social andworkplace problem. Figure 2 assumes,for the sake of illustration, that a giveninvestor takes too little, rather than toomuch, risk and refers to the result asthe loss from insufficient risk-taking.

Implementation reasons for

underperformance

A. Too much active management

Many plan sponsors offer only activelymanaged funds to DC investors. Othersprovide both active and indexed choic-es, but offer them in a format that failsto highlight the advantages of usingindexed and risk-controlled active fundsversus concentrated active funds as theprincipal sources of exposure to an assetclass. The typical result is a participantportfolio that is far too skewed towardactive management, causing the portfo-

lio to have a higher-than-appropriatelevel of residual risk and higher-than-appropriate fees. Moreover, since planparticipants tend to choose “hot” funds,and “hotness” in funds is often style ormomentum driven, participants’ portfo-lios often contain active bets that arehighly correlated across managersrather than being diversified in termsof their active risk.

Active management is difficult enoughfor institutional investors, who haveaccess to extensive manager databasesand can interview managers personally;nonetheless, active managers rarelybeat the indices in a consistent manner.(This is understandable because activemanagement is arithmetically a zero-sum game before fees are taken intoaccount; after fees and other investorcosts, the sum total of all active man-agers must underperform the index.5) It is difficult to imagine that individualDC plan participants can win at thisgame. To the extent that DC participantsshould invest with active managers, itshould be as part of the prepackaged-fund mixes referred to earlier.

B. Quality of active management

is too low

If DC participants invested in averageactive managers, the damage would belimited because these managers inaggregate would deliver index-like per-formance, minus fees and transactioncosts. However, the problem is worse


8

than that. Many DC sponsors select theirfund provider on the basis of ancillaryservices (recordkeeping and administra-tion), rather than on the basis of expect-ed active management performance. Tothe extent successful active managerscan be identified, and this method mostcertainly does not do so, one should notbe too surprised that the resulting man-ager choices tend to be subpar perform-ers. Plan participants would be betteroff in an all-indexed position if thesponsor is unable to put the rightresources and talent into the selectiondecision, free of ancillary constraints.

Figure 2 refers to the underperformanceof active management, relative to an ap-propriately chosen benchmark, as asource of loss in typical DC plans. Un-derperformance aside, the presence oftoo much active management in a plancauses another type of “loss,” namelythe loss in utility from assuming activerisk with no expected return payoff, sowe display this subtle—but real—effectas a separate category in the exhibit.

C. High commissions and fees

High investor costs, including managerfees and the cost of transacting, are yetanother source of underperformance inDC plans. Mutual funds, which current-ly form the basis of practically all DCinvesting, have fee structures which canbe twice as high as those of otherwisecomparable institutional funds.6 If notproperly monitored and managed, fees

and transaction-related costs can easilysum to 2% or more—occasionally muchmore—per year (Figure 2). Transactioncosts have several components, of whichdirect costs (brokerage commissions andbid-asked spreads) are only a small part;market impact and the opportunity costof delays in trading and missed tradesare hidden costs that can be a multipleof the direct costs.7 While the lack of aninvestment-strategy focus is not directlyresponsible for the high level of investorcosts in DC plans, the neglect of a policycreates a culture in which high costs areeasily hidden or ignored.

D. Hidden administrative costs

Finally, many sponsors choose mutualfunds with retail price structures inreturn for the mutual fund agreeing toprovide, at low or no cost, the record-keeping and administration required bythe plan. This is most typical at planswhere the sponsor has agreed to paythe recordkeeping costs, while the invest-ment management fees are chargedback to the plan’s participants. The“free” recordkeeping feels like a smartbudgeting move. This has been carriedto its logical next step in many cases,with sponsors demanding “rebates”from independent fund vendors to beused by the sponsor as a contribution to administration costs.

This is not a healthy arrangement. Thereis a huge difference between retail andinstitutional pricing for comparable

9


funds, and DC plans typically have thesize to gain access to low-cost institu-tional pricing structures. The excessfees, which constitute the price actuallybeing paid by the participants for admin-istration, are very high and get higherthe larger the per-person account bal-ance. By comparison, most third-partyadministrators would be charging on aper head basis, not based on the assetsunder management.8

It doesn’t take much reflection to seethat these corporations have negotiateda benefit for themselves—a low or zeroadministration fee in return for a hiddencost that participants have to pay. It’s a fiduciary issue; the sponsor has suc-cumbed to a conflict of interest to thedetriment of its plan participants.9 Inthe meantime, excess fees are fullyequivalent to lower rates of return.

2. Where did we go wrong? Root causes of the problem

DC fund structure as

a historical artifact

In the early days of DC plans—that is,through the 1970s and much of the1980s—many of these plans were pro-vided by insurance companies. Insurersearly on invented the guaranteed invest-ment contract (GIC), and it was naturalthat fixed-interest annuities and GICsdominated the market. Because neitherannuities nor GICs are marked-to-mar-ket (in the sense of adjusting portfolio

valuations for interest-rate fluctua-tions), these types of products give theparticipant a mistaken impression oflittle or no risk. In fact, the word “guar-anteed” in GIC is not, as widely mar-keted, a guarantee against risk, butmerely a guarantee by the insurancecompany issuing the contract not toreduce the stated rate of return. Thereis no third-party guarantor in the normalsense of the word. Guaranteed invest-ment contracts can be risky, as wasproven conclusively by the defaults ofExecutive Life and Mutual Benefit Life,both major carriers with top ratings.10

Some funds also contained companystock, which was seen as aligning theinterests of employer and employee,and was sometimes naively regardedas “free” from the point of view of thecorporate treasury. To this motley mix,a money-market fund and a few equityor balanced mutual funds were some-times added.

Amazingly, the GIC-centered structuredescribed above was dominant well intothe 1990s, by which time many defined-benefit plans—sometimes at the samecompanies!—had been very progressiveand well managed for years.11 (Longbefore that time, the defined-benefitplan community had established theprimacy of investment policy and strat-egy over manager selection issues, andwas starting to diversify into interna-tional, small-capitalization and value


10

stocks, multiple styles of fixed income,and other carefully chosen asset class-es.) Around 1987, Fidelity Investmentsbegan a strong play to convert sponsorsto a DC structure based on mutual-fundchoice, but that approach took a fewyears to catch on. By the early 1990s,however, this mutual-fund revolution inpersonal investing had made a remark-able inroad into DC plans. By the endof the 1990s, participants in most planshad many funds to choose from, but little useful help in building sensible,efficient portfolios.

As a result of this evolution, the DCworld that emerged during the 1990s isvery much fund oriented, not strategyoriented. Participants still have almostno access to strategy advice or to pre-packaged, well-diversified and efficientportfolios of asset classes.

Lifecycle or lifestyle funds, which evolvedin 1990s as a potential solution to theproblems we have identified, are capableof making up this DB/DC gap, but havenot yet successfully done so. Most arepoorly engineered by portfolio managerswho know more about picking stocksthan about total portfolio investmentstrategy. Furthermore, when lifecyclefunds are offered, they are usually com-municated poorly, and the concept thatsuch offerings provide a complete invest-ment strategy in a single fund is rarelyunderstood by participants with anydegree of success.

This haphazard evolutionary path clearlyhas not produced a structure conduciveto participants building efficient portfo-lios. The fund-oriented approach pushedby the mutual fund houses (which haslargely replaced the GIC-centeredapproach of the insurance industry)gives investors access to many specialtyfunds, which at best divert the investorfrom the asset allocation work to bedone. At worst, the profuse fund choicesencourage high fees, poor diversification,and uncompensated risk-taking by inves-tors chasing recent past performance.

In most DC plans, then, there is noemphasis on making sure all the build-ing blocks needed to construct efficientportfolios are even available for use,much less that they are used appropri-ately by each participant. Moreover,index funds, which are very effective at providing asset class exposure withlow residual risk and substantial cost-savings, when used as a core invest-ment, are used lightly and without supporting information.12

Failure of participant education

As DC sponsors built their plans, theyrecognized that participants would haveto be educated on the principles ofinvesting. To this end, much earnestlabor has been performed. The outputof written, spoken and electronicallydelivered educational programs, as well as asset allocation and financialplanning software, has been massive.

11


Despite this effort, the majority of par-ticipants today still have a poor grasp ofinvestment basics. Many participantsdo not know what an asset class is,much less the expected return and riskdifferences among them. They typicallyunderstand little about the need fordiversification and a long-term perspec-tive on asset allocation and fund perfor-mance. Few participants are aware ofthe fees being charged, or of the costand performance differences betweenactive, indexed, and blended investmentstyles. Moreover, participants oftenhave unrealistic market expectationsthat drive their investment decisions.

Even though educators are usually competent and investors try to exercisecommon sense, the academic principlesunderlying investment theory are tech-nical and require extensive learning.Unfortunately, plan participants rarelybudget the amount of time needed tobecome proficient as an investor. As we noted at the outset, chief investmentofficers and other professionals spendyears learning these principles, andstill do not always get everything right.DC participants, whose career-basedknowledge is usually unrelated toinvestment knowledge, are not about to master this trade in a few hours ofemployer-sponsored (and usuallyoptional) training.

As evidence of the inadequacies of par-ticipant education, and as proof that DCinvestments are managed well belowthe normal standards of institutionalinvesting, simply look at the number of funds chosen by participants in anyDC plan with which the reader is famil-iar. Normally, most employees will haveselected just one, two or three invest-ment choices as suggested by the data inFigure 3, which details the experienceof one prominent corporate plan sponsorand represents what we’ve learned fromsponsors with whom we’ve met. Withthis degree of portfolio concentration,clearly the message of diversification

DIVERSIFICATION OF 401(k) ACCOUNTS

0

5

10

15

20

25

30

PE

RC

EN

TA

GE

(%

)

1 Fund 2 Funds 6 Funds5 Funds4 Funds3 Funds CompanyŁ

Stock Only

Figure 3

As of December 31, 1999.

Source: BGI, February 2000.


12

is not getting through, even at the mostbasic level, and concepts of optimal diver-sification, so important to sound invest-ment strategy, are not even within reach.

3. The cure

Outline of a desirable product mix:

“Do it for them”

Any number of fixes for the problemsidentified in the previous sections canbe imagined, but a shot through theheart of the problem is the simplest andbest solution: Build well-engineered,complete investment strategies—fami-lies of pre-mixed asset allocation funds.Then, encourage participants to buy intothese funds by communicating that suchfunds are fundamentally different, andfundamentally better, than traditionalfund choices.13 For this message to besuccessfully conveyed and acted upon,these funds must be marketed and intro-duced to the participant as a premierinvestment selection, elevating these op-timal strategies to a prominent position.

What follows is a design for a DC prod-uct mix that puts strong emphasis onthe approach we believe is best for mostinvestors. Because pre-mixed asset allo-cation funds do not meet the needs anddesires of every investor that a DC planis expected to serve, our design keepsother choices available.

A. The solution for most investors: pre-mixed efficient portfolios

As we just indicated, the core of ourapproach is to offer, and aggressivelymarket, pre-mixed portfolios of assetsat different risk levels. Following insti-tutional practice, the portfolios are con-structed by:

• First, identifying relevant asset classes,setting capital market assumptions,and computing the efficient frontier

• Second, from among the differentinvestment strategies represented by each point on the efficient frontier,choosing the one having the targetedor budgeted risk level

• Finally, “staffing” each asset class in the mix with an appropriate fundor funds.

(This, of course, is just a bare outline ofthe institutional approach to portfolioconstruction; we provide more detailbelow in the section on implementation.)

Usually five of these strategic portfolios,spaced along the efficient frontier fromconservative to aggressive, should beenough to match the preferences ofalmost all individual investors. Figure 4(see page 13) shows illustrative mixesfor five risk choices.

13


Following Waring and Castille (1998),each asset class will be “staffed” with a mixture of indexed and active funds.14

The proportions would be decided onthe twin considerations of the sponsor’sconfidence in its skill at selecting activemanagers, and the need to meet a “bud-get” for active risk. It is important tokeep products simple and costs low. For example, a major asset class such asUS large-capitalization equities could berepresented entirely by an index fundfor the lowest cost and lowest activerisk. Or, if there were some confidenceby the sponsor in its ability to selectactive managers, it might be representedby 50% in an index fund plus 25% eachin two actively managed funds. Thiswould produce an active risk in the 2%region, which could be more or lessdepending on the specific characteristicsof the active managers chosen. In con-trast, an all-active manager structurewill have upwards of 4% active risk atthe asset class level if it includes two or three traditional active managers of average individual risk.15

A pre-mixed approach is well suited for participants who lack the time, theknowledge or the interest to developtheir own strategies. Likewise, thisapproach also serves the interest of thetruly sophisticated investor, who appre-ciates well-engineered investmentstrategies and is happy to take advan-tage of them when offered. Finally, the

pre-mixed approach helps the sponsorto communicate with confidence thatinstitutional “best practices” are beingoffered to participants, and that partici-pants can reasonably expect to earnreturns comparable to those earned bythe world’s best-managed investmentinstitutions. Participants will have effi-cient portfolios, maximizing expectedreturn at their preferred level of invest-ment risk, just as large investors do.

Figure 4

EX

PE

CT

ED

RE

TU

RN

R ISK

PRE -MIXED FUNDS

A

B

C

D

E

C B

I

B

C

I

L

S

C

LB

IS

LB

IS

LI

S

Asset mixes are for illustrative purposes only.

C—Cash

B—Diversified domestic fixed income

L—US large-cap equities

S—US small-cap equities

I—International equities

L

B. For people who want to determinetheir own asset-class mix: building

blocks by asset class

Of course, sponsors often cannot requireparticipants to invest in these pre-mixedportfolios without other options, andthere is a strong minority of participantsthat want to build their own portfolio. Toencourage the “mix-your-own” investorto remain oriented to mixing a strategyrather than just picking a bunch offunds, it is useful to offer a set of well-designed asset-class building blockssuch as those outlined in the previousparagraph. Index funds, carefully chosenfunds of funds, or other funds designedto be clean implementation vehicles forthe key asset classes, with low costs andlow active risks, are appropriate here,and it makes sense to have the sameones used in the pre-mixed strategiesavailable for the mix-your-own investor.

C. Specialty and legacy funds

Another category much lower ininvestment utility, but often politicallyrequired, consists of certain specialtyfunds and legacy funds. It is difficult to“take away” existing funds from partici-pants in the DC world, and yesterday’spoorly chosen funds may have an audi-ble constituency that can’t be ignored.The current literature often refers tothese active investors who make up thisaudible constituency as “sophisticated,”but the term is probably a misnomer.Certainly it is a misnomer if one acceptsthe prevailing wisdom in institutionalinvesting that strategy is paramount,


14

and that rapid fund switching and placing of large “bets” are unlikely to be productive over any reasonably longperiod of time. For the truly sophisticatedinvestor, the focus will be on a mix ofthe major asset classes held with a highdegree of stability over time: an efficientinvestment strategy.

For sponsors that need to serve thisactive group, however, specialty funds—such as real estate, growth- or value-biased funds, technology funds, andpopular legacy funds that don’t fit wellinto the strategic framework—can beoffered if positioned at an eye levelbelow the basic strategic options,which in effect are on the “top shelf.”This helps to minimize confusion forthe majority of the population, who arebetter off with their attention focusedon strategic solutions.

Finally, to accommodate participantswho insist on an even greater variety of choices, some sponsors may feelcompelled to offer (or to continue tooffer) a “window” into a whole mutualfund family, or even a brokerage win-dow. It is probably just as well to avoidthese where possible, as the high-risk,high-cost investing often promotedthrough these routes is not well suitedto the long-term investment require-ments of a personal retirement plan.While potentially exceptional returnscan be generated, they are unlikely tobe sustained over long periods of time.Again to minimize confusion with the

15


basic strategic mix and asset classchoices, such windows should be com-municated at an eye level below the top shelf, and care should be taken toinform participants of the higher risksand costs often associated with themarkets available in these windows.

Implementation details:

Incorporating institutional

“best practices”

The very structure of our proposed setof five mixes on the efficient frontierhas an institutional flavor; but to deliverbest results, the process needs to mimicinstitutional best practices to the great-est extent possible. “Institutionalizing”the process consists of adopting profes-sional approaches to:

• Identify the opportunity set of assetclasses

• Develop expected returns, risks, andcorrelations

• Identify suitable risk levels• Allocate among asset classes • Select funds or managers • Rebalance the portfolio and evaluate

performance

Identification of opportunity set of assetclasses. While typical DC plan imple-mentation skips the strategic tasks andgoes right to the selection of funds, well-managed institutions begin by figuringout what broad asset classes need rep-resentation. The institutional investortypically considers all of the major

components of world market capitaliza-tion to be in the initial opportunity set.Judgment is then used to modify thisopportunity set to take liquidity andinvestability considerations into account.For example, emerging market equitiesand high-yield bonds (which have someliquidity problems) often make the finallist, while farmland (which is illiquidand difficult to invest in, despite a largemarket capitalization) almost never does.

Development of expected returns, risks,and correlations. The professional insti-tutional investor then creates estimatesof the return and risk characteristics ofeach asset class, as well as the expect-ed correlation of each asset class withevery other, so that the efficient fron-tier can be calculated. While not allinstitutional portfolios are managed bydirect application of these quantitativemethods, such methods inform thestrategic process of every well-man-aged institutional portfolio.

Identification of suitable risk levels.Defined-benefit pension plans start bymodeling their liabilities, which consistlargely of post-retirement payments tobeneficiaries. The goal of the plan is tomaximize the plan surplus (assets minusliabilities), subject to various risk avoid-ance parameters. The typical choice of arisk level for plan assets is an outgrowthof an asset-liability study that takes allthese factors into consideration.


16

Individual investors also have “liabili-ties,” consisting of their post-retirementcost of living, but every individual isdifferent, and (in contrast to the defined-benefit world) the sponsor must comeup with more than one mix. Ratherthan study plan participants’ liabilities,the DC sponsor can simplify the prob-lem by recognizing that individuals’differences—while originating in fac-tors such as personality, time horizon,career prospects, the holding of otherassets outside the DC plan, and soforth—largely come down to differ-ences in risk tolerance. The number of distinct variations—which are bestthought of as risk levels—is probablyno more than five. These risk levels canbe mapped into efficient portfolios ofthe asset-class building blocks.

Thus, the many and varied clients of theDC plan sponsor can be well satisfiedby a small number of asset mixes. Likesock sizes, a few well-managed assetmixes positioned at intervals along theefficient frontier are good enough; thesefew positions on the “risk tolerance dial”will satisfy all but the most exactingcustomers.16 Certainly, the results willbe a huge improvement over thoseobtained from current practice.

Asset-class allocation. Once the risklevel has been determined, asset alloca-tion is a matter of identifying the asset-class mix providing the highest level ofexpected return for the risk taken. Themean-variance optimization approach

that we touched on earlier provides aquantitative solution to the problem andis the method of choice for most institu-tions. While the specification of an exactmethod for asset-class choice is beyondthe scope of this article, we emphasizeonce again that DC plans should emulatethe institutional best practice, which isto first get the asset-class allocationright, then “staff” each asset class with a mix of appropriately chosen managers.

Fund selection. Unlike many individuals,well-managed institutions have devel-oped fund- and manager-selection dis-ciplines that are more sophisticated and effective than buying the previousperiod’s hot strategy. The first “screen”used in selecting a fund is to determinewhether it really gives the investorexposure to the asset class it is sup-posed to represent. This is followed bya second “screen” which considers thefollowing criteria: the various measuresof risk, the magnitude and consistencyof past return relative to a well-chosenbenchmark and the likelihood that goodpast performance will be repeated, thepeople and process involved in manag-ing the fund, and the reasonableness offees and transaction costs.

Since the surest way of increasinginvestment return is to reduce costs,fees and transactions costs are critical.Because mutual fund fees are retail, andthus high, they don’t reflect or takeadvantage of the quantity buying powerof a large DC plan. It is better to use

17


institutional commingled funds, whichare much less expensive and typicallyget even less so with greater size.Transaction costs, including marketimpact costs, go up with turnover andsize, and can easily exceed 1% per year,all well-hidden in the returns.17 While itis popular for marketers of funds withgood recent historic track records topoint out that their fees have been wellearned, Sharpe’s argument, discussedabove, makes it clear that only the mostskillful of active managers will surmounttheir fee-and-cost hurdle over time. Thus,institutional best practice focuses veryclosely on reducing fees and turnover.

In our proposal for DC plans, each asset-class “building block,” or fund of funds,should be assembled using these criteria.Because active fund managers rarelybeat their appropriately chosen bench-mark consistently over long periods oftime, and because fees are an importantconsideration, index funds figure promi-nently, a feature also reflective of bestinstitutional practice.18 Active fundsshould be included only to the extentthe sponsor has a strong belief that thefund will outperform its benchmarkindex. In such cases, as noted earlier,an asset-class building block might consist of 50% in an index fund and25% in each of two carefully selectedactive managers. Tactical asset alloca-tion components can be used as well.They are best regarded as active man-agers managing against the asset-allo-cation benchmark, and should be used

only within budgeted levels of activerisk and, like all active managers, onlywhen there is a strong belief in contin-ued above-benchmark performance.

Rebalancing and performance evaluation.The creator of the pre-mixed portfoliosneeds to perform ongoing maintenanceon them so that the plan participantdoes not have to. One element of main-tenance is regular rebalancing to thepolicy weights. In addition, changes inasset allocation and fund selection canbe made when there are opportunitiesto improve any of the underlyingassumptions used to create the existingstrategies. The last quarter century’smany advances in quantitative perfor-mance evaluation and attribution areeasily adapted to the management ofpre-mixed portfolios for DC plans.

Having talked much about institutional“best practices,” we should be clear thatthere is widespread agreement on whatthese are. While plenty of room for dis-agreement about subsidiary details ofthe process exists, there is a strongconsensus that one should select theasset classes, develop long-term assump-tions, calculate the efficient frontier,manage costs tightly, and mix activeand indexed approaches—but choosingactive only if it is believed to be superi-or. All the first-order issues have broadacceptance in actual institutional prac-tice, and should be incorporated into DCpractice for the benefit of individualinvestors. The democratization of invest-


18

ment strategies previously accessibleonly to the wealthy proceeds from theseprinciples. As we said earlier, it’s allabout asset allocation, appropriatenessof investments, and fees.

Educating the participant

The task is not only to design a well-engineered DC plan structure, but to get people to buy into it. No one can be forced to invest in the structure wehave proposed, no matter how well theinvestment vehicle has been designed.To this end, participant education, which,as we noted earlier, has largely been afailure, needs to be recast. Even wherelifecycle funds have been offered, exist-

ing education paradigms don’t tell par-ticipants why these funds have a specialplace in the investment universe (toprovide a complete, optimal investmentstrategy with one fund purchase deci-sion), so people don’t buy them. Theyappear to be just one more set of fundson a long list, indistinguishable fromsingle-purpose funds and having nothingspecial about them.

In contrast, Figure 5 provides the frame-work for a better approach to participanteducation that highlights pre-mixed assetallocation funds. The exhibit leads theeye to these funds first, which, as wehave emphasized, will serve the partici-

Conservative Aggressive

Participants choose from the following complete investment strategies:

Build-your-own strategy

Asset class building blocks

US large-cap equities

US small-cap equities

International equities

Diversified domestic Ł

fixed income

Cash

Specialty funds

Brokerage window

Company stock

Active mutual funds

Sector funds

Stocks

Bonds

Cash

Mix A Mix EMix DMix CMix B

Select from theŁ

following choices:

Figure 5

19


pant as well as we know how. This promi-nent positioning is well justified becausethe pre-mixed funds represent complete,well-engineered investment strategies,superior in both kind and character tothe other available options. They are onthe efficient frontier, and thus have risk/return characteristics superior to thoseof any single-purpose fund. They are theoutcome of a thoughtful effort to mixsingle-purpose funds into an optimalblend using technology and approachesrepresentative of the best practices ofthe institutional investor.

For those who don’t want a pre-mixedstrategy, we focus them on a mix-your-own strategy, where asset-class buildingblocks identical or similar to those usedin the pre-mixed choices can be mixed bythe participant as desired. By calling it a“mix-your-own strategy,” the message ismade clear that it should be a strategy,and that it should be a diversified mix.For the remaining small percentage ofparticipants who are not interested ineither option, almost anything else canbe offered as long as it doesn’t muddythe primary message, which is thatstrategic choices are strongly preferred.

The education program must emphasizeto participants that they can enjoy thesame returns enjoyed by institutionalinvestors, at appropriate risk levels, onlyby investing in a mean-variance efficientportfolio. By orienting the investmentvehicles and the communications pro-gram to issues of investment strategy

rather than to issues of fund selection,the DC sponsor is far more likely tomotivate participants to choose one ofthe pre-mixed portfolios such as portfo-lios A through E in Figure 5, therebyincreasing their odds of successfullyavoiding the hazards of investing.

3. Every participant a chiefinvestment officer

While “every participant a chief invest-ment officer” is an admirable sentiment,the realization thus far has been terri-ble. Defined-contribution plan assetshave consistently underperformed theirdefined-benefit plan brethren becauseof the failure to focus on investmentstrategy. A DC plan that is designed to be top-down and strategy-oriented,rather than fund-oriented, will deliverhigher expected-return-to-risk ratios.

Why hasn’t the existing system servedDC participants well? Much effort hasbeen expended trying to turn partici-pants into chief investment officers,including risk-tolerance questionnaires,asset allocation and financial planningsoftware, and a vast bulk of educationalmaterials. These efforts rarely workbecause participants simply lack thetime, the knowledge, or the interest tobecome professional-quality experts ininvestment strategy. Recent experiencewith consultants indicates that it costs$35,000 to $250,000 in consulting feesand staff time for a corporate pension


20

sponsor to develop an investmentstrategy for their defined-benefit plan.Replicating the quality and profession-alism of the defined-benefit result wouldrequire more resources than the vastmajority of individuals would or couldmuster for this effort. Moreover, it isunlikely that many participants are will-ing—or able—to give more than a tokenamount of time to making their choices.

What we advocate—to refocus DC planinvesting on the strategy decision ratherthan the fund-choice decision—can beaccomplished with only two require-ments. First, the family of pre-mixedstrategic asset allocation funds must be well-engineered and worthy of beingput forward as representing the bestpractices of sophisticated investors.Second, the communications programneeds to concentrate on the strategicdecision and de-emphasize the manyfunds that might be available.

In addition to the long-term risk andreturn benefits, other benefits ofprepackaging a well-engineered familyof pre-mixed strategy funds include:

• professional engineering of the optimalstrategic asset mix itself

• professional implementation, including- determination of the active/

index mix- professional choice of active

managers, if used

- professional tactical decision making, if used

• systematic rebalancing• institutional-level fees

…and much more. And while the focusshould remain firmly on the strategicchoices, there are options available forparticipants who insist on determiningthe asset-class mix themselves (fund-of-funds building blocks by asset class),and for participants who insist on deter-mining the fund mix (sponsor-selectedactive and indexed funds).

At great cost to participants, DC-planinvesting has retained the baggage of itshistorical development as an insurance,rather than an investment product.Meanwhile, institutional investors havedeveloped a highly successful technolo-gy for investing money. By emphasizingasset allocation, selecting appropriateinvestments, and managing fees careful-ly, the framework advocated in thisarticle adapts these institutional “bestpractices” to DC management in a waythat is highly beneficial to participants.

It’s a win-win proposition: participantsare more assured of achieving theirretirement goals, and sponsors are moreassured of achieving their employeerelations goals.

21


Endnotes

1 A Watson Wyatt study, summarized in “Investment returns:defined benefit vs. 401(k),” Watson Wyatt Insider, June 1998,found that for the years 1990-1995, the 50th percentile of thedistribution of differences between defined-contribution anddefined-benefit plan returns was a return difference of 2.0%.The size of the gap today may be the same or different;regardless, we believe the gap remains substantial becausethe underlying causes have not been addressed.

2 Assuming that one’s career begins at age 25 and ends at 65,mid-career is at age 45. (Workers tend to earn higher incomesin the later years of their careers, but that is at least partial-ly offset by the longer time for which earlier retirement plancontributions are invested.) Assuming further that the retireeconsumes his or her income from age 65 to death at 85, mid-retirement is at age 75. Thus, the average investment-holdingperiod is 30 years (45 to 75).

3 Some participants underallocate to equities out of a consciousdesire to avoid risk (in the sense of volatility). However, iftheir returns turn out to be inadequate to fund their retire-ment, these investors find out too late that risk has dimen-sions that are not captured by asset-only volatility.

4 A January 21, 2000, news item on InvestorForce.com reads,“A new report from the Investment Company Institute andEmployee Benefit Research Institute finds 401(k) partici-pants are doing a better job of allocating their assets, with49.8% of total balances in equity funds, 17.7% in companystock, 11.4% in GICs, 8.4% in balanced funds, 6.1% in bondfunds and 4.7% in money markets at the end of 1998.” Toarrive at the equity total, we assume that half of the alloca-tion to balanced funds is in equities.

5 Sharpe, William F., “The arithmetic of active management,”Financial Analysts Journal, January/February 1991. The con-cept of active management as a zero-sum game is much older(Sharpe said in the 1960s), but his 1991 article is the bestpresentation of it.

6 The almost exclusive reliance on mutual funds is probablyunwise; institutional commingled funds and institutionalseparate accounts do the same job much more inexpensive-ly, and in many cases, much better.

7 See Wagner, Wayne H., and Mark Edwards, “Best execu-tion,” Financial Analysts Journal, January/February 1993.Updated January 1998 in Plexus Group (Los Angeles, CA)Commentary #54.

8 Fixed expenses are about the same whether one has $1,000or $150,000 in an account. Quarterly statements and annualreports still need to be issued, and phone and Internet sup-port needs to be supplied. By paying for fund administrationon a per-head basis, the administrative cost (expressed as apercentage of account assets) can be made much more rea-sonable for large accounts.

9 We recommend that the sponsor manage the administrationand investment management fees to the lowest reasonabletotal cost. This cost can be charged entirely to the participantif the sponsor desires, or all or just a part of it can be paidfor by the sponsor. Charging costs, and nothing more, to theparticipant is not a fiduciary problem, but a benefits decision.

10 See Waring, M. Barton, “GICs: The large print giveth and thefine print taketh away,” Investing, Summer, 1991, Journal ofInvesting, Summer, 1992.

11 A cynic would note that, in contrast to DC assets, the poormanagement of defined-benefit plan assets potentially hasa direct effect on the company’s balance sheet, explainingthe relative care with which the assets have been managed.

12 Although an S&P 500 index fund is widely offered in DCplans, it is seldom positioned to the participants as it is sosuccessfully used in defined benefit plans—that is, as thecore large-capitalization US equity holding (or, at the veryleast, as a benchmark around which such a core holdingcan be built). Again, while better informed education pro-grams could address this, it is one more detail for the par-ticipant to stumble over. Well-engineered lifestyle types ofsolutions, however, can readily utilize index funds andother core (low residual-risk) funds.

13 Although this description is technically the most accurate,we yield to the less informative but more popular nomencla-ture of “lifestyle” or “lifecycle” funds at many places in thispaper. We note, however, that these terms completely fail todescribe the potential, from an investment strategy perspec-tive, that a well-engineered asset allocation fund can havefor the individual investor. This is the technology of choicefor maximizing expected return at a given level of risk.

14 Waring, M. Barton, and Charles Castille, “A framework foroptimal manager structure,” Investment Insights, BarclaysGlobal Investors, Volume 2, number 1 (June 1998). Anupdated version is forthcoming in the Journal of PortfolioManagement.

15 Active risk is measured by the annualized standard devia-tion of the time series created by subtracting the return ofthe relevant benchmark from the actual portfolio return.

16 Morningstar has criticized the lifecycle approach as beinginsufficiently customized. We take the opposite point ofview. The likelihood of our approach being adopted hingeson it being simple, cost effective and easily communicated.The benefits of going beyond five portfolios spaced alongthe efficient frontier are incremental and probably are notworth the added expenses, while the added complexitywould discourage the implementation of the program.

17 Transaction costs are incurred through new plan contribu-tions, fund redemptions, fund switching and rebalancing.Prepackaged fund mixes should be managed carefully tominimize costs incurred through trading of securities.Transaction costs can be minimized by using low-turnoverstrategies; index funds, in particular, have very lowturnover because they need to trade only to invest newcash flows, redeem shares, or track changes in the index.

18 Waring and Castille op.cit.

Australia Barclays Global InvestorsACN 001 804 566Level 1, 111 Harrington Street, Sydney, NSW 2000P.O. Box N43, Grosvenor Place, NSW 1220Michael Ohlsson, 02 9272 2209

Canada Barclays Global InvestorsBCE Place, 161 Bay Street, Suite 2500P.O. Box 614, Toronto, ON M5J 2S1Marg Franklin, 416 643 4041Geri James (DC), 416 643 4061

Hong Kong Barclays Global InvestorsSuite 2901, 29th Floor, One International Finance Centre1 Harbour View Street, Central, Hong KongBrian J. Haskin, 852 2295 5228

Japan Barclays Nikko Global InvestorsEbisu Prime Square Tower1-1-39 Hiroo, Shibuya-ku, Tokyo, 150-0012Noriyuki Kawana, 3 5469 4130

Barclays Trust & Banking Company (Japan)Ebisu Prime Square Tower1-1-39 Hiroo, Shibuya-ku, Tokyo, 150-0012

The Netherlands Barclays Global InvestorsStrawinskylaan 3051, 1077 ZX AmsterdamP.O. Box 59366, 1040 KJ AmsterdamHilary Smith, 020 301 2366

United Kingdom Barclays Global Investors†

Murray House, 1 Royal Mint Court, London, EC3N 4HHNigel Williams (Europe), 020 7668 8199Miles O’Connor (UK), 020 7668 8046Philip Nash (Continental Europe), 020 7668 8166

United States Barclays Global Investors 45 Fremont Street, San Francisco, CA 94105P.O. Box 7101, San Francisco, CA 94120-7101Lee Harbert, 415 597 2676Jon Geller (DC Eastern US), 508 624 4955Trennis Jones (DC Public Funds), 512 327 8681Christopher Knef (DC Western US), 415 908 7612

Barclays Global Investors 333 West Wacker Drive, Suite 2020Chicago, IL 60606Margaret Caldwell (DC Central US), 312 422 1315

www.barclaysglobal.com

Investment Insights

Published by Barclays Global Investors45 Fremont Street, San Francisco, CA 94105

Please direct story ideas, comments

and questions to:

Barton Waring, co-executive editor415 597 2064 phone415 597 2124 [email protected]

Gail N. Radzevich, co-executive editor415 597 2073 phone415 908 7111 [email protected]

Kristina Gedvila, editor415 597 2484 phone415 908 7111 [email protected]

∫

The funds referred to in this publication are managed by Barclays GlobalInvestors, N.A. Barclays Global Investors, N.A., a national bank, is an affiliate of Barclays Bank PLC. For ease of reference, “Barclays Global Investors”,“Barclays” or “BGI” may be used to refer to the firm. Barclays Global Investorsprovides investment advisory, fiduciary and custody services to various institutional investors. Funds custodied with Barclays Global Investors arenot insured by the Federal Deposit Insurance Corporation.

Benchmark index returns, unless otherwise indicated, have been calculated by Barclays Global Investors. This document is intended for professional and non-private customers only. The price of investments may go up or down,and can be affected by exchange rate movements. These funds are not depositsof, or guaranteed by, Barclays Global Investors or its affiliates and involvesinvestment risk, including loss of principal.

Past performance is not necessarily a guide to the future. Approved for issuein the UK by Barclays Global Investors Limited, regulated by IMRO.

†Barclays Global Investors LimitedRegistered in London, EnglandRegistered Number 796793Regulated by IMRO

© 2000, Barclays Global Investors, N.A.

Volume3 Number1 April 2000 Investment Insights · taking inappropriate amounts of risk, either too much or too little. These shortcomings can have a tremendous impact on a participant’s

Documents