Target date funds, lifecycle and other similarly structured funds have been the subject of recent media attention, challenging whether these fund structures can deliver the performance necessary for typical retirement plans.  The issue has come to the forefront since the enactment, in 2006, of the Pension Protection Act, which allows Plan Sponsors to establish default investment options, but requires these options to meet higher fiduciary standards.  Managers offering a default investment option must be able to demonstrate – in a legal context – that the investment options they present are ones that can be expected to provide Plan participants with retirement income security.

A recent study by the Compass Institute concludes that target date funds, lifecycle funds, and balanced funds do not meet the needs of retirement Plan participants.  The study’s title is The Paradox of Asset Allocation for Retirement Plan Participants: A Blessing or a Curse?, and is available through Compass’ web site.  We had the opportunity to speak with Elliot Fineman, a Senior VP at Compass and author of the study, about the results.

Background

Compass is a 15-year old research organization, devoted to investment strategy, serving corporate-sponsored retirement and individual IRA plans.  In addition to research, Compass offers a definitive investment strategy participants in Plans use to manage their retirement portfolios.   Compass’ history traces to the early 1990s, when they sought to identify an optimal strategy for a fixed set of mutual funds in a retirement plan that would lead to the highest growth with the least risk over market cycles.  Fineman uses an analogy to blackjack, where a definitive card-counting strategy has been identified that will work in the long run, with limited decks of cards over deck cycles.  Compass identified and built the strategy they now offer – called Unrestricted Objective Adaptive Asset Allocation (UO-AAA) – and the current study provides ten years of performance data and comparison to benchmarks.

Fineman’s own background is in the consulting industry, where he was a strategic marketing advisor to a number of top tier international consulting firms.  He was lured out of retirement by Compass, and is motivated by two factors.  First, he believes that retirees and Plan participants are “heading off the cliff,” because they are being presented with investment options (e.g., target date funds) that do not work.  Second, he is confident that the UO-AAA strategy provides measurably superior results.  Fineman was skeptical about Compass’ methodology when he was first approached, and spent four months subjecting the UO-AAA to rigorous stress testing, before ultimately convincing himself that it offered sustainable superior performance.

The Secret Sauce – A Black Box to Produce Alpha

Fineman admits that UO-AAA is not the most elegant name for Compass’ methodology, but believes that this lack of elegance belies its sophistication.  The strategy is a risk management approach involving periodic reallocations of assets across a range of mutual funds, based on a positional model of investment performance.  Because the model does not assign any target percentages for asset classes, it is unrestricted.  The objectivity derives from the fact that there is no attempt to predict future market behavior, and therefore has nothing to do with market timing; it is manager independent.  Its black box nature resembles a quantitative strategy in that allocations are based on identifying the funds best positioned at the time of reallocation compared to the other funds in the retirement plan.  The identification of best-positioned funds is determined by a mathematical model, with no attempt to incorporate subjective market forecasts.  The model can be used with any group of mutual funds, provided that the set of funds adequately spans the range of asset classes typically found in a properly constructed retirement plan.  This characteristic gives the UO-AAA its adaptive nature.

The core of UO-AAA is a computer model that determines for a Plan participant at discrete intervals the best positioned funds of those in the Plan.  Because of intellectual property concerns, Fineman will not disclose the inner workings of the model.  But he explained that Compass has systematically identified a group of leading indicators, such as CPI or GDP indices which, in combination with technical data, accurately identify the best positioned funds over rolling two-to-three month periods.  Compass combines this with a historical price analysis of the funds used in the existing Plan, so that the sensitivity of each fund to changes in the leading indicators is modeled.  Each participant’s fund assets are reexamined every five weeks (a time period that Compass has determined to be optimal) and the assets reallocated among the funds in the Plan.

Returning to the blackjack analogy, Fineman explains that a card counting blackjack player will bet heavily when he or she knows the cards in the deck are in their favor, and they will bet minimally when they are not.  While these actions do not guarantee success in the short run, they will in the long run, i.e., over deck cycles.  The same rationale applies to UO-AAA and the investment recommendations from the model; a bet is placed on an asset allocation for a short (five week) interval, with the confidence that these bets will produce optimal results with the least amount of risk when repeated over long time intervals, i.e., market cycles.

The number of funds in a participant’s portfolio depends on the number of funds available in the Plan.  A Plan with 20 funds will typically result in investments in five or six best positioned funds for typical participant portfolios.

Just as with the definitive blackjack card counting strategy, the age and risk tolerance of the participant does not affect the identification of best positioned funds.  The model is designed to produce, with minimum risk, the highest possible rate of return – ideally a rate of return (which the study identifies as minimally 12% average annual return) sufficient for the Plan value to have grown to 20 times the participant’s salary at the time of retirement.  This is enough to support the participant’s retirement assuming the final Plan value amount is placed in an annuity (or other fixed income instrument) that pays 5% interest per year.  This will allow the participant to have their final salary every year for as long as they live through interest, without depletion of principal.

Fineman notes that “the use of age and risk tolerance are bandied about as the gospel, but that is just plain wrong when it comes to a definitive risk management strategy.”  He further adds that “the sad thing is that it is emotionally appealing, but it just doesn’t work.”

Compass’ Comparison of UO-AAA versus Target date funds

The critical component of the Compass study is the comparison of the UO-AAA strategy to a range of alternative asset allocation strategies in retirement plans.  Compass lumps these alternative strategies into a group it calls Formulaic Asset Allocation (FAA), which includes target date funds, lifecycle funds, balanced funds, and Monte Carlo simulation.  Compass has been providing its UO-AAA services to participants in large corporations, including multinational Fortune 1000 companies.  Having provided UO-AAA to these participants for over ten years, Compass is able to measure the performance of the UO-AAA against a series of benchmarks using actual data, without any backtesting.

However, Fineman notes that as the UO-AAA model is an objective strategy—it is not manager dependent and therefore always makes the same decisions based on the data entered—it can be perfectly backtested.  A fund manager cannot tell you what decision they would have made on a specific date, whereas the objective UO-AAA model can.  He says that this is extremely important for Plan Sponsors who want to know how UO-AAA would have performed for their specific Plan over time and can be expected to perform over the life of a typical participant’s Plan.

Plan Sponsors have established Plans which contain a set of mutual funds encompassing multiple asset classes.  The UO-AAA strategy is applied to the established Plan.  When UO-AAA is applied to an IRA, the fund collection comes from the same fund company.  This allows transfers of assets among funds in the IRA without transaction costs.  In a sponsored retirement plan there are no transaction costs when allocating between the different fund companies that are in the Plan.  In the case where the frequency of reallocations without penalty are limited to 60 or 90 days, UO-AAA still far outperforms the FAA funds but cannot yield the optimum results afforded by a 5-week reallocation cycle.  Most fund companies also offer a series of mutual funds in the FAA group.  Compass compared the actual performance of the UO-AAA strategy against the FAA funds over the ten year period from the end of 1996 to the end of 2006, a span which includes two up markets that bracket the dot com crash earlier this decade.

The overall result is that UO-AAA significantly outperformed FAA over the ten year period, as well as over a seven-year market cycle that started with a down market in 2000 followed by an up market.  The UO-AAA model not only outperformed FAA in down markets, but significantly outperformed FAA in up markets.  Fineman claims the two fundamental problems with FAA are its failure to allocate enough funds to equities in up markets, which he likens to “driving with the brakes on” and its requirement to keep a large percentage of the portfolio in equities, losing money for as long as a down market persists.

The study measures the riskiness or volatility of the UO-AAA strategy vs. the FAA strategy.  The UO-AAA strategy is seen to be less volatile than the FAA strategy showing a 22% lower standard deviation of average annual return.   Graphs provided show the time series of returns over the ten year period, and a cursory visual examination shows the volatility of the UO-AAA strategy is less than that of FAA.

The study provides detailed metrics on relative performance.  One key result was obtained by measuring UO-AAA versus FAA performance in 12 separate Plans.  In each Plan the UO-AAA was compared to the top performing FAA fund.  Two scenarios were considered – one where the participant had the good fortune to begin contributions at the start of the up market in 1996, and the second where the participant had the misfortune to begin at the start of the down market in 2000.  Both scenarios were measured through 2006, and the results of the two scenarios were averaged across the 12 Plans.  The UO-AAA outperformed the best FAA funds, producing a14.1% average annual return versus 6.4% for the best FAA funds. Both the 14.1% annualized return of UO-AAA and the 6.4% annualized return of the Target Date (FAA) funds are before the fees charged for their management.  The cost of the UO-AAA strategy is typically 1% of assets, in a managed account environment where participant assets are automatically reallocated every five weeks.  Target date funds typically charge a 50 basis point fee on assets.

These results are insensitive to the fund family considered–FAA funds from a variety of families were tested, and all fared poorly. 

Significantly, the UO-AAA strategy works equally well with actively and passively managed funds, but the Compass study did not compare results along these dimensions.  The only requirement is that the funds employed in the Plan adequately span the range of asset classes.

Implications for Advisors

Fineman admits that high—and ultra-high—net worth investors are not the intended market for FAA funds.  But he believes that advisors to Plan Sponsors and individuals need to be aware of the price investors will pay should they follow an FAA strategy, whether they are asked about this at cocktail parties or whether they serve on the boards of corporations or foundations that may adopt FAA funds or advise the use of FAA strategies. 

Fineman states that the Compass results are not a function of the FAA funds in the study.  He believes that FAA funds are flawed by design, stemming from the “cruise control” nature of fixed asset allocations that are insensitive to market conditions.  Fineman also contends that UO-AAA will significantly beat a single index fund, such as the S&P 500, over market cycles, and that the S&P 500 will not produce sufficiently high enough returns for a participant to retire and live off 5% annual interest of final Plan value.  For the 10-year period from the end of 1996 through the end of 2006, UO-AAA averaged, for the two scenarios discussed above, 14.1% average annual return vs. 3.0% for the S&P Single Index Fund.  “The UO-AAA strategy is necessary to achieve the needed rates of return required for participant retirement income security,” claims Fineman.

For advisors that serve as fiduciaries for retirement plans, UO-AAA is designed to meet the legal requirements of prudent investing that are part of the Pension Protection Act. 

Advisor Fiduciary Caution

Fineman points out that, as a consequence of the study, “it is now a matter of public record that FAA is a failed approach that does not prudently serve the -Plan participant’s best interest.  No one will be able to claim that they did not know.”

Combined with the provisions of the Pension Protection Act of 2006, which provide that a complaining participant can challenge in a court of law the quality of advice the Plan sponsor made available to the Plan, Fineman claims there is a “ticking fiduciary liability time bomb.  Not only Plan Sponsors, but their advisors, and investment managers will be exposed as they are co-fiduciaries.  They will be unable to answer in a court of law why they advised the Plan Sponsor to only provide advice that was known to fail and did not include an advice option that was known to succeed.”

By including UO-AAA as an advice option Fineman says that “this ticking fiduciary exposure is eliminated.”

Our Analysis

A debate is growing over the value of target date funds.  One the one hand, some contend – like Fineman - that these funds "are a great example of how marketing trumps actual thought in the financial services business."  A real danger exists if fund companies use target date funds as a way to repackage poorly-performing high-expense actively managed funds, and add an additional layer of fees.  On the other hand, target date funds are intended to replace what in many cases has been a confusing array of investment choices, many of which would be inappropriate for a given plan participant.  In these cases, target date funds represent meaningful improvement over the choices previously afforded to participants.  But will they provide a participant with retirement income security?   This debate will continue, and with improved analytical tools (see our accompanying article) to measure the performance of target date funds, the quality of information available to measure the effectiveness of these investment vehicles will grow.

The Compass study contributes to the growing body of information surrounding the effectiveness of target date plans.  For advisors considering the UO-AAA strategy in a plan for which he or she serves as an advisor or fiduciary, there are two key questions that need to be answered.  First, advisors must understand the theory behind the UO-AAA black box.  This will not be an easy analysis, but it is critical that advisors understand the principles underlying the UO-AAA investment methodology.  Advisors need to ask whether the theory is based on market timing or momentum, and assess the implications if this is the case.  By the same token, advisors need to understand how the black box behind a target date fund operates, since these funds have the same transparency issues as the UO-AAA strategy.  Second, advisors need to scrutinize the allocations of the benchmark target date funds used in the Compass study.  Target date funds are still in their infancy, and the funds that existed at the beginning of this study (in 1996) may have employed significantly different allocations and glide paths, as compared to the funds available today.  Recent data from Morningstar shows that, even for target date funds with very long (2045) time horizons, the equity allocation varies between 60% and 99%.  The equity allocation is surely the key asset allocation decision in a target date fund, having the greatest influence over future risk and return.  With this great a range in equity allocations, it is possible that the benchmarks used in Compass’ study (which had a maximum equity allocation of 80%) were unnecessarily conservative.  At the same time, while a 99% equity allocation would serve the participant well during an up market, advisors must keep an eye on risk considerations and assess the impact of a constant all-equity allocation during down markets.

A recent study by Lipper highlights both the difficulty in calibrating the performance of target-date funds, as well as the dangerous conclusions that can result from imperfect analysis.  Lipper’s study looks at a group of 2030 target date funds, but uses only three years of historical data for performance analysis.  Target date funds are in theory designed for the long term, to weather both up and down markets.  Any conclusions based on the last three years of data (which represent a consistent up market) are meaningless at best and misleading at worst.  Performance of target date funds must look at long time horizons, and the ten years of data in the Compass study provides a meaningful -  and perhaps  conclusive—set of metrics for an investment vehicle that is designed to last as long as 40 years.

Display article as PDF for printing.

Would you like to send this article to a friend?