There’s a real disconnect between the expectations of pension fund investors and the returns they can reasonably hope to achieve. How can we bridge the performance/expectations gap?

The gap between what a pension fund investor can reasonably expect to achieve and what the pension fund investor wants or needs to achieve has been widening over the past decade. What’s even more worrisome is that fund investment objectives and management expectations often reflect the “reach” rather than the more achievable “grasp”—a precarious situation, since a shortfall relative to objectives can have nasty consequences for the plan sponsor’s cash flows and financial statements.

Plan sponsors have been lucky, since the most recent period has provided sufficient investment return for most pension plans to achieve their objectives. According to the Morneau Sobeco Performance Universe of Pension Managers’ Pooled Funds, the median balanced pooled fund return for the 10 years ending on June 30, 2007 was 8.4%. If this level of performance were sustainable, there would be no problem. However, savvier plan sponsors do not expect to obtain returns like this in the future.

The widening performance/expectations gap is the result of two interrelated factors: declining interest rates and decreasing risk tolerance.

The Low Interest Rate Dilemma

The typical method for calculating a long-term return expectation is to use a “building block” approach. For example, take a typical defined benefit pension plan with a longterm return objective of inflation plus 4% (after expenses) and a typical asset mix of 60% stocks and 40% bonds. If bonds yield 2% net of inflation and we assume a 3% equity risk premium so that we expect a 5% long-term real return on equities, a 60/40 asset mix gives a long-term expected real return of 3.8% (before expenses). Shifting the asset mix more to stocks increases the expected return, but only by 0.3% for every additional 10% of the fund invested in stocks (10% times the 3% equity risk premium). If we assume expenses of 0.4%, we would need to invest about 80% in stocks to provide the expected return of inflation plus 4% after expenses.

One obvious problem with this approach is that we end up with a long-term return expectation that varies with the current interest rates. Perhaps the equity risk premium that is added on top of the bond return expectation should move inversely with interest rates to stabilize the long-term return expectation. Using a model that dynamically adjusts the expected equity risk premium based on the spread between the earnings yield on stocks and bond yields would be an interesting approach.

However, increasing the expected equity risk premium as bond yields decline is somewhat counterintuitive. Even if the equity risk premium is fixed, it already represents a larger proportional increment over a lower bond yield than over a higher bond yield (e.g., taking the 3% equity risk premium from the example above, 3% is three-quarters of a 4% yield but is only half of a 6% yield). And since actuaries strive to use conservative, long-term assumptions for going-concern valuations, the tendency has actually been to decrease the equity risk premium assumption with lower bond yields, which increases the effect of current movements in bond yields on the overall long-term return expectation and thereby increases funding volatility.

Clearly, a more stable, long-term return assumption would be desirable. However, any approach that uses long-term historical data and ignores current conditions is flawed—to be relevant, historical data must be adjusted for the current level of interest rates, level of inflation and economic conditions. And these adjustments tend to bring us back to the “building blocks” method, since the yield on a long-term bond is the market’s prediction of the bond’s long-term return.

In other words, the long term is composed of a series of shorter terms. Even the equity risk premium, which is typically derived from historical data, is debatable in that it implies a stable relationship between stock returns and bond returns. Since there have been 20-year periods in the recent past in Canada in which bonds have provided a higher return than stocks, we have to question how long term our return assumptions really need to be and how much we can afford to be wrong in the shorter term.

Decreasing Risk Tolerance

Pension fund management involves a trade-off between attempting to achieve a high long-term return to deliver the pension promise cost-effectively and trying to manage shorter-term funding and accounting volatility so that the plan sponsors are not affected too negatively by the fact that they are delivering this pension promise.

Risk tolerance tends to decrease as interest rates decrease. As the interest rates decrease, the solvency and accounting discount rates (tied to current long bond rates) also decline, while the going-concern discount rate (based on a somewhat more stable long-term return expectation) declines less rapidly. This means that interest rate movements have a bigger impact on funding and accounting results in a low interest rate environment, magnifying the importance of shorter-term results.

Other factors also have an impact on the plan sponsor’s risk tolerance. For example, as the pension plan matures and becomes larger relative to the sponsoring organization, the potential impact of the volatility becomes greater, which can also lower the plan sponsor’s risk tolerance.

Declining risk tolerance tends to widen the performance/expectations gap. A focus on managing shorter-term funding volatility comes at the expense of focusing on achieving long-term returns, which, in turn, translates into a shift in policy asset mix to a greater weighting in bonds or other asset classes with lower expected returns.

The dilemma caused by a lower-return environment is that the need to manage shorter-term volatility increases as our ability to achieve adequate long-term returns decreases. In our earlier example, when real bond yields decline to 2%, we would need an 80% allocation in stocks to expect a 4% real return after expenses. However, the expected standard deviation for such an asset mix would be greater than 10% per annum, meaning that we would expect a major shortfall relative to the long-term return objective in a significant number of periods. Even with a more conservative stock/bond asset mix, the expected standard deviation is above 8%, bringing into question our ability to use any conventional asset mix to manage shortterm funding volatility.

Fully mastering the competing objectives of achieving an adequate long-term return and surviving shorter-term funding volatility is extremely difficult. But the balance that the plan sponsor strikes between these two objectives will determine the asset mix policy, which will, in turn, determine the expected long-term return.

From a plan governance perspective, it’s important to ensure that the long-term return expectation outlined in the plan’s Statement of Investment Policies and Procedures supports the discount rate used to value the liabilities. The investment objectives should be linked to the funding objectives. In a 2% inflation world, a 6.5% discount rate implies a 4.5% real return objective, net of expenses. Given the example discussed above, an objective of 4.5% net of expenses would require an investment of 100% in equities—and, of course, the ability to withstand the funding volatility caused by such an approach.

Key Investment Terms and Definitions

Equity Risk Premium can be viewed as the extra return the plan sponsor believes that stocks will achieve over bonds in the long term.

Real Return is the part of the return that is not explained by inflation. For example, if the nominal return is 6% and inflation is 2%, then the real return is 4%.

Standard Deviation is a measure of dispersion in a probability distribution. The larger the standard deviation, the greater the spread of values within the distribution.

Bridging the Performance/Expectations Gap

How can we bridge the performance/expectations gap? We have to either lower our expectations or increase our expected return. While it may be the more practical option, lowering our expectations may have serious funding ramifications, since the discount rate used to value the pension liabilities on a going-concern basis should reflect the long-term return assumption. A decrease in this assumption of 0.5% would lead to a significant increase in liabilities (roughly 10% to 15%). Lowering the long-term return assumption means immediately recognizing 40-plus years of expected lower performance and being prepared to immediately fund for this future lower level of performance.

Therefore, the search is on for ways to increase the expected return to close the gap between expectations and requirements. Methods for increasing the investment return fall into two major categories, which, in the jargon of the investment profession, are called alpha (investment manager added value) and beta (market exposure).


The gap between what the markets are likely to give investors and what they need is often filled in by alpha, which is the value added by the investment manager above the market returns. One of the most unrealistic aspects of the pension industry is that every pension plan sponsor expects to outperform the market. This is reflected by the fact that many actuaries will actually include a value-added increment when determining discount rates. But this approach doesn’t make sense: if all investors—taken as a group—make up the market, then outperformance for one investor comes at another’s expense. For pension fund investors to continue to outperform the market, they would need a continuous supply of foolish investors willing to take the opposite side of the trades they make to outperform.

This isn’t to say that some investment managers won’t outperform their benchmarks—in any given period, many will. The lucky plan sponsors will take the superior returns as confirmation that their investment policy is sound. The unlucky plan sponsors will conclude that the underperformance can be resolved by simply selecting a better investment manager.

But investment outperformance isn’t sustainable, and there is no proven method of selecting better-performing investment managers. Consequently, when establishing projected returns for planning purposes, we should not assume any added value from the investment manager.

This approach contradicts the common industry practice of establishing a performance target—for example, a weighted average of index returns reflecting the policy asset mix plus 1% value-added. The fact that plan sponsors actually expect to achieve a return considerably in excess of the market return is a reflection of the superior marketing skills of investment professionals. They have a vested interest in persuading plan sponsors that higher-than-market returns are attainable—but only with the provider’s assistance. In fact, there is no conclusive way to prove that any investment manager value-added isn’t simply a random chance occurrence. Alpha is the “emperor’s new clothes” of the investment business: everyone should be able to see that it isn’t there, but no one wants to point it out.

The only way to reconcile the long-term fund performance objective with the investment manager performance target is to realize that the former should be based upon what we rationally expect to achieve, while the latter is based upon what we hope to achieve if the investment manager turns out to be as good as promised. Realizing that the value-added component in the investment manager performance standard represents an increment above and beyond what statistical probability would predict as the most likely outcome may help plan sponsors avoid unrealistic expectations and plan more effectively.

This argument doesn’t just apply to investment in public or more efficient markets. Even with an arguably less efficient market, such as private equity, if the benchmark being used truly reflects the overall market return, then we would expect the performance of all of the market participants to average out to this return. The potential for greater added value is counterbalanced by the potential for greater lost value.

In addition, getting access to this perceived added value comes at a significant cost. Kenneth French, a finance professor at the Tuck School of Business at Dartmouth in Hanover, N.H., has suggested that to the extent that persistent manager skill may exist in a market, management fees can be expected to increase to the point where the incremental value will accrue to the benefit of the investment manager, rather than the pension fund. Essentially, plan sponsors shouldn’t plan on using alpha strategies to increase the return objective, since the costs of implementing these programs may actually decrease the expected return.


Since alpha strategies cannot be relied upon to add value, then all long-term return should be expected to come from the market return and the extent to which the investor is exposed to it. This observation is supported by a study by Ibbotson and Kaplan, who found that “on average across funds, asset allocation policy explains a little more than 100% of the level of returns.” So if there is a gap between what we expect to achieve and what we need to achieve, this gap can best be bridged by finding higher-returning markets in which to invest and/or allocating a larger percentage of fund assets to these higher-returning markets. These markets will typically involve equity ownership of some type—for example, publicly traded stocks, private equity or real estate.

However, there is an obvious issue with pension funds attempting to eliminate the performance/expectations gap by investing more aggressively. There is a positive correlation between expected return and volatility: markets that provide higher returns also tend to provide returns that are less certain.

While moving toward investing in markets with higher-return expectations may help make the long-term return objective more achievable, it also moves the investor further away from its other objective of managing shorter-term funding volatility. Therefore, the only alternative to lowering the long-term return objective—and accepting higher funding costs—is to find ways to lessen the plan sponsor’s risk tolerance and relax the focus on short-term funding volatility, so that the plan can be managed more aggressively.

Plan fiduciaries also need to be realistic about the level of return that they expect from the market. This means stripping away the value-added claims, ignoring stellar past-performance numbers based on samples that aren’t statistically valid, and dispassionately examining how the market works. Undertaking this analysis will often result in adjusting performance expectations to more modest levels.

It’s interesting that plan sponsors are willing to pay large sums of money (often 2% of the return or more) in hopes of obtaining alpha, but want to avoid paying too much for beta—when beta is the only reliable source of returns. For a lot of non-publicly traded markets, beta cannot be easily or cheaply replicated, and exposure to better-performing markets may be worth the cost.

Active Versus Passive Management

The conclusion that the long-term return expectation should be based on exposure to markets with no adjustment for active management isn’t necessarily an argument for passive management. Paradoxically, at a time when investors are so heavily focused on achieving alpha, most opportunities to achieve exceptional value-added will quickly be arbitraged away. But if a lot of institutional investors stop pursuing alpha, more value-added opportunities may emerge, providing a stronger overall argument for active management.

Aside from the potential for higher returns, active management may provide other benefits, such as lower absolute volatility, than an index portfolio. Many active managers outperform in down-markets, but underperform in upmarkets, which may be evidence of a better-diversified portfolio. A cap-weighted index doesn’t usually provide an efficient portfolio; while the indexes serve as proxies for the underlying markets they represent, they may not be the best representations of these markets. A focus on achieving market exposure cost-effectively with a more desirable net return pattern may lead more plan sponsors to consider the merits of quantitative techniques such as fundamental indexation or even lower-cost active management in the future.

However, to the extent that active management increases the fund’s management costs, the long-term performance expectation for an actively managed fund (on an after-fees basis) should be lower than the long-term performance expectation for a passively managed fund with the same asset mix. It’s the cost of playing the game: it’s simply more expensive to try to achieve higher-than-expected performance. If these higher costs are accepted, plan fiduciaries have a duty to keep them as low as possible and to maximize the possibility of earning them back or exceeding them through active management.

Future Planning

When undertaking planning for pension funds, plan sponsors should base their policies on expected market returns without any additional expectations for investment manager brilliance. If the investment manager then outperforms, we have achieved betterthan- expected investment performance. But we shouldn’t count on it in advance.

To achieve a higher return, we need greater exposure to higher-returning markets. And to be able to tolerate this greater exposure to higher-returning markets, we need to find ways to lengthen the investment time horizon and make short-term funding volatility less important. To avoid future funding crises, plan sponsors should base their return expectations on what they can reasonably expect to achieve, not on what is likely beyond their grasp.

Robin Pond is a partner with Morneau Sobeco.

For a PDF version of this article, click here.

© Copyright 2007 Rogers Publishing Ltd. This article first appeared in the December 2007 edition of BENEFITS CANADA magazine.


Copyright © 2020 Transcontinental Media G.P. This article first appeared in Benefits Canada.

Join us on Twitter