Less pain, more pleasure: navigating the utility curve
by Verne Sedlacek, President and CEO, Commonfund

Propelled by the devastation of record-setting drops in endowment values, distribution policies are once again moving to the top of trustees’ agendas at many nonprofit organizations. But, is this really a case of spending reform déjà vu —or a lasting call for fundamental change in how educational endowments and other operating nonprofits manage distributions from long-term asset pools?

Even before the collapse of asset values, the passage of the Uniform Prudent Management of Institutional Funds Act (UPMIFA) in 43 states plus the District of Columbia (as of this writing) has put the question of distribution more squarely on trustees’ shoulders, as the new legislation incorporates a more flexible spending stan­dard. (For more information on UPMIFA refer to “Freedom isn’t free” in the previous issue of Mission Matters, Spring/Summer 2009.)

The intergenerational equity debate

In general today, conversations in board rooms are dividing into two camps. The first camp says that with the substantial drop in endow­ment values, distributions should be reduced by an amount greater than the formula would require. This drastic action is needed to preserve the purchasing power of the asset pool as much as possible given the horrendous market environment and the view by many that returns over the next several years or more will be below long-term historical averages. The second camp argues that the needs of the institution have increased dramatically because of other factors and that now, more than ever, spending should be above the calculated rate because it is truly needed to support the mission of the institution. These choices strike at the core of the classic debate over intergen­erational equity: how to ensure that both today’s mission and tomorrow’s mission will be treated as fairly and equitably as possible.

Historically, the decision to spend now or later has relied on a very simple formula that takes the agonizing task of making this trade-off calculation away from governing boards. This is the spend­ing formula. We agree wholeheartedly that a spending formula is an essential guidepost for maintaining intergenerational equity, but the formula used by most institutions that have discretion over spending is critically flawed.

According to Commonfund Benchmarks Study® surveys conducted by Commonfund Institute, approximately 75 percent of educational endowments use a spending formula that smoothes spending over a period that typically is between three and five years, with most using three years, or 12 quarters. This method was first promulgated in the late 1960s as most institutions began to move to the total return concept of endowment management. This “technology” is 40 years old. I often ask myself the question, What technology in use today is four decades old? I really haven’t been able to think of many things today where the technology hasn’t changed dramatically except, perhaps, for the number 2 lead pencil. The rolling average method was developed during a time when endowments generally supported a much lower portion of the oper­ations of an educational institution or community foundation than they typically do today. What, then, is wrong with 40-year-old thinking? Certainly, some things get better as they get older; wine and antiques come to mind, and I’d like to think that I have grown better with age (although some would certainly argue with that). But, the time has probably come to retire the rolling average method of calculating the amount to distribute from endowment pools.

There are three fundamental problems with the rolling average method that heretofore has served us so well. First, when thinking about a perpetual pool of assets, three years is an extremely short time-frame. It hardly takes into account the concept that the long term should be measured in decades and not years. Second, it creates volatility in the distribution rate. This is not a problem on the upside, but it is a serious problem when the nominal dollars distrib­uted are reduced. The third flaw with the rolling average method is that it puts all the risk on the mission of the institution and none on the asset pool, since it does not account for inflation. For example, consider two extremes: If you only pay out the return of the endow­ment, then all risk is placed on the current operation. Conversely, if you increase spending every year by the rate of inflation irrespective of endowment return, then all risk is placed on the sustainability of the endowment (and, hence, the future beneficiaries of the mission).

“If it ain’t broke, don’t fix it”

Why then, given these flaws, do the majority of institutions with flexible spending authority still use the rolling average method? Quite simply because it has delivered so well for the last 20 years. It has worked over this period of time since—except for a few hiccups—institutions have ridden successive waves of bull markets that have enabled the three-year rolling average method to deliver an increase in the nominal dollars spent year in and year out. The combination of good markets and the generosity of donors has led to an increase in the nominal dollars distributed from endowments that, for the most part, has been significantly higher than the rate of  inflation. In the Commonfund white paper Why Do We Feel So Poor?, published in April 2003, we stated that the high returns of the 1990s had led to an excess dividend over the rate of inflation by as much as 40 percent annually by the early 2000s. This excess dividend was then built into operations to become a permanent liability owed by the endowment to the mission. After a couple of difficult market years in 2001 and 2002, when we saw small decreases in the stated distribution rate, the industry once again returned to increases in the nominal distribution rate that were substantially in excess of inflation. There was no need to change spending methods because what matters—today’s mission and the future mission—seemed to be satiated.

We now know this was an illusion based on the misplaced belief that market values could never fall 20 or 25 percent. What seemed, at the time, like big losses of 3 percent and 6 percent, respectively, in FY2001 and FY2002 are now what we call “the good
old days.” Since those very same days, we at Commonfund have been advocating that the nonprofit industry consider other ways to manage spending. We fundamentally believe that each institution is different, as characterized by unique needs, risk profiles and financial resources. If this is true, why do so many organizations use exactly the same method to calculate spending? Or should they look to their own unique institutional challenges and attributes? For instance, a community foundation that funds most of its good works through an endowment should have a much different approach than an educational endowment that contributes only a small portion to the operation of its institution. An endowment that funds scholarships will have a different risk profile than that of an unrestricted gift.

Over the last seven years, we have been actively promoting a spending method that includes inflation in determining distri­butions. This approach, called a hybrid or banded inflation spending policy, seeks to remedy two of the three fundamental problems of
the rolling average method: volatility of the distribution and balanc­ing risk between the long-term value of the endowment and the institution’s mission.

Utility theory and spending

To understand how to view spending in a new light, it is appropriate to recall utility theory from our study of microeconomics. Utility theory at its core is a simple concept: Consumption of goods and services can be measured by the relative satisfaction gained. Changes in utility, that is, the benefit derived/lost from increases or decreases in consumption, are typically expressed as “utils.”

At the Commonfund Forum held in March 2009 (the bottom of the equity market), the theory of the kinked utility curve was dis­cussed. The basic premise behind this concept is that nonprofits (and others) should be willing to give up outsized gains to reduce the possibility of significant losses. The function of the theory basically states that the slope of desirability or pleasure expressed numerically as utils will diminish as returns increase above a certain level. Conversely, the slope of pain or suffering will increase as the returns turn significantly negative.

An illustrative example is presented in the graphic “On the investment side” on the facing page. Here you can see the kinked utility function for a theoretical institution. Annual returns are on the x axis and utils describing pain and pleasure are on the y axis. The return target range of 5 to 8 percent is the neutral point. The slope of the line between a gain of 10 percent and a loss of 10 percent is flat. What this says is that between those two points, the trade-off between returns and increased pleasure or increasing pain is ratable. Expressed a different way, moving from a plus 10 percent return to a plus 5 percent return is painful, but no more painful than moving from plus 5 percent to 0 percent. If returns increase from 10 percent to 20 percent, however, the steepening curve means that utils or pleasure increase at a higher rate than going from a 0 percent return to a 10 percent return. After the curve hits a return of more than 20 percent the curve kinks downward, or flattens, meaning that the excess return over 20 percent does not have the same rate of impact as returns of less than 20 percent. This phenomenon reflects a dimin­ishing marginal utility for returns above a certain level.

The more important part of the curve is on the left-hand side. Losses in excess of 10 percent cause the curve to kink downward very sharply, meaning the pain increases dramatically as losses exceed 10 percent.

Achieving balance between losses and gains

The implied return algorithm here is that this entity would be willing to give up returns greater than plus 20 percent if it could keep from ever having a return of less than minus 10 percent. Of course, while this is very nice in theory, it is difficult to execute in actual portfolio construction. (This basic premise was part of the theory behind the mid-1980s use of portfolio insurance; it worked until it blew up in the crash of October 1987.) It is always easy to give up upside, but it is very difficult to protect on the downside.

The next question to be asked is why a long-term pool should be willing to forgo upside to protect on the downside since by definition the purchase of downside protection will reduce returns over a very long-term time horizon. The answer has several parts. First, there is the concept of compounding. It takes five years to recover from a 25 percent fall in equity values assuming annual appreciation in value of 6 percent after the drop. If you never lose too much, you never have to recover from a deep hole. The second rationale for forgoing upside opportunity to protect against the downside is, quite simply, that it helps trustees and senior management sleep at night. (A very legitimate reason in the short term.) An obvious third reason is the realization that there is no long term if you don’t survive the short term. But the fourth and most important reason relates to the mission of one’s institution: Losses in the asset pool instill volatility into the operations of the institution’s mission because of the interplay between short-term asset values and the amount of mission-based expenses that can be funded from the asset pool. While asset values have generally increased over the last 25 years, that has not been an issue—outside of the fact that too much was spent.

A utility curve for spending

If the fourth reason for changing the return profile based on the utility curve is the most important, why then shouldn’t we look to use the spending method as a way to express the utility curve for returns without giving up long-term opportunities to generate those excess returns? Here again, each institution would have its own trade-off between pleasure and pain. The graphic “On the spending side” on page 9 outlines a hypothetical utility curve for spending. The x axis represents the annual change in nominal spending and the y axis represents the utils associated with each of the changes in spending levels.

In this chart, a 2 percent increase in the distribution amount is the neutral point. As the increase moves from 2 percent to 3 percent, the slope of the line increases, which means there is a higher level of pleasure relative to an increase from 3 percent to 4 percent. Once the increase approaches 4 percent, however, the slope of the line actually turns negative. In utility curve language, this means that moving from a 4 percent increase to a 5 percent increase may actually reduce pleasure, or utility. This could mean that a higher increase in spend­ing creates inefficiency in the operation, i.e., spending on programs that do not actually help the mission. On the left side of the curve, a decrease in the change of nominal spending from 2 percent to 1 percent is not any more painful, on a relative basis, than a decrease from 1 percent to 0 percent. As this institution begins to cut spend­ing, however, the slope of the pain level kinks downward significantly. As we all know, making cuts is much more painful than the pleasure received from modest increases.

So, how can nonprofits apply this utility curve to the design of a more effective spending policy—one that reduces the volatility of spending, mitigates the pain of deep cuts and manages to effectively balance the needs of today with the demands of future generations?

First, recall that every institution is different, so there is no silver bullet. But, as an example, with this utility curve an institution can develop a spending policy that optimizes the utils trade-off. For instance, this curve would lead one to a spending policy that calls for using the three-year rolling average method at 5 percent as the underlying calculation, but the nominal increase would never be more than 3.5 percent in any given year and the decrease would never be more than 1 percent. Applying this method historically would have resulted in a significant reduction in the volatility of the distribution both on the upside and the downside. Spending would have been far less in the 1990s and mid-2000s (which, admittedly, would have raised the ire of various constituents who harp on the accumulated wealth of endowments). But, the kind of draconian cuts faced by many institutions over the next several years would not have become today’s reality. In addition to the bands in distributions, any amounts over or under a straight 5 percent rolling average could be set aside in a stabilization reserve and spent down or replenished over 10 to 20 years.

Conclusion

The purpose of this article is to scratch the surface of several questions surrounding spending in the current environment and to encourage boards and staffs to critically assess their spending policies. We believe that the traditional rolling three-year average spending policy is deeply flawed because it risks highly volatile spending patterns and it can lead to bias in addressing the needs of current and future beneficiaries. Using a longer period to calculate average spending levels can help but, more fundamentally, fiduciaries need to “reset” their spending approach. UPMIFA has already increased the fiduciary burden on trustees and senior management. We believe that institu­tional leadership must make overt and fundamental decisions about spending both in terms of rate and method. The alternative— making arbitrary changes in distributions because of short-term stimuli —exposes institutions and their leadership to criticism from regulatory agencies and a range of constituents. 

feedback: mmeditor@cfund.org

1 Source: Commonfund Benchmarks Study® Educational Endowment Reports