Spending Policy in a Muted-Return Environment
The decade following the financial crisis (2010-2019) was great to investors, as markets generated both strong absolute and risk-adjusted returns. However, the next decade is likely to look much different, as investors face one of, if not the most, muted forward-looking return environments for a traditional stock and bond portfolio. As we entered the new decade, interest rates were at historically low levels (the 10-year Treasury yield was about 1.9%) and U.S. equity valuation levels were at historically high levels (the Shiller CAPE 10 was about 30, putting it at the 96th percentile of all levels since 1871).
While the crisis caused by the coronavirus did lower equity valuations (the CAPE 10 had fallen to about 25 by the end of March), interest rates fell sharply, offsetting that benefit. Since yields and valuations are the best predictors of future returns, going forward we are likely to experience low real fixed income and equity returns. With a real expected return for stocks of about 4%, and with TIPS yields below zero at the end of the first quarter, a typical 60/40 portfolio has a real expected return of only about 2%.
For foundation and endowment managers, this has implications not only in terms of asset allocation but also on spending policy, which determines the amount of funds that are distributed annually to support the institution’s mission and operations. With lower future expected returns on the horizon, endowment managers are questioning whether standard annual spending rates – which typically fall between 3 to 5% of endowment value – will be sustainable going forward and whether certain spending formulas might be better suited for the muted forward-looking environment investors currently face.
A focus on total return
Endowment management changed drastically in the late 1960s after the Ford Foundation released pioneering research suggesting endowments should focus on “total return” as opposed to just income. The foundation illustrated that over the long term, an endowment focused on total return – including both capital appreciation and income – would generate significantly more investment growth and thus provide larger payouts to beneficiaries. Before the release of this research, most endowments took a more conservative approach – only distributing interest and dividend income. As a result, endowments tended to allocate mostly to fixed income investments.
It wasn’t until 1972 when the National Conference of Commissioners on Uniform State Laws developed a model statute on endowment management called the Uniform Management of Institutional Funds Act (UMIFA), which encouraged a total-return approach, that equity allocations became more prevalent in endowment allocations. UMIFA was adopted by most states and later revised in July 2006 by the National Conference of Commissioners – the revised statute, the Uniform Prudent Management of Institutional Funds Act (UPMIFA), is now the statute currently adopted by most states.
This total-return approach brought about two new challenges for endowment managers. Endowments became more focused on capital appreciation. This led to larger allocations to risky assets, increasing the need for risk management. Second, new spending policies were needed that better aligned with this approach.
Endowment management is a difficult balancing act
Endowment managers must determine an appropriate asset allocation and spending policy that balance two conflicting goals: desired short-term spending stability and long-term preservation of purchasing power. A portfolio with lower risk is generally preferred for spending stability, while preservation of spending power typically comes from taking greater levels of risk. Those conflicting investment goals suggest two different long-term asset allocations. To help balance those competing goals, endowment managers have developed spending policies designed to smooth annual distributions to allow institutions to implement riskier portfolio allocations.
Before settling on specific spending and investment policies, an institution’s leadership needs to have a realistic conversation about budget needs, determining a reasonable amount to withdraw from the endowment each year. Generally, an organization should not consistently withdraw more than 5% from the endowment annually – with more prudent levels of spending closer to 3% given the outlook for markets. This is assuming an endowment’s assets are truly restricted or treated as such – meaning they are expected to be around in perpetuity. If this is not the case, an institution can spend as much as desired. However, it must understand that continually spending more than 5% per year will likely deplete the corpus over time.
After determining what level of spending is needed to balance the annual budget, leaders then need to have another conversation focusing on risk tolerance. Portfolio volatility will impact annual spending amounts and can have a material impact on the long-run success of the endowment. The more an endowment relies on distributions to support operations, the less risk the organization can tolerate. However, an institution that relies heavily on its endowment to support operations will also typically require a higher annual spending rate, suggesting a riskier asset allocation. This is what makes endowment management difficult and why many institutions have historically defaulted to a 60/40 blend of stocks and bonds – because it reflects the general consensus for what a reasonable balance is between risky and safe investments. This fails to recognize that because stocks are so much more volatile than the typical high-quality bonds used for fixed income allocations, as much as 85 to 90% of the risk of a 60/40 portfolio comes from equities! In other words, the typical balanced 60/40 portfolio is not very balanced at all(at least not in terms of risk, which is what matters).
After spending and risk tolerance levels have been determined, endowment managers need to determine an asset allocation. That allocation needs to have a high likelihood of being able to cover required annual spending, inflation and fees while staying within the tolerance for risk. It also needs to be one that the organization will stick with through the virtually inevitable difficult market environments that will occur –the S&P 500 has now experienced three drawdowns of at least 35% in the last 20 years.
Endowment managers need to be especially conscious of an institution’s openness to tracking variance and complexity. Higher levels of either of those increases the likelihood that an organization will give up on investment strategies after they’ve gone through poor stretches of performance – which all risky asset classes and strategies inevitably experience.
Portfolio volatility matters
Endowment managers often cite an endowment’s perpetual investment horizon as an attribute that allows them to focus more on long-term investment growth, accepting more risk. The thought is that over the long term, investment returns will be strong enough to cover needed spending while still preserving the purchasing power of the portfolio. However, portfolio volatility and the sequence of returns experienced play a pivotal role in the endowment’s ultimate success.
The higher the volatility in returns, the more difficult it is to retain purchasing power. As the endowment’s real rate of return and spending rate converge, two points deserve emphasis. Convergence between spending and returns increases the risk of the endowment losing real value over time. Second, this risk increases with portfolio volatility nonlinearly. Therefore, as expected returns are revised downward while spending rates stay the same, portfolio volatility needs to be reduced to avoid increasing the likelihood of the endowment losing real value. Many endowment managers, however, are doing the opposite: They are increasing allocations to risky assets to make up for lower expected returns while maintaining the same level of spending.
The sequence of returns achieved also has a material impact on an endowment’s long-term ability to preserve purchasing power, as most endowments must continue to spend after large market drawdowns. Sustaining spending levels after equity market declines is a necessary evil if an institution requires stable distributions. Not only will that spending be a higher percentage of the portfolio after a large drawdown, but spending could be expected to increase even more if the drawdown is accompanied by reduced donor contributions. This combination can erode principal quickly, leaving the endowment vulnerable to never getting back “above water.” Institutions with greater financial flexibility, and those that prioritize the long-term preservation of purchasing power, should view this negatively and take steps to avoid this situation. Establishing operating reserves and adopting a spending policy designed to reduce spending after large market drawdowns are ways that institutions typically address these concerns.
Types of common spending policies
Spending policy is comprised of two components: a spending rate and a spending formula. The spending rate is generally either a fixed dollar amount adjusted by inflation or a percentage of assets that gets distributed from the endowment each year. The spending formula determines how and when the spending rate is applied.
Different spending policies serve different purposes. Some prioritize preservation of principal, while others emphasize spending stability. The most common spending policies have been designed to focus on spending stability by smoothing annual distributions. These policies allow endowment managers to take more investment risk to support higher spending rates.
Most institutions use some variation of the following common spending policies:
- Dollar amount grown by inflation
This policy is straightforward and emphasizes stability in annual spending. An institution selects an initial dollar amount that meets its budgetary needs and then adjusts this amount each year going forward – most often by the annual change in inflation typically measured by the Consumer Price Index (CPI). The main drawback of this policy is its high likelihood of eroding purchasing power over the long term, creating the potential to deplete the corpus entirely.
- Percentage of portfolio
This policy calculates spending as a percentage of endowment assets at a specific point in time. It can be a simple formula where spending is based on the endowment value on a specific day of the year, usually fiscal year-end, or include a smoothing component that applies the percentage to an average of endowment values for a range of dates. For example, a common policy is a three-year moving average where the spending rate is applied to the average ending endowment value for the prior three fiscal years – this helps stabilize spending by reducing the impact of short-term market swings.
The main criticism of the moving-average policy is that endowments tend to underspend in strong markets and overspend in down markets, leaving the endowment susceptible to eroding purchasing power after bad periods. However, this policy tends to be the most commonly used – 73% of colleges and universities reported using this policy in the 2017 NACUBO-Commonfund Study of Endowments.
- Percentage of portfolio with caps and floors
This policy builds on the percentage of portfolio policy by adding caps and floors to the formula. The caps and floors help address the main criticism of the moving-average policy – that endowments tend to overspend in down markets and underspend in up markets. The caps and floors create a spending range that is applied to the endowment value and significantly helps in preserving the purchasing power of the portfolio by limiting spending after a large drawdown.
- Hybrid formula
Hybrid methods are considered the most complex and aren’t as widely used. However, they receive a lot of attention because some of the largest university endowments employ them. A hybrid policy combines two or more different spending policies and weights the importance of each depending on the overall goals of the institution, allowing for greater flexibility in prioritizing the objectives of an endowment.
Hybrid policies typically combine the dollar amount grown by inflation and percentage of portfolio formulas. This means that a portion of spending is driven by recent market performance and a portion that is very predictable. The higher the preference for spending predictability, the more weight that will be given to the dollar amount grown by inflation policy.
For example, 80% of annual spending can be based on the amount distributed last year adjusted for inflation and the remaining 20% determined as a percentage of assets. This is similar to Yale University’s policy (Yale’s investment office is considered one of the best investors in the country and pioneered the often-imitated “Endowment Model”).
The drawbacks to the typical hybrid method are its relative complexity and large weighting to the prior year’s spending, which makes the endowment susceptible to eroding purchasing power.
Improving upon the common moving-average formula
The percentage of portfolio with smoothing policy has become the workhorse for endowment managers since the shift to a total-return approach. As its popularity grew over the years, so too have its critics.
The moving-average formula’s sensitivity to equity returns makes it particularly problematic in a prolonged low-return environment. Intuitively, we know that if the goal of a spending policy is to maintain purchasing power over time, increasing spending during periods when portfolio values are falling is illogical. However, that is exactly what the moving-average policy does. As a result, this flaw in the formula renders portfolios more susceptible to sequence-of-returns risk. Generally, larger than expected returns result in underspending, promoting real appreciation, while lower than expected returns result in overspending, promoting real depreciation.
In their August 2018 paper, Evaluating Spending Policies in a Low-Return Environment, authors Peng Wang, Laura Chapman, Steven Peterson and Jon Spinney highlight the shortcomings of the popular moving-average formula and suggest institutions can improve spending stability by using a variation of the hybrid formula they term the “snake-in-the-tunnel” (SIT) approach. The authors argue that this approach is more responsive to market movements, thus lowering payout risk, while allowing institutions to maintain control over the stability of investment policy. While generally much less popular, the SIT approach is more commonly used by larger institutions that typically rely on endowment contributions for a larger proportion of their annual operating budget. The SIT approach uses the dollar amount grown by inflation policy as its base – an initial annual spending amount is determined and then adjusted each year by inflation. The SIT method sets a predetermined band of caps and floors, as a percentage of endowment value, that the total payout is not allowed to go above or below – helping to keep spending more sustainable long term.
Wang, Chapman, Peterson and Spinney evaluated long-term risk and return relationships for different spending formulas using Monte Carlo simulation. They used a simple 70% equity and 30% fixed income portfolio for their analysis with an expected return of around 7.5% and generated 15,000 simulations over 30 years. The authors first ran the analysis for the common three-year moving-average formula for a range of annual spending rates.
The authors found that there are diminishing benefits in terms of total payout over time as an institution moves beyond a certain annual spending rate. Total payout increases marginally as spending is increased from 4% to 6%, with total payout decreasing once spending hits 7% per year. Additionally, lowering the spending rate from 5% to 4% considerably lowers the risk of eroding purchasing power and the risk that real payout will be cut year-over-year. This 4% spending rate is lower than the typical 5% rate used by many institutions and more in line with current forward-looking return expectations.
Wang, Chapman, Peterson and Spinney then compared these results to the SIT approach. They started with an annual spending amount set at 5% of the beginning endowment value and grew this amount by 2.5% per year, with cap and floor bands set at 3 to 7%. The annual payout amount was reset to 5% of endowment value if spending fell below or above the specified band. The authors showed that the SIT approach yields very similar results as the moving-average policy. Both policies have similar levels of shortfall risk, ending terminal values, and total payouts over the long term. However, an important difference between the two policies is that the SIT approach offers a much smaller risk of having to cut real payout year-over-year. The probability of having to cut real payout using the SIT method compared to the moving-average approach is 3% versus 23% – demonstrating that the moving-average model is much more sensitive to returns than the SIT model.
Just as the selection of the annual spending rate is very important in the moving-average policy, the precise selection of the bands for the SIT policy is also material in maximizing its benefits. The authors found that a 3 to 7% band may be the most effective range because the risk curve flattens out meaningfully beyond this range. Total payout over the period analyzed was approximately the same for the 3 to 7% range as that of the 4 to 6% range but with considerably less year-over-year risk in having to cut spending.
A new policy designed to preserve principal
While popular, the way the moving-average and hybrid policies distribute capital is mechanically flawed. Distributions calculated under these policies are negatively correlated with the change in real portfolio value over time, meaning these policies tend to spend more after poor market returns and spend less after strong market returns. This highlights a preference toward stability in distributions as opposed to preservation of principal.
James Yaworski added to the research in his March 2019 paper, Spending Policy Customization for Institutional Preferences. Yaworski identified the negative correlation between real portfolio values and recommended spending amounts as the primary shortcomings of common spending policies. Yaworski outlined a new spending formula, the “purchasing-power rule,” which is designed to sustain real portfolio values in a reliable manner. He suggests that an ideal spending formula would recognize whether the portfolio has gained or lost purchasing power and would adjust spending in a way that helps the portfolio retain it.
The purchasing-power rule calculates spending by initially multiplying the endowment value by the desired long-term spending rate. Going forward, that value is multiplied by an adjustment factor each year, which divides the current market value of the portfolio by the purchasing power the portfolio should have. The adjustment factor allows spending rates to fluctuate in accordance with portfolio volatility. If the portfolio has appreciated by 20%, the spending level will be 20% higher. If the portfolio has lost 20% of its purchasing power, spending will decline by 20%. Thus, spending is mean reverting to real purchasing power and reflects the current state of the portfolio.
Yaworski ran Monte Carlo simulations for popular spending formulas and for his purchasing-power rule. His analysis covered 100 years and set both spending and real expected return at 6%, which is based on the historical real return of a simple 60/40 stock and bond portfolio. He showed that the purchasing power rule produces far more attractive long-term outcomes than other popular methods when focusing on preservation of principal. Median outcomes for the moving-average and hybrid formulas lose over half of real portfolio value, while the median outcome using the purchasing power rule maintains 86% of real value. While Yaworski’s new rule is far superior at preserving purchasing power, it comes at a cost of greater annual volatility in distributions, making the purchasing power formula generally unusable in isolation for many institutions.
The purchasing-power rule alone is at one extreme, allowing for the majority of purchasing power to be retained but subjecting an organization to frequent variability in distribution amounts. Other popular policies are at the other extreme, providing more stable distributions. However, they are susceptible to eroding principal. Yaworski showed that using a blended formula, combining the hybrid and purchasing-power formulas, is preferable to using any one formula in isolation. A policy that roughly equal weights these two policies allows for a similar level of purchasing power to be retained, compared with using the purchasing-power rule in isolation, while significantly reducing payout volatility. In general, the lower the tolerance for volatility in spending, the less weight the purchasing power rule should receive.
Endowment management is a tough task, as managers must balance two conflicting goals: spending stability and preservation of principal. This task has been made more difficult given today’s muted return expectations. This has endowment managers reexamining not only investment policy but spending policy as well. Every institution is unique, with its own investment goals and mission. Thus, both investment and spending policy should be tailored to fit the institution’s unique situation.
That said, every organization should be mindful of the following:
Focus most on spending rates and examine if current spending rates can be reduced to better align with current return expectations. A good approach to backing down spending is to do so gradually over several years so as not to disrupt programming too much in any single year.
Examine if there are additional ways to increase the expected return of the portfolio without materially increasing its risk profile. This is generally best achieved through factor tilts (such as size, value and quality) in equity holdings and using diversifying alternative strategies. Some alternative strategies we recommend at our firm, Buckingham Strategic Wealth, are managed futures, market neutral, reinsurance, alternative lending and variance risk premium.
For smaller organizations that do not rely heavily on endowment distributions, the common three-year moving-average spending formula with a conservative spending rate of around 3% is a good place to start. This policy is easy to understand and administer and helps provide stable annual endowment income. However, larger institutions that rely on endowment spending for a larger portion of their operating budget should consider more complex spending policies, such as the hybrid approaches described above. While these formulas are more complex to administer, they can be better tailored to meet the objectives of an institution’s endowment.
Larry Swedroe is the chief research officer for Buckingham Strategic Wealth and Buckingham Strategic Partners. Tim Jost is an advisor with Buckingham Strategic Wealth and Buckingham Strategic Partners.