Pension executives now know that risk and expected return must be considered in tandem. In the risk/reward ratio -- more accurately the reward/risk ratio -- risk is the dominant factor.
Investment policies are written and shaped mainly by risk tolerances. Most of the risk constraints, imposed because of the risk tolerance of the plan sponsor, come about after some financial disaster affecting the client or the market place in general.
But risk is often thought of improperly, or calculated incorrectly, because it is based on volatility, not the true objectives of the client. This often can lead to inappropriate asset allocation, improper asset management and inaccurate performance measurement.
Initially, risk was thought of as losing money, or losing principal. This concept of risk has deep roots. Many investment policies even today stress the preservation of principal, no matter what the long-term objective or benchmark. Preservation of principal would be a most difficult and contradictory task, if the long-term investment objective were to at least equal the Standard & Poor's 500 stock index, or to equal the sponsor's long-term liabilities.
William F. Sharpe shared in the Nobel Memorial Prize in Economic Sciences for his work that resulted in the first measure of risk -- beta. In 1966 he introduced a way of measuring the risk-adjusted performance of mutual funds, initially calling it the reward-to-variability ratio. It soon became known as the Sharpe Ratio. It consists of the return of a portfolio minus the risk-free rate divided by the standard deviation of portfolio returns.
The risk-free rate was calculated as the return on the shortest Treasury bill. The translation here was that all portfolios must be compared with the risk-free rate to understand the "risk-adjusted return" of that portfolio. So, the lowest-risk security was the one with the lowest return volatility --the 30-day T-bill. The Sharpe Ratio has become one of the fundamental measurements. Most practitioners use absolute risk such that higher volatility means higher risk, no matter what the objective of the portfolio.
I met with Bill Sharpe Sept. 16, 1993, and asked him two questions. First, if a client objective is to fund a 10-year liability (e.g. a lottery), what is the least risky asset to meet that objective? He answered: A 10-year zero-coupon bond that matched the future value of that liability.
The second question was: If a client objective is the S&P 500 index return, what is the least risky asset? He answered: An S&P 500 index fund. He agreed that, given the above objectives, cash could be a more risky investment for that client than long-duration bonds or equities.
My conclusion: No generic definition of risk is valid. Risk is best defined as the uncertainty of meeting the client objective. Only the client objective can determine and measure risk. The best way to measure risk is to compare the behavior of any asset or portfolio return to the behavior of the client objective. This requires a custom index that best represents the client objective.
Four months later, Mr. Sharpe introduced an enhanced version of his Sharpe Ratio. The revised formula is the return of a portfolio, minus the return of the objective divided by the standard deviation of the differential return.
Thus, he agrees that risk is a relative measurement based on the
client's true objectives. The least risky asset is now the one that can meet the client objective with the most certainty, not the lowest volatility.
For most investors, funding liabilities (e.g. pensions, debt service) is the sole purpose for investing assets. Even individuals have liabilities as their primary objective (e.g. weddings, education, etc.). However, most investment portfolios use generic market indexes as their benchmark or objective. The reason is that liabilities are not normally calculated into an index, but are thought of as some future payment. Until liabilities are priced into a present or market value it is difficult, if not impossible, for an asset manager to perform prudently against such an ill-defined objective.
One solution to this problem is the Ryan Labs Liability Index, which can be customized to a client's actuarial liability schedule. It provides transparent pricing of the liabilities, thereby calculating present value term structures, growth rates and full spectrum of portfolio statistics. More than any generic market index, this liability index best represents the client objectives.
Once the objective is accurately measured, one can measure the risk of the asset side. Liabilities can be subdivided into short, intermediate, long and very long. Long assets should be measured against long liabilities. Figure 1 illustrates the risk/reward behavior of asset classes vs. the Ryan Labs Liability Index. Assets above the line outperformed liabilities; those below underperformed. Notice that bond indexes behave like intermediate liabilities; equities behave like long liabilities; small-cap and international stocks behave like very long liabilities. Asset allocation should get its shape from the term structure of liabilities. The percent in intermediate liabilities is the allocation to bonds. The percent in long liabilities is the allocation to large-cap and midcap stocks; and the percent in very long liabilities is the allocation to small-cap and international stocks.
Measuring the growth rate behavior of each asset portfolio relative to the growth rate behavior of the liability (e.g. stocks vs. long liabilities) would be the only way to determine the true risk/reward profile of that portfolio. Over a proper time horizon, if any portfolio outperformed a market index, but underperformed liabilities, didn't the client lose? Wasn't this portfolio risky vs. the true objective? The no-risk portfolio would be a liability index fund that matched the liabilities with certainty. Such a fund would consist of bonds that match the duration and amounts of the liability payment schedule. Instead of matching future values (dedication) or an average duration of the liabilities (immunization), it is important the liability index fund match the term structure for each year of liabilities in present value dollars. As a result, the core portfolio for a liability objective is a liability index fund.
In summary, risk must be measured correctly, if a fund is to adopt the asset allocation appropriate to it. And for risk to be measured correctly, it must be measured as the relative movement of asset values vs. the movement of the value of liabilities.
Ron Ryan is president of Ryan Labs Inc., New York