I began working with Bob Jarrow, who wrote his thesis under Robert Merton, in 1995. Even before that, riding the train from our then-office into Tokyo nearly every day, I spent the time reading Merton's collected works, Continuous Time Finance. It was an honor yesterday to haul that tattered volume to Robert Merton's office at the Harvard Business School for an autograph and a chat on the topics of the day. The conversation turned to the challenges that universities are facing in this difficult environment, from Harvard's budget cuts to the elimination of the swim team at the University of Washington. As it turned out, Merton was re-editing a 1993 paper on exactly the key risk issue involved. The implications of that paper apply to all financial institutions and pension funds, in addition to university endowments. This post explains why.
The original Robert C. Merton paper "Optimal Investment Strategies for University Endowment Funds," appeared in a 1993 volume published by the National Bureau of Economic Research entitled Studies of Supply and Demand in Higher Education, edited by C. T. Clotfelter and M. Rothschild. It's been republished in the second edition of Continuous Time Finance. In the paper, after a review of the implications of diversification in a capital asset pricing model context, Merton argues that the efficient frontier and standard "best asset allocation" conclusions shouldn't apply to university endowments because their "endowment assets" (i.e. the securities portfolio) were not their only source of cash flow driven by market returns and other macro factors. The same point is true of pension funds, securities firms, banks, and insurance companies. That's why this paper by Merton has such important implications beyond the initial audience in higher education that he targeted.
In a university context, in addition to endowment assets, cash flows are also driven, in Merton's terms, by non-endowment assets. There are many implicit university assets that generate cash in this context, like alumni donations, which are heavily driven by market returns. Other more subtle drivers of cash flow mentioned by Merton are the implicit rental income on existing university real estate, inflation driving salaries, particularly those of tenured faculty, and planned expenditures of endowment funds for the operations of the university. The reason for Merton's article was the near-universal focus of endowment managers on endowment returns to the exclusion of all else. The result is that benchmark returns are misspecified and asset allocation is inconsistent with the total implicit portfolio of university assets, including both endowment funds and non-endowment funds. As Merton put it yesterday, correctly allocating assets with the full view of total university assets dominates the more limited quest for "plus alpha" performance within the endowment assets alone. For most universities, they are already "long" the stock market because donations to the university are driven by capital gains on common stocks and commercial and residential real estate. The last time I did these numbers, the annual change in unrestricted gifts to some prestigious U.S. universities was 3% plus 0.6 times the one year return on the Standard & Poor's 500 index. If a university ignores this implicit asset, they will end up being dramatically overweighted in common stocks in the endowment. Similarly, say the university has a planned outlay of hundreds of millions of dollars of new construction of campus buildings that will be funded from the endowment. If this is ignored in asset allocation of endowment assets, again, they will be overweighted in common stocks and underweighted in fixed income investments with cash flows that match the planned outlays for construction. In typical Merton fashion, he provides a beautiful continuous time proof of this in the article.
The implications of asset allocation that includes all implicit assets and liabilities of an organization are important in a much broader context. Consider the following examples:
Example 1: Oil price hedging via engineering in the auto industry
Oil prices and gasoline prices drive demand for various auto models up and down dramatically. Toyota hedged this volatility with a broad product line from gas-guzzling trucks to the hybrid Prius. General Motors did not
Example 2: Market driven revenues in the securities industry
Merger and acquisition revenues and underwriting revenues rise and fall in a highly correlated fashion with stock index levels and interest rates. The same is true for revenues from stock brokerage and investment management. Securities firms who take this into account will take a much more conservative liability structure and capital structure than firms that myopically focus only on "securities" risk in their portfolio, just like the endowment managers who focus only on endowment returns instead of the university's total assets, including non-endowment assets
Example 3: Market driven pension plan expenses
Institutions of many types manage their risk excluding any cash in or cash out from their pension fund, a major pool of assets. "Too complicated," as an explanation for ignoring the pension fund, is another way of spelling m-i-s-t-a-k-e.
Example 4: The impact of the bank's own credit risk on deposit gathering
As mentioned in other posts on this blog, our brochure Kamakura Risk Manager--In Depth details the run-off of commercial paper at Countrywide Financial and the 63% drop in "customer accounts" at Northern Rock plc in Newcastle in the face of their credit related problems. Most bank risk managers have not taken this potential cash outflow into account in their balance sheet strategies, and the result is that they are postured much more aggressively than they would be if they took the "flight risk" of their liabilities explicitly into account.
There are many other examples of similar myopia in risk management strategy. Robert Merton's "Optimal Investment Strategies for University Endowment Funds" is another gem from a Nobel laureate who makes a habit of gem production. It is a pleasure to call your attention to this work. Comments and suggestions are welcome at info@kamakuraco.com.
Donald R. van Deventer
Kamakura Corporation
Honolulu, May 5, 2009