What's Happening with the Pension Fund? -- Part 17

by Charles Schwartz, Professor Emeritus, University of California, Berkeley
schwartz@physics.berkeley.edu                              February 24, 2003

>> This series is available on the Internet at http://socrates.berkeley.edu/~schwrtz

Further Evidence of Skullduggery

A new document and a new set of data pried from the UC Treasurer allow us to examine and to critique more closely the University's recent dramatic shift in investment strategy. This new evidence provides strong reinforcement for the suspicion that this change was a huge swindle, perpetrated in secret meetings upon a gullible Board of Regents.

     A newly "declassified" document from UC's Office of the Treasurer presents the detailed rationale for the regents' recent dramatic shift in equity investment strategy. This 31 page document was the background paper for the "Item for Action 603X" at the October 29, 2002 meeting of the Regents' Committee on Investments. That matter was discussed and approved at a Closed Session, and my first request for this documentation was ignored by UC officials; but this new revelation has come in (partial) response to our lawyer's formal request to UC's lawyers. [In a later issue I plan to report more on this continuing legal effort to get UC's top brass to comply with the state's Public Records Act and Open Meetings law.]

     In this paper I explain how numerical data was mis-used to sell the regents on this new policy, which affects $15 Billion in UC investments for the pension fund, endowment fund, etc. I find, in summary, that this was something close to a fraud.

     First, here are the opening paragraphs of the document, Agenda Item 603X.

Implementation of Multiple Active Investment Management Programs

The Treasurer recommends that the Committee on Investments recommend to The Regents that The Regents approve a change in the investment strategy employed to manage the U.S. Equities in all of The Regents' funds. In place of a single internally managed strategy, the Treasurer recommends that an externally managed multiple manager and multiple strategy structure be employed in the future in an effort to manage risk and add value over and above the appropriate investment benchmarks established by The Regents' investment policy.
Historically, the Office of the Treasurer has actively managed the U.S. Equity exposure using internal staff resources. The Treasurer recommends that this practice cease due to the riskiness of relying on a single active equity strategy as well as long-term poor investment return relative to the historical benchmark return.

     The bulk of the document looks to be a Power Point presentation, with many captions rather than text; but it does include graphs and tables of numerical data, and this will be the focus of my critique.

UCRP Relative Performance - Raw Returns vs. Risk-Adjusted Returns

     The first page of Treasurer David Russ' data is a graph: "UCRP Total Fund, Internal Equity and Bonds - CUMULATIVE ALPHA (Annualized) 7/1992 to 6/2002 - Relative to Benchmark". The line drawn in this graph for UCRP Internal Equity shows mostly negative values, indicating underperformance relative to the benchmark; and on page 6 we read the Treasurer's summary:

This is an important point that must be emphasized: the long term relative returns from internally managed US Equity have been below the historical benchmark (emphasis in original)      Actually, this graph of Cumulative Alpha for the equity performance was presented at the (open) meeting of the regents' Committee on Investments on September 18, 2002; and in my paper Part 13 (issued November 4) I made a strong objection to this way of handing the data. I said that modern portfolio theory says one should evaluate on the basis of Risk-Adjusted Returns and not just the Raw Returns as Treasurer Russ had done there. The textbook definition of ALPHA (also called Jensen's measure) is one of several risk-adjusted procedures; and the correct formula for calculating it is given in the Appendix of this paper.

     In order to carry out the calculation of the correct (risk-adjusted) ALPHA, I needed to get the ten-year history of monthly returns for the internal equity investments of the UC Treasurer's Office. After repeated requests for this data, it was finally sent to me and I have now done the calculations. See Figure 1.


     The light-weight line in Figure 1, labeled Raw Returns, shows the same results as given by Treasurer Russ in his graph of Cumulative Alpha. (In reading these graphs, one should ignore the far left-hand portions since those data represent averages over short time intervals, starting at 7/92, and exhibit too much fluctuation for reliable interpretation.) Looking at the middle and right-hand portions, one does see that this Raw Returns line is consistently below the horizontal axis and this leads Russ to conclude that UC Internal Equities have consistently underperformed relative to the benchmark. That conclusion is clearly indicated by that set of data.

     Now look at the heavy-weight line in Figure 1, labeled Risk-Adjusted Returns. These are new data points, not seen before, and they present the correct calculation of ALPHA. These points lie sometimes below the horizontal axis and sometimes above it. Cumulative Alpha, correctly calculated, shows no consistent behavior relative to the given benchmark. Averaged over some time intervals it comes out net positive (outperforming the benchmark) and averaged over other time intervals it comes out negative (underperforming).

     Thus, I conclude that the analysis used by Treasurer Russ to justify eliminating the Internal Equity management program was false, or at best incomplete and misleading. Unfortunately, this new evidence could not be gathered and presented before the Regents acted, secretly and precipitously, to fire the entire equity investment staff.

     What I cannot avoid noticing about the data shown in Figure 1 is the steep decline (for both lines) that is seen over the last year. Since these points each represent an average result over all previous years (starting at 7/92), this rapid decline in Cumulative Alpha implies an even steeper decline in the short term performance over the past couple of years. Now we all know about the sharp decline in the stock market over these past two years; but these graphs show that UC's investments did an even steeper nosedive, since they declined relative to the market-benchmark. And this decline was so strong that it wiped out the gains shown in the preceeding years. I feel obliged to suggest that maybe there was some other cause for this very poor performance. After all, UC's internal investment staff has been in place for many years; and the data show their commendable performance up until quite recently. What has changed in UC's investment picture over the last couple of years? The answer immediately comes to mind: UC hired a new Investment Consultant (Wilshire Associates) and a new Treasurer (David Russ). Maybe, when the regents eliminated the internal equity investment staff, they fired the wrong people.

     Note. I have previously questioned the legitimacy of Russ' using the S&P 500 Index as the true "historical benchmark" for UC's equity investments prior to 2000. While I continue to maintain that objection, for the present study I put this aside and used the S&P 500 Index for the calculations.

What About the Bond Portfolio?

     The presentation of Item 603X includes some information on the Bond portion of UCRP investments and in Figure 2 you can see the two sets of results for Cumulative Alpha for this portfolio. This is a different story from that of the equities. The Raw Returns for internally managed Bonds over the past decade show a quite sizeable overperformance, relative to their benchmark. While the equity record shows a lower risk compared to the benchmark; the bond record shows a higher risk compared to its benchmark. Higher risk and higher returns -- how does one weigh these two facts and reach an intelligent evaluation of performance? By calculating the Risk-Adjusted Returns. These new data points in Figure 2 (the heavy-weight line) lie below the Raw Returns line, which can be read as paying a penalty for having taken the larger level of risk. But that Risk-Adjusted line still lies comfortably above the zero line, which means that they still outperformed their benchmark even after adjusting for the greater risk. In other words, that greater risk was very worthwhile.


     What did Treasurer Russ have to say about the Bond performance in his presentation to the regents? His summary, on page 6, says, "Overall returns slightly in excess of the policy portfolio benchmark were achieved by value-added Fixed Income performance." His adjective "slightly" refers to the 2-3% excess shown for the Raw Returns in Figure 2; but the 1-2% shortfall shown for the equities, in Figure 1, was not considered "slight", indeed this was the main point of his whole analysis. Do you detect some bias here? On page 8 he says, "Lower expectation for fixed income value-added in future." However, he gives no reason for this recommendation, implying, I guess, that the previous good performance on bonds was merely good luck. The Risk-Adjusted data for bonds, seen in Figure 2, indicates that there was more than luck involved in that success: substantial credit is due to exceptional skill of the internal investment staff.

     (As I have said before, I believe that Mr. Russ is mostly the front man in this campaign to shift UC's investments out into the private sector, while the dominant forces come from Wilshire Associates, the regents' outside advisors, and one or more members of the Board itself.)

Picking the New Managers

     What is the basis for Russ' recommendation that the Regents shift to multiple external managers for equity investments? Since he used long term comparison with the benchmark to condemn the internal management, one might ask him to show that external managers have a better track record, relative to a comparable benchmark. But no such data is offered; and I know why. My textbook says,

One explanation for the lagging popularity of risk-adjusted performance measures is the generally negative cast to the performance statistics. In nearly efficient markets it is extremely difficult for analysts to perform well enough to offset costs of research and transaction costs. Indeed, we have seen that most professionally managed equity funds generally underperform the S&P 500 index on both risk-adjusted and raw return measures.[emphasis added]
From: Bodie, Kane and Marcus, "Investments", 5th edition, page 812.
     Instead, Russ gives rather different standards for External Managers on pages 13,14 of his presentation to the regents: "Emphasize Top Quartile performing managers. Careful manager selection should raise overall portfolio results." He then gives industry-wide data showing that the spread between the top quartile manager and the median manager is about 1.5 percentage points for U.S. Large-Cap Equities, measuring average annual compound returns over the decade 1991-2000. I find this most provocative for two reasons.

     First, I can compare UCRP's internal equity management to this new statistic. The UC Treasurer's Annual Report for fiscal year 2000 gives a comparison of ten-year annualized returns on Common Stocks as follows:

UCRP .................................... 17.4%
CRA Equity Only Median ... 16.8%

And two footnotes explain:
"Capital Resource Advisors (CRA), formerly SEI, measures investment returns on approximately 5,500 portfolios, with $364 billion in assets. These returns are gross returns and are before any investment management fees, which would be approximately 0.50% of average annual market value."
"UCRP's total returns are net of (after) investment management and administrative expenses of 0.04% of average annual market value."
Thus, we see that UCRP exceeded the median of other equity managers by (17.4-16.8+0.5)% = 1.1%, which puts it close to the top quartile, using Russ' data!

     Second, Russ doesn't explain how he intends to select the managers for UC's equity investments who will perform in the top quartile. Maybe he just thinks that you pick the ones who were in the top quartile last year. But readers of my last paper (Part 16) know that that is a fool's strategy, as illustrated by some data found at the website of Russell.com, the "Global Leaders in Multi-Manager Investing."


     The new data discussed above strongly reinforces my previous guess that the Regents were the victims of a con game when they voted to fire their own investment staff and adopt the Treasurer's new strategy using external managers. One further set of crucial evidence remains hidden, despite my repeated requests under California's Public Records Act: the minutes and tape recordings of the closed meetings on October 29 and on November 13, when the full Board of Regents approved this new policy. UC's lawyers have so far refused to release these documents, making general claims about the public interest being better served by keeping those deliberations hidden from public scrutiny. We shall see.

Appendix -- Technical Notes

Formula for calculating ALPHA, the Risk-Adjusted Return of your Portfolio compared to the Market Index (or your chosen Benchmark):

a = <P - F> - b <M - F>
P is the return for your portfolio
M is the return for the Market Index
F is the return for the risk-free investment (T Bills)

<X> is the (arithmetic) average of X over some given time period (To convert monthly averages to annual averages, multiply by 12.)

b = Cov (P-F,M-F)/Cov (M-F,M-F)
and the definition of covariance is Cov(X,Y) = <(X - <X>)(Y - <Y>)>.
Cov(X,X) is also called the variance of X = (Standard Deviation of X)2.

     I have also done the calculation using a geometric, rather than arithmetic averaging procedure; the results are practically the same. For example, calculating the arithmetic average of the difference of raw returns for UCRP Internal Equity over the full ten-year period (annualized), I get the result -1.25% while Russ gives the geometric averaged result as -1.14%. This difference (0.11%) is negligible in the present context.

     If it should happen that b = 1, then the above formula simplifies to a = <P> - <M>, which is the difference of the "Bare Returns", given by Russ. However, in the case of UCRP Internally managed Equity investments over the past decade or more, the portfolio fluctuations in monthly returns (volatility) are found to be less than the fluctuations in the S&P 500, the Benchmark used by Russ. This means that the coefficient b will be less than 1, and a is therefore increased, as shown in the graph for Risk-Adjusted Returns. Finding out how much this correction amounts to required access to the full set of data, which has only now been made available.