The following is in response to a commentary from Research Affiliates, The Folly of Peer Group Analysis, posted on March 19:
Dear Editor,
Rob Arnott never ceases to pique my interest. I have just done a comparable study to that described by Arnott in his recent article. Attached is a four-slide report. The first slide shows that for all 604 large-blend mutual funds in the peer group, the average return marginally beat the S&P 500 index over five and 10 years. The margin in the 10-year trailing rolling period is only 9 basis points and in the five-year trailing rolling period only five basis points. In the three-year trailing rolling period the margin grows to 60 basis points and in the one-year period ending 12/31/2009 the margin for the average fund grew to 162 basis points.
The second slide shows the S&P 500 index to be below peer median in the one- and three-year trailing rolling periods ending 12/31/2009. The index goes to 46th percentile in the five-year trailing rolling period and the 42nd percentile in the 10-year trailing rolling period.
My study makes no adjustments for survivorship but filters the peer group by eliminating funds based upon their holding in excess of 10% of their portfolios in non-correlated asset classes (bonds, foreign stocks, cash, etc.), so it is purer than either Morningstar or Lipper peer groups. Also, only one share class per fund is used and the results are after fees and expenses. The peer group does contain both index and actively managed funds. This study was created using MPI Stylus using 24-month regression periods.
Mr. Arnott appears correct, in my opinion, as to his conclusions about the published distortions relating to performance relative to peers insofar as funds are concerned. If only actively managed funds had been used the return spreads would probably have been higher in favor of the active managers and the peer group percentile for the index lower, as the index obviously bested all other index funds by the margin of the index funds’ expense ratios – however low they are.
However, if the same study is conducted using only Separate Account Managers, a different picture presents. Now the return spread favors the managers by 17 basis points in 10 years, 76 basis points in five years, and 103 basis points in three years. The peer group comparison when using Separate Account managers, gross of fees is starker. Using the eVestment Alliance database and peer groups, the S&P 500 index presents in the 90th percentile of peers for the 10-year trailing rolling period and fares no better than the 50th percentile for the one-year trailing rolling period. Different databases, different peer group methodology, different results. Also, gross of fees versus net of fees and expenses comparisons could account for the differences.
I am confident there are some lessons here:
- There is no substitute for practitioners doing their own homework based upon their own core beliefs in a statistically robust manner consistent with sound practices. Also, just because it is published doesn’t mean it is correct or unbiased.
- The debate between passive investing and active management is alive and well and there will always be a sub-set of active managers (as small as it might be) that will provide higher returns (and added value) than their correct relative indexes – after fees and expenses – even in more efficient asset classes.
- The challenge is to harness correctly selected indexes to measure and understand what these managers/funds did and how they did it relative to better parsed peer groups – like Surz PODs.
I want to thank Research Affiliates for their response below. Inasmuch as I function in a client universe of qualified plan fiduciaries, I believe that a higher standard is called for which is why I go the extra mile and do not accept publishers’ peer groups or their index selections without my own empirical research. I agree wholeheartedly that individuals – especially “do-it-yourselfers” – will always be misled by what they perceive to be useful information. The bottom line lesson for the serious investor might be: data is not information, information is not wisdom.
Henry Schwarzberg, JD, CIMA®, AIF®
Mobile AL
