The Vanguard Index 500 is a proxy for the Standard & Poor's 500 Stock index and should, therefore, have the same risk and return characteristics. Then, why did the performance results last quarter indicate the proxy earned an excess return with less risk? A very cogent question.
It strikes at the heart of style-based performance as opposed to the more traditional approach. To help in understanding this difference, I have reproduced last quarter's results for the Vanguard Index 500.
Style beta: 0.9
Risk-adjusted omega returns:
1 year: 18.6%
5 years: 13.5%
(The one- and five-year results are risk-adjusted returns. They are called omega returns because they are the end result of some highly technical analysis, and omega comes at the end of the Greek alphabet.)
The Vanguard 500 did not earn 160 basis points more than the S&P 500.
On a risk-adjusted basis, it earned 160 basis points more than the "returns-based style" would indicate. Those who think of the S&P 500 as "the ol' gray mare" have found "she ain't what she used to be."
For the past five years, this old plow horse has been acting more like a race horse. To use the language of style analysis, the proxy has behaved as if it was 27% large value, 58% large growth and 6% foreign stock. The remainder behaves like fixed income, possibly because of utility stocks.
It is important to realize different indexes will produce different results.
Using the Morgan Stanley Capital International Europe Australasia Far East Index instead of separate indexes for Europe and the Pacific basin would result in a 7% weighting to foreign. Using the Frank Russell indexes instead of Dimensional Fund Advisors results in no foreign stocks at all. In fact, about 3% of the companies in the S&P 500 are foreign (e.g., Royal Dutch Petroleum Co.).
Returns-based style analysis does not care if the proxy is really a race horse. If it looks like a race horse and runs like a race horse, it will be categorized as a race horse until proven otherwise.
We chose DFA indexes because they are the only ones to our knowledge that have data going back to 1974. For our proposes, data should at least go back to 1974. We hope the other index providers will adjust to this need.
By the same token, the style beta of 0.9 did not indicate the Vanguard 500 took 10% less risk than the S&P 500. The style beta captures the extent to which each manager took more or less risk than the manager's style would indicate.
The Vanguard 500 no doubt took the same risk as the S&P 500. But both of them took less risk than is inherent in a style benchmark with 58% growth and 6% foreign stock.
In addition, both market indexes took less risk then the style benchmark did for 30 years.
Why not just use five years of data instead of 30? Because risk measures are very unstable over short-term intervals like five years. If we can identify the manager's style, we can calculate the inherent risk associated with that manager's style using 30 years of data.
The resulting increase in reliability is the main justification for using style analysis in this manner.
Frank Sortino is director of the Pension Research Institute, Menlo Park, Calif.