Research under way with Gary Miller, chief financial officer of LCG Associates, indicates calculations for downside risk for set periods in history might understate the "inherent" downside risk of investing in various equity styles.
We believe better estimates of risk might be found by building on the style analysis research of Bill Sharpe at Stanford University. In the Aug. 7 issue, I discussed the problem of calculating downside risk and concluded the bootstrap procedure developed by Bradley Effron at Stanford University coupled with the curve-fitting technique developed by J. Aitchison and J.A. Brown at Cambridge University provided the best estimates of downside risk. However, even these esoteric procedures will not be enough to describe the inherent risk we face if the historic period used is not representative of a particular investment style.
For the five years ended Dec. 31, 1995, the return for the large-cap growth style was 17.6%, the downside risk was 4.1% and the probability of falling below the average return of the S&P 500 was 26%.
Using style data from Dimensional Fund Advisors, the inherent downside risk for the large growth style index is three times greater than predicted by the five-year historical data, and the chance of falling below the S&P long-term average return rises to 53%. We changed to DFA data because it includes the down markets of 1969 and 1974.
PRI is working with Mobius Group as well as LCG Associates to bring this improved calculation later this year, but it should be noted that downside risk calculations shown in this study may seriously understate the inherent risk associated with various styles of management.
The Pension Research Institute downside risk analysis starts with the equity funds Pensions & Investments ranks as the 100 most popular among defined contribution funds.
Morningstar Ondisk, from Morningstar Inc., Chicago, is used to calculate the raw five-year returns; funds with less than five years of data are excluded.
There is concern, however, that downside risk calculations for set periods in history might understate the "inherent" downside risk of investing in various equity styles.
The graph on page represents the realized risk-return characteristics inherent for each strategy when the market was overvalued. The line from 8.75% through the market index represents the Sortino Ratio for the S&P 500 index. The Large Value, Midcap and Small stock indexes provided investors a better risk-return tradeoff than a passive strategy of owning the market. Large-cap value managers took the least risk and had the best risk-adjusted performance. Small stocks had the highest return, but also took the highest risk.
Given the market is currently overvalued by 15%, and assuming the market will either become fairly valued or undervalued, we can generate pictures of the uncertainty associated with each strategy at this point in time.
Preliminary results lead PRI to believe the use of style analysis could greatly improve the ability to calculate all risk measures, and particularly, downside risk. Also, this technique should improve the ability to predict the future performance of a given style by differentiating between realized performance and projected performance.
Mr. Sortino is director of the Pension Research Institute, San Francisco, (415) 323-6111.