Strategy Source Number of Managers Event driven TASS   179 Equity long/short TASS   622 Convertible arbitrage TASS   71 Equity market neutral TASS   177 Fixed income arbitrage TASS   89 Tactical trading Barclays   1,355 Sources: TASS Research; Barclays CTA index.   managers used in the study to a more relevant subset. In particular, we selected only those managers that had reported at least nine consecutive months of performance, and excluded managers for which we had missing monthly performance data points, fit is important to keep in mind that managers reported returns over differing time periods and that most managers did not report returns over the entire period indicated in Table 26.3). For each strategy, we then developed samples of equally weighted portfolios. Our objective was to determine how many managers were necessary within each strategy to match the volatility of the strategy index. For the purposes of this analysis, we decided to consider the individual components of the relative value sector. That is, rather than focus on relative value, we looked at convertible arbitrage, fixed income arbitrage, and equity market neutral. For event driven, equity long/short, convertible arbitrage, equity market neutral, and fixed income arbitrage, the CSFB/Tremont index was used, while the Barclays CTA index was used for tactical trading. (The CSFB/Tremont indexes use only a subset of the managers in the TASS Research database.) We further restricted the sample to use only those managers who had a complete history of data for the three-year period from June 1998 through May 2001. On the basis of these data, we formed 1,000 samples of 5, 10, and 20 managers for each strategy, and calculated portfolio risk characteristics. Table 26.4 summarizes our analysis. For each hedge fund strategy and each portfolio size (measured by number of managers), we show the median and the mean portfolio volatility, as well as the median and mean manager volatility. The table also shows the average correlation of excess returns between the managers for each hedge fund strategy. For comparison purposes, we also show the corresponding index volatility, calculated over the same time period. (The volatility differences between the indexes and the portfolios can be explained, in part, by the weighting schemes-the indexes are approximately capitalization weighted while the portfolios are equally weighted.) As Table 26.4 clearly illustrates, an investor does not need to hold all of the managers in each hedge fund sector to approximate index-level volatility. Part of the explanation for this lies in the low levels of manager-specific correlation for some of the sectors. For example, the average correlation between manager returns in the fixed income arbitrage sector is around 0.19. Clearly, a low level of correlation across managers can help to reduce the volatility in a hedge fund portfolio. While this result is good news for investors, it is also cautionary. Our results