The Russell SEW indexes have outperformed cap-weighted indexes over the entire period under review (1/1/1979-6/30/2010) in the domestic U.S. indexes and in international markets. However, the outperformance was accompanied by higher volatility.
We present Sharpe ratios over three time periods for the Russell 1000, 2000 and 3000 Indexes for the cap-weighted, SEW and CEW indexes (see Figure 10). With the exception of the Russell 2000, over the longest sample time period (1979 to June 2010), the risk-adjusted returns for the SEW were higher than for the cap-weighted and CEW indexes.
For the period 1996 to 2010, which includes the outperformance of cap-weighted indexes during the technology boom, absolute returns and Sharpe ratios are higher for the SEW indexes.
Where Is the Outperformance Coming From?
To analyze where the outperformance of the SEW indexes is coming from and to gain additional insight, we perform a Fama-French three-factor regression analysis and calculate contribution to returns by sector and size.
Fama-French Factor Regression
We regress the returns of the Russell 1000 and the Russell 2000 SEW indexes using the Fama-French three-factor model of market, size and style for each decade on the Russell 1000 and Russell 2000 cap-weighted indexes. The dependent variable is the excess return of the respective SEW index, minus the risk-free rate. The small-capitalization factor (SmB) for the Russell 1000 SEW index is measured as the excess return of the Russell Midcap Index over the Russell Top 200 Index, and for the Russell 2000 SEW index as the excess return of the Russell 2000 over the Russell 1000. The high minus low (HmL) style factor is taken as the value minus growth of the respective cap-weighted indexes. The regression results are presented in Figures 11 and 12.