where e, is the least squares residual for the tth observation, therefore measures the standard deviation of the specific(unsystematic) risk-portfolio risk that is not responsive to market fluctuations. A large standard error of the residual, say, s = 15%per month, would indicate that a substantial amount
of change in the portfolio j risk premium could not be explained by changes in the market risk premium. Further, since the R2 value from regression computer output indicates what proportion of the variation in the dependent variable is explained by variation in the right-hand or independent variables, in the CAPM context of Eq. (2.17), R2 measures the market (systematic) portion of total risk. On the other hand, 1 - R2 is the proportion of total risk that is specific (unsystematic). William F. Sharpe [1985, p. 167] notes that for an individual company a typical R2 measure from a CAPM equation is about .30 but that as one diversifies across companies' assets into a larger portfolio, the R2 measure increases, owing to the reduction of specific risk through diversification. It is important to note that, since in the bivariate regression model R2= P2xx high R2 values do not necessarily correspond with large estimates of Beta. To see this, note that from Eq. (2.15),