[We examine whether explicitly controlling for the fixed-income exposure of mutual funds affects conclusions drawn in performance assessment. We focus on daily return data from two hybrid mutual fund samples. Comparing abnormal performance estimates from the Carhart (1997) model to extensions designed to correct for bond holdings, we find that the estimates within one of our samples change from positive to significantly negative. Additional evidence indicates that cash flows to the funds are more closely correlated with the traditional Carhart measure, clearly indicating that the absence of bond indices misleads investors who use a fund's risk-adjusted performance as the basis for investment decisions.]

We examine whether explicitly controlling for the fixed-income exposure of mutual funds affects conclusions drawn in performance assessment. We focus on daily return data from two hybrid mutual fund samples. Comparing abnormal performance estimates from the Carhart (1997) model to extensions designed to correct for bond holdings, we find that the estimates within one of our samples change from positive to significantly negative. Additional evidence indicates that cash flows to the funds are more closely correlated with the traditional Carhart measure, clearly indicating that the absence of bond indices misleads investors who use a fund's risk-adjusted performance as the basis for investment decisions.