Theory suggests that earnings response coefficients (ERCs) are positively a
ssociated with expected earnings growth and negatively associated with equi
ty risk. Dhaliwal and Reynolds (1994) (DR) hypothesize that equity beta fai
ls to capture a default risk component of equity risk and demonstrate that
ERCs are negatively associated with two measures of default risk-bond ratin
gs and debt/equity ratios-in a regression model that contains equity beta.
Bond ratings and debt/equity ratios are associated with expected earnings g
rowth. This paper examines how the association between ERCs and default ris
k is impacted by the inclusion of expected earnings growth in the model. Th
e relation between ERCs and bond ratings is not significant, while the asso
ciation between ERCs and debt/equity ratios is weakened but is still signif
icant. These findings suggest part of the reason for the negative associati
on between ERCs and default risk in DR is that their default risk proxies a
lso reflect expected earnings growth. In fact, there is no incremental asso
ciation between ERCs and default risk when equity beta, bond ratings, and e
xpected earnings growth are in the model.