The results for the two subsamples demonstrate two phenomena. First, the earnings-
smoothing effect is documented in both subsamples, which suggests that both
positive and negative earnings changes may be subject to smoothing. This finding suggests
that the conjecture in research on distressed companies that these firms employ
single-sided income smoothing cannot be generalized to my sample of financially
healthy firms (see also fn. 25). Second, the slope estimates for the subsample of firms
with negative SEPS, exceed (in absolute values) the corresponding estimates for the subsample
of firms with positive SEPS, (the differences are significant at the 10 percent
level for one-tailed tests). This implies that the earnings-smoothing and the debt-equity
effects are more pronounced for negative SEPS, firms.