Our model perform well within the sample. the r-squares for the panel and cross-sectional regressions are 45% and 65% respectively. the f-statistics indicate that our models are well specified. With respect to control variables our results are consistent with the findings of prior empirical studies. As in duffee 1998 and Collin-dufresne Goldstein and martin 2001 both t-bill yield and treasury spread relate negatively to credit spreads. The treasury spread volatility, as predicted by theory (Acharya and Carpenter, 2002), associates positively with credit spreads. Similar to Elton et al.2001, we find that a firm size significantly affects its credit spreads. Larger firms have smaller spreads. As expected, all measures of default and recovery risk affect credit spreads adversely. As expected, more levered and less profitable firms have larger credit spreads. Finally, firms with lower credit quality and larger profit volatility have wider credit spreads. In sum, our results suggest that bondholders view dividend payouts as positive or negative depending upon the levels of dividend. At the lower payout levels, dividend send a positive signal to bondholders about a firm future prospects, and hence results in lower cost of debt. However, at higher levels of payouts – consistent with agency-cost-of-debt hypothesis – bondholder view payouts negatively and demand greater return. Thus, for larger payout firms, higher levels of dividends result in higher cost of debt.