Given that our data are a large panel of newly issued corporate bonds, we follow the convention of recent studies (Chen et al, 2007 Guntay and Hackbarth, 2010; Nejadmalayeri and Singh, in press) to estimate our model using two different methods. First, we estimate Panel regressions with industry and year dummies where we employ year and industry dummies to control for and clustering and time dependencies. This approach ensures that autocorrelation is not influencing the results and that the results and not conditional on observations being from a particular sample year. Second, to confirm robustness of our findings, we estimate pooled– sectional regressions. Following cross Johnson (2003)