How do changes in the availability of bank credit affect the way that firms manage their working capital, which is essential to their day-to-day operations? Further, are these effects, if they exist, either larger or smaller for firms that are more dependent on their access to bank financing? These questions are interesting and relevant because bank loans are often the primary source of working capital financing for many firms. We address these questions and in doing so, confront two empirical issues. First, we use novel instruments to capture the effect of exogenous shocks to bank capital and thereby their supply of credit to commercial and industrial firms. For example, we use the write-offs of residential real estate and farm loans. Such write-offs reduce a bank’s capital and therefore its ability to supply credit to businesses. As the recent financial crisis demonstrates, the write-off of these loans preceded a drop in commercial and industrial loans. Second, we address a statistical issue often ignored in prior studies of different elements of a firm’s working capital. Many studies scale a working capital account, say inventories, by total assets or some other aggregate measure to adjust for firm size and then estimate a linear regression model for their conditional mean. Such measures are doubly bounded random variables for which a linear regression model is invalid. See Cox (196) or Papke and Wooldrige (1996) for further discussion of this point. To present our evidence, we organize this study as follows. In Section 2, we identify prior research that is relevant to our research focus. Section 3 describes our samples and the variables that we create to conduct our study. Section 4 discusses the statistical issue that we address and provides a baseline analysis of how different factors influence a firm’s working capital management. Section 5 presents an analysis of the effects of changes in the supply of bank credit on corporate working capital policies. Section 6 then discusses the implications of our evidence for both what significantly influences corporate working capital policies, and for how changes in the availability of bank credit influences them. Section 7 concludes with a summary of our findings. Using data on U.S. corporations from 2000 through 2016, we derive the following major conclusions. First, one derives different implications depending on whether one recognizes or ignores the doubly bounded nature of the dependent variables. Such evidence supports the statistical arguments in Papke and Wooldridge (1996) and others. Consequently, we base our subsequent evidence on their panel data quasi-likelihood model (Papke and Wooldridge (2008)).