In many countries, bankruptcy is associated with low recovery by creditors. We develop a model of corporate credit markets in such an environment. Corporate credit is provided by either a bond market or risk-averse banks. Restructuring of insolvent firms happens out of court if in-court bankruptcy is inefficient, giving banks an advantage over bondholders. Riskier borrowers will use bank loans anywhere, but also bonds when bankruptcy is efficient. The model matches empirical debt mix patterns better than fixed-issuance-cost models. Across systems, efficient bankruptcy should be associated with more bond issuance by high-risk borrowers. This effect is small or absent for safe firms. We find that both predictions hold both cross-country and using insolvency reforms as natural experiments. Our empirical estimates suggest that a one-standard-deviation increase in the efficiency of bankruptcy is associated with an increase in the stock of corporate bonds equal to 5% of firm assets. This is equivalent to two thirds of the difference between the US and other countries. Most external financing to corporations is debt, i.e. financing with a fixed horizon and a predetermined repayment schedule. Among the dimensions in which corporate debt contracts vary, such as maturity and seniority, one stands out as particularly salient: the difference between intermediated debt (such as bank loans) and non-intermediated debt (bonds and commercial paper). Bank loans are associated with more screening, monitoring and intervention, but are held in much more concentrated fashion (Diamond 1991). Corporations use both loans and bonds in large amounts, and individual firms often use both simultaneously, showing willingness to switch. 1 However, there are strong cross-country differences in the mix of debt use: for European listed companies the amount of loans outstanding is more than twice the amount of bonds. US firms, on the contrary, have less loans outstanding than bonds. Figure 1 shows aggregate amounts of bond debt, bank loans, and other debt for publicly listed corporations in North America, Europe, and Asia in 2010, demonstrating the very large geographical differences. 2 This international variation in the corporate debt mix, not previously documented in detail, cannot easily be explained by standard models of bank-bond choice, which focus on the superior monitoring ability of banks (Diamond 1991) or the fixed costs associated with bond issuance (e.g. Bhagat and Frost 1986). These forces do not vary in an obvious way across countries, especially not on the scale that could plausibly cause these wide differences in debt mix. After all, firms need monitoring both in France and Canada, and banks presumably screen and monitor their small and large borrowers with approximately the same technology.We propose instead that a better explanation of the broad cross-country patterns is offered by variation in the efficiency with which insolvency is resolved. Countries exhibit substantial differences in how creditors in insolvent f...