Examining bank behavior around Federal Reserve stress tests, we find that stress test banks increase capital ratios at the starting point for annual stress testing significantly more than non-stress test banks. These trends are completely reversed (and economically significant) in the other quarters. Further, the differences between stress test and non-stress test banks seen in stress test years do not occur in 2010, when the Fed did not conduct a stress test. Results show that, as they enter the stress test, stress test banks lower dividends significantly more than non-stress test banks. Finally, stress test banks spend significantly more on lobbying than non-stress test banks. The results suggest that stress test banks may be managing financial performance and investing in political spending to improve their chances of passing stress tests.

Using detailed loan level data, we examine bank lending to corporate customers relying on principal suppliers. Customers experience larger loan spreads, higher intensity of covenants and greater likelihood of requiring collateral when they depend more on the principal supplier for inputs. The positive association between the customer’s loan spread and its dependence on the principal supplier is less pronounced when the bank has a prior loan outstanding with the principal supplier, and when the bank has higher market share in the industry. Longer relationships between the customer and its principal supplier, and between the bank and the principal supplier, mitigate lending constraints. The evidence is consistent with corporate suppliers serving as an informational bridge between the lender and the customer.