Despite the importance of banks’ role as delegated monitors, little is known about how non-price terms of loan contracts are structured to optimize information production in a lending relationship. Using a large sample of corporate loans, this paper examines the effect of relationship lending on covenant choice. Consistent with information asymmetry theories, covenant tightness is relaxed over the duration of a relationship, especially for opaque borrowers. In contrast, the effect of lending relationship intensity on the number of covenants included in a loan follows an inverted U shape. I discuss potential explanations for this finding.

From 1973 to 2014, the common stock of U.S. banks with loan growth in the top quartile of banks over a three-year period significantly underperformed the common stock of banks with loan growth in the bottom quartile over the next three years. After the period of high growth, these banks have a lower return on assets and increase their loan loss reserves. The poorer performance of fast-growing banks is not explained by merger activity. The evidence is consistent with banks, analysts, and investors being overoptimistic about the risk of loans extended during bank-level periods of high loan growth.

Analyzing hand-collected credit agreements for a sample of middle-market firms over 2010–2015, we find that one-third of all loans are directly extended by nonbank financial intermediaries. Two-thirds of such nonbank lending can be attributed to bank regulations that constrain banks’ ability to lend to unprofitable and highly levered borrowers. Firms with negative EBITDA and debt/EBITDA greater than six are 32% and 15% more likely to borrow from nonbanks. These firms pay significantly higher interest rates, especially following the 2013 leveraged loan guidance revisions. Nonbank borrowers also receive different nonprice terms compared to firms borrowing from banks.

From 1973 to 2014, the common stock of U.S. banks with loan growth in the top quartile of banks over a three-year period significantly underperformed the common stock of banks with loan growth in the bottom quartile over the next three years. After the period of high growth, these banks have a lower return on assets and increase their loan loss reserves. The poorer performance of fast-growing banks is not explained by merger activity. The evidence is consistent with banks, analysts, and investors being overoptimistic about the risk of loans extended during bank-level periods of high loan growth. Received September 14, 2016; editorial decision May 28, 2017 by Editor Itay Goldstein.

Are some banks prone to perform poorly during crises? If yes, why? In this paper, we show that a bank's stock return performance during the 1998 crisis predicts its stock return performance and probability of failure during the recent financial crisis. This effect is economically large. Our findings are consistent with persistence in a bank's risk culture and/or aspects of its business model that make its performance sensitive to crises. Banks that relied more on short‐term funding, had more leverage, and grew more are more likely to be banks that performed poorly in both crises.