This paper provides large-sample evidence on the debt covenant hypothesis using Dealscan, a database of private corporate lending agreements. The study finds strong support for the hypothesis that managers make accounting choices to reduce the likelihood of violating accounting-based debt covenants. The data show that private lenders use debt covenants as "trip wires" to monitor borrowers, and that violations occur relatively frequently, about 30% of all loans. The study also finds that violations are not necessarily associated with financial distress, and that violations are often waived for healthy firms. The study also provides evidence that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that covenants in private lending agreements are set tightly, and that violations are common. The study also finds that the consequences of covenant violations vary depending on the borrowers' economic circumstances. The study uses a large sample of loans to provide direct evidence on the relation between covenant slack and firm leverage. The study finds that the leverage variable is a relatively poor proxy for closeness to covenants. The study also finds that the distribution of covenant slack is not smooth, with a noticeable discontinuity around zero, suggesting that managers are trying to avoid covenant violations. The study finds that the discontinuity is more pronounced in the net worth sample. The study also finds that after an initial violation, managers have little incentive or ability to avoid covenant violations. The study finds that the evidence is consistent with the debt covenant hypothesis. The study also finds that the results are not fully explained by the way that loan covenants are set by lenders ex ante. The study finds that the evidence is consistent with the idea that managers have incentives to avoid covenant violations. The study also finds that the evidence is consistent with the idea that private lenders use covenants as early warning signals. The study finds that the evidence is consistent with the idea that private lenders set covenants tightly. The study finds that the evidence is consistent with the idea that violations are common. The study finds that the evidence is consistent with the idea that the consequences of violations vary depending on the borrowers' economic circumstances. The study finds that the evidence is consistent with the idea that violations are often waived for healthy firms. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent withThis paper provides large-sample evidence on the debt covenant hypothesis using Dealscan, a database of private corporate lending agreements. The study finds strong support for the hypothesis that managers make accounting choices to reduce the likelihood of violating accounting-based debt covenants. The data show that private lenders use debt covenants as "trip wires" to monitor borrowers, and that violations occur relatively frequently, about 30% of all loans. The study also finds that violations are not necessarily associated with financial distress, and that violations are often waived for healthy firms. The study also provides evidence that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that covenants in private lending agreements are set tightly, and that violations are common. The study also finds that the consequences of covenant violations vary depending on the borrowers' economic circumstances. The study uses a large sample of loans to provide direct evidence on the relation between covenant slack and firm leverage. The study finds that the leverage variable is a relatively poor proxy for closeness to covenants. The study also finds that the distribution of covenant slack is not smooth, with a noticeable discontinuity around zero, suggesting that managers are trying to avoid covenant violations. The study finds that the discontinuity is more pronounced in the net worth sample. The study also finds that after an initial violation, managers have little incentive or ability to avoid covenant violations. The study finds that the evidence is consistent with the debt covenant hypothesis. The study also finds that the results are not fully explained by the way that loan covenants are set by lenders ex ante. The study finds that the evidence is consistent with the idea that managers have incentives to avoid covenant violations. The study also finds that the evidence is consistent with the idea that private lenders use covenants as early warning signals. The study finds that the evidence is consistent with the idea that private lenders set covenants tightly. The study finds that the evidence is consistent with the idea that violations are common. The study finds that the evidence is consistent with the idea that the consequences of violations vary depending on the borrowers' economic circumstances. The study finds that the evidence is consistent with the idea that violations are often waived for healthy firms. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with the idea that the leverage variable is a relatively poor proxy for closeness to covenants. The study finds that the evidence is consistent with