Private Debt Covenants

Ilia Dichev and Doug Skinner used a database called Dealscan to examine large samples of private-debt covenants, which are particularly important for private-debt agreements because the covenants in these cases are set tightly. As a consequence, most covenant violations, and the greatest avoidance of these violations, tend to occur with private debt. In contrast, covenants are few and loose for public debt, such as bonds, and violations are rare. Dichev and Skinner have reported their research findings in their paper, Large-sample Evidence on the Debt Covenant Hypothesis.

The size and type of samples and the covenant detail provided in the database enabled the authors to construct tests that provide direct evidence to support the debt-covenant hypothesis. They proved managers take actions to avoid debt covenant violations. This conclusion was based on findings that showed an unusually small number of firm/quarters with financial measures just below covenant thresholds and an unusually large number that just met or beat covenant thresholds.

This was particularly true before an initial covenant violation, thereby supporting the idea that initial violations are more costly than succeeding ones. Dichev and Skinner also found that such violations are common-about 30 percent of loans had at least one violation -but for most firms, this did not reflect financial distress. Further, because private-debt covenants are set relatively tightly, the authors deduced that private lenders use these violations as a screening device, often waiving violations or resetting covenants without imposing serious consequences on the borrower.

The authors’ testing also enabled them to measure covenant-slack directly, so they examined firm leverage, which often is used as a proxy for closeness to covenants. They found that while covenant slack and leverage are correlated in the "right" way, the economic effect of this correlation is small, which implies that leverage is not a good proxy for closeness to covenants.

Read their working paper at http://eres.bus.umich.edu/docs/workpap/wp01-013.pdf or contact Doug Skinner at dskinner@bus.umich.edu.