Print
E-MailTom Griffin is an Assistant Professor of Finance at Villanova University, Villanova School of Business, Greg Nini is an Associate Professor of Finance at Drexel University, LeBow College of Business, and David Smith is the Virginia Bankers Association Eminent Professor of Commerce at the University of Virginia, McIntire School of Commerce. This post is based on their recent paper.
In our paper, Losing Control? The 20-Year Decline in Loan Covenant Violations, we show that the annual proportion of U.S. public firms that report a financial covenant violation dropped by nearly 70% over the last 20 years. Given that these “tripwires” serve as an important tool for lenders seeking to protect their financial claim prior to payment default, the secular trend has drawn concern from policymakers and industry professionals. For example, a member of the U.S. Senate Banking Committee warned that “the large leveraged lending market exhibits many of the characteristics of the pre-2008 subprime mortgage market. These loans are generally poorly underwritten and include few protections for lenders.”
We highlight that the role of financial covenants in incomplete contracting theory is not to grant decision rights to lenders in as many states as possible, but rather to allocate control to the party with greatest incentive to make a value-maximizing decision. Borrowers retain decision rights in normal states because their payoff structure generally incentivizes joint surplus maximization, but lenders have the right to intervene when incentive conflicts are likely to bias borrowers toward inefficient behavior. Since agency problems worsen as borrowers approach financial distress, the transfer of control rights is contingent on a signal that imperfectly captures the underlying economics of the borrower.
Adopting terminology from medical diagnostic testing, a more restrictive covenant package is beneficial because it has a lower probability of a false negative outcome, in which the borrower is distressed but fails to violate a covenant. However, we stress that a more restrictive covenant package is costly because it creates a higher probability of a false positive outcome, in which the borrower violates despite not being financially distressed.
We model covenant design as a process in which loan parties optimally select the level of restrictiveness – the covenant threshold – to minimize the expected total cost of false positive and false negative outcomes. Based on this model, we identify two fundamental factors that influence the optimal covenant threshold. First, the threshold depends on the ratio of the expected costs of false positives to false negatives, which we refer to as the “preferences” of the loan parties. The optimal threshold is less restrictive when this ratio is larger, reflecting a willingness to forego early detection of some distressed borrowers in exchange for fewer inconsequential violations. Second, the optimal threshold depends on the ability of financial covenant packages to discriminate between distressed and non-distressed borrowers, which we refer to as the covenant “technology.” Better technology allows covenants to catch more truly distressed borrowers – true positives – without concomitantly increasing the number of false positives.
Our empirical analysis produces several stylized facts. About one-half of the entire decline in covenant violations over the sample period can be explained by a drop in the false positive rate, which falls nearly 90% from the late 1990s to 2016. In other words, the largest component of the observed trend represents a substantial drop in the number of false positive violations, where lenders waive the violation without a consequential renegotiation. Conversely, the frequency at which truly distressed borrowers violate a covenant – the true positive rate – remains relatively constant over the first two-thirds of the sample period but begins to decline following the global financial crisis. By the end of the sample, the true positive rate falls by one-third relative to its earlier level and explains about 15% of the overall decline in violations. Finally, the rate of distress varies cyclically but, on average, declines during the latter part of the sample period. The decline in the distress rate explains another 30% of the total decrease in covenant violations.
This evidence helps alleviate concerns about excessively loose covenants by showing that the decline in violations is not predominantly driven by an inability of covenants to catch distressed firms early. The optimal threshold appears to have loosened over time as developments in the loan market increased the relative cost of a false positive (e.g., the rise of nonbank lenders, which resulted in larger/more diverse syndicates) and improved covenants’ signal-to-noise ratio (e.g., the typical loan in the late 1990s had three or four covenants written on a combination of balance sheet and cash flow metrics, while loans today rely on one or two covenants that benchmark performance against borrower EBITDA).
We believe our analysis may benefit regulators charged with monitoring the stability of the financial system, including the Federal Reserve and the U.S. Treasury’s Financial Stability Oversight Council. These agencies frequently examine the nonprice terms of corporate credit to gauge the level of risk-taking and potential threats to financial stability. Our paper offers a conceptual and empirical framework for interpreting changes in financial covenants and highlight that more restrictive financial covenants can create both benefits and costs.
The complete paper is available for download here.
