David C. Smith
One cannot read news about markets today without coming across at least one headline warning about the distress faced by many corporate borrowers and the risks to investors (and the economy) from a possible wave of defaults. For many of us, this has uncomfortable echoes of the Global Financial Crisis, and some have suggested that we are poised for a repeat of that event. Many of these articles point to “bank loans” (or “leveraged loans”) as the tip of the spear because of the fear that banks increasingly have been making loans with fewer and fewer covenants, making the loans more and more risky. They reason that, like the wave of defaults that triggered the meltdown in mortgaged-backed securities in 2008, bank loan defaults will be the trigger of another credit crisis.
Some interesting new research by McIntire Finance Professor David C. Smith and his co-authors casts some doubt on this logic. They find that from 1997 to 2016, the number of financial covenants in bank loan agreements dropped steadily. But instead of weakening the standards, banks substituted covenants that were more likely to accurately signal when a company was in distress. Overall, Smith and his co-authors find that much of the reduction in covenants reflected the better accuracy of the new covenants. They also found that firms that violated the new covenants were no more distressed than before but that banks took more severe action when a covenant breach disclosed real distress.
Smith and his co-authors begin by documenting the diminishing returns on additional financial covenants in a loan contract. They find that by increasing from one covenant to two, the rate of detecting actual distress goes from 63% to 73%, while the rate of violations that are “foot faults”—that is, don’t reveal actual distress—increases only slightly, from 31% to 33%. But when a third covenant is added to a loan contract, there is very little additional benefit. They find that the rate of detecting distress goes from 73% to 80%, but foot faults jump from 33% to 47%. In the vernacular of a medical diagnostic test, Smith and co-authors find that by moving from one to two financial covenant in a loan contract, the set of covenant “tests” increase early detection of “true positives” (financial distress) without a big increase in “false positives” (foot faults), but the balance tilts as a contract goes from two to three covenants, not false positives far outweigh true positives.
Perhaps more strikingly, they find that from 2010 to 2016, as banks used fewer covenants, the number of foot faults fell much more dramatically than they would have expected. What could account for that?
For one thing, Smith and his co-authors find that, while banks reduced the number of covenants, they also shifted to more reliable covenants. Historically at least two-thirds of covenants were so-called “balance sheet” covenants—that is, promises by the company to maintain certain accounting ratios of assets and debt. But by 2016, only 25% of covenants were balance sheet covenants. Instead, banks had begun using “cash flow” covenants, which assess a company’s ability to make debt payments from the cash flows it generates from operations. Smith and his co-authors find that balance sheet covenants produce many more foot faults than do cash flow covenants, and that by switching from balance sheet to cash flow covenant, a lender can rely on fewer but higher-quality covenants.
Recently the financial press has reported an increase by banks in “waiving” violations of cash flow covenants or allowing firms to use last year’s cash flows to evaluate compliance with covenants. The banks defend these actions because, they say, the reduction in cash flow associated with the COVID-19 shutdown is short lived and doesn’t truly represent the credit worthiness of the borrower. Relying on Smith’s work, we can see that probably simply reflects the higher confidence that banks can have in cash flow covenants over balance sheet covenants. Their work also suggests that “looser” covenants do not necessarily imply “weaker” covenants. They show that cash flow covenants can be sensitive enough to distress signals even in a more relaxed form.
While any investor should always be concerned with purported “loosening” of lending standards, Smith and his co-authors provide investors with the framework to assess whether the current lessening of covenant restrictions is good news or bad.
