Credit growth seems to be the best predictor of future financial instability. This is the lesson that Jorda et. al (2011) draw from looking at aggregate data from 14 countries over the past 140 years. Mian & Sufi (2016) refine this result and find that growth in household credit leads to lower output growth subsequently.
How does this work? A period of increased credit supply allows households to increase their borrowing, as US households did before 2008. Once repayment becomes difficult, because of rising interest rates for example, some households restrict their consumption while others default on their loan. Using data from US counties, Mian & Sufi (2012) find evidence in favour of a decline in consumption: the higher the household leverage in a county before the housing bust, the higher is the decline in employment at local businesses in the non-tradable sector. Households who defaulted on their bank loans damaged the balance sheets of banks. Banks with impaired balance sheets decrease their lending to firms. Gabriel Chodorow-Reich (2014) has shown this link between the losses of banks and the decrease in employment of a firm during the Great Recession.
Why does this matter for macroeconomics? First, traditional financial accelerator models focused on firms and investment seem to be less promising to account for the financial cycle compared to models including a household sector. Second, the financial sector, in particular banks, plays an important transmission role between the household and firms.
Loan covenants and the macro economy
In my work I focus on a particular shock transmission channel from the banking sector to firms. How can banks reduce lending to firms when they are faced with losses? Banks can refuse to give out a new loan or to renew an existing one. Empirically however, only a small fraction of loans arrives at maturity in a given period and at the beginning of a recession banks have promised large amounts of credit to firms in the form of long-
term credit line contracts. Credit lines for firms function like credit cards for individuals: they allow firms to flexibly access an amount of funding within a certain limit, provided that the firm does not violate the contract conditions. In the following, I will explain what these contract conditions or “loan-covenants” are, and argue that they allow banks to cut funding for firms in a downturn and thus affect firms’ investment and employment immediately.
Financial loan covenants are limits to a firm’s accounting ratios and are can be found in almost all credit line contracts. Typical examples are:
. the ratio of debt to earnings cannot exceed 2.75:1
· the ratio of total liabilities to physical capital cannot be larger than 2:1
The firm is required to satisfy these covenants until the repayment of the loan. In case of a violation, the bank has the right to ask for the immediate repayment of the entire loan. But, because banks have an interest in keeping the firm from filing for bankruptcy, they rarely do so. Most frequently, banks cut access to additional credit, increase the interest rate, impose limits on investment, or ask for a partial repayment of the loan. For a subset of Compustat firms, the empirical corporate finance literature has found large real effects of covenant violations: firms decrease employment, investment, and acquisitions. This suggests that firms cannot easily substitute bank credit lines by other forms of funding when the bank decreases its supply of credit to the firm.
Building on this literature I have used a text-search algorithm to find covenant violations of almost all US public firms in their filings with the Securities Exchange Commission (SEC). My findings are: first, covenant violations are more frequent during recessions. Second, firms violating a covenant make up for a significant part of the aggregate decrease in employment growth during both the 2001 and 2008 recessions.
Compared to the existing macro literature on financial frictions with constraints on net worth or working capital, loan covenants include conditions based on earnings, income, and cash-flow that affect a different and potentially broader set of firms. Importantly, financial covenants are not conditional on the aggregate state. In the data, I find that most firms violate a covenant after one particularly bad income realization. An economy-wide downturn might therefore, cause firms to violate income covenants and lose their access to credit even though the firm’s net worth remains large.
Why are covenants included in loan contracts in the first place? Bankers try to avoid a firm’s default as good as they can. Therefore, they use a firm’s earnings, net worth, and a lot of other accounting ratios, as imperfect signals for how likely the firm will be able to repay the loan and interest. For more information, refer to classical micro theory papers about covenants, such as Dewatripont & Tirole (1994).
To formalize the mechanism described above I build a model: earnings in one period are an imperfect signal about the next period’s earnings. The management of a firm needs to obtain a credit line from a bank to finance future investment. The contract terms include how much the firm needs to repay and whether the firm or the bank are in control at any point in time before repayment. The management is protected by limited liability and therefore tends to invest too much relative to the frictionless first-best amount. When banks are in control of the firm, they tend to under-invest because they cannot recoup all the gains from their investment as the management might shirk when the bank is in control. When the productivity distribution is dispersed enough, there is an optimal income covenant threshold. For income realizations below the threshold, the bank is in control; for realizations above the threshold, the management is in control of the firm. This rule minimizes the losses from both frictions because over (under-)investment matters less when repayment period productivity is likely to
be good (bad).
An aggregate downturn occurs when the productivity distribution becomes more spread out. More firms are pushed below the income covenant threshold and under the more conservative investment strategy of banks. This decreases aggregate investment and next period’s output. There is, however, a second effect going in the opposite direction: because the signal is less informative during a recession firms decrease and banks increase their investment. Understanding the transmission of financial shocks to the real economy is important. I hope that my work on firm financing and loan covenants can contribute to this effort. More generally, I think there are a lot of interesting unanswered questions concerning household credit as well as the interaction of household and firm credit.
by Konrad Adler