The importance of accounting changes in debt contracts: the cost of flexibility in covenant calculations☆
Introduction
Prior research has focused on the importance of accounting changes in debt contracts either by examining whether borrowers change accounting methods to reduce the probability of covenant violations or by examining stock price reactions to mandated accounting changes. Although these ex post studies consider the importance of accounting changes, they examine only one part of the contracting process. In this paper, we assess the ex ante importance of accounting changes in debt contracts by estimating the price borrowers willingly pay to include the effects of accounting changes in the calculations of covenant compliance. Our ex ante approach produces insights not provide by ex post studies of covenant violations. Specifically, we estimate the amount paid by borrowers to retain the flexibility voluntary accounting changes provide and to avoid the duplicate record keeping costs associated with excluding accounting changes from the calculation of covenant compliance.
Watts and Zimmerman (1990) describe the importance of accounting changes in the contracting process. Borrowers can use voluntary accounting changes ex post to avoid covenant violations, which increases the moral hazard and adverse selection costs associated with the contracts. In contrast, standard setters impose mandatory accounting changes externally and, therefore, these changes impose only limited additional moral hazard costs on the contracting parties. Mandatory accounting changes nevertheless impose additional contracting costs because they increase the costs of investigating and resolving inadvertent violations, and they increase the costs associated with delays in covenant violations that predict default.
Jensen and Meckling (1976) hypothesize that lenders ex ante anticipate the moral hazard and adverse selection costs associated with voluntary accounting changes and protect themselves against this possibility. We expect this protection to come in the form of higher interest rates for contracts that allow voluntary accounting changes to affect covenant compliance. Borrowers are willing to pay the higher interest rate if they sufficiently value the flexibility provided by voluntary accounting changes. Similarly, lenders will protect themselves from the increased contracting costs associated with including mandatory changes by charging the borrower a higher interest rate when mandatory accounting changes are included in covenant compliance calculations. Borrowers may be willing to incur the higher interest rate cost if it is less than the expected duplicate record keeping costs associated with excluding mandatory accounting changes.
To provide evidence on whether borrowers are willing to pay higher interest rates to retain accounting flexibility and to avoid duplicate record keeping costs, we examine the effect of excluding accounting changes on the rate charged on the loan. We control for borrower and loan characteristics known to determine loan pricing. We also control for the borrower's choice to exclude accounting changes since the difference in rates charged when these changes are excluded will not capture the treatment effect if a systematic difference exists in the loan rates that would otherwise be charged to the borrower. To control for this potential problem, we use a Heckman (1976) selectivity correction variable derived from the determinants of the choice to exclude accounting changes.
After controlling for other characteristics known to affect loan pricing and the self-selection problem, we find that excluding voluntary accounting changes from the calculation of covenant compliance results in an average reduction in the loan spread over LIBOR of 84 basis points. We also find that excluding mandatory accounting changes from the calculation of covenant compliance results in a reduction of the average loan spread of 71 basis points. These findings indicate that accounting changes are an important consideration in the debt contracting process. They suggest that borrowers are willing to pay substantially higher interest rates to retain accounting flexibility that may help them avoid covenant violations and to avoid duplicate record keeping costs. They also suggest that lenders protect themselves from the effects of accounting changes either by charging a higher rate when accounting changes are included in the calculation of covenant compliance, or by restricting a firm's ability to make accounting changes when the contracting costs associated with the change are relatively large. These results are consistent with Watts and Zimmerman's hypothesis that accounting changes are important in the lending process and that lenders consider the effects of accounting changes before entering into a contract.
Section 2 includes the background for our study. We describe our sample in Section 3. We develop our research design in Section 4. Section 5 provides descriptive statistics. We discuss our empirical results in Section 6, and Section 7 presents our conclusions.
Section snippets
Background
The inclusion of voluntary versus mandatory accounting changes in the calculation of debt covenant compliance may exert different effects on the probability of covenant violations. Specifically, the inclusion of voluntary accounting changes will reduce the probability of covenant violation but the effect of including mandatory changes on the probability of covenant violations is uncertain. However, the inclusion of either type of accounting change in the calculation of covenant compliance is
Sample selection and classification of contracts
We obtain our initial sample of credit agreements from the Lexis-Nexis database of corporate information using a keyword search to identify private loan agreements entered into during 1994–1996.5
Model development and research design
Based on the reasoning of Jensen and Meckling (1976), we expect that lenders will adjust the spread charged on the loan in accordance with whether covenant compliance calculations exclude or include the effects of voluntary and mandatory accounting changes. We expect lenders to reduce the spread charged on the loan if the borrower is willing to reduce the potential moral hazard costs by excluding voluntary accounting changes from the calculation of covenant compliance. Borrowers are willing to
Descriptive statistics
Table 1 shows the means and standard deviations of the variables we use to measure loan, borrower and lender characteristics. Loans that exclude both mandatory and voluntary accounting changes are more likely to require security, to be entered into by borrowers with worse credit ratings, and to have a longer maturity than those that include either voluntary changes or both voluntary and mandatory changes. Given our expectation that moral hazard costs and the costs of delayed covenant violations
Results
Table 2 reports the results of our regression model comparing the loan spreads for contracts that exclude voluntary accounting changes to the loan spreads for contracts that include voluntary accounting changes in the calculation of covenant compliance. After controlling for other factors known to affect loan pricing and for the selectivity correction, we find the spread on loans that exclude voluntary accounting changes is 84.53 basis points lower than for those where these changes are allowed
Conditional versus unconditional analysis
We examine the decision to exclude one type of accounting change while holding constant the decision about the other type of change to avoid difficulties in identifying structural equations for the potentially joint decision of excluding the two types of accounting changes. In sensitivity tests, we make unconditional comparisons of the decisions to exclude mandatory and voluntary accounting changes by including all observations in each model. Specifically, we compare the 122 agreements that
Conclusion
We examine the ex ante importance of accounting changes in debt contracts by estimating the price that borrowers willingly pay to retain the flexibility provided by voluntary accounting changes and to avoid the duplicate record keeping costs associated with excluding accounting changes from the calculation of covenant compliance.
We find that the spread on loans that exclude voluntary accounting changes is 84 basis points lower than for those that include those changes. This result suggests that
References (20)
- et al.
Collateral, loan quality, and bank risk
Journal of Monetary Economics
(1990) - et al.
Theory of the firmmanagerial behavior, agency costs and ownership structure
Journal of Financial Economics
(1976) Mandated accounting changes and debt covenants the case of oil and gas accounting
Journal of Accounting and Economics
(1984)- et al.
The effects of the bankruptcy reform act of 1978 on small business loan repricing
Journal of Financial Economics
(1986) Debt-covenant violations and managers’ accounting responses
Journal of Accounting and Economics
(1994)- Blackwell, D.W., Winters, D.B., 1997. Banking relationships and the effect of monitoring on loan pricing. The Journal...
- Blackwell, D.W., Nolend, T.R., Winters, D.B., 1998. The value of auditor assurance: evidence from loan pricing. Journal...
Contract costs, bank loans, and the cross-monitoring hypothesis
Journal of Financial Economics
(1992)Structure and pricing of large bank loans
Federal Reserve Bank of San Francisco Economic Review
(1995)- English, W.B., Nelsen, W.R., 1999. Bank risk rating of business loans. Working paper, Federal Reserve Board of...
Cited by (151)
Actions speak louder than words: Environmental law enforcement externalities and access to bank loans
2023, Journal of Banking and FinanceTax avoidance and the cost of debt for SMEs: Evidence from Spain
2023, Journal of Contemporary Accounting and EconomicsMultiple directorships and the extent of loan loss provisions: Evidence from banks in South Asia
2021, Journal of Contemporary Accounting and EconomicsImpairment or amortization of goodwill? An analysis of CFO perceptions of goodwill accounting
2021, European Management JournalTrust and use of covenants
2021, Research in International Business and FinanceAccruals quality and the cost of debt: Evidence from Vietnam
2021, International Review of Financial AnalysisCitation Excerpt :The expected sign of the coefficient on leverage is positive. Nevertheless, some prior research such as Minnis (2011), Francis et al. (2005), Beatty, Ramesh, and Weber (2002), Vander Bauwhede et al. (2015), document a negative coefficient on this variable. A possible explanation is that firms that can borrow at lower rates tend to borrow larger amounts.
- ☆
We would like to thank Robert Bowen and Angela Davis (the referees), Paul Asquith, Paul Fischer, Bob Holthausen, S.P. Kothari, Richard Leftwich, Thomas Lys (the editor), Jody Magliolo, Ed Maydew, Karl Muller, Katherine Schipper, Linda Vincent, and seminar participants at the University of Chicago, University of Florida and the Pennsylvania State University for helpful comments. Anne Beatty gratefully acknowledges financial support from PricewaterhouseCoopers.