What determines outstanding corporate debt mix? Evidence from fractional multinomial logit estimation

ABSTRACT

This study employs the fractional multinomial logit setting proposed by Papke and Wooldridge (1996) to examine factors driving the choice among nonbank private (144A) debt, bank loans and public debt made by 988 nonfinancial firms during 1993–2007. We document that the majority of firm-level factors have persistent effects on corporate outstanding debt mix across economic conditions. We also highlight the importance of macroeconomic variables on firms’ borrowing decisions as predicted by Diamond (1991). Finally, we document a substitution effect among debt financing sources due to credit rating downgrades, which is inconsistent with Rauh and Sufi (2010).

KEYWORDS:

• Corporate debt mix
• public debt
• bank loan
• 144A debt

JEL CLASSIFICATION:

• G32
• G30

Acknowledgements

We thank Jason Fink for the data set of firms’ founding year and the seminar participants at the University of North Carolina at Charlotte, 2012 Midwest Finance Association Annual Meeting, and 2012 Financial Management Association Meeting for comments on the earlier drafts of the article. We also thank the anonymous referees for their valuable suggestions.

Notes

¹ We thank a referee for bringing this to our attention.

² Although the coefficients on interest rate volatility changed when partitioning the sample, this is likely due to the reduced range of the interest rate volatility variable caused by the partitioning.

³ Note that the coefficient reverses, but this indicates a continuation in the pattern because it is now after the omitted year rather than before. For example, assume the firm is 50% public debt/50% private debt in the omitted year. Suppose the year before the omitted year we have 60% public/40% private debt. This would yield a positive coefficient on public debt and a negative coefficient on private debt. Suppose the year after the omitted year we have 40% public/60% private debt. This would yield a negative coefficient on the public debt and a positive coefficient on the private debt. However, both indicate a trend towards less public debt and more private debt, though the coefficient switches sign.