Accounting Fraud and the Pricing of Corporate Liabilities: Structural Models with Garbling

Abstract

Structural models produce credit spreads which are too low relative to those observed in the market. This problem is particularly relevant for: i) short maturities; ii) firms with low leverage and volatility. Duffie and Lando (2001) suggested that allowing for accounting noise could solve issue i). We present a model including both unbiased noise (estimate errors) and biased reporting (deliberate fraud), showing that fraud - unlike noise - is able to solve issue ii). In particular, our model and simulations show that: i) noise increases the value of equity at the expense of debt, while fraud risk affects both in the same direction; ii) the impact of noise on credit spreads is increasing in the maturity of debt, while the impact of fraud is more relevant for the short end of the credit spread term structure; iii) the impact of noise increases with leverage and asset volatility of the obligor, while this is not the case for the risk of fraud, which turns out to be more relevant in explaining credit spreads paid by safer firms. Therefore, any institutional feature able to reduce the risk of fraud would be mostly beneficial for high credit standing firms and for all of their claim holders.

Keywords: Accounting, Information, Asset pricing, Credit spreads

JEL Classification: D82, G12, G30, M41