ABSTRACT This study investigates whether companies engage in audit opinion shopping activities by exerting influence over an audit firm's decision to switch the engagement partner (“partner‐level opinion shopping”) in the Chinese setting, where the identities of engagement partners are publicly disclosed. Adopting the empirical framework developed by Lennox [2000], we show evidence that companies successfully engage in partner‐level opinion shopping. Further, partner‐level opinion shopping is more likely to be successful if a company is economically important to an audit firm, and it is less likely to be successful if the audit firm is formed as a partnership rather than a corporation. We also find that companies successfully engaging in partner‐level opinion shopping exhibit significantly lower earnings quality. Finally, we directly compare audit records between incoming and outgoing partners and find that, for companies that successfully improve audit opinions after partner switching, incoming partners have a significantly higher propensity to issue clean opinions than their outgoing counterparts.

Abstract This study examines the relation between performance covenants in private debt contracting and conservative accounting under adverse selection. We find that under severe adverse selection (i.e., high information asymmetry), accounting conservatism and performance covenants act as complements to signal that the borrower is unlikely to appropriate wealth from the lender. No such relation obtains in a low information asymmetry regime. We further show that in the high information asymmetry regime, borrowers with high levels of conservatism and tight performance covenants generally enjoy lower interest rate spreads than borrowers with low levels of conservatism and loose performance covenants. Consistent with our signaling theory, in the high information asymmetry regime, borrowers with high levels of conservatism and tight performance covenants are less likely to make abnormal payouts to shareholders. Our empirical results are robust to alternative measures of conservatism and covenant restrictiveness.