An equilibrium model of attestation by public accountants

Attestation by public accountants fundamentally is a statistical test of the null hypothesis that a firm's financial statements fairly present the firm's financial condition. A prominent framework guiding the auditor's choice of tests is single-person decision theory. However, since the auditor is conducting a test of a strategic opponent having the motivation and the means to mislead the auditor's inferences, a game theoretic framework may be more appropriate.;My analysis begins by developing a strategic statistical testing model subsuming decision-theoretic testing and capturing the distinction between testing for hidden action and testing a report of hidden information. I characterize a general functional representation for the auditor's optimal trade off between Type I and Type II testing errors, enabling me to derive and compare, in closed-form, optimal decision-theoretic tests and equilibrium tests.;Decision theoretic testing is shown to be a substantively incomplete model relative to a game theoretic testing in that it prescribes choosing the optimal test given a subjectively assessed prior probability over the auditee's strategies, but it gives no guidance as to the correct prior to assess. Game theoretic analysis endogenously derives the only correct equilibrium "prior" as an explicit function of payoffs and properties of the nondegenerate audit technology.;When my auditor-auditee game is embedded in a market setting beset by an adverse selection problem, I show that changes in auditor liability levels for audit failures leads to a trade off between the rate of audit failures and the level of social trading gains. In contrast to a decision theoretic analysis, my strategic analysis reveals conditions under which increasing liability for audit failure could have the perverse effect of decreasing social trading gains and increasing the audit failure rate.;The primary objectives of this thesis are to characterize equilibrium behavior in testing settings in which the auditee strategically responds to the auditor's tests and to identify how equilibrium audit tests compare with those prescribed by decision theory. A secondary objective is identify implications of my equilibrium analysis regarding regulatory policies designed to increase auditors' liability for audit failures.