The accounting and management literature reflects an increasing interest in relative performance evaluation (RPE) measures, often called ‘benchmarking’. In this paper we illustrate how stochastic frontier estimation (SFE) can be used to estimate benchmark performance standards which control for differences in the environments of the benchmarked operating units. Our application of SFE uses cross-sectional data reported by school districts in the state of Missouri for the year 1990–1991. The results suggest that the districts may have had as much as $394 million of excess expenditures in their operations.

This study uses experimental methods to assess how different legal regimes affect the frequency and amounts of settlements in a three person setting (one plaintiff and two defendants). The four legal regimes investigated differ along two dimensions: (1) allocation of liability, either ‘joint and several’ or proportionate liability, and (2) liability mappings, either unconditional or conditional mappings. We find that the lowest settlement frequencies occur in regimes with joint and several allocation rules. In addition, joint and several rules have the most adverse effects on the defendants' wealth.

This paper contains evidence of a significant negative stock price reaction to media disclosures of ‘subject to’ qualified audit opinions. Disclosures of qualifications in the financial news media (the Wall Street Journal and the Broad Tape) are rare relative to the frequency of audit qualifications. Other studies do not detect an impact of qualified opinions on stock prices. None of the explanations for the difference in the results between this study and prior studies is consistent with the data. We are unable to draw strong inferences because we cannot identify the selection process that produces the sample of media disclosures.

We investigate whether announcements of ‘subject to’ audit opinions and disclaimers of opinions affect stock prices. The results indicate that many firms experience negative abnormal performance prior to the release of qualified opinions, and that the magnitude of prior abnormal performance differs across types of qualifications. However, there is little evidence of a stock price effect when qualifications are disclosed publicly. It is difficult to construct powerful tests of the announcement effect of a qualified opinion for three reasons. First, the announcement date of the qualification is not easily identified. Second, measuring the unanticipated component of the announcement requires a model of market expectations. Third, controls must be employed for concurrent disclosures. The problems concerning event date identification have ramifications for other accounting event studies, particularly studies of disclosures typically contained in the annual report or 10-K.