This study explains the challenges associated with the Heckman (1979) procedure to control for selection bias, assesses the quality of its application in accounting research, and offers guidance for better implementation of selection models. A survey of 75 recent accounting articles in leading journals reveals that many researchers implement the technique in a mechanical way with relatively little appreciation of important econometric issues and problems surrounding its use. Using empirical examples motivated by prior research, we illustrate that selection models are fragile and can yield quite literally any possible outcome in response to fairly minor changes in model specification. We conclude with guidance on how researchers can better implement selection models that will provide more convincing evidence on potential selection bias, including the need to justify model specifications and careful sensitivity analyses with respect to robustness and multicollinearity. Data Availability: Data used are available from public sources identified in the study.

We investigate whether short sellers and analysts differ in their use of information that is predictive of future returns. We find that short interest is significantly associated in the expected direction with all 11 variables examined. In contrast, analysts tend to positively recommend stocks with high growth, high accruals, and low book-to-market ratios, despite these variables having a negative association with future returns. We then investigate the profitability of using short interest in trading. We find abnormal returns (1.11 percent per month) from a zero-investment strategy that (1) shorts firms with highly favorable analyst recommendations (buy signal) but high short interest (sell signal), and (2) buys firms with highly unfavorable analyst recommendations (sell signal) but low short interest (buy signal). Short interest, therefore, appears to capture predictive information that can be used by investors in trading against analysts’ recommendations to increase returns.

Prior research shows that an audit supervisor’s active intervention in a subordinate’s judgment distorts that judgment. However, subordinates’ judgments are only one input into audit team judgments. How do supervisors finalize audit team judgments after actively intervening in their subordinates’ judgments? In an experiment using audit teams, supervisors with weaker or stronger goals to reach a client-preferred conclusion either were or were not asked to first actively coach a subordinate’s judgment (i.e., active intervention) before reviewing it and finalizing the audit team’s judgment. Supervisors’ intervention influenced subordinates’ inputs, which, in turn, supervisors incorporated into their final judgments. More interestingly, intervention biased supervisors’ final judgments, controlling for supervisor directional goal strength and for concurrent effects on subordinates’ inputs. However, supervisors distorted their judgments less as they perceived a larger technical knowledge advantage over subordinates. In a second experiment, auditors appear aware of the bias-reducing knowledge advantage effects but unaware of the bias-increasing active intervention effects. We discuss implications for audit team judgments and audit quality control.

In this paper we investigate the reputational penalties to managers of firms announcing earnings restatements. More specifically, we examine management turnover and the subsequent employment of displaced managers at firms announcing earnings restatements during 1997 or 1998. In contrast to prior research (Beneish 1999; Agrawal et al. 1999), which does not find increased turnover following GAAP violations or revelation of corporate fraud, we find that 60 percent of restating firms experience a turnover of at least one top manager within 24 months of the restatement compared to 35 percent among age-, size-, and industry-matched firms. Moreover, the subsequent employment prospects of the displaced managers of restatement firms are poorer than those of the displaced managers of control firms. Our results hold after controlling for firm performance, bankruptcy, and other determinants of management turnover, and suggest that both corporate boards and the external labor market impose significant penalties on managers for violating GAAP. Also, in light of resource constraints at the SEC, our findings are encouraging as they suggest that private penalties for GAAP violations are severe and may serve as partial substitutes for public enforcement of GAAP violations.

We analyze the effect of social norms and enforcement on the dynamics of taxpayer compliance. Specifically, we develop two models to evaluate the movement between classes of compliant and noncompliant taxpayers. Our analysis suggests that the effect on compliance of changing enforcement levels depends on whether the taxpayer population is initially compliant or noncompliant. Compliant populations are insensitive to changes in enforcement policies until enforcement becomes sufficiently lax, when we observe a sudden shift to high levels of noncompliance in equilibrium. In contrast, relatively noncompliant populations respond to increased enforcement by gradually increasing compliance. Then, when enforcement becomes sufficiently harsh, we find a sudden shift in equilibrium to very high levels of compliance. After the taxpayer population shifts from compliance to noncompliance, or vice versa, our models predict that returning to the previous enforcement policy will not cause the population to return to its previous state. On the whole, our models' results help explain why taxpayer compliance varies across time and across geographic regions, even under similar enforcement regimes.

Accounting researchers (and potentially others) generally select rather simple, lower-order, time-series models to develop proxies for earnings persistence. However, measures of persistence produced by such models are not related to characteristics of the firm's economic environment that are expected to influence earnings persistence. Using a sample of 162 calendar year-end New York Stock Exchange firms, we document the cross-sectional relations between a set of relatively constant, firm-specific, economic characteristics that are theoretical determinants of persistence and measures of earnings persistence derived from both lower-order and higher-order Autoregressive, Integrated, Moving-Average (ARIMA) models. When lower-order ARIMA models are used to generate measures of earnings persistence, the cross-sectional regression models measuring the association between persistence and economic determinants of persistence yield very low adjusted R2s. In sharp contrast, when differenced, higher-order ARIMA models are used to measure earnings persistence, adjusted R2s are in the 10–12 percent range. Moreover, independent variables such as capital intensity, barriers-to-entry, and product-type are all significant in the directions suggested by economic theory. Our results are consistent with Lipe and Kormendi (1994) who argue that higher-order ARIMA models do a better job of capturing the valuerelevance of current period earnings than lower-order models.

The article presents a report on teaching methods and programmed instructions in the U.S. During the past few years, in the process of continuing search to discover means for increasing instructional efficiency and for developing new materials to cope with the rapidly expanding demands for additional education and training by a burgeoning number of learners, the experimentation and development of programed learning has received the attention of educators and training directors across the country. Programed instruction has found wide-spread usage in industry and in the armed forces in various aspects of training for specific assignments or tasks, in which predetermined performance levels or behavioral patterns had been established. In numerous situations in both education and industry, programed instruction has proved to be effective, and under some circumstances more so than conventional methods. However, at this time research findings, especially at the college level, are totally inadequate to serve as a basis for making decisions on the adoption of programed learning for use throughout any specific course or program.

The path leading to a degree with a major in accounting is a rugged one at best. Obstacles are to be encountered at every turn. Many an eager student has dashed confidently down this path only abruptly to encounter a sizeable obstacle known as partnership dissolution resulting from the admission of a new partner. Managerial Accounting as a course of study represents a recent addition to the accounting curriculum of many schools of higher education. The addition of this course reflects the current demand for an approach to accounting that emphasizes the utilization of accounting data for management planning and control. At the present time many accounting departments are in the process of reevaluating their entire accounting curriculum to determine to what extent present offerings may continue to serve their traditional purpose as well as fulfill the need for an approach which takes into greater consideration the managerial aspect. The non-accounting major terminates his study of accounting with the Managerial Accounting course. The accounting major, almost without exception, takes the same Managerial Accounting course as the non-major which means that he takes the course between Introductory Accounting and Intermediate Accounting.

In the single-IV model, researchers commonly rely on t-ratio-based inference, even though the literature has quantified its potentially severe large-sample distortions. Building on Stock and Yogo (2005), we introduce the tF critical value function, leading to a standard error adjustment that is a smooth function of the first-stage F-statistic. For one-quarter of specifications in 61 AER papers, corrected standard errors are at least 49 and 136 percent larger than conventional 2SLS standard errors at the 5 percent and 1 percent significance levels, respectively. tF confidence intervals have shorter expected length than those of Anderson and Rubin (1949), whenever both are bounded.