مقایسه مقررات خود تنظیمی و قانونی : شواهد از حرفه حسابداری

The accounting profession in the United States recently shifted from self-regulation by peer review to statutory regulation by the Public Company Accounting Oversight Board (PCAOB). Using this shift, I compare outcomes from self-regulation and statutory regulation for the same group of firms. I find that firms choosing their own reviewers, and firms choosing reviewers likely to be connected through prior relationships, tend to receive peer review opinions more favourable than their subsequent PCAOB reports, suggesting that some firms obtained ‘friendly’ reviews in the peer review era. On the other hand, reviewers with relevant industry knowledge are less likely to give such favourable reviews. Further, reviewers from the same geographic area are likely to give peer reviews that are more negative than the subsequent PCAOB reports. Additional analysis suggests that peer reviewers from similar industry or geographic areas bring greater firm-specific expertise to the reviewing process. In the PCAOB regime, I find that firms inspected later tend to receive PCAOB reports more favourable than their peer reviews, suggesting some trends over time in PCAOB reporting. Overall, the findings help in understanding the influences on each approach to regulation, and suggest a nuanced understanding of both approaches as having strengths as well as weaknesses.

The accounting profession in the United States has historically been self-regulated by peer review, through the American Institute of Certified Public Accountants’ (AICPA) Peer Review Program (PRP). Under this system, an accounting firm would typically find another firm to review its quality control procedures and performance of audit engagements. Self-regulation by peer review extends well beyond accounting, to other professions such as law, medicine and actuarial science (Ogus, 1995). Peer review is also the primary mechanism by which research is selected for publication in academic journals, and by which agencies such as the National Science Foundation and National Institutes of Health evaluate research proposals and decide how to allocate their resources. Peer review in accounting has long been criticized for lacking independence. It has been described as “shrouded in secrecy” (Fogarty, 1996), “toothless” (Coffee, 2001), “clubby” (Public Oversight Board, 2002) and even “incestuous” (Williams, 2002).1 These allegations contributed to the demise of self-regulation for US accounting firms. In 2002, the Sarbanes–Oxley Act (SOX) created the Public Company Accounting Oversight Board (PCAOB) and charged it with inspecting public company auditors, marking the onset of statutory regulation for US accounting firms. Statutory regulation under the PCAOB, however, has also not been free of criticism. Some commentators believe that the PCAOB was created, at least partially, from the government’s need after Enron to show that something was being done (Kinney, 2005 and Radin, 2007).2 Accounting firms inspected by the PCAOB criticize it for a “form over substance” approach (Daugherty & Tervo, 2010). Whether the PCAOB’s full-time inspectors have the in-depth expertise (or the incentives to update expertise quickly) that partners and managers currently practicing at a peer firm will have, is also unclear (Glover et al., 2009 and Palmrose, 2006). Both self-regulation and statutory regulation, hence, are believed to have strengths and weaknesses. There is very little empirical evidence, however, on how outcomes differ under self-regulation and statutory regulation. Comparing review outcomes across these alternative systems of regulation, and identifying when and why outcomes differ, can highlight the unique influences on each system and help to understand the relative merits and demerits of self-regulation vis-a-vis statutory regulation. Comparing outcomes for the same firms under self-regulation and statutory regulation is difficult, as the two systems usually do not co-exist. In the US, the recent shift from self-regulation by peer review to statutory regulation by the PCAOB offers a quasi-experimental setting to compare outcomes. In the sample of firms with both a peer review from the last round of the AICPA Peer Review Program pre-SOX, as well as a PCAOB inspection report from the first round of inspections post-SOX, each firm was first peer reviewed and then inspected by the PCAOB within a certain period of time. This allows us to compare outcomes from these two different regulatory approaches for the same group of firms. As the debate between self-regulation and statutory regulation is often characterized as a trade-off between independence and expertise (Glover et al., 2009, Gunningham and Rees, 1997 and Palmrose, 2006), I identify variation across reviewers in these dimensions, and examine how it affects the likelihood of peer reviewers being more (or less) favourable than PCAOB inspectors, relative to the scenario where peer reviewers and PCAOB inspectors agree on their overall assessment of the firm. While peer reviews and PCAOB inspections agree for a majority of the sample, peer reviews are more favourable than PCAOB reports (‘favourable peer reviews’) for 25.3% of the sample, and less favourable than PCAOB reports (‘unfavourable peer reviews’) for 12.8% of the sample. Favourable peer reviews are more likely for firms that choose another firm directly to review them, and for firms that choose a review team drawn from their private CPA association, as opposed to firms that choose an AICPA-assembled review team. Favourable peer reviews are also weakly more likely for firms that review other firms. Assuming that the model controls adequately for changes in firm characteristics between the peer review and PCAOB inspection, this is consistent with claims that some peer reviews in the AICPA regime were “friendly”. Favourable peer reviews are less likely for firms whose reviewers have industry knowledge relevant to the reviewed firm’s client base; such reviewers are more likely to agree with the PCAOB. Further, unfavourable peer reviews are more likely when the reviewer is located close to the reviewed firm. This finding is consistent both with close peer reviewers having more firm-specific knowledge due to their proximity, and alternatively with competition for a common pool of clients inducing them to strategically give negative reviews. In further tests, I find that peer review opinions from close reviewers are better able to predict SEC enforcement actions against the reviewed firm in the future than are opinions from distant reviewers, suggesting tentatively that close reviewers have some superior knowledge that helps them to better identify problems at the firms they review. Finally, while I find no evidence that larger audit firms receive PCAOB reports less favourable than their peer reviews, I find preliminary evidence that firms inspected later tend to receive PCAOB reports more favourable than their peer reviews, suggesting that the PCAOB may have reported deficiencies more aggressively in initial years of inspections. The findings reveal some interesting contrasts between outcomes from self-regulation and statutory regulation. There is evidence that some peer reviewers may have been friendly (consistent with critiques of peer review), but also that some peer reviewers could have brought superior knowledge to the process through greater familiarity with local clients and practices. The patterns over time in PCAOB reporting, on the other hand, suggest influences on PCAOB inspections that also deserve further study. The transition between regimes, therefore, may be more nuanced than a pure trade-off of expertise for independence. Even though this evidence is too preliminary to have policy implications, finding that both approaches to regulation have different strengths supports claims that a combination may work better than either extreme (Doyle, 1997 and Sinclair, 2003). This in turn, echoes suggestions for a “peer-enhanced” inspection model that would augment independent federal inspections with peer expertise (Glover et al., 2009). This study contributes to three streams of literature: first, to the growing literature on governance of the accounting profession. While evidence is mixed on whether peer review opinions correlate with indicators of audit quality (Casterella et al., 2009 and Gunny and Zhang, 2009), the opinions are perceived to be informative (Hillary & Lennox, 2005). Conversely, PCAOB reports correlate predictably with audit quality indicators (Gunny et al., 2009 and Gunny and Zhang, 2009) but are perceived to be uninformative (Lennox & Pittman, 2010). By highlighting that both systems have strengths and weaknesses, this study may help to bridge the mixed findings by suggesting more refined comparisons across regimes, of subsamples of firms or reviewers. Second, by examining how institutional features of the AICPA Peer Review Program affect review outcomes, this study contributes to an interdisciplinary body of work that studies effectiveness of peer review systems and of various ways of designing them, in law (Fortney, 1995 and Fortney, 1997), medicine (Jefferson, Wager, & Davidoff, 2002), psychology (Blackburn & Hakel, 2006), sociology (Bakanic, McPhail, & Simon, 1987), and the sociology of science (Cole et al., 1978 and Travis and Collins, 1991). Third, it contributes to the interdisciplinary literature on the relative merits and demerits of self-regulation and statutory regulation (Coffee, 2001, Green and Hrab, 2003, Gunningham and Rees, 1997 and Ogus, 1995). While the arguments for self-regulation are that it is speedy, flexible and sensitive to market circumstances, criticisms invariably tend towards bias, favouritism, and capture of the system by self-interested parties. This study provides large-scale empirical evidence on some of these strengths and weaknesses. This evidence is also of interest to regulators, as self-regulation still remains the principal regulatory mechanism for accounting firms in the US with only privately held clients, for the accounting profession in other countries, and for other professions. Within the US, efforts are underway at the state level to restructure peer reviews for accounting firms with only privately held clients, including proposals to eliminate firms’ ability to choose their own reviewers so as to reduce conflicts of interest (NYSSCPA, 2005).3 The evidence in this study has the potential to inform this debate.

The accounting profession in the United States was self-regulated by peer review until 2003, when statutory regulation by the Public Company Accounting Oversight Board (PCAOB) was introduced. I use the shift from peer reviews to PCAOB inspections to compare outcomes across self-regulation and statutory regulation for the same group of firms. The results, summarized in Table 9, indicate that firms that choose their own reviewers and choose reviewers likely to be connected through prior relationships tend to receive peer reviews more favourable than their PCAOB reports. Peer reviewers with similar industry expertise as the reviewed firm, on the other hand, are likely to agree with the PCAOB, while reviewers located in the same area tend to be more unfavourable than the PCAOB. Further tests show that peer review opinions from reviewers in the same industry or geographic area are informative about future audit failures, suggesting that these reviewers bring some superior knowledge to the review process. The results suggest a nuanced understanding of both peer reviews and PCAOB inspections as having strengths and weaknesses. While PCAOB inspectors may be less susceptible to “friendly” reviews than peer reviewers, peer reviewers may be able to bring greater familiarity with local clients and practices that may not be easily accessible to PCAOB inspectors. The study is subject to many limitations besides those already noted. There is relatively little variation in reviewer type, raising concerns about the robustness of the documented effects of reviewer type. The fact that a vast majority of firms get ‘firm-on-firm’ reviews could limit our ability to comment on the effectiveness of other methods of identifying reviewers. More broadly, the low explanatory power of the models shows that our understanding of the factors affecting peer reviews and PCAOB inspections is far from complete. Both outcomes could be influenced by unspecified features of the review team or the reviewed firm, allowing for alternative interpretations of the findings.29 Furthermore, the research design is non-experimental. The fact that PCAOB inspections cover only public clients, that peer reviews and PCAOB inspections occur at different points in time and are separated by the Sarbanes–Oxley Act, complicates inferences. While I control for some audit firm characteristics and changes therein, the lack of public data on audit firms makes any such controls rudimentary. SOX, in particular, changed the audit environment considerably, and affected each audit firm differently. The multi-faceted and pervasive impact of SOX cannot be comprehensively controlled for, and is a fundamental limitation of the approach in this study. While it is tempting to interpret the results as suggesting that peer reviews (or PCAOB reports) are “too” positive or too negative for certain firms, no such inference can be made, as there is no objective benchmark for the “right” number of deficiencies expected from any review, done either by the PCAOB or by a peer (Wallace, 1991). Rather, this study can speak only to the more modest objective of comparing outcomes from two different systems of regulation and offering explanations as to why one system produces more positive (or negative) outcomes, relative to the other, in various situations. Finally, while peer review exists in many professions, its strengths and weaknesses lie ultimately in its institutional features. Hence, results from the AICPA’s Peer Review Program may not generalize to other settings. Future research could examine the effectiveness of peer review in other settings where it continues to be the principal regulatory mechanism. Furthermore, our understanding of the PCAOB is still in its infancy. With data from subsequent rounds of PCAOB inspections becoming available, research could examine in more detail the influences on PCAOB inspections, contributing to the debate on self-regulation versus statutory regulation in accounting.