Final Group Report and Presentation

Journal of Accounting and Economics 70 (2020) 101318

Contents lists available at ScienceDirect

Journal of Accounting and Economics journal homepage: www.journals .e lsevier .com/

journal-of -accounting-and-economics

PCAOB international inspections and Merger and Acquisition outcomes

Yongtae Kim a, b, *, Lixin (Nancy) Su c, Gaoguang (Stephen) Zhou d, Xindong (Kevin) Zhu e

a Santa Clara University, USA b Korea Advanced Institute of Science and Technology (KAIST), South Korea c Lingnan University, Hong Kong d Hong Kong Baptist University, Hong Kong e City University of Hong Kong, Hong Kong

a r t i c l e i n f o

Article history: Received 27 February 2018 Received in revised form 13 March 2020 Accepted 27 March 2020 Available online 3 April 2020

JEL classification: G34 M41 M49

Keywords: PCAOB international inspection Audit quality Merger and acquisition

* Corresponding author. Leavey School of Busines E-mail addresses: y1kim@scu.edu (Y. Kim), nancy

(X. (Kevin) Zhu). 1 We refer to public accounting firms as auditors 2 See https://pcaobus.org/International/Registratio

https://doi.org/10.1016/j.jacceco.2020.101318 0165-4101/© 2020 Elsevier B.V. All rights reserved.

a b s t r a c t

This study examines how PCAOB international inspections of non-U.S. auditors affect in- ternational Merger and Acquisition (M&A) outcomes. We find that clients of inspected auditors are more likely to become acquisition targets after the public disclosure of au- ditor's inspection report. We also find that deal completion is more likely and deal announcement returns are higher if deals involve targets with auditors for which in- spection reports are available. Engagement deficiencies and unremediated quality control deficiencies identified in inspection reports weaken the positive effect of PCAOB oversight on M&A outcomes. Collectively, our results suggest that PCAOB oversight reduces infor- mation uncertainty in M&A deals.

© 2020 Elsevier B.V. All rights reserved.

1. Introduction

The creation of the Public Company Accounting Oversight Board (PCAOB) and the requirement that the PCAOB inspect auditors of Securities and Exchange Commission (SEC) registered companies are core provisions of the Sarbanes-Oxley Act (SOX).1 Under SOX and PCAOB rules, auditors must be registered with the PCAOB and are subject to PCAOB inspections. These requirements apply to both U.S. and non-U.S. auditors.2 Extant studies show that the PCAOB international inspection program improves inspected auditors’ audit quality of U.S.-listed clients (Lamoreaux, 2016; Krishnan et al., 2017). Studies also find that PCAOB international inspections have spillover effects on non-U.S.-listed foreign firms and their auditors (Aobdia and Shroff, 2017; Fung et al., 2017). Our study redirects the focus and examines the real economic benefits of the PCAOB international

s, Santa Clara University, Santa Clara, CA 95053, USA. su@ln.edu.hk (L. (Nancy) Su), stephenzhou@hkbu.edu.hk (G. (Stephen) Zhou), xindozhu@cityu.edu.hk

and their clients as audit clients or firms throughout the paper. n/Pages/default.aspx.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 1013182

inspection program. Specifically, we examine the effect of the PCAOB international auditor inspection program on global acquisitions that involve non-U.S.-listed foreign target firms.

Merger and Acquisition (M&A hereafter) activities are among the most significant corporate events that affect various stakeholders, foster managerial discipline, and help allocate capital in an economy.3 Information uncertainty about the value of targets and expected synergy from deals is an important concern that hinders M&A activities. As such, it is not surprising that several studies examine the effect of financial reporting quality on M&A outcomes (e.g., Skaife and Wangerin, 2013; Amel-Zadeh and Zhang, 2015; McNichols and Stubben, 2015). Auditors exert significant influence on clients’ financial reporting quality. Prior studies consider audit quality from two perspectives: actual and perceived audit quality (DeFond and Zhang, 2014). Because information uncertainty regarding the target firms affects the M&A outcomes, both perceived and actual audit quality of the targets can shape the M&A outcomes. Due to the unobservable nature of audit quality and identification problems (DeFond and Zhang, 2014), however, isolating the effect of audit quality on M&A activities and outcomes is a challenging task. Disentangling the auditor effect arising from the assurance role and that arising from theM&A advisory role is also difficult.4

The PCAOB international inspections of non-U.S. auditors offer an ideal setting to examine the effect of the quality of financial statement audits on M&A outcomes for several reasons. First, compared with the U.S. setting, the international setting allows a more powerful test of the effect of audit quality on M&A outcomes. U.S. auditors are unlikely to experience a significant audit quality improvement from PCAOB inspections because their audit quality is already relatively high due to other forms of scrutiny in place and a strict litigation environment (Francis and Wang, 2008; Shroff, 2019). Second, because PCAOB international inspections are largely exogenous to non-U.S. auditors' non-U.S.-listed clients, we can better isolate the effect of audit quality from the effect of the firms’ innate characteristics and financial reporting systems. Clients of never- inspected auditors also work as a control group and help identify the effect of audit quality on M&A outcomes. Finally, because PCAOB inspections improve the actual and perceived quality of financial statement audits (Fung et al., 2017; Krishnan et al., 2017), we can attribute their effect on M&A outcomes to the assurance role of auditors and disentangle it from their advisory role in M&As (Louis, 2005; Xie et al., 2013).

Although PCAOB international inspections of non-U.S. auditors aim to improve audit quality for U.S.-listed firms registered with the SEC (i.e., SEC registrants), the benefits can spill over to the inspected auditors' home-country clients. PCAOB in- spectionsmay serve as a signal of high audit quality not just for SEC registrants whose audit engagements are inspected by the PCAOB, but also for non-SEC-registered clients. Inspections ensure that the audit process conforms to certain minimum standards of quality and independence (Aobdia and Shroff, 2017). As auditors improve their audit process for SEC registrants to conform to the standards set by the PCAOB, they can apply such improvements to audits of their other clients. Thus, in- vestors may consider inspected auditors' audit quality to be higher not only for their SEC registrants but also for their non- SEC-registered, home-country clients. Furthermore, PCAOB inspections examine not only specific engagements but also auditor-level quality control systems, which helps improve the audit quality for all clients, including non-U.S. auditors' non- U.S.-listed clients whose audits are not directly inspected by the PCAOB. Thus, PCAOB oversight helps build investors' trust in audits for inspected auditors’ home-country clients.

Due to the uncertainty about investment opportunities and the cost of collecting private information, reliable accounting information expandsmanagers' information sets and alters their investment decisions (Roychowdhury et al., 2019). Acquirers seeking M&A opportunities rely on financial information to identify and evaluate potential targets. Poor financial reporting quality of targets hampers acquirers' ability to evaluate targets. To the extent that acquirers perceive that PCAOB international inspections improve audit quality (not only for SEC-registered clients but also for other home-country clients) and reduce information uncertainty about potential targets and acquisition synergy, we expect that clients of PCAOB-inspected auditors are more likely to receive M&A bids. We also posit that the likelihood of M&A completion is higher for deals that involve targets with inspected auditors than those that involve targets with uninspected auditors. PCAOB oversight increases acquirers’ confidence in the financial reporting quality of targets, lowering the likelihood of deal termination due to concerns about misreporting. The reduced information uncertainty of targets also enables acquirers to identify better investment opportunities, leading to higher deal quality. Thus, we predict that the combined returns of the acquirer and the target around the deal announcement, a popular proxy for deal quality, are higher for deals that involve targets with PCAOB-inspected auditors.

Using data from 35 countries that have at least one PCAOB inspected auditor, we show that compared to firms with uninspected auditors, those with inspected auditors are more likely to receive an acquisition bid. We also find that the likelihood of deal completion is greater and deal announcement returns are higher if deals involve targets with inspected auditors. We find that the positive effect of PCAOB oversight on M&A outcomes appears after the public disclosure of the inspection report, rather than immediately after the inspection. Considering that changes in audit quality (for example, through improvements in audit processes) typically occur after completion of the PCAOB inspection, the evidence suggests that perceived increases in audit quality, rather than actual changes, affect M&A outcomes. The effect of PCAOB oversight on

3 Moeller et al. (2005), for example, estimate that their sample companies spent around US $3.4 trillion (in 2001 U.S. dollars) on acquisitions from 1980 to 2001.

4 Auditors play an assurance role in M&As to the extent that acquirers rely on audited financial statements to make M&A decisions. They also play an advisory role as both acquirers and targets seek advice from auditors. Auditors also play an information intermediary role, as discussed in Cai et al. (2016).

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 3

M&A outcomes is more pronounced for targets located in countries with weaker legal institutions and for deals with greater information uncertainty.

We further test whether the findings disclosed in inspection reports matter. PCAOB inspects both specific audit en- gagements and auditor-level quality control systems and reports findings from both inspections. Part I findings refer to audit engagement deficiency, and Part II findings refer to quality control deficiency. We find that the likelihood of receiving a bid, the likelihood of deal completion, and deal quality are higher following an inspection report that contains no audit engagement deficiency. We find lower benefits when the report contains an engagement deficiency. We also find that an inspection report with an unremediated quality control deficiency is associated with relatively lower levels of the likelihood of receiving a bid and the likelihood of deal completion. Finally, we find that post-acquisition outcomes, reflected in di- vestitures and goodwill impairments, are better for M&As involving targets with PCAOB-inspected auditors, providing additional evidence on the positive effect of PCAOB oversight on M&A deal quality.

Our study contributes to the literature in several ways. First, our study adds to the debate on the value and effectiveness of the PCAOB international inspection program by showing the real economic benefits of such oversight in the global M&A market. Proponents of PCAOB inspections believe that the inspection program can improve audit quality because it helps overcome the independence problem in the peer-review regime. Proponents also view PCAOB inspections as certifying audit quality in the international setting, which is otherwise unknown due to the lack of comparably rigorous public oversight in most non-U.S. jurisdictions (Aobdia and Shroff, 2017). Opponents of PCAOB inspections, however, question the effectiveness of such inspections for several reasons. In particular, they question the inspectors' expertise (DeFond, 2010). An important issue in this debate is how best to quantify and assess the economic impact of PCAOB inspections. Recent studies render some support for the benefits of PCAOB inspections by showing an increase of inspected auditors' audit quality for both SEC registrants (Lamoreaux, 2016; Krishnan et al., 2017) and non-U.S.-listed clients (Fung et al., 2017), and an expansion of market share among non-U.S. clients (Aobdia and Shroff, 2017). In this paper, we show that real economic benefits accrue to the non- U.S.-listed clients of these inspected auditors. The results also suggest that the PCAOB inspection program leads to more efficient capital allocation, as evidenced by improved M&A deal quality. Given the economic significance of M&As and their impact on various stakeholders, our study contributes to the debate by documenting an important benefit of PCAOB inter- national inspections in global capital markets. We note, however, that because not all costs of the PCAOB international in- spection program are observable to us and we examine just one of many potential benefits of the program, our results cannot speak to the program's overall net benefit/cost.

Second, this study contributes to the literature that examines the effect of audit quality on M&A outcomes. Auditors play various roles (e.g., assurance, advisory, and information intermediary roles) in M&As. Focusing on auditors' advisory role, Louis (2005) finds that acquirers with non-Big N auditors significantly outperform those with Big N auditors in M&As and attributes the results to non-Big N auditors' superior local knowledge and close relationships with their clients. Focusing on auditors' assurance role, Xie et al. (2013) show that firmswith Big N auditors aremore likely to become acquisition targets and have a higher deal-completion rate. In the global audit market, differentiating audit quality based on the Big N/non-Big N classification is problematic, because auditors’ incentives to provide high quality assurance services vary across institutional characteristics (Francis and Wang, 2008). Exploiting PCAOB international inspections as events that bring changes to audit quality, our study provides new evidence on the effect of audit quality on acquisition outcomes. The evidence supports the assurance role of auditors in M&As.

Third, our study contributes to the literature that examines the real effect of auditing. Kausar et al. (2016) find that by hiring auditors voluntarily, small private firms can reduce information asymmetry and financing frictions, which leads to greater investment and financing. Bae et al. (2017) document that hiring large or industry-specialist auditors can improve client firms' investment efficiency, as such auditors provide informational advantages arising from their knowledge and resources. Shroff (2019) shows that non-U.S.-listed clients' access to external financing, as well as their capital investments, increase after the disclosure of PCAOB inspection reports on these clients’ non-U.S. auditors. We add to this literature by showing the positive effects of PCAOB oversight on outcomes from global M&A activities. While both Shroff (2019) and our study examine the spillover effect of PCAOB international inspections on non-U.S.-listed firms, we examine the benefits that accrue to these firms as acquisition targets, whereas Shroff (2019) examines their financing and investment activities.

Finally, this study contributes to the debate on the effectiveness of public regulation in the international audit setting. Regulation economists argue that regulations may not be able to achieve the desired goals because of resource constraints, regulatory capture, and political pressure (e.g., Demsetz, 1968; Stigler, 1971). La Porta et al. (2006), for example, show that private enforcement of securities laws benefits stock market development more than public enforcement does. Recent studies, however, claim that regulation has positive effects, as it enhances enforcement (Jackson and Roe, 2009) and provides a positive externality (e.g., Aobdia and Shroff, 2017; Fung et al., 2017; Minnis and Shroff, 2017).5 Against this backdrop, we examine whether PCAOB international inspections generate a positive externality in the context of M&As. By showing the positive externality of U.S. audit oversight to non-U.S.-listed firms in global capital markets, our study provides novel evidence that sheds light on the effectiveness of public regulation.

5 Minnis and Shroff (2017) provide an excellent discussion on the costs and benefits of regulations in disclosure and auditing.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 1013184

The rest of the paper is organized as follows: Section 2 reviews the literature and develops hypotheses. Section 3 provides the sample selection procedure and descriptive statistics. Section 4 outlines the research design and presents the empirical results. Section 5 concludes the paper.

2. Literature review and hypothesis development

2.1. PCAOB inspections

SOX Section 104 requires the PCAOB to inspect registered auditors of SEC-registered companies (issuers). The PCAOB started inspecting U.S. auditors from 2003 and non-U.S. auditors from 2005. PCAOB inspections are classified into two major categories: audit engagements performed by auditors and auditor-level quality control systems. Typically, each inspection report consists of four parts. Part I discusses the audit engagement deficiencies, Part II concerns quality control criticisms, Part III suggests some specific approaches that an audit committee might consider for initiating or enhancing inspection-related discussions with an auditor, and Part IV consists of the auditor's responses to the PCAOB's draft inspection report. In almost all cases, Part III is redacted. While the PCAOB reports significant audit engagement deficiencies and related information in Part I and makes it public, the quality control deficiency information in Part II is not released to the public and remains confi- dential.6 If auditors cannot remediate auditor-level quality control deficiencies within 12 months after receiving the in- spection report, the PCAOB discloses them in Part II of the inspection report by issuing a follow-up explicit quality-control criticism and updating the original report.7

2.2. Prior studies on PCAOB inspections

Early studies cast doubt on the effectiveness of PCAOB inspections.8 Recent empirical studies, however, generally support the notion that PCAOB inspections are effective in the U.S. (Carcello et al., 2011; Gramling et al., 2011; Gunny and Zhang, 2013; DeFond and Lennox, 2017). Carcello et al. (2011), for example, show that audit clients’ abnormal accruals drop significantly in the two years following the first PCAOB inspection. Other studies show audit market consequences of PCAOB inspections in the U.S. (Abbott et al., 2012; Nagy, 2014; Aobdia, 2018). Gipper et al. (2019) show that the stock market response to earnings news increases following PCAOB oversight, suggesting that regulatory oversight improves financial reporting credibility.

Studies also find that PCAOB inspections have a significant effect on non-U.S. auditors. The PCAOB's inspections of non- U.S. auditors introduce another layer of scrutiny and public oversight (i.e., PCAOB) from the U.S., where audit regulatory oversight is arguably stricter than that in most other countries. Lamoreaux (2016) documents that auditors in jurisdictions allowing PCAOB inspection access exhibit higher audit quality for foreign clients cross-listed in U.S. markets than those not subject to PCAOB inspection. Krishnan et al. (2017) further show lower abnormal accruals and greater value relevance of accounting numbers in U.S.-listed foreign clients of inspected auditors following PCAOB initial inspections. Using a sample of non-U.S.-listed foreign clients of PCAOB-registered non-U.S. auditors, Fung et al. (2017) show that the financial reporting quality of these firms increases after the initial PCAOB inspections, suggesting that the benefits of PCAOB inspections can spill over to non-U.S.-listed clients whose audits are not inspected by the PCAOB. Aobdia and Shroff (2017) show that non- U.S. inspected auditors benefit from PCAOB inspections, as their market shares in non-U.S.-listed clients increase after the release of PCAOB inspection reports, but audit engagement deficiencies reduce the benefits. Shroff (2019) finds that non- U.S.-listed firms audited by PCAOB-inspected auditors raise significantly more external capital and increase capital ex- penditures following the public disclosure of their auditors' PCAOB inspection report. Our study extends prior studies by examining the effect of PCAOB inspections on global M&A activities that involve non-U.S.-listed clients of PCAOB-inspected non-U.S. auditors.

2.3. The role of auditors in M&As

The M&A process can be broadly divided into three stages, which are characterized by different degrees of reliance on public information (Skaife and Wangerin, 2013; Chen et al., 2018). In the first stage, acquirers conduct preliminary due diligence, which is based primarily on public information, to screen and assess targets.9 Based on the results of the preliminary due diligence, acquirers may short-list targets and sign a confidentiality agreement with targets to start in- depth due diligence. Although acquirers can access private information at this stage, they continue to rely on public information, as targets may provide biased private information, and costs and time constraints could limit the extraction

6 Part I includes audit deficiencies of inspected auditors, explanatory information from the PCAOB, and the inspection process for the largest auditors (CAQ, 2012).

7 Readers may refer to Aobdia and Shroff (2017) and Shroff (2019) for more details about the PCAOB international inspection program. 8 Glover et al. (2009) argue that PCAOB inspections are “inefficient and dysfunctional.” Lennox and Pittman (2010) show that audit clients do not perceive

PCAOB reports as being informative. Using the PCAOB's censure of Deloitte in 2007 as an event, Boone et al. (2014) show that Deloitte's audit quality was not significantly distinguishable from that of the other Big 4 auditors during the three-year windows before and after the sanction, respectively.

9 The same process applies regardless of whether the M&A is initiated by an acquirer or a target.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 5

of private information. After in-depth due diligence, acquirers may sign an acquisition agreement with targets and subsequently disclose the acquisition to the public, entering into the transactional due-diligence stage. This stage verifies the accuracy of the representations and warranties in the financial statements and other reports in the acquisition agreement.

As financial statements are important sources of information in the M&A process and falsified financial statements constitute a condition for deal termination, considerable research attention has been given to the consequences of financial reporting quality in M&As. Skaife and Wangerin (2013) find a significant association between targets' poor reporting quality and the probability of renegotiation and deal termination. McNichols and Stubben (2015) posit that high quality accounting information reduces uncertainty in the valuation of the target and leads to better bidding decisions in acquisitions. They find that acquirers’ returns around deal announcements are higher when targets have higher quality accounting information.

The M&A process for international M&As and cross-border M&As (Campa and Moschieri, 2008; Rothenbuecher and von Hoyningen-Huene, 2008; Dickey, 2013) is largely the same as that for M&As between U.S. domestic firms. Erel et al. (2012) find that the likelihood of cross-border mergers increases with the country-level accounting disclosure quality, confirming the role of financial reporting quality in global M&As. Similarly, Francis et al. (2016) argue that the difference in accounting standards across countries exacerbates information uncertainty in cross-border M&As, and find that such a difference is significantly negatively associated with the volume of cross-border M&As.

The role of auditors in M&As is more encompassing than their role in periodic financial reporting, as auditors play not only an assurance role but also an advisory role in M&As. Louis (2005) argues that non-Big N auditors have superior knowledge of local markets and better relationships with their clients. Consistent with this argument, he shows that acquirers with non-Big N auditors significantly outperform those with Big N auditors at merger announcements and that this effect is more sig- nificant when targets are privately held, supporting the advisory role of auditors in M&As. In contrast, Xie et al. (2013) focus on the assurance role and find that firms with Big N auditors are more likely to become M&A targets and are more likely to ultimately be acquired in M&A deals. In addition, Cai et al. (2016) and Dhaliwal et al. (2016) contend that auditors who audit both acquirer and target firms work as an information intermediary and reduce uncertainties in M&A deals. The demand for high audit quality is stronger in the international setting because investors are skeptical about the effectiveness of audit monitoring regimes outside the U.S. (Francis and Wang, 2008; Gul et al., 2013; El Ghoul et al., 2016). This study extends prior studies by examining whether PCAOB inspections of non-U.S. auditors can help facilitate M&As and improve M&A outcomes in the global M&A market.

2.4. Hypothesis development

Roychowdhury et al. (2019) review the literature on the effect of financial reporting and disclosure on corporate in- vestments and summarize the roles of financial reporting and disclosure in two broad categories: financial reporting and disclosure 1) resolve agency frictions and 2) expand managers' information set and help managers learn about investment opportunities. To the extent that private information collection is costly and managers' information set regarding investment opportunities is incomplete, managers can learn from (their own or others') financial reporting to expand their information set, reduce information uncertainty, and change their investment decisions. In the setting of M&As, acquirers seeking M&A opportunities learn from potential targets' financial information to identify and evaluate many aspects of potential targets, including financial prospects, risks, and valuations. As such, poor financial reporting quality among potential targets creates difficulties for acquirers (Martin and Shalev, 2016). High-quality audits improve the credibility of financial reports and, in turn, boost acquirers' confidence about using financial statements to assess M&A opportunities (Xie et al., 2013). Prior studies find that audit quality is influenced by the strength of the enforcing institutions that govern auditors (Francis andWang, 2008; Gul et al., 2013; El Ghoul et al., 2016). Oversight by the PCAOB, an internationally recognized enforcing agency, can serve as a certification of audit quality (Aobdia and Shroff, 2017) and thus increase prospective acquirers’ confidence in using financial statements to identify targets and assess their value.

PCAOB international inspections target inspected auditor's work on SEC registrants, as well as their auditor-level quality control systems. Although engagement-level inspections are performed on the audit engagements of SEC registrants only, the benefits of PCAOB oversight can spill over to the inspected auditors' other home-country clients. Auditors that are subject to PCAOB inspections would improve their audit process for SEC registrants to conform to the standards set by the PCAOB. Such improvements can be applied to audits of their non-U.S.-listed clients whose audit engagements are not directly inspected by the PCAOB, improving audit quality of such engagements. To the extent that market participants understand this spillover effect, PCAOB international inspections can provide a signal of high audit quality not only for inspected auditors' SEC registrants but also for their other home-country clients. In addition, improvements made to auditor-level quality control systems in response to PCAOB international inspections would help improve audit quality for all clients, including non-U.S. auditors' non-SEC-registered, home-country clients. If acquirers perceive the financial reporting quality of inspected auditors' clients to be more reliable, uncertainty regarding the prospective target's value and

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 1013186

expected synergy from acquisitions will be greatly reduced, making acquirers' evaluations easier. Thus, we expect the likelihood of receiving an acquisition bid to be higher for non-U.S.-listed firms with PCAOB-inspected auditors than for other firms.

During the inspection fieldwork, PCAOB inspectors discuss identified material accounting errors or significant auditing deficiencies with the auditors being inspected; they encourage the auditor to correct the problems and improve its audit process, even before the problems are reflected in the inspection reports (PCAOB, 2005; Krishnan et al., 2017). Therefore, undergoing a PCAOB inspection itself can signal the financial reporting quality of firms with PCAOB-inspected auditors. In addition, PCAOB inspection reports either expose deficiencies of both the inspected audit engagements and auditor-level quality control systems or provide a clean bill of health. Thus, the disclosure of inspection reports also signals financial reporting quality. Hanlon and Shroff's (2020) survey of inspectors from auditor public oversight boards in 20 countries suggests that public disclosure of inspection reports is one of the most important tools to induce changes in auditor behavior. Following Shroff (2019), we test the two treatment effects, the inspection treatment effect and the report disclosure treatment effect, separately. We state our hypotheses in an alternative form, as follows:

H1a. Compared to firms with uninspected auditors, firms with PCAOB-inspected auditors are more likely to receive an M&A bid, following the PCAOB inspection fieldwork.

H1b. Compared to firms with uninspected auditors, firms with PCAOB-inspected auditors are more likely to receive an M&A bid, following the disclosure of a PCAOB inspection report.

Acquirers continue to rely on financial information tomake relevantM&A decisions, even after an acquisition agreement is signed. Although targets affirm the accuracy of financial information in the acquisition agreement, the agreement cannot guarantee unbiased financial statements as targets have incentives to manipulate accounting numbers to obtain favorable terms in M&A deals. PCAOB oversight increases acquirers' confidence in the targets' financial reporting quality, lowering the likelihood of deal termination due to concerns about misreporting. As discussed before, acquirers may update their per- ceptions about the target firms’ financial reporting quality after the auditor undergoes PCAOB inspections or after the public disclosure of the PCAOB inspection report. Thus, we state our second set of hypotheses in an alternative form by separating the two treatments, as follows:

H2a. Compared to deals that involve targets with uninspected auditors, M&A deals that involve targets with PCAOB- inspected auditors, following the inspection fieldwork, are more likely to be completed.

H2b. Compared to deals that involve targets with uninspected auditors, M&A deals that involve targets with PCAOB- inspected auditors, following the disclosure of inspection report, are more likely to be completed.

Lastly, the reduced information uncertainty associated with targets with PCAOB-inspected auditors enables acquirers to identify better investment opportunities. Uncertainties negatively affect M&A quality by preventing acquirers from identi- fying profitable investment opportunities (Goodman et al., 2014). Poor quality financial information of targets increases uncertainty about the targets' earnings potential, financial health, and risk. By improving the financial reporting quality of firms with inspected auditors, PCAOB inspections expand acquirers’ information set and reduce uncertainties, which in turn allow them to identify better investment opportunities. Thus, we expect that PCAOB inspections lead to higher deal quality for M&As that involve targets with inspected auditors. Following Bradley et al. (1988) and Cai et al. (2016), we use the combined abnormal returns of the acquirer and the target around the deal announcement as a proxy for deal quality. We predict that announcement returns are greater for deals that involve targets with PCAOB-inspected auditors than for deals that involve targets with uninspected auditors. Our final set of hypotheses is stated in an alternative form by separating the two treat- ments, as follows:

H3a. Compared to announcement returns of deals that involve targets with uninspected auditors, announcement returns of deals that involve targets with PCAOB-inspected auditors, following the inspection fieldwork, are greater.

H3b. Compared to announcement returns of deals that involve targets with uninspected auditors, announcement returns of deals that involve targets with PCAOB-inspected auditors, following the disclosure of inspection report, are greater.

Although it is reasonable to expect that PCAOB inspections help facilitate M&As and improve M&A outcomes, there are some arguments against our predictions. First, many opponents are concerned about whether PCAOB inspectors have the cutting-edge, industry-specific expertise needed to detect deficiencies, as they are not practicing certified public accountants (CPAs) (Glover et al., 2009; Lennox and Pittman, 2010).10 Second, critics of PCAOB inspections are concerned about whether PCAOB inspectors can substantially improve audit quality, as they may rigidly comply with PCAOB rules and standards and

10 Hanlon and Shroff (2020), however, find that the vast majority of inspectors worked as auditors at Big-N accounting firms and reached the title of manager or senior manager, suggesting that inspectors have significant audit experience.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 7

simply focus on documentation and substantiation. Third, geographic distance and differences among cultures and business practices may impair effective communication between inspectors and non-U.S. auditors and, in turn, may negatively affect the effectiveness of inspections (Fung et al., 2017). Finally, acquirers may not be aware of PCAOB inspections and may hire their own agencies to verify targets’ financial statements. These factors may reduce the usefulness of PCAOB inspections in global M&As and lead to the lack of a relation between PCAOB inspections and M&A outcomes. We therefore turn to the data to test our predictions.

3. Data and sample selection

We first collect all PCAOB inspection reports of non-U.S. auditors from the PCAOB website, with the last inspection report dated 2015.11 In total, we collect 504 PCAOB reports corresponding to 261 non-U.S. auditors from 45 countries, disclosed in 2006e2015. We treat each Big 4 auditor in the different countries as a different auditor. The public release date of each in- spection report is published on the PCAOB website, together with the inspection reports. By reading the inspection reports, we can identify each period of inspection fieldwork. Typically, an inspection fieldwork period lasts one to twoweeks, and the time lag between the completion of inspection fieldwork and the public release of the inspection report is 1.5þ years, on average (Aobdia and Shroff, 2017; Shroff, 2019). We identify PCAOB inspection reports with engagement deficiencies in the Part I findings and/or auditor-level quality control deficiencies in the Part II findings by reading the inspection reports. Engagement deficiencies can be identified as soon as the inspection reports are publicly available on the PCAOB website. Following Hermanson et al. (2007) and Bishop et al. (2013), we classify the inspection reports into three types concerning the auditor-level quality control deficiencies: a clean report without quality control deficiencies (Type A); a report with an explicit quality control criticism disclosed to the public, as the auditor has failed to remedy the deficiencies within 12 months (Type C); or a report with remediated quality control deficiencies (Type B). When PCAOB inspectors do not find auditor-level quality control deficiencies, they disclose the inspection report without Part II.12 Thus, we can differentiate Type A and the other two types as soon as the inspection report is disclosed. As the PCAOB waits 12 months for the auditor to remediate quality control deficiencies before it publicly discloses any unremediated quality control deficiencies, we cannot distinguish between Type B and Type C reports until later. The PCAOB inserts a comment of “QC criticisms now public,” by which we identify the Type C report, and updates the original inspection report by adding a section of “Issues Related to Quality Controls” in Part II. Following Nagy (2014) and Aobdia and Shroff (2017), we infer the disclosure dates of Type C findings as the dates of the PDF files of the updated inspection reports. We identify Type B reports as those that are not defined as Type A or Type C by the conclusion of our data collection.13

We separately extract all firm-year observations available in the Compustat Global database from 2002 to 2016. We start our sample period in 2002, following the enactment of SOX and the creation of the PCAOB. The lead time before the first inspection report in 2006 ensures that we have observations before the first set of PCAOB inspections. Our sample period ends in 2016, which allows at least one year of the post-inspection period for our final inspection report, dated 2015. We obtain auditor identity data from the Capital IQ database because Compustat Global provides auditor identity for only about one-half of the firm-year observations (e.g., Fung et al., 2017) and miscodes auditor identity in some countries (e.g., India, Japan, Pakistan, and South Korea) (Francis and Wang, 2008).14

We identify international M&A bids for public acquirers and public targets from the Securities Data Company (SDC) database. Following Cai et al. (2016), we only consider bids in which the deal value is more than one million U.S. dollars or more than 1% of the acquirer's market value prior to the deal announcement date. We remove those countries without any PCAOB-inspected auditors.15,16 Since wewant to examine the spillover effect of PCAOB inspections, we also drop foreign firms

11 The first PCAOB international inspection report was issued in 2006 for non-U.S. auditors from Canada and the U.K. (https://pcaobus.org/Inspections/ Reports/Pages/default.aspx). 12 In some cases, Part II contains the auditor's response to the PCAOB draft inspection report but no mention about quality control deficiencies. We classify these reports as Type A reports. 13 The PCAOB requires the auditor to remedy quality control deficiencies within 12 months, but, on average, the time lag between the initial report disclosure date and the Type C disclosure date (that is, the date that PDF is updated) is two years in our sample. Therefore, we cannot identify a report as Type B simply when 12 months has passed since the disclosure of the original report and the PDF file are not updated. 14 Following Shroff (2019), we clean the auditor identities, as the auditor names are not uniformly coded in the Capital IQ. First, we collect the information of worldwide member firms' names and addresses disclosed in the Big 4's official websites. For example, PwC in Israel is called Kesselman & Kesselman C. P. A.s. EY in Argentina is called Pistrelli, Henry Martin y Asociados S. R. L. Second, as some Big 4s do not disclose the worldwide member firms' local names or historical names before merging with the Big 4, we search the historical names of the Big 4 via search engines in countries with inspected auditors, especially in countries with a higher likelihood of having local names, i.e., Argentina, Brazil, India, Israel, Japan, and Mexico. For auditor information in Japan and Korea, we also refer to the NEEDS (Nikkei Economic Electronic Databank System) CD-ROM and KIS-VALUE database, respectively. We find that AF Ferguson & Co was the fifth chartered accountant firm to join Deloitte in India (Deloitte Haskins & Sells), after CC Chokshi & Co, Fraser & Ross, PC Hansotia & Co, and SB Billimoria & Co. Another example is in Japan, ChuoAoyama Audit Corporation was the Japanese affiliate of PwC from 2000 to 2006. 15 This happens when none of the auditors in the country has any engagements with a client listed in the U.S. markets or when the country denies PCAOB inspections. 16 Hong Kong allows PCAOB inspections only to the extent that the inspections do not involve a review of audit work relating to a client's operations in China. In other words, since China does not allow the PCAOB to inspect its auditors, Hong Kong does not allow inspections of Hong Kong auditors with respect to their clients' China operations. In a sensitivity analysis, we treat firms audited by Hong Kong auditors as firms with uninspected auditors and inferences from the results (untabulated) remain unchanged.

Table 1 Sample description. Panel A reports sample selection for firm-year level analyses, Panel B reports sample selection for deal-level analyses, and Panel C and D report the sample distribution across countries and years, respectively, based on 70,503 firm-year observations and 1285 deals. In Panel C, the mean values of TARGET, COMPLETION and COMBINECAR are the average values calculated based on country-level means. In Panel D, the mean values of TARGET, COMPLETION and COMBINECAR are the average values calculated based on annual means.

Obs. Firms Countries

Panel A: Sample selection for firm-year level analyses Merge of Compustat Global, Capital IQ, and SDC from 2002 to 2016 500,869 48,461 126 Less: Countries without any PCAOB-inspected auditors (109,105) (10,479) (83) Less: Firm-years of SEC registrants (6214) (453) (0) Less: Firms that switched auditors (108,578) (8283) (0) Less: Observations in financial industries (70,259) (6944) (0) Less: Observations with missing control variables (136,210) (10,098) (8) Test of the likelihood of receiving a bid 70,503 12,204 35

Deals Target firms Target countries

Panel B: Sample selection for deal-level analyses Deals with public acquirers and public targets in SDC from 2002 to 2016 11,063 9792 108 Less: Deals with U.S. target firms (3312) (3122) (1) Less: Deals that cannot be matched with Compustat Global and Capital IQ (3455) (2711) (28) Less: Deals with target firms from countries without any PCAOB-inspected auditors (389) (365) (38) Less: Deals with non-U.S. target firms of SEC registrants (40) (39) (0) Less: Deals with target firms that switched auditors (99) (58) (0) Less: Deals with target firms in financial industries (661) (603) (0)

3107 2894 41 Test of the likelihood of deal completion Less: Deals with missing control variables (1822) (1669) (6)

1285 12251 35 Test of announcement returns Less: Deals with missing control variables (2027) (1862) (6)

1080 10321 35

Column (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

Country No. of firm-year obs.

TARGET No. of Deals

COMPLETION COMBINECAR No. of unique auditors inspected

No. of inspections2 No. of inspections with engagement deficiencies

No. of inspections with TYPEA findings

No. of inspections with TYPEB findings

No. of inspections with TYPEC findings

Panel C: Distribution across countries Argentina 283 0.007 3 0.667 0.035 4 15 6 7 8 0 Australia 5085 0.034 202 0.738 0.027 10 22 12 6 15 1 Bermuda 83 0.024 2 1.000 �0.031 4 15 4 8 7 0 Brazil 202 0.054 14 0.786 �0.007 5 14 11 2 11 1 Canada 2390 0.031 291 0.863 0.023 21 72 55 8 60 4 Chile 324 0.019 6 0.667 0.024 4 12 8 3 9 0 Colombia 74 0.012 1 1.000 0.000 4 6 4 1 5 0 Denmark 364 0.008 4 1.000 0.041 1 1 0 1 0 0 Finland 536 0.017 10 0.900 0.086 1 1 1 0 1 0 France 1906 0.013 28 0.964 �0.007 2 2 2 0 2 0 Germany 2325 0.011 26 0.692 0.042 3 3 3 0 3 0 Greece 670 0.009 8 0.625 0.032 2 2 0 1 1 0 Hong Kong 4904 0.005 24 0.708 0.001 1 1 1 0 1 0 India 9538 0.004 44 0.682 �0.002 9 15 8 3 11 1 Indonesia 302 0.007 3 1.000 0.079 2 4 2 0 4 0

Y.K im

et al./

Journal of

A ccounting

and Econom

ics 70

(2020) 101318

8

Israel 943 0.012 13 0.923 0.033 7 20 14 4 16 0 Japan 9357 0.021 250 0.976 0.023 6 16 13 3 12 1 Korea 928 0.008 12 0.917 �0.005 5 11 6 5 5 1 Malaysia 4202 0.007 32 0.938 0.031 4 4 1 1 2 1 Mexico 265 0.008 2 1.000 0.044 5 16 13 2 13 1 Netherlands 348 0.034 13 0.846 0.022 3 3 3 0 3 0 New Zealand 707 0.008 6 0.500 0.022 3 6 1 4 2 0 Norway 926 0.028 31 0.742 0.012 4 7 4 2 5 0 Peru 131 0.008 1 1.000 0.022 3 4 0 2 2 0 Philippines 633 0.005 3 0.667 0.025 4 6 2 3 3 0 Russia 339 0.021 7 0.571 �0.069 4 7 4 3 4 0 Singapore 3470 0.010 39 0.743 0.036 5 9 1 6 3 0 South Africa 595 0.018 15 0.467 �0.021 7 14 4 10 4 0 Spain 501 0.010 5 1.000 0.072 2 2 2 0 2 0 Sweden 1379 0.018 29 0.862 0.027 2 2 2 0 2 0 Switzerland 1150 0.022 25 0.880 0.035 5 7 7 0 7 0 Taiwan 7375 0.004 41 0.610 0.035 4 11 3 6 5 0 Thailand 1985 0.006 11 0.727 �0.056 1 2 0 2 0 0 Turkey 259 0.012 3 0.667 �0.014 2 2 2 0 2 0 United Kingdom 6024 0.011 81 0.753 0.001 8 15 8 4 11 0 Total/mean 70,503 0.015 1285 0.802 0.018 157 349 207 97 241 11

Column (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

Year No. of firms

TARGET No. of Deals

COMPLETION COMBINECAR No. of auditors initially inspected

No. of inspections

No. of inspections with engagement deficiencies

No. of inspections with TYPEA findings

No. of inspections with TYPEB findings

No. of inspections with TYPEC findings

Panel D: Distribution across years 2002 2240 0.028 26 1.000 0.005 0 0 0 0 0 0 2003 2517 0.020 65 0.814 0.023 0 0 0 0 0 0 2004 2762 0.027 106 0.895 0.006 0 0 0 0 0 0 2005 3397 0.029 106 0.848 0.028 0 0 0 0 0 0 2006 4113 0.026 143 0.852 0.014 2 2 0 1 1 0 2007 4443 0.017 138 0.848 0.020 7 7 6 2 5 0 2008 4759 0.015 114 0.789 0.015 18 19 6 13 6 0 2009 4933 0.014 98 0.866 0.020 6 8 4 0 7 1 2010 5006 0.010 100 0.810 0.031 31 35 30 5 26 4 2011 5191 0.009 82 0.853 0.029 45 90 42 35 50 5 2012 5568 0.007 71 0.789 0.006 12 40 19 14 26 0 2013 6103 0.008 48 0.688 0.022 7 52 36 10 42 0 2014 6203 0.010 66 0.864 0.036 21 54 42 7 46 1 2015 6441 0.009 71 0.662 0.018 8 42 22 14 28 0 2016 6827 0.007 51 0.745 0.023 0 0 0 0 0 0 Total/mean 70,503 0.015 1285 0.821 0.020 157 349 207 97 241 11

Notes: 1. There are several reasons why we have more “deals” than “target firms”. First, we do not limit our sample deals to 100% acquisitions (i.e., acquisitions that result in an acquirer obtaining 100% of target's equity), and thus the target of a previous successful deal can be acquired again when the acquirer obtains the remaining shares that it does not own already (representing 80 percent of repeat targets). Second, the target of an incomplete deal may become a target again in subsequent deals (representing 20 percent of repeat targets). 2. We identify inspections by inspection reports. By construction, the total number of inspections is equal to the total number of inspection reports for each country.

Y.K im

et al./

Journal of

A ccounting

and Econom

ics 70

(2020) 101318

9

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131810

that are cross-listed in the U.S. as they are SEC registrants whose audits are subject to PCAOB inspection. To ensure that what we document is not due to PCAOB oversight changing the auditor market share (Aobdia and Shroff, 2017) and firms switching auditors, we remove firms that switch auditors in our sample. We also remove observations in financial industries and remove observations with missing data on the control variables in Eq. (1). These procedures result in a sample of 70,503 firm- year observations from 35 countries, corresponding to 12,204 unique firms and 349 PCAOB inspection reports for 157 non- U.S. auditors. All continuous variables arewinsorized at the 1% and 99% levels. Panel A of Table 1 presents the sample selection procedure for the panel data analyses that are used to test H1a and H1b. We also use a matched sample of firms with and without inspection reports for additional analyses.

Panel B of Table 1 presents the sample screening procedure for the deal-level analyses that are used to test H2a-H3b. For the tests of deal outcomes (the likelihood of deal completion and announcement returns), we use deal-level data. Since a target typically appears only once in M&A deals, we cannot use panel data for these tests. For the tests of deal outcomes, we compare deals with inspected auditors and those with uninspected auditors to examine the effect of PCAOB inspections. Independent from the procedure in Panel A of Table 1, we begin with all deals with public acquirers and public targets in the SDC from 2002 to 2016. Again, we only consider deals with a deal value of more than one million U.S. dollars or more than 1% of the acquirer's market value prior to the deal announcement date. We drop deals involving U.S. targets and those that cannot be matched with Compustat Global and Capital IQ, deals involving targets from a country without any PCAOB- inspected auditors, deals with non-U.S. targets being SEC registrants, deals with targets that had ever switched auditors between 2002 and 2016, and deals with targets of financial industries. We further drop observations with missing control variables, which leads to a sample of 1285 deals from 35 countries for the test of the likelihood of deal completion and a sample of 1080 deals from 35 countries for the test of announcement returns. All continuous variables in the deal-level analyses are winsorized at the 1% and 99% levels.

Panel C of Table 1 presents the M&A activities and outcomes, by country, for the sample used in the analysis of the likelihood of receiving a takeover bid (i.e., 70,503 firm-year observations) and the sample used in the analysis of the likelihood of deal completion (i.e., 1285 deals). This panel shows that while a large number of firm-year observations are drawn from India, Japan, Taiwan, and the U.K., Canada, Japan, and Australia are the countries with the largest number of deals, altogether accounting for around 57.82% of the total deals in our sample period.17 Among the 349 PCAOB inspections involving 157 auditors in our final sample, 207 PCAOB reports contain various types of engagement deficiencies. The statistics also reveal that 97 reports do not contain any auditor-level quality control deficiencies (Type A); 241 reports show auditor-level quality control deficiencies that had been remediated (Type B), and 11 reports publicly disclose auditor-level quality control de- ficiencies in Part II of the inspection reports (Type C). Similar breakdowns of different types of Part II findings are reported in prior studies (Bishop et al., 2013; Shroff, 2019).18

We also provide our sample distribution across years in Panel D. The number of deals is largest in 2006e2008. On average, the probability of becoming a target ranges from 0.7% to 2.9%, while the completion rate is lowest (66.2%) in 2015 and highest (100%) in 2002. The average combined announcement returns are positive in all years. The last few columns of Panel D also present the number of inspections, inspection reports with engagement deficiency and with Type A, B, and C Part II findings by year.

4. Empirical tests

4.1. Likelihood of receiving a bid

As we focus on the spillover effect of the PCAOB international inspection program, we remove the foreign firms cross- listed in U.S. markets (SEC registrants) from the (prospective) target sample. We further differentiate two treatment effects concerning PCAOB inspections: the inspection treatment, which occurs after the firm's auditor undergoes the PCAOB in- spection fieldwork, and the report disclosure treatment, which occurs after the inspection report becomes publicly available on the PCAOB website.

To test H1a and H1b, we examine the likelihood of a firm receiving an M&A bid after its auditor has been inspected by the PCAOB and after the inspection report is publicly disclosed. This analysis is conducted on a full panel of firm years with auditors that are not subject to PCAOB inspections (because they do not have SEC registrants as clients), as well as those with auditors who eventually undergo PCAOB inspection at some point in our sample period (because they had SEC registrants as clients). To account for the differences between these two groups of firms, we construct an indicator, INSPECT, which equals one for firms with auditors that undergo PCAOB inspections and zero for firms with auditors that do not receive PCAOB inspections during our sample period. To test the two treatment effects concerning PCAOB inspections, we define

17 In a sensitivity test, we remove each country one at a time, and the tenor of our results (untabulated) remains unchanged, suggesting that our findings are unlikely to be driven by any one country. 18 Bishop et al. (2013), for example, report the percentages of Type A reports and Type B þ C reports for 175 first-time inspections between 2006 and February 2011 as 32% and 68%, respectively. In our sample of 349 reports between 2006 and 2015 for both first-time and subsequent inspections, the percentages are 28% and 72%, respectively. Bishop et al. (2013) do not further break down Type B and C reports. Shroff (2019) finds there are 7 Type C reports in his sample period from 2002 to 2014.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 11

POSTINSPECT as an indicator variable that turns on (i.e., its value becomes one) between the completion of PCAOB inspection fieldwork and the public disclosure of the inspection report. If there are multiple inspections concerning an auditor, then this variable turns on multiple times during the sample period. We define POSTREPORT as an indicator variable that turns on after the public disclosure of the inspection report. If there are multiple inspections concerning an auditor, this variable turns on after the disclosure of the auditors’ first PCAOB inspection report. Because all inspection reports are publicly available and accessible on the PCAOB website, starting from the first report, the effect of the inspection report begins after the first in- spection report is publicly available. Our research design mirrors that in Shroff (2019). Fig. 1 illustrates the timeline of the two treatment indicators, POSTINSPECT and POSTREPORT.

To test H1a and H1b, we estimate the following logistic regression model:

LnðProbðTARGET ¼ 1Þ=ð1� ProbðTARGET ¼ 1ÞÞÞit ¼ a0 þ a1INSPECTit þ a2POSTINSPECTit þ a3POSTREPORTit þ a4CARit�1 þa5GRi;t�3 to t�1 þ a6SGROWTHi;t�3 to t�1 þ a7LIQUIDITYi;t�3 to t�1 þa8LEVERAGEi;t�3 to t�1 þ a9INDBIDSit�1 þ a10ASSETSIZEit�1 þ a11MBit�1 þa12PTEit�1 þ a13ROAit�1 þ a14INTANGit�1 þ a15R&Dit�1 þ a16INVESTit�1 þCountry * Year fixed effectsþ Industry fixed effects þε:

(1)

where i and t represent the firm (prospective target) and year, respectively. The dependent variable, TARGET, is an indicator variable that equals one if a firm receives at least one bid in a given year, and zero otherwise.19 POSTINSPECT equals one for the fiscal years beginning after the completion of a company's auditor's PCAOB inspection and ending before the disclosure of the inspection report, and zero otherwise. POSTREPORT equals one for the fiscal years beginning after the disclosure of a com- pany's auditor's first inspection report, and zero otherwise. INSPECT captures the time-invariant difference between client firms with inspected and non-inspected auditors. Since the time-invariant difference is controlled for, the coefficient on POSTINSPECT (POSTREPORT) can be interpreted as the incremental effect of the treatment of receiving an inspection (a report). A positive marginal effect of POSTINSPECT (POSTREPORT) would suggest that the likelihood of receiving an M&A bid is higher after the PCAOB inspection (disclosure of inspection report).

Following prior studies (e.g., Palepu, 1986; Dhaliwal et al., 2016), we include several firm-level control variables that are potentially related to the probability of a firm receiving a bid. First, we include a proxy of managerial effectiveness, measured by the one-year cumulative abnormal return (CAR), as Dhaliwal et al. (2016) predict that managerial ineffectiveness increases the probability of receiving a bid. Second, prior studies predict that firms with an imbalance between growth opportunities and financial resources, as well as illiquid and leveraged firms, are more likely to be targets (e.g., Palepu, 1986; Ambrose and Megginson, 1992). Hence, we include a growth-resource “mismatch” indicator (GR), sales growth (SGROWTH), and two proxies of financial constraints (LIQUIDITY and LEVERAGE) as control variables. Third, we include an indicator (INDBIDS) that equals one if there was a bid in an industry in the prior year, and zero otherwise, to control for time-clustered acquisition activities (see Andrade et al., 2001). Fourth, we control for the effect of firm size, measured by the natural logarithm of a firm's book value of total assets (ASSETSIZE); the effect of firm value, measured by a firm's market-to-book ratio (MB) and price-to- earnings ratio (PTE); and profitability, measured by a firm's return on assets (ROA). We also control for the effect of invest- ment, measured by the intensity of intangible assets (INTANG), R&D expenditure (R&D), and investment intensity (INVEST). Finally, we include industry-fixed effects to control for heterogeneity across industries, and country x year-fixed effects to control for time-varying country-specific factors, such as the adoption of International Financial Reporting Standards (IFRS) and other country-level changes in regulations and enforcements. In all tests, test statistics are computed with robust standard errors clustered at the auditor level.20 See the Appendix for detailed definitions of the variables.

Table 2 Panel A presents descriptive statistics for observations with inspected auditors (INSPECT ¼ 1) and uninspected auditors (INSPECT ¼ 0), and the difference between these two groups. As shown in Panel A, non-U.S. firms with inspected auditors are significantly more likely to become targets than those with uninspected auditors (the difference ¼ 0.6%). We also observe that firms with auditors that are inspected by the PCAOB are larger (ASSETSIZE) and priced higher (PTE and MB), have more intangible assets (INTANG), invest more in R&D (R&D), and have less in capital expenditure (INVEST). The completion rate and deal announcement returns are larger (albeit insignificantly) for deals that involve targets with au- ditors for which PCAOB inspection reports are available (REPORTAUDITOR ¼ 1). Non-U.S. firms with inspected auditors are also more likely to reside in countries with higher levels of rule of law (RULE). We also observe that deals involving targets with auditors for which PCAOB inspection reports are available (REPORTAUDITOR ¼ 1) are less likely to experience goodwill write-offs.

We report the descriptive statistics for key variables for the observationswith inspected auditors and uninspected auditors and the difference between these two groups in ourmatched sample in Panel B of Table 2.We discuss ourmatching procedure

19 In a sensitivity test, if the acquisition is not completed in the year of the deal announcement, we also code the year subsequent to the announcement as TARGET ¼ 1, and inferences from the results (untabulated) remain the same. For example, the marginal effect of POSTREPORT that corresponds to the model of Table 3 Panel A is 0.005, with a z-statistic of 2.90. 20 Alternatively, we cluster standard errors at the country, industry, or target-firm level, and the results (untabulated) remain unchanged. That is, we find an insignificant marginal effect of POSTINSPECT and a significant marginal effect of POSTREPORT.

Fig. 1. Timeline for PCAOB inspection and report. This figure illustrates the timeline of two treatment indicators, POSTINSPECT and POSTREPORT. POSTINSPECT turns on (i.e., its value becomes one) between the completion of PCAOB inspection field work and the public disclosure of the inspection report. If there are multiple inspections concerning an auditor, then POSTINSPECT turns on multiple times during the sample period. As shown in the figure, POSTINSPECT ¼ 0 in the Pre-inspection period I and the Pre-inspection period II, and POSTINSPECT ¼ 1 in the Post-inspection period I and the Post-inspection period II. POSTREPORT turns on after the public disclosure of the first inspection report. If there are multiple inspections concerning the auditor, POSTREPORT turns on after the disclosure of the auditor's first inspection report and stays on. As shown in the figure, POSTREPORT ¼ 0 before the first report year (Pre-report period) and POSTREPORT ¼ 1 for all the years after the first report year (Post-report period).

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131812

later in this section. The result shows that firms with inspected auditors are significantly more likely to become targets (the difference¼ 0.6%). The difference in thematched sample is largely the same as that in our full sample. We omit the discussion of the remaining descriptive statistics for the sake of brevity.

Table 3 presents the results of testing H1a and H1b. The marginal effect from estimating Eq. (1) based on the full sample is presented in Column (1) of Panel A. We find that the effect of POSTINSPECT is insignificant, suggesting that inspections alone do not alter acquirers' perceptions about the financial reporting quality of inspected auditors' clients. The marginal effect of POSTREPORT, however, is positive (0.006) and significant (z-value of 3.03), suggesting that the probability of receiving anM&A bid is higher for clients with PCAOB-inspected auditors than that for clients with uninspected auditors, after the release of PCAOB inspection reports. The standard deviation of the probability of receiving a bid is 0.113 in our sample. Thus, an increase in the likelihood of receiving a bid due to the PCAOB inspection report disclosure is equivalent to 5.3% (0.006/0.113) of the standard deviation of the likelihood of receiving a bid. Considering that one standard deviation (i.e., 0.059) increase in the investment intensity (INVEST) increases the probability of a firm receiving a bid by 0.0017 (0.059*0.028), representing an increase of 1.5% (0.0017/0.113) of the standard deviation of receiving a bid, the incremental effect of PCAOB inspection report on the likelihood of receiving a bid is economically significant. The results suggest that the positive effect of PCAOB oversight on acquirers' perceptions about the potential target's accounting quality appears after the public disclosure of the inspection report, rather than immediately after the inspection. Regarding the control variables, we find that firms with larger size, higher profitability, higher intangible intensity, and higher investment intensity are more likely to receive an M&A bid. The results are robust to the linear probability model. For example, the coefficient on POSTREPORT is 0.005 in the linear probability model, which is close to the marginal effect of 0.006 reported in Column (1) of Table 3, Panel A.

In ourmain empirical specification to test the likelihood of receiving a bid (i.e., Eq. (1)), the indicator variable for years after the public disclosure of the PCAOB inspection report (POSTREPORT) can be defined for firms with inspected auditors (INSPECT ¼ 1) only, because the PCAOB inspection event does not exist for firms with uninspected auditors (INSPECT ¼ 0). Thus, in this specification, we are essentially comparing the post-inspection report period of firms with inspected auditors to the pre-inspection report period of these firms, while accounting for the time-invariant difference between firms with inspected auditors and firms with uninspected auditors. As an alternative, we employ a difference-in-differences (DiD) research design, using a matched sample of firms with inspected auditors and those with uninspected auditors. Specifically, we match the treatment group of firms with inspected auditors and the control group of firms with uninspected auditors based on country, industry, year (the year of inspection report disclosure), and size (closest based on total assets).21 This procedure yields 16,718 observations from 1085 matched pairs of treatment and control firms for [-3, þ3] years around the year of the report disclosure. We assign the actual inspection report release year of the treatment firm to the matched control firm as its pseudo-inspection report release year, even though its auditor has never been inspected. We then re-specify Eq. (1) using the traditional DiD approach, as follows:

LnðProbðTARGET ¼ 1Þ=ð1� ProbðTARGET ¼ 1ÞÞÞit ¼ a0 þ a1TREATit þ a2POSTit þ a3TREATit * POSTit þ Controlsit þ Country * Year fixed effectsþ Industry fixed effectsþ ε:

(2)

21 The differences in target sizes of the sample of treatment observations and the sample of control observations after the matching drop from 0.727 to 0. 192, suggesting that the matching on firm size is successful.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 13

where i and t represent the firm (prospective target) and year, respectively. POST is an indicator variable that equals one if year t falls in the post-inspection report release period (either the actual post-inspection report release years for deals involving targets with PCAOB-inspected auditors or the pseudo post-inspection report release years for the matched control firms), and zero otherwise.22 The coefficient on the interaction of TREAT and POST, a3, captures the incremental changes in the dependent variables from the pre- to post-inspection period for the treatment targets, compared to the changes in the dependent variables for the matched control targets over the same time period. As shown in Column (2) of Panel A in Table 3, the marginal effect of TREAT*POST is significantly positive (0.008, z-value of 3.45), suggesting that the likelihood of receiving a bid increases by 0.8% after the disclosure of PCAOB inspection reports. The Ai and Norton (2003) interaction effect of TREAT*POST is also significantly positive (0.007, z-value of 2.20), further supporting the nontrivial and statistically significant effect of having a PCAOB inspection report on the likelihood of receiving a bid. The results are robust to the linear probability model. For example, the coefficient on TREAT*POST is 0.007 in the linear probability model, which is close to the marginal effect of 0.008 reported in Column (2) of Table 3, Panel A.

To mitigate the reverse causality concern and further examine when the disclosure effect takes place, we examine the dynamic effect of the disclosure of PCAOB inspection report on the probability of receiving an M&A bid. We replace POST- REPORT in Eq. (1) with four indicator variables: YEARt-2REPORT equals one for clients of inspected auditors in the second year prior to the first inspection report year; YEARt-1REPORT equals one for clients of inspected auditors in the year prior to the first inspection report year; YEARtREPORT equals one for clients of inspected auditors in the first inspection report year; and YEARtþnREPORT equals one for clients of inspected auditors in all years after the first inspection report year. We report the results in Panel B of Table 3.

As shown in the table, the marginal effects of YEARt-2REPORT and YEARt-1REPORT are insignificant, whereas the marginal effects on YEARtREPORT and YEARtþnREPORT are significantly positive (0.004 with a z-value of 2.09 and 0.004 with a z-value of 1.77, respectively). Taken together, the results based on this dynamic model provide additional evidence that the likelihood of receiving a bid increases after the release of PCAOB inspection reports. More importantly, the results confirm that, before the inspection report, there is no discernible difference between target auditors with and without PCAOB inspection reports.

To better illustrate the differences based on thematched sample, we present in a graph the likelihood of receiving a bid for the treatment and control groups separately over time (Panel A of Fig. 2) and the differences in the likelihood between the treatment and control groups over time (Panel B of Fig. 2). We also plot the expected likelihood of receiving a bid by esti- mating Eq (1) with the control variables (Panel C of Fig. 2).23 All these graphs show little difference in the likelihood of receiving a bid between the treatment and matched control firms prior to the inspection report year, lending support to the parallel trend assumption. Yet there is a substantial increase in the likelihood of receiving a bid for the treatment firms compared to the matched control firms after the inspection report year.

4.2. The likelihood of deal completion

To test H2a and H2b, we examine whether the likelihood of deal completion increases after the target's auditor has been inspected by the PCAOB and after the disclosure of the inspection report, respectively. To capture the inspection treatment, we turn the indicator variable INSPECTAUDITOR on if the target's auditor has been inspected but the report is not disclosed before the deal announcement year. To capture the report treatment, we turn the indicator variable REPORTAUDITOR on if the target's auditor has been inspected and the inspection report has been disclosed publicly by the PCAOB before the deal announcement year. We estimate the logistic regression model as follows:

LnðProbðCOMPLETION ¼ 1Þ=ð1� ProbðCOMPLETION ¼ 1ÞÞÞit ¼ a0 þ a1INSPECTAUDITORit þ a2REPORTAUDITORit þ a3ACQINSPECTit þa4ACQREPORTit þ a5COMMONAUDITORit þ a6VALUEit þ a7SAMESICit þa8TOEHOLDit þ a9TENDERit þ a10HOSTILEit þ a11TERMIFEEit a12OFBIDDERSit þ a13CASHONLYit þ Acquirer Country * Target Country * Year fixed effectsþ Acquirer Industry fixed effects þTarget Industry fixed effects þε:

(3)

We estimate this model using deal-level data. i and t represent the deal and year, respectively. The dependent variable, COMPLETION, is an indicator variable that equals one if the deal is completed, and zero otherwise. A positivemarginal effect of

22 POST is not subsumed by country*year fixed effects in this regression, because it is defined differently for each auditor in the same country. 23 In Panel C of Fig. 2, we set the expected likelihood of receiving a bid to zero without confidence intervals for the matched control firms in year t-3, because no firm in our matched control sample receives a takeover bid in year t-3 and the expected likelihood of receiving a bid cannot be calculated for this group of firms in year t-3.

Table 2 Descriptive statistics. The table reports descriptive statistics of variables used in firm-level and deal-level analyses. See Appendix for variable definitions.

Column (1) (2) (3) (4) (5) (6) (7) (8)

Variable Mean Median Std. Dev. Mean Median Std. Dev. Mean t-statistic

Panel A: Descriptive statistics for main analyses

Test of the likelihood of receiving a bid

INSPECT ¼ 1 (N ¼ 38,222) INSPECT ¼ 0 (N ¼ 32,281) Test for difference

TARGET 0.017 0.000 0.124 0.011 0.000 0.099 0.006*** 6.65 CAR �0.035 �0.036 0.537 �0.029 �0.025 0.507 �0.006*** �16.00 GR 0.387 0.000 0.487 0.400 0.000 0.490 �0.013*** �3.68 SGROWTH 0.140 0.067 0.284 0.165 0.091 0.300 �0.025*** �11.12 LIQUIDITY �0.133 �0.139 0.224 �0.184 �0.194 0.236 0.051*** 29.55 LEVERAGE 1.401 0.914 1.653 1.530 0.974 1.799 �0.129*** �9.86 INDBIDS 0.949 1.000 0.219 0.930 1.000 0.254 0.019*** 10.59 ASSETSIZE 5.524 5.365 1.802 4.797 4.668 1.708 0.727*** 51.66 MB 1.476 1.150 0.927 1.401 1.053 0.940 0.075*** 10.67 PTE 9.583 8.225 19.261 8.469 6.454 19.288 1.113*** 7.64 ROA 0.038 0.054 0.120 0.037 0.051 0.116 0.001 0.80 INTANG 0.404 0.032 0.832 0.321 0.011 0.775 0.083*** 13.52 R&D 0.027 0.000 0.064 0.011 0.000 0.042 0.016*** 38.06 INVEST 0.049 0.031 0.056 0.051 0.030 0.062 �0.002*** �4.90

Test of the likelihood of deal completion

REPORTAUDITOR ¼ 1 (N ¼ 413) REPORTAUDITOR ¼ 0 (N ¼ 872) Test for difference

COMPLETION 0.829 1.000 0.377 0.823 1.000 0.382 0.006 0.25 ACQINSPECT 0.484 0.000 0.500 0.233 0.000 0.423 0.251*** 8.58 ACQREPORT 0.439 0.000 0.497 0.115 0.000 0.319 0.324*** 11.75 COMMONAUDITOR 0.208 0.000 0.406 0.283 0.000 0.451 �0.075*** �2.80 VALUE 4.288 3.992 1.057 4.334 4.011 1.097 �0.046 �0.69 SAMESIC 0.600 1.000 0.491 0.577 1.000 0.494 0.023 0.77 TOEHOLD 0.358 0.000 0.480 0.351 0.000 0.478 0.007 0.22 TENDER 0.187 0.000 0.390 0.401 0.000 0.490 �0.214*** �7.57 HOSTILE 0.013 0.000 0.114 0.013 0.000 0.114 0.000 0.01 TERMIFEE 0.639 0.000 1.657 0.520 0.000 1.467 0.119 1.28 OFBIDDERS 1.047 1.000 0.225 1.087 1.000 0.323 �0.040** �2.20 CASHONLY 0.353 0.000 0.478 0.421 0.000 0.494 �0.068** �2.28

Test of announcement returns

REPORTAUDITOR ¼ 1 (N ¼ 347) REPORTAUDITOR ¼ 0 (N ¼ 733) Test for difference

COMBINECAR 0.025 0.032 0.070 0.018 0.032 0.083 0.007 1.35 ACQASSETSIZE 7.691 8.120 1.274 7.817 8.120 1.233 �0.126 �1.53 ACQMB 1.186 0.753 1.031 1.203 0.920 0.923 �0.016 �0.26 TARASSETSIZE 5.477 5.451 1.743 5.422 5.437 1.691 0.055 0.50 TARMB 1.281 1.071 0.867 1.474 1.176 0.980 �0.193*** �3.14

Cross-sectional and other firm-year level analyses

INSPECT ¼ 1 (N ¼ 38,222) INSPECT ¼ 0 (N ¼ 32,281) Test for difference

RULE 1.347 1.519 0.570 0.847 0.982 0.810 0.500*** 93.05

Cross-sectional and other deal-level analyses

REPORTAUDITOR ¼ 1 (N ¼ 413) REPORTAUDITOR ¼ 0 (N ¼ 872) Test for difference

CROSSBORDER 0.383 0.000 0.487 0.396 0.000 0.489 �0.013 �0.45 GWWRITEOFF 0.140 0.000 0.347 0.197 0.000 0.398 �0.057*** �2.61 DIVESTURE 0.087 0.000 0.282 0.099 0.000 0.299 �0.012 �0.72

Panel B: Descriptive statistics for matched sample analyses

Test of the likelihood of receiving a bid

TREAT ¼ 1 (N ¼ 8432) TREAT ¼ 0 (N ¼ 8286) Test for difference

TARGET 0.009 0.000 0.090 0.003 0.000 0.051 0.006*** 4.86 CAR 0.004 0.010 0.427 �0.002 �0.013 0.408 0.006 0.90 GR 0.434 0.000 0.496 0.390 0.000 0.488 0.044*** 5.78 SGROWTH 0.085 0.036 0.229 0.096 0.033 0.251 �0.011*** �2.95 LIQUIDITY �0.124 �0.136 0.240 �0.157 �0.176 0.213 0.033*** 9.61 LEVERAGE 1.187 0.733 1.430 1.356 0.969 1.375 �0.169* �1.78 INDBIDS 0.970 1.000 0.170 0.968 1.000 0.176 0.002 0.83 ASSETSIZE 5.379 5.257 1.343 5.187 5.160 1.619 0.192*** 5.68 MB 1.201 0.932 0.779 1.092 0.905 0.663 0.109*** 9.78 PTE 8.543 6.992 18.201 9.275 7.508 16.886 �0.732*** �2.70 ROA 0.034 0.043 0.105 0.040 0.041 0.086 �0.006*** �4.23 INTANG 0.284 0.016 0.725 0.310 0.017 0.790 �0.026** �2.24 R&D 0.029 0.003 0.065 0.016 0.002 0.042 0.013*** 16.00 INVEST 0.038 0.022 0.048 0.038 0.023 0.046 0.000 0.55

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131814

Table 3 PCAOB inspections and the likelihood of receiving a bid. This table reports the results of estimating the effect of PCAOB inspections on the likelihood of receiving a bid. Column (1) of Panel A reports the results of estimating logistic model using the full sample. Marginal effects are reported. Column (2) of Panel A reports the results of difference-in-differences (DiD) analysis using thematched sample. For each firm audited by an inspected auditor, based on the year of public disclosure of the inspection report, we find a matched control sample from those firms audited by uninspected auditors in the same year, country, industry, and with similar firm size. We employ a seven-year window [-3, þ3] around the report year for the DiD analysis. We report both the regular logit marginal effect and the Ai and Norton (2003) interaction effect estimate for the interaction term of TREAT*POST. Panel B reports the dynamic model based on the report year as the event date using the full sample. See Appendix for variable definitions. Z-statistics are computed based on robust standard errors clustered at the firm's auditor level.*,**, and*** represent two-tailed significance at the 10%, 5%, and 1% level, respectively. N denotes the number of firm- year observations.

Column (1) (2)

Dependent Variable ¼ Prob. (TARGET) Prob. (TARGET)

Marginal effect z-statistic Marginal effect z-statistic

Panel A: Full sample analysis and difference-in-differences analysis using a matched sample INSPECT �0.002 �0.95 POSTINSPECT �0.000 �0.30 POSTREPORT 0.006*** 3.03 TREAT 0.002 0.78 POST 0.000 0.13 TREAT*POST 0.008*** 3.45 CAR �0.001 �0.64 0.005*** 3.23 GR 0.001 1.23 �0.001 �0.75 SGROWTH 0.001 0.88 �0.003 �1.27 LIQUIDITY �0.002 �0.77 �0.002 �0.58 LEVERAGE 0.001 0.97 �0.001 �1.56 INDBIDS �0.001 �0.58 �0.000 �0.13 ASSETSIZE 0.001* 1.71 �0.003*** �4.37 MB �0.001 �0.83 �0.003** �2.11 PTE �0.000 �0.29 0.000** 1.98 ROA 0.005* 1.69 �0.007 �0.99 INTANG 0.001** 2.07 0.001 1.12 R&D 0.012 1.39 0.024 1.45 INVEST 0.028*** 3.42 0.059*** 4.76 Country*Year fixed effects Yes Yes Industry fixed effects Yes Yes N 70,503 16,718 Pseudo R-squared 0.118 0.115 Ai and Norton (2003) Marginal Effect for TREAT*POST 0.007** 2.20

Dependent Variable ¼ Prob. (TARGET)

Marginal effect z-statistic

Panel B: Dynamic effects analysis using the full sample INSPECT �0.002 �0.62 POSTINSPECT 0.001 0.57 YEARt-2REPORT ¡0.001 ¡0.68 YEARt-1REPORT ¡0.002 ¡1.13 YEARtREPORT 0.004** 2.09 YEARtþnREPORT 0.004* 1.77 CAR �0.001 �0.68 GR 0.001 0.91 SGROWTH 0.001 0.57 LIQUIDITY �0.002 �0.99 LEVERAGE 0.000 0.62 INDBIDS �0.002 �0.47 ASSETSIZE 0.001* 1.91 MB �0.000 �0.53 PTE �0.000 �0.34 ROA 0.005 1.12 INTANG 0.002** 2.32 R&D 0.012 1.15 INVEST 0.028*** 2.85 Country*Year fixed effects Yes Industry fixed effects Yes N 70,503 Pseudo R-squared 0.118

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 15

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131816

INSPECTAUDITOR (REPORTAUDITOR) suggests that the likelihood of deal completion is higher for deals involving targets with auditors that underwent PCAOB inspections (for which inspection reports are publicly available) than for deals involving other targets.

Shroff (2019) finds that PCAOB oversight can reduce agency conflicts and financial constraints, and thus it allows acquirers to have cheaper financing to engage in more and better M&A deals. We therefore control for the effect of acquirer auditors’ PCAOB inspections by including two indicator variables ACQINSPECT and ACQREPORT, defined similarly as INSPECTAUDITOR and REPORTAUDITOR. We also control for the common auditor effect by including an indicator for the target and the acquirer sharing the same auditor (COMMONAUDITOR), as Cai et al. (2016) find that common auditors have a significant effect on deal outcomes. We also add a vector of deal-level control variables shown to be related to M&A outcomes, following prior studies: the value of transaction (VALUE) (Dhaliwal et al., 2016); an indicator for same industry deals (SAMESIC) (Andrade and Stafford, 2004); three indicator variables for toeholds (TOEHOLD), tender offer (TENDER), and hostile deals (HOSTILE) (Schwert, 2002; Betton et al., 2009); termination fees (TERMIFEE) (Bates and Lemmon, 2003; Officer, 2003); bid competition (OFBIDDERS) (Bradley et al., 1988); and an indicator variable of cash-only payment (CASHONLY) (Eckbo and Langohr, 1989). We also control for acquirer country x target country x year, acquirer industry, and target industry fixed effects. The acquirer country x target country x year fixed effects control for changes in bilateral investment flows across countries over time.

We present the results in Table 4. As shown, the marginal effect of INSPECTAUDITOR is insignificant, while the marginal effect of REPORTAUDITOR is statistically significant (0.040, with z-value of 2.07). Given that the unconditional probability of deal completion is 82.5% for the full sample, this marginal effect represents a 4.8% (4.0%/82.5%) increase in the likelihood of deal completion. In comparison, a tender offer, TENDER, increases the likelihood of deal completion by 5.7%, which represents a 6.9% (5.7%/82.5%) increase from the sample mean. The results reported in Table 4 support (reject) H2b (H2a). That is, the likelihood of deal completion increases after the release of the PCAOB inspection report concerning the target's auditor. The effects of the control variables on the likelihood of deal completion are generally consistent with those in prior studies (e.g., Dhaliwal et al., 2016). In particular, we find that the likelihood of completion is higher for deals with higher levels of toehold, tender offers, higher termination fees, and lower for deals with hostile offers, higher bid competition, and cash only pay- ments. The results are robust to the linear probability model. For example, the coefficient on REPORTAUDITOR is 0.043 in the linear probability model, which is close to the marginal effect of 0.040 reported in Table 4.

4.3. Deal announcement returns

Finally, to test H3a and H3b, we estimate the following OLS regression:

COMBINECARit ¼ a0 þ a1INSPECTAUDITORit þ a2REPORTAUDITORit þ a3ACQINSPECTit þa4ACQREPORT þ a5COMMONAUDITORit þ a6ACQASSETSIZEit�1 þa7ACQMBit�1 þ a8TARASSETSIZEit�1 þ a9TARMBit�1 þ a10VALUEit þa11SAMESICit þ a12TENDERit þ a13TERMIFEEit þ a14OFBIDDERSit þa15CASHONLYit þ Acquirer Country * Target Country * Year fixed effectsþ Acquirer Industry fixed effects þTarget Industry fixed effectsþ ε:

(4)

where i and t represent the deal and year, respectively. COMBINECAR is the cumulative abnormal returns in the three-day window around the deal announcement, averaging the cumulative abnormal returns of the target and the acquirer.24 The test variables are the same as those defined in Eq. (3). Following Dhaliwal et al. (2016), we include a set of control variables that are likely to affect announcement returns. In addition to deal characteristics, we control for the acquirer size andmarket- to-book ratio (ACQASSETSIZE and ACQMB), as well as the target size and market-to-book ratio (TARASSETSIZE and TARMB) in the year prior to the deal announcement. Acquirer country x target country x year, acquirer industry, and target industry fixed effects are also included.

Table 5 presents the results of estimating Eq. (4). As shown in the table, the coefficient of INSPECTAUDITOR is insignificant, while the coefficient on REPORTAUDITOR is significantly positive (coefficient ¼ 0.014 with a t-value of 2.17). The magnitude of the coefficient represents a 70% (1.4%/2.0%) increase in the deal announcement returns after the disclosure of PCAOB in- spection reports, given that the average combined announcement return is 2.0%. The economic significance is comparable to that of another determinant of deal announcement returns documented in prior studies (e.g., Cai et al., 2016; Dhaliwal et al., 2016), i.e., tender offer (TENDER). In particular, the coefficient on TENDER represents a 60% (1.2%/2.0%) higher announcement returns for tender offers. The results support (reject) H3b (H3a) and suggest that PCAOB oversight helps reduce information

24 We also test the portfolio returns weighted by the value of the acquirer and the target, and the results (untabulated) are similar. For example, the coefficient of REPORTAUDITOR that corresponds to the model of Table 5 is 0.011, with a t-value of 2.02.

Fig. 2. Differences in the likelihood of receiving a bid for the treatment group and control group across various time periods for the matched sample. Panel A of this figure shows the time trends in annual averages of the likelihood of receiving a bid for [-3, 3] years for the treatment sample (in solid line) and the matched control sample (in dashed line), respectively. The x-axis represents time relative an auditor's PCAOB report year (year t) and the y-axis represents the likelihood of receiving a bid. Panel B of this figure shows the differences in annual averages of the likelihood of receiving a bid for [-3, 3] years between the treatment sample and the matched control sample. The x-axis represents time relative an auditor's PCAOB report year (year t) and the y-axis represents the differences in the likelihood of receiving a bid. Panel C of this figure shows the expected likelihood of receiving a bid with the control variables as well as 95% confidence interval for [-3, 3] years for the treatment sample (in solid line) and the matched control sample (in dashed line), respectively.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 17

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131818

uncertainty in M&A deals and allows acquirers to identify better investment opportunities, thereby improving the deal quality.

4.4. Cross-sectional analyses

In this section, we examine whether the effect of PCAOB inspection varies across country-level legal institutions and different levels of information uncertainty.25 Specifically, we re-estimate Eqs. (1), (3) and (4) after partitioning the full sample into two subsamples based on country-level legal institutions and deal-level information uncertainty. Table 6 reports the results.We present themarginal effects/coefficients of POSTREPORT and REPORTAUDITOR for each subsample and compare the differences in the effects of PCAOB inspection report disclosures between the subsamples.

The value of PCAOB international inspections would be greater when home-country legal institutions are weaker. Following prior studies (e.g. Lamoreaux, 2016; Aobdia and Shroff, 2017), we obtain a measure of country-level legal in- stitutions (RULE) from Kaufmann et al. (2009). Based on the median value of RULE, we classify (prospective) targets into two subsamples: those from countries with strong and weak legal institutions, respectively. Results are reported in Panel A of Table 6. As in Tables 3e5, the results show that PCAOB inspections alone have no significant effect on M&A outcomes. The results also show that PCAOB inspection reports have significantly positive effects on M&A outcomes in countries with weaker rule of law, but their effect is insignificant in countries with stronger rule of law. The difference in the marginal effects of POSTREPORT between the two subsamples is significant at the 5% level for the likelihood of receiving a bid.26 The difference of themarginal effects or coefficients on REPORTAUDITOR in the two subsamples is significant for the deal completion rate, but insignificant for the deal announcement returns. These results indicate that PCAOB oversight generates greater real economic benefits for targets located in countries with weak legal institutions, where the institutional infrastructure is ineffective in forcing auditors to provide high quality assurance services (Aobdia and Shroff, 2017).

We next examine whether the benefit of PCAOB oversight differs across deals with different levels of information un- certainty. We employ two proxies to capture deal-level information uncertainty: cross-border vs. within-country M&As and diversifying vs. non-diversifying M&As. Prior studies argue that the cost of acquiring information about foreign equity markets is a significant barrier to international capital mobility (Brennan and Cao, 1997) and that information asymmetry can result in less foreign investment (Gordon and Bovenberg, 1996). Hence, compared to within-country M&As (CROSSBORDER ¼ 0), cross-border M&As (CROSSBORDER ¼ 1) are likely to benefit more from PCAOB oversight that reduces information uncertainty. We find, as shown in Panel B of Table 6, that the effect of PCAOB inspections is more pronounced for cross-border M&As. The difference is statistically significant for both the likelihood of deal completion and announcement returns as dependent variables. Acquirers from different industries face higher information uncertainty than those from the same industry as the targets. Hence, the benefits of PCAOB inspections are likely to be higher for cross-industry M&As (SAMESIC¼ 0) than for within-industry M&As (SAMESIC¼ 1). The results shown in Table 6, Panel C indicate that the effects of PCAOB oversight are more pronounced for cross-industry M&As than for within-industry M&As, consistent with our prediction.

4.5. Content of PCAOB inspection reports

Having found that the positive effect of PCAOB oversight on M&A outcomes appears after the public disclosure of the inspection report, rather than immediately after the inspection, as shown in Tables 3e5, we also conduct analyses regarding the inspection report contents. Specifically, we examine whether the effects of PCAOB oversight on M&A activities and outcomes vary depending on the contents of inspection reports. If acquirers use PCAOB reports to assess targets’ financial reporting quality, we conjecture that acquirers perceive inspected auditors having deficiencies reflected in the PCAOB reports as being low audit quality, and therefore, deals involving inspected auditors with deficiencies have unfavorable M&A outcomes.

We focus on the inspection report findings pertaining to both engagement-level (Part I) and auditor-level quality control (Part II). To test whether Part I findings matter, for the analysis using panel data (i.e., the test of the likelihood of receiving a bid), we construct an indicator variable for engagement deficiencies, POSTREPORT_ENGDEF, which equals one if by the beginning of the fiscal year, at least one PCAOB inspection report for the auditor is available and the most recent report identifies an engagement deficiency. When we include this indicator in Eq. (1), the main effect of POSTREPORT captures the effect of receiving a PCAOB inspection report without engagement deficiencies and POSTREPORT_ENGDEF captures the in- cremental effect of receiving a report with engagement deficiencies identified in the inspection report. The sum of the effects

25 We do not conduct a cross-sectional analysis with respect to Big 4 vs. non-Big 4, because the majority (94%) of inspected auditors are Big 4 affiliates, which would make a cross-sectional analysis less meaningful. 26 Based on the matched sample, we find that, consistent with the results based on the full panel data, the effect of having an inspection report is significantly positive in countries with weaker rule of law but insignificant in countries with stronger rule of law. While the magnitude of the marginal effects of the inspection report is similar, the difference in the inspection-report effects between the two subsamples is not statistically significant in the matched sample. The matched sample is possible only for the panel data analysis, i.e., an analysis of the probability of receiving a bid.

Table 4 PCAOB inspections and the likelihood of deal completion. This table reports the results of estimating the effect of PCAOB inspections on the likelihood of deal completion. Marginal effects are reported. See Appendix for variable definitions. Z-statistics are computed based on robust standard errors clustered at the target's auditor level.*,**, and*** represent two-tailed significance at the 10%, 5%, and 1% level, respectively. N denotes the number of deals included in the analysis.

Dependent Variable ¼ Prob. (COMPLETION)

Marginal effect z-statistic

INSPECTAUDITOR �0.031 �0.93 REPORTAUDITOR 0.040** 2.07 ACQINSPECT �0.006 �0.23 ACQREPORT 0.010 0.19 COMMONAUDITOR 0.021 1.22 VALUE �0.001* �1.88 SAMESIC 0.001 0.06 TOEHOLD 0.513*** 5.05 TENDER 0.057*** 3.86 HOSTILE �0.170* �1.86 TERMIFEE 0.018* 1.79 OFBIDDERS �0.081*** �6.22 CASHONLY �0.045** �2.53 Acquirer Country*Target Country*Year fixed effects Yes Acquirer Industry fixed effects Yes Target Industry fixed effects Yes N 1285 Pseudo R-squared 0.547

Table 5 PCAOB inspections and deal announcement returns. This table reports the effect of PCAOB inspections on deal announcement returns. See Appendix for variable definitions. T-statistics are computed based on robust standard errors clustered at the target's auditor level. *, **, and *** represent two-tailed significance at the 10%, 5%, and 1% level, respectively. N denotes the number of deals included in the analysis.

Dependent Variable ¼ COMBINECAR

Coefficient t-statistic

INSPECTAUDITOR �0.002 �0.32 REPORTAUDITOR 0.014** 2.17 ACQINSPECT �0.004 �0.41 ACQREPORT �0.003 �0.35 COMMONAUDITOR 0.004 0.57 ACQASSETSIZE �0.002 �0.96 ACQMB 0.002 0.53 TARASSETSIZE 0.002 0.98 TARMB �0.003 �0.70 VALUE 0.000** 2.48 SAMESIC 0.005 0.57 TENDER 0.012* 1.77 TERMIFEE �0.002 �0.89 OFBIDDERS 0.013** 2.08 CASHONLY 0.005 0.62 Acquirer Country*Target Country*Year fixed effects Yes Acquirer Industry fixed effects Yes Target Industry fixed effects Yes N 1080 Adjusted R-squared 0.155

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 19

of POSTREPORT and POSTREPORT_ENGDEF captures the total effect of PCAOB inspection reports with engagement deficiencies. Similarly, in the deal-level analyses, we construct an indicator variable REPORTAUDITOR_ENGDEF that turns on if prior to the deal announcement year, at least one PCAOB inspection report for the target's auditor is available before the deal announcement year and the most recent report identifies engagement deficiencies, and include it in Eqs. (3) and (4). As shown in Panel A of Table 7, the main effects of POSTREPORT and REPORTAUDITOR remain positive and significant, suggesting that clean inspection reports signal high audit quality, and therefore, they improve M&A outcomes by reducing information

Table 6 Cross-sectional analyses. This table reports cross-sectional differences in the effect of PCAOB inspections by comparing the marginal effects/coefficients of POSTREPORT/REPORTAUDITOR across the various subsamples. Marginal effects are reported for logistic regression models. All variables are defined in Appendix. Z-statistics/t-statistics are computed based on robust standard errors clustered at the firm's auditor level for firm-level analyses and at the target's auditor level for deal-level analyses. The superscripts *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively, in a two-tailed test.

Column (1) (2) (3) (4) (5) (6)

Dep. Var. ¼ Prob. (TARGET) Prob. (TARGET) Prob. (COMPLETION) Prob. (COMPLETION) COMBINECAR COMBINECAR

Lower RULE Higher RULE Lower RULE Higher RULE Lower RULE Higher RULE

Mar. Eff. z-stat. Mar. Eff. z-stat. Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat. Coef. t-stat.

Panel A: Acquisitions with (prospective) target firms from countries of weaker vs. stronger rule of law (RULE) INSPECT 0.001 0.05 0.004 1.35 POSTINSPECT 0.001 0.29 0.003 1.19 POSTREPORT 0.007*** 3.81 0.001 0.32 INSPECTAUDITOR 0.009 0.59 �0.031 �0.53 0.003 0.34 �0.004 �0.40 REPORTAUDITOR 0.043*** 3.24 0.020 0.42 0.018* 1.76 0.007 0.98 ACQINSPECT �0.037* �1.81 0.077 1.32 �0.007 �0.49 0.004 0.34 ACQREPORT 0.033 1.63 �0.034 �0.47 �0.006 �0.43 �0.002 �0.12 COMMONAUDITOR �0.019 �1.10 0.032 0.72 0.004 0.57 0.001 0.30 CAR �0.003*** �2.70 0.001 1.20 GR �0.001 �0.85 0.004** 2.32 SGROWTH 0.001 0.31 0.003** 2.51 LIQUIDITY 0.000 0.08 0.003 0.98 LEVERAGE 0.001*** 3.15 �0.001* �1.65 INDBIDS 0.003 1.18 �0.004 �1.15 ASSETSIZE 0.000 0.38 0.001** 2.07 MB 0.000 0.28 �0.001 �0.12 PTE 0.000 1.01 �0.000 �1.42 ROA �0.011** �2.38 0.007 1.25 INTANG �0.000 �0.41 0.003*** 3.98 R&D 0.027** 2.27 �0.003 �0.20 INVEST 0.010 1.05 0.049*** 5.60 VALUE 0.000 1.38 �0.002*** �3.85 0.001** 2.16 0.001*** 3.33 SAMESIC 0.003 0.85 �0.005 �0.16 0.007 0.63 �0.000 �0.06 TOEHOLD 0.056*** 3.07 0.522*** 5.51 TENDER 0.070*** 4.89 0.043*** 3.21 0.007 0.88 0.012 0.91 HOSTILE �0.205* �1.97 �0.241*** �3.83 TERMIFEE 0.011* 1.95 0.009 1.12 �0.003 �0.33 �0.002 �0.81 OFBIDDERS �0.052** �2.51 �0.090*** �4.93 �0.020 �1.08 0.018 1.19 CASHONLY �0.031*** �5.05 �0.082* �1.91 0.011 1.20 �0.006 �0.67 ACQASSETSIZE 0.002 0.93 �0.004 �0.76 ACQMB 0.005 1.07 0.000 0.16 TARASSETSIZE �0.002 �0.58 0.002 0.42 TARMB �0.007 �1.24 �0.001 �0.15 Diff. in POSTREPORT 0.006** (p ¼ 0.02) Diff. in REPORTAUDITOR 0.023*** (p < 0.01) 0.011 (p ¼ 0.21) Country*Year F.E. Yes Yes Industry F.E. Yes Yes Acquirer Country*Target

Country*Year F.E. Yes Yes Yes Yes

Acquirer Industry F.E. Yes Yes Yes Yes Target Industry F.E. Yes Yes Yes Yes N 36,810 33,693 637 648 571 509 Pseudo/Adjusted R-squared 0.147 0.145 0.567 0.547 0.185 0.149

Column (1) (2) (3) (4)

Dep. Var. ¼ Prob. (COMPLETION) Prob. (COMPLETION) COMBINECAR COMBINECAR

CROSSBORDER ¼ 0 CROSSBORDER ¼ 1 CROSSBORDER ¼ 0 CROSSBORDER ¼ 1

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat. Coef. t-stat.

Panel B: Local acquisitions vs. cross-border acquisitions (CROSSBORDER) INSPECTAUDITOR �0.035 �0.78 0.013 0.52 0.001 0.12 �0.003 �0.07 REPORTAUDITOR 0.040 1.06 0.079* 1.85 0.008 0.65 0.049* 1.75 ACQINSPECT �0.001 �0.03 �0.058 �1.17 0.002 0.27 0.103 1.63 ACQREPORT 0.023 0.43 �0.088 �0.10 0.006 0.36 �0.125 �1.53 COMMONAUDITOR 0.036 1.13 0.003 0.05 0.008 1.47 0.017 0.36 VALUE �0.001*** �3.16 0.001 0.27 0.000** 2.24 �0.001 �0.15 SAMESIC �0.033 �1.39 �0.047 �0.27 0.005 0.40 0.013 0.24 TOEHOLD 0.567*** 6.73 0.165 1.11 TENDER 0.043** 2.17 0.079 0.83 0.005 0.45 0.002 0.16

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131820

Table 6 (continued )

Column (1) (2) (3) (4)

Dep. Var. ¼ Prob. (COMPLETION) Prob. (COMPLETION) COMBINECAR COMBINECAR

CROSSBORDER ¼ 0 CROSSBORDER ¼ 1 CROSSBORDER ¼ 0 CROSSBORDER ¼ 1

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat. Coef. t-stat.

HOSTILE �0.269*** �3.03 �0.173 �0.60 TERMIFEE 0.019** 2.32 �0.019 �1.43 �0.001 �0.19 �0.011** �2.09 OFBIDDERS �0.090*** �4.41 0.083 0.79 0.010 0.90 0.029 0.99 CASHONLY �0.086*** �2.78 0.015 0.95 0.003 0.43 0.079 1.55 ACQASSETSIZE 0.001 0.63 �0.037** �2.31 ACQMB 0.000 0.10 �0.016 �0.49 TARASSETSIZE 0.000 0.17 0.008 0.79 TARMB �0.003 �0.38 0.016*** 4.49 Diff. in REPORTAUDITOR ¡0.039*** (p < 0.01) ¡0.041** (p ¼ 0.02) Acquirer Country*Target Country*Year F.E. Yes Yes Yes Yes Acquirer Industry F.E. Yes Yes Yes Yes Target Industry F.E. Yes Yes Yes Yes N 782 503 741 339 Pseudo/Adjusted R-squared 0.485 0.629 0.120 0.437

Column (1) (2) (3) (4)

Dep. Var. ¼ Prob. (COMPLETION) Prob. (COMPLETION) COMBINECAR COMBINECAR

SAMESIC ¼ 0 SAMESIC ¼ 1 SAMESIC ¼ 0 SAMESIC ¼ 1

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat. Coef. t-stat.

Panel C: Within-industry vs. across-industry acquisitions (SAMESIC) INSPECTAUDITOR �0.035 �1.56 �0.013 �0.36 0.013 0.73 0.000 0.13 REPORTAUDITOR 0.052*** 3.68 ¡0.006 ¡0.69 0.026* 1.86 0.008 1.58 ACQINSPECT 0.005 0.20 0.012 0.38 0.033 1.35 �0.003 �0.28 ACQREPORT �0.037 �1.63 �0.006 �0.13 �0.016 �0.52 �0.002 �0.26 COMMONAUDITOR �0.013 �0.50 0.062* 1.72 0.011 0.82 0.001 0.13 VALUE 0.001*** 2.68 �0.001* �1.80 0.002** 2.49 0.000 1.03 TOEHOLD 0.152*** 4.26 0.515*** 4.87 TENDER 0.073*** 5.28 0.019 0.88 0.019 1.21 0.008 1.41 HOSTILE �0.003 �0.12 �0.637*** �7.72 TERMIFEE 0.010*** 4.58 0.011* 1.76 �0.004 �0.53 0.001 0.60 OFBIDDERS 0.063*** 3.87 �0.041*** �3.69 0.023 0.47 0.007* 1.83 CASHONLY 0.001 0.17 �0.045** �1.98 �0.003 �0.22 0.005 0.69 ACQASSETSIZE 0.000 0.13 0.003 1.55 ACQMB �0.006 �0.72 0.007** 2.34 TARASSETSIZE �0.003 �0.70 �0.001 �0.80 TARMB 0.002 0.22 �0.008** �2.07 Diff. in REPORTAUDITOR 0.058*** (p < 0.01) 0.018* (p ¼ 0.09) Acquirer Country*Target Country*Year F.E. Yes Yes Yes Yes Acquirer Industry F.E. Yes Yes Yes Yes Target Industry F.E. Yes Yes Yes Yes N 537 748 460 620 Pseudo/Adjusted R-squared 0.600 0.537 0.245 0.166

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 21

uncertainty in M&As. When the inspection report contains engagement deficiencies, however, the positive effect of having a PCAOB inspection report on M&A outcomes weakens, as evidenced by the significantly negative marginal effects of POST- REPORT_ENGDEF and REPORTAUDITOR_ENGDEF. We also test the total effects of receiving an inspection report with engage- ment deficiency and show the F-test results at the bottom of the panel. Although the total effect of receiving an inspection report with engagement deficiency still increases the likelihood of receiving a bid, there are no benefits in other M&A outcomes.

To examine the effect of the Part II findings, we construct and include two indicator variables, POSTREPORT_TYPEA and POSTREPORT_TYPEC, in Eq. (1), which turn onwhen by the beginning of the fiscal year, at least one PCAOB inspection report for the auditor is available and the most recent report is identified as a Type A and Type C report, respectively. As explained earlier, a Type A report is identified as soon as the inspection report is disclosed, and a Type C report is identified later when the auditor does not remediate the quality control deficiencies and the PCAOB discloses it publicly. When Type A and Type C

Table 7 Content of PCAOB inspection report. This table reports the effects of PCAOB inspection report findings on M&A outcomes. Panel A examines the presence of audit engagement deficiency in Part I findings of PCAOB reports. Panel B examines the presence of quality control deficiency in Part II findings of PCAOB reports. Marginal effects are reported for logistic regression models. All variables are defined in Appendix. Z-statistics/t-statistics are computed based on robust standard errors clustered at the firm's auditor level for firm-level analyses and at the target's auditor level for deal-level analyses, respectively. The superscripts *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively, in a two-tailed test.

Column (1) (2) (3)

Dep. Var. ¼ Prob. (TARGET) Prob. (COMPLETION) COMBINECAR

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat.

Panel A: Engagement deficiency reported in Part I findings INSPECT 0.001 0.65 POSTINSPECT �0.001 �1.56 POSTREPORT 0.005** 2.38 POSTREPORT_ENGDEF ¡0.002* ¡1.75 INSPECTAUDITOR �0.018 �1.03 0.003 0.43 REPORTAUDITOR 0.070** 2.18 0.023*** 2.76 REPORTAUDITOR_ENGDEF ¡0.065** ¡2.02 ¡0.027** ¡2.12 ACQINSPECT 0.012 0.59 �0.003 �0.27 ACQREPORT 0.013 0.45 0.003 0.45 COMMONAUDITOR 0.030 1.53 0.006 0.83 CAR �0.002** �1.99 GR 0.001 1.07 SGROWTH 0.001 1.06 LIQUIDITY �0.002 �0.91 LEVERAGE 0.000 1.03 INDBIDS �0.001 �0.55 ASSETSIZE 0.000 1.52 MB �0.000 �0.47 PTE �0.000 �0.39 ROA 0.007** 2.01 INTANG 0.002*** 3.06 R&D 0.013 1.32 INVEST 0.031*** 4.34 VALUE �0.000*** �2.85 0.001*** 2.67 SAMESIC �0.016 �0.88 0.008 1.00 TOEHOLD 0.445* 1.90 TENDER 0.029* 1.80 0.013** 2.08 HOSTILE �0.195*** �2.80 TERMIFEE 0.009** 2.08 �0.002 �0.76 OFBIDDERS �0.069*** �4.62 0.009 1.39 CASHONLY �0.037** �2.46 0.004 0.51 ACQASSETSIZE �0.002 �1.05 ACQMB 0.003 0.85 TARASSETSIZE 0.001 0.62 TARMB �0.003 �0.57 POSTREPORT þ POSTREPORT_ENGDEF 0.003* (p ¼ 0.08) REPORTAUDITOR þ REPORTAUDITOR_ENGDEF 0.005 (p ¼ 0.83) �0.004 (p ¼ 0.66) Country*Year F.E. Yes Industry F.E. Yes Acquirer Country*Target Country*Year F.E. Yes Yes Acquirer Industry F.E. Yes Yes Target Industry F.E. Yes Yes N 70,503 1285 1080 Pseudo/Adjusted R-squared 0.119 0.547 0.159

Column (1) (2) (3)

Dep. Var. ¼ Prob. (TARGET) Prob. (COMPLETION) COMBINECAR

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat.

Panel B: Inspection report type in Part II findings INSPECT �0.001 �0.66 POSTINSPECT �0.001 �0.33 POSTREPORT 0.003* 1.76 POSTREPORT_TYPEA 0.004** 2.07 POSTREPORT_TYPEC ¡0.010* ¡1.80 INSPECTAUDITOR �0.025 �0.78 �0.004 �0.43 REPORTAUDITOR 0.017 0.71 0.015* 1.71 REPORTAUDITOR_TYPEA 0.025 0.57 0.003 0.21 REPORTAUDITOR_TYPEC ¡0.251*** ¡3.58 ¡0.037* ¡1.75 ACQINSPECT �0.005 �0.15 �0.003 �0.34 ACQREPORT 0.006 0.13 �0.001 �0.18 COMMONAUDITOR 0.020 0.85 0.006 0.89

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131822

Table 7 (continued )

Column (1) (2) (3)

Dep. Var. ¼ Prob. (TARGET) Prob. (COMPLETION) COMBINECAR

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat.

CAR �0.001 �0.65 GR 0.001 1.12 SGROWTH 0.001 0.83 LIQUIDITY �0.002 �0.73 LEVERAGE 0.000 0.87 INDBIDS �0.001 �0.57 ASSETSIZE 0.000 1.36 MB �0.000 �0.75 PTE �0.000 �0.36 ROA 0.005 1.31 INTANG 0.001** 2.52 R&D 0.012 1.28 INVEST 0.027*** 3.81 VALUE �0.000* �1.79 0.001** 2.20 SAMESIC �0.017 �1.05 0.007 0.83 TOEHOLD 0.451*** 6.97 TENDER 0.041*** 2.90 0.013 1.59 HOSTILE �0.229*** �4.39 TERMIFEE 0.011 1.63 �0.002 �0.75 OFBIDDERS �0.067*** �5.03 0.008 0.83 CASHONLY �0.045** �2.55 0.003 0.40 ACQASSETSIZE �0.002 �0.51 ACQMB 0.003 0.70 TARASSETSIZE 0.001 0.45 TARMB �0.003 �0.61 POSTREPORT þ POSTREPORT_TYPEA 0.007*** (p < 0.01) POSTREPORT þ POSTREPORT_TYPEC �0.007*** (p < 0.01) REPORTAUDITOR þ REPORTAUDITOR_TYPEA 0.042 (p ¼ 0.29) 0.018 (p ¼ 0.15) REPORTAUDITOR þ REPORTAUDITOR_TYPEC �0.234*** (p < 0.01) �0.022 (p ¼ 0.52) Country*Year F.E. Yes Industry F.E. Yes Acquirer Country*Target Country*Year F.E. Yes Yes Acquirer Industry F.E. Yes Yes Target Industry F.E. Yes Yes N 70,503 1285 1080 Pseudo/Adjusted R-squared 0.119 0.547 0.155

Table 8 PCAOB inspection and post-acquisition consequences. This table reports the effect of PCAOB inspections on deal quality using post-acquisition goodwill write-offs (GWWRITEOFF) or divestitures (DIVESTITURE) as alternative proxies for deal quality. Marginal effects are reported. See Appendix for variable definitions. Z-statistics are computed based on robust standard errors clustered at the target's auditor level. *, **, and *** represent two-tailed significance at the 10%, 5%, and 1% level, respectively. N denotes the number of deals included in the analyses.

Column (1) (2)

Dependent Variable ¼ Prob.(GWWRITEOFF) Prob.(DIVESTITURE)

Mar. Eff. z-stat. Mar. Eff. z-stat.

INSPECTAUDITOR �0.018 �0.64 �0.028 �1.55 REPORTAUDITOR ¡0.071** ¡2.10 ¡0.033*** ¡3.52 ACQINSPECT 0.049 1.47 0.052 1.63 ACQREPORT �0.004 �0.07 �0.021 �1.02 COMMONAUDITOR �0.013 �0.62 0.010 0.53 ACQASSETSIZE 0.021 1.07 0.013*** 2.93 ACQMB 0.017 1.24 0.013*** 2.76 VALUE 0.018* 1.78 �0.009* �1.78 SAMESIC 0.011 0.64 �0.007 �0.13 TENDER 0.033 1.51 �0.018 �1.20 TERMIFEE 0.004 0.47 0.004*** 5.93 OFBIDDERS �0.045* �1.91 �0.032 �1.35 CASHONLY �0.006 �0.28 �0.040 �1.30 Acquirer Country*Target Country* Year fixed effects Yes Yes Acquirer Industry fixed effects Yes Yes Target Industry fixed effects Yes Yes N 1285 1285 Pseudo R-squared 0.416 0.372

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 23

Table 9 PCAOB inspections and international M&A outcomes: robustness checks. This table reports the robustness checks on the effect of PCAOB inspections onM&A outcomes. Marginal effects are reported for logistic regression models. All variables are defined in Appendix. Z-statistics/t-statistics are computed based on robust standard errors clustered at the firm's auditor level for firm-level analyses and target's auditor level for deal-level analyses. The superscripts *, **, and *** denote significance at the 10%, 5%, and 1% levels, respectively, in a two-tailed test.

Column (1) (2) (3)

Dep. Var. ¼ Prob. (TARGET) Prob. (COMPLETION) COMBINECAR

Mar. Eff. z-stat. Mar. Eff. z-stat. Coef. t-stat.

Panel A: Remove the firms/targets audited by uninspected auditors in the sample period POSTINSPECT 0.002 1.57 POSTREPORT 0.005** 2.45 INSPECTAUDITOR �0.021 �0.45 �0.001 �0.21 REPORTAUDITOR 0.041** 2.12 0.013* 1.91 ACQINSPECT �0.025 �1.13 0.004 0.49 ACQREPORT 0.031 0.69 �0.002 �0.17 COMMONAUDITOR 0.051 1.00 0.001 0.31 CAR 0.001 0.79 GR 0.002 1.45 SGROWTH 0.003 0.20 LIQUIDITY �0.002 �0.75 LEVERAGE �0.001 �1.17 INDBIDS �0.001 �0.34 ASSETSIZE 0.000 0.34 MB 0.000 0.52 PTE �0.000* �1.85 ROA 0.010** 2.02 INTANG 0.005*** 4.73 R&D 0.017* 1.75 INVEST 0.047*** 5.66 VALUE �0.002*** �2.90 0.002 1.43 SAMESIC �0.064** �2.37 �0.000 �0.12 TOEHOLD 0.222*** 3.42 TENDER 0.024 0.58 0.003 0.47 HOSTILE �0.198*** �6.76 TERMIFEE 0.008 0.99 �0.000 �0.17 OFBIDDERS �0.061*** �3.76 0.006 0.67 CASHONLY �0.021 �1.61 0.001 0.17 ACQASSETSIZE �0.003 �1.15 ACQMB 0.006* 1.89 TARASSETSIZE 0.000 0.24 TARMB �0.006** �2.16 Country*Year F.E. Yes Industry F.E. Yes Acquirer Country*Target Country*Year F.E. Yes Yes Acquirer Industry F.E. Yes Yes Target Industry F.E. Yes Yes N 38,222 979 805 Pseudo/Adjusted R-squared 0.100 0.511 0.199

Column (1) (2)

Dep. Var. ¼ Variable Prob. (COMPLETION) COMBINECAR

Mar. Eff. z-stat. Coef. t-statistic

Panel B: Remove the deals with SEC registrants as acquirers INSPECTAUDITOR �0.038 �1.30 �0.005 �0.63 REPORTAUDITOR 0.040* 1.72 0.015** 2.26 ACQINSPECT �0.008 �0.25 0.000 0.09 ACQREPORT 0.010 0.23 �0.005 �0.37 COMMONAUDITOR 0.010 0.43 0.006 0.97 VALUE �0.001 �1.39 0.001*** 3.06 SAMESIC �0.015 �0.82 �0.001 �0.12 TOEHOLD 0.473*** 8.15 TENDER 0.055*** 2.76 0.014 1.41 HOSTILE �0.202*** �2.90 TERMIFEE 0.015 1.36 �0.001 �0.41 OFBIDDERS �0.083*** �3.52 0.021** 2.21 CASHONLY �0.055*** �2.83 �0.003 �0.41 ACQASSETSIZE �0.000 �0.08 ACQMB 0.004 1.18 TARASSETSIZE 0.001 0.38

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131824

Table 9 (continued )

Column (1) (2)

Dep. Var. ¼ Variable Prob. (COMPLETION) COMBINECAR

Mar. Eff. z-stat. Coef. t-statistic

TARMB �0.005 �0.96 Acquirer Country*Target Country*Year F.E. Yes Yes Acquirer Industry F.E. Yes Yes Target Industry F.E. Yes Yes N 1159 972 Pseudo/Adjusted R-squared 0.529 0.153

Panel C: Remove the deals with acquirers from countries without any inspected auditors INSPECTAUDITOR �0.012 �0.45 �0.001 �0.16 REPORTAUDITOR 0.043** 2.23 0.014** 2.17 ACQINSPECT �0.009 �0.24 �0.005 �0.45 ACQREPORT 0.012 0.29 �0.002 �0.21 COMMONAUDITOR 0.022 0.81 0.003 0.39 VALUE �0.001 �0.95 0.001** 2.41 SAMESIC �0.015 �0.86 0.007 0.76 TOEHOLD 0.616*** 7.23 TENDER 0.058*** 3.42 0.015** 2.50 HOSTILE �0.185*** �7.15 TERMIFEE 0.019** 2.52 �0.002 �1.00 OFBIDDERS �0.076*** �8.56 0.016** 2.56 CASHONLY �0.055*** �3.39 0.003 0.45 ACQASSETSIZE �0.002 �0.75 ACQMB 0.003 0.88 TARASSETSIZE 0.001 0.47 TARMB �0.003 �0.73 Acquirer Country*Target Country*Year F.E. Yes Yes Acquirer Industry F.E. Yes Yes Target Industry F.E. Yes Yes N 1207 1009 Pseudo/Adjusted R-squared 0.525 0.161

Panel D: Remove the deals with acquirers with auditor changes INSPECTAUDITOR 0.004 0.13 �0.007 �0.85 REPORTAUDITOR 0.038** 2.15 0.016* 1.69 ACQINSPECT 0.004 0.12 �0.003 �0.29 ACQREPORT 0.016 0.45 0.001 0.17 COMMONAUDITOR 0.034 1.17 0.000 0.08 VALUE �0.000 �0.76 0.001** 2.30 SAMESIC �0.039** �2.22 �0.006 �0.34 TOEHOLD 0.685*** 5.91 TENDER 0.047** 2.55 0.006 0.76 HOSTILE �0.190*** �6.75 TERMIFEE 0.006** 2.30 �0.002 �0.62 OFBIDDERS �0.093*** �4.33 0.018* 1.68 CASHONLY �0.045** �2.33 �0.001 �0.11 ACQASSETSIZE �0.003 �0.92 ACQMB 0.005 1.22 TARASSETSIZE 0.001 0.49 TARMB �0.005 �1.18 Acquirer Country*Target Country*Year F.E. Yes Yes Acquirer Industry F.E. Yes Yes Target Industry F.E. Yes Yes N 955 777 Pseudo/Adjusted R-squared 0.569 0.208

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 25

indicators are included in Eq. (1), themain effect of POSTREPORTcaptures the effect of having a PCAOB inspection report that is neither Type A nor Type C. POSTREPORT_TYPEA captures the incremental effect of a Type A inspection report, whereas POSTREPORT_TYPEC captures the incremental effect of a Type C inspection report. Similarly, for the deal-level analyses, we construct and include two indicator variables, REPORTAUDITOR_TYPEA and REPORTAUDITOR_TYPEC, in Eqs. (3) and (4). As shown in Panel B of Table 7, the main effects of POSTREPORT and REPORTAUDITOR remain positive and significant for the likelihood of receiving a bid and announcement returns, but having an inspection report with an explicit quality control

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131826

criticism (Type C) results in relatively lower levels of the likelihood of receiving a bid and the likelihood of deal completion.27,28

4.6. Robustness tests

In our main analysis, we use combined cumulative abnormal announcement returns as an ex-ante measure of deal quality. We check the robustness of the results to post-acquisition performance as an alternative measure of deal quality. Following Goodman et al. (2014), we use post-acquisition goodwill write-offs (GWWRITEOFF) and divestitures (DIVESTITURE) as mea- sures of deal quality. A lower likelihood of goodwill write-offs or divestitures represents a higher deal quality. The results, based on Eq. (4) with GWWRITEOFF and DIVESTITURE as dependent variables, are presented in Table 8. We find that deals involving targets with inspected auditors whose PCAOB inspection reports are disclosed to the public experience less goodwill write-offs or divestitures, suggesting that PCAOB oversight, specifically the public disclosure of inspection reports, expands acquirers’ information sets about targets, allowing them to identify better targets, which, in turn, results in better post-acquisition performance, as captured by less goodwill write-off or divestitures.

We conduct a number of other robustness tests to ensure that the inferences we obtain from the main findings are not subject to other explanations and are robust to alternative research designs. The results are presented in Table 9. First, we restrict our sample to observations involving inspected auditors only in an additional test to mitigate the potential con- founding effect arising from the systematic difference between inspected and uninspected auditors (Aobdia and Shroff, 2017). The result shown in Panel A of Table 9 indicates that our inferences are unchanged using the reduced sample. Second, we perform deal-level analyses after removing deals involving SEC registrants as acquirers to rule out the possible confounding effect fromPCAOB inspection of acquirers' auditors onM&A outcomes. The results reported in Panel B of Table 9 show that our main inferences are unaffected, suggesting that our results are unlikely to be driven by the effect of PCAOB inspection to acquirers’ auditors. Third, we perform another set of deal-level analyses after excluding deals with acquirers from countries without any inspected auditors. Note that in ourmain analysis, we already excluded deals with targets from countries without any inspected auditors. As shown in Panel C of Table 9, the effects of having a PCAOB inspection report on the probability of deal completion and the announcement returns remain significant with the further reduced sample. Finally, we carry out deal-level analyses based on a reduced sample after excluding acquirers that change auditors during our sample period. Note that in our main analyses, we already exclude deals with targets that change auditors during our sample period. Although the sample is substantially smaller, the result in Panel D of Table 9 shows that the main inferences from the results remain unchanged.

5. Conclusion

In this study, we examine how PCAOB international inspections of non-U.S. auditors affect international M&A outcomes. Our results reveal that firms audited by PCAOB-inspected auditors are more likely to receive an acquisition bid, and con- ditional on receiving the bid, deals involving such firms as targets are more likely to be completed, following the public release of their auditors' PCAOB inspection reports. We also show that the announcement returns of deals involving targets audited by auditors whose PCAOB inspection reports are publicly available prior to the deal announcement are higher than those of deals involving targets with uninspected auditors. In addition, we find that deals involving targets audited by auditors whose PCAOB inspection reports are publicly available exhibit superior post-acquisition performance. Taken together, the evidence suggests that PCAOB oversight improves acquirers' perceptions of the target's financial reporting quality, reducing information uncertainty in M&A transactions, and that the positive effect of PCAOB oversight on M&A outcomes appears after the public disclosure of the inspection report, rather than immediately after the inspection. The effect of PCAOB oversight on M&A outcomes is more pronounced for targets from countries with weak legal institutions and for deals with greater information uncertainty. Additional analyses suggest that engagement deficiencies and unremediated

27 We also conduct these tests on the matched sample for the probability of receiving a bid. In the analysis of engagement deficiencies reported in the Part I findings, similar to the result based on the full sample, the effect of having an inspection report is positive when no engagement deficiency is found, and this effect weakens significantly when engagement deficiencies are reported. While the effect of the disclosure of an inspection report is still positive even with the engagement deficiency in the full panel data, in the matched sample result, the effect of the disclosure of an inspection report becomes insig- nificant when an engagement deficiency is found. In the analysis of the Part II findings, inferences from the results are virtually the same between the analysis based on the full panel and the analysis based on the matched sample. That is, the likelihood of receiving a bid is higher following an inspection report that contains no quality control deficiency (Type A). The effect of PCAOB oversight, however, is negative for an inspection report with an explicit quality control criticism (Type C). 28 If there are multiple inspection reports available for an auditor, we further explore whether the changes in the content of inspection reports from the previous report to the most recent report matter. Untabulated results show that when an inspection report has two consecutive engagement deficiencies, the positive effects of PCAOB oversight on M&A outcomes disappear. Regarding Part II findings, we find that having an inspection report with two consecutive quality control deficiencies results in a lower likelihood of receiving a bid and that reports with deteriorating contents (that is, from having no or remediated quality control deficiency to having quality control deficiencies) have no positive effect on M&A outcomes. These results should be inter- preted with caution, however, because a low test power arising from a small number of Part II findings with quality control deficiencies may contribute to insignificant results.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 27

auditor-level quality control deficiencies disclosed in inspection reports significantly weaken the positive effect of PCAOB oversight on M&A outcomes.

Our study provides novel evidence on the real economic benefits of the PCAOB international inspection program. Our findings also provide strong support for the relation between audit quality and capital allocation efficiency. As such, our study is informative to regulators, investors, and other market participants across the globe.

Acknowledgement

We thank Joanna Wu (the editor), an anonymous referee, Inder Khurana, Ke Wang (discussant), and the conference/ workshop participants at 2018 European Accounting Association Conference, 2018 MIT Asian Conference in Accounting, and Shanghai University of Finance and Economics for valuable comments. All errors are our own.

Appendix. Variable definitions

Panel A: Variable definitions for firm-year level analyses to test the likelihood of receiving a bid

Variable name Definition

Dependent variable and variables of interest TARGET An indicator variable that equals one if a firm received at least one bid in a given fiscal year, and zero otherwise. INSPECT An indicator variable that equals one if a firm's auditor undergoes PCAOB inspections during the sample period, and zero

otherwise. POSTINSPECT An indicator variable that equals one for the fiscal years beginning after the completion of a company's auditor's PCAOB

inspection and ending before the disclosure of the inspection report, and zero otherwise. POSTREPORT An indicator variable that equals one for the fiscal years beginning after the disclosure of a company's auditor's first

inspection report, and zero otherwise. TREAT An indicator variable that equals one if the firm's auditor has been inspected by the PCAOB and an inspection report has

been released by the PCAOB during the sample period in the matched sample, and zero otherwise. POST An indicator variable that equals one for years after the firm's auditor has been inspected by the PCAOB and an

inspection report has been released by the PCAOB in the matched sample, and zero otherwise. We assign the actual report year of the treatment firm as a pseudo report year to thematched control firm even though control firms' auditors have never been inspected during our sample period.

Control variables CAR The market-adjusted one-year cumulative abnormal return in the fiscal year t-1. GR An indicator variable that equals one if a firm has the combination of low growth-high liquidity-low leverage or high

growth-low liquidity-high leverage in the fiscal year t-1, and zero for all other combinations. SGROWTH The average annual sales growth rate over the previous three years prior to the fiscal year t. LIQUIDITY The average ratio of net liquid assets to total assets over the previous three years prior to the fiscal year t, where net

liquid assets are calculated as cash plus marketable securities less current liabilities. LEVERAGE The average ratio of long-term debt to equity over the previous three years prior to the fiscal year t. INDBIDS An indicator variable that equals one if there were bid activities in the firm's country-industry in the fiscal year t-1, and

zero otherwise. ASSETSIZE The natural logarithm of total assets in the fiscal year t-1. MB The market value of total assets divided by the book value of total assets in the fiscal year t-1. PTE The share price divided by the earnings per share in the fiscal year t-1. ROA The return on total assets in the fiscal year t-1. INTANG The ratio of intangible assets to total assets in the fiscal year t-1. R&D R&D intensity, calculated as the ratio of R&D expenditure to total assets in the fiscal year t-1. INVEST Investment intensity, measured as capital expenditure divided by total assets in the fiscal year t-1.

Panel C: Variable definitions for other tests

Cross-sectional analyses RULE An annual country-level measure of “the perceptions of the extent to which agents have confidence in and abide by the rules of

society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence” (Kaufmann et al., 2009), taken from the World Bank.

CROSSBORDER An indicator variable that equals one if the acquirer and the target are from different countries, and zero otherwise. Content of report analyses POSTREPORT_ENGDEF An indicator variable that equals one if, by the beginning of the fiscal year, at least one PCAOB inspection report for the firm's

auditor is available, and the most recent inspection report identifies engagement deficiency in Part I findings, and zero otherwise.

REPORTAUDITOR_ENGDEF An indicator variable that equals one if, before the deal announcement year, at least one of the target auditor's inspection reports has already been released by the PCAOB in the year, and if the target's most recent inspection report identifies engagement deficiency in Part I findings prior to the deal announcement year, and zero otherwise.

POSTREPORT_TYPEA An indicator variable that equals one if, by the beginning of the fiscal year, at least one PCAOB inspection report for the firm's auditor is available, and the most recent inspection report does not contain quality control deficiency in Part II findings, and zero otherwise.

POSTREPORT_TYPEC An indicator variable that equals one if, by the beginning of the fiscal year, at least one PCAOB inspection report for the firm's auditor is available, and the most recent inspection report includes quality control criticism regarding Part II findings, and zero otherwise.

REPORTAUDITOR _TYPEA An indicator variable that equals one if, before the deal announcement year, at least one of the target auditor's inspection reports has already been released by the PCAOB in the year, and if its most recent inspection report does not contain quality control deficiency in Part II findings, and zero otherwise.

REPORTAUDITOR _TYPEC An indicator variable that equals one if, before the deal announcement year, at least one of the target auditor's inspection reports has already been released by the PCAOB in the year, and if its most recent inspection report receives quality control deficiency in Part II findings, and zero otherwise.

Post-acquisition consequences analyses GWWRITEOFF An indicator variable that equals one if a firm records a large decrease in goodwill in the three-year period following an

acquisition that generates a larger increase in goodwill, and zero otherwise. A large decrease/increase in goodwill is defined as a decrease/increase in goodwill greater or equal to one percent of total assets.

Panel B: Variable definitions for deal-level analyses to test the likelihood of deal completion and announcement returns

Variable name Definition

Dependent variables and variables of interest COMPLETION An indicator variable that equals one if the deal is completed, and zero otherwise. COMBINECAR The market-adjusted three-day cumulative abnormal returns around the deal announcement date, averaging the cumulative

abnormal returns of the target and the acquirer. INSPECTAUDITOR An indicator variable that equals one if the target's auditor has been inspected by the PCAOB and an inspection report has not

been released by the PCAOB in the year prior to the deal announcement year. REPORTAUDITOR An indicator variable that equals one if the target's auditor's inspection report has been released by the PCAOB in the year prior

to the deal announcement year. Control variables Test of the likelihood of deal completion ACQINSPECT An indicator variable that equals one if the acquirer's auditor has already been inspected by the PCAOB and an inspection report

has not been released by the PCAOB in the year prior to the deal announcement year, and zero otherwise. ACQREPORT An indicator variable that equals one if the acquirer's auditor's inspection report has already been released by the PCAOB in the

year prior to the deal announcement year, and zero otherwise. COMMONAUDITOR An indicator variable that equals one if the target and the acquirer share the same auditor in the deal announcement year, and

zero otherwise. VALUE The natural logarithm of deal value in million U.S. dollars reported by SDC. SAMESIC An indicator variable that equals one if the target and the acquirer share the same two-digit SIC code, and zero otherwise. TOEHOLD An indicator variable that equals one if the acquirer has an ownership stake in the target at the time of deal announcement, and

zero otherwise. TENDER An indicator variable that equals one if a bid is structured as a tender offer, and zero otherwise. HOSTILE An indicator variable that equals one if a bid is classified as hostile, and zero otherwise. TERMIFEE The target-payable termination fee in million U.S. dollars. The fee is set equal to zero in deals without termination fees (see

Dhaliwal et al., 2016). OFBIDDERS The number of bidders that have made public bids in an auction, where an auction is defined as all public bids on a target within

365 days of an initial bid. CASHONLY An indicator variable that equals one if a deal is paid entirely in cash, and zero otherwise. Test of announcement returns ACQASSETSIZE The natural logarithm of the acquirer's total assets in the year prior to deal announcement year. ACQMB The market value of the acquirer's market value of total assets divided by the book value of the acquirer's total assets in the year

prior to deal announcement year. TARASSETSIZE The natural logarithm of the target's total assets in the year prior to deal announcement year. TARMB The market value of the target's total assets divided by the book value of the target's total assets in the year prior to deal

announcement year.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131828

DIVESTURE An indicator variable that equals one if an acquisition has a subsequent divestiture (i.e. the target has the same four-digit SIC code as the firm divested and the divested firm's parent is the acquirer in the original M&A) in the three-year period following an acquisition, and zero otherwise.

Dynamic effects analyses YEARt-2REPORT An indicator variable for clients of inspected auditors in the second year prior to the first inspection report year. YEARt-1REPORT An indicator variable for clients of inspected auditors in the year prior to the first inspection report year. YEARtREPORT An indicator variable for clients of inspected auditors in the year in the first inspection report year. YEARtþnREPORT An indicator variable for clients of inspected auditors in the years after the first inspection report year.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 101318 29

References

Abbott, L.J., Gunny, K., Zhang, C.T., 2012. When the PCAOB talks, who listens? Evidence from client firm reaction to adverse, GAAP deficient PCAOB in- spection reports. Audit J. Pract. Theor. 32 (2), 1e31.

Ai, C., Norton, E.C., 2003. Interaction terms in logit and probit models. Econ. Lett. 80 (1), 123e129. Ambrose, B.W., Megginson, W.L., 1992. The role of asset structure, ownership structure, and takeover defenses in determining acquisition likelihood. J.

Financ. Quant. Anal. 27 (4), 575e589. Amel-Zadeh, A., Zhang, Y., 2015. The economic consequences of financial restatements: evidence from the market for corporate control. Account. Rev. 90

(1), 1e29. Andrade, G., Mitchell, M., Stafford, E., 2001. New evidence and perspectives on mergers. J. Econ. Perspect. 15 (2), 103e120. Andrade, G., Stafford, E., 2004. Investigating the economic role of mergers. J. Corp. Finance 10 (1), 1e36. Aobdia, D., 2018. The impact of the PCAOB individual engagement inspection processdpreliminary evidence. Account. Rev. 93 (4), 53e80. Aobdia, D., Shroff, N., 2017. Regulatory oversight and auditor market share. J. Account. Econ. 63 (2), 262e287. Bae, G.S., Choi, S.U., Dhaliwal, D.S., Lamoreaux, P.T., 2017. Auditors and client investment efficiency. Account. Rev. 92 (2), 19e40. Bates, T.W., Lemmon, M.L., 2003. Breaking up is hard to do? An analysis of termination fee provisions and merger outcomes. J. Financ. Econ. 69 (3),

469e504. Betton, S., Eckbo, B.E., Thorburn, K.S., 2009. Merger negotiations and the toehold puzzle. J. Financ. Econ. 91 (2), 158e178. Bishop, C.C., Hermanson, D.R., Houston, R.W., 2013. PCAOB inspections of international audit firms: initial evidence. Int. J. Audit. 17 (1), 1e18. Boone, J.P., Khurana, I.K., Raman, K.K., 2014. Did the 2007 PCAOB disciplinary order against Deloitte impose actual costs on the firm or improve its audit

quality? Account. Rev. 90 (2), 405e441. Bradley, M., Desai, A., Kim, E.H., 1988. Synergistic gains from corporate acquisitions and their division between the stockholders of target and acquiring

firms. J. Financ. Econ. 21 (1), 3e40. Brennan, M., Cao, H., 1997. International portfolio investment flows. J. Finance 52 (5), 1851e1880. Cai, Y., Kim, Y., Park, J.C., White, H.D., 2016. Common auditors in M&A transactions. J. Account. Econ. 61 (1), 77e99. Campa, J.M., Moschieri, C., 2008. The European M&A Industry: Trends, Patterns, and Shortcomings. Working Paper. University of Navarra. Carcello, J.V., Hollingsworth, C., Mastrolia, S.A., 2011. The effect of PCAOB inspections on Big 4 audit quality. Res. Account. Regul. 23 (2), 85e96. Center for Audit Quality (CAQ), 2012. Guide to PCAOB Inspections. Center for Audit Quality, Washington, DC. Chen, C.W., Collins, D.W., Kravet, T.D., Mergenthaler, R., 2018. Financial statement comparability and the efficiency of acquisition decisions. Contemp.

Account. Res. 35 (1), 164e202. DeFond, M.L., 2010. How should the auditors be audited? Comparing the PCAOB inspections with the AICPA peer reviews. J. Account. Econ. 49 (1), 104e108. DeFond, M.L., Lennox, C.S., 2017. Do PCAOB inspections improve the quality of internal control audits? J. Account. Res. 55 (3), 591e627. DeFond, M.L., Zhang, J., 2014. A review of archival auditing research. J. Account. Econ. 58 (2), 275e326. Demsetz, H., 1968. Do competition and monopolistic competition differ? J. Polit. Econ. 76 (1), 146e148. Dhaliwal, D.S., Lamoreaux, P.T., Litov, L.P., Neyland, J.B., 2016. Shared auditors in mergers and acquisitions. J. Account. Econ. 61 (1), 49e76. Dickey, J.C., 2013. Top 10 Issues for Cross-Border M&A and Strategic Investments in 2013. April 4, 2013. Dykema Gossett PLLC. Available at: https://www.

dykema.com/resources-alerts-Top-10-Issues-for-Cross-Border-M-and-A-and-Strategic-Investments-in-2013_4-2013.html. Eckbo, B.E., Langohr, H., 1989. Information disclosure, method of payment, and takeover premiums: public and private tender offers in France. J. Financ.

Econ. 24 (2), 363e403. El Ghoul, S., Guedhami, O., Pittman, J.A., 2016. Cross-country evidence on the importance of Big Four auditors to equity pricing: the mediating role of legal

institutions. Account. Org. Soc. 54, 60e81. Erel, I., Liao, R.C., Weisbach, M.S., 2012. Determinants of cross-border mergers and acquisitions. J. Finance 67 (3), 1045e1082. Francis, J.R., Huang, S.X., Khurana, I.K., 2016. The role of similar accounting standards in cross-border mergers and acquisitions. Contemp. Account. Res. 33

(3), 1298e1330. Francis, J.R., Wang, D., 2008. The joint effect of investor protection and Big 4 audits on earnings quality around the world. Contemp. Account. Res. 25 (1),

157e191. Fung, S., Raman, K.K., Zhu, X., 2017. Does the PCAOB international inspection program improve audit quality for non-US-listed foreign clients? J. Account.

Econ. 64 (1), 15e36. Gipper, B., Leuz, C., Maffett, M., 2019. Public Audit Oversight and Reporting Credibility: Evidence from the PCAOB Audit Inspection Regime. The Review of

Financial Studies, Forthcoming. Glover, S.M., Prawitt, D.F., Taylor, M.H., 2009. Audit standard setting and inspection for US public companies: a critical assessment and recommendations for

fundamental change. Account. Horiz. 23 (2), 221e237. Goodman, T., Neamtiu, M., Shroff, N., White, H., 2014. Management forecast quality and capital investment decisions. Account. Rev. 89 (1), 331e365. Gordon, R., Bovenberg, A., 1996. Why is capital so immobile internationally? Possible explanations and implications for capital income taxation. Am. Econ.

Rev. 86 (5), 1057e1075. Gramling, A.A., Krishnan, J., Zhang, Y., 2011. Are PCAOB-identified audit deficiencies associated with a change in reporting decisions of triennially inspected

audit firms? Audit J. Pract. Theor. 30 (3), 59e79. Gul, F.A., Zhou, G.S., Zhu, X.K., 2013. Investor protection, firm informational problems, Big N auditors, and cost of debt around the world. Audit J. Pract. Theor.

32 (3), 1e30. Gunny, K.A., Zhang, T.C., 2013. PCAOB inspection reports and audit quality. J. Account. Publ. Pol. 32 (2), 136e160. Hanlon, M., Shroff, N., 2020. Insights into Auditor Public Oversight Boards: whether, How, and Why They “work.” Working Paper. MIT Sloan School of

Management. Hermanson, D.R., Houston, R.W., Rice, J.C., 2007. PCAOB inspections of smaller CPA firms: initial evidence from inspection reports. Account. Horiz. 21 (2),

137e152. Jackson, H.E., Roe, M.J., 2009. Public and private enforcement of securities laws: resource-based evidence. J. Financ. Econ. 93 (2), 207e238. Kaufmann, D., Kraay, A., Mastruzzi, M., 2009. Governance Matters VIII: Aggregate and Individual Governance Indicators 1996-2008. The World Bank. Kausar, A., Shroff, N., White, H., 2016. Real effects of the audit choice. J. Account. Econ. 62 (1), 157e181. Krishnan, J., Krishnan, J., Song, H., 2017. PCAOB international inspections and audit quality. Account. Rev. 92 (5), 143e166.

Y. Kim et al. / Journal of Accounting and Economics 70 (2020) 10131830

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 2006. What works in securities laws? J. Finance 61 (1), 1e32. Lamoreaux, P.T., 2016. Does PCAOB inspection access improve audit quality? An examination of foreign firms listed in the United States. J. Account. Econ. 61

(2), 313e337. Lennox, C., Pittman, J., 2010. Auditing the auditors: evidence on the recent reforms to the external monitoring of audit firms. J. Account. Econ. 49 (1),

84e103. Louis, H., 2005. Acquirers' abnormal returns and the non-Big 4 auditor clientele effect. J. Account. Econ. 40 (1), 75e99. Martin, X., Shalev, R., 2016. Target firm-specific information and acquisition efficiency. Manag. Sci. 63 (3), 672e690. McNichols, M.F., Stubben, S.R., 2015. The effect of target-firm accounting quality on valuation in acquisitions. Rev. Account. Stud. 20 (1), 110e140. Minnis, M., Shroff, N., 2017. Why regulate private firm disclosure and auditing? Account. Bus. Res. 47 (5), 473e502. Moeller, S., Schilingemann, F., Stulz, R.M., 2005. Wealth destruction on a massive scale? A study of acquiring-firm returns in the recent merger wave. J.

Finance 60 (2), 757e782. Nagy, A.L., 2014. PCAOB quality control inspection reports and auditor reputation. Audit J. Pract. Theor. 33 (3), 87e104. Officer, M.S., 2003. Termination fees in mergers and acquisitions. J. Financ. Econ. 69 (3), 431e467. Palepu, K.G., 1986. Predicting takeover targets: a methodological and empirical analysis. J. Account. Econ. 8 (1), 3e35. Public Company Accounting Oversight Board (PCAOB), 2005. Annual Report. PCAOB, Washington, DC. Rothenbuecher, J., von J., 2008. The rise of emerging markets in mergers and acquisitions. A.T. Kearney. Roychowdhury, S., Shroff, N., Verdi, R.S., 2019. The effects of financial reporting and disclosure on corporate investment: a review. J. Account. Econ. 68 (2e3),

101246. Schwert, G.W., 2002. Hostility in takeovers: in the eyes of the beholder? J. Finance 55 (6), 2599e2640. Shroff, N., 2019. Real Effects of PCAOB International Inspections. The Accounting Review, forthcoming. Skaife, H.A., Wangerin, D.D., 2013. Target financial reporting quality and M&A deals that go bust. Contemp. Account. Res. 30 (2), 719e749. Stigler, G.J., 1971. The theory of economic regulation. Bell J. Econ. Manag. Sci. 2 (1), 3e21. Xie, Y., Yi, H.S., Zhang, Y., 2013. The value of Big N target auditors in corporate takeovers. Audit J. Pract. Theor. 32 (3), 141e169.

__MACOSX/group assignment for quantitative analysis/professor gave us/._paper_sample.pdf