UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Essays on foreign firms listed in the United States Zhang, Lei (Ray) 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2018_september_zhang_lei.pdf [ 1.29MB ]
Metadata
JSON: 24-1.0368915.json
JSON-LD: 24-1.0368915-ld.json
RDF/XML (Pretty): 24-1.0368915-rdf.xml
RDF/JSON: 24-1.0368915-rdf.json
Turtle: 24-1.0368915-turtle.txt
N-Triples: 24-1.0368915-rdf-ntriples.txt
Original Record: 24-1.0368915-source.json
Full Text
24-1.0368915-fulltext.txt
Citation
24-1.0368915.ris

Full Text

 ESSAYS ON FOREIGN FIRMS LISTED IN THE UNITED STATES  by  Lei (Ray) Zhang B.B.A in Business Administration, Simon Fraser University, 2012   A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Business Administration)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  July 2018 © Lei (Ray) Zhang, 2018   ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: ESSAYS ON FOREIGN FIRMS LISTED IN THE UNITED STATES  submitted by Lei (Ray) Zhang  in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Business Administration  Examining Committee: Prof. Dan Simunic Co-supervisor Prof. Ralph Winter Co-supervisor  Prof. Rafael Rogo Supervisory Committee Member Prof. Vadim Marmer University Examiner Prof. Thomas Davidoff University Examiner  Additional Supervisory Committee Members:  Supervisory Committee Member  Supervisory Committee Member   iii  Abstract This thesis is a collection of two essays on US listed foreign firms, which represent a significant proportion of firms trading in US markets. These firms, like listed US firms, are subject to the litigation system in the US and are subject to monitoring by the US regulators, such as the Securities Exchange Commission (SEC) and the Public Company Accounting Oversight Board (PCAOB). Because of US regulators’ tripartite mission to protect investors, maintain efficient markets, and facilitate capital formation, it is not obvious how the US regulators monitoring of foreign firms compares with US firms. In addition, it is not trivial how the foreign auditors react to the legal environment in the US compared with the US auditors.  The first chapter of this thesis explores whether SEC’s monitoring activities differ for US versus foreign firms and whether monitoring varies based on attributes of the home country’s institutions. We find that, on average, SEC monitors foreign firms with less frequency and SEC provides increased monitoring for those foreign firms where SEC monitoring is most valuable to US. In addition, SEC reduces monitoring when it can rely on public and private enforcement in the foreign firm’s home country. Our study highlights the heterogeneity in SEC monitoring of foreign firms.   The second chapter examines the effects of cross-listings in the United States on the pricing and quality of foreign auditors. The Bonding hypothesis suggests that foreign auditors can leapfrog their home countries’ weak legal institutions and provide quality similar to US auditors for their cross-listed clients because they abide by U.S. legal requirements. However, recent evidence, such as the first chapter of these essays, suggests that the U.S. legal enforcement is weaker for foreign entities because U.S. regulators focus their resources on domestic firms. This study finds that despite the nuances of the functional convergence hypothesis, foreign auditors are found to provide   iv  quality at least as good as the U.S. non-Big4 auditors. The findings of this paper can mitigate some recent concerns about the quality of foreign auditors practicing in the U.S. cross-listing market.          v  Lay Summary This thesis examines the level of bonding of foreign firms and foreign auditors to the US. I find that, on average, foreign firms are subject to less monitoring frequency than US firms. In addition, foreign firms are subject to less monitoring frequency when the foreign firm’s home country’s public and private enforcement is strong. I document heterogeneity in SEC monitoring of foreign firms. Meanwhile, despite the nuances of the bonding level to the US, foreign auditors are found to provide quality at least as good as the U.S. non-Big4 auditors.              vi  Preface The research essay in chapters 1 is based on the research paper with James Naughton (Northwestern University), Rafael Rogo (Indiana University) and Jayanthi Sunder (University of Arizona) that is forthcoming at Review of Accounting Studies. For the co-authored project, every author worked on all aspects of the paper. However, my contribution is more concentrated in research question identification and empirical analysis. I spent relatively little time in the aspect of manuscript writing. The research projects in chapters 2 were identified and conducted solely by myself.                   vii  Table of Contents Abstract .................................................................................................................................... iii Lay Summary ............................................................................................................................. v Preface ...................................................................................................................................... vi Table of Contents ...................................................................................................................... vii List of Tables ............................................................................................................................. ix Acknowledgements .................................................................................................................... xi Introduction .............................................................................................................................. 34 Chapter1: SEC Monitoring of Foreign Firms' Disclosures in the Presence of Foreign Regulators .............................................................................................................. 35 1.1 Introduction ..........................................................................................................35 1.2 Literature review and hypotheses development .................................................41 1.3 Institutional Setting and Measurement of SEC Monitoring ...............................43 1.4 Sample and data ...................................................................................................47 1.5 Research Design and Results ...............................................................................51 1.6 Conclusion ..............................................................................................................63 1.7 Tables ...................................................................................................................65 Chapter 2: Auditor Quality of U.S.-Listed Foreign Firms ............................................ 80 2.1 Introduction ..........................................................................................................80 2.2 Literature review, Institutional Background and hypothesis development ........................82 2.3 Testing Hypothesis 1 ................................................................................................91 2.4 Testing Hypothesis 2 ................................................................................................96 2.5 Testing Hypothesis 3 .............................................................................................. 101 2.6 Conclusion ............................................................................................................ 102 2.7 Tables ................................................................................................................. 103 Conclusion ...............................................................................................................................117 Bibliography ............................................................................................................................118   viii  Appendix A: Chapter 1 ......................................................................................................121 A.1 Variable Definitions ..................................................................................................... 121 A.2 Figure -- Overview of Domestic versus Foreign Private Issuer Reporting and Disclosure Requirements ............................................................................................... 127 Appendix B: Chapter 2 ......................................................................................................128 B.1 Variable Definitions ..................................................................................................... 128 B.2 Stylished Model ....................................................................................................... 130    ix  List of Tables Table 1.1 Sample Composition ................................................................................................65 Table 1.2 Location of Primary Exchange relative to Headquarters .......................................66 Table 1.3 Sample Composition by Country .............................................................................68 Table 1.4 Variation in SEC Monitoring based on Home Country Enforcement ....................69 Table 1.5 Variation in SEC Monitoring with Country Fixed Effects .....................................71 Table 1.6 Alternate Measure of SEC Oversight to Mitigate Partial Observability Concerns .................................................................................................................................................72 Table 1.7 Alternative Measures of SEC Monitoring Intensity ...............................................73 Table 1.8 Variation in SEC Monitoring based on US Investor Exposure ..............................76 Table 1.9 Comparison of Monitoring Intensity for US versus Foreign Firms .......................78 Table 2.1 – Sample Selection ................................................................................................ 103 Table 2.2 – Sample Distribution by Country ........................................................................ 104 Table 2.3 – Analysis of Audit Fees (Test of Hypothesis 1) ................................................... 107 Table 2.4 – Analysis of Audit Fees with First Differences Regression – Supplemental Robustness Tests (Test of Hypothesis 1) ............................................................................... 108 Table 2.5 – Analysis of Audit Fees with U.S. Non-Big4 auditors Variation – Supplemental Robustness Tests (Test of Hypothesis 1) ............................................................................... 109 Table 2.6 – Sample Selection of Restatement ....................................................................... 110 Table 2.7 – Analysis of Restatement Probability (Test of Hypothesis 2) ............................. 111 Table 2.8 – PCAOB Sample Selection ................................................................................... 112 Table 2.9 – Descriptive Statistics of PCAOB Analysis Sample ............................................ 113 Table 2.10 – PCAOB Analysis ............................................................................................... 114 Table 2.11 – Market Reaction to the Announcement of Change of Auditors ....................... 115   x  Table 2.7.12 – Analysis of Restatement Probability with country variation (Test of Hypothesis 3) ......................................................................................................................... 116    xi  Acknowledgements I would like to express my sincere thanks to my supervisory committee members – Dr. Dan Simunic, Dr. Rafael Rogo and Dr. Ralph Winter. In addition, I would like to thank other Accounting department faculty members as well as my fellow PhD students for useful feedback. I would not have continued my PhD program without the gentle guidance and warm encouragement of the co-chair of my committee, Dan Simunic. I really admire Dan’s commitment and generosity towards research and his students. Dan’s curiosity in research topics kept me motivated. Under Dan’s guidance, I learned to be more persistent in research.  I would like to give my special thanks to Rafael Rogo for his invaluable guidance throughout my PhD program. It is my great privilege to work with him closely in different projects from year one till now. I can never thank him enough.    Last but not least, this work is dedicated to my wife Cynthia Li, who is my great companion and supports me through all of the most difficult moments in my PhD life.34  Introduction This thesis is a collection of two essays exploring the corporate governance environment of US listed foreign firms. The functional convergence hypothesis suggests that when foreign firms are listed in the US, they are bonded to the US corporate governance environment. However, a handful of prior studies suggests that SEC’s enforcement actions against foreign firms may occur at a different rate compared to US firms. The first essay, therefore, investigates the forces that shape SEC’s monitoring of foreign firms’ disclosures. The second essay investigates the impact of cross-listings on foreign auditors.  Each chapter is designed to be self-contained. I have a more detailed discussion of the research question and contribution in the introduction section of each chapter. 35  Chapter1: SEC Monitoring of Foreign Firms' Disclosures in the Presence of Foreign Regulators 1.1  Introduction Foreign firms represent a significant proportion of firms trading in US markets, and like their US counterparts, are subject to SEC oversight.1 While SEC’s tripartite mission—to protect investors, maintain fair, orderly, and efficient markets, and facilitate capital formation—equally applies to foreign firms, there are some moderating factors. First, foreign firms are subject to home-country regulatory oversight in addition to SEC oversight, and second, US investors are typically exposed to only a portion of the shares issued by foreign firms. Thus, it is hard to predict how SEC oversight will vary across foreign firms. For example, while intensive monitoring of foreign firms would enhance investor protection, it could impose onerous reporting requirements on foreign firms already subject to rigorous oversight by home-country regulators, thus hurting the capital formation objective of the SEC. At the same time, reduced monitoring by the SEC could hurt the investor protection objective when US investors have exposure to the foreign firm, particularly since foreign firms generally have lower quality accounting earnings than comparable US firms (Lang, Raedy and Wilson, 2006).  We provide evidence on how the SEC varies its monitoring of foreign firms based on the quality of the home-country regulator and US investor exposure. Finding that the SEC incorporates these factors is important for two reasons. First, it provides evidence to support the emerging regulatory philosophy of substituted compliance proposed by Tafara and Peterson (2007). The                                                  1 For example, the New York Stock Exchange had a total global market value of $26 trillion in 2006, which included 424 non-US issuers valued at $10 trillion. Source: http://www.nyse.com/press/1190629848623.html   36  objective of the substituted compliance framework is to enable SEC to achieve its investor protection objective more efficiently by recognizing improvements made in regulatory monitoring in some foreign countries. Second, understanding the determinants of SEC monitoring intensity for foreign firms is important because it has implications for the bonding hypothesis (e.g., Piotroski and Srinivasan, 2008; Doidge, Karolyi and Stulz, 2004). Finding that the SEC relies on foreign regulators and incorporates the exposure of US investors suggests that there is heterogeneity in the strength of bonding. We use comment letters issued by the SEC as a proxy for SEC oversight because they are the typical outcome of an SEC review of financial statements. As we outline in more detail in Section 4, our main analyses do not distinguish the types of issues raised in the comment letter as our focus is on whether there was an SEC review, not whether that review identified substantive issues. We categorize firms as foreign if they are identified as a foreign private issuer by the SEC. This approach ensures that our analyses focus on firms that are subject to the same SEC reporting and disclosure requirements. We provide a more detailed description of foreign private issuers and their reporting and disclosure requirements in Section 1.3.  Consistent with the substituted compliance framework, we find that the intensity of SEC monitoring is lower when foreign firms originate from countries with more robust private and public enforcement of shareholder rights. We measure private enforcement using the disclosure requirements and liability standards indices developed by La Porta, Lopez-de-Silanes, and Shleifer (2006), as these measures capture the ease with which investors can recover damages in response to misleading disclosures. We measure public enforcement using both the formal public enforcement index developed by La Porta et al. (2006) and the resource-based measures of public enforcement based on the size of the regulator’s budget and staff developed by Jackson and Roe 37  (2009), as these measures capture the strength of the regulatory rules and the resources available to regulators to implement these rules. Our empirical specifications control for the underlying financial condition of the firm and its accounting quality, to partially address the alternative explanation that lower comment letter frequency is attributable to better quality firms rather than stronger home-country enforcement. SEC comment letters indicate that the SEC has reviewed the firm's filings and identified a disclosure issue based on the SEC’s disclosure requirements. If a firm does not have a comment letter, it could be because the firm wasn’t reviewed, or it could be because the firm was reviewed and no disclosure issue was identified. These aspects of the comment letter process suggest that there are three empirical concerns with using the incidence of comment letters to proxy for the intensity of SEC monitoring. First, because disclosure requirements are not necessarily comparable across all firms, it is possible that the variation in home-country accounting and disclosure standards is related to the variation in the issuance of comment letters, conditional on SEC review, even when SEC monitoring is held constant.2  We mitigate this concern with two empirical specifications that incorporate country fixed effects, thus allowing us to control for unobserved country specific factors such as home-country accounting and disclosure standards. The first specification takes advantage of the fact that some foreign firms trade only on US exchanges.3  These firms are subject to a much lower level of, or possibly no, public and private enforcement in their home country. We find that these firms are subject to more intensive SEC monitoring when compared to other firms from the same country                                                  2 If weaker standards are associated with lower levels of compliance, then weaker standards would be associated with a higher incidence of comment letters. However, if weaker standards simply imply reduced disclosure requirements, technical compliance rates for foreign issuers may be higher and there would be fewer SEC comment letters issued. 3 An example of this type of foreign private issuer is Alibaba, which is only listed on the NYSE. 38  that are listed both in the US and abroad, but find no statistically significant difference in SEC oversight between single-listed foreign firms and comparable US firms. These results suggest that differences in SEC oversight across countries are driven by the presence of a foreign regulator, rather than unobserved country-level factors. The second specification uses increases in regulatory staff in 14 European countries to capture changes in enforcement and regulatory oversight. We find that, controlling for the country effects, improvements in oversight by the foreign regulator result in less intensive monitoring by the SEC. Overall, these results provide additional evidence that the home country’s public and private enforcement regimes influence the intensity of SEC monitoring. The second concern with using the incidence of comment letters to proxy for the intensity of SEC monitoring is that we do not observe cases in which the SEC reviews firm disclosures but does not issue a comment letter, either because the reviewer did not identify a disclosure issue or because there were no disclosure issues. We address this concern by repeating our tests using an alternate measure of monitoring intensity that also takes the value of one in those cases where there were no comment letters for a given firm in the preceding three years. We do this because the Sarbanes Oxley Act (SOX) requires the SEC to review each firm at least once every three years. Our results using this alternative measure of SEC monitoring intensity are unchanged, thus providing some assurance that our results are not driven by missing comment letters.  The third concern with using the incidence of SEC comment letters is that it may be a noisy proxy for the intensity of SEC monitoring because some reviews entail substantially more effort and resources than others. The ideal measure of SEC oversight would capture the amount of resources committed by the agency for each review, presumably measured in dollars. Using the incidence of comment letters as a proxy for this underlying construct is primarily a concern if there 39  is a systematic difference in the effort required to conduct a review across countries and this difference in effort affects the number of reviews conducted by the SEC. To mitigate this concern, we repeat our analyses using the length of the comment letter and the number of financial filings covered by the comment letter to better capture the intensity of regulatory effort needed to conduct the review. We also use an alternative measure that identifies whether firms were the recipients of more than one comment letter over a three-year period, to better capture the incidence of discretionary reviews. Across each of these approaches, our conclusions are unchanged. All of these specifications indicate that stronger home country institutions are associated with less intensive monitoring by the SEC, consistent with implicit cooperation across regulators. Overall, our various robustness tests suggest that the incidence of comment letters is a reasonable proxy for the intensity of SEC monitoring because each alternative specification generates similar conclusions. 4  Our use of the incidence of comment letters to proxy for monitoring intensity is also consistent with several other recent studies (e.g., Bens, Cheng and Neamtiu, 2015; Kubick, Lynch, Mayberry and Omer, 2016; Cunningham, Johnson, Johnson and Lisic, 2017; Heese, Khan and Ramanna, 2017). However, even though the incidence of comment letters is well suited to our setting, we do not suggest that the incidence of comment letters is a valid proxy for SEC monitoring intensity in every setting. While the recognition of foreign regulator quality is consistent with the capital formation objective of the SEC and monitoring efficiency, it might be at odds with the countervailing objective of investor protection. Next, we examine whether the SEC focuses its monitoring to provide additional protection to investors on US exchanges. In particular, we examine whether the                                                  4 Our conversations with SEC staff also revealed that the volume of comment letters closely tracks SEC review activity. In other words, when more resources are provided for reviews, more comment letters are produced. 40  relative size of the US listing has an effect on the variation in SEC monitoring based on home-country institutions. Consistent with the idea that the SEC protects US investors, we find that within cross-listed firms from a given foreign home country, SEC monitoring is higher for foreign firms that have a higher fraction of shares listed in the US. Collectively, our analyses suggest that the SEC reduces monitoring intensity when it can rely on the public and private enforcement institutions in the foreign firm’s home country. In contrast, the SEC provides increased monitoring when US investors have greater investment exposure.  We add to these analyses by directly comparing the level of oversight of foreign firms to that of US firms. These analyses are complicated by the fact that US firms are not easily comparable to foreign firms on both economic and financial reporting dimensions. Nonetheless, we undertake these additional analyses because the bonding hypothesis suggests that the foreign firms cross-listed in the US will be monitored like US domestic firms, whereas our analyses suggest that US firms will be monitored more intensively by the SEC because there is not another regulator to rely on. Using both full-sample and matched-sample analyses, we find that, on average, foreign firms are subject to less intensive monitoring by the SEC than comparable US firms. However, when foreign firms are listed only on US exchanges, and little if any foreign monitoring is present, we do not find a statistically significant difference in the level of SEC monitoring between these foreign firms and comparable US firms. Collectively, these results support our conclusion that SEC reduces its monitoring intensity when it can rely on a foreign regulator. Our primary contribution is showing that there is implicit cooperation between the SEC and foreign regulators, as current monitoring by the SEC considers the strength of the home-country private and public enforcement regimes. Our findings also add nuance to the bonding 41  hypothesis. Prior studies have generally found that foreign cross-listed firms on US exchanges exhibit a valuation premium and enjoy a lower cost of capital when compared with similar firms that do not cross-list (Doidge et al., 2004; Hail and Luez, 2006). These studies suggest that these outcomes are consistent with effective monitoring of foreign firms by SEC. While providing direct evidence that SEC monitors foreign firms, we add nuance to these conclusions by documenting that SEC varies its monitoring effort based on attributes of the home country and exposure of US investors rather than viewing all foreign firms as homogenous. In addition, our evidence suggests that bonding may be strengthened when there is more trading of the foreign firm’s shares by US investors, as SEC monitoring increases with trading by US investors. This paper proceeds as follows. In Section 1.2, we summarize the existing literature and present our hypotheses. We summarize the disclosure and reporting requirements for foreign firms in Section 1.3, and present our data collection in Section 1.4. Our research design and empirical results are presented in Sections 1.5. We conclude in Section 1.6. 1.2 Literature review and hypotheses development  Our analysis is motivated in part by the emerging regulatory philosophy of substituted compliance proposed by Tafara and Peterson (2007). The objective of the substituted compliance framework is to enable SEC to achieve its investor protection objective while recognizing improvements made in regulatory monitoring in some foreign countries. Most foreign firms listed in the US are subject to monitoring by their own domestic securities regulator in addition to the SEC. To avoid inefficient duplication of monitoring effort, especially given limited budgets, the SEC could optimally lower its own level of monitoring and leave a share of the monitoring to the foreign firm’s domestic regulator, especially when the institutions in the foreign country are strong. 42  Similarly, the foreign regulator could also reduce its monitoring intensity relative to other local firms not listed on US exchanges in reliance on SEC oversight. A reduction in SEC monitoring is consistent with implicit, rather than explicit, cooperation between the SEC and foreign regulators. Explicit cooperation does not rely on the strength of home-country institutions, but rather is based on agreements between regulators that describe how information is shared.5 To the extent that the SEC implicitly relies on public and private enforcement regimes in other countries, there will be variation in SEC monitoring for foreign firms from different countries. This yields our first hypothesis (stated in the alternative): H1: SEC monitoring intensity will be higher for foreign firms from countries with weak public and private enforcement compared to foreign firms from countries with strong public and private enforcement.  One component of the SEC’s tripartite mission is to protect US investors. When US investors hold large stakes in foreign firms that are cross-listed in the US, the potential losses to US investors are amplified. Therefore, even after controlling for the strength of the foreign firm’s home-country institutions and the quality of the foreign firm’s accounting, there may be additional variation in the level of SEC monitoring based on the exposure of investors on US exchanges. In fact, if the exposure is high enough then the SEC could intensify its oversight regardless of the quality of the home-country regulator. Our second hypothesis (stated in the alternative) is as follows: H2: SEC monitoring intensity will be higher for foreign firms with greater US investor exposure                                                    5 Explicit agreements between regulators typically take the form of bi-lateral or multi-lateral memorandum-of-understanding. Unlike a treaty, a memorandum of understanding is simply a statement of intent (known as “soft law”) and is not enforceable under international law. These memoranda identify the scope, cost, permissible uses, and confidentiality obligations associated with information sharing.  43  Our two hypotheses suggest that there are countervailing forces driving the level of SEC oversight. While SEC may reduce monitoring intensity when the foreign firm has strong home-country institutions, it may increase monitoring intensity when US investors have greater investment exposure. These forces do not necessarily work in isolation. In particular, the combination of high exposure and weak institutions could intensify SEC oversight, whereas the combination of low exposure and strong institutions could weaken SEC oversight. We test for these interaction effects in our empirical analyses in Section 5.  1.3 Institutional Setting and Measurement of SEC Monitoring  Our foreign firm sample consists of firms designated as foreign private issuers by the SEC. In this section, we describe the different types of foreign firms and how they are categorized for reporting and disclosure requirements. We outline these institutional details for two reasons. First, it allows the reader to compare our approach to studying foreign firms with other studies that use international data. Second, we believe these institutional details are helpful in understanding the design and implications of our empirical analyses. Federal securities laws define a foreign issuer as a company incorporated under the laws of any foreign country. A foreign private issuer is a subset of these companies. A foreign private issuer is any issuer incorporated outside the US, unless (i) more than 50 percent of the outstanding voting securities of the issuer are directly or indirectly held of record by US residents; and (ii) any one of the following: (a) the majority of the executive officers or directors of the issuer are US citizens or residents; or (b) more than 50 percent of the assets of the issuer are located in the US; or (c) the business of the issuer is administered principally in the US.6                                                  6 See https://www.sec.gov/divisions/corpfin/internatl/foreign-private-issuers-overview.shtml for more information. 44  In simple terms, a foreign private issuer is a company that the SEC considers to be truly foreign, rather than a US firm merely operating out of a foreign jurisdiction. Under federal securities laws, only foreign private issuers are eligible for regulatory concessions, which include relaxed disclosure requirements. For example, while US issuers must file interim quarterly reports (10-Qs) that contain unaudited financial and other prescribed information, foreign private issuers need only file interim reports that are required under the home country’s laws.  Similarly, foreign private issuers may produce an annual report using either home-country GAAP or IFRS, and provide a reconciliation on Form 20-F. If a foreign issuer does not qualify as a foreign private issuer, it is subject to the same reporting and disclosure requirements as a domestic issuer (i.e., a company incorporated in the US). A brief overview of the disclosure requirements for domestic issuers and foreign private issuers is provided in Appendix A.2.  The disclosure requirements of foreign private issuers that are traded on major US exchanges is independent of the type of security traded—a foreign private issuer may offer any type of security that a US domestic issuer is permitted to offer in addition to using American Depositary Receipts (“ADRs”). An ADR is a negotiable instrument issued by a US depository bank that represents an ownership interest in a specified number of securities that have been deposited with a custodian, typically in the issuer’s country of origin. ADRs can represent one or more shares, or a fraction of a share, and are offered as either “unsponsored” or “sponsored” programs. “Unsponsored” ADR programs are issued by a depository bank without a formal agreement with the foreign private issuer whose shares underlie the ADR. Unsponsored ADRs are only permitted to trade in over-the-counter markets.  “Sponsored” ADRs are depositary receipts that are issued pursuant to a formal agreement, known as a depository agreement, between the depository bank and the foreign private issuer. 45  There are three levels of sponsored ADR programs. Level I ADRs do not involve new capital raising and can only be traded in the US over-the-counter market. In order to establish a Level I ADR, a foreign private issuer must qualify for an exemption under Rule 12g3-2(b) of the Exchange Act. Both unsponsored ADRs and Level I ADRs are exempt from SEC reporting and disclosure requirements. Both Level II and III ADRs are traded on a US exchange, such as the NYSE or NASDAQ, with the difference that Level III involves new capital raising and Level II does not. Both Level II and III are subject to the SEC reporting and disclosure requirements.  Since our sample of foreign firms consists of firms designated by the SEC as foreign private issuers, it excludes (i) global US firms, which are incorporated in the US, but also have listings on foreign exchanges (e.g., IBM) (ii) reverse merger firms, which are incorporated in the US, but headquartered and operate out of a foreign jurisdiction (e.g., China Green Agriculture, as discussed in Lee, Li, and Zhang (2015)), (iii) foreign issuers that are incorporated outside the US but fail to meet the SEC foreign private issuer requirements (e.g., Valeant Pharmaceuticals, a Canadian firm that has executive offices in the US and therefore is treated as a domestic issuer by the SEC), and (iv) foreign firms with unsponsored shares or Level 1 ADRs that trade on the OTC markets because the firms are exempt from reporting requirements by the SEC (e.g., Vestas, as discussed in Iliev, Miller, and Roth (2014)). The underlying construct that we are trying to capture in our empirical tests is the total resources expended by the SEC on monitoring activities, and we proxy for this underlying construct using the incidence of SEC comment letters issued in response to annual and interim financial statement filings of a firm. 7  Comment letters are issued by the SEC's Division of                                                  7 We exclude comment letters on other filings such as prospectuses for capital issues. 46  Corporation Finance in response to a periodic review of a firm's annual financial statements and related filings. These letters contain requests from the SEC to the firm to provide additional information, modify the submitted filing, or alter future filings. The SEC allocates considerable resources to the enforcement of disclosure standards. The SEC’s 2006 Audit Report #401 states that about $125 million were allocated to the Division of Corporate Finance that has 515 staff, of which 80 percent are assigned to review filings.  A few recent studies have used comment letter data, although not as a proxy for the intensity of SEC monitoring. Rather, these studies have examined capital market consequences of the comment letter process by identifying those comment letters with substantive content. Johnston and Petacchi (2016) find that comment letter resolutions are associated with a better information environment and less disagreement among investors and analysts. Dechow et al. (2015) document that insider trading is significantly higher than normal levels prior to the public disclosure of SEC comment letters relating to revenue recognition. Ryans (2015) uses classifications based on textual analysis to identify important comment letters, and shows that these letters are associated with lower future performance and undisclosed financial reporting deficiencies. We do not attempt to distinguish between different types of comment letters because we believe that the intensity of SEC monitoring is reflected in whether the SEC reviewed the firm’s filings, not whether the SEC found material issues when it reviewed the firm’s filings.  While not a perfect measure, we suggest that comment letters provide a reasonable proxy for the level of SEC monitoring because a very high percentage of SEC reviews result in a comment letter.8 In addition, our discussions with SEC staff indicated that an increase in SEC                                                  8 The percentage of reviews that produce a comment letter has declined in recent years, which may be due to a shift in the focus of the SEC to only issue comment letters in response to material issues. For example, see http://www.auditanalytics.com/blog/comments-pending-companies-without-recent-comment-letters/.   47  monitoring intensity would entail an increase in the number of reviews, and that increasing the number of reviews would also increase the number of comment letters issued, thus indicating that there is a strong positive correlation between our measure and the underlying construct. Nonetheless, there are several potential shortcomings associated with our measure. In particular, not all reviews require the same amount of SEC resources to complete, and not all SEC reviews generate comment letters, either because the SEC did not identify an issue or because there was no issue with the firm’s disclosures. We believe that these shortcomings generate noise in our empirical proxy, rather than a systematic bias that varies across countries in tandem with the strength of those country’s regulatory institutions. To provide some assurance that this is the case, we present in Section 5 the results of a series of robustness tests that are designed to mitigate this concern. 1.4 Sample and data We start with the list of foreign private issuers published by the SEC for each year from 2004 through 2012.9 We match foreign private issuers identified by the SEC to Compustat by manually comparing firm names. Of the 1,085 firms on the SEC lists of foreign private issuers, there are 168 firms that are not on Compustat. We exclude these firms because we do not have the necessary data to generate the variables used in our analyses. We restrict the sample period to firm years with fiscal year ends after August 1, 2004 and before December 31, 2012 because this date range corresponds with the availability of comment and response letter data, which the SEC began to release for disclosure filings made after August 1, 2004.10                                                   9 https://www.sec.gov/divisions/corpfin/internatl/companies.shtml 10 https://www.sec.gov/news/press/2005-72.htm 48  The sample period spans two major SEC rule changes related to foreign cross-listed firms.  First, effective November 15, 2007, the SEC eliminated the 20-F reconciliation to U.S. GAAP for foreign registrants preparing financial statements in accordance with IFRS. Second, on September 5, 2008, an amendment to Rule 12g3-2 allowed foreign firms to be cross-listed involuntarily (Iliev et al., 2014). To ensure that our results are not influenced by the 20-F reconciliation change, we conduct a robustness test using only data for the period prior to the elimination of the 20-F reconciliation, and separately using only data for the period after the elimination of the 20-F reconciliation. In untabulated analyses, we find that our conclusions are unchanged. We do not believe our results are significantly influenced by Rule 12g3-2, as we focus our analyses on Foreign Private Issuers, which were not directly affected by this rule change. The number of foreign private issuer firm-year observations with available data on Compustat is 4,808.  Next, we search for comment and response letters on the Audit Analytics database for the period August 1, 2004 through December 31, 2012. This search produces 13,555 comment letters that are in response to a financial filing. For foreign firms, this means that the comment letter was in response to either a 20-F filing or a 6-K filing that contained interim financial reports, which we identified through manual inspection.11 For each of these letters, we extract the firm’s CIK, the subject of the letter, and the date of the corresponding financial filing. We collect control variables from Compustat and CRSP by first matching the CIK from our search of Audit Analytics to GVKEY (Compustat) and PERMNO (CRSP). All control variables are defined in Appendix A.1. These variables control for aspects of the firm’s accounting quality, attributes of the firm’s auditor, and other aspects of the firm’s financial condition that prior research has suggested may be                                                  11 Eliminating the small number of comment letters for 6-K filings and focusing exclusively on 20-F comment letters does not influence our results or change any of our conclusions. 49  associated with comment letter frequency (e.g., Cassell, Dreher and Myers, 2013). The resulting sample contains 4,808 unique firm-years, including 1,500 firm-years in which one or more filings by the foreign firm resulted in a comment letter. In Table 1, we report the descriptive statistics of variables of interest and control variables for the final sample of foreign-firm years.  We next identify each foreign firm’s home country using the location of the exchange where the firm’s non-US shares are traded. When a given firm trades on multiple non-US exchanges, we select the country with the exchange with the highest trading volume measured in US dollars. Information on the various exchanges and the trading volume on each exchange were collected from Capital IQ. We use the location of the principle non-US exchange, rather than simply relying on the location of the firm’s headquarters, because we expect that the local securities regulator will focus on those firms that are traded on the local exchange, rather than those firms with only a local physical presence.12  We believe that the location of the firm’s headquarters is well suited to capturing the intensity of other types of oversight. For example, environmental compliance or labor laws are likely to be driven by firm’s physical presence. A subset of our analyses uses foreign private issuers who are only listed on a US exchange. We identify these firms using data retrieved from BNY Mellon, which has researched and identified those firms where the US listing is the only listing for that foreign firm.13 There are 84 firms (402 firm-year observations) in our sample that are identified by BNY Mellon as single-listed firms.                                                  12 In untabulated results, we find that our conclusions are unchanged when we conduct our analyses using the foreign firm’s headquarters instead of the location of the primary non-US exchange, suggesting that there is little difference across the two measures of home country. We conjecture that part of the reason the results are similar when we use the firm’s headquarters is because the location of the primary non-US exchange and the firm’s headquarters are the same country for approximately three-quarters of foreign firms in our sample.   13 https://www.adrbnymellon.com/ 50  The distribution of firm-year observations by home country, where home country is measured using either the location of the primary exchange or the location of the firm’s headquarters, is shown in Table 1.2. Using the location of the primary non-US exchange rather than the location of the firm’s headquarters has a significant effect on two countries—China and Germany. The number of firm-year observations where China is the home country regulator decreases from 487 to 22. This occurs because most of the Chinese firms are also listed on exchanges outside of China, such as Hong Kong, and the majority of their shares trade on these exchanges. In contrast, the number of firm-year observations where Germany is the home-country regulator increases from 88 to 830. This occurs because many firms, particularly those in Europe, have their primary listings in Germany. The changes in the number of firm-year observations are generally modest for the other countries in our sample. Approximately half of the countries in our sample have a small number of firm-year observations. In robustness tests, we confirm that our results are unchanged if countries with less than 30 firm-year observations are excluded from our analyses. In addition to firm level data, we also collect data on attributes of the home country’s institutions for the foreign firms in our sample. We collect country level indices for private enforcement (disclosure requirements and liability standards) and formal public enforcement from La Porta et al. (2006). We also collect resource based indices of public enforcement from Jackson and Roe (2009). We collect two resource based measures, budget and staff. The budget based measure equals the natural log of the security regulator’s 2005 budget, and the staff based measure equals the security regulator’s 2005 staff headcount divided by the country’s population in millions. We use the country-specific growth in staff reported by Christensen et al (2016) to identify countries that experienced growth in regulatory staff during the sample period.  51  The data by country on the percentage of firm years with comment letters and measures of public and private enforcement are summarized in Table 1.3. Firms with a primary non-US listing in Canada and Germany comprise the largest proportion, followed by firms primarily traded on exchanges in Israel and the United Kingdom. There is variation in comment letter frequency across countries. For example, Canada has a comment letter frequency for financial filings of only 26 percent, compared with 50 percent for Italy. As discussed in more detail in La Porta et al. (2006) and Jackson and Roe (2009), there is variation in the country level indices that capture public and private enforcement. For example, the disclosure requirements index is 0.67 and the liability standards index is 0.22 in Italy, suggesting that private enforcement is relatively weak in Italy. In contrast, the disclosure requirements index is 0.92 and the liability standards index is 1.00 in Canada, suggesting that private enforcement is relatively strong in Canada. 1.5 Research Design and Results 1.5.1 Association between Home-Country Institutions and SEC Monitoring We examine whether there is variation in the intensity of SEC monitoring across foreign firms based on the strength of public and private enforcement in the foreign firm’s home country using the following specification: SEC_Reviewj,t = β0 +  β1 Enforcementc + ∑βj Controlsj,t   + Ind_FE + Year_FE + εj,t  (1) The coefficient of interest in equation (1) is β1, which measures how SEC monitoring of foreign firms is related to the quality of home-country enforcement. Our first hypothesis implies that there is implicit sharing of monitoring between the SEC and foreign institutions, with the result that SEC monitors foreign firms from countries with strong public and private enforcement 52  less than those from countries with weak public and private enforcement. This implies that the coefficient β1 < 0.  The dependent variable, SEC_Reviewj,t, is an indicator variable set equal to 1 if firm j received a comment letter for a period t filing, and 0 otherwise. The variable of interest in equation (1) is Enforcement, which we proxy for using five different measures for the strength of the home country’s public and private enforcement regime. We measure private enforcement using the disclosure requirements and liability standards developed by La Porta et al. (2006), as these measures capture the ease with which investors can recover damages when firm disclosures are false or misleading. We measure public enforcement using the formal public enforcement index developed by La Porta et al. (2006) and two resource-based measures of public enforcement developed by Jackson and Roe (2009), as these measures capture the strength of the regulatory rules and the resources available to regulators to implement these rules. The two resource based measures of public enforcement are the natural log of the foreign regulator’s budget and the foreign regulator’s staff headcount scaled by the country’s population in millions. Definitions of each index used to proxy for Enforcement are provided in Appendix A.1 Panel B. We control for a comprehensive set of variables that prior literature has argued are associated with comment letter issuance. In broad terms, the prior literature has found that disclosure and financial reporting quality, the quality of the firm’s auditor, and the financial condition of the firm are associated with comment letter frequency (e.g., Cassell et al., 2013). Cassell et al., (2013) argue that material weaknesses and restatements signify previous accounting failures and therefore are more likely to attract SEC scrutiny. We therefore control for Material_Weakness and Restatement. In addition, we control for accounting quality using Small_NI, and IFRS. The choice of these control variables is consistent with other cross-country 53  studies (e.g., Lang, Raedy, and Wilson, 2006) and reflect data availability considerations in a sample of foreign firms. In addition to capturing aspects of accounting quality, IFRS also indicates whether the firm is reporting under an accounting standard that could either enhance or limit the SEC’s ability to generate comments relative to US GAAP. We control for the quality of the firm’s auditor using Auditor_Big4, Auditor_2Tier, Auditor_Tenure, Auditor_Dismiss, Auditor_Resigned. We control for aspects of the firm’s financial condition and information environment using RetVol_High, Market_Cap, Firm_Age, Sales_Growth, Altman_Z, Ext_Financing, Restructuring, M&A, Litigation_Risk, Loss, Num_Segments, Institutional_Holding. We control for the economic condition of the country using GDP. Each of these variables is defined in Appendix A.1, Panel C. Equation (1) also includes year and industry fixed effects, where we classify industry based on Fama-French 17-industries using SIC codes from Compustat. We estimate equation (1) using an OLS specification because nonlinear models tend to produce biased estimates in panel data sets with a short time series and many fixed effects, leading to an incidental parameters problem and inconsistent estimates (see e.g., Ai and Norton, 2003). The standard errors are clustered by country-industry groups. To ensure that our results are not sensitive to this research design choice, we repeat our analyses using both logit and probit specifications, and obtain the same conclusions. The results from equation (1) are provided in Table 1.4. The coefficients on each of the measures of public and private enforcement are negative and statistically significant (t-statistics range from 2.693 to 5.374), indicating that SEC monitoring is lower when the firm originates from a country with strong enforcement. The economic magnitude of Enforcement in each of these specifications is also economically meaningful. For example, the 25th percentile of the Disclosure Requirements Index is 0.42 and the 75th percentile is 0.75, indicating that a movement from the first to the third quartile of the Disclosure Requirements Index reduces the likelihood a comment 54  letter by approximately 4.9 percentage points.14 Overall, the results in Table 1.4 indicate that the SEC is at least partially relying on the foreign firm’s home-country regulator in setting its monitoring intensity, consistent with our first hypothesis. In untabulated analyses, we find similar results if we combine our measures of enforcement using principal component analysis and use the first component as our variable of interest. There are three empirical concerns with using the incidence of comment letters to proxy for the intensity of SEC monitoring. First, because disclosure requirements are not necessarily comparable across all firms, it is possible that variation in the issuance of comment letters is related to variation in accounting and disclosure standards to which the foreign firm is subject, not to variation in SEC monitoring. Second, we have partial observability because we do not observe cases in which the SEC reviews the firm disclosures but does not issue a comment letter, either because an issue was missed by the SEC or because there was no issue with the firm’s filings. Third, simply using the incidence of SEC comment letters may be a noisy proxy for the intensity of SEC monitoring because some reviews entail significantly more resources than others. We conduct additional analyses to address each of these concerns in the following subsections. 1.5.1.1 Analyses with Country Fixed Effects  A potential concern with the results in Table 1.4 is that the home-country enforcement metrics could be capturing some underlying country characteristics. To the extent that these country characteristics are responsible for strong regulatory institutions and are associated with better quality firms, we would be spuriously attributing reduced SEC monitoring to the quality of                                                  14 The coefficient on Disclosure Requirements is -0.149 (t-statistic = 4.171). Therefore, we calculate the reduction in likelihood as (0.75 -  0.42) x -0.149 = -0.0492. 55  regulatory institutions rather than firm quality in specific countries. We address this concern by conducting two sets of tests that examine variation in SEC monitoring within country.  In the first test, we focus on a subset of countries where there was a documented increase in the regulatory staff budget, and test whether this increase is associated with a corresponding decrease in SEC monitoring intensity. The specification we use is as follows: SEC_Reviewj,t = β0 +  β1 Post * Staff_Growthc + ∑βj Controlsj,t   + Ind_FE + Year_FE   + Country_FE + εj,t       (2) The coefficient of interest in equation (2) is β1, which identifies whether changes in SEC monitoring of foreign firms is associated with changes in public enforcement. Because equation (2) includes country fixed effects, the coefficient β1 captures the differential effect by comparing firms from the same country that are subject to different levels of enforcement, thus mitigating concerns related to unobservable country factors. To the extent that the SEC relies on the strength of the home-country regulatory quality, increases in the strength of that regulatory quality should be correlated with decreases in SEC monitoring intensity. This implies that the coefficient β1 < 0. Staff_Growth is a binary variable, taken from Table 1.4 in Christensen et al (2016) that identifies whether there was an increase in staff resources from 2005 to 2008. Post is a binary variable that takes the value 1 for the years 2008 and later, which corresponds with the years when the increase in enforcement became effective. The main effects for both Post and Staff_Growth are subsumed by the year and country fixed effects, respectively. This analysis is restricted to 14 countries, since data in Christensen et al. (2016) is only for European countries. Of these countries, Christensen et al. (2016) report that 7 experienced an increase in staff resources and 7 did not.  The results in Column (1) of Table 1.5 show that the coefficient on Staff_Growth is negative and statistically significant (coefficient = -0.119, t-statistic = 2.05), consistent with our first hypothesis and the 56  main results in Table 1.4. This coefficient indicates that an increase in staff resources reduces the likelihood a comment letter by approximately 11.9 percentage points In the second test, we create variation in home-country monitoring within a given country by distinguishing between foreign private issuers from that country that are listed in both the home country and the US with foreign private issuers from that country that are only listed in the US. We use the following empirical specification: SEC_Reviewj,t = β0 +  β1 Single_Listedj + ∑βj Controlsj,t   + Ind_FE + Year_FE   + Country_FE + εj,t       (3) The coefficient of interest is β1, which identifies whether firms that are listed only on a US exchange are subject to a differential level of monitoring when compared to firms from the same country that are also listed on the foreign exchange. Single_Listed is a binary variable that takes the value of 1 for foreign private issuers who are listed only in the US. Since foreign firms not listed on a foreign exchange do not necessarily have another securities regulator overseeing the firm’s financial disclosures, we expect that foreign private issuers listed only in the US will be subject to higher intensity monitoring by SEC. The results in Column (2) of Table 1.5 are consistent with this prediction, as the coefficient on Single_Listed is positive and significant (coefficient = 0.053, t-statistic = 1.736). In economic terms, the coefficient on Single_Listed indicates that foreign firms listed only on US exchanges are 5.3 percentage points more likely to receive comment letters relative to other cross-listed firms from the same country. Overall, the results in Table 1.5 mitigate the concern that our conclusions from Table 1.4 are driven by unobservable country specific factors.  57  1.5.1.2 Partial Observability of SEC Monitoring Our measure of SEC monitoring is based on the presence of a comment letter issued by the SEC. This approach results in partial observability because sometimes there is an SEC review without a comment letter being issued. To the extent that the likelihood of receiving a comment letter conditional on an SEC review is related to firm quality, our inferences could be affected. In particular, the documented relation between SEC monitoring intensity and the quality of home-country regulation may arise because firms from weaker regulatory regimes are of poorer quality, rather than the implicit sharing of monitoring effort by the SEC. As a result, these firms would be more likely to receive a comment letter, conditional on SEC review. We address this concern by constructing an alternative measure of SEC monitoring which assumes that a review is completed every three years. We do this because the Sarbanes Oxley Act (SOX) requires the SEC to undertake some sort of review of every firm at least once in three years, which could involve a partial review of specific disclosures to full-fledged review of the entire filing. Therefore, if a firm does not have a comment letter during a 3-year period, it is likely that there was an SEC review but that no issues were discovered as part of that review. We re-estimate equation (1) replacing the dependent variable with SEC_Review_Alt, which is set equal to 1 for period t when (a) a firm receives a comment letter for a period t filing or (ii) when firms have not received a comment letter for the period t-2 through period t filings, and 0 otherwise.  The results reported in Table 1.6 are consistent with those in Tables 1.4 and 1.5. We continue to find that foreign firms with stronger home country enforcement were less likely to receive a comment letter, conditional on an SEC review. The coefficients in Table 1.6 are smaller than those in Table 1.5. For example, coefficient on the Disclosure Requirements index drops from 0.149 to 0.105. This suggests that there are slightly more SEC reviews that do not produce a 58  comment letter in countries with strong enforcement, consistent with the notion that firms from countries with strong regulatory regimes are of higher quality. However, since each coefficient continues to be negative and both statistically and economically significant, the results in Table 1.6 continue to support our conclusion that there is implicit cooperation by the SEC with foreign regulators. Taken together, the results in Tables 1.5 and 1.6 give us confidence that our results are unlikely to be entirely explained by unobservable country quality or firm quality and provide robust evidence that home-country regulatory quality affects SEC oversight of foreign firms.  1.5.1.3 Alternative Measures of SEC Monitoring Our main analyses use the incidence of comment letters as a proxy for SEC monitoring intensity. One potential issue with this approach is that it does not capture variation in the effort required by the SEC to undertake each review. In addition, many of the observed comment letters may be due to the triennial review cycle required by SOX, rather than an active decision to increase monitoring of a particular firm. We use three variations in how we measure the intensity of SEC monitoring to mitigate these concerns.  First, we replace SEC_Review in equation (1) with SEC_Review_Words, a variable that equals the number of words in the first comment letter issued as part of a dialogue between the SEC and the firm. Second, we replace SEC_Review in equation (1) with SEC_Review_Filing, a variable that equals the number of financial filings with comment letters. Both of these variables better capture the effort required to conduct a review, as longer comment letters or those that cover more financial filings likely took more effort to prepare. Third, we replace SEC_Review in equation (1) with SEC_Review_2, an indicator variable that takes the value 1 if the firm received a comment letter in year t and at least one comment letter in the previous two years. This variable better captures the discretionary review that is beyond the required level. The results of all three 59  specifications are provided in Table 7. Each specification confirms the results in Table 1.4. The coefficients on each measure of enforcement in Panels A through C are negative and statistically significant. In addition, the economic significance of the results is comparable to those in Table 1.4. Overall, these analyses suggest that our results are robust to variations in the measurement of SEC monitoring intensity. 1.5.2 Association between US Investor Exposure and SEC Monitoring Our second hypothesis predicts that SEC monitoring intensity will be higher for foreign firms with greater US investor exposure. We test this hypothesis using the following specification: SEC_Reviewj,t = β0 +  β1 US Exposurej,t +  ∑βj Controlsj,t   + Industry_FE  + Year_FE + Country_FE + εj,t             (4) The coefficient of interest is β1, which identifies whether firms with greater US exposure are subject to a differential level of monitoring. The inclusion of country fixed effects ensures that this comparison is across firms within the same country. US_Exposure is equal to the percent of the firm’s market capitalization that is traded on US exchanges, where data on the total and US market capitalization is obtained from either Bloomberg or by manually inspecting the firm’s 20-F filing. The results are reported in Table 1.8 Panel A. Column (1) reports the estimates for equation (4) without country fixed effects and column (2) includes the fixed effects. In both specifications, the likelihood of a comment letter is increasing in the market capitalization of the firm, and in the percent of that market capitalization that is held by US investors. In economic 60  terms, moving from the first to the third quartile in US_Exposure is associated with a 2.2 percentage point increase in the likelihood of a comment letter being issued for the firm. Next, we examine how US Exposure and Enforcement interact to influence SEC oversight. For ease of exposition and interpretation, we divide our sample into terciles using independent sorts based on the dollar amount of US exposure (i.e., the product of US_Exposure and MarketCap) and each metric of home country enforcement, and then construct two-way comparisons of the high and low groupings across each grouping using a modified version of Equation (4) in which we allow for interaction between indicator variables capturing the high and low groups. The coefficients from these analyses are reported in Table 1.8 Panel B. The values within each grouping (e.g., low-low) represent the difference in monitoring intensity between that group of firms and the group of firms in the middle tercile. For example, the low-low value of 0.004 indicates that the likelihood of a comment letter is 0.4 percentage points higher for firms with US exposure and enforcement in the bottom tercile relative to firms in the middle tercile. Overall, the results in Table 1.8, Panel B suggest that the strength of the foreign regulator influences the way in which US trading is incorporated into the SEC’s level of monitoring intensity. In particular, we find that SEC monitoring is significantly lower for firms with low US exposure and high quality home-country institutions compared with firms in the middle tercile. This is the group in which the benefit from sharing monitoring efforts is high without a significant sacrifice of US investor protection. In contrast, SEC monitoring is significantly higher for firms with high US exposure and low quality home country institutions compared with firms in the middle tercile, consistent with the notion that these firms warrant the most oversight by the SEC. In general, we do not find significant differences in the intensity of SEC monitoring when we 61  compare firms that are high in both US investor exposure and enforcement with those that are low in both US investor exposure and enforcement.  Collectively, our analyses show that there are two countervailing forces. On the one hand, the SEC reduces monitoring intensity when it can rely on the public and private enforcement institutions in the foreign firm’s home country. On the other hand, the SEC provides increased monitoring when US investors have greater investment exposure in the foreign firms.  1.5.3 Comparison of Foreign Firms with US Firms To the extent that the SEC shares monitoring effort with foreign regulators, US firms may be subject to more intensive SEC oversight than foreign firms because the SEC may be the sole regulator for these firms. Therefore, we provide additional evidence to support our first hypothesis by comparing SEC oversight for foreign relative to US firms. We use the following specification:  SEC_Reviewj,t = β0 +  β1Foreign_Firmj + ∑βj Controlsj,t     + Industry_FE + Year_FE + εj,t     (5) We estimate this specification using both a full sample of US firms as well as two matched samples. In the first matched sample, we match each foreign firm to a US firm using an exact match based on year and industry (Fama-French 17-industries), and then choose the US firm that is closest in size based on total market cap to the foreign firm. In the second matched sample, we match each foreign firm to a US firm using an exact match based on year, industry (Fama-French 17-industries) and Small_NI status (a measure of accounting quality), and then choose the US firm that is closest in size based on total market cap to the foreign firm. The inclusion of the additional criterion based on Small_NI status in the second matched sample is to reduce variation across firms in accounting quality. Foreign_Firm is a binary variable that takes the value of 1 for foreign firms, 62  and 0 for US firms. The results in Table 1.9 show that foreign firms are subject to a lower level of monitoring than comparable US firms. The coefficient on Foreign_Firm is negative and statistically significant in each specification (t-statistics range from 4.634 to 5.803), and in economic terms it indicates that the likelihood of receiving a comment letter is 8.0 to 12.5 percentage points lower for foreign firms compared to US firms. This result is consistent with our first hypothesis which states that when SEC can share its monitoring effort with another regulator, it does not monitor as much. Next, we focus on the subsample of foreign private issuers that are only listed on US exchanges to provide additional support for this conclusion. Consistent with our first hypothesis, we predict that the SEC will monitor these firms to the same extent as they do US firms because these firms are not subject to oversight by another securities regulator. The results are reported in Table 1.9 Panel B. The coefficient on Single_Listed is insignificant in both specifications (coefficient = -0.015 with t-statistic = 0.841 in column (1); coefficient = -0.012 with t-statistic = 0.692 in column (2)), indicating that there is not a statistically significant difference in the intensity of SEC monitoring between foreign private issuers that are only listed on US exchanges and domestic issuers. There are several challenges in comparing US firms with foreign firms in the context of SEC oversight. For example, as described in Section 1.3, foreign private issuers are subject to some reporting exemptions and therefore to the extent that there are fewer filings made by foreign firms, there may be fewer comment letters. We conduct several robustness tests to mitigate the concern that these differences are driving our results. First, we use the classification of comment letters into topics as identified by Audit Analytics to narrow our sample to topics that are most likely to be equally relevant to both US and foreign firms. More specifically, we conduct one 63  analysis where the dependent variable covers all topics related to disclosure including MD&A, disclosures about risk, internal controls, and accounting rules, and a second analysis where the dependent variable focuses more narrowly on comment letters that discuss accounting rules and accounting disclosures. In untabulated analyses, we find similar results to Table 1.9 Panel A for both sets of analysis. We also restrict the sample of foreign firms to those that provide quarterly reporting. In untabulated analyses, we find that the periodicity of financial statement filings does not appear to drive our results.  1.6 Conclusion Using the existence of an SEC comment letter as a proxy of SEC monitoring, we find significant variation in the level of SEC monitoring across foreign firms based on the strength of their home-country public and private enforcement and the level of US investor exposure. Because comment letters are only issued in response to a disclosure issue, we use a number of alternative specifications to mitigate the concern that shortcomings in our empirical approach are influencing our conclusions. We use country fixed effect specifications, which document differences in SEC monitoring between foreign single-listed and foreign cross-listed firms, to mitigate the concern that our results are due to a country level omitted variable such as local accounting standards. We also use a specification which documents that changes in foreign regulatory intensity are associated with changes in SEC regulatory intensity. In addition, we use a number of different approaches to generate our proxy for SEC monitoring intensity, including adjustments for SEC reviews that did not generate a comment letter and adjustments to reflect the length and content of comment letters, to ensure that our conclusions are not driven by differences in measurement error in our proxy across countries. Collectively, our various robustness tests suggest that the incidence of comment 64  letters is a reasonable proxy for the intensity of SEC monitoring and our conclusions are unchanged across each specification. Overall, our findings add nuance to the arguments used in the bonding literature by recognizing that SEC may not view all foreign firms as homogenous and thus SEC may vary its monitoring effort based on characteristics of the home country institutions. One consequence of the implicit sharing of regulatory duties we document is that foreign firms are, on average, subject to less intensive monitoring than comparable US firms. This result suggests that SEC is taking a calculated risk by relying on strong foreign institutions to protect US investors. While we find that SEC appears to be optimizing its effort in seemingly efficient ways, it is not clear if this strategy is optimal. We call on future research to evaluate whether this strategy achieves optimal monitoring or whether it imposes some costs on US investors in terms of poorer reporting quality for foreign firms. More broadly, our results suggest that we need to rethink the role of country level regulators in a global economy with increasing numbers of global firms. To the extent that country level regulators implicitly share regulatory duties with foreign counterparts, coordination across jurisdictions will be increasingly important. Traditionally, the resources allocated to the SEC have ignored the resources allocated by foreign governments to their local regulator. However, to the extent that the SEC shifts regulatory duties to these regulators, it suggests that the resources allocated to the SEC should reflect this practice. In addition, while foreign firms from weak jurisdictions benefit the most from a US listing, the regulatory cost associated with these firms is also the highest, suggesting that perhaps the cost of listing in the US should vary based on the strength of the home country regulator.  65  1.7 Tables Table Error! No text of specified style in document..1 Sample Composition   This table provides the descriptive statistics for each of the variables used in our analyses. Each variable is defined in Appendix A.1  Variable N Mean P25 P50 P75 SEC_Review 4,808 0.312 0 0 1 US_Exposure 4,191 0.358 0.110 0.285 0.564 Controls for Accounting Quality Small_NI 4,808 0.08 0 0 0 IFRS 4,808 0.239 0 0 0 Material_Weakness 4,808 0.17 0 0 0 Restatement 4,808 0.103 0 0 0 Auditor Related Controls     Auditor_Big4 4,808 0.929 1 1 1 Auditor_2Tier 4,808 0.02 0 0 0 Auditor_Tenure 4,808 4.131 2 4 6 Auditor_Dismiss 4,808 0.084 0 0 0 Auditor_Resigned 4,808 0.023 0 0 0 Other Controls      RetVol_High 4,808 0.243 0 0 0 MarketCap 4,808 10,776 227 1,432 10,541 Firm_Age 4,808 9.215 6 9 12 Loss 4,808 0.422 0 0 1 Sales_Growth 4,808 0.32 0.039 0.145 0.310 Bankruptcy 4,808 5.66 4 6 8 Num_Segments 4,808 1.475 1 1 1 Ext_Financing 4,808 0.01 -0.029 0 0.018 Restructuring 4,808 0.308 0 0 1 M&A 4,808 0.154 0 0 0 Litigation Risk 4,808 0.311 0 0 1 Inst_Ownership 4,808 0.133 0.004 0.045 0.200         66  Table Error! No text of specified style in document..2 Location of Primary Exchange relative to Headquarters   This table provides a breakdown of the sample of foreign firms and the location of the firm’s headquarters and the primary non-US exchange.  Country Firm years based on Exchange Firm years based on Headquarters United Arab Emirates 1 1 Argentina 102 102 Australia 101 94 Austria 7 3 Belgium 17 15 Belize 0 1 Bermuda 0 76 Brazil 97 97 Canada 1,114 1,148 Switzerland 0 12 Chile 107 107 China 158 580 Colombia 5 5 Cayman Islands 0 14 Cyprus 0 1 Germany 1,019 98 Denmark 15 18 Spain 40 42 Finland 15 15 France 127 131 United Kingdom 328 294 Greece 51 126 Hong Kong 0 47 Hungary 6 7 Indonesia 18 18 India 84 104 Ireland 0 76 Israel 490 572 Italy 62 54 Japan 185 187 South Korea 76 85 Luxembourg 0 38 Monaco 0 7 Mexico 178 149 Marshall Islands 0 8 Netherlands 116 152 Norway 53 17 New Zealand 8 10 Panama 0 13 Peru 19 19 Philippines 13 13 Papua New Guinea 0 6 Portugal 11 10 Russian Federation 23 30 Singapore 0 21 Sweden 21 13 Turkey 9 9 Taiwan 68 93 Venezuela 3 3 South Africa 61 67 67  Total 4,808 4,808 68  Table Error! No text of specified style in document..3 Sample Composition by Country   Country Firm years Percent Comment Letters Disc Req. Liability Standard Rules Enforc. Budget Resources  Staff Resources United Arab Emirates 1 0.00% . . . . . Argentina 102 35.29% 0.50 0.22 0.58 15,994  3.46 Australia 101 31.68% 0.75 0.66 0.90 89,217  34.44 Austria 7 42.86% 0.25 0.11 0.17 34,464  9.97 Belgium 17 41.18% 0.42 0.44 0.15 27,276  13.76 Brazil 97 36.08% 0.25 0.33 0.58 31,729  2.68 Canada 1,114 26.21% 0.92 1.00 0.80 82,706  38.93 Chile 107 37.38% 0.58 0.33 0.60 66,093  9.93 China 158 24.05% . . . . . Colombia 5 40.00% 0.42 0.11 0.58 42,660  3.94 Germany 1,019 29.05% 0.42 0.00 0.22 12,903  4.43 Denmark 15 66.67% 0.58 0.55 0.37 25,940  10.85 Spain 40 42.50% 0.50 0.66 0.33 29,873  8.5 Finland 15 53.33% 0.50 0.66 0.32 45,937  11.23 France 127 37.80% 0.75 0.22 0.77 28,851  5.91 United Kingdom 328 42.07% 0.83 0.66 0.68 80,902  19.04 Greece 51 29.41% 0.33 0.50 0.32 60,111  12.16 Hungary 6 50.00% . . . 79,996  10.75 Indonesia 18 33.33% 0.50 0.66 0.62 5,576  1.97 India 84 39.29% 0.92 0.66 0.67 . 0.43 Israel 490 23.47% 0.67 0.66 0.63 145,673  18.78 Italy 62 50.00% 0.67 0.22 0.48 61,239  7.25 Japan 185 32.43% 0.75 0.66 0.00 15,754  4.32 South Korea 76 39.47% 0.75 0.66 0.25 80,192  11.55 Mexico 178 28.09% 0.58 0.11 0.35 49,864  5.19 Netherlands 116 35.34% 0.50 0.89 0.47 131,285  23.53 Norway 53 37.74% 0.58 0.39 0.32 25,109  20.78 New Zealand 8 12.50% 0.67 0.44 0.33 37,539  8.95 Peru 19 31.58% 0.33 0.66 0.78 108,353  5.32 Philippines 13 30.77% 0.83 1.00 0.83 65,848  4.29 Portugal 11 72.73% 0.42 0.66 0.58 75,562  14.5 Russian Federation 23 34.78% . . . . . Sweden 21 52.38% 0.58 0.28 0.50 21,988  7.19 Turkey 9 44.4% 0.50 0.22 0.63 58,893 6.17 Taiwan 68 30.9% 0.75 0.66 0.52 44,090 12.53 Venezuela 3 33.3% 0.17 0.22 0.55 . . South Africa 61 49.2% 0.83 0.66 0.25 49,291 3.52  69  Table Error! No text of specified style in document..4 Variation in SEC Monitoring based on Home Country Enforcement This table examines whether the intensity of SEC monitoring is different across foreign firms based on the strength of home country enforcement. The dependent variable is SEC_Review, an indicator variable set equal to 1 if the firm received a comment letter for a period t filing, and 0 otherwise. The five measures of home country enforcement are described in Appendix A.1. A negative (positive) coefficient on any of the proxies for the strength of home country enforcement indicates that foreign firms with stronger home country enforcement were subject to lower (higher) SEC monitoring. The control variables are listed and defined in Appendix A.1.  Each of the continuous variables is winsorized at 1% and 99% to mitigate outliers. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry.              (1) (2) (3) (4) (5) Private Enforcement      Disclosure_Requirements -0.149***      (-4.171)     Liability_Standards  -0.112***      (-5.305)    Public Enforcement      Rules_Enforcement   -0.132***      (-3.515)   Staff_Resources    -0.003***      (-5.374)  Budget_Resources     -0.034***      (-2.693) Controls for Accounting Quality     Small_NI 0.077*** 0.080*** 0.072*** 0.078*** 0.084***  (3.098) (3.251) (2.885) (3.091) (3.449) IFRS 0.020 0.020 0.029 0.025 0.023  (1.028) (1.026) (1.446) (1.263) (1.117) Material_Weakness 0.025 0.024 0.026 0.026 0.031*  (1.494) (1.414) (1.562) (1.574) (1.820) Restatement 0.015 0.017 0.015 0.010 0.014  (0.784) (0.855) (0.738) (0.473) (0.710) Auditor Related Controls      Auditor_Big4 -0.044 -0.039 -0.043 -0.039 -0.047  (-1.286) (-1.128) (-1.230) (-1.119) (-1.297) Auditor_2Tier -0.048 -0.046 -0.053 -0.046 -0.052  (-0.953) (-0.927) (-1.057) (-0.938) (-1.016) Auditor_Tenure 0.000 0.001 -0.001 -0.000 -0.000  (0.121) (0.185) (-0.224) (-0.019) (-0.043) Auditor_Dismiss -0.022 -0.022 -0.017 -0.021 -0.021  (-1.047) (-1.051) (-0.798) (-0.991) (-0.982) Auditor_Resigned 0.059 0.063 0.055 0.063* 0.052  (1.548) (1.633) (1.450) (1.656) (1.333) Other Controls      RetVol_High 0.014 0.012 0.016 0.015 0.013 70   (0.712) (0.611) (0.829) (0.777) (0.676) Table 1.4 (continued)     MarketCap 0.042*** 0.043*** 0.040*** 0.039*** 0.041***  (8.712) (8.682) (8.135) (7.865) (8.044) Firm_Age -0.000 0.000 -0.001 -0.000 -0.001  (-0.126) (0.066) (-0.248) (-0.076) (-0.247) Loss 0.027 0.028* 0.026 0.029* 0.024  (1.652) (1.707) (1.584) (1.736) (1.419) Sales_Growth 0.004 0.005 0.005 0.008 0.004  (0.484) (0.597) (0.568) (0.830) (0.413) Bankruptcy -0.001 -0.002 -0.001 -0.001 -0.002  (-0.371) (-0.514) (-0.274) (-0.288) (-0.555) Num_Segments -0.013* -0.013** -0.013* -0.014** -0.013*  (-1.885) (-2.009) (-1.907) (-2.094) (-1.922) Ext_Financing -0.063 -0.057 -0.062 -0.054 -0.051  (-1.138) (-1.035) (-1.146) (-0.999) (-0.929) Restructuring -0.010 -0.013 -0.009 -0.008 -0.013  (-0.684) (-0.911) (-0.658) (-0.588) (-0.873) M&A 0.022 0.026 0.024 0.027 0.017  (1.227) (1.445) (1.315) (1.476) (1.003) Litigation Risk 0.072*** 0.072*** 0.072*** 0.067*** 0.067***  (2.971) (2.909) (2.969) (2.830) (2.752) GDP -1.209** -2.159*** -2.305*** -1.842*** -2.480***  (-2.059) (-3.567) (-2.910) (-2.791) (-2.775) Inst_Ownership 0.006 0.020 0.008 0.058 -0.017  (0.159) (0.545) (0.201) (1.494) (-0.386)             Industry FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Observations 4,620 4,620 4,620 4,623 4,539 R2 0.133 0.135 0.133 0.135 0.134         71    Table Error! No text of specified style in document..5 Variation in SEC Monitoring with Country Fixed Effects   This table examines whether the intensity of SEC monitoring is different across foreign firms based on the strength of home country enforcement. Both columns include country fixed effects, which allows us to examine the variation in SEC monitoring within country. The dependent variable is SEC_Review, an indicator variable set equal to 1 if the firm received a comment letter for a period t filing, and 0 otherwise. The coefficient of interest in the first column is on the interaction of Staff_Growth, a binary variable which takes the value of one if there is an increase in staff resources, and Post, a binary variable which takes the value of one for 2008 and later. A positive (negative) coefficient indicates that an increase in enforcement was associated with an increase (decrease) in SEC monitoring intensity. The coefficient of interest in the second column is on Single_Listed, a binary variable which takes the value of one if the foreign firm is only listed on a US exchange. A positive (negative) coefficient indicates that single listed firms are subject to higher (lower) levels of SEC monitoring. Each control variable is described in Table 4. Coefficients on control variables are not shown for ease of presentation. We include country-, industry- and year-fixed effects in the regression, but do not report the coefficients. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry.        (1) (2)    Staff_Growth*Post -0.119**   (-2.050)  Single_Listed  0.053*   (1.736)       Control Variables          Accounting Quality Yes Yes        Auditor  Yes Yes        Other  Yes Yes Industry FE Yes Yes Country FE Yes Yes Year FE Yes Yes Observations 1,867 4,808 R2 0.212 0.145      72  Table Error! No text of specified style in document..6 Alternate Measure of SEC Oversight to Mitigate Partial Observability Concerns This table examines whether the intensity of SEC monitoring is different across foreign firms based on the strength of home country enforcement. The dependent variable is SEC_Review_ADJ, an indicator variable set equal to 1 if the firm received a comment letter at time t or if it did not receive a comment letter for a period t-2 through period t filing (i.e., for the previous three years), and 0 otherwise. Each measure of home country enforcement and each control variable is described in Table 4. A negative (positive) coefficient on any of the measures of home country enforcement indicates that foreign firms with stronger home country enforcement were less (more) likely to receive a comment letter, conditional on an SEC review. Each control variable is described in Table 4. Coefficients on control variables are not shown for ease of presentation. We include industry- and year-fixed effects in the regression, but do not report the coefficients. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry.              (1) (2) (3) (4) (5) Private Enforcement      Disclosure_Requirements -0.105***      (-4.006)     Liability_Standards  -0.084***      (-5.232)    Public Enforcement      Rules_Enforcement   -0.082***      (-2.782)   Staff_Resources    -0.002***      (-4.969)  Budget_Resources     -0.024**      (-2.457)         Control Variables             Accounting Quality Yes Yes Yes Yes Yes        Auditor  Yes Yes Yes Yes Yes        Other  Yes Yes Yes Yes Yes Industry FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Observations 4,620 4,620 4,620 4,623 4,539 R2 0.077 0.078 0.076 0.078 0.075 73  Table Error! No text of specified style in document..7 Alternative Measures of SEC Monitoring Intensity  Each Table below uses an alternative measure of SEC monitoring intensity. In Panel A, we proxy for SEC monitoring intensity using SEC_Review_Words, a variable that equals the sum of the number of words in each comment letter issued as part of a conversation between the SEC and the firm. In Panel B, we proxy for SEC monitoring intensity using SEC_Review_Filing, a variable that equals the number of financial filings with comment letters. Both of these variables better capture the effort required to conduct a review, as longer comment letters or those that cover more financial filings likely took more effort to prepare. In Panel C, we proxy for SEC monitoring intensity using SEC_Review_2, an indicator variable that takes the value 1 if the firm received a comment letter in year t and at least one comment letter in the previous two years. This variable better captures the discretionary review that is beyond the required level. A negative (positive) coefficient on any of the proxies for the strength of home country enforcement in each Panel indicates that foreign firms with stronger home country enforcement were subject to lower (higher) SEC monitoring. Each control variable is described in Table 4. Coefficients on control variables are not shown for ease of presentation. We include industry- and year-fixed effects in the regression, but do not report the coefficients. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry. Panel A: Length of Comment Letter             (1) (2) (3) (4) (5) Private Enforcement       Disclosure_Requirements -1.046***      (-3.995)     Liability_Standards  -0.787***      (-5.033)    Public Enforcement      Rules_Enforcement   -0.920***      (-3.376)   Staff_Resources    -0.021***      (-5.259)  Budget_Resources     -0.227**      (-2.523)             Control Variables             Accounting Quality Yes Yes Yes Yes Yes        Auditor  Yes Yes Yes Yes Yes        Other  Yes Yes Yes Yes Yes Industry FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Observations 4,620 4,620 4,620 4,623 4,539 R2 0.136 0.138 0.135 0.138 0.137       74    Table 1.7 (continued)  Panel B: Number of financial filings with comment letter during fiscal year t as dependent variable             (1) (2) (3) (4) (5) Private Enforcement      Disclosure_Requirements -0.178***      (-3.512)     Liability_Standards  -0.132***      (-4.766)    Public Enforcement      Rules_Enforcement   -0.146***      (-2.954)   Staff_Resources    -0.003***      (-4.440)  Budget_Resources     -0.039**      (-2.293)         Control Variables             Accounting Quality Yes Yes Yes Yes Yes        Auditor  Yes Yes Yes Yes Yes        Other  Yes Yes Yes Yes Yes Industry FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Observations 4,620 4,620 4,620 4,623 4,539 R2 0.127 0.128 0.126 0.127 0.130    75  Table 1.7 (continued)  Panel C: Two or more comments letters in a three-year period              (1) (2) (3) (4) (5) Private Enforcement      Disclosure_Requirements -0.125**      (-2.418)     Liability_Standards  -0.089***      (-2.982)    Public Enforcement      Rules_Enforcement   -0.153***      (-2.935)   Staff_Resources    -0.003***      (-2.985)  Budget_Resources     -0.021      (-1.481)         Control Variables             Accounting Quality Yes Yes Yes Yes Yes        Auditor  Yes Yes Yes Yes Yes        Other  Yes Yes Yes Yes Yes Industry FE Yes Yes Yes Yes Yes Year FE Yes Yes Yes Yes Yes Observations 2,959 2,959 2,959 2,962 2,903 R2 0.146 0.147 0.148 0.149 0.145    76  Table Error! No text of specified style in document..8 Variation in SEC Monitoring based on US Investor Exposure   Panel A: Analysis of the impact of US investor exposure on regulatory oversight conditional on home country enforcement  This table examines whether the intensity of SEC monitoring is different for foreign firms compared to US firms. The dependent variable is SEC_Review, an indicator variable set equal to 1 if the firm received a comment letter for a period t filing, and 0 otherwise. The coefficient of interest is US_Exposure, which equals the percentage of the firm’s overall market cap that is traded on US exchanges. A negative (positive) coefficient on US_Exposure indicates that foreign firms which greater exposure to the US were subject to lower (higher) SEC monitoring. Each control variable is described in Table 4. Coefficients on control variables are not shown for ease of presentation. We include industry- and year-fixed effects in the regression, but do not report the coefficients. We include country fixed effects in Column (2). ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry.          (1) (2)    US_Exposure 0.066*** 0.050**  (2.988) (1.999)       Control Variables Yes Yes        Accounting Quality Yes Yes        Auditor  Yes Yes        Other  Yes Yes Country FE No Yes Industry FE Yes Yes Year FE Yes Yes Observations 4,191 4,191 R2 0.136 0.150   77  Table 1.8 (continued)   Panel B: Combined Effect of Enforcement and US Investor Exposure This panel reports the mean values of SEC monitoring intensity across the low and high tercile of US Exposure and each proxy for enforcement. We indicate statistical significance of differences across cells with t-tests.  Disclosure  Requirements  Liability Standards   US  Low High    Low High       Exposure  (a) (b) (b)-(a)   (a) (b) (b)-(a)      Low  (i) 0.004 -0.072 -0.076**  (i) -0.007 -0.135 -0.128***      N=446 N=427   N=446 N=319      High  (ii) 0.092 0.016 -0.074*  (ii) 0.077 -0.023 -0.100**      N=445 N=427   N=445 N=319       (ii)-(i) 0.088*** 0.088** 0.012  (ii)-(i) 0.084** 0.112** -0.16                      Rules Enforcement  Staff Resources  Budget Resources US  Low High    Low High    Low High  Exposure  (a) (b) (b)-(a)   (a) (b) (b)-(a)   (a) (b) (b)-(a) Low  (i) 0.008 -0.092 -0.100***  (i) -0.005 -0.103 -0.098***  (i) 0.015 -0.126 -0.111*** N=449 N=459   N=468 N=463    N=419 N=510  High  (ii) 0.114 0.020 -0.094***  (ii) 0.089 0.029 -0.60  (ii) 0.093 -0.012 -0.105*** N=449 N=459   N=467 N=462    N=419 N=510    (ii)-(i) 0.106*** 0.112** 0.012   (ii)-(i) 0.094** 0.132*** 0.034   (ii)-(i) 0.078** 0.114*** 0.027    78  Table Error! No text of specified style in document..9 Comparison of Monitoring Intensity for US versus Foreign Firms Panel A: Comparison of Monitoring Intensity for US versus Foreign Firms This table examines whether the intensity of SEC monitoring is different for foreign firms compared to US firms. The dependent variable is SEC_Review, an indicator variable set equal to 1 if the firm received a comment letter for a period t filing, and 0 otherwise. The coefficient of interest is on Foreign_Firm, an indicator variable set equal to 1 if firm is classified as a foreign private issuer by the SEC for fiscal year t, and 0 otherwise. A negative (positive) coefficient on Foreign_Firm indicates that foreign firms were subject to lower (higher) SEC monitoring. Column (1) uses the full sample of US firms, Column (2) creates a matched sample based on year, industry and size, and Column (3) creates a matched sample using year, industry, accounting quality and size. Each control variable is described in Table 4. Coefficients on control variables are not shown for ease of presentation. We include industry- and year-fixed effects in the regression, but do not report the coefficients. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry.    Full Sample  Matching 1 (year, ind, size) Matching 2 (year, ind, AQ, size) (1) (2) (3)     Foreign_Firm -0.080*** -0.121*** -0.125***  (-4.634) (-5.405) (-5.803)     Control Variables           Accounting Quality Yes Yes Yes        Auditor  Yes Yes Yes        Other  Yes Yes Yes Industry FE Yes Yes Yes Year FE Yes Yes Yes Observations 41,540 9,616 9,574 R2 0.101 0.145 0.151        79  Table 1.9 (continued)  Panel B: Analyses of Foreign Private Issuers Listed only on US Exchange This table examines whether the intensity of SEC monitoring is different for foreign firms based on whether the foreign firm’s shares are single-listed or cross-listed. The dependent variable is SEC_Review, an indicator variable set equal to 1 if the firm received a comment letter for a period t filing, and 0 otherwise. Single_Listed is an indicator variable set equal to 1 if foreign firm’s shares only trade on a US exchange, and 0 otherwise. Foreign_ CrossListed is an indicator variable set equal to 1 if foreign firm’s shares are cross-listed on both a US and foreign exchange, and 0 otherwise. A positive (negative) coefficient on Single_Listed indicates that foreign firms whose shares only trade on a US exchange are subject to a higher level of SEC monitoring than the other firms in the sample. Column (1) compares foreign single-listed with US firms using the sample of US firms and foreign firm whose shares are only traded on a US exchange. Column (2) compares foreign single-listed and foreign cross-listed firms with US firms using the full sample of US firms. All variables are as defined in Appendix A.1. Each control variable is described in Table 4. Coefficients on control variables are not shown for ease of presentation. We include country-, industry- and year-fixed effects in the regression, but do not report the coefficients. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% level, respectively, using two-tailed tests and standard errors clustered by country-industry.       (1) (2)      Single_Listed -0.015 -0.012  (-0.841) (-0.692)    Cross_Listed  -0.083***   (-4.589)       Control Variables          Accounting Quality Yes Yes        Auditor  Yes Yes        Other  Yes Yes Country FE Yes Yes Industry FE Yes Yes Year FE Yes Yes Comparison Group US US Observations 37,134 41,540 R2 0.102 0.102     80  Chapter 2: Auditor Quality of U.S.-Listed Foreign Firms2.1 Introduction In the literature on the audit production process, the general consensus is that the legal regime of a country affects audit pricing and quality. Where legal regimes are strong, auditors are likely to devote more resources to detect potential financial statement errors and irregularities because they anticipate higher legal liability costs; this applies to both Big4 affiliates and smaller auditors. Other studies show that auditors charge higher audit fees for firms that are cross-listed in countries whose legal regimes are stronger than the auditors’ home country. This behavior, widely known as the cross-listing premium, is likely a reaction to the strong legal and regulatory environment of the listing country. The functional convergence hypothesis states that foreign auditors provide quality similar to U.S. auditors for their cross-listed clients because they abide by U.S. securities laws and regulations. However, recent evidence suggests that the legal enforcement may be weaker for foreign firms listed in the United States (e.g. Sigel, 2004; Naughton et al., 2016). Thus, the magnitude of the cross-listing premium is unknown. In this empirical study, I examine whether foreign auditors provide quality and charge fees similar to U.S. auditors for the foreign firms listed on the NYSE and NASDAQ.    Foreign companies listed in the U.S. are interesting and much studied because they play an important role in the U.S. capital market. For example, the Chinese Internet giant Alibaba raised $25 billion through an initial public offering (IPO) in 2014—the largest IPO in history. Despite the increasing importance and popularity of foreign firms listed in the United States, concerns exist regarding the financial reporting quality of foreign firms (e.g., Lang, Raedy, and Wilson, 2006; Srinivasan et al., 2015). The auditors of these foreign firms are of particular interest 81  because of 1) the important role auditors play in mitigating the information asymmetry between the investors and the foreign firms, and 2) concerns that the foreign auditors are of lower quality, as evidenced by the recent PCAOB sanctions against foreign Big4 affiliates in countries such as Brazil and Indonesia.  In order to develop a comprehensive understanding of the research question, this study first provides institutional details concerning the different groups of auditors in the foreign listing market. Based on the institutional details, I make the following arguments: First, whether foreign auditors charge higher (or lower) audit fees than U.S. auditors depends on the relative weight of the expected liability resulting from the expected enforcement regime and audit operating costs of the audit production process. In addition, if a foreign auditor charges a fee premium over a U.S. auditor, the audit quality must be higher because of the lower audit operating costs. Finally, U.S. auditors are more likely to provide lower quality if the foreign country varies considerably from the United States in terms of geographical distance, languages spoken, and rule of law.  I also develop a simple analytical model in which the expected liability and actual audit cost of an auditor play a crucial role in determining the auditor’s equilibrium fees and effort. Five observations are derived from the model, which support the above arguments. These observations offer insights into the audit process of foreign firms listed in the United States. I then test the empirical predictions by using audit fees and restatements, as well as PCAOB inspection report data, as proxies for audit quality. The results show that U.S. Big4 auditors charge the highest audit fees and provide the best audit quality. In addition, foreign Big4 affiliates charge higher audit fees and provide more assurance than U.S. non-Big4 auditors and foreign non-Big4 auditors. No significant differences are identified in this study between smaller U.S. non-Big4 and foreign non-Big4 auditors. Moreover, U.S. non-Big4 auditors are more likely to provide lower quality when 82  the client is located in a country that is very different from the United States. These findings suggest that despite the potential nuances of the bonding hypothesis1516, foreign auditors provide audit quality similar to their counterparts in the United States.  The remainder of this paper is organized as follows. Section 2.2 presents a literature review, the institutional background, and the hypotheses. Sections 2.3, 2.4, and 2.5 explain the empirical procedures used to test Hypotheses 1, 2, and 3, respectively, and provide the results. Section 2.6 presents concluding remarks. Appendix B.2 presents a stylized model supporting the hypothesis development and Appendix B.3 presents a graphical representation of the hypothesis 2.2 Literature review, Institutional Background and hypothesis development 2.2.1 Literature Review Foreign firms listed in the United States have attracted considerable scholarly interest in accounting and finance, and prior studies examine various aspects of the topic. In particular, they document that foreign firms have higher trading costs (e.g., Eleswarapu and Venkataraman, 2006), exhibit lower earnings quality (e.g., Lang et al., 2004), provide more readable annual statements (Lundholm et al., 2015), and have received less SEC attention (Siegel, 2006; Naughton et al., 2016). However, relatively few papers consider the auditing aspect of such firms (except for                                                  15 Bonding hypothesis is first developed in the finance literature to explain the cross-listing premium (i.e. companies listed in the US have higher valuations than their peers at home). The central idea of bonding hypothesis is that by listing in the US, a company from a country with low investor protections can "bond" itself with the US, where investor protections are high. Thus, bonding hypothesis suggests that listing in the US improves corporate governance of the company, which, in turn, could lower the company's cost of capital. 16 There are several hypotheses developed within the literature of accounting and finance to explain the cross-listing premium. For example, the investor recognition hypothesis suggests that foreign companies listed in the US have higher valuation because cross listing enables companies to attract and maintain US investor attention.  83  Asthana et al. [2015], as discussed later). My study contributes to the literature by studying the audit costs and auditor quality of foreign cross-listed firms.   This study is also related to the audit literature that examines the effect of a country’s legal liability regime on audit fees and audit quality. Choi et al. (2008), Francis and Wang (2008), and Choi et al. (2009) focus on the effect of legal regimes on Big4 auditor fee premiums and audit quality across different countries, as well as the effect of audit fees paid by cross-listed and non-cross-listed firms in international settings. By contrast, my study focuses on foreign listings in the U.S. market. This study also contributes to the emerging literature concerning the failure of the PCAOB in monitoring foreign auditors. Whereas Carcello (2011) and Gu (2012) consider the market’s reaction to the announcement of the PCAOB’s failure to inspect foreign auditors, this study focuses on the determinants of audit fees and audit quality. This study is also related to that of Asthana et al. (2015), who examine the differences between U.S.-based Big N and foreign Big N auditors; however, my study also considers the differences between foreign Big4 versus U.S. non-Big4 auditors and foreign non-Big4 versus U.S. non-Big4 auditors.  This study adds values to the recent debate about the “bonding” issues of Chinese firms listed in the United States (Chen et al., 2016; Ghosh and Wagner, 2014; Lee et al., 2014). Compared with firms from other countries, Chinese firms are more likely than companies from other countries to hire local Big4 auditors. Some have argued that the traditional bonding argument failed for U.S.-listed Chinese companies due to a lack of audit quality and audit firm oversight (e.g., Carcello et al., 2014). This paper provides insight into the quality of the local (Chinese) Big4 auditors, and the findings have implications for regulators and investors. The findings should also interest academics by providing a new perspective on the relative importance of government regulation and supervision and Big4 self-governance. 84  Finally, this study contributes to the recent debate about the existence of Big4 (N) effects. A large body of auditing literature concludes that Big 4 auditors provide higher audit quality and charge fees than non-Big N auditors (e.g. DeAngelo 1981 and Choi et al. 2008). However, recently, a high profile study by Lawrence et al. (2011) suggests that prior findings are subject to estimation errors and propensity score matching on client characteristics eliminates the Big4 effect. The findings of Lawrence et al. (2011) have attracted considerable attention. To validate Lawerence et al. (2011)’s findings, Defond et al (2016) examine random combinations of propensity score matching design and find that the majority of these design choices support a Big4 effect. These authors conclude that it is premature to suggest that propensity score matching eliminates the Big4 effect. My study contributes to the recent debate concerning the existence of Big4 effects by further examing the Big4 effect between foreign Big4 affliates and U.S. non-Big4 auditors.  2.2.2 Institutional Background Four categories of potential auditors are available in the foreign listing market. A foreign issuer can hire a U.S. Big4 office, a foreign Big4 affiliate, a U.S. non-Big4 auditor, or a foreign non-Big4 auditor as a public accountant. Under the Sarbanes-Oxley Act of 2002, foreign accounting firms that have registered with the PCAOB are subjected to the same PCAOB inspections in the same manner as U.S. firms. As of November 2015, over 900 non-U.S. audit firms from more than 85 countries had registered with the PCAOB. Several types of non-U.S. audit firms exist, ranging from foreign Big4 offices (e.g., KPMG LLP of Toronto, Canada) to small local audit firms (e.g., Guangzhou Good Faith CPA LTD of Guangzhou, China).  The U.S. Big4 auditors are expected to be the highest quality providers. A large body of academic and industry evidence suggests substantial differences between Big 4 and non-Big4 85  auditors in the United States. In addition, U.S. Big4 auditors have expert knowledge regarding U.S. security regulations and are audited by the PCAOB on an annual basis.   The differences between U.S. Big4 offices and their foreign counterparts is debatable. On one hand, every Big4 audit firm markets itself as one auditor globally. For example, the official website of Ernst and Young states that “we’re not merely a loose collection of national practices—we are a global organization, unified in our approach.” In addition, anecdotal evidence suggests that companies perceive the U.S. and foreign Big4 affiliates similarly. For example, on October 15, 2010, China Natural Resources, Inc. dismissed GHP Horwath, P.C., a U.S. non-Big4 auditor, as its principal independent registered public accounting firm and engaged Ernst & Young Hong Kong as its principal independent registered public accounting firm. However, the auditor location of Ernst & Young was not mentioned in the 6-K disclosure17. On the other hand, substantive differences may exist between the U.S. Big4 and their foreign affiliates. One factor supporting this is that the Big4s are legally independent from each other. For example, PwC states that “[i]n many parts of the world, accounting firms are required by law to be locally owned and independent. Although regulatory attitudes on this issue are changing, PwC member firms do not and cannot currently operate as a corporate multinational. The PwC network is not a global partnership, a single firm, or a multinational corporation.” Some have argued that foreign accounting firms simply borrow the name and reputation of the Big4 by becoming affiliate firms (Guo 2016). The fall of ChuoAoyama in 2006 demonstrates the legal independence of Big4 networks. As the former Japanese affiliate of PwC, ChuoAoyama was one of Japan's largest audit firms. In May 2006, the Japanese Financial Services Agency (FSA) suspended ChuoAoyama from providing statutory                                                  17 An alternative explanation is that this is an obfuscation strategy 86  auditing services for 2 months because of its heavy involvement in the Kanebo fraud. PwC acted quickly after the suspension of ChuoAoyama by setting up PricewaterhouseCoopers Aarata and moving some of ChuoAoyama's accountants to the new firm. ChuoAoyama was eventually dismissed in 2007. However, there is little other evidence on this issue apart from Asthana et al. (2015), who find that the U.S. Big4 outperform their foreign affiliates. More empirical evidence is required.  As an alternative to the Big4, companies could hire a U.S. non-Big4 to be their auditor. The advantages of U.S. non-Big4 auditors are clear: They charge lower audit fees than U.S. Big4 auditors, and although they are smaller and do not carry the Big4 names, they are located in the United States and are constantly under the surveillance of the SEC and the PCAOB. However, they have a cost disadvantage compared with the foreign auditors because they are geographically and culturally removed from their foreign clients.  Another option for companies is to hire a foreign non-Big4 auditor. This group of auditors is the least known to U.S. investors; however, these auditors tend to be larger auditors in the foreign country. In some cases, they are similar in size to, if not larger than, the foreign Big4 affiliates in the foreign country. Moreover, the foreign non-Big4 auditors may value the opportunity to conduct audits in the United States. For example, Aobdia and Shroff (2016) document that PCAOB-inspected non-U.S. auditors witness, on average, a 4% to 6% increase in their market share after the publication of PCAOB inspection reports, although this increase is only significant for auditors with low levels of deficiencies. Thus, foreign non-Big4 auditors may have an incentive to provide high audit quality. 87  2.2.3 Empirical Predictions The audit quality and pricing of Big4 and non-Big4 firms are known to vary systematically across countries (Choi et al., 2008). In general, they vary as a function of the strictness of the local legal regime. A client’s cross-listing is expected to change local auditor behavior because cross-listing changes the legal regime of the client and its auditor, a theory known as the bonding hypothesis. There is evidence that the audit fees and quality of cross-listed firms are higher than those of non-cross-listed firms. Moreover, the differences in audit fees and quality increase with the differences between the rigor of the local legal regime and the U.S. regime (Choi et al., 2009). Because foreign auditors abide by U.S. securities laws and regulations, foreign auditors may seem to provide similar quality and charge similar fees to U.S. auditors for their cross-listed clients. That is, foreign Big4 affiliates provide similar quality and charge similar fees to U.S. Big4 auditors, and provide better quality and higher fees than U.S. non-Big4 auditors, whereas foreign non-Big4 auditors provide similar quality and fees to U.S. non-Big4 auditors.   However, multiple differences exist between foreign and U.S. auditors. Specifically, foreign auditors’ input cost and expected legal liability may differ from those of U.S. auditors. Foreign auditors have lower marginal audit costs than U.S. auditors because they speak the same language, are more familiar with the business culture, and are geographically closer to the cross-listed clients. It is also unclear whether cross-listing mimics the legal regime of U.S. firms. U.S. regulators may have difficulty monitoring auditors in foreign jurisdictions because of political barriers. For example, the PCAOB is denied the ability to conduct inspections in countries such as China. In addition, prior studies show that U.S. regulators may have incentives to reduce their monitoring of foreign firms compared with U.S. firms (Siegel, 2005; Naughton et al., 2016). One possible explanation for the reduced level of monitoring is that U.S. regulators focus their 88  resources to protect U.S. investors. If the difference in the monitoring level is substantive, then the foreign auditors should have smaller incentives to provide a high audit quality. In sum, for a fixed amount of audit fees, foreign auditors provide a high level of assurance because they have a cost advantage. However, whether foreign auditors provide better (or worse) quality than their U.S. counterparts requires further investigation.  U.S. Big4 auditors charge the highest audit fees for multiple reasons. First, they have the highest expected legal liability because of the size of their portfolio in the United States and because they are constantly under the inspection of the SEC and PCAOB. This is consistent with the notion that a larger auditor provides higher audit quality, as suggested by DeAngelo (1981). They have to charge higher fees for the expected legal liability. Furthermore, U.S. Big4 auditors have higher audit costs for cross-listed clients than their foreign affiliates because of their geographical distance to the overseas clients.  Whether the audit fee premiums charged by foreign Big4 auditors are higher or lower than those charged by their U.S. non-Big4 counterparts is unclear. In addition to the lower audit costs, foreign Big4 auditors may be subject to weak legal enforcement in the United States (i.e., there is nuance to the bonding hypothesis). The lower expected legal liability would be attributed to the U.S. regulator having lower incentives to monitor the foreign Big4 auditors or having difficulty monitoring auditors in foreign jurisdictions. Therefore, foreign auditors may have fewer incentives to work hard and, in turn, expect lower audit fees. However, Big4 auditors always market themselves as one firm with consistent quality in their advertisements. The audit failure of a foreign Big4 office may have spillover effects for global Big4 offices, giving the global Big4 head offices incentives to enforce the monitoring of audit efforts around the globe. Therefore, if the nuances of the potential of the bonding hypothesis are small, then foreign Big4 affiliates may work harder and 89  charge higher fees than the U.S. non-Big4 auditors. In sum, whether audit fees are higher or lower for foreign Big4 affiliates depends on several opposing factors, and the magnitude of each cannot be predetermined. Thus, Hypothesis 1a is formulated as follows: H1a: There is no difference between the audit fees charged by U.S. non-Big4 auditors and foreign Big4 auditors.  Whether the audit fee premiums charged by foreign non-Big4 auditors are higher or lower than those charged by their U.S. non-Big4 counterparts is also unclear. Foreign non-Big4 auditors have lower audit costs and the expected legal liability may be lower; however, foreign non-Big4 auditors may also incur greater losses in the event of an audit failure for a cross-listed firm than U.S. non-Big4 auditors. This is because it may be costlier (because of language barriers and documentation complexity) for foreign non-Big4 auditors to register with the PCAOB. Evidence exists that registration with the PCAOB is valuable to foreign non-Big4 auditors because it can provide credibility in their respective home countries. For example, as noted earlier, Aobdia and Shroff (2016) document that PCAOB-inspected non-U.S. auditors witness, on average, a 4% to 6% increase in their market share after the publication of PCAOB inspection reports, although this increase is only significant for auditors with low levels of deficiencies. Therefore, Hypothesis 1b is formulated as follows: H1b: There is no difference between the audit fee premiums charged by U.S. non-Big4 auditors and foreign non-Big4 auditors. In a rational equilibrium, auditors who charge higher audit fees must provide higher quality audits because investors/companies will not pay a higher price for a product (in this case, a level of assurance) that is not of higher quality. However, friction may exist in the audit market and U.S. investors may have a home bias against foreign auditors. It is possible that investors are willing to 90  pay, indirectly through management, for the higher audit fees of U.S. auditors without obtaining a higher level of assurance. In other words, it is possible that U.S. auditors charge higher audit fees but produce a lower audit quality. The alternative does not hold. If foreign Big4 auditors charge a fee premium over U.S. non-Big4 auditors, then the foreign Big4 auditors may provide higher audit quality than the U.S. non-Big4. The intuition behind this is simple: because foreign auditors have lower audit costs than U.S. auditors, high total costs imply more hours of work. Thus, Hypothesis 2a is formulated as follows: H2a: U.S. non-Big4 only provide higher audit quality than foreign Big4 auditors if the U.S. non-Big4 charge a fee premium over the foreign Big4 affiliates.  By the same logic, if foreign non-Big4 auditors charge a fee premium over U.S. non-Big4 auditors, then the foreign non-Big4 auditors provide higher audit quality than the U.S. non-Big4. Hypothesis 2b is thus formulated as follows: H2b: U.S. non-Big4 only provide higher audit quality than foreign non-Big4 auditors if the U.S. non-Big4 charge a fee premium over the foreign non-Big4 affiliates. Finally, I examine how the audit quality of U.S. auditors changes across countries with different characteristics. If a foreign country varies considerably from the United States with regard to language, business culture, or geographical distance, U.S. auditors would provide lower audit quality because they face higher operating costs. However, because U.S. Big4 auditors can utilize the staff resources of their foreign affiliates through global Big4 networks, they may be able to more effectively mitigate this situation than the non-Big4 auditors. Therefore, Hypotheses 3a and 3b are formulated as follows: H3a: U.S. Big4 audit quality does not vary across firms from different countries. H3b: U.S. non-Big4 audit quality decreases when the barriers to audit increase.  91  Appendix B.2 provides a stylized model of the hypothesis development. Appendix B.3 offers a graphical representation of the hypothesis.  2.3 Testing Hypothesis 1 2.3.1 Research Design for Testing Hypothesis 1 To test the first hypothesis, I use a sample of foreign firm-years with data available between 2000 and 2013. Following the literature, I classify firms as foreign if their headquarters are outside of the United States (using the LOC variable from Compustat). I use the data from the intersection of Audit Analytics, COMPUSTAT Fundamentals Annual, and CRSP. Financial firms (SIC 6000-6999) and utilities firms (SIC 4900-5000) are excluded. Table 2.7.1 shows the results of the sample selection procedure. The final sample consists of 6,689 firm-years and 1,201 unique firms.  I apply the following regression model to test Hypothesis 1: 𝐿𝑜𝑔 𝑜𝑓 𝐴𝑢𝑑𝑖𝑡 𝐹𝑒𝑒𝑠𝑡 =   0 +  𝛽1𝑈𝑆_𝐵𝑖𝑔4𝑡+  𝛽2𝐹𝑜𝑟𝑒𝑖𝑔𝑛_𝐵𝑖𝑔4𝑡+  𝛽3𝐹𝑜𝑟𝑒𝑖𝑔𝑛_𝑛𝑜𝑛_𝐵𝑖𝑔4𝑡  + 𝛽4𝐿𝑜𝑔 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑡 +     𝛽5𝑅𝑒𝑐𝑒𝑖𝑣𝑎𝑏𝑙𝑒/𝐼𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑡     + 𝛽6 𝑅𝑒𝑡𝑢𝑟𝑛 𝑜𝑛 𝐴𝑠𝑠𝑒𝑡𝑠𝑡  +𝛽7 𝐿𝑜𝑠𝑠𝑡 +  𝛽8 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑡 +     𝛽9 𝐿𝑖𝑡𝑖𝑔𝑎𝑡𝑖𝑜𝑛 𝑟𝑖𝑠𝑘𝑡   + 𝛽10  𝐵𝑢𝑠𝑦 𝑠𝑒𝑎𝑠𝑜𝑛𝑡 +𝛽11𝐺𝑒𝑜𝑠𝑒𝑔𝑚𝑒𝑛𝑡𝑡 +  𝛽12 𝐵𝑢𝑠𝑠𝑒𝑔𝑚𝑒𝑛𝑡𝑡 + 𝛽13 𝑅𝑒𝑡𝑢𝑟𝑛 𝑣𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦𝑡  +  𝛽14 𝐺𝑜𝑖𝑛𝑔_𝑐𝑜𝑛𝑐𝑒𝑟𝑛 + 𝛽15 10𝐾 +  𝛽𝑗  𝐶𝑙𝑖𝑒𝑛𝑡𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡 +  𝛽𝑘 𝑌𝑒𝑎𝑟 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡      (14)       All the variables are defined in Appendix B.1. I include three categorical variables in the regression to represent three of the auditor types; the category that is not included in the regression, the U.S. non-Big4 auditors, is the benchmark of the OLS analysis. Two variables are of particular interest. The first test variable, Foreign_Big4, is an indicator variable that equals 1 for foreign firms that use foreign Big4 auditors, and 0 otherwise. The coefficient of Foreign_Big4 thus captures the audit fee premium of foreign Big4 auditors relative to U.S. non-Big4 auditors. The second variable of interest is the Foreign_non_Big4 indicator that equals 1 for foreign firms that use foreign non-92  Big4 auditors, and 0 otherwise. It captures the audit fee premium of foreign non-Big4 auditors relative to U.S. non-Big4 auditors. The regression includes 12 firm-specific control variables. The first 10 variables are proxies for client size, complexity, and risk (e.g., Simunic 1980; Hay et al. 2006). I also include return volatility as an additional proxy for client risk that is not captured by the traditional control variables, and I include a 10-K dummy to control for the extra audit work required because of filing differences1819. Finally, I include fixed-effect indicator variables for countries and years to control for potential variations in audit fees across countries and over time. Because of the expectation that audit fees are positively related to client size, complexity, and risk factors, all of the coefficients from 𝛽5to 𝛽15 are expected to be positive, except 𝛽6. In addition, as explained in an earlier section, 𝛽6 is expected to be positive because of the extra work of reviewing quarterly filings. I test Hypothesis 1 on a full sample of foreign firms. The coefficients of interest are 𝛽2 to 𝛽3. Foreign firms that use U.S. non-Big4 auditors constitute the benchmark sample in the regression. The coefficients of 𝛽2 and 𝛽3 are the estimations of audit fee premiums relative to U.S. non-Big4 auditors charged by foreign Big4 auditors and foreign non-Big4 auditors. In the regression model, the coefficient of US_Big4 is expected to be significantly positive (𝛽1 > 0), as studies in the literature suggest that U.S. Big4 auditors charge higher audit fees. In addition, as stated in Hypothesis 1, this study has no expectations about the coefficients on Foreign_Big4 and                                                  18 Certain foreign firms are required to comply with U.S. domestic issuers’ continuous filing requirements such as filing quarterly financial statements using U.S. GAAP, disclosure of insider trading, and filing annual 10-K statements. A foreign firm can always choose to increase its financial statement disclosure by voluntarily filing 10-K.   19 The result of the paper is quantitively the same if I use the number of words of the annual filings to control for the audit effort induced by idiosyncratic firm factors (Brad, et al. 2014). 93  Foreign_non_Big4 because evidence is lacking concerning the relative wealth of foreign Big4 auditors, foreign non-Big4 auditors, and U.S. non-Big4 auditors.  2.3.2 Descriptive Statistics Table 2.2 presents the distribution of firm-year observations and auditor choices for all countries from 2000 to 2013. Foreign Big4 auditors, on average, are more likely to be hired by foreign firms (5,424 firm-years out of 6,689). Table 2.3 suggests that differences can be found across each group of auditors. U.S. Big4 auditors, on average, charge higher audit fees than all other firms, whereas foreign non-Big4 auditors charge the lowest audit fees. There are also several differences across the control variables. For example, U.S. Big4 auditors have bigger clients than all other firms, and U.S. non-Big4 auditors have the smallest clients among the four groups of auditors. Because of these client differences, I also conduct analyses using within-firm analysis and change analysis approaches.   2.3.3 Results for Hypothesis 1 Table 3 presents the first test of Hypothesis 1. Column 1 of Table 2.3 indicates that the coefficients of the U.S. Big4 (𝛽1= 0.355) and foreign Big4 (𝛽2= 0.364) auditor variables are positive and significant, whereas the coefficient of the foreign non-Big4 auditor variable is marginally negatively significant (𝛽3= –0.173). In addition, the signs of the coefficients of the control variables are mostly in the direction identified in the literature, and the clear majority of these signs are statistically significant. Column 2 of Table 2.7.3 reveals stronger results (Adjusted R-square = 69.3% in Model 1 and 88.7% in Model 2), as derived after the inclusion of firm fixed effects in the regression model to control for firm-specific but time-invariant components of audit fees. Importantly, the coefficient of the foreign Big4 (𝛽2= 0.579) auditor variables remain positive 94  and significant. Recall that the Foreign Big4 is a dummy variable, of which the U.S. non-Big4 auditors serve as the baseline; thus, the magnitude of the coefficient is the difference in the log of audit fees between the foreign Big4 auditors and the U.S. non-Big4 auditors. Therefore, this study finds that the audit fees are approximately 43.9% (calculated as an exponential of 0.364 – 1) to 78.4% (calculated as an exponential of 0.579 – 1) higher for firms hiring foreign Big4 auditors than for firms hiring U.S. non-Big4 auditors. These results indicate that the differences between the audit fees charged by foreign Big4 auditors and U.S. non-Big4 auditors are both economically and statistically significant. A notable finding is that the coefficient of the Foreign non-Big4 (𝛽3 = −0.008) dummy is no longer significant after firm fixed effects are controlled, which suggests that there is only marginal evidence that U.S. non-Big4 auditors charge higher fees than their peers overseas. One thing to note that Model 2 also suggests that the U.S. Big4 charges higher audit fees than foreign Big4 auditors after controlling for time-invariant but firm specific characteristics.  Overall, the results in Table 2.3 are inconsistent with Hypothesis 1a and suggest that foreign Big4 auditors also charge higher audit fees than U.S. non-Big4 auditors, even though the foreign Big4 auditors have a cost advantage over U.S. non-Big4 auditors. However, the results in Table 2.3 are consistent with Hypothesis 1b, suggesting that there is no evidence that U.S. non-Big4 auditors and foreign non-Big4 auditors charge different fees.  2.3.4 Supplemental Robustness Tests for Hypothesis 1  I conduct two more robustness tests to confirm the preceding results. Lawrence et al. (2011) suggest that the Big4 effect largely reflect client characteristics and researchers should explore alternative methodologies that separate client characteristics from audit-quality effects. To improve the identification of controlling for clientele differences, I estimate the audit fee 95  regression using first differences. Specifically, rather than conducting a cross-sectional analysis, this test focuses on a subsample of firms that have used both foreign Big4 and U.S non-Big4 auditors during the sample period. Because the clients act as their own control, there is more confidence that the changes in audit fees are due to the auditors rather than other firm-specific factors such as firm size. Column 1 of Table 2.4 shows the results of the first-differences regression, which are consistent with those in Table 2.3: The foreign Big4 dummy (coefficient = 0.679) is significantly positive. Companies that change their auditors from U.S. non-Big4 to foreign Big4 auditors likely have to pay significantly higher audit fees. This evidence rejects Hypothesis 1a. Additionally, there is marginal evidence that companies that change their auditors from U.S. non-Big4 to foreign non-Big4 auditors pay higher audit fees (coefficient = 0.278), which is inconsistent with results in Table 2.3. To overcome potential concerns that the results of the first-differences estimation are driven by firms that do not change auditors, I refine the estimation by estimating the regression within the firms that change their auditors from U.S. non-Big4 to foreign Big4 (or vice versa). Column 2 of Table 2.4 shows the inferences for this subsample, indicating that the coefficient for changing auditors from U.S. non-Big4 auditors to foreign Big4 affiliates is positive and significant (coefficient = 0.535, t = 3.38), which suggests that foreign Big4 auditors, on average, charges 70.7% higher (calculated as e0.535 – 1) audit fees than U.S. non-Big4 auditors. Again, this evidence rejects Hypothesis 1a. 2.3.5 Variations Within U.S. Non-Big4 Auditors  It is noteworthy that not all U.S. non-Big4 auditors are the same. Some of them are large regional firms and some are small “local” auditors. As suggested in the literature (e.g., DeAngelo 1981), larger auditors provide better quality and charge higher fees, because auditors with a greater 96  number of clients have more to lose in an audit failure. An interesting cross-sectional analysis, therefore, is to explore whether foreign Big4 affiliates charges higher audit fees than U.S. non-Big4 auditors of all sizes. Accordingly, I measure auditor size by three categories: total assets, total market capitalization of clients, and total number of audit clients. I rank U.S. non-Big4 auditors as large or small using the three categories. Table 2.5 presents the regression results derived from re-estimating the equation by adding an indicator variable for large U.S. non-Big4 auditors. Table 2.5 presents the consistent results that, ceteris paribus, foreign Big4 affiliates charge higher audit fees than large U.S. non-Big4 auditors, which in turn charge higher fees than foreign non-Big4 auditors and small U.S. auditors.   In sum, this consistent and robust evidence rejects Hypothesis 1a, as it suggests that foreign Big4 auditors charge a higher audit fee premium than all U.S. non-Big4 auditors. There is mixed evidence on the differences in audit fees charged between U.S. and foreign non-Big4 auditors, neither supporting nor rejecting Hypothesis 1b. 2.4 Testing Hypothesis 2 2.4.1 Research Design for Testing Hypothesis 2 Hypothesis 2 predicts that if foreign auditors charge a fee premium over U.S. non-Big4 auditors, then the foreign auditors provide higher audit quality than the U.S. non-Big4 auditors because of their low marginal costs (i.e., higher fees suggest more audit hours). Given the robust evidence that foreign Big4 auditors charge higher audit fees than U.S. non-Big4 auditors, and considering the existence of inconclusive evidence on the fee differences between foreign non-Big4 and U.S. non-Big4 auditors, Hypothesis 2 is restated as follows: H2a: Foreign Big4 auditors provide higher audit quality than U.S. non-Big4 auditors. 97  H2b: There is no difference between the audit quality of foreign and U.S. non-Big4 auditors. This study tests Hypothesis 2 using the following probit regression model: 𝑃𝑟(𝑅𝐸𝑆𝑇𝐴𝑇𝐸𝑀𝐸𝑁𝑇𝑖, 𝑡 , 𝐴𝐶𝐶𝑂𝑈𝑁𝑇𝐼𝑁𝐺𝑖, 𝑡 )  =  0+ 1𝑈𝑆_𝐵𝑖𝑔4𝑡+  2𝐹𝑜𝑟𝑒𝑖𝑔𝑛_𝐵𝑖𝑔4𝑡+  3𝑈𝑆_𝑛𝑜𝑛_𝐵𝑖𝑔4𝑡 +  𝑖𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑗 ++ 𝑗 𝐶𝑙𝑖𝑒𝑛𝑡𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡 +   𝑘 𝑌𝑒𝑎𝑟 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡𝑠𝑡    (15)  where RESTATEMENT, from Audit Analytics, is an indicator variable equal to 1 if a firm-year’s statement is subsequently restated (and done so because of accounting issues), and 0 otherwise. To be consistent with Hypothesis 1, the control variables are the same as in Model 14. In addition, like the cross-section analysis of Hypothesis 1, I also conduct a test in which small U.S. non-Big4 auditors (ranked by total assets) are the benchmark in the regression by including a large U.S. non-Big4 auditor indicator. In the second test, the three variables of interest are foreign Big4, foreign non-Big4, and large U.S. non-Big4 auditors. I conduct an F-test to examine whether the coefficients of the three variables of interest (i.e., foreign Big4, foreign non-Big4, and large U.S. non-Big4 auditors) are significantly different from each other in the three samples.  Table 2.7 shows the results from the probit estimation of (15). The results support Hypothesis 2a. Column 2 of Table 2.7 suggests that foreign Big4 auditors provide higher audit quality, as measured by ex-post restatement probability, than large U.S. non-Big4 auditors, which provide higher quality than small U.S. non-Big4 auditors. The results in Table 2.7 suggest that foreign non-Big4 auditors provide higher audit quality than U.S. non-Big4 auditors, which is inconsistent with Hypothesis 2b.   98  2.4.2 Supplemental Robustness Tests for Hypothesis 2 One concern about using the ex-post restatement as the measure of audit quality is that restatement is a joint outcome of the existence of an accounting irregularity and the firm reporting the error truthfully. The two-step process implies that the lower rate of restatement may be a result of better audit quality (i.e., a lower rate of audit failure) or non-timely detection and reporting of irregularity. For example, Srinivasan et al. (2015) find evidence that home country enforcement affects foreign firms’ likelihood of reporting restatement, and that for U.S.-listed foreign firms, less frequent restatements could be a signal of opportunistic reporting rather than better accounting quality. Although I attempt to mitigate this concern by including country fixed effects in regression Model 15, the concern of restatement remains valid.   In addition, using the ex-post restatement as the measure of audit quality has endogeneity concerns. Specifically, Datar et al. (1992) suggests that audit quality is an increasing function of the company’s type. Therefore, it could be that the lower restatement rate is a mere reflection of the clientele effect rather than the quality of the auditor (i.e. the production process of the auditor). In an ideal world, to be able to comment on the relative quality of auditors, one needs random assignment of clients to different auditors. One solution to overcome the problem is by utilizing the fact PCAOB is mandated to examine the audit process of auditors. In theory, PCAOB’s examination focus on the production process of the auditor and should be independent of the actual outcome of the audit. Therefore, analyzing the report of PCAOB could help to mitigate the concern of endogeneity. I re-test Hypothesis 2 using a unique measure of audit process quality derived from the PCAOB’s 99  inspection report20. Specifically, I measure the audit process quality of each individual auditor for a given year by the number of audit deficiencies of individual engagements identified during the PCAOB inspection process (given in the Part I findings of the PCAOB report). The unique feature of this measure is that board inspections are designed to identify and address weaknesses and deficiencies related to how an auditor conducts audits. I run the following regression for the sampling period between 2007 and 201321:  𝑁𝑢𝑚𝑏𝑒𝑟𝑜𝑓𝐴𝑢𝑑𝑖𝑡𝐷𝑒𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑖𝑒𝑠𝑡=   𝛼1𝑈𝑆_𝐵𝑖𝑔4𝑡+  𝛼2𝐹𝑜𝑟𝑒𝑖𝑔𝑛_𝐵𝑖𝑔4𝑡+  𝛼3𝐹𝑜𝑟𝑒𝑖𝑔𝑛_𝑛𝑜𝑛_𝐵𝑖𝑔4𝑡+  𝛼4𝑈𝑆_𝑛𝑜𝑛_𝑙𝑎𝑟𝑔𝑒_𝐵𝑖𝑔4𝑡+   𝛼𝑖𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑗 +   𝛼𝑘 𝑌𝑒𝑎𝑟 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡𝑠𝑡       (16) where 𝑁𝑢𝑚𝑏𝑒𝑟𝑜𝑓𝐴𝑢𝑑𝑖𝑡𝐷𝑒𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑖𝑒𝑠  is defined as the number of audits reviewed where deficiencies were identified in the PCAOB report, scaled by the number of clients included in the scope of the PCAOB's inspection. Two sets of independent variables (i.e., the average and median client characteristics of the auditor) are included to control for the auditor’s portfolio. 𝑁𝑢𝑚𝑏𝑒𝑟𝑜𝑓𝐴𝑢𝑑𝑖𝑡𝐷𝑒𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑖𝑒𝑠 is an inverse measure of audit process quality.  Column 1 of Table 2.10 reports the univariate estimation of the regression. Columns 2 and 3 also report the multivariate estimation conducted using the average and median client characteristics of the auditors. Consistent with the findings in Table 2.7, the coefficient of foreign Big4 auditors is smaller than large U.S. non-Big4 auditors, which is smaller than small U.S. non-                                                 20 The PCAOB is an organization established through the Sarbanes–Oxley Act of 2002 to oversee the audit of public companies. One mandate of the PCAOB is to inspect the audit processes of public accounting firms. U.S. Big4 accounting firms (and other auditors with more than 100 clients) are inspected by the PCAOB annually, whereas other auditors are inspected by the PCAOB at least once every 3 years. As part of the inspection process, the PCAOB issues a report about the inspected auditor’s audit process. The Part I Finding of the report summarizes all the major findings of the audit process. For more details, please see Aobdia (2016). The data of PCAOB’s inspection report are from Audit Analytics. 21 The sampling period is from 2007 because that is the first year for which data pertaining to the variable “number of clients included in the scope of PCAOB's inspection” are widely available.  100  Big4 auditors that give the baseline of the OLS regression, although the differences between foreign Big4 auditors and large U.S. non-Big4 auditors is not statistically significant. However, the results in Table 2.8 also indicate that foreign non-Big4 auditors provide lower audit quality than large U.S. non-Big4 auditors (coefficient difference: 0.159; P-value less than 1%). This differs from the results in Table 2.8, suggesting that clients with foreign non-Big4 auditors are less likely to restate than those with large U.S. non-Big4 auditors.   One could also infer quality of different groups of auditors by examing market’s perception of these auditors. Specially, if market reacts positively (negatively) to company’s announcement of auditor switching from one group to another, it could be inferred that the new auditor is of higher (lower) quality. Table 2.11 present evidence of market perception of auditor switches22. The evidence of in Table 2.11 is consistent with the findings of prior tables. Market reacts positively when the company switches from a US non-Big4 auditor to a foreign Big4 auditor which suggests that market perceives Foreign Big4 auditor to be of higher quality. Market reacts positively when the company switches from a foreign Big4 to a US Big4 auditor which suggests that market perceives Foreign Big4 auditor to be of lower quality than their US counterpart. Lastly, there is no systematic evidence between the perceived quality of US and foreign non-Big4 auditors.  In summary, the restatement, the PCAOB report and the market perception analysis suggest that foreign Big4 auditors provide higher audit quality than l U.S. non-Big4 auditors. These findings provide support for Hypothesis 2a. In addition, mixed evidence is obtained regarding the quality differences between foreign non-Big4 auditors and large U.S. non-Big4 auditors. Hence,                                                  22 In order to have accurate inferences, I only include the sample which the firms announce auditor dismissal and auditor appointment on the same day, even though this requirement constraint the sample considerably.  101  there is insufficient evidence to determine quality differences between these two groups of auditors, failing to reject Hypothesis 2b.  2.5 Testing Hypothesis 3 2.5.1 Research Design for Testing Hypothesis 3 Hypothesis 3 could be tested by modifying Model 15 used to test Hypothesis 2. Specifically, because this hypothesis is based on U.S. auditors, this study uses the Foreign Big4 auditors as the benchmark of the analysis by including three dummy variables (US_Big4, US_non-Big4, and Foreign_non-Big4) in the regression. The 12 controls are included. Three proxies (geographical distance, language, and client’s home country’s rule of law) are used to measure Distance, which represents the barriers that U.S. auditors face. The geographical distance and rule of law dummy variables are equal to zero if the geographical distance and rule of law of the client’s home country are below the median in the population. The non-English dummy is equal to zero if the country’s official language is English and equal to one otherwise. The model includes an interaction term of U.S. auditor indicators and the Distance indicators. Specifically, this study tests Hypothesis 3 using the following probit regression model: 𝑃𝑟(𝑅𝐸𝑆𝑇𝐴𝑇𝐸𝑀𝐸𝑁𝑇𝑖, 𝑡 , 𝐴𝐶𝐶𝑂𝑈𝑁𝑇𝐼𝑁𝐺𝑖, 𝑡 )  =  0+  1𝑈𝑆_𝐵𝑖𝑔4𝑡+2 𝑈𝑆_𝐵𝑖𝑔4𝑡 ∗𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 +  3𝑈𝑆_𝑛𝑜𝑛_𝐵𝑖𝑔4𝑡  +4𝑈𝑆_𝑛𝑜𝑛_𝐵𝑖𝑔4𝑡 ∗ 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒  + 5𝐹𝑜𝑟𝑒𝑖𝑔𝑛_𝑛𝑜𝑛_𝐵𝑖𝑔4𝑡 +  𝑖𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑗 + + 𝑗 𝐶𝑙𝑖𝑒𝑛𝑡𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡 +   𝑘 𝑌𝑒𝑎𝑟 𝐹𝑖𝑥𝑒𝑑 𝐸𝑓𝑓𝑒𝑐𝑡𝑠𝑡  (17)  The variables of interest are 2 and 4. Table 2.11 shows the results from the probit estimation of (17). The interaction terms of US_Big4 and Distance are non-significant across the three proxies, thus suggesting that the U.S. Big4’s audit quality does not vary across different proxies for country characteristics. This result supports Hypothesis 3a. The interaction term of 102  US_non_Big4 and Distance is negative and significant, suggesting that U.S. non-Big4 audit quality decreases with wider differences in country characteristics between the United States and the foreign country, supporting Hypothesis 3b.    2.6 Conclusion In this study, I examine the auditor aspect of foreign listings in the United States. The functional convergence hypothesis, which suggests that foreign auditors can bypass their home countries’ weak legal institutions and provide quality similar to U.S. auditors for their cross-listed clients because they must abide by U.S. securities laws and regulations, is tested. I find that despite prior studies documenting nuances to the functional convergence hypothesis, foreign Big4 auditors provide superior audit quality and charge higher audit fees than U.S. non-Big4 auditors of all sizes. No evidence has been found of any systematic differences in audit quality and fees between U.S. and foreign non-Big4 auditors. The quality of U.S. non-Big4 auditors decreases the further away the clients are from the United States, although this phenomenon does not exist for U.S. Big4 auditors. These results are robust according to various research specifications, such as change analysis and within-firm analysis.  The results suggest that the despite the problems of weak legal institutions in foreign countries, small U.S. portfolios, and the PCAOB’s barriers to monitoring, foreign Big4 auditors offer higher audit quality than U.S. non-Big4 auditors, which provide similar audit quality to foreign non-Big4 auditors. The findings differ from the traditional view that U.S. audit firms always provide higher audit quality than foreign auditors, particularly those that are not inspected by the PCAOB because of legal barriers (see Carcello et al. 2014). These findings have important implications for practitioners, regulators and other researchers.  103  2.7 Tables Table 2.1 – Sample Selection All Foreign Firms in Compustat and CRSP from 2000 to 2013              8,997  Less: Firm years with missing Audit Analytics data -               515  Less: Firm years with stock price below $1 -               464 Less: Financial and Utilities firms -            1,329    Total number of fyears               6,689   Total number of firms              1,201    104  Table 2.2 – Sample Distribution by Country  Country US_Big4 US_Non_Big4 Foreign_Big4 Foreign_Non_Big4 Total Argentina 1 0 66 0 67 Australia 3 1 71 8 83 Austria 0 0 5 0 5 Bahamas 18 0 19 0 37 Belgium 0 0 22 0 22 Belize 0 0 3 0 3 Bermuda 66 14 126 0 206 Brazil 0 0 79 1 80 Canada 29 21 1309 146 1505 Cayman Islands 10 1 16 0 27 Chile 0 1 91 0 92 China 3 232 579 142 956 Colombia 0 1 5 0 6 Curacao 12 0 0 0 12 Cyprus 0 0 2 0 2 Denmark 0 1 12 0 13 Dominican Republic 0 0 1 0 1 Finland 0 0 22 0 22 France 25 7 162 0 194 Germany 6 0 108 6 120 Greece 0 4 115 0 119 Guadeloupe 0 0 7 0 7 Hong Kong 14 23 160 28 225 Hungary 0 0 7 0 7 Iceland 6 0 1 0 7 India 0 0 101 14 115 Indonesia 0 0 23 0 23 Ireland 81 0 109 10 200 Israel 9 5 637 24 675 Italy 0 0 65 0 65 Japan 0 0 185 11 196 Korea (South) 0 2 65 0 67 Luxembourg 6 0 48 0 54 Macao 0 4 0 12 16 Marshall Islands 0 0 8 0 8 Mexico 0 0 133 28 161 Monaco 0 3 31 0 34 Netherlands 61 1 149 1 212 105  Table 2.2 (continued)   New Zealand 0 0 17 0 17 Norway 0 0 27 0 27 Panama 0 0 9 0 9 Papua New Guinea 0 0 8 0 8 Peru 0 0 17 0 17 Philippines 0 0 16 0 16 Portugal 0 0 12 0 12 Russia 0 0 63 0 63 Singapore 30 0 46 1 77 South Africa 0 0 87 0 87 Spain 0 2 32 7 41 Sweden 0 3 33 1 37 Switzerland 34 0 66 0 100 Taiwan (China) 0 7 96 6 109 Thailand 0 2 0 0 2 Turkey 0 0 12 0 12 United Arab Emirates 0 0 1 0 1 United Kingdom 58 4 334 8 404 Uruguay 0 0 2 0 2 Total 472 339 5424 454 6689     106  Table 2.2 (continued) – Descriptive Statistics   U.S. Big 4 auditors U.S. Non-Big 4 auditors  Foreign Big 4 auditors Foreign Non-Big 4 auditors Variable N Mean Median SD N Mean Median SD N Mean Median SD N Mean Median SD Log of Audit Fees 472 14.56 14.65 1.27 339 12.22 12.25 0.87 5424 13.42 13.48 1.94 454 11.90 12.01 1.45 Log of Assets 472 7.52 7.70 1.74 339 4.76 4.89 1.05 5424 7.30 7.32 2.37 454 5.12 4.94 1.62 Receivable Inventory Intensity 472 0.21 0.21 0.14 339 0.29 0.26 0.22 5424 0.19 0.15 0.15 454 0.26 0.23 0.20 Return on Assets 472 0.03 0.05 0.17 339 0.00 0.06 0.25 5424 0.00 0.04 0.18 454 -0.03 0.02 0.21 Loss 472 0.22 0.00 0.41 339 0.29 0.00 0.45 5424 0.30 0.00 0.46 454 0.41 0.00 0.49 Leverage 472 0.55 0.54 0.25 339 0.37 0.33 0.26 5424 0.46 0.46 0.23 454 0.34 0.31 0.23 Litigation Risk 472 0.24 0.00 0.43 339 0.35 0.00 0.48 5424 0.34 0.00 0.47 454 0.28 0.00 0.45 Busy Season 472 0.74 1.00 0.44 339 0.78 1.00 0.41 5424 0.81 1.00 0.39 454 0.69 1.00 0.46 Business Segments 472 2.42 2.00 1.58 339 1.96 1.00 1.32 5424 2.44 1.00 1.93 454 2.03 1.00 1.64 Geo Segments 472 3.87 3.00 3.06 339 1.53 1.00 1.45 5424 2.97 2.00 2.48 454 2.16 1.00 1.80 MB 472 3.14 2.25 4.65 339 2.29 1.22 3.80 5424 2.60 1.83 3.10 454 2.13 1.57 2.52 Return Volatility 472 0.12 0.10 0.07 339 0.21 0.17 0.12 5424 0.12 0.10 0.08 454 0.17 0.15 0.11 Going Concern 472 0.02 0.00 0.14 339 0.11 0.00 0.31 5424 0.02 0.00 0.15 454 0.07 0.00 0.25 Form_10K 472 0.82 1.00 0.38 339 0.76 1.00 0.43 5424 0.09 0.00 0.28 454 0.29 0.00 0.45  107  Table 2.3 – Analysis of Audit Fees (Test of Hypothesis 1)     Model One Firm Fixed Effect VARIABLES Expected Sign Coefficient Standard Error Coefficient Standard Error             US Big4 + 0.355*** 0.103 0.764*** 0.093 Foreign Big4 ? 0.364** 0.138 0.579*** 0.090 Foreign non-Big4 ? -0.173* 0.091 0.049 0.086 Log of Assets + 0.463*** 0.040 0.470*** 0.078 Receivable Inventory Intensity + 0.488*** 0.181 0.135 0.315 Return on Assets - -0.148 0.132 -0.335*** 0.059 Loss + 0.007 0.047 0.004 0.027 Leverage + 0.489** 0.194 0.129 0.148 Litigation Risk + 0.388*** 0.077 0.340*** 0.076 Busy Season + -0.022 0.072 0.466** 0.200 Business Segments + 0.046*** 0.015 0.004 0.017 Geo Segments + 0.067*** 0.013 0.015 0.012 MB + 0.005 0.008 -0.005 0.004 Return Volatility + -0.103 0.201 -0.049 0.201 Going Concern + 0.287* 0.157 0.134* 0.069 Form_10K + 0.379*** 0.066 -0.057 0.061 Constant  8.276*** 0.244 7.901*** 0.539       Observations  6,689 6,689 Adjusted R-squared  0.693 0.887 Country FE  YES NO Firm FE  NO YES Year FE   YES YES F-Test Between Coefficients        US Big4 – Foreign Big4  -0.09      0185*** Foreign_Big4 – Foreign_nBig4    0.537*  - 0.530*** Note: Coefficient estimates and standard errors are adjusted for country clusters. *, **, and *** indicate significance at the 5%, 1%, and 0.1% levels, respectively.   108  Table 2.4 – Analysis of Audit Fees with First Differences Regression – Supplemental Robustness Tests (Test of Hypothesis 1)              Large Sample Sample with Changing auditors VARIABLES Expected Sign Coefficient Standard Error             ∆ US Big4 + 0.817*** 0.102   ∆ Foreign Big4 ? 0.679*** 0.126   ∆ Foreign non-Big4 ? 0.278* 0.145   chg USnB4toFB4 ?   0.535** 0.158 ∆ Log of Assets + 0.394*** 0.033 0.551 0.433 ∆ Receivable Inventory Intensity + 0.052 0.112 0.091 0.269 ∆ Return on Assets - -0.321*** 0.058 0.223 0.184 ∆ Loss + -0.022 0.021 -0.189 0.251 ∆ Leverage + -0.021 0.051 1.556*** 0.381 ∆ Litigation Risk + -0.003 0.055 0.402*** 0.110 ∆ Busy Season + 0.141 0.304   ∆ Geo Segments + 0.011 0.011 0.024 0.066 ∆ Business Segments + 0.012 0.009 0.033 0.088 ∆ MB + -0.002 0.002 0.003 0.004 ∆ Return Volatility + 0.038 0.077 0.659 0.437 ∆ Going Concern + 0.041 0.055 -0.337*** 0.083 ∆ Form_10K  -0.092 0.114   Constant  0.104*** 0.025 0.187*** 0.020       Observations  5,307 117 0.253 YES Adjusted R-squared  0.0732 Year FE   YES F-Test Between Coefficients        US Big4 – Foreign Big4    0.138* Foreign_Big4 – Foreign_nBig4      0.401*** Note: Coefficient estimates and standard errors are adjusted for country clusters. *, **, and *** indicate significance at the 5%, 1%, and 0.1% levels, respectively.  109   Table 2.5 – Analysis of Audit Fees with U.S. Non-Big4 auditors Variation – Supplemental Robustness Tests (Test of Hypothesis 1)     Total Assets Total Market Capitalization Total number of Clients VARIABLES Expected Sign Coefficient Standard Error Coefficient Standard Error Coefficient Standard Error               US Big4 + 0.416*** 0.116 0.477*** 0.112 0.454*** 0.107 Foreign Big4 ? 0.425*** 0.153 0.488*** 0.148 0.465*** 0.141 Foreign non-Big4 ? -0.112 0.087 -0.051 0.088 -0.073 0.090 US non-Big4 large + 0.101*** 0.032 0.196*** 0.027 0.172*** 0.016 Log of Assets + 0.463*** 0.040 0.463*** 0.040 0.463*** 0.040 Receivable Inventory Intensity + 0.490*** 0.181 0.491*** 0.181 0.492*** 0.181 Return on Assets - -0.148 0.132 -0.146 0.132 -0.149 0.132 Loss + 0.007 0.047 0.008 0.047 0.007 0.047 Leverage + 0.487** 0.194 0.485** 0.193 0.487** 0.193 Litigation Risk + 0.387*** 0.077 0.386*** 0.077 0.388*** 0.077 Busy Season + -0.022 0.072 -0.022 0.072 -0.022 0.072 Business Segments + 0.047*** 0.015 0.047*** 0.015 0.046*** 0.015 Geo Segments + 0.067*** 0.013 0.067*** 0.013 0.067*** 0.013 MB + 0.005 0.008 0.005 0.008 0.005 0.008 Return Volatility + -0.098 0.200 -0.100 0.200 -0.100 0.200 Going Concern + 0.286* 0.157 0.287* 0.157 0.285* 0.157 Form_10K + 0.382*** 0.066 0.385*** 0.065 0.383*** 0.065 Constant  8.217*** 0.234 8.148*** 0.238 8.179*** 0.242 Observations  6,689 6,689 6,689 Adjusted R-squared  0.693 0.697 0.697 Country FE  YES YES YES Year FE   YES YES YES F-Test Between Coefficients        Foreign Big4 – US_non-Big4_large    0.324**   0.292**   0.293** Foreign_nBig4 – US_non-Big4_large   - 0.213** - 0.247** - 0.245** Note: Coefficient estimates and standard errors are adjusted for country clusters. *, **, and *** indicate significance at the 5%, 1%, and 0.1% levels, respectively. 110  Table 2.6 – Sample Selection of Restatement  All Foreign Firms from Table One              6,689 Less: Firms years from Countries without restating sample -               154   Total number of firm years for Hypothesis Two                                      6,535       Total number of firms              1,170    111  Table 2.7 – Analysis of Restatement Probability (Test of Hypothesis 2)      VARIABLES Expected Sign Coefficient Standard Error Coefficient Standard Error             US Big4 - -0.620*** 0.133 -0.692*** 0.140 Foreign Big4 - -0.383*** 0.126 -0.456*** 0.135 US non-Big4 large  ?   -0.124*** 0.026 Foreign non-Big4 - -0.252** 0.099 -0.322*** 0.104 Log of Assets - 0.009 0.035 0.009 0.035 Receivable Inventory Intensity + 0.357 0.288 0.354 0.288 Return on Assets - -0.057 0.136 -0.057 0.136 Loss + 0.083 0.064 0.082 0.064 Leverage + 0.230 0.174 0.234 0.173 Litigation Risk + -0.003 0.063 -0.001 0.064 Busy Season + -0.022 0.089 -0.023 0.088 Business Segments + 0.022 0.020 0.022 0.020 Geo Segments + -0.025** 0.011 -0.025** 0.011 MB + -0.001 0.008 -0.002 0.008 Return Volatility + -0.009 0.159 -0.008 0.159 Going Concern + 1.012*** 0.220 1.007*** 0.220 Form_10K + 0.570*** 0.086 0.568*** 0.086 Constant  -1.923*** 0.340 -1.856*** 0.342       Observations  6,535 6,535 Pseudo R-squared  0.128 0.128 Country FE  YES YES Year FE   YES YES F-Test Between Coefficients        Foreign_Big4 — US non-Big4 large   −0.272** Foreign_nBig4_large —  US non-Big4 large    −0.164** Note: Coefficient estimates and standard errors are adjusted for country clusters. *, **, and *** indicate significance at the 5%, 1%, and 0.1% levels, respectively.  112  Table 2.8 – PCAOB Sample Selection PCAOB Inspection Report Data from 2007 to 2013               1,694  Less: Auditor fyears not in the compustat sample  -               657 Less: Auditor fyears not in the foreign firm sample -               661   Total number of auditor years                   376    Total number of auditors                  188      113  Table 2.9 – Descriptive Statistics of PCAOB Analysis Sample  U.S. Big 4 auditors U.S. Non-Big 4 auditors  Foreign Big 4 auditors Foreign Non-Big 4 auditors Variable N Mean Median SD N Mean Median SD N Mean Median SD N Mean Median SD Number of Deficient Audits 28 0.31 0.31 0.12 70 0.46 0.47 0.31 196 0.34 0.33 0.35 82 0.52 0.50 0.40 Average Log of Assets 28 7.23 7.21 0.23 70 3.44 3.58 1.84 196 8.15 8.23 1.89 82 2.99 2.98 2.59 Average Receivable Inventory Intensity 28 0.24 0.23 0.03 70 0.30 0.28 0.13 196 0.20 0.18 0.13 82 0.20 0.17 0.18 Average Return on Assets 28 0.06 0.00 0.14 70 -5.20 -0.48 12.69 196 -0.13 0.03 1.26 82 -2.98 -0.16 10.53 Average Loss 28 0.28 0.26 0.05 70 0.55 0.51 0.24 196 0.24 0.04 0.29 82 0.68 1.00 0.41 Average Leverage 28 0.60 0.59 0.03 70 12.91 0.83 33.74 196 0.56 0.56 0.28 82 3.84 0.50 17.49 Average Litigation Risk 28 0.27 0.26 0.03 70 0.30 0.29 0.21 196 0.24 0.10 0.32 82 0.26 0.00 0.36 Average Busy Season 28 0.77 0.77 0.02 70 0.69 0.71 0.22 196 0.74 1.00 0.38 82 0.50 0.52 0.43 Average Business Segments 28 2.41 2.41 0.17 70 1.64 1.63 0.44 196 3.33 3.00 2.11 78 2.13 1.06 1.93 Average Geo Segments 28 3.00 2.94 0.23 70 1.99 1.65 1.40 196 5.30 4.04 3.20 56 4.85 2.75 4.33 Average MB 28 4.29 4.02 1.99 70 2.64 1.80 30.81 196 2.37 1.85 8.74 81 1.98 1.77 33.21 Average Return Volatility 28 0.11 0.10 0.03 70 0.17 0.15 0.07 196 0.11 0.10 0.06 49 0.26 0.22 0.13 Average Going Concern 28 0.02 0.02 0.01 70 0.30 0.26 0.25 196 0.04 0.00 0.12 82 0.39 0.20 0.43 Median Log of Assets 28 7.24 7.23 0.23 70 3.66 3.80 1.79 196 8.13 8.27 1.96 82 3.08 3.07 2.61 Median Receivable Inventory Intensity 28 0.18 0.18 0.02 70 0.25 0.24 0.17 196 0.18 0.15 0.14 82 0.18 0.13 0.19 Median Return on Assets 28 0.03 0.03 0.01 70 -0.21 -0.07 0.45 196 0.02 0.03 0.11 82 -0.54 -0.12 1.01 Median Loss 28 0.00 0.00 0.00 70 0.53 1.00 0.50 196 0.21 0.00 0.39 82 0.67 1.00 0.45 Median Leverage 28 0.58 0.57 0.03 70 0.60 0.52 0.35 196 0.52 0.52 0.17 82 0.78 0.44 1.00 Median Litigation Risk 28 0.00 0.00 0.00 70 0.14 0.00 0.33 196 0.21 0.00 0.38 82 0.23 0.00 0.39 Median Busy Season 28 1.00 1.00 0.00 70 0.86 1.00 0.34 196 0.78 1.00 0.41 82 0.54 0.75 0.48 Median Business Segments 28 1.50 1.50 0.51 70 1.06 1.00 0.31 196 3.23 3.00 2.46 82 2.09 1.00 2.16 Median Geo Segments 28 2.25 2.00 0.44 70 1.54 1.00 1.26 196 5.04 4.00 3.31 82 4.80 2.25 4.36 Median MB 28 1.77 1.77 0.28 70 1.78 1.06 7.44 196 2.29 1.69 4.47 82 5.04 1.75 14.97 Median Return Volatility 28 0.09 0.08 0.02 70 0.16 0.15 0.08 196 0.11 0.09 0.06 82 0.26 0.21 0.13 Median Going Concern 28 0.00 0.00 0.00 70 0.27 0.00 0.43 196 0.01 0.00 0.11 82 0.40 0.00 0.47                   114  Table 2.10 – PCAOB Analysis     No Controls Control for Average Client Characteristics  Control for Median Client Characteristics VARIABLES Expected Sign Coefficient Standard Error Coefficient Standard Error Coefficient Standard Error               US Big4 - -0.144*** 0.010 -0.224*** 0.051 -0.206*** 0.045 Foreign Big4 - -0.109*** 0.034 -0.197** 0.075 -0.153** 0.071 Foreign non-Big4 ? 0.085** 0.038 0.005 0.062 0.032 0.054 US non-Big4 large -   -0.133*** 0.026 -0.127*** 0.023 Log of Assets +   0.010 0.019 0.004 0.018 Receivable Inventory Intensity +   0.084 0.164 0.174 0.153 Return on Assets -   0.002 0.002 0.051 0.063 Loss +   0.004 0.061 0.000 0.050 Leverage +   0.001 0.001 0.020 0.054 Litigation Risk +   0.013 0.072 -0.012 0.072 Busy Season +   0.111 0.067 0.090 0.057 Business Segments +   -0.009 0.014 -0.011 0.013 Geo Segments +   0.006 0.008 0.007 0.008 MB +   -0.001 0.001 0.001 0.003 Return Volatility +   0.390 0.295 0.266 0.274 Going Concern +   0.022 0.139 0.055 0.081 Constant  0.447*** 0.003 0.309** 0.137 0.320** 0.128         Observations  376 376 376 Adjusted R-squared  0.096 0.128 0.133 Year FE   YES YES YES F-Test Between Coefficients        Foreign _Big4 – US_nBig4_large    - 0.064 -0.026 US_nBig4_large – Foreign_Non_Big4       0.138**     - 0.159*** Note: Coefficient estimates and standard errors are adjusted for country clusters. *, **, and *** indicate significance at the 5%, 1%, and 0.1% levels, respectively. 115  Table 2.11 – Market Reaction to the Announcement of Change of Auditors    Note: If a company switch its auditor from a Foreign (non) Big4 auditor to a US (non) Big4 auditor, its return is coded as negative of the raw return (– CAR).            Trading day/Event window N Mean abnormal return t-test     Cumulative abnormal returns (-1, +1)     Change_USB4_to_ForeignBig4 4 -0.0291 -3.32** Change_USnB4_to_ForeignBig4 17 +0.0289 1.97** Change_USnB4_to_ForeignnBig4 7 -0.0979 -0.963     116  Table 2.7.12 – Analysis of Restatement Probability with country variation (Test of Hypothesis 3)     Geographic Distance Non English Weak Rule of Law VARIABLES Expected Sign Coefficient Standard Error Coefficient Standard Error Coefficient Standard Error               US Big4  -  -0.290*** (0.0948) -0.249*** (0.0596) -0.265*** (0.0705) US_Big4_High_Distance ? 0.113 (0.216) -0.00199 (0.104) 0.0205 (0.106) US non-Big4  ? -0.538*** (0.150) -0.0931 (0.139) -0.0953 (0.142) US non-Big4_High_Distance + 1.151*** (0.122) 0.683*** (0.0621) 0.677*** (0.0634) Foreign non-Big4 + 0.164** (0.0826) 0.154* (0.0793) 0.153* (0.0794) Log of Assets - 0.00372 (0.0311) 0.00752 (0.0334) 0.00712 (0.0336) Receivable Inventory Intensity - 0.285 (0.296) 0.348 (0.288) 0.337 (0.289) Return on Assets + -0.0533 (0.136) -0.0788 (0.139) -0.0969 (0.141) Loss - 0.0731 (0.0622) 0.0722 (0.0620) 0.0654 (0.0621) Leverage + 0.249 (0.170) 0.234 (0.175) 0.238 (0.176) Litigation Risk + -0.0132 (0.0603) -0.00924 (0.0609) -0.00865 (0.0619) Busy Season + -0.0324 (0.0855) -0.0236 (0.0880) -0.0238 (0.0885) Business Segments + 0.0236 (0.0197) 0.0229 (0.0199) 0.0195 (0.0214) Geo Segments + -0.0262** (0.0115) -0.0258** (0.0116) -0.0241** (0.0116) MB + -0.00118 (0.00776) -0.000711 (0.00800) -0.000403 (0.00806) Return Volatility + 0.0108 (0.149) -0.00946 (0.162) -0.00643 (0.156) Going Concern + 0.962*** (0.227) 1.019*** (0.222) 1.008*** (0.221) Form_10K        + 0.584*** (0.0902) 0.566*** (0.0856) 0.571*** (0.0873) Constant  -2.202*** (0.321) -2.262*** (0.336) -2.248*** (0.333)                 Observations  6,535 6,535 6,535 Adjusted R-squared  0.133 0.130 0.129 Country FE  YES YES YES Year FE   YES YES YES Note: Coefficient estimates and standard errors are adjusted for country clusters. *, **, and *** indicate significance at the 5%, 1%, and 0.1% levels, respectively.   117  Conclusion This thesis examines the research questions regarding foreign firms listed in the United States. The functional convergence hypothesis suggests that when foreign firms are listed in the US, they are bonded to the US corporate governance environment. Therefore, one would expect that SEC oversight is similar for domestic and foreign firms. One would also expect foreign auditors behave similarly as their US counterparts.  Since the essays are presented in self-contained chapters, exhaustive discussion of the contribution and position of each essay in the broader accounting literature is provided in the conclusion specific to each chapter.  The first essay of this thesis is a co-author paper that is forthcoming at Review of Accounting Studies. It uses the existence of an SEC comment letter as a proxy of SEC monitoring, find foreign firms are subject to lower monitoring intensity US firms. The findings add nuance to the arguments used in the bonding literature which assumes SEC view all foreign firms as homogenous.  The second essay of this thesis examines the effects of cross-listings in the United States on the pricing of audits and quality of foreign auditors. I find that despite the nuances of the functional convergence hypothesis (such as the first essays) and negative public impression, foreign auditors are found to provide quality at least as good as the U.S. non-Big4 auditors. Specifically, foreign Big4 affiliates charge higher fees and provide better quality than U.S. non-Big4 auditors, and foreign non-Big4 auditors provide similar quality as their counterparts in the United States. My findings can mitigate some recent concerns about the quality of foreign auditors practicing in the U.S. cross-listing market.   118  Bibliography  Aobdia, D., & Shroff, N. (2016). Regulatory Oversight and Auditor Market Share, Working Paper  Asthana, S. C., Raman, K. K., & Xu, H. (2015). US-Listed Foreign Companies' Choice of a US-Based versus Home Country-Based Big N Principal Auditor and the Effect on Audit Fees and Earnings Quality. Accounting Horizons, 29(3), 631-666.  Badertscher, B., Jorgensen, B., Katz, S., & Kinney, W. (2014). Public equity and audit pricing in the United States. Journal of Accounting Research, 52(2), 303-339.  Bens, D. A., M. Cheng, and M. Neamtiu. 2016. The Impact of SEC Disclosure Monitoring on the Uncertainty of Fair Value Estimates. The Accounting Review 91 (2): 349-375.  Burnett, Brian M., Bjørn N. Jørgensen, and Troy J. Pollard. (2016). The Stock Market Reaction to Losing or Gaining Foreign Private Issuer Status. Journal of Accounting and Public Policy (forthcoming).  Carcello, J. V., Carver, B. T., & Neal, T. L. (2011). Market reaction to the PCAOB's inability to conduct foreign inspections. Available at SSRN 1911388.  Carcello, J. V., Carver, B. T., Lennox, C. S., & Neal, T. L. (2014). When bonding fails: Audit firm oversight of US-listed Chinese, Working Paper  Carson, E. (2009). Industry specialization by global audit firm networks. The Accounting Review, 84, 355-382.  Cassell, C., L. Dreher, and L. Myers. 2013. Reviewing the SEC’s review process: 10-K comment letters and the cost of remediation. The Accounting Review 88 (6): 1875-1908.  Chen, K. C., Cheng, Q., Lin, Y. C., Lin, Y. C., & Xiao, X. (2016). Financial Reporting Quality of Chinese Reverse Merger Firms: The Reverse Merger Effect or the Weak Country Effect?. The Accounting Review, 91(5), 1363-1390.  Choi, J. H., Kim, J. B., Liu, X., & Simunic, D. A. (2008). Audit pricing, legal liability regimes, and big 4 premiums: Theory and Cross‐country evidence*. Contemporary Accounting Research, 25(1), 55-99.  Choi, J. H., Kim, J. B., Liu, X., & Simunic, D. A. (2009). Cross-listing audit fee premiums: Theory and evidence. The Accounting Review, 84(5), 1429-1463.  Christensen, H., L. Hail and C. Leuz. 2016. Capital-Market Effects of Securities Regulation: Prior Conditions, Implementation, and Enforcement. Review of Financial Studies 29:2885-2924  Cunningham, L., B.A. Johnson, E.S. Johnson, and L.L. Lisic. 2017. The switch up: An examination of changes in earnings management after receiving SEC comment letters. Working paper, University of Tennessee, George Mason University, and Virginia Polytechnic University.  DeAngelo, L. E. (1981). Auditor size and audit quality. Journal of accounting and economics, 3(3), 183-199.  Dechow, P. M., A. Lawrence and J. Ryans. 2015. SEC Comment Letters and Insider Sales. Accounting Review, forthcoming.  DeFond, Mark, David H. Erkens, and Jieying Zhang. "Do client characteristics really drive the Big N audit quality effect? New evidence from propensity score matching." Management Science (2016). 89   119  Doidge, C., G.A. Karolyi, and R. Stulz. 2004. Why are foreign firms listed in the US worth more? Journal of Financial Economics 71: 205–238.  Deloitte. 2013. SEC comment letters—including industry insights constructing clear disclosures. Available at: http://www.deloitte.com/assets/Dcom-UnitedStates/  Eleswarapu, V. R., & Venkataraman, K. (2006). The impact of legal and political institutions on equity trading costs: A cross-country analysis. Review of Financial Studies.  Francis, J. R., & Wang, D. (2008). The joint effect of investors protection and big 4 audits on earnings quality around the world. Contemporary Accounting Research, 25, 157-191.  Ghosh, A. A., & Wagner, E. P. (2014). Are US Cross Listed Chinese Companies Tainted by the Poor Audit Quality of Reverse Mergers? The Role of Depository Banks, Working Paper  Guo, Yiheng. "Audit Quality Control: Big Four Global Member Firms and Audit Services." (2016), Working Paper  Hay, D. C., Knechel, W. R., & Wong, N. (2006). Audit fees: A Meta‐analysis of the effect of supply and demand attributes. Contemporary Accounting Research, 23(1), 141-191.  Hail, Luzi, and Christian Leuz. 2006. International differences in the cost of equity capital: Do legal institutions and securities regulation matter? Journal of accounting research 44.3: 485-531.  Heese, J., M. Khan and K. Ramanna. 2017. Is the SEC captured? Evidence from comment-letter reviews. Journal of Accounting and Economics. 64.1: 98-122.  Iliev, P., D. Miller, and L. Roth. 2014. Uninvited US investors? Economic consequences of involuntary cross-listing. Journal of Accounting Research 52: 473–519.  Jackson, H., and M. Roe. 2009. Public and private enforcement of securities laws: Resource-based evidence. Journal of Financial Economics 93: 207–238.  Johnston, R. and R. Petacchi. 2016. Regulatory Monitoring of Financial Reporting: Securities and Exchange Commission Comment Letters. Contemporary Accounting Research, forthcoming.  Kubick, T. R., D. P. Lynch, M. A. Mayberry and T. C. Omer. 2016. The Effects of Regulatory Scrutiny on Tax Avoidance: An Examination of SEC Comment Letters. The Accounting Review 91 (6): 1751-1780.  La Porta, R., F. López de Silanes, A. Shleifer, 2006. What Works in Securities Laws? Journal of Finance 61, 1-32.  Lang, M., J. S. Raedy, and W. Wilson. 2006. Earnings management and cross listing: Are reconciled earnings comparable to US earnings? Journal of Accounting and Economics 42.1: 255-283.  Lawrence, A., Minutti-Meza, M., & Zhang, P. (2011). Can Big 4 versus non-Big 4 differences in audit-quality proxies be attributed to client characteristics?. The Accounting Review, 86(1), 259-286.  Lee, C. M., Li, K. K., & Zhang, R. (2014). Shell games: The long-term performance of Chinese reverse-merger firms. The Accounting Review, 90(4), 1547-1589.  Lenz, H., & James, M. L. (2007). International audit firms as strategic networks—The evolution of global professional service firms. In G. Cliquet, M. Tuunanen, G. Hendrikese, & J. 90   120  Windsperger (Eds.), Economics and management of networks (pp. 367-392). Heidelberg, Germany: Physica-Verlag.  Lundholm, R. J., Rogo, R., & Zhang, J. L. (2014). Restoring the tower of Babel: How foreign firms communicate with US investors. The Accounting Review, 89(4), 1453-1485.  Magnan, M. L. (2008). Discussion of “Audit Pricing, Legal Liability Regimes, and Big 4 Premiums: Theory and Cross‐country Evidence”. Contemporary Accounting Research, 25(1), 101-108.  Naughton, J. P., Rogo, R., Sunder, J., & Zhang, R. (2016). SEC monitoring of foreign firms’ disclosures. Review of Accounting Studies, forthcoming Piotroski, J. D. and S. Srinivasan. 2008. Regulation and Bonding: The Sarbanes-Oxley Act and the Flow of International Listings. Journal of Accounting Research, 46: 383–425.  Ryans, J. 2015. Textual Classification of SEC Comment Letters. Working Paper, University of California, Berkeley.  Saito, Y., & Takeda, F. (2014). Global audit firm networks and their reputation risk. Journal of Accounting, Auditing & Finance, 29(3), 203-237.  Securities and Exchange Commission (SEC). 2012a. Study on the cross-border scope of the private right of action under Section 10(b) of the Securities Exchange Act of 1934. Washington, D.C.  Securities and Exchange Commission (SEC). 2012b. Study on the cross-border scope of the private right of action under Section 10(b) of the Securities Exchange Act of 1934. Washington, D.C.  Securities and Exchange Commission (SEC). 2013. Division of Corporation Finance: Filing Review Process. Available at: http://www.sec.gov/divisions/corpfin/cffilingreview.htm  SEC Insight Inc. 2004. Response to “Proposal to Publicly Disclosure Staff Comment Letters and Non-Confidential Response Letters. September 30, 2004. Available at: http://www.sec.gov/news/press/s72804/secinsight093004.pdf  Shnister, N., 2010. A free pass for foreign firms? An assessment of SEC and private enforcement against foreign issuers. Yale Law Journal, 119, 1638-1701.  Skinner, D. J., & Srinivasan, S. (2012). Audit quality and auditor reputation: Evidence from Japan. The Accounting Review, 87(5), 1737-1765.  Simunic, D. A. (1980). The pricing of audit services: Theory and evidence. Journal of Accounting Research, 161-190.  Siegel, J. 2005. Can foreign firms bond themselves effectively by renting US securities laws? Journal of Financial Economics, 75(2), 319-359.  Srinivasan, S., A. Wahid, and G. Yu. 2015. Admitting mistakes: Home country effect on the reliability of restatement reporting. Accounting Review, forthcoming.  Tafara, E., and R. J. Peterson. 2007. A blueprint for cross-border access to US investors: A new international framework. Harvard International Law Journal, 48(1): 31-68.   121  Appendix A: Chapter 1 A.1 Variable Definitions Variable Description Data Source Panel A: Dependent Variables and Main Variable of Interest SEC_Review Indicator variable set equal to 1 if the firm received a comment letter for a period t filing, and 0 otherwise. Audit Analytics Sec_Review_Words The number of words in each comment letter issued as part of a conversation between the SEC and the firm. Audit Analytics SEC_Review_Filing Count variable equal to the number of financial filings with comment letters. Audit Analytics SEC_Review_2 Indicator variable that takes the value 1 if the firm received a comment letter in year t and at least one comment letter in the previous two years, and 0 otherwise. Audit Analytics SEC_Review_Alt Indicator variable equal to 1 for period t when (a) a firm receives a comment letter during period t or (ii) when a firm has not received a comment letter for the period t-2 through period t filing, and 0 otherwise. Audit Analytics Foreign_Firm Indicator variable set equal to 1 if firm is classified as a foreign private issuer by the SEC for fiscal year t, and 0 otherwise. SEC Website Single_Listed Indicator variable set equal to 1 if foreign firm’s shares only trade on a US exchange, and 0 otherwise. BNY Mellon Cross_Listed Indicator variable set equal to 1 if foreign firm’s shares are cross-listed on both a US and foreign exchange, and 0 otherwise. BNY Mellon US_Exposure The percent of the firm’s market capitalization that is traded on US exchanges, determined as the ratio of the US market capitalization to the total market capitalization of the firm. Bloomberg or by manually inspecting the firm’s 20-F filing 122  Variable Description Data Source Panel B: Measures of Public Enforcement Rules_Enforcement Formal Public Enforcement Index. The arithmetic mean of: (1) supervisor characteristics index; (2) its rule-making power index; (3) its investigative powers index; (4) orders authority index; and (5) criminal authority index, as La Porta, Lopez-de-Silanes, and Shleifer (2006) describe. La Porta, Lopez-de-Silanes, and Shleifer (2006)   Staff_Resources  Staff Resource Based Public Enforcement Index. The number of the securities regulators’ staff in 2005, divided by the country’s population in millions based on the extended sample, as Jackson and Roe (2009) describe. Jackson and Roe (2009)  Budget_Resources  Budget Resource Based Public Enforcement Index. The securities regulators’ 2005 budget divided by the country’s GDP based on the extended sample, as Jackson and Roe (2009) describe. Jackson and Roe (2009)  Panel C: Measures of Private Enforcement Disclosure_ Requirements The index of disclosure equals the arithmetic mean of: (1) nature of liability on a prospectus; (2) extent compensation must be disclosed; (3) shareholders’ disclosure; (4) extent inside ownership must be disclosed; (5) extent irregular contracts must be disclosed; (6) and the extent that related party and irregular transactions must be disclosed, as La Porta, Lopez-de-Silanes, and Shleifer (2006) describe. La Porta, Lopez-de-Silanes, and Shleifer (2006)  Liability_   Standards The index of liability standards equals the arithmetic mean of: (1) liability standard for the issuer and its directors; (2) liability standard for the distributor; and (3) liability standard for the La Porta, Lopez-de-Silanes, and Shleifer (2006)  123  Variable Description Data Source accountant, as La Porta, Lopez-de-Silanes, and Shleifer (2006) describe.  Panel E: Controls for Accounting Quality Material_Weakness Indicator variable set equal to 1 if the internal control audit opinion (under SOX Section 404) or the management certification (under SOX Section 302) as reported in Audit Analytics is qualified for a material weakness in any years of t, , t – 1, or t – 2,  and 0 otherwise. Audit Analytics Restatement Indicator variable set equal to 1 if the firm filed a 10-K restatement in any years of t, t – 1, or t – 2, and 0 otherwise.   Audit Analytics IFRS Indicator variable equal to 1 for period t when the firm reports using IFRS. Audit Analytics Small_NI Indicator variable set equal to 1 if for firm-years when annual net income scaled by total assets is between 0 and 0.01, a 0 otherwise (see Lang, Raedy, and Wilson, 2006). Compustat Panel F: Auditor Related Controls Auditor_Big4 Indicator variable set equal to 1 if the firm’s auditor is a Big 4 audit firm, and 0 otherwise. Audit Analytics Auditor_2Tier Indicator variable set equal to 1 if the firm’s auditor is a second tier audit firm(i.e., BDO Seidman, Crowe Horwath, Grant Thornton, or McGladrey & Pullen), and 0 otherwise. Audit Analytics Auditor_Tenure The number of years during which the auditor has audited the firm. Audit Analytics Auditor_Dismiss Indicator variable set equal to 1 if the auditor was dismissed in any years of t, t – 1, or t – 2, and 0 otherwise. Audit Analytics 124  Variable Description Data Source Auditor_Resigned Indicator variable set equal to 1 if the auditor resigned in any years of t, t – 1, or t – 2, and 0 otherwise. Audit Analytics Panel G: Other Variables RetVol_High Indicator variable set equal to 1 if the volatility of abnormal monthly stock returns (equal to the monthly return [RET] minus the value weighted return [VWRTD]) is in the highest quartile in a given fiscal year, and 0 otherwise. Return volatility is calculated as the standard deviation of monthly stock returns for the 36-month period ending in the last month of the fiscal year. CRSP MarketCap The natural log of market capitalization, calculated as shares outstanding at fiscal year-end (CSHO) times the share price at fiscal year-end (PRCC_F). Compustat Firm_Age The age of the firm. Compustat Loss Indicator variable set equal to 1 if net income (NI) is negative in any years of t , t – 1, or t – 2, and 0 otherwise. Compustat Sales_Growth The mean of sales growth in years t, t – 1, and t – 2, where sales growth is measured as the percentage change (in decimal form) in annual sales (REVT). Compustat Bankruptcy The decile ranks of Altman’s Z-score, where companies with the weakest financial health are assigned to decile 10. Altman’s Z-score is measured following Altman (1968). Compustat Num_Segments The number of business segments reported in the Compustat segments database. Compustat Ext_Financing The sum of equity financing and debt financing scaled by total assets, measured in t+1, following Ettredge et al. (2011). Equity financing equals the sales of common and preferred stock Compustat 125  Variable Description Data Source (SSTK) minus the purchases of common and preferred stock (PRSTKC) and dividends (DV). Debt financing equals long-term debt issued (DLTIS) minus long-term debt reduction (DLTR) minus the change in current debt (DLCCH). Restructuring Indicator variable set equal to 1 for non-zero restructuring costs as reported on a pre-tax basis (RCP) in any years of t , t – 1, or t – 2, and 0 otherwise. Compustat M&A Indicator variable set equal to 1 for non-zero mergers or acquisitions as reported on a pre-tax basis (AQP) in year t, and 0 otherwise. Compustat Litigation Risk Indicator variable set equal to 1 if the firm SIC code is one of the following: 2833-6, 3570-7, 3600-74, 5200-5961, or 7370-4, and 0 otherwise.  Compustat Inst_Ownership Total institutional holdings minus institutional holdings held by institutions categorized as ‘‘transient’’ in the quarter immediately preceding fiscal year-end divided by the total shares outstanding as of fiscal year-end, winsorized to 1.00 (following D’Souza et al., 2010). We identify transient institutions using data from http://acct3.wharton.upenn.edu/faculty/bushee/IIclass.html Thomson (S34) GDP The log of 2005 gross domestic product in US dollars. Derived from GDP variables downloaded from World Bank Data and Statistics website, as Jackson and Roe (2009) describe. Jackson and Roe (2009)    127  A.2 Figure -- Overview of Domestic versus Foreign Private Issuer Reporting and Disclosure Requirements Obligation Domestic Issuer Foreign Private Issuer  Exchange Act Registration Forms    Form 10, which requires SEC-specified disclosure regarding the U.S. domestic issuer and is subject to SEC review.  Form 20-F, which requires SEC-specified disclosure regarding the Foreign Private Issuer and is subject to SEC review.   Exchange Act Reporting Forms  Form 10-K for annual information required by the SEC, including annual audited financial statements. Form 10-Q for interim period financial and other information. Form 8-K for disclosure of specified material events.  Form 20-F for annual information, including annual audited financial statements. Form 6-K for all other material information disclosed by the Foreign Private Issuer according to home-country or stock exchange requirements.   Annual Reporting Form 10-K prescribes specific disclosures and must be filed within 60-90 days after fiscal year end. Form 20-F prescribes specific disclosures and must be filed within 4 months after fiscal year end.  Quarterly Reporting Must file quarterly reports on Form 10-Q.   Not required. Periodic Reporting Must file Form 8-K generally within 4 business days of event to be reported.  Prescribes specific disclosures to be made.  Form 6-K to be furnished promptly, after information is made public in home jurisdiction. No prescribed specific disclosures. Foreign Private Issuers that produce interim financial statements due to home country requirements disclose those statements in the US using Form 6-K.  Required Accounting Standards Financial statements typically prepared in accordance with U.S. GAAP.  Must reconcile to U.S. GAAP, unless financial statements are prepared in accordance with IFRS. 128  Appendix B: Chapter 2 B.1 Variable Definitions Variable Definition Business Segments Number of business segments Busy Season Indicator variable; equal to 1 if the client’s fiscal year-end month is December (the busy season for audits), and 0 otherwise Foreign Big4 Indicator variable; equal to 1 if the client’s auditor is a foreign Big4 audit firm, and 0 otherwise Foreign non-Big4 Indicator variable; equal to 1 if the client’s auditor is a foreign non-Big4 audit firm, and 0 otherwise Foreign 10k Indicator variable; equal to 1 if the firm files a 10-K statement, and 0 otherwise Geo Segments Number of foreign segments Going Concern Indicator variable; equal to 1 if the firm received a growing concern opinion, and 0 otherwise Leverage Total long-term debt (Compustat DLTT) divided by total assets (Compustat AT) Litigation Risk Indicator variable; equal to 1 if the firm SIC code is one of the following: 2833-6, 3570-7, 3600-74, 5200-5961, or 7370-4, and 0 otherwise Log of Assets Natural logarithm of total assets (Compustat AT) 129  Variable Definition Log of Audit Fees Natural logarithm of annual audit fees Loss Indicator variable; equal to 1 if net income (Compustat NI) is less than 0, and 0 otherwise Market-to-book (MB) Market Cap (Compustat PRCC_F times CSHO) divided by equity of the firm (Compustat AT minus Compustat LT) Receivable Inventory Intensity Sum of account receivables and inventory (Compustat RECT and INVT) divided by total assets (Compustat AT)  Return on Assets Net income (Compustat N) divided by total assets (Compustat AT) Return Volatility Standard deviation of daily returns for the 12-month period ending on the last month of the fiscal year US Big4 Indicator variable; equal to 1 if the client’s auditor is a US Big4 audit firm, and 0 otherwise   130  B.2 Stylished Model This section presents a stylized model explaining how audit costs, audit fees, and audit quality may be related to an auditor’s country of origin. Observations based on this model are formulated to help understand and to develop the empirical predictions of this paper. Consider a single-period model in which the audit market is segmented into different groups of auditors. The audit market is perfectly competitive within its own segment (i.e., Big4s compete against other Big4s, and non-Big4s compete against other non-Big4s, but Big4s and non-Big4s do not compete against each other). Investors may hire an auditor to issue an audit report with respect to the project type. This project is good with probability P23, and with probability 1 – P, the project is bad. An auditor is hired at the beginning of the period, and the auditor charges a fixed fee at the end of the period and issues an audit report. The auditor may be viewed as a vehicle to extract information on the project type. The auditor exerts effort q ∈ [0, 1] in the audit process and obtains a binary signal about the project type (i.e., signal ∈ {good, bad}). The level of audit effort is not publicly observable when the auditor’s report is issued, and it becomes observable only in the litigation process if the project turns out to be bad. The auditor might not be able to detect a bad project because of technological limitations or inadequate effort. Consistent with the literature, the relation between the audit report and the true type of project is: Pr( | , ) 1g good q   and Pr( | , )b bad q q  (1) where g and b denote the auditor’s signal that the project is good and bad, respectively, and q denotes the level of effort and the probability of detecting the failed project conditional on a failed project and effort q. This assumption implies that the auditor will make no mistakes in detecting a                                                  23 Note: that the probability of the project is not dependent on the auditor choice (i.e. no adverse selection) 131  good project (i.e., there are no Type I errors), but the auditor may fail to detect a bad project depending on his or her audit effort (i.e., Type II errors may occur). The auditor is assumed to be independent and to report the observed signal truthfully. The auditor faces a variable cost K for his or her effort. Audit failure occurs when the project fails but the auditor issues a “good” report; that is, the auditor fails to detect material misstatements in the financial statements. The probability of audit failure conditional on the project’s failure is 1 – q. Therefore, q can also be interpreted as the quality of the audit. The investors will sue the auditor if an audit failure occurs, and the probability of the auditor being found liable is r (1 – q), where r ∈ [0, 1] represents the strength or strictness of the financial reporting regime. In this model, r is assumed to be a constant because all foreign firms are under the legal regime of the United States. One can interpret r (1 – q) as the probability of the auditor being found negligent and liable in court in the event of an audit failure. An auditor who is found liable is required to pay the entire loss I or his or her wealth, whichever is smaller (i.e., min [I, W]). The concern that it is more difficult for the Securities and Exchange Commission (SEC), PCAOB, and U.S. court to enforce the penalty on the foreign auditors is captured by the differences in wealth W under the radar of the SEC, PCAOB, and U.S. court (i.e. Wforeign < Wdomestic).  The auditor’s objective is to exert an effort level that maximizes his or her profit. The auditor’s revenue is the audit fee, denoted by F, minus the total expected cost, which is the sum of the costs of the audit effort plus the expected liability payments: 21( ) ( )2F EL q q K D   (2) where the expected liability the investors can obtain from the auditor who exerts effort q is ( ) (1 )(1 )[ (1 )]min( , )EL q p q r q I W     (3). 132  Note that because of limited liability, the auditor’s liability is W if W I , or I  if I W . Because auditors’ incentive to work is limited by individual wealth and because auditors have heterogeneous wealth levels, the auditor’s wealth is the key determinant of his or her audit effort. This assumption is consistent with those in the literature (e.g., Choi et al. 2008; Choi et al. 2009; Ye and Simunic 2011).  There are two innovative elements in the present model. First, there is an assumption of distance (D), which represents the transaction costs for the auditor to audit the investment. In the real world, D would represent factors of influence auditor’s direct out-of-pocket costs or auditor’s marginal labor productivity. Traveling cost is an example of factors that influence auditor’s direct out-of-pocket costs. For the same travelling budget, the local auditor could afford sending five staffs to do the field work while the U.S. auditors would only afford three. In addition to the out-of-pocket costs, for the same hour of audit work, local auditor may be more efficient because of its knowledge of the local business. In sum, Distance (D) affects the equilibrium such that the per unit cost of effort, q, of an auditor in the United States is higher than the ones of an auditor in the foreign country (i.e., K(DUS) > K(Dforeign)). An easy way to interpret D is as the geographical distance between the U.S. and the foreign country.   The second innovative elements are that I remove the restrictions that the wealth of Big4 auditors are not universal. While each Big4 auditors always market itself as one global firm around the world, each Big4 network consists of local offices which are separate legal entities. Each local office is liable only for its own acts or omissions and they are not legal partners with each other. Conceptually, wealth of Big4 auditors consists of two components: the reputation of their network, which is similar across the globe and the “money in their bank account”, which is a local 133  component. It is assumed that the U.S. Big4 have more “money in their bank account” than their foreign affiliates.  In equilibrium, the auditor maximizes profit by choosing an optimal amount of auditor effort:   2 21 1arg max ( ) ( ) (1 )(1 )[ (1 )]min( , ) ( , _ )2 24qF EL q q C D F p q r q I W q K D Audito type       . Additionally, in equilibrium, the auditor has zero profit; therefore,  21 ( , _ ) (1 )(1 )[ (1 )]min( , )2F q K D Audito type p q r q I W      (5).  From (5), the envelope theorem implies that if an auditor has a larger amount of wealth and higher marginal costs, he or she will charge higher audit fees. This finding leads to the following observation. Observation 1: The audit fee associated with a large U.S. audit firm is higher than that of any foreign auditor because of greater expected liability (greater wealth) and higher marginal costs. This study next examines a more general setting, in which wealth and cost differences vary between domestic and foreign auditors: In equilibrium, the auditor chooses the optimal level of q. Setting 0Fq yields  ( , _ ) 2 (1 )min( , )(1 ) 0FqK D Audito type r p I W qq     (6)  and * 2 (1 )min( , )( , _ ) 2 (1 )min( , )r p I WqK D Auditor type r p I W  (7). When (7) is substituted into (5), (8) is developed as follows: 134  * (1 ) min( , ) ( , _ ) 11 2( , _ ) 2 (1 )min( , )(1 )min( , ) ( , _ )r p I W K D Auditor typeFK D Auditor type r p I Wr p I W K D Auditor type   (8) From (8), it is clear that the audit fee is a function of W and K such that higher auditor wealth and marginal audit costs are associated with higher audit fees. This finding leads to the following observation. Observation 2: Whether a U.S. auditor charges higher (or lower) audit fees than a foreign auditor depends on the relative weight of the expected legal liability and audit operation costs of the audit production process.  Equation (7) could be rearranged to * 2 (1 ) min( , ) 1( , _ )( , _ ) 2 (1 ) min( , )12 (1 ) min( , )r p I WqK D Auditor typeK D Auditor type r p I Wr p I W   (9)  which can be used to interpret the audit quality differences between U.S. auditors and foreign auditors. From (9), one can interpret that audit quality is lower when the marginal audit cost K is high and when the auditor wealth W is low. This result is intuitive because the auditor faces a trade-off between actual audit costs and the expected legal liability costs, and these two costs are substitutes.   One can investigate the relative quality of the U.S. and foreign auditors by examining the equilibrium quality using (9):  **( , )12 (1 ) min( , )10( , )12 (1 ) min( , )foreigndomesticforeigndomesticK D Foreignr p I WqK D Domesticqr p I W . 135  From (10), it is clear that whether a foreign auditor has higher (or lower) quality depends on the relative differences in the marginal costs between the foreign and domestic auditors and on the magnitude of wealth differences between the foreign and domestic auditors. It is therefore unclear whether U.S. auditors have superior quality than foreign auditors because the foreign auditors have lower audit costs from being closer to the clients but also have lower wealth. Equation (10) thus leads to the following observation. Observation 3: Whether a U.S. auditor produces a higher (or lower) audit quality than a foreign auditor depends on the relative weight of the U.S. auditor’s size and the foreign auditor’s cost advantage during the audit production process.  Even though q, the ex-ante level assurance, is not publicly observable when the auditor’s report is issued, one may infer q from the audit fee data, which are publicly available. Specially, if the foreign auditor charges higher audit fees than the U.S. auditor (i.e., 𝐹𝑓𝑜𝑟𝑒𝑖𝑔𝑛_𝑎𝑢𝑑𝑖𝑡𝑜𝑟  >𝐹𝑑𝑜𝑚𝑒𝑠𝑡𝑖𝑐_𝑎𝑢𝑑𝑖𝑡𝑜𝑟), then under the assumption that the foreign auditor has a cost advantage over the U.S. auditor (i.e., 𝐾𝑓𝑜𝑟𝑒𝑖𝑔𝑛_𝑎𝑢𝑑𝑖𝑡𝑜𝑟< 𝐾𝑑𝑜𝑚𝑒𝑠𝑡𝑖𝑐_𝑎𝑢𝑑𝑖𝑡𝑜𝑟 ), from (8), one could infer that 1 1(1 )min( , ) (1 )min( , )foreign domesticdomestic foreignW Wr p I W r p I W    Equation (10) could be re-arranged as:  **( , ) 2 (1 )min( , ) min( , )*( , ) 2 (1 )min( , ) min( , )foreigndomestic domesticforeign domestic foreignK D Foreign r p I Wq I Wq K D Domestic r p I W I W   Because foreign domesticW W  and K(DUS) > K(D foreign), It is obvious that equation (11) is smaller than 1. This inference leads to the following interpretation of (11):  **( , ) 2 (1 )min( , ) min( , )* 1 11( , ) 2 (1 )min( , ) min( , )foreigndomestic domesticforeign domestic foreignK D Foreign r p I Wq I Wq K D Domestic r p I W I W    136  Note that this is a sufficient, but not necessary condition (i.e., the foreign auditor can produce a higher-quality audit for a smaller fee because this auditor has a cost advantage). This yields the following observation.  Observation 4: If the foreign auditor charges a fee premium over the U.S. auditor, the audit quality must be higher.   Finally, exploring how the audit quality of U.S. versus foreign auditors differs across countries is appealing. Specifically, the question as to how the quality of U.S. auditors changes when the transaction costs are high warrants exploration. This could be explored by taking the partial derivatives of (10) with respect to distance, D24.**min( , ) 1 ( , )* *min( , ) ( , ) 2 (1 ) min( , )( , ) 2 (1 ) min( , )min( , )*min( , ) [ ( , ) 2 (1 ) mdomesticforeign domesticforeign domesticforeigndomesticforeignqq I W K D foreignD I W K D Domestic r p I W DK D Foreign r p I WI WI W K D Domestic r p       2( , )* (12)in( , )]domesticK D domesticI W D Note that ( , )KD ForeignD= 0 because a foreign auditor’s distance to their clients does not vary with the country’s distance (e.g. geographical distance) from the United States, but( , )KD DomesticD> 0 because U.S. auditors’ marginal costs increase as the distance between the United States and the foreign country increases. Thus, (12) can be simplified as  **2( , ) 2 (1 ) min( , )min( , ) ( , )* * 0(13)min( , ) [ ( , ) 2 (1 ) min( , )]domesticforeign foreigndomesticforeign domesticqq K D Foreign r p I WI W K D domesticD I W K D Domestic r p I W D                                                         24 Note that I use distance and difference interchangeably in this paper. D would represent the physical distance from U.S. to the foreign country or the differences in languages or legal environment.  137  This yields the following observation.  Observation 5: U.S. auditors are more likely to provide lower quality when the foreign country is far from the United States.   The rationale behind this observation is that, in contrast to foreign auditors who have local expertise, when a foreign country varies considerably from the United States in language, business culture, or geographical distance, the U.S. auditors face higher operating costs. This leads to lower quality audits25.                                                      25 An assumption of this observation is that foreign auditors’ wealth does not vary across different countries. If this assumption is relaxed, such that foreign auditors’ wealth decreases as the difference from the United States increases (i.e., it is more difficult for the U.S. court and regulator to charge a foreign auditor), then it is uncertain how the relative audit quality would change across different countries because these two forces (i.e., change of auditor wealth and marginal costs) would offset each other. I expect that the former is immaterial compared with the latter.  138  B.3 Graphical Representation of the hypothesis       

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0368915/manifest

Comment

Related Items