Early Evidence from Canadian Firms Choice Between IFRS and U.S. GAAP *

Transcription

1 Early Evidence from Canadian Firms Choice Between IFRS and U.S. GAAP * Brian M. Burnett ** Indiana University Elizabeth A. Gordon Temple University Bjorn N. Jorgensen University of Colorado at Boulder Cheryl L. Linthicum University of Texas at San Antonio * We thank Vic Anand, Alan Jagolinzer, Karen Jones, Grace Pownall, Shiva Rajgopal, Gregory Waymire, Eunyoung Whang (IAS discussant), and seminar participants at Baylor, Drexel, Emory, and the 2013 International Accounting Section Mid-year Meeting for helpful comments and suggestions. We thank Larry Ochoa and Claire Veal for helpful research assistance. ** Orfalea College of Business, Cal Poly, San Luis Obispo, 1 Grand Avenue, San Luis Obispo, CA Ph: (805) burnettb@calpoly.edu.

2 Early Evidence from Canadian Firms Choice Between IFRS and U.S. GAAP Abstract For fiscal years starting on or after January 1, 2011, Canada abandoned Canadian Generally Accepted Accounting Principles (GAAP) and adopted International Financial Reporting Standards (IFRS) as the dominant accounting standard, yet de facto allowed firms to adopt U.S. GAAP in lieu of IFRS. We document the unintended consequence that more Canadian firms report under U.S. GAAP after Canada adopted IFRS. We find firms more likely to choose IFRS are larger, in the developmental stage, have more R&D expenditures, fewer U.S. operations or fewer U.S. shareholders. Further, the likelihood of IFRS adoption decreases if stockholders' equity under U.S. GAAP exceeds Canadian GAAP in the year before IFRS adoption. Second, consistent with an increase in information asymmetry around Canada's IFRS adoption, bid-ask spreads increase for smaller Canadian firms not listed in the U.S. These results inform U.S. regulators' current decision whether to allow U.S. firms the choice between U.S. GAAP and IFRS. JEL Code: M41 Keywords: IFRS, U.S. GAAP, Liquidity, Accounting Choice, Event Study. 2

3 And it is actually true, in some respects, that the Canadian environment probably most resembles the situation that the U.S. is going to be in if it makes a similar kind of decision. -Tricia O Malley, Former AcSB and IASB member, speaking of the Canadian decision to adopt IFRS 1. Introduction In the quote from the U. S. Securities and Exchange Commission (SEC) July 2011 Roundtable on International Financial Reporting Standards (IFRS), Ms. O Malley is referring to Canada s decision to adopt IFRS and the potential implications of its experience as the U.S. considers IFRS. As implied, the environments in Canada and the U.S. share many similarities such as high-quality financial reporting and accounting standards, a market-orientation, strong enforcement, and legal systems. In our paper, we exploit the Canadian setting where standard setters adopted IFRS as domestic GAAP but gave companies a de facto choice to use U.S. Generally Accepted Accounting Principles (U.S. GAAP). 1 We explore the determinants and consequences of the choice. We find an unintended consequence that more Canadian companies report using U.S. GAAP after the adoption of IFRS than before, which supports the notion that companies will make the choice based on factors they perceive as important. As such, Canada s adoption of IFRS offers a unique setting to examine the debate about firms having a choice among accounting standards, especially as the U.S. considers whether and how to adopt IFRS. With the emergence of IFRS and U.S. GAAP as the predominant world accounting standards, academics and standard setters are contemplating the advantages and disadvantages of full convergence or the coexistence of two accounting standards. Despite the large-scale adoption of IFRS by over 100 countries in recent years, the vast majority of IFRS-adopting firms never reported under U.S. GAAP since they previously reported only under the GAAP of their 1 Technically, only firms registered with the SEC were permitted to use U.S. GAAP instead of IFRS by Canadian securities regulators. However, we note that firms had a de facto choice because firms could, though at some cost, register with the SEC and report with U.S. GAAP. In fact, Nimin Energy Corporation and some other Canadian firms publicly stated that they initiated cross-listing in the U.S. to be permitted to adopt U.S. GAAP.

4 home country. Further, in other IFRS adopting countries, companies were not given the choice to adopt either IFRS or U.S. GAAP as they were in Canada. To our knowledge, previous academic research considers only two experimental settings that permit empirical examination of the effects of IFRS adoption relative to U.S. GAAP: (1) Germany s Neue Markt exchange that allowed the choice of U.S. GAAP or International Accounting Standards (IAS), the predecessor to IFRS, and (2) U.S. cross-listed firms that were required to reconcile differences between their IFRS and U.S. GAAP. Germany s Neue Markt exchange allows for an examination of the choice of accounting standards in a market where domestic GAAP was considered low quality. Leuz (2003) finds insignificant differences in measures of information asymmetry and market liquidity between Neue Markt firms choosing IAS or U.S. GAAP. 2 At that time, however, enforcement of standards was limited in Germany. Germany s enforcement body did not implement a proactive review of financial statements until 2005 (Christensen et al., 2011), which makes it difficult to infer whether the insignificant differences in information asymmetry between IAS and U.S. GAAP documented in Leuz (2003) are due to the similarity of the standards or lack of enforcement of the standards. The IFRS to U.S. GAAP reconciliations of foreign private issuers (FPIs) provide a direct comparison of reporting under IFRS and U.S. GAAP. 3 Gordon et al. (2011) compare the earnings attributes of IFRS and U.S. GAAP holding fixed the underlying cash transactions of the firm. They conclude that while some earnings attributes do not differ due to accounting standards, other differences due to financial reporting incentives persist even after the adoption of IFRS. 4 2 Hwang and Lin (2010) note that some firms that voluntarily chose to report under U.S. GAAP have voluntarily switched to IFRS, which replaced German GAAP when the EU adopted IFRS. For example, Euromicron Group, a German manufacturer of fiber optic components, abandoned U.S. GAAP for IFRS in During the period from 2004 through 2007, these foreign private issuers were required to reconcile from IFRS to U.S. GAAP. 4 Further current research studies the effect of discontinuing the reconciliation requirements for foreign private issuers. See, among others, Jiang et al. (2010) and Kim et al. (2012). 2

5 While informative, the generalizability of the above findings to the Canada and the U.S. may be limited. First, due to self-selection bias arising because firms chose to voluntarily report under U.S. GAAP when they listed on Neue Markt and in the U.S., respectively. Furthermore, if financial reporting incentives are an integral determinant of earnings quality then it is difficult to draw inferences from samples of foreign, predominantly European, firms that face different reporting environments than North American firms. For example, most countries in the European Union have civil law legal origin, the exceptions being two common law countries: Ireland and the United Kingdom. Ball et al. (2000) document that legal origin is a codeterminant of earnings attributes. Since both Canada and the U.S. have common law, the applicability to North American firms of inferences regarding IFRS adoption from a sample of firms located in the European Union s predominantly civil law countries with lower intensity of enforcement may be limited. To mitigate these concerns, we study a recent sample of IFRS adopting firms based in Canada which is a predominantly a common law country. Since Canadian and U.S. regulators have historically designated each other a special consideration due to their proximity in accounting standards and levels of enforcement procedures, any inferences based on a sample of Canadian firms are more likely to carry over to U.S. firms than prior studies with lower enforcement prior to IFRS adoption. With the more recent adoption of IFRS, companies are also using updated IFRS standards that have benefited from the IASB-FASB convergence project. In summary, the Canadian setting has three distinct advantages: (1) Canada permitted the choice between IFRS and U.S. GAAP, (2) the Canadian and U.S. environments are more similar than other countries might be, and (3) Canada recently adopted IFRS. 3

6 We examine two main aspects of IFRS adoption in Canada. First, we describe and investigate initial reporting choices Canadian firms make. 5 While Canada adopted IFRS on or after January 1, 2011, standard setters gave companies the de facto option to use IFRS or U.S. GAAP. Those Canadian firms that had cross-listed in the U.S. and maintained their FPI status did not need special permission from regulators to use U.S. GAAP. Other Canadian firms needed regulatory approval to use U.S. GAAP. Interestingly, we find more firms report under U.S. GAAP after Canada adopted IFRS. Before Canada s IFRS adoption, 48 out of 245 crosslisted companies reported under U.S. GAAP, or about 20%. After, 45 of these continued to report under U.S. GAAP suggesting they were minimizing the costs of switching standards. Of the 197 cross-listed companies that previously reported under Canadian GAAP, we document that at least 29 or about 15%, voluntarily choose U.S. GAAP over IFRS, implying that companies differ in the perceived costs and benefits of their choice of accounting standards. 6 We investigate the determinants of the choice of IFRS versus U.S. GAAP, finding cross-listed companies choosing IFRS are larger, in the developmental stage, have more R&D expenditures, fewer U.S. operations or fewer U.S. shareholders. Further, consistent with the impact of the accounting standards on reported results playing an important role on firms standard choice, we find the likelihood of IFRS adoption decreases if stockholders' equity under U.S. GAAP exceeds Canadian GAAP in the year before IFRS adoption. 5 Competition is likely to be ongoing as Canadian firms assess whether IFRS or U.S. GAAP best meets their reporting needs. For example, Encana initially switched to IFRS from Canadian GAAP, but then announced that for fiscal 2012 it was switching to U.S. GAAP, stating Adopting U.S. GAAP will make it easier for investors to compare Encana s financial performance with its peer companies, most of which are based in the United States. Consistent with current practice, Encana will report its 2011 year-end financial results in February 2012 in accordance with International Financial Reporting Standards (IFRS). Starting in April 2012, Encana will report its first quarter results using U.S. GAAP. 6 In addition to these 29 firms, five or fewer Canadian firms may have been mandated to switch to U.S. GAAP if the U.S. SEC determined that these firms no longer qualified for foreign private issuer status. 4

7 Second, we investigate whether market liquidity changes with the adoption of IFRS versus U.S. GAAP. Consistent with research examining IFRS adoption in other countries (e.g., Daske et al. 2008; Christensen et al. 2011), liquidity is a proxy for economic outcomes because theory predicts higher transparency reduces information asymmetries which in turn increases liquidity (e.g., Glosten and Milgrom 1985; Diamond and Verrecchia 1991; Verrecchia 2001). We examine three common measures of liquidity: share turnover, zero returns days, and bid-ask spread. If IFRS is more (less) informative to the capital markets than U.S. GAAP or Canadian GAAP thereby reducing (increasing) information asymmetry, then we expect that liquidity will increase (decrease) after the adoption of IFRS. If the accounting standards are similarly informative, we expect no difference. We use two samples to examine the liquidity implications: (1) U.S. cross-listed firms and (2) non-u.s. cross-listed firms. In the U.S. cross-listed sample, we find no differences in liquidity before and after the adoption of IFRS, implying those companies interacting with the U.S. market do not face negative consequences when adopting IFRS. In the non-u.s. cross-listed sample, we find mixed results using the three different liquidity measures. Results of bid-ask spreads suggest lower liquidity but share turnover and zero returns data suggest no difference. Further investigation shows that the lower bid-ask spreads are concentrated in smaller Canadian firms. Our paper contributes to the debate about permitting two accounting standards to coexist. When given a choice of two high quality accounting standards, we find that companies will base their decision on their perceived costs and benefits of the accounting standard. The unintended consequence of Canada adopting IFRS that more Canadian companies report using U.S. GAAP supports the notion that companies will make the choice based on factors they perceive as important. Specifically, we find that Canadian FPIs are more likely to choose U.S. GAAP if they 5

8 have more operations in the U.S., fewer operation in IFRS reporting countries, less influence in their local Canadian market (in terms of size), and have less complex accounting. We find no evidence that cross-listed companies are penalized or rewarded by the market for having the choice of standards. Similarly, Canadian firms switching from Canadian GAAP to IFRS are also not adversely affected, except for smaller firms. Regulators, standards setters, and listed companies should be interested in our findings. As U.S. regulators consider allowing or requiring IFRS, the Canadian experience suggests that when moving from one set of high quality accounting standards to another in a high enforcement regulatory regime and given a de facto choice, companies gain as they tradeoff their firmspecific costs and benefits. Our results also suggest that the move from one high quality accounting standard to one (or two) other high quality accounting standard(s) has limited capital market effects. Standard setters should be interested in the financial statement users assessment of changing from one standard to another. Finally, if U.S. companies are given the choice, U.S. managers should also be interested in how the capital markets perceive the change. The paper proceeds as follows. Section 2 summarizes the arguments for and against full convergence or permitting two accounting standards to exist side by side in a market. Section 3 offers a brief literature review. Section 4 investigates Canadian firms choice of accounting standards. Section 5 analyzes the liquidity effects of IFRS adoption in Canada. Section 6 concludes and suggests avenues for future research. The Appendix further details the Canadian path to IFRS and its market consequences. 6

9 2. Competition among accounting standards Academic literature discusses the costs and benefits of full convergence or continued competition between IFRS and U.S. GAAP accounting standards. Dye and Sunder (2001) detail arguments for and against competition among accounting standard setters, and for and against allowing U.S. corporations to issue financials in accordance with standards promulgated by either FASB or International Accounting Standards Board (IASB). A core argument in favor of competition among standards is that competition increases the long run efficiency of accounting standards. The standard-setting organizations are likely to be more responsive to constituents demand for better standards than with a monopolist standard setter. A countervailing concern, however, is that competition may lead to a race to the bottom. Although motivated by this debate, this paper cannot speak directly to long run effects (if any) of absence of convergence but merely to firms choice of accounting standards and the resulting unintended consequences. While arguments have been made both for and against competition, the SEC has followed a strategy of converging U.S. GAAP with IFRS and is contemplating full adoption of IFRS. In September 2002, the IASB and FASB signed a Memorandum of Understanding, The Norwalk Agreement, pledging to make their existing financial reporting standards fully compatible as soon as practicable. Convergence, though, has also been criticized for a lack of competition. Kothari et al. (2010) assert, that, rather than converging U.S. GAAP with IFRS, competition between the FASB and the IASB would allow GAAP to better respond to market forces absent regulation of the choice of accounting standards. As academics debate the merits of single versus multi-gaap reporting environments, the SEC is considering whether to permit the more than 10,000 U.S. domiciled filers to (a) continue the currently required reporting under U.S. GAAP, (b) require the adoption of IFRS, or (c) allow 7

10 U.S. companies to choose between either IFRS or U.S. GAAP. The original roadmap for the consideration of mandatory adoption of IFRS by U.S. issuers was published in 2005, and deliberations continue. To illustrate, consider the following quote from SEC Chairman Christopher Cox, March 6, 2007 (SEC 2007): That original commitment [to the Roadmap] meant that IFRS and U.S. GAAP would someday compete freely in America's capital markets, and that two accounting systems would operate side by side at least until the process of convergence concludes with actual convergence, and there is truly one global accounting standard and seamless international comparability of reporting. It meant that issuers, markets, and investors would have a choice because they, not the government, will decide between IFRS and GAAP. More recently, the SEC hosted a Roundtable in July 2011 to discuss IFRS, in which Richard Larsen of Duff and Phelps, LLC commented, I think that, in many ways, the last several years, or post-norwalk Agreement, that the healthy tension between the FASB and the IASB has created better standards (SEC 2011). Canada s recent experience with IFRS adoption provides potential insights about what might happen if the U.S. fully adopts IFRS or permits firms to choose between the two standards. As Tricia O Malley, a former Canadian Accounting Standards Board (AcSB) member and IASB member, stated, Canada is the canary in the coal mine on behalf of this whole process [in the US] (SEC 2011). Canada permitted competition among accounting standards for Canadian firms registered with the SEC. These firms were allowed to choose between the IFRS and U.S. GAAP. Such competition is consistent with a long-standing tradition for competition between provinces in Canada concerning its legal standard setting (Daniels, 1991). In summary, Canada s IFRS adoption experience provides unique insights about the unsettled debate over whether the SEC should permit U.S. firms to choose between IFRS and U.S. GAAP. 8

11 3. Related research on IFRS The large scale adoption of IFRS has prompted accounting researchers and regulators to consider the preferred attributes of accounting standards. 7 Some argue that adoption per se is more a label than an actual change in financial reporting quality. While the prevailing accounting standards (prior GAAP or IFRS) might be of importance, the surrounding institutions and enforcement mechanisms that help shape managers financial reporting incentives could be equally or even more important. Given the flexibility within every accounting standard, the financial reporting incentives could a priori play as an important role as the rules and standard themselves. The use of IFRS by over 100 countries has led to a number of insightful academic studies, mainly focused in the European Union (EU). For example, Armstrong et al. (2010) study the European market reaction to IFRS adoption announcements. Daske et al. (2008) investigate the firm characteristics of early IFRS adopters in the EU. Henry et al. (2009) and Gordon et al. (2011) consider the effect of FPIs adoption of IFRS. For example, Henry et al. study the reconciliations between IFRS and U.S. GAAP for 75 US-listed European firms. They find that net income reported using IFRS is statistically significantly higher than net income reported using U.S. GAAP. Further, the reconciliation from IFRS to U.S. GAAP was value relevant. However, the U.S. SEC removed the reconciliation requirement for IFRS filers for reporting periods beginning in Therefore, while some non- U.S. firms can report using either IFRS or U.S. GAAP, U.S. firms have no choice but to provide the SEC with financial statements prepared using U.S. GAAP. Hail et al. (2010a,b) provide a conceptual discussion of the economic arguments for and against the adoption of IFRS, and detail political factors influencing the consideration of IFRS adoption in the United States. The authors conclude with a discussion of future scenarios for 7 For recent survey papers see, Hail et al. (2010a,b) and Kothari et al. (2010). 9

12 U.S. accounting standards, including the elimination of U.S. GAAP, or potential co-existence of U.S. standards and IFRS in U.S. capital markets. Canadian IFRS adoption The Canadian Accounting Standards Board (AcSB) sets accounting standards for Canadian entities outside of the public sector. The Board actively considered harmonization and convergence of Canadian GAAP with IFRS and U.S. GAAP. Prior to 2004, two of the AcSB s main objectives were to eliminate or minimize differences first with U.S. GAAP and second with International Accounting Standards (Discussion Paper of Accounting Standards in Canada: Future Directions June 24, 2004). To work to eliminate the differences with U.S. GAAP, each year the AcSB performed a detailed review of differences between Canadian GAAP and U.S. GAAP for a random sample of Canadian firms that reported reconciliations from Canadian GAAP to U.S. GAAP. The AcSB then developed standards that eliminated or minimized these differences. After considering their constituent s input, on February 10, 2005, the AcSB proposed adopting IFRS in full to its oversight body, the Accounting Standards Oversight Council, while allowing entities that wanted to use U.S. GAAP to do so. 8 On March 31, 2005, the AcSB sought comment on its proposal to adopt IFRS (Leuz and Wysocki 2006). On April 15, 2005, the U.S. SEC introduced a possible road map to eliminate the reconciliation requirement for IFRS. On January 10, 2006, the AcSB ratified its plan to adopt IFRS over a five-year transition period, while allowing SEC registrants to continue reporting with U.S. GAAP. On November 15, 2007, the U.S. SEC voted to eliminate the reconciliation requirement for FPIs reporting under IFRS. An important final step in Canada s IFRS adoption was the AcSB s confirmation of the IFRS 8 For a more thorough discussion of why Canada adopted IFRS, please see the Appendix and the AcSB s 2011 report, Adoption of International Financial Reporting Standards: Background and Basis for Conclusions. 10

13 changeover with the announcement that Canadian publicly accountable enterprises would be required to adopt IFRS for fiscal years beginning on or after January 1, Canadian GAAP and U.S. listed companies In adopting IFRS, the AcSB required all publicly accountable enterprises to apply IFRS after firm years beginning on or after January 1, The provincial securities regulators, who have authority over the application of accounting standards, gave Canadian companies crosslisted in the U.S. the option to choose IFRS or U.S. GAAP. The provincial securities regulators also permitted firms to petition for special permission to use U.S. GAAP without listing in the United States. The provincial securities regulators required companies to begin reporting under IFRS or U.S. GAAP for the first quarter of 2011 (e.g., Ontario Securities Commission 2011). Prior to Canada s adoption of IFRS, Canadian and U.S. regulators determined that Canadian GAAP and U.S. GAAP were allowable alternatives for cross-listed companies under the Multi-jurisdictional Disclosure System (MJDS). 9 Canadian regulators accepted U.S. GAAP for domestic reporting. U.S. regulators accepted Canadian GAAP for FPIs, without reconciliation to U.S. GAAP. 10 Canada was the first country for which the SEC accepted domestic GAAP reporting for FPIs. In November 2007, the U.S. Securities and Exchange Commission exempted all non-us-based firms that report under IFRS from reconciliation. Therefore, as Canadian firms switch to IFRS, they will maintain the exemption from reconciliation requirements. 9 Canadian regulators also permitted these Canadian firms to report under U.S. GAAP. 10 This exemption from reconciliation based on the similarity in the standards is consistent with research by Webster and Thornton (2005) finding no overall difference in accrual quality between Canadian firms reporting under Canadian GAAP and U.S. firms reporting under U.S. GAAP. 11

14 The distinction between the SEC s requirements for FPIs and registrants is important in our sample selection and research design. 11 FPIs are not required to report under U.S. GAAP. They could use their domestic reporting standards and reconcile to U.S. GAAP (where required). For those FPIs reporting under IFRS, the SEC eliminated the reconciliation in November, Some Canadian firms listed in the U.S. do not qualify to be FPIs, and must then follow the same higher reporting requirements as U.S. companies, including using U.S. GAAP. Notwithstanding the previous arguments, minor differences in accounting rules between U.S. GAAP and Canadian GAAP did persist prior to and during our sample period. Bandyopadhyay et al. (1994) investigate a sample of firms that were listed both on Toronto and on a U.S. stock exchange between 1983 and Overall, they find that earnings scaled by market capitalization are 2% lower under U.S. GAAP than Canadian GAAP. Some of their main source of differences in accounting rules pertain to foreign exchange gains or losses on foreign long-term debt, early extinguishment of debt, extraordinary items, and interest capitalization of self-constructed assets [see Table 1 on page 265]. 11 The U.S. SEC defines a FPI as any foreign issuer that does not meet either of the following two conditions: (i) More than 50 percent of the outstanding voting securities of such issuers are directly or indirectly owned of record by residents of the U.S.; and (ii) any of the following: (A) The majority of the executive officers or directors are U.S. citizens or residents; (B) More than 50 percent of the assets of the issuer are located in the U.S.; or (C) The business of the issuer is administered principally in the U.S.. 12 Their sample is non-random since firms choose to cross-list in the U.S., due to higher anticipated need for bonding with U.S. institutions, capital or visibility. 12

15 4. Firms Accounting Standard Choice Data Our sample consists of Canadian firms in Compustat that were required to choose IFRS or U.S. GAAP by May 31, The AcSB required Canadian companies with fiscal years beginning on or after January 1, 2011 to adopt IFRS and begin reporting under IFRS in the first quarter of Because of the MJDS, the provincial securities regulators allowed SEC registrants to choose between IFRS and U.S. GAAP, and similarly required them to begin reporting using one of these two accounting standards for the first quarter of 2011 (e.g., Ontario Securities Commission 2011). We obtain financial data and prices from Compustat. Table 1, Panel A, details our sample selection. Of the 1,445 Canadian firms in Compustat, 135 insurance and investment companies deferred their decision until 2013 and the remaining 1,310 firms choose either IFRS or U.S. GAAP. Specifically, 92 adopted U.S. GAAP and 1,218 adopted IFRS. Table 1, Panel B, presents the accounting standards firms adopt conditional on their USlisting status and previous accounting standard. We focus on three important categories of firms. The first category includes the 48 firms that were listed in the U.S. and reporting with U.S. GAAP prior to Many of these firms subsequently lost their FPI status and became required by the SEC to report using U.S. GAAP, which is why all but three firms continue to use U.S. GAAP. The second category represents firms listed in the US, but initially reporting under 13 We examined whether firms go private rather than adopt IFRS, and document that 23 firms went private between 2009 and This suggests that the costs of IFRS adoption did not lead to a significant number of going-private transactions. 14 Due to delayed International Accounting Standards Board (IASB) s projects involving investment companies, insurance contracts, and accounting for rate-regulated entities, the provincial securities regulators allowed investment companies, insurance companies, and rate-regulated entities to delay adoption of IFRS until fiscal years beginning on or after January 1, 2013 (2012 for rate-regulated entities). Our sample includes rate-regulated entities that made their adoption choice by May 31, On January 10, 2006, Canada formally announced it would abandon Canadian GAAP in favor of IFRS. If a firm s inception is in 2006 or later, we label the firm based on its accounting standard used in its first year. 13

16 Canadian GAAP. These firms are arguably the most interesting because they were permitted to choose between IFRS or U.S. GAAP. Of the 197 firms in this category, we note that five firms lost their FPI status and were required by the SEC to begin reporting with U.S. GAAP. The remaining 192 firms had a choice between IFRS and U.S. GAAP 29 (15%) chose U.S. GAAP, while 163 (85%) chose IFRS. The third category represents firms not listed in the US, which meant that they had to report under Canadian GAAP prior to For these firms to be able to adopt U.S. GAAP instead of IFRS, they would have (1) become an SEC registrant or (2) obtained special permission from Canadian securities regulators. Thirteen firms in this category did adopt U.S. GAAP eight became SEC registrants and five obtained special permission from securities regulators. 16 The vast majority of firms in this category, 99%, adopted IFRS. Table 1, Panel B, documents an unintended consequence of Canada s adoption of IFRS. Before Canada adopted IFRS, we observe that 48 out of 245 cross-listed companies previously reported under U.S. GAAP, or about 20%. After Canada adopted IFRS, the number of firms reporting under U.S. GAAP nearly doubled to 92. Nevertheless, a majority of firms choose IFRS. Determinants of firms accounting standard choice We examine the determinants of U.S.-listed Canadian firms choice between IFRS and U.S. GAAP. We focus on U.S.-listed Canadian firms because they were given a choice between IFRS and U.S. GAAP, without having to incur additional costs to obtain special permission to use U.S. GAAP or register with the SEC. We exclude those Canadian firms that originally reported under U.S. GAAP because the SEC required many of these to report under U.S. GAAP 16 The five firms that obtained special permission to adopt U.S. GAAP without becoming SEC registrants are rateregulated entities that were allowed to adopt U.S. GAAP until January 1, 2015 due to uncertainty under IFRS related to rate-regulated accounting. 14

17 as they did not qualify as FPIs. This results in a sample of U.S.-listed firms that originally reported under Canadian GAAP and then had a choice between IFRS and U.S. GAAP. We are not the first to examine firms choices between accounting standards. Leuz and Verrecchia (2000) study the determinants of German firms choices between German GAAP and International Reporting standards (U.S. GAAP and IAS). They find the choice of an International Reporting standard is positively associated with firms performance, measured as return on assets (ROA), and financing needs, measured as capital intensity and a listing in the U.S. or UK. Leuz (2003) studies firms trading in Germany s New Market during the years 1999 and 2000 where German firms chose between U.S. GAAP and IAS, the predecessor to IFRS. His primary motivation was to provide evidence about the quality of IAS relative to U.S. GAAP. Assuming U.S. GAAP is associated with higher quality corporate disclosure, he hypothesizes that the choice of U.S. GAAP is a function of firm size (+), financing needs (+), and firm performance (+/-). He finds that the choice of U.S. GAAP is significantly and positively associated with financing needs, which he notes, is consistent with a perception at that time that U.S. GAAP was preferable for firms with large future financing needs because it allowed them better access to the U.S. capital markets. He does not find the choice of U.S. GAAP is significantly associated with firm size or firm performance. We believe reexamining the choice between IFRS and U.S. GAAP is important. The SEC is contemplating whether to require U.S. public companies to (a) retain U.S. GAAP, (b) adopt IFRS, or (c) permit firms to choose between the two standards. As the most similar capital market to the US, Canada is the nearest setting, both geographically and in terms of accounting standards and financial reporting incentives, to best possibly understand what U.S. firms might choose if permitted to choose between IFRS and U.S. GAAP. Finally, the study of Canadian 15

18 firms listed in the U.S. enables more robust analysis than a study of accounting choice by non- U.S. listed companies. The financial reporting incentives of firms are similar to U.S. firms; they are all listed in the U.S. and provided reconciliations from Canadian GAAP to U.S. GAAP in their SEC filings prior to adopting either IFRS or U.S. GAAP. We model the choice between adopting IFRS or U.S. GAAP as a function of the firmspecific costs and benefits of adoption. Specifically, we use the following probit regression to examine this choice: Prob(IFRS=1) = F(β 0 + β 1 SE Comparability t-1 + β 2 R&D Intensity t-2 + β 3 IFRS vs. US Operations t-2 + β 4 US Ownership t-2 + β 5 Leverage t-2 + β 6 DSE t-2 + β 7 Size t-2 + β 8 ROA t-2 + ε t ) (1) We include industry fixed effects and use robust standard errors. We calculate the variables in our model two years prior to firms adoption of IFRS or U.S. GAAP because the decision to adopt IFRS requires a two-year transition period, and this is likely when most firms made the choice between the two standards. As firms weighed the costs and benefits of choosing between the accounting standards, they commonly disclosed three primary determinants informing their decisions: (1) the impact on reported results, (2) comparability with industry peers, and (3) the needs of key stakeholder groups (e.g., shareholders, lenders, etc.). 17 SE Comparability and R&D Intensity proxy for the impact on reported results. IFRS vs. US Operations focuses on comparability with peer firms. US Ownership considers stakeholders needs. Additionally, we include proxies for the relative costs of IFRS adoption versus US GAAP and control for firm size and performance. 17 As an example, Magna International Inc. stated that in making its decision between the two standards, the board of directors considered many factors, including, but not limited to (i) the changes in accounting policies that would be required and the resulting impact on our reported results and key performance indicators, (ii) the reporting standards expected to be used by many of our industry comparables, and (iii) the financial reporting needs of our market participants, including shareholders, lenders, rating agencies and market analysts. 16

19 The impact of U.S. GAAP relative to IFRS on reported results is likely a primary determinant of firms choice between the two standards. Specifically, firms are more likely to choose IFRS than U.S. GAAP when IFRS portrays the firm in a more favorable light than U.S. GAAP, and vice versa. Ideally, we could observe IFRS and U.S. GAAP reported numbers for the same firm in the same year prior to their choice of accounting standard. Unfortunately, we are unable to observe IFRS and U.S. GAAP reported numbers for firms that choose US GAAP. 18 We are, however, able to observe Canadian GAAP and U.S. GAAP reporting in the year prior to their choice of accounting standard for all but 10 firms because most firms provided reconciliations to U.S. GAAP. We posit that when Canadian GAAP is less comparable to U.S. GAAP in an unfavorable way, firms are more likely to choose U.S. GAAP. When Canadian GAAP portrays a firm in a more favorable light than U.S. GAAP, firms are more likely to choose IFRS. We examine the comparability stockholders equity under Canadian GAAP and U.S. GAAP as a cumulative summary measure of which standard portrays the net assets of a firm in a more positive light. We calculate the variable SE Comparability as Canadian GAAP stockholders equity less U.S. GAAP stockholders equity scaled by the absolute value of Canadian GAAP stockholders equity. 19 We expect that SE Comparability will be positively associated with choosing IFRS. We further consider the impact of the standard choice on reported results by focusing on a key difference between IFRS and U.S. GAAP. Under U.S. GAAP, R&D is generally expensed as incurred, with capitalization of software being a notable exception. In contrast, IFRS requires 18 For firms that choose IFRS, IFRS adoption requires retroactive disclosure of the prior year under IFRS making it possible to observe firms reporting under Canadian GAAP and IFRS. Further, all but ten of these firms provided a reconciliation to U.S. GAAP for the year prior to IFRS adoption. Thus, in the year prior to IFRS adoption, for firms that adopt IFRS, we can observe their reporting under Canadian GAAP, IFRS, and US GAAP. 19 Ten firms in this sample have negative stockholders equity. We obtain qualitatively similar results if we delete these firms from our analysis. 17

20 capitalization of certain R&D expenses. As a consequence, we predict that firms that reported high R&D expenses under Canadian GAAP (R&D Intensity) are more likely to switch to IFRS. Firms are more likely to choose IFRS (U.S. GAAP) when that standard provides for enhanced comparability with a firms peer group. 20 Canadian firms with significant operations in the U.S. (outside the U.S.) are likely to have peer firms that use U.S. GAAP (IFRS). We use IFRS vs. US Operations to measure whether more of a firm s foreign operations are in or outside the U.S. IFRS vs. US Operations is calculated as the proportion of firm s assets located outside of Canada and the U.S. less the proportion of firm s assets located in the U.S. 21 We expect that IFRS vs. US Operations is positively associated with choosing IFRS, consistent with a higher likelihood of a firm s peers using IFRS. We consider the role of stakeholders needs in the between the two standards by focusing on shareholders. Shareholders are likely the key constituent affecting this decision. Accordingly, we proxy for this using US Ownership which is the percentage of common shares held by U.S. institutional investors. We expect that US Ownership will be negatively related to choosing IFRS. Since changing accounting standards is costly, we consider two types of firm-specific costs associated with the adoption of IFRS versus U.S. GAAP. In either case, adopting a new accounting standard likely entails renegotiation of contracts, where the contractual terms are based on accounting numbers. For Canadian firms, renegotiation costs are likely lower when adopting U.S. GAAP than IFRS because each of these firms was already providing a 20 For example, Canadian Pacific Railway Limited stated, CP commenced reporting its financial results using U.S. GAAP, which is consistent with the current reporting of all other North American Class I railways. 21 Alternatively, if we include two variables, one for U.S. operations and one for operations outside the US, we obtain qualitatively similar results for these variables and the other variables in our model. 18

21 reconciliation from Canadian GAAP to U.S. GAAP in their U.S. filings. 22 Leverage, long-term debt divided by total assets, captures the costs involved in renegotiating debt covenants. IFRS adoption may move firms closer or farther from their debt covenants depending on the effect of IFRS adoption. For example, IFRS adoption permits firms to revalue their property, plant and equipment to fair market value. If a revaluation results in a higher property, plant, and equipment value then IFRS adoption may reduce a firm s leverage ratio and therefore reduce debt-covenant renegotiation costs. Second, we consider that many of the sample firms are considered Development Stage Enterprises (DSEs) 23, which are focused on establishing a new business where either their primary operations have not yet begun or no significant revenues have been earned. Willenborg (1999) argues that the financial statements of a typical DSE contain little meaningful accounting information and do not have more audit-intensive accounts such as revenues, inventory and accounts receivable. Therefore, the costs to adopt IFRS for DSEs are relatively lower than for non-dses due to lower accounting complexity. DSE is equal to one if a firm is a DSE, and zero otherwise. Finally, we control for Size, the log of market value, which captures size-related differences, such as the information environments. ROA, net income divided by total assets, is our measure of performance. Since IFRS and U.S. GAAP are largely viewed as of similar quality in European Union countries with low enforcement (e.g., Leuz 2003; Barth et al. 2008; 22 Bandyopadhyay et al. (1994) find that earnings scaled by market capitalization are 2% lower under U.S. GAAP than Canadian GAAP. Since the time of Bandyopadhyay et al. (1994) s study, U.S. GAAP and Canadian GAAP have become even more similar. From 1995 to 2004, the AcSB focused on harmonizing Canadian GAAP with U.S. GAAP by adopting standards that reduced differences between the two accounting standards (Discussion Paper of Accounting Standards in Canada: Future Directions June 24, 2004). 23 ASC , IFRS 3. 19

22 Gordon et al. 2011), a priori it is unclear whether and how Size or ROA will relate to firms choices between IFRS and U.S. GAAP in Canada and U.S. with higher level of enforcement. Empirical results for determinants of firms accounting standard choice Table 2 presents descriptive statistics for the 170 US-listed firms with a choice between IFRS and U.S. GAAP and with necessary financial information. 24 The differences in SE Comparability and R&D Intensity are consistent with differences in reported results playing an important role in choosing between IFRS and U.S. GAAP. The mean and median of SE Comparability are statistically significantly larger for firms that choose IFRS based on a t-test of difference in means and the non-parametric Wilcoxon signed-rank test, respectively. The mean of for firms that choose U.S. GAAP indicates that on average stockholders equity of firms that choose U.S. GAAP is larger under U.S. GAAP than Canadian GAAP. The positive mean and median indicate that the opposite is true for firms that adopt IFRS. This is consistent with firms choosing the standard that portray themselves in the most favorable light. The nonparametric Wilcoxon signed-rank test indicates R&D Intensity is higher for firms that choose IFRS. This is consistent with firms with these development expenses preferring to capitalize them. The low average and median for R&D Intensity is attributable to the low percentage of firms that have R&D expenses (18 percent). The differences in IFRS vs. US Operations suggest that comparability with peer firms plays an important role in firms choices between the two standards. The mean (median) of (-0.15) for IFRS vs. US Operations for firms that choose US GAAP indicates on average they have more operations in the U.S., while the mean of 0.16 of IFRS vs. US Operations for 24 The number of observations in the analysis is 170, not the 197 firms listed in the U.S. and previously reporting with Canadian GAAP reported in Table 1, Panel B, because we exclude 27 firms. Specifically: five firms lost foreign private issuer status and were required to adopt U.S. GAAP by the SEC, nine firms did not separately disclose U.S. assets, three rate-regulated entities since their rate-regulated status perfectly explains their choice of U.S. GAAP, and 10 firms did not disclose reconciliation to U.S. GAAP. 20

23 firms that choose IFRS indicates that firms that choose IFRS have more operations in IFRSbased countries. Neither the mean nor the median of US Ownership are statistically significantly different, contrary to our expectation. However, this is likely due to measurement error in our proxy since institutional investors tend to invest in large companies. Multivariate testing that controls for size is necessary to control for this weakness in our proxy. The t-test of difference in means and the non-parametric Wilcoxon signed-rank test indicate that more DSE firms choose IFRS than U.S. GAAP, consistent with lower accounting complexity reducing the cost of IFRS adoption. The descriptive statistics indicate that firms choosing U.S. GAAP are not statistically significantly different from firms that choose IFRS in terms of Leverage, Size, or ROA. Table 3 reports the pairwise correlations. The Pearson and Spearman correlations indicate that IFRS is positively and statistically significantly correlated with SE Comparability and IFRS vs. US Operations, suggesting that the impact of reported results and comparability with peer firms are important determinants of firms choice between the two standards. The high and statistically significant Pearson (Spearman) correlation of 0.50 (0.57) between US Ownership and Size highlights the measurement error in US Ownership caused by institutional investors preference for investing in large stocks and confirms the need to control for size when using our proxy. Table 4 reports coefficients, z-statistics, and marginal effects for the probit regression analysis of firms standard choice. We calculate the marginal effect of each independent variable as π(x) = Φ(x'β), where Φ is the cumulative distribution function of the standard normal distribution and x and β are the vector of independent variables and corresponding coefficient 21

24 estimates from equation 1. We use the values zero and one for indicator variables and the first and third quartiles for continuous X variables, with the remaining X variables set equal to their mean values. We then compute the difference in π(x) at these two values of each X variable. Our proxies for the impact of the standard choice on reported results indicate that this is a significant determinant in firms decisions. SE Comparability is positive and statistically significant at the five percent level. The marginal effect indicates that as Canadian GAAP portrays the net assets of the firm in a more favorable light than US GAAP, firms are 2.1 percent more likely to adopt IFRS when moving from the first quartile to the third quartile SE Comparability. In contrast, when U.S. GAAP portrays the net assets of the firm in a more favorable light than Canadian GAAP, firms are more likely to adopt U.S. GAAP. 25 R&D Intensity is positive and significant, suggesting that firms with higher research and development activities prefer the capitalization approach under IFRS to U.S. GAAP. The marginal effect is low at 0.0 percent; but, as mentioned previously, only 18 percent of firms have R&D expense, which means both the first and third quartile of R&D Intensity is zero. The results for IFRS vs. US Operations are consistent with comparability being an important determinant in firms decisions between the two standards. The positive and significant coefficient on IFRS vs. US Operations indicates that firms with more operations in IFRS countries than in the U.S. are more likely to choose IFRS; firms with more operations in the U.S. than outside the U.S. are more likely to choose U.S. GAAP. The marginal effect 25 We acknowledge that just because Canadian GAAP reports higher net assets than US GAAP does not mean that IFRS will result in higher reported net assets than US GAAP. For the firms that adopt IFRS, we are able to provide evidence consistent with our hypothesis. As noted in footnote 17, for all but 10 firms that chose IFRS we have the reported stockholders equity under IFRS and U.S. GAAP for the year prior to IFRS adoption. Sixty-two percent of the time IFRS is higher than U.S. GAAP and the mean and median stockholders equity are statistically significantly higher under IFRS than U.S. GAAP. 22

25 indicates that firms at the third quartile of IFRS vs. US Operations are 6.2 percent more likely to adopt IFRS than firms at the first quartile of IFRS vs. US Operations. The results indicate that investor needs are a determinant in firms decisions in the choice between IFRS and U.S. GAAP. Firms with higher ownership by U.S. investors are more likely to choose U.S. GAAP than IFRS as evidenced by the negative and significant coefficient on US Ownership. Firms at the third quartile of US Ownership are 4.0 percent less likely to adopt IFRS than firms at the first quartile. The results for Leverage and DSE provide some evidence about the role of costs in firms decisions between IFRS and U.S. GAAP. The coefficient on Leverage is not statistically significant. However, the positive and significant coefficient on DSE is positive and significant is consistent with lower IFRS adoption costs for DSEs relative to non-dses due to less complex accounting. DSEs are 6.4 percent more likely to adopt IFRS than non-dses. Size is significantly, positively correlated with the likelihood that firms adopt IFRS. The marginal effect is economically significant at 12.4 percent. This is consistent with a fixed cost component to IFRS adoption that makes adoption relatively less costly for larger firms. Performance as measured by ROA is not statistically significantly associated with firms standard choice. 5. Liquidity effects of IFRS adoption in Canada We extend our analysis of the capital-market effects of IFRS adoption in Canada by examining whether its adoption resulted in changes in the liquidity of Canada s capital markets. Consistent with research examining IFRS adoption in other countries (Daske et al. (2008); Christensen et al. 2011), we focus on liquidity as a proxy for economic outcomes because theory 23