The Reputational Costs of Tax Avoidance and the Under-Sheltering Puzzle

Transcription

1 The Reputational Costs of Tax Avoidance and the Under-Sheltering Puzzle John Gallemore University of North Carolina Edward L. Maydew University of North Carolina Jacob R. Thornock University of Washington January 2012 Corresponding author. We thank Darren Bernard, Nicole Cade, Michelle Hanlon, Bradford Hepfer, Kevin Markle, Zoe-Vonna Palmrose, Phil Quinn, Steven Savoy, Terry Shevlin, Michelle Shimek, Brady Williams, and Ryan Wilson for insightful comments. We are grateful to Michelle Hanlon and Joel Slemrod for sharing tax shelter data and Bob Bowen, Andy Call, and Shiva Rajgopal for sharing Fortune Magazine reputation data.

2 The Reputational Costs of Tax Avoidance and the Under-Sheltering Puzzle ABSTRACT: We investigate whether firms and their top executives bear reputational costs from engaging in aggressive tax avoidance activities. Prior literature has posited that reputational costs partially explain why so many firms apparently forgo the benefits of tax avoidance, the so-called under-sheltering puzzle. We employ a database of 113 firms that were subject to public scrutiny for having engaged in tax shelters, representing the largest sample of publicly identified corporate tax shelters analyzed to date. We examine the reputational costs that prior research has shown that firms and managers face in cases of alleged misconduct: increased CEO and CFO turnover, auditor turnover, lost sales, increased advertising costs and decreased media reputation. Across a battery of tests, we find no consistent evidence that firms or their top executives bear significant reputational costs as a result of being accused of engaging in tax shelter activities. Moreover, we find no decrease in firms tax avoidance activities after being accused of tax shelter activity, and no evidence that high reputation firms avoid engaging in tax shelters in the first place. We conclude that the under-sheltering puzzle does not appear to be explained by reputational costs, even for firms accused of being at the aggressive end of the tax avoidance spectrum. Keywords: reputation; tax shelters; tax avoidance; under-sheltering. 1

3 I. INTRODUCTION This study investigates whether firms and their top executives bear reputational costs from engaging in aggressive tax avoidance activities. At least two decades of empirical tax research has shown that firms engage in a wide range of strategies for tax avoidance purposes. 1 Recent studies suggest that for many firms, tax avoidance appears to be highly effective at reducing the firms tax payments and increasing their after-tax earnings. Dyreng, Hanlon and Maydew (2008) find that more than twenty-five percent of the publicly traded U.S. firms in their sample are able to reduce their taxes to less than 20 percent of their pretax earnings, and are able to sustain such low rates of taxation over periods as long as ten years. Tax avoidance strategies are abundant and include a wide variety of activities such as shifting income into tax havens (Dyreng and Lindsey 2009), using complex hybrid securities (Engel, Erickson and Maydew 1999), and engaging in other tax shelters (Wilson 2009). While the evidence indicates there is wide variation in tax avoidance across firms, the extant literature has a difficult time explaining this variation. What is puzzling is not that some firms engage in tax avoidance, but rather why some firms engage in it enthusiastically while others appear to shun it. For example, while showing that some firms engage in substantial tax avoidance, Dyreng et al. (2008) also find that approximately one-fourth of firms pay taxes in excess of 35 percent of their pre-tax income over a ten year period. Given a U.S. federal corporate tax rate of 35 percent, these firms appear to be engaging in little or no sustainable tax avoidance. The question of why so many firms do not avail themselves of tax avoidance opportunities has been coined the under-sheltering puzzle (Desai and Dharmapala 2006; Hanlon and Heitzman 2010; and Weisbach 2002). 1 For reviews of the literature on tax avoidance, see Hanlon and Heitzman (2010), Maydew (2001), and Shackelford and Shevlin (2001). 2

4 Reputational costs are often posited as a factor that limits tax avoidance activities, particularly the most aggressive tax strategies. For example, the Commissioner of the Internal Revenue Service (IRS) asserts that aggressive tax strategies can pose a significant risk to corporate reputations and the general public has little tolerance for overly aggressive tax planning (Shulman, 2009). However, empirical evidence on the reputational costs of tax avoidance is scarce. The most compelling evidence to date is Hanlon and Slemrod (2009) and Graham, Hanlon and Shevlin (2011). Hanlon and Slemrod (2009) examine the stock price responses of firms accused of engaging in tax shelters. They find evidence of small but significant stock price declines associated with public revelation of tax shelter behavior. They are careful to acknowledge that there are many possible determinants of the negative returns, of which reputation costs are only one. With the exception of some tests on retail firms, they leave extensive testing of reputational costs for future research. Graham et al. (2011) survey tax executives and find that almost half agree that potential harm to their firm s reputation is a very important factor in deciding whether or not to implement a tax planning strategy. This evidence is consistent with managers perceiving some potential tax shelters as being too risky from a reputational standpoint, causing them to avoid using those shelters. Beyond this important initial evidence, there is much we do not know about reputational costs of tax avoidance. Among tax shelters that are actively used in the marketplace, we still have an imperfect understanding of why some firms implement them while others do not (i.e., the under-sheltering puzzle). In particular, we do not know if reputation plays a role in which firms implement a given tax shelter. Further, when firms are publicly scrutinized for having engaged in aggressive tax avoidance, do they actually bear reputational costs, as might have been feared ex ante? In their review of tax research, Hanlon and Heitzman (2010) call for research on 3

5 the under-sheltering puzzle, specifically posing the following questions: Why do some corporations avoid more tax than others? How do investors, creditors, and consumers perceive corporate tax avoidance? These are interesting questions worthy of study (p. 11, 20). This study answers the call for research, focusing on the extent to which reputational costs can explain the under-sheltering puzzle. If there are reputational costs of tax-avoidance, they are most likely to manifest in a) the firms with the most reputation to lose, and b) the most aggressive forms of tax avoidance. 2 Accordingly, we analyze a sample of firms identified in prior studies as engaging in aggressive tax shelters, most of which are large companies that are industry leaders and household names. Our study combines the samples of several prior studies of tax shelter behavior, namely, Graham and Tucker (2006), Hanlon and Slemrod (2009), and Wilson (2009). We then supplement the sample with firms that were publicly identified as having engaged in the corporate owned life insurance (COLI) shelter (Brown 2011). After imposing data requirements, our sample constitutes 113 firms revealed during the period 1995 to 2005 as having engaged in tax shelters. To our knowledge, this is the largest sample of publicly-identified corporate tax shelters analyzed to date. In this sample, representing some of the most extreme observed cases of tax avoidance, we find little evidence of reputational costs following public tax shelter scrutiny. We examine a wide set of potential reputational costs, such as CEO and CFO turnover and auditor turnover, that firms and/or their managers might face following the public revelation of using a tax shelter. We find no evidence of increased CEO, CFO, or auditor turnover in the three years following tax shelter revelation relative to propensity-matched control firms. 2 We adopt the terminology of Hanlon and Heitzman (2010) and use the term tax avoidance to refer to a continuum of tax planning strategies, where clearly legal tax reduction strategies (e.g., holding tax-exempt municipal bonds) are at one end and terms such as tax shelters and aggressive are used to describe strategies closer to the other end of the spectrum. 4

6 Next, we assess whether customers exert a reputational cost on shelter firms. Here also, the evidence does not suggest a reputational cost from tax sheltering. Neither sales, sales growth nor advertising expense has a differential change for revealed shelter firms relative to control firms. This finding suggests that customers do not seem concerned about tax sheltering. Moreover, in a subsample of retail firms, in which the potential reputational effect of tax shelter involvement could be highest, we find no evidence of reputational costs. We also examine how the revelation of a tax shelter influences the public reputation of a firm, as measured by the Fortune Magazine lists for Most Admired Companies and Best Companies to Work For, following Bowen, Call and Rajgopal (2010). We find no evidence that being caught in a tax shelter lowers the likelihood of making one or both of the Fortune Magazine lists, relative to the matched control sample. These results are consistent with tax shelter public scrutiny having little or no effect on firm reputation. We next consider the possibility that high reputation firms avoid tax shelters ex ante, which would help explain why we find little ex post reputational costs of tax sheltering. The univariate results actually show the opposite to be true firms with high reputations are more frequently users of tax shelters than are low reputation firms. In logistic regressions of tax shelter usage on proxies for reputation and other factors known to be associated with tax shelter usage (Wilson 2009), we find that reputation does not significantly influence of the likelihood of tax shelter participation. Overall, the results do not support the conclusion that high reputation firms ex ante avoid tax shelter usage. Firms also face reputational consequences with the tax authorities. When the IRS learns that a firm has engaged in activities it considers to be tax shelters, its policy allows it to expand the scope of information that it requests from the firm, which can lead to the discovery of other 5

7 aggressive tax avoidance. 3 Accordingly, we examine whether firms accused of engaging in tax shelters become more conservative in the future regarding their future tax avoidance, perhaps as a result of increased IRS scrutiny. If so, we would expect to observe the effective tax rate (ETR) of tax shelter firms to increase following scrutiny of their activities. The data, however, show that the ETR of tax shelter firms on average remains approximately the same before and after discovery, and this is true regardless of whether ETRs are measured on a GAAP or cash basis. These results suggest that firms identified as tax shelter users do not suffer significant reputational costs even at the hands of the tax authorities. In fact, univariate tests show a decline in tax shelter firm ETRs following the tax shelter revelation, relative to control firms, consistent with the reputational costs being so minor that some firms are emboldened to increase their tax avoidance. 4 Taken as a whole, the results suggest that the reputational costs of tax avoidance are unlikely to explain the under-sheltering puzzle. This finding holds despite focusing on the most extreme cases of tax avoidance and examining a wide set of potential reputational costs. If these extreme cases do not result in reputational costs, then it is unlikely that more mainstream tax avoidance results in significant reputational costs. By casting doubt on this often-conjectured explanation for the under-sheltering puzzle, it appears that under-sheltering is more of a puzzle than ever. The results are subject to some interrelated caveats. First, though we use tax shelters to focus on aggressive tax strategies, by definition our sample consists of tax strategies that were 3 These are sometimes referred to as listed transactions. One example of additional information that the IRS could request if it discovered the firm had engaged in listed transactions during the sample period was the documentation behind the firm s tax reserve, which represents the firm s self-assessment of the tax risk associated with its various activities. This was highly controversial and hotly contested. 4 In supplemental analyses we examine more possible reputation effects. We find no evidence that firms report higher litigation expense following the revelation of a tax shelter. Nor do we find evidence that a tax shelter revelation increases the level of short interest in a revealed shelter firm, relative to control firms. 6

8 actually implemented and discovered. It is entirely possible that firms avoid the most egregious tax strategies due to reputational concerns, and we do not observe those unused strategies. Indeed, such an effect would be consistent with survey evidence in Graham et al. (2011). Our study speaks to the reputational costs of tax shelters that are implemented by at least some firms, and whether reputational concerns affect how widely they are adopted (or not adopted). Second, firms and managers self-select into tax shelters, which means that it is possible that high reputation firms forgo the tax benefits of sheltering to avoid jeopardizing their good name. However, we test this directly and find that tax shelter incidence is just as prevalent among highly reputable firms. This again is consistent with tax sheltering not resulting in reputational costs, even for high reputation firms. Third, we can only speak to the reputational costs that we examine. Reputation is a broad concept that means different things to different people. We have endeavored to use a multi-faceted approach to consider reputational costs imposed by a variety of parties (e.g., shareholders, customers, and taxing authorities) that could manifest in various ways (e.g., managerial turnover, lost sales, increased taxes). Despite our best efforts, it is possible that there are reputational costs in some form that we have not examined. Finally, we are not claiming to have solved the under-sheltering puzzle. We are simply providing empirical evidence of whether reputational considerations explain the apparent under-sheltering. Our results cast doubt on reputational costs being a significant determinant of observed tax shelter usage, and we welcome future research on reputational effects and tax avoidance. 7

9 II. PRIOR LITERATURE Literature on Tax Avoidance and Tax Shelters There is a vast and growing empirical literature on tax avoidance, dating back at least as far as Scholes, Wilson and Wolfson (1990) and continuing to the present day. From this literature, several general themes emerge that are relevant to this study. To begin, there is ample evidence that tax avoidance is pervasive and adaptable. Research has shown that firms of all shapes and sizes engage in all manner of tax avoidance strategies, ranging from simple strategies like holding tax-exempt municipal bonds, to complex strategies such as debt-equity hybrid securities (Engel et al. 1999), cross-border avoidance strategies (Dyreng and Lindsey 2009), intangible holding companies (Dyreng, Lindsey and Thornock 2011), and COLI (Brown 2011). COLI shelters are a useful example of a tax avoidance strategy that many viewed as aggressive and that resulted in adverse scrutiny for firms that engaged in them. The COLI shelter involved firms taking out life insurance policies on their rank-and-file employees, and then receiving the death benefits if and when the employee died. The COLI shelter relied on asymmetric tax treatment between the firm s interest expense when borrowing against the life insurance policy and the income from investments made via the life insurance policy. The COLI shelter was the subject of unflattering coverage in the media, including the Wall Street Journal, which identified the companies that engaged in COLI alongside pictures of their actual deceased employees. Following the adverse media scrutiny, Congress moved into action, held hearings, and enacted tax law changes to curtail the shelter. The IRS challenged the tax benefits the companies had already claimed on their tax returns, a process that played out over many years. If there are reputational costs of tax avoidance, they are most likely to be seen in extreme cases like this one. 8

10 While many firms take advantage of tax avoidance opportunities, many others do not avail themselves of these opportunities. For example, Dyreng et al. (2008) report that approximately one-fourth of firms are able to maintain long-run effective tax rates below twenty percent. What is perhaps more interesting is that approximately one-fourth of firms have effective tax rates greater than thirty-five percent, not just for a single year, but over extended periods as long as ten years. With a U.S. federal tax rate of thirty-five percent and with lower rates in most other developed countries, this suggests that many firms are not actively engaging in tax avoidance (Markle and Shackelford 2011). The forces that are curtailing more widespread tax avoidance are not well understood. On average, it is reasonable to assume that firms are engaging in the optimal amount of tax avoidance. A maintained assumption in the literature, and in this paper, is that the presence of costs of tax avoidance that at some point outweigh the benefits. However, such costs have been difficult to identify. 5 The lack of tax avoidance more generally, especially in non-regulated settings and settings with no financial reporting trade-off, has been dubbed the under-sheltering puzzle. Essentially, the question is: what is holding those firms back from taking advantage of known tax avoidance opportunities being used by other firms? Reputational costs are often conjectured to be an important factor constraining tax avoidance, particularly the more aggressive forms of tax avoidance. As noted in Hanlon and Slemrod (2009), some firms, for example, are reported to apply a Wall Street Journal test to their tax avoidance activities, whereby they forgo such activities that would look unsavory if they were linked to the firm on the front page of the Wall Street Journal. However, firms also 5 The costs identified are financial reporting costs and regulatory costs, which apply to a subset of tax avoidance strategies and a subset of firms, respectively. See Frank, Lynch and Rego (2009) and Badertscher, Philips, Pincus and Rego (2009) for examples of financial reporting costs of tax avoidance and Mills, Nutter and Schwab (2010) for an example of the regulatory costs of tax avoidance. 9

11 have a clear incentive to engage in activities that increase their after-tax profitability, including tax avoidance. This tension is exemplified by a statement made by Wal-Mart s vice president for tax policy, expressing the pressure he faced from the CFO, who rides herd on us all the time that we have the world s highest tax rate of any major company. The statement was made under oath in litigation about a tax avoidance activity that, while reducing Wal-Mart s taxes (at least prior to being challenged), later landed Wal-Mart on the front page of the Wall Street Journal (Drucker 2007, A1). There is also anecdotal evidence that companies want to be perceived as being good corporate citizens that pay their fair share of taxes. General Electric, for example, has been criticized by the New York Times for paying no taxes to the U.S. government in 2011 despite being one of the largest companies in the world by earnings and by market capitalization. 6 GE immediately responded by pointing out its other contributions to society, such as being a major employer and exporter, its prior tax payments, as well as a tallying up a broader measure of its tax burden that includes payroll, property, and sales taxes paid. 7 These examples are consistent with the conjecture that firms perceive reputational costs from aggressive tax avoidance, especially when subjected to media scrutiny. To date, however, there is little in the way of empirical evidence on the validity of that claim. The closest evidence on the reputational effects of tax shelters is provided by Hanlon and Slemrod (2009), who examine the change in firms market values following public revelation that the firms engaged in tax sheltering, and by Graham et al. (2011) who survey tax executives in regards to their tax planning activities. Hanlon and Slemrod (2009) find that when tax shelter participation is revealed in the news media, the average decrease in market value is between and accessed September 30, accessed September 30,

12 percent, depending on the sample. 8 Graham et al. (2011) survey tax executives and find that nearly half agree that potential harm to their firm s reputation is very important when deciding what tax planning strategies to implement. Moreover, responding to a question about tax disclosures, twenty-seven percent of executives surveyed responded that risk of adverse media attention was very important in reducing their firm s willingness to be tax aggressive. Other research examines the role of political costs in determining effective tax rates, where political costs can be viewed as a type of reputational cost or at least related to reputational costs. Zimmerman (1983) hypothesizes that accounting choices are often driven by political costs, where taxes are a form of political costs that may influence the strategies undertaken by management. He shows that large firms have higher effective taxes, which he asserts are a function of the higher political costs faced by these firms. Mills et al. (2010) examine the influence of political costs on the tax avoidance in a sample of federal contractors. They find that federal contractors that are highly sensitive to political costs have higher effective tax rates, consistent with political costs driving the tax strategy of federal contractors. An interesting finding in their study is that only small federal contractors appear to have high effective tax rates; effective tax rates of firms subject to political scrutiny are decreasing in firm size, a finding the study attributes to larger firms having more political power. While firms may prefer to avoid public scrutiny and criticism of their tax sheltering activities, they also have a direct incentive to reduce their firms effective tax rates, as evidenced by the earlier quote from Wal-Mart s VP of taxes. Indeed, there is empirical evidence that some firms treat their tax departments as profit centers rather than cost centers (Robinson, Sikes and 8 By comparison, when news of financial statement fraud is reported, the average firm loses 38 percent of its market value (Karpoff, Lee and Martin, 2008). 11

13 Weaver 2010) and that firms incentivize their tax directors to reduce their effective tax rates (Armstrong, Blouin and Larcker 2011). Prior Literature on Reputational Costs We also draw on research on reputational costs in non-tax settings. In general, the extant literature examines the reputational costs of malfeasance on both firms and their managers. Various types of corporate misconduct, including financial statement fraud, earnings restatements, product recalls, environmental violations, and defense procurement, were in many cases followed by significant reputational penalties, including top management turnover, litigation, settlements, and large declines in market capitalization. For example, Karpoff, Lee, and Martin (2008a) track the careers of over 2,000 managers who allegedly engaged in financial misrepresentation. They find that nearly all (93 percent) of these managers are fired and have a more difficult time finding new employment. In addition, they lose a significant amount of wealth due to SEC fines and the decline in value of their holdings. Desai, Hogan and Wilkins (2006) examine whether managers of firms suffer reputational costs following earnings restatements. They find that top managers of firms that restate their earnings experience more turnover than managers at a control sample of firms. Bowen, Call and Rajgopal (2010) examine firms that are subject to exposure by whistleblowers. They find a number of adverse effects following the whistle-blowing announcement, including significant negative abnormal returns, restatements, shareholder lawsuits, and negative future operating performance. Prior research has shown that firms accused of accounting improprieties (Karpoff, Lee and Martin 2008b; Dechow, Sloan and Sweeney 1996), recalling defective products (Jarell and Peltzman 1985), and violating environmental regulations (Karpoff, 12

14 Lott and Wehrly 2005) are subject to adverse effects following revelation of the alleged misdeeds. However, there is also evidence that managers and firms are not always severely punished for improprieties. Beneish (1999) finds that managers of firms with extreme earnings overstatements do not have higher employment losses than managers of a matched sample of firms. Agrawal, Jaffe and Karpoff (1999) find no evidence that managers and directors of firms charged with fraud have higher rates of turnover. Finally, Srinivasan (2005) finds that outside directors face labor market penalties following accounting restatements, but not penalties through regulatory actions or litigation. In sum, there is evidence of significant reputational costs accruing to both firms and managers in a range of non-tax settings. This evidence suggests that it is reasonable to expect that tax avoidance, particularly of the extreme kind, can be accompanied by adverse reputational costs if it is discovered and subject to outside scrutiny. However, some evidence suggests the opposite is the case, in which managers go unpunished following instances of inappropriate behavior. Thus, whether tax avoidance leads to reputational penalties is an empirical question. III. DATA To assess firms reputational costs of aggressive tax avoidance, we concentrate on firms that were subjected to media scrutiny for having participated in a tax shelter. Our focus on tax shelters is motivated by the following interrelated reasons. Foremost, focusing on tax shelters gives us the best chance of finding reputational costs of tax avoidance, if they exist. Tax avoidance activities range from mundane strategies that are unlikely to attract adverse attention, to aggressive strategies that are contrary to the intent of Congress and potentially illegal. Tax 13

15 shelters tend to be at the extreme end of the tax avoidance spectrum. Second, other proxies for tax avoidance, such as low effective tax rates and high book-tax differences, are volatile (Dyreng et al. 2008), subject to measurement error when using financial statement information (Hanlon 2003), and vary with the macroeconomy (Seidman 2010). Third, by focusing on tax shelters that attracted media scrutiny, we give the best chance for the public (including shareholders and governments) to form a negative opinion about the firm. Given that outside stakeholders are the most likely to impose reputational penalties on tax aggressive firms, it is important to have an observable signal of tax aggressiveness. In summary, tax shelters give us the best chance of finding a reputational effect of tax avoidance, if one exists. However, if we observe reputational costs in the sample of tax shelter firms, we need to be careful about generalizing the results to less aggressive forms of tax avoidance. Conversely, if we do not observe reputational costs in this extreme sample, then it is unlikely that more mundane forms of tax avoidance result in significant reputation costs. Tax shelters are secretive by nature. As a result, it is difficult to identify a comprehensive sample of firms that are engaged in sheltering. Prior research has identified small samples of public cases of tax sheltering, including Graham and Tucker (2006, 44 observations), Hanlon and Slemrod (2009, 108 observations), Wilson (2009, 33 observations). To create a broad sample of tax shelter observations, we combine the samples in these prior studies of tax sheltering with 61 additional observations of the COLI shelter from Tax Notes. To the best of our knowledge, we have the largest public tax shelter database employed in current and past research. 9 9 Other researchers (e.g., Lisowsky, Robinson and Schmidt, 2011) have obtained larger samples of tax shelter transactions using confidential IRS data. Although that sample is useful for many research questions, the tax shelters in those studies are unobservable to the public and thus would not be suitable for tests of reputational penalties exerted by outside stakeholders. 14

16 We require the data to meet only minimal conditions to remain in the sample. Table 1 presents the details of our sample composition. We begin with 184 shelter observations from prior research and add 61 additional shelter observations from a Tax Notes article on the COLI tax shelter (Sheppard 1995). Of the resulting 245 observations, we remove 69 observations that are duplicated across the databases. 10 We then remove foreign firms (3 observations), those for which the initial date of public scrutiny is unclear or missing (15 observations), pre-1993 tax shelters (10 observations), those with insufficient data (17 observations) and those that upon closer inspection are not clearly tax shelters (2 observations). 11,12 Finally, we remove observations for which we cannot obtain a matched control firm (6 observations), according to the matching process described below. In the end, we are left with 113 shelter observations over the period We use a propensity-score matching technique that matches revealed shelter firm-years (i.e., the treatment group) to control firm-years that have a similar probability of having engaged in a tax shelter. For each sample firm, we calculate a predicted likelihood (i.e., propensity score) of having engaged in a tax shelter using the variables and point estimates from the reduced model in Lisowsky (2010). 13 We then match treatment firms to control firms in the same industry that have the closest propensity score in the year before the tax shelter revelation. The matching is done without replacement and thus each treatment firm has a unique control firm Since this study is conducted at the firm-year level, if there is more than one tax shelter revelation in a year we count that as one observation. 11 We begin our sample period in 1993 because SFAS 109 changed the rules regarding the financial reporting of taxes and we want to maintain similar accounting treatment of income taxes for all our observations. 12 Both incidents involved transfer pricing rather than tax sheltering. 13 Because not all of the variables in Lisowsky are readily available, we use the reduced prediction model from that paper, which is similar to the model in Wilson (2009). 14 We require that potential control firms have data for each analysis. We require that control firms have a minimum of 5 years of data for the one-year regressions (CEO and CFO turnover, sales and ad expense, Fortune reputation lists). We make no such requirement of treatment firms. Thus in the one-year regressions, there is not necessarily a one-to-one mapping of firm-years between the treatment and control groups. 15

17 We estimate all empirical analyses using a differences-in-differences design, which compares differences in reputational costs for the revealed shelter firms to their propensitymatched control firms before and after the shelter firm is caught. This design choice offers several advantages. First, it helps us to isolate the effect of being caught engaging in a tax shelter, separate from the characteristics of a tax shelter user. We expect the reputational effects of tax sheltering to manifest only for those firms that are caught relative to a control set of firms that have similar characteristics. Second, by evaluating differences between the revealed shelter firms and the matched control firms, we account for unobserved changes over time, such as changes in competitive and macroeconomic forces, which can confound interrupted time-series tests. We employ financial statement data from Compustat and executive turnover data from Execucomp in our analyses. To maximize the sample, we follow Dyreng and Lindsey (2009, p. 1296) and set the following variables to zero if missing: advertising expense, research and development expense, tax loss carryforwards, intangible assets, special items, and long-term debt. We also use their methodology to correct for errors in foreign tax expense, foreign pre-tax income, pre-tax domestic income, total pre-tax income, federal current tax expense, and worldwide current tax expense. Table 2 contains the descriptive statistics (Panel A) and correlations (Panel B) for revealed shelter firms and their matched control firms. For simplicity, the variables measured in this table are measured in the first year in which the tax shelter was subject to public scrutiny. The statistics in Panel A suggest that revealed shelter firms are large, reputable firms. In particular, these firms have mean (median) total assets of $10.4 billion ($10.1 billion). Thus, we have met our objective of identifying a sample of firms that are well known household names 16

18 (i.e., they have a reputation to protect). Moreover, tests of differences between the treatment and matched control firms suggest that they are similar for most of the variables we consider. IV. RESEARCH DESIGN AND RESULTS In this section, we detail the empirical tests and results for each of the reputational costs of being revealed as a tax shelter user. Given that reputational costs can take many forms, we examine reputation costs using a variety of approaches, including changes in managerial and auditor turnover, changes in sales revenue and advertising costs, changes in firm reputation as measured by media lists (e.g., Fortune s Most Admired ), differences in sheltering by high reputation firms, and changes in effective tax rates. In addition, we examine whether the reputational effect of tax sheltering is greater for subsamples likely to be more sensitive to reputational costs, such as retail firms and firms that had a low ex-ante probability of engaging in a tax shelter. The basic research design follows a differences-in-differences methodology in which a reputation variable is regressed against an indicator for the treatment firm (CAUGHTFIRM), an indicator for the tax shelter revelation year (CAUGHTYEAR) and the interaction of the two. Specifically, the general empirical specification will take the following form:. (1) In equation (1), REPUTATION is one of several proxies for firm and manager reputational costs, measured in year t for a given firm i. CAUGHTFIRM is an indicator variable equal to one for firms that were revealed to have been in a tax shelter, and zero for the control firms. CAUGHTYEAR is an indicator variable set equal to one, for both treatment and control firms, in the year it was revealed the treatment firm had a tax shelter and zero otherwise. 17

19 The set of control variables differs depending on the specific reputational cost being examined. All variables are defined in Appendix A. In all analyses that follow, we account for residual correlation by clustering the standard errors at the firm level and all continuous variables are winsorized at the 1 percent and 99 percent levels within each group (treatment and control). Does tax shelter scrutiny lead to top executive turnover? To examine whether firms that are revealed as having participated in a tax shelter experience higher CEO or CFO turnover subsequent to the revelation, we estimate equation (1) separately for CEO and CFO turnover. TURNOVER equals one if the CEO or CFO of firm i changed in year t, and zero otherwise. For this test, the variable of interest in equation (1) is β 3, the coefficient on the interaction of CAUGHTFIRM and CAUGHTYEAR, which for this test we predict to be positive. A positive β 3 indicates a greater incidence of CEO or CFO turnover following being accused of being in a tax shelter, relative to control firms over the same time period. A positive coefficient on the interaction of CAUGHTFIRM and CAUGHTYEAR is consistent with both firms and their top executives bearing a reputational penalty for getting caught in a tax shelter. In the implementation of equation (1), the set of control variables are those that previous research has shown to influence CEO or CFO turnover (Engel, Hayes and Wang 2003; Gilson 1989; Hennes, Leone, and Miller 2008; Menon and Williams 2008). SIZE is the natural log of average total assets, measured at the beginning and ending of the fiscal year (Compustat variable name AT). ABNORMAL RETURN is the stock return minus the return to the value-weighted CRSP index, summed over twelve months. ROA is return on assets, measured as the ratio of net income (Compustat NI) to average total assets. LEV is financial leverage, measured as ratio of 18

20 long-term debt (Compustat DLTT) to average total assets. CEO RETIRE is an indicator variable equal to zero if the CEO is aged 64 or older, and zero otherwise. In the turnover tests, all control variables are measured in the year before the turnover variable. Figure 1 plots the frequency of CEO and CFO turnover in event time around the year in which the tax shelter was revealed (year 0). For each event year, we present the number of CEOs (panel A) or CFOs (panel B) that left the company. The darkly shaded columns measure the turnover frequency for revealed shelter firms and the lightly shaded colums measure the turnover frequency for matched control firms. For both CEOs and CFOs, there is very little difference in turnover frequency between the revealed tax shelter firms and the control firms before or after the tax shelter revelation (i.e., in years 0, 1 and 2). Table 3 presents the results of logistic regressions estimating equation (1), in which the dependent variable is an indicator variable equal to one if the firm experienced CEO or CFO turnover that year and zero otherwise. 15 In column (1), we find no evidence that revealed shelter firms experience a higher likelihood of of CEO turnover than a matched control sample. In column (2), we examine turnover of CFOs, and find no evidence that CFOs experience higher turnover following a tax shelter revelation. In column (2), three of five control variables are significant at the five percent level, which suggests that the empirical model has sufficient power for these variables. Specifically, the coefficients on ABNORMAL RETURN and ROA are negative and significant and that on SIZE is positive and significant. Overall, in neither the univariate tests 15 Ai and Norton (2003) and Greene (2010) show that several mainstream statistical packages incorrectly calculate interaction effects in non-linear models such logit regressions. However, Kolasinski and Siegel (2010) argue that the interaction effects presented by these statistical packages are economically meaningful. Our interaction effects throughout the paper are calculated as usual, but inferences are robust to using the inteff procedure in Stata developed by Ai and Norton. 19

21 (Figure 1) nor the multivariate tests (Table 3) do we find evidence that a reputational cost of tax shelter revelation manifests in increased CEO or CFO turnover. 16 Does tax shelter scrutiny lead to audit turnover? The promotion of tax shelters often occurs through the firm s external auditor, at least during portions of our sample period (Maydew and Shackelford 2007). Even when the auditor is not involved in designing the tax shelter, the firm will often seek the auditor s agreement to the accounting treatment of the transaction prior to implementation. Thus, if the revealed shelter firms face reputation costs of tax avoidance, they may hold their auditor responsible when the shelter goes bad. Figure 1, Panel C presents the frequency of auditor turnover for revealed shelter firms relative to a matched control sample in event time surrounding the year of revelation. The figure shows that for every year, the frequency of auditor turnover for revealed shelter firms is equal to or less than that of the control firms. Moreover, in the year of revelation, none of the revealed shelter firms experienced an auditor turnover. Untabulated analyses show that the accounting firm KPMG, relative to the other big accounting firms, experienced no change in its audit and non-audit fees following allegations that it had promoted tax shelters to high net worth individuals. 17 We interpret all of this as evidence that auditor-related reputational effects do not appear sufficiently large to explain the undersheltering puzzle. 16 One possibility for this finding is that the legal proceedings for a tax shelter are sufficiently longer than the threeyear window we evaluate. However, we argue that reputational costs are likely to arise immediately after the shelter becomes public, which is consistent with the findings in prior research that companies and managers are immediately punished following revelations of corporate misconduct (e.g., Karpoff et al, 2008). 17 We note that KPMG was agreed to pay $456 million in fines, penalties, and restitution as part of its settlement with the U.S.. See for details. 20

22 Does tax shelter scrutiny influence sales revenue and advertising expense? We next examine whether firms accused of engaging in tax shelters suffer lost sales from customers and whether they had to increase advertising as a result. The intuition is that firms prefer to avoid negative media coverage to maintain an image as a good corporate citizen, particularly with their customers. Firms, especially those that depend on retail consumers, expend significant resources to build and defend their brand names and reputation. The first set of tests examine whether firms accused of engaging in tax shelters suffer lost sales revenue following the adverse media coverage relative to their matched control firms. The second set of tests focuses on possible responses to the adverse media coverage in the form of increased expenditures on advertising. The regressions are specified as follows equation (1), with SALES and AD EXPENSE as the dependent variables. SALES and AD EXPENSE are measured in levels and changes, where the level of sales (advertising expense) is measured as sales revenue (advertising expense), divided by average of total assets. Growth in sales (growth in advertising expense) is measured as the contemporaneous sales revenue (advertising expense) less that from the prior year, divided by average total assets. When SALES is the dependent variable, a negative coefficient on the interaction of CAUGHTFIRM and CAUGHTYEAR would indicate a reduction of sales revenue for firms following adverse media coverage accusing them of being in a tax shelter. If firms respond to the negative publicity from the media coverage by increasing their advertising expenditures, then we would expect a positive coefficient on the interaction of CAUGHTFIRM and CAUGHTYEAR when AD EXPENSE is the dependent variable. Both findings would be consistent with firms suffering reputational costs from aggressive tax avoidance. 21

23 We include a number of control variables in this implementation of equation (1) including SIZE, PPE, PPE, LEV, INTANGIBLE ASSETS, R&D EXPENSE, AD EXPENSE, NOL DUMMY, NOL, SPECIAL ITEMS, EXTRAORDINARY ITEMS, FOREIGN INCOME DUMMY, and FOREIGN INCOME. All variables are from the Compustat annual database and are measured over the same measurement window as the dependent variables. SIZE is the natural log of average total assets. PPE is the firm s property, plant and equipment, scaled by average total assets. ΔPPE is the change in the firm s property, plant and equipment, scaled by average total assets. LEVERAGE is the average long-term debt, scaled by average total assets. INTANGIBLE ASSETS is the firm s average intangible assets scaled by average total assets. R&D EXPENSE and AD EXPENSE are the firm s research and development and advertising expense, respectively, scaled by average total assets. We include two variables for the presence of net operating losses: NOL DUMMY, an indicator equal to one if the firm had a tax loss carryforward on its balance sheet at the beginning of year, and ΔNOL, which is the change in the firm s tax loss carryforward during the year, scaled by average total assets. SPECIAL ITEMS and EXTRAORDINARY ITEMS are the firm s special and extraordinary items scaled by average total assets. Finally, we include two variables to account for foreign operations: FOREIGN INCOME DUMMY is an indicator variable equal to one if the firm reported foreign income and zero otherwise, and FOREIGN INCOME is the firm s foreign income scaled by average total assets. Table 4 presents the results of estimating equation (1) when SALES and AD EXPENSE are the dependent variables. In columns (1) and (2), in which the dependent variables are SALES and SALES, we see that neither sales nor sales growth decrease for revealed shelter firms following the tax shelter revelation. In columns (3) and (4), in which the dependent variables are AD EXPENSE and AD EXPENSE, we see that advertising expense is no different for treatment 22

24 firms than for control firms, nor is it different for either group following the revelation of tax sheltering. We note that, in each of these models, the coefficients on many of the control variables are significant, indicating that the estimation had sufficient power to yield significance. For example, in column (1), nine of a possible thirteen coefficients on control variables are significant. However, the variable of interest, which is the interaction of CAUGHTFIRM and CAUGHTYEAR, is insignificant. Thus, across all models, we find no evidence of a reputational effect of tax shelter revelation that manifests in the form of reduced sales or increased advertising expense. Does tax shelter scrutiny negatively influence firm reputation in the media? We next examine the impact of tax shelter scrutiny on direct measures of the firm s overall reputation in the media. To do so, we obtain the Most Admired company and Best Company to Work For lists, which are compiled by Fortune magazine. 18 Following Bowen et al. (2010), we use a firm s presence on the Fortune lists as a proxy for a high overall reputation. Specifically, we estimate equation (1) with ADMIRED and ADMIRED&BEST as the dependent variables, where ADMIRED is an indicator set equal to one if the firm makes the Fortune Most Admired list, and zero otherwise. ADMIRED&BEST takes on a value of one if the firm makes either of the Fortune lists, and is set equal to zero otherwise. We use the same control variables as in the sales and advertising expense regressions. If firms suffer significant costs to their overall reputation in the media from tax shelter behavior, then we expect the coefficient on the interaction of CAUGHTFIRM and CAUGHTYEAR to be negative. We note that all of the firms in the sample are the subject of media scrutiny for their tax shelter, since that was a requirement to 18 For 2011, the Fortune lists can be found at: and accessed January 12,

25 be in the sample. Thus, this test is assessment of whether the scrutiny over one particular activity (i.e., tax shelter usage) has adverse effects on the firm s overall reputation in the media. In Table 5, we compare the reputation of tax shelter revelation firms to that of the propensity-matched control sample of firms. We estimate multivariate equation (1) separately for the Most Admired firms because we have a broader sample going back to 1993, which yields 3,585 firm-year observations. We then estimate equation (1) for the combination of Most Admired and Best Companies to Work For, which covers , yielding 1,313 firmyear observations. We find that, relative to the control sample, firms with tax shelters experience no significant change in their reputation once the tax shelter is made public. This holds for both samples (i.e., columns (1) and (2)). Many of the other determinants of reputation, such as SIZE and PPE, are statistically significant in the direction one would expect. As with the evidence presented earlier, the analysis here indicates that tax shelter scrutiny does not significantly reduce a firm s overall reputation. Does reputation affect the propensity to engage in a tax shelter? The prior tests have examined the ex post consequences to the firm s reputation from public scrutiny of tax shelter involvement. In Table 6, we examine whether the firm s reputation affects the ex ante probability of engaging in a tax shelter. Panel A of Table 6 examines the frequency with which firms on the Fortune Best Company or Most Admired lists are identified as having engaged in a tax shelter, compared to publicly traded firms that do not make either of the Fortune lists. The results do not indicate that higher reputation leads to lower tax shetler usage. If anything, the results in Panel A indicate higher tax shelter usage among high 24