DOI: 10.1111/1475-679X.12101
Journal of Accounting Research
Vol. 54 No. 1 March 2016
Printed in U.S.A.
Public Pressure and Corporate
Tax Behavior
S C O T T D . D Y R E N G ,∗ J E F F R E Y L . H O O P E S ,†
A N D J A R O N H . W I L D E‡
Received 30 July 2014; accepted 30 September 2015
ABSTRACT
We use a shock to the public scrutiny of firm subsidiary locations to investigate whether that scrutiny leads to changes in firms’ disclosure and corporate
tax avoidance behavior. ActionAid International, a nonprofit activist group,
levied public pressure on noncompliant U.K. firms in the FTSE 100 to comply with a rule requiring U.K. firms to disclose the location of all of their
subsidiaries. We use this setting to examine whether the public pressure led
scrutinized firms to increase their subsidiary disclosure, decrease tax avoidance, and reduce the use of subsidiaries in tax haven countries compared to
∗ Fuqua School of Business, Duke University; † Fisher College of Business, Ohio State University; ‡ Tippie College of Business, University of Iowa.
Accepted by Christian Leuz. We thank Martin Hearson and Chris Jordan of ActionAid
International for extensive help understanding ActionAid’s investigation of the FTSE 100
firms, and for providing us with data used in this paper. We thank PwC, U.K. for helping us with technical details of the U.K. tax system. We thank John Campbell, Travis Chow
(discussant), Dane Christensen, Cristi Gleason, Michelle Hanlon, Jim Hines (discussant),
Ken Merkley, Aaron Nelson, Tom Omer, Ed Outslay (discussant), Jeff Pittman, Steven Savoy,
Lisa De Simone, Jake Thornock, Connie Weaver, and Ryan Wilson for comments on a previous draft of this paper. We also thank workshop participants at the University of Iowa,
Vienna University of Economics and Business, the 2014 Midwest Accounting Research Conference at Ohio State University, the University of Michigan 2014 MTAXI Conference, the
2014 Taxation in a Global Economy Research Symposium, and the University of Oxford
2014 Summer Tax Symposium. We also appreciate financial support from the Center for International Business Education & Research at Ohio State University. This paper previously
circulated under the title, “Real Costs of Subsidiary Disclosure: Evidence from Corporate
Tax Behavior.” An Online Appendix to this paper can be downloaded at http://research.
chicagobooth.edu/arc/journal-of-accounting-research/online-supplements.
147
C , University of Chicago on behalf of the Accounting Research Center, 2016
Copyright 148
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
other firms in the FTSE 100 not affected by the public pressure. The evidence
suggests that the public scrutiny sufficiently changed the costs and benefits
of tax avoidance such that tax expense increased for scrutinized firms. The
results suggest that public pressure from outside activist groups can exert a
significant influence on the behavior of large, publicly traded firms. Our findings extend prior research that has had little success documenting an empirical relation between public scrutiny of tax avoidance and firm behavior.
JEL codes: H25; H26; H20; G39
Keywords: Real effects of disclosure; corporate taxation; corporate reputation
1. Introduction
Research examining the determinants and consequences of tax avoidance
in publicly traded firms has grown dramatically in the past decade (Hanlon
and Heitzman [2010]). Despite the substantial research on tax avoidance,
little is known about how firms respond to public scrutiny of their corporate
tax avoidance behavior. Decision makers within the firm potentially view
public scrutiny of tax avoidance activities negatively because of fears that
the public pressure could result in backlash against the firm or its products
from investors, politicians, and customers (Ernst & Young [2014], Graham
et al. [2014]), particularly if the firm is a government contractor or otherwise indirectly supported by government programs. On the other hand,
decision makers potentially view public scrutiny of tax avoidance positively
as it could signal to investors that fewer of the firm’s resources will be lost
to the government because the firm has strategically arranged its affairs to
reduce tax payments. Moreover, if the firm is compliant with tax rules and
regulations, the risks associated with tax avoidance could be small (Dyreng,
Hanlon, and Maydew [2014]). On net, firms are likely to trade off the benefits of tax avoidance with the political, reputational, and proprietary costs
that arise with increased public scrutiny.1 In this study, we use a shock to
the public scrutiny of firm subsidiary locations to investigate whether that
scrutiny leads to changes in firms’ corporate tax avoidance activities.
Disclosure of the location and identity of specific subsidiaries can reveal information about corporate tax behavior given the significant operating implications and tax consequences associated with the jurisdictions
in which firms locate their operations (Dyreng and Lindsey [2009], Creal
1 The recent “inversion wave” in the United States is an example of firms making a tradeoff
between lower tax bills and increased public scrutiny, especially by politicians. See, for
example, two letters written by Senator Dick Durban to AbbVie and Walgreens, available at
http://www.sdurbin.enate.gov/newsroom/press-releases/durbin-appalled-by-taxpayer-subsidized-abbvies-decision-to-renounce-its-american-citizenship and http://www.durbin.senate.
gov/newsroom/press-releases/durbin-to-walgreens-renouncing-your-american-corporatecitizenship-hard-to-defend, respectively.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
149
et al. [2014]). In particular, subsidiary disclosure is important because it
provides external parties with information about firms’ use of tax havens
and geographic exposure.
In contrast to U.S. regulations that only require disclosure of significant
subsidiaries, the United Kingdom’s Companies Act of 2006 (“Companies
Act”) requires firms to disclose the name and location of all subsidiaries,
regardless of size or materiality. Although the U.K. law went into effect
in 2006, ActionAid International, a global nonprofit dedicated to ending
poverty worldwide, discovered in 2010 that approximately half of the firms
in the FTSE 100 were not disclosing the name and location of all subsidiaries.2 ActionAid’s finding was prima facie evidence that the Companies
House, the U.K. body charged with making required disclosures available to
the public, was not enforcing the subsidiaries disclosure requirement. More
importantly, the fact that some firms chose not to comply with the law suggests that the cost of disclosing detailed information on subsidiaries (e.g.,
political costs, proprietary costs, and administrative burdens) was greater
than the benefit of a more complete information environment for the
noncompliant firms.3 Stemming from noncompliant firms’ revealed preferences toward nondisclosure, we assume that noncompliant firms have
higher costs of disclosing tax avoidance compared to the firms that complied with the disclosure law.
The ActionAid discovery was part of an attempt to publicly shame FTSE
100 firms for having subsidiaries located in tax haven countries in an effort
to increase these firms’ willingness to pay taxes.4 ActionAid’s shaming campaign encompassed two stages: (1) pressuring FTSE 100 firms to become
compliant with U.K. law requiring firms to disclose the name and location
of each of their subsidiaries, and (2) broadly publicizing all of FTSE 100
firms’ lists of subsidiaries to highlight tax haven use. In the initial stage,
beginning in 2010, ActionAid directly pressured individual firms to comply with the law. This pressure involved threatening the possibility of negative publicity for noncompliance with the disclosure rules and reminding
firms of ActionAid’s previous success in garnering negative media attention
2 The FTSE are the 100 largest market cap firms trading on the London Stock Exchange.
These 100 firms represent over 75% of total market cap traded on the London Stock
Exchange.
3 While the subsidiary disclosure portion of the Company Act of 2006 was unenforced, the
Companies Act of 2006 did provide for explicit monetary fines for noncompliance. However,
these fines, set at a maximum of £1,000 for initial noncompliance (for the company and
each officer in default), would have been inconsequential for the FTSE 100 firms we examine
(Companies Act of 2006, Section 410 (4) and 410 (5)).
4 ActionAid did not attempt to differentiate between subsidiaries located in nations for nontax reasons and those used for tax avoidance purposes. For example, it is plausible that a U.K.
firm could have a subsidiary in Ireland, a country labeled as a tax haven, for a variety of nontax
reasons ranging from supply chain efficiencies to lower operating costs.
150
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
aimed at specific firms.5 ActionAid also contacted the Companies House
and asked that it enforce the subsidiary disclosure regulation for noncompliant firms (see the letter in the appendix). Once ActionAid was presumably satisfied that FTSE 100 firms were in compliance with the disclosure
rule, in the second stage, ActionAid released a report that included the full
list of subsidiaries for the firms, presumably after the firms became compliant. This second stage attracted additional attention from the media and
Parliament (e.g., Choy, Lai, and Ng [2014]). As another part of this second
stage, ActionAid recollected and published the full lists of FTSE 100 firms’
subsidiaries in 2012.
The fact that a large proportion of FTSE 100 firms initially failed to comply with the subsidiary disclosure rule suggests that these firms likely perceived subsidiary disclosure as costly. Recent survey evidence suggests that
tax-related disclosures can impose political costs by altering interactions
with tax authorities and damaging government contracting relationships.
Further, tax-related disclosures can impose reputational costs by provoking customer retaliations and organized boycotts, which can require companies to spend resources to reassure key shareholders (Ernst & Young
[2014]).6 Ernst & Young (EY) reports that 89% of the largest respondent
firms indicate that they are somewhat or significantly concerned about media coverage of firms’ tax activities. A recent PwC report illustrates just how
costly public pressure can be: “We’re living in a world of 24-hour news and
Twitter, a world where information is amplified and distributed in seconds
and . . . where complex issues are brutally summarized. Great damage can
be done before a company has a chance to explain their position. Public
opinion, even if it’s based on inaccurate information, is powerful” (PwC
[2012]).7
5 In email correspondence with the authors, ActionAid explained that it used these tactics. For one example, see http://www.theguardian.com/business/2010/nov/29/sabmillerindia-africa-actionaid-report. For more recent examples of ActionAid’s work in this area,
see http://www.actionaidusa.org/eu/2013/12/barclays-must-stop-promoting-use-tax-havensafrica-actionaid-report, noting Barclays’s use of tax havens in Africa, and how it deprives
African school children of tax money. This particular attack on Barclays was part of a
larger campaign, #taxpaysfor, by ActionAid, in which people posted on Twitter what vital
services taxes paid for, and how tax haven use would deprive the government of funds to
be put to those purposes. Over 500 tweets reference #taxpaysfor (see https://twitter.com/
search?f=realtime&q=%23taxpaysfor&src=typd).
6 ActionAid’s 2011 report titled “Addicted to Tax Havens” garnered significant media coverage, resulting in members of the U.K. Parliament sponsoring and signing two early day
motions to press U.K. government officials to confront tax haven use, as well as resulting in abnormal market returns for firms targeted in the report (Choy, Lai, and Ng [2014]). The report
is available at http://www.actionaid.org/sites/files/actionaid/addicted to tax havens.pdf.
7 Even analysts have picked up on the scrutiny faced by companies as a result of their tax
avoidance. For example, in a July 2012 analyst report by GlobalData on SABMiller, the report
notes “As a part of ongoing operations, the company is directly and indirectly subject to lawsuits and other claims. In November 2010, ActionAid, an international charity organization
claimed that SABMiller avoided taxes in its developing markets of India and Africa. Action-
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
151
In this study, we use the ActionAid campaign to examine whether public pressure to comply with subsidiary disclosures rules made tax avoidance more costly.8 We investigate whether public pressure to disclose the
location of all subsidiaries sufficiently changed the costs and benefits of
tax avoidance and using tax haven operations such that it altered firms’
tax avoidance behavior. We begin by examining whether firms that did not
initially comply with the subsidiary disclosure rule (noncompliant firms)
exhibit significant increases in subsidiary disclosure following ActionAid’s
initial public pressure. We find that noncompliant firms did increase their
subsidiary disclosures immediately following the public pressure. Moreover,
the subsequent disclosures reveal disproportionately higher levels of tax
haven usage compared to the previous incomplete disclosures, suggesting
strategic nondisclosure of tax havens in the prepressure period.
Next, using a difference-in-differences research design, we find that noncompliant FTSE 100 firms report higher than expected effective tax rates
(ETRs) following the public scrutiny, indicating a decrease in tax avoidance
relative to FTSE 100 firms that were not affected by the scrutiny (compliant
firms). Specifically, our estimates suggest a 2.7 percentage point increase in
the ETRs of noncompliant firms, representing approximately £34.34 million in higher tax expense, compared to the ETRs of compliant firms in
the years following the initial public pressure to comply with the subsidiary
disclosure law.
We conduct several additional tests to confirm the connection between
public pressure and corporate tax behavior. First, we conduct a placebo
test in which we change the time period of the ActionAid scrutiny to one
of the three years before it occurred. In this placebo test, we find no evidence of significantly higher ETRs in either of the two years prior to the
public pressure period, mitigating concerns that our documented result is
merely capturing the effect of differing ETR trends among compliant and
noncompliant firms in the prepressure period. Second, we find that noncompliant firms decreased the proportion of their subsidiaries located in
tax havens relative to compliant firms, consistent with the increase in ETRs
Aid’s report alleges that the company avoided an estimated £20 m of taxes in Africa and India
every year. The report also revealed different systems utilized by the company, such as tax
haven places, valuable trademarks for African beers in Europe as well as procurement services
via a Mauritius-based subsidiary. It also avoids tax through the payment of management fees.
Though the company rejected these claims, it could create a negative image as well as increase
government scrutiny on the company’s operations in other countries.”
8 One common question regarding tax haven usage is why a given firm would ever need
more than one subsidiary in a tax haven, or multiple subsidiaries in the same haven nation.
There are several reasons. First, tax benefits from havens may be a function of how tax laws in
different countries interact with each other, often necessitating subsidiaries in various havens
(for example, the “Double Irish” was so named because it requires two Irish subsidiaries to
obtain the desired tax result (Duhigg and Kocieniewski [2012]). Second, many subsidiaries
are historical artifacts of acquisitions, etc., and it is simply easier to maintain two subsidiaries
with similar functions rather than combine them. Finally, holding different assets or projects
within distinct legal entities may reduce legal risk.
152
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
being driven by noncompliant firms curtailing tax avoidance strategies that
involve tax haven subsidiaries.
Third, we provide evidence that the decrease in tax avoidance is stronger
in the subsample of noncompliant firms with a decrease in the percentage of subsidiaries located in small (“dot”) tax haven countries—countries
where subsidiaries are unlikely to have operational substance (Desai,
Foley, and Hines [2006]).9 These results suggest that noncompliant firms
responded to negative public scrutiny by decreasing subsidiary use in locations where they would incur high disclosure costs (e.g., political and
reputation costs arising from increased scrutiny from taxing authorities,
customer and political outcry, or market penalties) and where it would be
relatively easy to dissolve subsidiaries without incurring significant operating costs. We also provide evidence that the effects of public pressure on
tax avoidance are most pronounced among the noncompliant firms that
hid a disproportionate percentage of tax havens in the prepressure period
and that operate in politically sensitive industries.
Finally, we conduct a test to validate our assumption that nondisclosure of
subsidiary information was a firm choice motivated by the relative costliness
of these disclosures. We examine market returns around a well-publicized
ActionAid report that highlighted tax haven use by FTSE 100 firms (Choy,
Lai, and Ng [2014]). We find that the short-window market returns for firms
that initially did not disclose their subsidiary list are significantly more negative than returns for firms that initially disclosed, highlighting one cost
(market penalties) of public scrutiny related to noncompliant firms’ use of
subsidiaries in tax haven nations. In combination, the results are consistent
with noncompliant firms failing to initially disclose the full list of their subsidiaries because they perceived costs to the disclosure, and that the costs
of disclosure were sufficiently large to alter their corporate tax avoidance
behavior.
This study contributes to the literature in several ways. First, our study
provides evidence of real consequences to disclosure of subsidiary locations. Existing theoretical research suggests that disclosure policies can encourage managers to behave myopically, leading to inefficient investment
(Edmans, Heinle, and Huang [2013], Gigler et al. [2014]). Although this
study is not a direct test of those theories, our results suggest that firms
make real changes to their corporate organizational structure to avoid disclosing too many subsidiaries in tax haven countries once the disclosure
requirements are enforced. This led to greater tax expense compared to
what would have been realized had the status quo been maintained. Thus,
our study has relevance for the tax literature that examines the economic
9 Desai, Foley, and Hines [2006] delineate between tax haven countries associated with tax
benefits, but few operational benefits (i.e., “dot” havens) and tax haven countries that provide
some tax benefits, but also have a sufficiently large workforce to potentially provide operational benefits (Big 7 tax havens). Typically, tax haven subsidiaries located in “dot” havens
suggest a tax-focused motive for the location decision.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
153
and managerial determinants of tax avoidance (Dyreng, Hanlon, and Maydew [2010], Holms [2011], Rego and Wilson [2012], Badertscher, Katz,
and Rego [2013], Brown and Drake [2013], McGuire, Wang, and Wilson
[2014]) and for the literature that suggests real consequences to changes
in disclosure policies (e.g., Kanodia [2007], Kanodia and Sapra [2015]).
Second, our evidence suggests that activist groups can have a meaningful influence on firm outcomes, improving our understanding of the
role of nontraditional monitors in overseeing firm behavior (Miller [2006],
Dyck, Volchkova, and Zingales [2008], Dyck, Morse, and Zingales [2010]).
Our evidence suggests that public pressure can increase both corporate
compliance with and government enforcement of existing laws. This evidence is consistent with other research indicating that firms respond to
pressure from external stakeholders (Smith [1995]), internal employees
(Wilde [2015]), and nonbinding shareholder votes (Ertimur, Ferri, and
Maber [2012]), and extends research that examines whether political costs
and reputational concerns influence corporate tax decisions (Hanlon and
Slemrod [2009], Gallemore, Maydew, and Thornock [2014], Graham et al.
[2014], Austin and Wilson [2015])
Finally, our study is informative to policy debates regarding the granularity of firm disclosures about geographic operations and the taxes paid to
the governments in which these operations are located (e.g., the Organisation for Economic Co-Operation and Development’s (OECD) initiative on
country by country reporting). Our evidence suggests that more disclosure
related to the geographic footprint of the firm can reduce tax avoidance
activities. This finding complements and extends the findings in Hope,
Ma, and Thomas [2013], who find changes in ETRs surrounding voluntary
changes in geographic segment reporting in the United States.
2. Background and Hypothesis
In 2006, the United Kingdom Parliament enacted changes in publicly
traded firms’ subsidiary disclosure requirements. Sections 409 and 410 of
the Companies Act mandate that U.K. companies disclose a complete list
of all subsidiaries either in their annual report or annual return.10 In 2010,
ActionAid International initiated a two-stage campaign aimed at shaming
large U.K. multinationals for their tax haven use.
In the first stage, ActionAid examined the annual returns and annual
reports of all firms included on the FTSE 100 to obtain data for its public shaming campaign. After obtaining the annual reports and annual returns, which are both publicly available, ActionAid noted that about half
10 The annual report refers to the financial statement report and other information similar
to the annual reports of publicly traded U.S. firms. U.K. law also requires firms to file an
“annual return,” which refers to a specific form that lists information about a firm’s directors,
address, shares, and shareholders. It is not the firm’s tax return. The annual return form can
be accessed, for a fee, at https://www.gov.uk/file-an-annual-return-with-companies-house.
154
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
of the FTSE 100 firms were not compliant. For example, many of the noncompliant firms indicated in their annual report that they would provide a
complete list of subsidiaries as an addendum to their annual return. However, upon further investigation, ActionAid found that many of the firms
provided no such list. After identifying noncompliant firms, ActionAid took
steps to motivate firms to provide the full list of subsidiaries. These steps included: (1) threatening firms with negative publicity related to their failure
to comply with disclosure law and (2) providing Companies House with a
list of noncompliant FTSE 100 firms and requesting that Companies House
compel such firms to provide the full list of their subsidiaries (see the appendix).11 ActionAid’s revelation of widespread noncompliance with the
disclosure mandate by some of the largest firms in the United Kingdom led
to an investigation by the U.K. Business Secretary (Holms [2011]).
After presumably feeling satisfied that all FTSE 100 firms were in compliance with the disclosure requirement, ActionAid initiated the second
stage of the campaign in which it published FTSE 100 firms’ subsidiary lists
and emphasized their widespread use of tax havens. This report, titled “Addicted to Tax Havens,” attracted significant media coverage and led to two
early day motions in the U.K. Parliament to press HMRC (Her Majesty’s
Revenue and Customs), the United Kingdom’s tax authority, to confront
U.K. multinationals on their haven use (Choy, Lai, and Ng [2014]). Then,
in 2012, ActionAid again collected and published the lists of FTSE 100
firms’ subsidiaries.12
We assume that noncompliant firms face higher costs of disclosing tax
avoidance compared to the compliant firms. In particular, we argue that
firms that initially complied with the disclosure requirement provided an
ex ante signal that they perceived a relatively low cost to the public knowing
about the location of their subsidiaries. However, firms that initially did not
disclose their full subsidiary list (i.e., noncompliant firms) provided an ex
ante signal that it was costly for them to provide information about the
location of their subsidiaries. To the extent that such costs are associated
with tax avoidance behavior, the exogenous and relatively sudden public
pressure to disclose a full list of subsidiaries likely led to changes in these
firms’ tax avoidance activities.
ActionAid’s efforts could affect the tax avoidance behavior of scrutinized
firms through two related forces. First, ActionAid intended to increase
public awareness of firms’ tax avoidance activities. Public scrutiny of tax
avoidance activities can increase the political and reputational costs of tax
avoidance in several ways, including shareholder penalties, tax enforcement actions, reputational damage, customer boycotts, and political
11 We use the list of noncompliant firms identified in ActionAid’s letter along with one
additional FTSE 100 firm that ActionAid specified in personal correspondence as the initial
firm that it identified and targeted for noncompliance.
12 This second report included the firms that comprised the FTSE 100 in the later period.
Most of these were part of the initial FTSE 100 at the time of the initial data collection effort.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
155
backlash (Hanlon and Slemrod [2009], Choy, Lai, and Ng [2014],
Graham et al. [2014]).
Two recent episodes in the United Kingdom illustrate some of these
costs. First, in 2012, numerous articles in the financial press revealed that
Starbucks paid no corporate income tax to the United Kingdom despite
strong growth in U.K. sales. The negative publicity resulted in verbal attacks from members of Parliament, customer boycotts of Starbucks stores,
significant drops in the firm’s reputation ratings, and numerous store closures (Christensen et al. [2014]). The reputational damage prompted Starbucks to voluntarily pay future taxes and relocate physical offices to the
United Kingdom, even though the company had purportedly been in compliance with U.K. tax law (Titcomb [2014]). A second example involves
Ethical Consumer magazine’s proposed boycott of Amazon, which was supported by eight members of the U.K. parliament. Ethical Consumer called
for a boycott based on the fact that Amazon paid only 0.1% of its U.K. sales
in U.K. taxes, noting that “if enough people boycott Amazon then we will
damage their business. Amazon’s market share and reputation matters.”13
In 2015, Amazon committed itself to abandoning the types of corporate
structures that succeeded in legally reducing their corporate tax obligation
in the United Kingdom (Bowers [2015]).14,15
These anecdotes suggest that public scrutiny can affect firm behavior.
Given that consumer boycotts often have negative implications for firm
value (Pruitt and Friedman [1986]), managers may be sensitive to customers who prefer to purchase goods and services from providers they perceive as good corporate citizens that pay their “fair share” of taxes. However,
recent work provides inconclusive evidence on the association between reputational concerns and corporate tax behavior. Prior research in other settings suggests that executives are concerned about protecting their reputation and finds that executives seek to influence media coverage (Westphal
and Deephouse [2010]), and recent survey evidence suggests that tax executives at multinational firms are concerned about the reputational costs
associated with corporate tax planning decisions (Graham et al. [2014]).
Yet, to date there is little archival evidence to support the claim that
reputational concerns influence tax activities (Gallemore, Maydew, and
Thornock [2014], Austin and Wilson [2015]).
13 See
http://www.ethicalconsumer.org/boycotts/boycottamazon.aspx.
and Venkatachalam [2015] provide a commentary on the costs of firms voluntarily
paying more tax as a result of reputational concerns.
15 Another example of taxes potentially tarnishing reputation comes from Bruce Packard,
an analyst for Seymour Pierce, noting in his January 3, 2012, Morning Meeting Flash report,
“However, we can’t help thinking that Barclays brand is being tarnished, last year the MP
Chuka Umanna, asked Bob Diamond how many off shore tax haven subsidiaries the bank
operated, and when the Chief Exec flannelled, the MP responded by suggesting the number
was over 300 . . . Barclays may argue that it is acting in the interests of shareholders to optimise ‘tax efficiency’ by reducing cash tax paid to the absolute minimum, but the share price
performance over the last decade suggests otherwise.”
14 Mintz
156
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
Tax-related disclosures can also directly influence enforcement by tax authorities. While it is unclear whether shareholders desire public disclosure
of the firm’s tax avoidance activities, tax authorities may demand the information. Research suggests that tax-related disclosures can prove useful to
the taxing authority in determining where to allocate scarce enforcement
resources (Mills [1998]). Thus, if the taxing authority can use the information to more efficiently audit firms’ tax positions, disclosure of information
related to the tax function may be costly to the firm. Such information disclosed for financial reporting purposes may provide new information for
tax authorities. For example, recent evidence suggests that the IRS finds
FIN 48 disclosures in the United States useful for tax enforcement purposes (Bozanic et al. [2014]). In a similar manner, mandatory subsidiary
disclosures could provide new information to tax authorities in the United
Kingdom and other jurisdictions. In fact, evidence suggests that such disclosures may have been useful to non-U.K. tax authorities investigating the
tax behavior of U.K.-based firms (Maylie [2011]).16 Whether disclosure requirements have an effect on tax avoidance behavior is uncertain, and there
is little academic evidence on the issue. One possible exception is Hope,
Ma, and Thomas [2013], who find firms that discontinued disclosure of geographic earnings after the implementation of SFAS 131 had lower ETRs
than other firms. However, whether the Hope, Ma, and Thomas [2013]
result relates to public scrutiny of subsidiary disclosure is unclear because
geographic segment information rarely corresponds to legal jurisdictions
that are meaningful for tax purposes.17
In summary, ActionAid’s investigation and subsequent public campaign
likely affect the corporate tax avoidance behavior of firms in the FTSE 100
in ways that correlate with the costs of such disclosure, especially for initially
noncompliant firms. This public scrutiny could directly affect corporate tax
avoidance behavior by increasing public awareness of tax avoidance activities or indirectly by requiring disclosure of more information related to the
tax function, which could in turn increase the costs of tax avoidance. Accordingly, we compare the tax avoidance activities of firms that were forced
to comply with the subsidiary disclosure requirement with the tax avoidance
activities of firms that were already in compliance with the requirement before the ActionAid investigation.
16 It is very unlikely that the increase in tax expense we observe is the actual result of tax
authority action against FTSE firms, as tax enforcement actions are a lengthy process, making
it unlikely that any immediate response we observe would be the direct result of tax authority
sanctions.
17 Only 0.87% of all segments listed in the Compustat Segment data can be associated with
a specific tax haven country, and only 0.45% can be associated with “dot” havens. However,
geographic segment disclosures may be influenced by taxes. For example, Leuz [2004, p. 181]
notes, in references to geographic segment reports, “tax authorities may use these disclosures
in transfer pricing disputes.”
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
157
3. Sample Selection and Descriptive Statistics
3.1.
SAMPLE SELECTION
Our primary sample consists of all firms listed on the FTSE 100 in 2010.
Using Compustat Global, we search for data for these 100 firms between
1997 and 2012, which results in a maximum sample size of 1,500 firm-years.
Our primary dependent variable is ETR, which we define as the book ETR
less the statutory tax rate. We require firm-year observations to have nonmissing values of income tax expense (TXT), and positive values of pretax income (PI) in order to compute ETR so that values of ETR will be
interpretable in our regression models (Dyreng, Hanlon, and Maydew
[2008]). We also require nonmissing values of Size, Leverage, Intangibles, Inventory Intensity, RD Intensity, Capital Intensity, Capex, Return on Assets, and
%Havens. Table 1 provides formal definitions for each of these variables.
To be included in the tax avoidance regression analyses, we require firms
to have all model data in at least the immediate pre- and postpressure periods (i.e., 2009 and 2010).
Because we have a small sample, our tests could be easily influenced by a
few influential observations. To mitigate this risk, we winsorize ETR at [0,1]
(prior to subtracting the statutory tax rate), and winsorize all other continuous variables at the 1st and 99th percentiles. We also remove observations
with a Cook’s D outlier statistic in the top 2% of the sample after estimating
our primary regression model (model 1, described below).18 These criteria leave a sample of 1,010 firm-year observations between 1997 and 2012,
corresponding to 77 unique firms listed on the FTSE 100 in 2010.
Our data set consists of 515 observations (corresponding to 39 unique
firms) that did not comply with subsidiary disclosure laws, and 495 observations (corresponding to 38 unique firms) that did comply. We create an
indicator variable, Incomplete Subs List, that is equal to one for the 515 firmyears corresponding to the 39 firms that were not complying with subsidiary
disclosure laws when ActionAid began its investigation. We also create an
indicator variable, Post Pressure, that is equal to one for all firm-years from
calendar year 2010 and later, and zero otherwise.
18 We also take additional measures to mitigate concerns that our results are driven by a
relatively small number of influential observations. First, we reestimate model (1) without removing influential observations based on the Cook’s D outlier statistic. Second, we reestimate
the model after removing the top 2% of influential observations based on the Cook’s D outlier statistic, the DFFIT outlier statistic, and/or studentized residuals. Third, we reestimate the
model using robust regression. In each of these analyses, we find qualitatively similar results
(i.e., the coefficient on Incomplete Subs List × Post Pressure retains the same sign and significance
as reported in the primary tests). We report the results of these additional tests in table A7,
panel A in the Online Appendix. These results mitigate concerns that research design choices
related to influential observations drive our results.
−0.006
0.288
0.510
0.198
8.865
0.177
0.179
0.065
0.012
0.311
0.061
0.106
0.242
1,010
1,010
1,010
1,010
1,010
1,010
1,010
1,010
1,010
1,010
1,010
1,010
1.000
0.000
8.835
0.153
0.078
0.042
0.000
0.234
0.046
0.091
0.222
0.285
Median
Mean
−0.003
N
1,010
0.500
0.399
1.793
0.147
0.208
0.070
0.031
0.274
0.062
0.079
0.109
0.127
0.125
Std Dev
P25
0.000
0.000
7.568
0.060
0.012
0.004
0.000
0.076
0.016
0.050
0.168
0.236
−0.052
P75
1.000
0.000
10.014
0.247
0.317
0.102
0.004
0.522
0.087
0.144
0.294
0.331
0.039
8.106
0.184
0.191
0.074
0.012
0.335
0.072
0.119
0.219
−0.012
−0.012
−0.013
0.278
0.286
0.249
Mean
7.970
0.175
0.108
0.052
0.000
0.261
0.059
0.103
0.208
−0.004
−0.005
0.000
0.286
0.292
0.260
Median
Firms with Incomplete
Subs List (N = 515)
9.655∗ ∗ ∗
0.169
0.167∗
0.055∗ ∗ ∗
0.012
0.286∗ ∗ ∗
0.050
0.092∗ ∗ ∗
0.267∗ ∗ ∗
0.007∗ ∗
0.014∗ ∗ ∗
−0.022
0.298∗ ∗
0.312∗ ∗ ∗
0.240
Mean
9.502∗ ∗ ∗
0.123∗ ∗ ∗
0.053∗ ∗ ∗
0.031∗ ∗ ∗
0.000
0.158∗ ∗ ∗
0.033∗ ∗ ∗
0.074∗ ∗ ∗
0.227
−0.008
−0.005
−0.018
0.284
0.294
0.248
Median
Firms with Complete
Subs List (N = 495)
ETR is (txt/pi), bound between [0,1] less the top corporate statutory tax rate for the United Kingdom in the given year. Incomplete Subs List is an indicator variable equal to one
for FTSE 100 firms that did not disclose the full list of their subsidiaries prior to the 2010 public pressure to disclose and equal to zero, otherwise. Post Pressure is an indicator variable
equal to one for firm-year observations ending during 2010 (i.e., following the beginning of the increase in public pressure) or later and equal to zero, otherwise. Size is the natural
log of total assets (at). Leverage is the long-term debt (dltt), scaled by total assets (at). Intangibles is intangible assets (intan), scaled by total assets (at). Inventory Intensity is inventory
(invt), scaled by total assets (at). RD Intensity is research and development expense (xrd), scaled by total assets (at). Capital Intensityt -1 is the net property, plant, and equipment
(ppent), scaled by lagged assets (at). Capex is capital expenditures (capx), scaled by lagged assets (at). Return on Assets is pretax income (pi), scaled by total assets (at). % Havens is
the percentage of total subsidiaries ultimately reported that are located in tax havens (as indicated in ActionAid’s initial report). We set missing values of xrd, intan, and capx to zero.
We winsorize continuous variables at the 1st and 99th percentiles.
∗∗∗ ∗∗
, , and ∗ denote statistical significance at the 1%, 5%, and 10% levels, respectively, using two-sided tests.
ETR
ETR Prepressure
ETR Postpressure
Unadjusted ETR
Unadjusted ETR Prepressure
Unadjusted ETR Postpressure
Incomplete Subs List
Post Pressure
Size
Leverage
Intangibles
Inventory Intensity
RD Intensity
Capital Intensityt -1
Capex
Return on Assets
% Havens
Full Sample
TABLE 1
Summary Statistics
158
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
3.2.
159
DESCRIPTIVE STATISTICS
Table 1 reports the summary statistics for our primary sample. We present
the statistics for the full sample and also report the statistics separately for
the firms in the noncompliant (Incomplete Subs List = 1) and compliant subsamples (Incomplete Subs List = 0). During the prepressure sample period,
noncompliant firms have an average ETR of −1.2%, which is significantly
lower than compliant firms’ average ETR of 1.4% (p-value < 0.01). Recall
that ETR is adjusted for the statutory tax rate, which changed over our sample period.19 Using mean (median) pretax income of £1,271.98 million
(£566.00 million) for the noncompliant firms in 2010, this difference represents about £33.07 million (£14.72 million) less in tax expense for the
average noncompliant firm compared to the average compliant firm in the
prepressure period.20 Lower ETRs suggest that noncompliant firms’ initial reluctance to disclose their subsidiary list may indeed be related to tax
avoidance behavior. Tellingly, the pattern reverses in the postpressure period such that noncompliant firms have higher ETRs compared to compliant firms—average ETRs of −1.3% and −2.2%, respectively—representing
a difference of about £11.45 million (£5.09 million) more in tax expense for
the average (median) noncompliant firm compared to the average (median) compliant firm, although the difference is statistically insignificant.21
Table 2 presents the correlation matrix for the variables described in
table 1. Consistent with the table 1 summary statistics, noncompliant firms
are smaller, more profitable, and have a lower percentage of tax havens
actually reported. Although the univariate tests suggest that noncompliant
firms tend to be smaller statistically, all firms in our sample are part of the
FTSE 100, which comprises the largest firms in the United Kingdom and,
thus, are likely to be associated with relatively higher levels of scrutiny from
external parties.
4. Empirical Tests and Results
4.1.
SUBSIDIARY DISCLOSURE
ActionAid’s public pressure campaign specifically targeted firms’ use
of tax haven subsidiaries. In this subsection, we examine whether firms
19 Unadjusted ETRs for noncompliant firms averaged 28.6% in the prepressure period,
which is significantly lower than compliant firms’ average unadjusted ETR of 31.2% (p-value
< 0.01).
20 Mean (median) PI for all firms in the sample is £1,404.15 million (£484.95 million)
over the sample period, suggesting a difference of about £36.51 million (£12.61 million) less
in tax expense for the mean (median) noncompliant firm compared to the mean (median)
compliant firm.
21 In computing the differences in tax expense, we use ETRs rounded to the nearest thousandth, consistent with the presentation in the tables. For consistency, we use PI (shares outstanding) in 2010 for the noncompliant firms in our primary sample to compute the tax savings (earnings per share effects) we discuss throughout the paper, including for additional
analyses that may have slightly smaller sample sizes.
−0.019
−0.049
0.062
−0.106
0.046
0.088
−0.025
−0.068
0.026
−0.019
−0.041
(1)
0.015
−0.446
0.090
0.075
0.127
0.011
0.122
0.210
0.204
−0.203
0.148
0.087
0.177
0.012
0.054
−0.027
−0.028
0.056
−0.003
(3)
−0.059
0.015
(2)
−0.078
(4)
−0.074
−0.126
−0.228
−0.020
−0.132
−0.282
−0.416
0.145
0.144
−0.432
0.132
(5)
0.247
0.154
0.027
0.351
0.235
0.027
−0.032
−0.173
0.051
0.063
−0.160
(6)
0.061
0.312
−0.270
−0.225
0.078
−0.014
0.004
0.057
0.161
−0.157
0.124
(7)
0.283
0.300
0.349
0.437
−0.166
0.021
0.136
−0.030
−0.240
0.060
−0.046
(8)
0.073
0.046
0.280
0.048
−0.031
0.000
0.001
−0.149
−0.141
0.215
0.110
(9)
0.807
0.298
−0.180
−0.139
0.089
−0.037
−0.110
0.262
−0.412
−0.031
−0.173
(10)
0.452
−0.138
−0.039
0.180
−0.034
−0.284
0.136
−0.297
0.053
−0.102
0.601
(11)
−0.088
−0.082
0.167
0.025
−0.403
0.057
−0.050
0.313
0.248
0.156
0.336
(12)
−0.044
−0.224
−0.002
0.114
−0.047
−0.003
−0.153
−0.006
−0.157
0.013
−0.118
This table tabulates the Pearson’s (above the diagonal) and Spearman’s (below the diagonal) correlations for our sample. We report significant (p-value < 0.05) coefficients in
bold. ETR is the ETR (txt/pi), bound between [0,1] less the top corporate statutory tax rate for the United Kingdom in the given year. Incomplete Subs List is an indicator variable
equal to one for FTSE 100 firms that did not disclose the full list of their subsidiaries prior to the 2010 public pressure to disclose and equal to zero, otherwise. Post Pressure is an
indicator variable equal to one for firm-year observations ending during 2010 (i.e., following the beginning of the increase in public pressure) or later and equal to zero, otherwise.
Table 1 defines all other control variables (Size, Leverage, Intangibles, Inventory Intensity, RD Intensity, Capital Intensityt -1 , Capex, Return on Assets, and % Havens). We winsorize continuous
variables at the 1st and 99th percentiles.
(1) ETR
(2) Incomplete Subs List
(3) Post Enforcement
(4) Size
(5) Leverage
(6) Intangibles
(7) Inventory Intensity
(8) R D Intensity
(9) Capital Intensityt -1
(10) Capex
(11) Return on Assets
(12) % Have ns
Variable
TABLE 2
Correlation Matrix
160
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
Number of Multiple Annual
Return Filings in Calendar Year
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
161
16
14
12
10
8
6
4
2
0
2006
2007
2008
2009
2010
2011
2012
2013
2014
Year
Incomplete Subs List Firms
Non-Incomplete Subs List Firms
FIG. 1.—Multiple annual return filings by year. This figure graphs the number of multiple
annual return filings for Incomplete Subs List Firms and for Non-Incomplete Subs List Firms, by year.
actually changed their subsidiary disclosure policies in response to the ActionAid pressure. One way firms could have responded was to release a
new, revised annual return. In figure 1 we plot the number of firms that
had multiple annual return filings in each year separately for firms that
did and did not comply with the disclosure rule in the prepressure period.
Figure 1 shows noncompliant firms were more likely to file multiple annual returns in 2010, the year of the ActionAid pressure, consistent with
the pressure driving disclosure changes.
We also analyze the location of these newly disclosed subsidiaries. If tax
haven usage truly motivated the initial noncompliance, then we would expect newly revealed subsidiaries to be disproportionately located in tax
havens.22 Table 3 lists the 25 most commonly disclosed countries in the
full disclosure in 2011, compared to the number of subsidiaries in these
countries in 2008.23 In 2008, there were a total of 1,122 subs listed in these
25 countries, 10.96% of which were in tax haven nations (highlighted in
gray). In 2011, after ActionAid had induced compliance with the disclosure rule, there were 7,251 total subsidiaries, 15.40% of which were located
in tax havens. Consistent with tax haven subsidiaries being excluded from
disclosures in the prepressure period, the number of disclosed tax haven
subsidiaries grew more quickly between 2008 and 2011 than the number
of disclosed subsidiaries in nonhaven countries. For the entire sample, and
using all subsidiary locations, figure 2 graphs the percentage of subsidiaries
located in tax havens, and an increase in the percentage of tax havens from
the prepressure period to the postpressure period is visually evident.
22 It is also possible that, if tax haven subsidiaries were smaller than other subsidiaries and if
firms merely imposed a materiality threshold in spite of the disclosure rule, we might see that
tax haven subsidiaries were disproportionately excluded.
23 We chose 2008 as the preyear because it is clearly removed from ActionAid’s actions.
Similar results obtained using 2009.
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
162
TABLE 3
Location of Disclosed Subsidiaries for Noncompliant Firms
Country
Number of Number of
Number of
Disclosed
Disclosed
Disclosed
Percentage Percentage
Subsidiaries Subsidiaries
Subsidiaries
Change
Change
for Fiscal
in ActionAid in ActionAid from 2008 from 2011
Year 2008 Report in 2011 Report in 2012 to 2011
to 2012
UNITED KINGDOM
USA
AUSTRALIA
SOUTH AFRICA
NETHERLANDS
JERSEY
CANADA
FRANCE
IRELAND
CHINA
GERMANY
HONG KONG
INDIA
BRITISH VIRGIN
ISLANDS
CHILE
LUXEMBOURG
ITALY
BRAZIL
KAZAKHSTAN
SPAIN
MEXICO
SINGAPORE
SWITZERLAND
RUSSIA
BELGIUM
Total subsidiaries in tax
havens
Total subsidiaries not in
tax havens
Percentage in havens
372
147
71
46
44
16
40
52
13
35
59
12
30
2
3,344
552
483
435
320
248
191
188
152
151
146
103
99
93
3,504
614
533
417
342
268
190
160
149
191
154
101
112
91
799%
276%
580%
846%
627%
1,450%
378%
262%
1,069%
331%
147%
758%
230%
4,550%
5%
11%
10%
−4%
7%
8%
−1%
−15%
−2%
26%
5%
−2%
13%
−2%
19
6
20
17
28
26
16
14
16
8
13
123
93
87
79
76
75
62
62
59
55
50
48
1,117
96
93
88
92
72
72
64
63
55
51
45
1,162
389%
1,350%
295%
347%
168%
138%
288%
321%
244%
525%
269%
1,296%
3%
7%
11%
21%
−4%
16%
3%
7%
0%
2%
−6%
2%
999
6,134
6,455
369%
6%
10.96%
15.40%
15.26%
351.57%
31.20%
This table lists the number of subsidiaries located in each country in 2008, in the ActionAid report in
2011, and in the ActionAid report in 2012. Listed are the 25 countries with the most disclosed subsidiaries
in the 2011 ActionAid report. The sample is limited to noncompliant firms that disclose at least some
subsidiaries in all three years. Countries identified by ActionAid as tax havens are highlighted in shaded
gray.
More repeat filings in 2010 for noncompliant firms and an increase
in the percentage of subsidiaries in tax havens suggest that the ActionAid pressure was indeed sufficient to change the disclosure behavior of
noncompliant firms in our sample. These findings also suggest that the
firms in our sample had systematically failed to disclose subsidiaries located
in tax haven countries.
4.2.
TAX AVOIDANCE
In the previous subsection, we document changes in disclosure following
public pressure. In this subsection, we investigate whether public pressure
related to subsidiary disclosure affects corporate tax avoidance activities.
Percentage of Subsidiaries in
Tax Havens
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
163
0.21
0.19
0.17
0.15
0.13
0.11
2008
2009
ActionAid Report in 2011
FIG. 2.—Percentage of subsidiaries in tax havens for firms that did not initially comply with
the disclosure rule. This figure depicts the percentage of subsidiaries that are incorporated in
tax havens for firms that did not initially comply with the disclosure rule in 2006 mandating
full disclosure of all subsidiary locations.
We use compliant and noncompliant firms and the pre- and postpressure
period to employ a difference-in-differences estimation to identify the effect of public scrutiny on firms’ tax avoidance activities. The differencein-differences research design allows us to control for time-invariant differences between treatment firms (noncompliant firms) and control firms
(compliant firms) as well as for economic trends common to both treatment and control firms. We estimate the following difference-in-differences
model using a panel of data from 1997 to 2012:
ETR it = βi + β1 Incomplete Subs List i + β2 Post Pressure t
+ β3 Incomplete Subs List i ∗ Post Pressure t + β4 Size it
+ β5 Leverage it + β6 Intangibles it + β7 Inventory Intensityit
+ β8 RD Intensityit + β9 Capital Intensityit−1 + β10 Capex it
+ β11 Return on Assets it + β12 %Haven it=2011 + it .
(1)
The dependent variable, ETR, is widely available for public firms in the
United Kingdom, and is a well-accepted summary measure of firms’ tax
avoidance behavior (e.g., Hanlon and Heitzman [2010)].24 β 1 reflects the
average prepressure difference in ETR between noncompliant and compliant firms. β 2 reflects compliant firms’ average difference in ETR between
24 While widely accepted, ETR does have certain well-known limitations. For example, ETR
does not capture all tax avoidance activity, such as the use of accelerated depreciation or other
tax savings mechanisms designed to delay the payment of tax, rather than completely avoid
it (Neubig [2006]). That firms may try to reduce their subsidiary disclosure requirements
without affecting their tax expense, and therefore earnings, is important, especially in light
of evidence that firms will undergo costs to avoid depressing earnings (Erickson, Hanlon, and
Maydew [2004]).
164
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
the pre- and postpressure periods. The coefficient of interest in the model
is β 3 , which is the difference-in-differences estimate of the effect of public
pressure on tax avoidance. The remaining coefficients capture the effects
of the control variables on ETR. The control variables are determinants
of tax avoidance documented in prior research, and we include them in
the model to help alleviate concerns that correlated omitted variables are
confounding our inferences. We control for firm size (Size) using the natural log of assets because larger firms potentially have higher political costs
(Zimmerman [1983]) or greater tax planning opportunities (Rego
[2003]). We control for pretax profitability (Return on Assets) and the use of
tax havens (%Havens) given prior work that suggests that firm performance
and tax haven usage are associated with ETRs (Rego [2003], Dyreng and
Lindsey [2009]).25 Following prior research, we also control for other firm
attributes that are potentially associated with firms’ tax planning activities,
including Leverage, Intangibles, Inventory Intensity, R&D Intensity, Capital Intensity, and Capex (Chen et al. [2010], Hoopes, Mescall, and Pittman [2012],
Hanlon, Hoopes, and Shroff [2014]). Table 1 provides definitions for each
variable. Given that the event affects all treatment firms in the same time
period, we are concerned that standard errors could be understated. In the
reported results, we use bootstrapped standard errors, which are arbitrarily
robust to different structures in the data.26
Next, as another way to corroborate that the inference is not due to
random chance, we conduct a bootstrapping placebo test by scrambling
the indicator variable for Incomplete Subs List across random firms, creating a placebo treatment. We maintain the true value for the Post Pressure
variable. Counting the number of times we estimate a coefficient β 3 using
the placebo treatment that is larger than β 3 using the true treatment is an
estimate of the probability the results would obtain if the assignment of
treatment to firms were random. We repeat this estimation 1,000 times for
our regression in model 1. In 35 out of 1,000 random samples, we estimate
coefficients that are larger than the 0.038 coefficient, corresponding to a
bootstrapped p-value of 0.035.27
25 By definition, subsidiary disclosure was incomplete for the noncompliant firms prior to
the public pressure period. Accordingly, we measure percentage of havens (%Havens) using
ActionAid’s initial complete subsidiary lists released following the public pressure, when firms’
subsidiary disclosures were presumably complete. Because it is not possible to observe a firm’s
true list of subsidiaries with publicly available information, it is possible that this list is incomplete or is measured with error. However, we expect nondisclosure to work against our finding
evidence consistent with the hypothesis.
26 We provide details on the bootstrapped standard errors in the Online Appendix. We
estimate bootstrapped standard errors using 1,000 iterations for each model. When we include
industry fixed effects in the regressions, one or more of the parameters may not be estimated
in some of the bootstrap iterations. We specify in the notes to the applicable tables when the
number of converged iterations is less than 1,000.
27 We also conduct the bootstrapping analysis including firm fixed effects. In 147 out of
1,000 random assignments of the treatment, we estimate a coefficient that is larger than 0.027,
165
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
TABLE 4
Public Pressure and Corporate Tax Avoidance
(1)
(2)
ETR
Pred.
Incomplete Subs List
Post Pressure
Incomplete Subs List × Post Pressure
Size
Leverage
Intangibles
Inventory Intensity
RD Intensity
Capital Intensityt -1
Capex
Return on Assets
% Havens
Intercept
Industry fixed effects
Year fixed effects
Firm fixed effects
Bootstrapped standard errors
Observations
Adj. R 2
+
ETR
Coef.
(p-Value)
−0.000
−0.021
0.038∗∗
0.021∗∗∗
−0.078∗∗
0.070∗∗
0.032
0.032
−0.027
0.408∗∗∗
0.006
−0.076
−0.228∗∗∗
Yes
Yes
No
Yes
1,010
0.217
(0.998)
(0.332)
(0.016)
(0.000)
(0.011)
(0.019)
(0.748)
(0.807)
(0.466)
(0.000)
(0.925)
(0.149)
(0.000)
Coef.
(p-Value)
−0.030
0.027∗
0.031∗∗∗
−0.009
0.030
0.015
−0.161
0.052
0.231∗∗
−0.088
(0.215)
(0.060)
(0.000)
(0.823)
(0.434)
(0.911)
(0.659)
(0.202)
(0.025)
(0.269)
−0.298∗∗∗
No
Yes
Yes
Yes
1,010
0.308
(0.000)
This table reports the results of tests examining the effect of public pressure on corporate tax behavior. ETR is the ETR (txt/pi), bound between [0,1] less the top corporate statutory tax rate for the United
Kingdom in the given year. Incomplete Subs List is an indicator variable equal to one for FTSE 100 firms that
did not disclose the full list of their subsidiaries prior to the 2010 public pressure to disclose and equal to
zero, otherwise. Post Pressure is an indicator variable equal to one for firm-year observations ending during
2010 (i.e., following the beginning of the increase in public pressure) or later and equal to zero, otherwise.
Table 1 defines all other control variables (Size, Leverage, Intangibles, Inventory Intensity, RD Intensity, Capital
Intensityt -1 , Capex, Return on Assets, and % Havens). We winsorize continuous variables at the 1st and 99th
percentiles. We compute p-values based on bootstrapped standard errors. We bold lines related to coefficients of interest. The superscripts asterisks ∗∗∗ , ∗∗ , and ∗ denote statistical significance at the 1%, 5%, and
10% levels, respectively, using two-tailed tests.
4.2.1. Results from Estimating Model (1). We present the results of this
test in table 4. In column (1), the coefficient on Incomplete Subs List ×
Post Pressure (β 3 ) is positive and significant (p-value < 0.01), indicating
that scrutiny of noncompliant FTSE 100 firms is associated with significantly higher ETRs relative to control firms, which were already compliant
with subsidiary disclosure requirements. The coefficient estimate of 0.038
corresponding to a bootstrapped p-value of 0.147. In table A7, panels B and C in the Online
Appendix, we report our results using other techniques to adjust standard errors (e.g., clustering by year clustering by firm, clustering by year and by firm, etc.). Conclusions are not sensitive to these choices. In addition, we draw similar conclusions from three-year ETR changes
analyses with and without bootstrapped standard errors (reported in table A6, panel B in the
Online Appendix). Finding similar results using a changes analysis limited to one observation
per firm (i.e., collapsing values in the three-year pre- and postperiods into average values for
the firm) further mitigates concerns that the primary results reflect the effects of biased standard errors in panel data leading to an overrejection of the null hypothesis (e.g., Bertrand,
Duflo, and Mullainathan [2004]).
166
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
suggests that public pressure related to subsidiary disclosure is associated
with ETRs that are 3.8% of PI (£48.34 million of tax expense) greater for
noncompliant firms compared to control firms.28 These results imply that
political and reputational costs of the ActionAid pressure were at least
£48.34 million, otherwise firms would have continued to avoid tax as before the pressure began.
In column (2) we reestimate the model using firm fixed effects in place
of industry fixed effects. We include both year and firm fixed effects in
this specification to verify that time-invariant firm attributes do not drive
our results. Note that this precludes the inclusion of Incomplete Subs List, a
time-invariant firm variable, and %Havens, which is also time-invariant and
is measured for the initial period in which both compliant and noncompliant FTSE 100 firms presumably disclosed the full list of their subsidiaries,
as compiled in ActionAid’s 2011 subsidiary report. In this specification, the
coefficient on Incomplete Subs List × Post Pressure (β 3 ) is 0.027, (p-value =
0.06), 1.1% of PI smaller than the estimate in column (1), suggesting the
importance of controlling for firm-specific heterogeneity in ETRs. Here,
the 0.027 coefficient estimate suggests that public pressure related to subsidiary disclosure is associated with ETRs that are 2.7% of PI (£34.34 million
of tax expense) greater for noncompliant firms compared to control firms.
The results of these tests suggest that the public scrutiny related to subsidiary disclosures reduced the corporate tax avoidance of noncompliant
firms compared to compliant firms. The relative increase in ETRs is both
statistically significant and economically meaningful. To put these estimates
into perspective, consider the case of Starbucks, U.K., which was purportedly willing to voluntarily pay £10 million in cash taxes to avoid negative
publicity (Pfanner [2012]). Compared to the Starbucks case, which involves
a non-U.K., non-FTSE 100 firm that is smaller than many of the firms in
our sample, our estimate of an average effect between £34.34 million and
£48.34 million in additional tax expense appears economically significant,
but reasonable. Further, these estimates are in line with ActionAid’s estimates of SABMiller’s alleged gains from tax avoidance. ActionAid estimates
that SABMiller avoids about £20 million in tax, an amount on par with
our estimates of the effect of public pressure on tax avoidance (Lawrence
[2010]).
4.2.2. Parallel Trends Assumption and Robustness of Model (1). Differencein-differences estimation requires that treatment and control firms exhibit parallel trends in the outcome variable in the period prior to the
treatment (Roberts and Whited [2012]). We conduct a series of tests to
28 The mean of PI in our sample is £1,271.98 million, and we use this figure to calculate
the £48.34 million (£1,271.98∗.038). The median value of PI is £566 million, generating an
average total tax expense increase of £21.5 million (£21.5∗0.038). To obtain an estimate in
per-share units, we multiply the coefficient estimate, 0.038, times PI, and divide this number
of shares on a firm by firm basis. The mean (median) value of this number is £0.067 (£0.019),
a modest, but nontrivial per-share impact of the ETR change.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
167
examine whether the parallel trends assumption is reasonable. First, we
modify model (1) by including separate year indicators to assess whether
noncompliant firms have significantly different ETRs in the periods preceding the public pressure activity. That is, we estimate the following model
using individual year indicator and interaction variables:
ETR it = βi + β1 Incomplete Subs List i + β2 Yr 2007t + β3 Yr 2008t
+ β4 Yr 2009t + β5 Yr 2010t + β6 Yr 2011t + β7 Yr 2012t
+ β8 Incomplete Subs List i ∗ Yr 2007t + β9 Incomplete Subs i ∗ Yr 2008t
+ β10 Incomplete Subs i ∗ Yr 2009t + β11 Incomplete Subs i ∗ Yr 2010t
+ β12 Incomplete Subs i ∗ Yr 2011t + β13 Incomplete Subs i ∗ Yr 2012t
+
βc C + it ,
(2)
c
where ETR and Incomplete Subs List are as defined above and in table 1.29
In model (2), the interactions of Incomplete Subs List and each of the postpressure years (i.e., β 11 , β 12 , and β 13 , respectively) comprise the variables
of interest, which we expect to be significantly positive if the public scrutiny
of FTSE 100 firms has a persistent effect on the tax avoidance of noncompliant firms relative to compliant firms. Model (2) also allows us to assess
whether there appears to be a significant difference in the ETRs of compliant and noncompliant firms in the years prior to the public scrutiny. If
the parallel trends assumption is satisfied, we expect the coefficients β 8 , β 9 ,
and β 10 to be insignificant.
Table 5 reports the results of our estimates of model (2). The regression results indicate that the coefficients in the two years leading up to the
public pressure are relatively small (0.018 and 0.007) and not statistically
significant (p-value > 0.10). In the three years following public pressure,
the coefficients more than double in magnitude and are generally statistically significant. We find weak evidence of higher ETRs among noncompliant firms in the period three years before the public scrutiny (i.e., in
2007, coefficient = −0.049). However, this difference does not persist. If
anything, the coefficients suggest that the trend in the ETR difference is
moving in the opposite direction in the preperiod compared to the treatment period, which mitigates concerns that some other factor leads to the
relatively greater ETRs of noncompliant firms compared to control firms
in the postpressure period.30
Figure 3, panel A plots the coefficients on the interaction terms for
2007 through 2012 from model (2). The figure shows that there is a visual
29 Model (2) includes year fixed effects for all years in the sample, but we tabulate 2007
through 2012.
30 In additional analysis reported in table A5 in the Online Appendix, we extend the analysis back to 1998 and find no evidence of persistently different ETRs between treatment and
control firms until the postpressure period.
168
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
TABLE 5
Public Pressure and Corporate Tax Avoidance: Analysis by Year
(1)
(2)
ETR
Pred.
Incomplete Subs List
Yr2007
Yr2008
Yr2009
Yr2010
Yr2011
Yr2012
Incomplete Subs List × 2007
Incomplete Subs List × 2008
Incomplete Subs List × 2009
Incomplete Subs List × 2010
Incomplete Subs List × 2011
Incomplete Subs List × 2012
Intercept
Control variables
Industry fixed effects
Year fixed effects
Firm fixed effects
Bootstrapped Standard errors
Observations
Adj. R 2
+
+
+
ETR
Coef.
(p-Value)
−0.007
−0.049∗∗
−0.022
0.001
−0.050∗∗
−0.033
−0.028
0.043∗
0.018
0.007
0.042∗∗
0.039†
0.054∗
−0.221∗∗∗
Yes
Yes
Yes
No
Yes
1,010
0.215
(0.608)
(0.032)
(0.382)
(0.977)
(0.018)
(0.259)
(0.285)
(0.091)
(0.455)
(0.821)
(0.043)
(0.190)
(0.061)
(0.000)
Coef.
(p-Value)
−0.057∗∗
−0.028
−0.013
−0.062∗∗
−0.040
−0.033
0.043∗
0.009
0.005
0.038∗
0.022
0.038†
−0.296∗∗∗
Yes
No
Yes
Yes
Yes
1,010
0.307
(0.039)
(0.334)
(0.696)
(0.012)
(0.179)
(0.194)
(0.094)
(0.727)
(0.885)
(0.066)
(0.397)
(0.120)
(0.000)
This table reports the results of tests examining the effect of public pressure on corporate tax behavior.
ETR is the ETR (txt/pi), bound between [0,1] less the top corporate statutory tax rate for the United Kingdom in the given year. Incomplete Subs List is an indicator variable equal to one for FTSE 100 firms that did
not disclose the full list of their subsidiaries prior to the 2010 public pressure to disclose and equal to zero,
otherwise. Table 1 defines and lists all other control variables (Size, Leverage, Intangibles, Inventory Intensity,
RD Intensity, Capital Intensityt -1 , Capex, Return on Assets, and % Havens). We winsorize continuous variables
at the 1st and 99th percentiles. We compute p-values based on bootstrapped standard errors. We bold lines
related to coefficients of interest. The superscript asterisks ∗∗∗ , ∗∗ , and ∗ denote statistical significance at
the 1%, 5%, and 10% levels, respectively, using two-tailed tests. † denotes statistical significance at the 10%
level using a one-tailed test when we have a prediction and the sign of the coefficient is consistent with the
prediction.
downward trend in the preperiod, although, as noted above, the data-points
that comprise the line in the prepressure period are statistically indistinguishable from zero in 2008 and 2009, and is only marginally significant
in 2007. Thus, while there is statistically no difference between treatment
and control firms in the prepressure period, there is a statistically positive
difference between treatment and control firms in the postpressure period.
Figure 2, panel B plots similar information, but instead of plotting the difference in coefficients between treatment and control firms, it plots each
group separately. As can be seen in the figure, the ETRs of the noncompliant and compliant firms converge in 2009 and then separate abruptly
in 2010 and later. We observe a drop in ETRs of compliant firms in the
postperiod, which is consistent with the broad trend of U.K. firms’ ETRs
decreasing in 2010 (results untabulated).
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
169
We also conduct a propensity score analysis in which we match each noncompliant firm to a control firm based on haven exposure and the control
variables in the primary analysis (e.g., Size, Profitability, Leverage, Intangibles,
Inventory Intensity, R&D Intensity, Capital Intensity, and Capex Size) in the
preperiod (2009). We allow any U.K. firm with sufficient data to be a potential match firm (not just firms from the FTSE 100). Because, by definition, we cannot observe true haven exposure of noncompliant firms in the
preperiod, for this analysis we obtain proprietary data on firm subsidiary
locations from LexusNexus.31 We supplement this tax haven exposure data
using the annual returns and annual reports assembled in the initial ActionAid data collection effort, which should reflect the initial haven exposure
once firms begin to comply with the disclosure requirements (although
these data are only available for FTSE 100 firms). We label this variable
%Havens1. We also match firms on ETRs in 2007, the year in which the
primary results suggest a marginally significant difference in ETRs between
the compliant and noncompliant firms. We require firms to have all model
data in 2009 and 2010 and nonmissing ETR in 2007–2010.
In table 6, panel A we tabulate the results of the propensity-score selection model and report the results of tests of differences in means
(paired t-test), medians (Wilcoxon signed-rank test), and distributions
(Kolmogorov–Smirnov test) to assess covariate balance for the propensity
score model. We find no differences between treatment and control firms
other than for %Havens1, which we include in the ETR regression model
along with the other tax avoidance control variables. In table 6, panel B we
estimate model (2) using the propensity score matched sample. We find no
significant difference in ETR between treatment and control firms in any
of the three prepressure years, but generally significant positive differences
in each of the postpressure years (p-values < 0.10 for all three years).32
In figure 3, panel C we plot the coefficients on the interaction terms
for 2007 through 2012 for model (2) using the propensity score sample.
The figure shows no difference between treatment and control firms in
the prepressure period, consistent with the trends of the two subsamples
31 The LexusNexus data are imperfect, as LexusNexus itself likely has to rely upon firm
disclosures for much of its information. In discussions with LexusNexus staff, they state that
they do not rely entirely upon firm disclosures for their information, but are unwilling to fully
disclosure where their information comes from.
32 In a number of untabulated tests using alternative control samples, we continue to find
qualitatively similar results to the primary analysis and only very rarely do we fail to find results
similar to the primary tests. We find that the coefficient on Incomplete Subs List × Post Pressure remains positive and significant (p-value < 0.10 for the primary specification without firm fixed
effects and p-value < 0.05 for the primary specification with firm fixed effects) using alternative samples based on different propensity-score matching models (e.g., omitting percentage
of havens or 2007 ETRs), but whether the treatment and control firms exhibit similar ETR
trends in the preperiod is sensitive to the propensity score model specification. Accordingly,
we tabulate the results of the sample for which the treatment and control firms exhibit the
most similar trends in the prepressure period.
FIG. 3.—Public pressure and corporate tax avoidance trends. The figure is a plot of the difference-in-differences coefficients from model (2) as well as the
separate ETRs from the model for the noncompliant (Incomplete Subs List = 1) and compliant (Incomplete Subs List = 0) firms. Panel A graphs the
coefficients obtained from the FTSE 100 firm sample (table 5), while panel C graphs the similar coefficients from the propensity-score matched sample
(table 6, panel B). Panel B graphs the fitted values of ETRs from table 5 for the primary sample while panel D does the same, but, using coefficients derived
from the propensity score matched sample (table 6, panel B).
170
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
171
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
TABLE 6
Public Pressure and Corporate Tax Avoidance: Propensity-Score Matched Sample Analysis
Panel A: Selection model
(1)
Incomplete Subs List
Size
Leverage
Intangibles
Inventory Intensity
RD Intensity
Capital Intensityt-1
Capex
Return on Assets
% Havens1
ETR07
Intercept
Coef.
(p-Value)
0.698∗∗∗
0.712
1.801
−1.016
3.807
0.951
0.931
8.218∗∗
−2.848
2.734
−8.457∗∗∗
(0.000)
(0.608)
(0.125)
(0.722)
(0.685)
(0.440)
(0.850)
(0.015)
(0.270)
(0.225)
(0.000)
Observations
Pseudo-R 2
Area under ROC curve
331
0.274
0.899
Incomplete Subs
Firms (N = 35)
Control Firms
(N = 35)
Test of Differences
(p-Values)
Covariate Balance
Mean
Median
Mean
Median
t-Test
Signed Rank
Distr.
Size
Leverage
Intangibles
Inventory Intensity
RD Intensity
Capital Intensityt -1
Capex
Return on Assets
% Havens1
ETR07
8.545
0.217
0.258
0.059
0.012
0.309
0.059
0.095
0.079
−0.016
8.427
0.228
0.241
0.039
0.000
0.252
0.048
0.097
0.059
−0.016
8.638
0.258
0.245
0.055
0.006
0.311
0.066
0.085
0.063
0.000
8.053
0.217
0.206
0.019
0.000
0.213
0.056
0.074
0.000
−0.020
0.740
0.280
0.799
0.767
0.341
0.974
0.619
0.524
0.000
0.573
0.873
0.333
0.860
0.325
0.478
0.822
0.987
0.393
0.000
0.712
0.486
0.486
0.486
0.683
0.486
0.976
0.976
0.198
0.683
0.320
Panel B: Year-by-year specification
Incomplete Subs List
Yr2007
Yr2008
Yr2009
Yr2010
Yr2011
Yr2012
Incomplete Subs List × 2007
Incomplete Subs List × 2008
Incomplete Subs List × 2009
Incomplete Subs List × 2010
Incomplete Subs List × 2011
(1)
(2)
ETR
ETR
Pred.
Coef.
(p-Value)
Coef.
(p-Value)
+
+
0.013
0.004
−0.003
0.000
−0.066∗∗∗
−0.047∗
−0.048∗∗
0.004
−0.006
0.001
0.054∗∗
0.050∗
(0.243)
(0.886)
(0.896)
(0.996)
(0.004)
(0.067)
(0.025)
(0.873)
(0.770)
(0.971)
(0.020)
(0.053)
0.007
0.000
0.000
−0.057∗∗
−0.041
−0.037∗
−0.004
−0.012
−0.006
0.033†
0.030
(0.791)
(0.994)
(0.989)
(0.016)
(0.106)
(0.078)
(0.872)
(0.573)
(0.861)
(0.149)
(0.229)
(Continued)
172
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
T A B L E 6—Continued
Panel B: Year-by-year specification
(1)
(2)
ETR
Pred.
Incomplete Subs List × 2012
Intercept
Control variables
Industry fixed effects
Year fixed effects
Firm fixed effects
Bootstrapped standard errors
Observations
Adj. R 2
+
Coef.
∗∗∗
0.072
−0.211∗∗∗
ETR
(p-Value)
(0.001)
(0.000)
Coef.
∗∗
0.050
−0.249∗∗∗
Yes
Yes
Yes
No
Yes
Yes
No
Yes
Yes
Yes
886
0.312
886
0.353
(p-Value)
(0.026)
(0.000)
Panel A of this table reports the results of the logit selection model we use to identify propensity score
matched control firms for the noncompliant (treatment) firms in the analysis. ETR07 is the firm’s 2007
ETR (txt/pi), bound between [0,1] less the top corporate statutory tax rate for the United Kingdom in
the given year. Incomplete Subs List is an indicator variable equal to one for FTSE 100 firms that did not
disclose the full list of their subsidiaries prior to the 2010 public pressure to disclose and equal to zero,
otherwise. Table 1 in the paper defines all other variables. We winsorize continuous variables at the 1st and
99th percentiles. We compute p-values based on bootstrapped standard errors. Panel B of this table reports
the results of tests examining the effect of public pressure on corporate tax behavior using a propensity
score-matched sample discussed in section 4.2.2 of the paper. ETR is the ETR (txt/pi), bound between
[0,1] less the top corporate statutory tax rate for the United Kingdom in the given year. Incomplete Subs List
is an indicator variable equal to one for FTSE 100 firms that did not disclose the full list of their subsidiaries
prior to the 2010 public pressure to disclose and equal to zero, otherwise. Post Pressure is an indicator variable
equal to one for firm-year observations ending during 2010 (i.e., following the beginning of the increase in
public pressure) or later and equal to zero, otherwise. Table 1 defines and lists all other control variables
(Size, Leverage, Intangibles, Inventory Intensity, RD Intensity, Capital Intensityt -1 , Capex, Return on Assets, and %
Havens). Table 1 in the paper defines all other variables. We winsorize continuous variables at the 1st and
99th percentiles. We compute p-values based on bootstrapped standard errors. We bold lines related to
coefficients of interest. In the panel B, column 1, one or more of the parameters may not be estimated in
seven of the bootstrap iterations, and we estimate standard errors using 993 iterations rather than 1,000. In
both panels, the superscripts asterisks ∗∗∗ , ∗ ∗ , and ∗ denote statistical significance at the 1%, 5%, and 10%
levels, respectively, using two-tailed tests. † denotes statistical significance at the 10% level using a one-tailed
test when we have a prediction and the sign of the coefficient is consistent with the prediction.
being parallel in the prepressure period. The panel also shows the significant difference between treatment and control firms in the postpressure
period. Figure 3, panel D plots similar information, but instead of plotting
the difference in trends between treatment and control firms, it plots each
group separately. As can be seen in the figure, the trends are similar in the
preperiod, but are quite different in the postperiod. Consistent with our
expectations, noncompliant firms exhibit substantially less tax avoidance
than compliant firms in the postpressure period. The figures highlight that
the difference between the treatment and control firms appears to stem
from compliant firm ETRs declining more than noncompliant firm ETRs.
One advantage of the difference-in-differences analysis is that it allows us to
difference out trends that influence treatment and control firms generally
in the pre- and postperiods. Although we focus on the difference, we note
that, in the absence of concurrent changes in the environment, one would
expect noncompliant (compliant) firms to exhibit increasing (flat) ETRs.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
4.3.
173
REAL CHANGES TO SUBSIDIARY LOCATIONS
In the previous subsections, we show that the public scrutiny applied by
ActionAid was sufficient to affect disclosure and to reduce corporate tax
avoidance behavior. A related question is whether firms make real changes
to their corporate structure by eliminating subsidiaries located in tax haven
jurisdictions. To investigate this possibility, we examine whether noncompliant firms reduce their use of tax haven subsidiaries relative to compliant firms in the year following the public pressure. Different from previous tests that investigate changes in disclosure, in this test, we assume that
the disclosures in 2011 and 2012 were complete, and that any observed
changes are due to actual changes in subsidiary locations. Specifically, we
estimate the following cross-sectional difference-in-differences model for
t = 2012:
Havens i = βi + β1 Subs i + β2 Incomplete Subs List i
+ β3 Incomplete Subs List i ∗ Subs i +
βc Ci + i ,
(3)
c
where Havens is the change in tax haven subsidiaries (as defined by
ActionAid) from 2011 to 2012, and Subs is the change in total subsidiaries
over the same time period.33 C is measured in fiscal year 2010. β 3 represents the extent to which the association between the change in haven subsidiaries to the change in total subsidiaries is different for firms that did
not initially disclose their complete subsidiary list. We expect increases in
haven subsidiaries to be smaller for a given increase in total subsidiaries for
Incomplete Subs List firms. Therefore, we expect β 3 to be negative.
Table 7 reports the results from estimating model (3).34 We report four
estimations of the model: (1) with no controls and no industry fixed effects,
(2) with no controls but including industry fixed effects, (3) including controls but no industry fixed effects, and (4) including controls and industry
fixed effects. Regardless of which model we estimate, we find that the coefficient on Incomplete Subs List × Subs is negative and generally statistically
significant. The coefficient magnitudes are also largely insensitive to the
inclusion of control variables.35
33 ActionAid undertook the effort to hand-collect the complete subsidiary lists from FTSE
100 firms’ individual Annual Reports and Annual Returns for both the period immediately
following ActionAid’s investigation and noncompliant firms’ purported subsequent disclosure
compliance, releasing the data in 2011, as well as the subsidiary disclosure period for the
following year, covering 2012. We thank ActionAid for providing the data.
34 Note that we estimate model (3) with a single cross-section of firms, and thus we use at
most 89 observations, corresponding to the firms for which ActionAid collected and publicized data in its initial subsidiary report (in 2011) and its subsequent report (in 2012); in the
subsequent report ActionAid did not include all firms included in the initial report, given
changes in FTSE 100 membership, etc. The tax avoidance analyses use the firms from the
ActionAid investigation that have data for the ETR tests (77 firms).
35 In table A8 of the Online Appendix, we also report the results of an additional test in
which we regress the percentage change in haven subsidiaries, defined as the change in haven
174
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
TABLE 7
Public Pressure and Changes in Tax Haven Subsidiaries
(1)
(2)
(3)
(4)
Coef.
(p-Value)
Havens
Coef.
(p-Value)
Havens
Coef.
(p-Value)
Havens
Coef.
(p-Value)
Havens
0.187∗∗∗
(0.000)
−2.097
(0.390)
−0.098†
(0.119)
3.526
(0.125)
0.189∗∗∗
(0.000)
−3.608
(0.188)
−0.115∗
(0.057)
−2.378
(0.599)
0.208∗∗∗
(0.000)
−0.386
(0.871)
−0.115∗
(0.055)
−17.945
(0.279)
0.216∗∗∗
(0.000)
−2.866
(0.413)
−0.130∗
(0.054)
−22.541
(0.392)
Control variables
Industry fixed effects
Year fixed effects
Firm fixed effects
Bootstrapped standard errors
No
No
N/A
N/A
Yes
No
Yes
N/A
N/A
Yes
Yes
Yes
N/A
N/A
Yes
Yes
Yes
N/A
N/A
Yes
Observations
Adj. R 2
89
0.461
89
0.417
81
0.446
81
0.379
Pred.
Subs
Incomplete Subs List
Incomplete Subs List × ΔSubs
Intercept
−
This analysis examines changes in subsidiaries in the period following the public pressure related to subsidiary disclosure for the FTSE 100 firms for which ActionAid publicly disclosed the data—one observation
per firm. Havens is the total change in subsidiaries in tax havens for the period, as defined by ActionAid.
Subs is total change in subsidiaries for the period. Incomplete Subs List is an indicator variable equal to one
for FTSE 100 firms that did not disclose the full list of their subsidiaries prior to the 2010 public pressure
to disclose and equal to zero, otherwise. Table 1 defines all other control variables. Given the limited sample size, we use Fama and French 12-industry classifications in this analysis. We compute p-values based on
bootstrapped standard errors. In column 2 (column 4), one or more of the parameters may not be estimated in 437 (404) of the bootstrap iterations, and we estimate standard errors using 563 (596) iterations
rather than 1,000. We winsorize continuous variables at the 1st and 99th percentiles. We bold lines related
to coefficients of interest.
∗∗∗ ∗∗
, , and ∗ denote statistical significance at the 1%, 5%, and 10% levels, respectively, using two-tailed
tests.
† denotes statistical significance at the 10% level using a one-tailed test when we have a prediction and
the sign of the coefficient is consistent with the prediction.
4.4.
CROSS-SECTIONAL VARIATION IN THE DIFFERENCE-IN-DIFFERENCES
ESTIMATOR
In this section, we investigate whether the effect of public scrutiny on
corporate tax avoidance varies in predictable ways. In the following three
subsections, we explore the role of undisclosed subsidiaries in tax havens
subsidiaries scaled by the change in total subsidiaries over the period (%Havens), on Incomplete Subs List and controls. As expected, in column 1, we find that the coefficient on Incomplete
Subs List is negative in the model, although only marginally statistically significant (two-tailed pvalue = 0.166). %Havens requires firms to have nonzero changes in total subsidiaries, which
is distinct from Decrease % Havens, defined in section 4.4.1. We find weaker results if we use
changes in the percent of havens instead of %Havens. In column 2, we find that this result is
especially pronounced in changes in “dot” haven nations, but, in column 3, fail to find a result
for the percentage change in Big 7 tax havens. This result is consistent with a relative decrease
in subsidiary use in countries where there is generally little economic reason to establish a
subsidiary (“dot” havens), whereas there was no discernable decrease in larger countries that
also happen to be tax havens (the “Big 7” havens).
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
175
in the preperiod (subsection 4.4.1), of reductions in the use of tax havens
in the postpressure period (subsection 4.4.2), and of politically sensitive
industries (subsection 4.4.3).
4.4.1. The Role of Hidden (Undisclosed) Tax Haven Subsidiaries in the Prepressure Period. As mentioned earlier, in table 3 we document the significant increase in noncompliant firms’ subsidiary disclosures and report the
disproportionate levels of undisclosed tax haven subsidiaries compared to
other subsidiaries in the prepressure period. We extend this analysis in the
Online Appendix and examine whether the effect of public pressure on
corporate tax avoidance is significantly more pronounced for firms that
failed to initially disclose relatively more tax havens (i.e., firms with a large
proportion of “hidden” tax havens, %Hidden Havens). In table A1 in the
Online Appendix we find that the coefficient on the interaction between
Incomplete Subs List × Post Pressure with %Hidden is positive and significant
(p-value < 0.05), suggesting that the public pressure effect on subsequent
tax avoidance is significantly more pronounced for firms that failed to initially disclose relatively more tax havens.
4.4.2. The Role of Tax Havens in the Postpressure Period. As highlighted in
section 4.2, we find that noncompliant firms reduce tax avoidance (higher
ETRs) after ActionAid publicly scrutinized their lack of compliance with
subsidiary disclosure laws. Moreover, the evidence presented in section 4.3
suggests that noncompliant firms changed their subsidiary locations to reduce their tax haven footprint. If dissolving subsidiaries in tax haven countries is an important factor in explaining the decrease in tax avoidance,
we should observe that the increase in ETRs is greater in firms that reduced their presence in tax havens. “Dot” haven subsidiaries are located
in countries with few operating advantages and are generally considered to
be established primarily for tax planning purpose (e.g., Desai, Foley, and
Hines [2006]). To probe this further, we split our sample based on whether
firms decrease the percentage of “dot” tax havens in the postpressure period (Decrease % Dot Havens = 1) or do not (Decrease % Dot Havens = 0)
and estimate model (1) on each subsample. We use the Chow [1960] test
to assess differences in tax avoidance outcomes in the postpressure period
for noncompliant firms relative to compliant firms.36 We report the results
in table 8. Consistent with subsidiary location decisions being associated
with tax avoidance outcomes, the coefficient on Incomplete Subs List × Post
Pressure is only significant for the observations experiencing a decrease in
36 We measure the percentage of havens as a ratio of total subsidiaries in each of the two
ActionAid subsidiary lists (i.e., the initial list released in 2011 and the subsequent list released
in 2012). We split firms into groups based on whether the percentage of dot haven subsidiaries
firms in the subsequent report is lower than the percentage in the first report (Decrease % Dot
Havens). Prior work often defines “dot” havens as those in tax haven countries that are not part
of the “Big 7” tax havens known for operational and tax benefits (e.g., Desai, Foley, and Hines
[2006] Given clear operating benefits as well as tax benefits, we also exclude Netherlands and
Delaware (i.e., the United States) from the “dot” haven designation.
176
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
TABLE 8
Corporate Tax Avoidance Analyses by Changes in Tax Haven Subsidiaries
(1)
(2)
Decrease % Dot Havens =
0
1
Pred.
Post Pressure
Incomplete Subs List × Post Pressure
Intercept
Control variables
Industry fixed effects
Year fixed effects
Firm fixed effects
Cluster by firm
Observations
Adj. R 2
Differences between coefficients on
Incomplete Subs List × Post Pressure:
+
Coef.
(p-Value)
ETR
Coef.
(p-Value)
ETR
0.003
(0.923)
0.006
(0.738)
−0.285∗∗∗
(0.006)
−0.112∗∗
(0.021)
0.057∗∗
(0.025)
−0.440∗∗∗
(0.003)
Yes
No
Yes
Yes
No
Yes
No
Yes
Yes
No
506
0.323
435
0.257
(2) − (1):
0.051∗
This analysis examines the effect of public pressure related to subsidiary disclosure on tax avoidance
for FTSE 100 firms that were initially compliant with disclosure requirements (i.e., not subject to public
pressure related to subsidiary disclosure) relative to those that were not. In these analyses we partition the
sample based on firms’ subsidiary changes in the initial year following the public pressure (i.e., from 2010
to 2011). Decrease %Dot Havens is equal to one if the firm reports a decrease in the percentage of “Dot” tax
haven subsidiaries in the postpressure period and is equal to zero, otherwise. “Dot” haven subsidiaries are
located in countries with few operating advantages and are generally considered to be established primarily
for tax planning purpose (e.g., Desai, Foley, and Hines [2006]). ETR is the ETR (txt/pi) bound between
[0,1] less the top corporate statutory tax rate for the United Kingdom in the given year. Incomplete Subs List
is an indicator variable equal to one for FTSE 100 firms that did not disclose the full list of their subsidiaries
prior to the 2010 public pressure to disclose and equal to zero, otherwise. Post Pressure is an indicator variable
equal to one for firm-year observations ending during 2010 (i.e., following the beginning of the increase
in public pressure) or later and equal to zero, otherwise. Table 1 defines all other variables. We include all
control variables used in the primary ETR analyses (e.g., the specifications with firm fixed effects in tables 5
and 6, column (2)). We winsorize continuous variables at the 1st and 99th percentiles. We compute p-values
based on bootstrapped standard errors. We bold lines related to coefficients of interest.
∗∗∗ ∗∗
, , and ∗ denote statistical significance at the 1%, 5%, and 10% levels, respectively, using two-sided
tests.
the percentage of their “dot” tax haven subsidiaries and the 0.051 difference in the coefficients on Incomplete Subs List × Post Pressure across the two
subsamples is significant (p-value < 0.05). The difference in coefficients for
the two samples suggests that the public pressure effect for firms decreasing their use of subsidiaries located in “dot” tax havens is associated with
£64.87 million more in average tax expense compared to other firms.
To shed further light on the mechanism linking public pressure to corporate tax rates, we examine the association between decreased use of “dot”
tax havens and changes in ETRs. We regress changes in three-year ETRs
on the Incomplete Subs List indicator, Decrease % Dot Havens, and the interaction between Incomplete Subs List and Decrease % Dot Havens. In table A6,
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
177
panel A in the Online Appendix, we report that the coefficient on Incomplete Subs List × Decrease % Dot Havens is positive and significant, indicating
that noncompliant firms that decrease their use of dot havens have a significantly greater decrease in tax avoidance (increase in ETRs) relative to
other firms.
4.4.3. Politically Sensitive Industries. As noted earlier, political costs
wherein the government imposes penalties on the firm either directly
(e.g., increased audits, changing tax laws) or indirectly (e.g., fewer awarded
contracts) are one potential cost of the public scrutiny initiated by ActionAid. If political costs are salient, we would expect firms’ political sensitivity
to affect the impact of public pressure on corporate tax avoidance. We
use the politically sensitive industry designations from Julio and Yook
[2012] to construct two subsamples based on whether the firm is in a
politically sensitive industry (i.e., tobacco products, pharmaceuticals,
health care, defense, petroleum and natural gas, telecommunications,
and transportation) or not. Then, we reestimate model (1) on the two
subsamples and use the Chow [1960] test to evaluate differences in the
difference-in-difference estimators (β 3 ) across the two subsamples. These
results are tabulated in the Online Appendix, table A2. The coefficient on
Incomplete Subs List × Post Pressure is positive and significant for the more
politically sensitive subsample (two-tailed p-value < 0.05), but insignificant
for the less politically sensitive subsample. Moreover, the effect of public
pressure on ETRs is marginally more pronounced among firms that are
in more politically sensitive industries (two-tailed p-value < 0.20) and the
difference in coefficients for the two samples suggests that the public pressure effect for politically sensitive firms is associated with £95.40 million
more in average tax expense compared to other firms. This greater tax
expense is consistent with political costs representing an important factor
influencing how firms respond to public pressure related to corporate tax
behavior.37
5. Additional Tests
5.1.
RETURNS ANALYSIS AROUND ACTIONAID’S REPORT ON TAX HAVEN
USAGE
A key assumption in this study is that the noncompliant FTSE 100 firms
consciously chose to withhold their full list of foreign subsidiaries in 2006
37 We also examine whether noncompliant firms were significantly more likely to have membership in an industry that traditionally benefits from government contracting, as defined in
Mills et al. [2013]. We estimate a linear probability model examining whether firms in industries commonly associated with government contracting are significantly more likely to be
noncompliant firms in our sample, controlling for the other variables in model (1). We find
that membership in these industries is positively associated with noncompliant subsidiary disclosure (p-value < 0. 10). For more details on this test, and for a table that reports the results
of the estimated linear probability model, see the Online Appendix, table A3.
178
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
Average Three-Day Buy-and-Hold
Return centered on ActionAid
Report
Non-FTSE 100 Firms
Complete Subs List
FTSE 100 Firms
Incomplete Subs List
FTSE 100 Firms
0
-0.002
-0.004
-0.006
-0.008
-0.01
-0.012
-0.014
-0.016
-0.018
-0.02
FIG. 4.—Buy-and-hold returns surrounding ActionAid report on tax haven usage. This graph
depicts the three-day buy and hold returns, centered on October 11, 2011, the date on which
ActionAid released a well-publicized report on FTSE 100 firms’ use of tax havens. The graph
depicts three groups of firms: (1) all U.K. firms not in the FTSE 100 (all firms on Compustat
Global daily security file (comp.g secd) with nonzero or missing returns with LOC = “GBR” or
FIC = “GBR”), (2) all FTSE 100 firms that ActionAid found to have been in compliance with
the subsidiary disclosure requirements, and (3) all FTSE 100 firms that ActionAid determined
were not initially disclosing their full subsidiary list.
because they faced a high cost of disclosure. In this section, we document
evidence consistent with nondisclosing firms being more sensitive to public
pressure than other firms. On October 11, 2011, ActionAid released a report titled “Addicted to Tax Havens,” which received widespread attention
and was the subject of general interest, many popular press articles, and
parliamentary discussion. Choy, Lai, and Ng [2014] document significant
negative returns for FTSE 100 firms on this date, consistent with market
participants anticipating negative consequences from the unwanted publicity. We examine whether the negative returns for FTSE 100 firms around
the October 11, 2011 event date are especially concentrated in firms that
did not disclose their full subsidiary list in 2006.
We calculate the three-day buy-and-hold returns for all publicly traded
U.K. firms available on Compustat Global from October 10 to October 12,
2011. Figure 4 graphs the averages of these returns for three groups of
firms: (1) non-FTSE 100 firms, (2) FTSE 100 firms that initially disclosed
their full subsidiary list, and (3) FTSE 100 firms that did not initially disclose
their full subsidiary list. The returns of the nondisclosing FTSE 100 firms
range from 63 to 83 basis points lower than the other two groups of firms.
In table 9, column (1) we tabulate regressions of the buy-and-hold return
on Incomplete Subs List, an indicator coded to equal one for FTSE 100 firms
that did not initially disclose their full subsidiary list. Including all U.K.
firms in the regression implies that the intercept term captures the market
return for FTSE 100 firms that were initially compliant as well as all other
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
179
TABLE 9
Buy-and-Hold Returns Surrounding ActionAid Report on Tax Haven Usage
(1)
(2)
Three-Day Buy-and-Hold Return
Centered on ActionAid Report Date
Prediction
Incomplete Subs List
–
Intercept
All U.K. and FTSE 100 firms
Only FTSE 100 firms
Observations
Adj. R 2
Coef.
(p-Value)
Coef.
(p-Value)
−0.006∗∗
(0.042)
−0.010∗∗∗
(0.000)
−0.008∗
(0.096)
−0.008∗∗
(0.047)
Yes
No
No
Yes
1,520
0.000
98
0.018
The dependent variable is the Three-Day Buy-and-Hold Return Centered on ActionAid Report Date, October
11, 2011. This report is the focus of Choy, Lai, and Ng [2014]. Incomplete Subs List is an indicator variable
equal to one for FTSE 100 firms that did not disclose the full list of their subsidiaries prior to the 2010
public pressure related to subsidiary disclosure and equal to zero, otherwise. We compute p-values based on
bootstrapped standard errors.
∗∗∗ ∗ ∗
, , and ∗ denote statistical significance at the 1%, 5%, and 10% levels, respectively, using two-tailed
tests.
non-FTSE 100 firms. Nondisclosing FTSE 100 firms had a return that was
0.6 percentage points lower than other firms in the market, which, when
taking into account the market cap of Incomplete Subs List firms, represents
a median effect of £27.22 million. The market may well be impounding
the estimated extra taxes paid, which we estimate to be between £34.34
million and £48.34 million. As reported in column (2), we observe a similar result when we estimate the model only for FTSE 100 firms. Here, we
find that noncompliant firms had a return that was 0.8 percentage points
lower than compliant FTSE 100 firms. FTSE 100 firms that did not initially
comply with the disclosure requirement experienced significant negative
market returns relative to other FTSE 100 firms in the window surrounding
widespread attention on FTSE 100 firm’s tax haven usage. These results are
consistent with our expectation of noncompliant firms having higher costs
of public scrutiny.
5.2.
THE ROLE OF MEDIA COVERAGE
Prior papers have documented that media coverage of corporate behavior can change firm outcomes (Dyck, Volchkova, and Zingales [2008]).
One mechanism through which ActionAid seeks to publicize corporate tax
avoidance and haven use is through media coverage. We conduct a number of tests to evaluate the effect of tax-related media coverage on the results. First, we use Factiva to search for all press articles in the U.K.-related
to tax avoidance for the FTSE 100 firms during our sample period. We
use the search term “tax w/10 (avoid∗ or shelter or dodg∗ or haven).” We
then count the number of articles returned with this search term for each
180
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
firm-year in our sample. We use these data to assess how tax-related media
coverage affects the results. The results of our analyses, tabulated in our
Online Appendix, table A4, suggest that media coverage is possibly one factor contributing to how firms responded to the public pressure in terms of
their corporate tax behavior, though the results are somewhat mixed. First,
we find that the results of our primary analysis persist after controlling for
tax-related media coverage. Second, we find that, when we split the sample
into groups of firms experiencing an increase in tax-related media coverage
from 2009 to 2010 versus those that did not, the effect of public pressure
on subsequent tax avoidance is greater (i.e., the ETR response is larger
in magnitude) for firms experiencing an increase in tax-related media
coverage, albeit the difference is not statistically significant. We interpret
these results as being consistent with tax-related media coverage enhancing, but not entirely driving, the effect of public pressure on corporate tax
avoidance.
5.3.
OTHER ROBUSTNESS TESTS
In this section, we discuss the results of several robustness tests that
we tabulate in the Online Appendix. First, table A6, panel B in the Online Appendix reports the results of a changes specification regression in
which we examine differences in three-year ETRs between the postpressure
(2010–2012) and prepressure (2007–2009) periods for noncompliant versus compliant firms. We regress this change in three-year ETR (ETR3)
on Incomplete Subs List, the three-year changes in our model (1) control
variables over the same time period and %Havens from model (1). We
find that the coefficient on Incomplete Subs List is positive and significant
(p-value < 0.10), mitigating concerns that the results do not only obtain in
fixed effects or difference-in-differences regressions. Second, we reestimate
model (1) on a sample limited to FTSE 100 firms domiciled in the United
Kingdom Although general scrutiny of FTSE 100 firms is likely associated
with significantly higher ETRs for noncompliant firms relative to compliant
firms in the postpressure period, the specific provisions requiring U.K. publicly traded firms to report the full list of their subsidiaries (Companies Act
of 2006, Sections 409 and 410) technically applies to U.K.-domiciled firms.
As shown in panel A of table A7, column (4), when we restrict the sample
to only U.K.-domiciled firms, we find that the coefficient on Incomplete Subs
List × Post Pressure is 0.028, and is statistically different from zero (p-value
< 0.10), suggesting that the small number of non-U.K.-domiciled firms in
the FTSE 100 do not drive the results.38 Finally, the Online Appendix also
contains discussion regarding why our results are unlikely to be driven by
the United Kingdom’s move to a territorial tax regime.
38 We also report the results of estimating model (1) on the shorter sample period beginning in 2006 (the year that Parliament enacted the subsidiary disclosure requirements) in
panel A of table A7, column (5) in the Online Appendix.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
181
6. Conclusion
In this paper, we use a shock to the public scrutiny of firm subsidiary
locations to investigate whether that scrutiny leads to changes in firms’ disclosure and corporate tax avoidance behavior. ActionAid International, a
nonprofit, conducted an investigation to identify which FTSE 100 firms
were not complying with rules requiring firms to disclose the full list of
their subsidiaries. ActionAid then petitioned the Companies House of the
United Kingdom to enforce the disclosure rule. We find firms that were
newly required to disclose a full subsidiary list decreased their tax avoidance and use of tax havens relative to firms that already disclosed a full subsidiary list. We find the decrease in tax avoidance for noncompliant firms
in the postpressure period is most pronounced for firms that experience a
decrease in the percentage of total subsidiaries located in tax haven countries. We also document some evidence that suggests that public scrutiny
imposed greater costs on firms in politically sensitive industries, suggesting political costs play a role. The evidence suggests that public pressure
related to subsidiary disclosure can impose significant political and reputational costs on and affect the tax avoidance activities and subsidiary location
decisions of large, publicly traded firms.
A few caveats should be noted. First, we have a relatively small sample,
and our results may not be generalizable to larger populations of firms.
Second, because we have a small sample, we may lack statistical power to
detect some effects. Where we have signed predictions, we note one-tailed
tests of statistical significance in the tables, but the two-tailed tests we also
report are sometimes not significant at conventional levels. Finally, one advantage of the difference-in-differences analysis we employ is that it allows
us to difference out trends that influence treatment and control firms generally in the pre- and postperiods. Although we focus on the difference,
we note that, in the absence of concurrent changes in the environment,
one would expect noncompliant (compliant) firms to exhibit increasing
(flat) ETRs. Instead, we find that noncompliant (compliant) firms have flat
(decreasing) ETRs, suggesting that some environmental changes occurred
during our sample period.
In sum, our findings suggest that activist groups such as ActionAid can
influence corporate outcomes, consistent with other research that shows
that firms respond to pressure from external stakeholders (Smith [1995],
Ertimur, Ferri, and Maber [2012]) and that suggests real consequences to
disclosure policies (e.g., Christensen et al. [2015], Granja [2014]). Our
study is informative to policy debates regarding how much firms should be
required to disclose with regard to the extent and location of their multinational operations. Our paper speaks to potential effects of increased geographic disclosure and should be useful as policy makers consider proposals
to expand disclosure requirements for multinational corporations.
182
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
APPENDIX
ActionAid Letter
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
183
REFERENCES
AUSTIN, C., AND R. WILSON. “Are Reputational Costs a Determinant of Tax Avoidance?” Working paper, University of Oregon, 2015.
BADERTSCHER, B. A.; S. P. KATZ; AND S. O. REGO. “The Separation of Ownership and Control
and Corporate Tax Avoidance.” Journal of Accounting and Economics 56 (2013): 228–50.
BERTRAND, M.; E. DUFLO; AND S. MULLAINATHAN. “How Much Should We Trust Differences-inDifferences Estimates?” The Quarterly Journal of Economics 119 (2004): 249–75.
BOWERS, S. “Amazon to Begin Paying Corporation Tax on UK Retail Sales.”
The Guardian, May 22, 2015. Available online at http://www.theguardian.com/
technology/2015/may/23/amazon-to-begin-paying-corporation-tax-on-uk-retail-sales
BOZANIC, Z.; J. L., HOOPES; J. R. THORNOCK; AND B. WILLIAMS. “IRS Attention.” Working paper,
Ohio State University, 2014.
BROWN, J. L., AND K. D. DRAKE. “Network Ties Among Low-Tax Firms.” The Accounting Review
89 (2013): 483–510.
CHEN, S.; X. CHEN; Q. CHENG; AND T. SHEVLIN. “Are Family Firms More Tax Aggressive than
Non-Family Firms?” Journal of Financial Economics 95 (2010): 41–61.
CHOW, G. C. “Tests of Equality Between Sets of Coefficients in Two Linear Regressions.” Econometrica 28 (1960): 591–605.
CHOY, S. K.; T. LAI; AND T. NG. “Do Treasure Islands Create Firm Value.” Working paper,
Shanghai University of Finance and Economics, 2014.
CHRISTENSEN, D. M.; D. S. DHALIWAL; S. BOIVIE; AND S. D. GRAFFIN. “Top Management Conservatism and Corporate Risk Strategies: Evidence from Managers’ Personal Political Orientation and Corporate Tax Avoidance.” Strategic Management Journal 36 (2014): 1918–38.
CHRISTENSEN, H. B.; E. FLOYD; L. Y. LIU; AND M. G. MAFFETT. “The Real Effects of Mandatory
Non-Financial Disclosures in Financial Statements.” Working paper, University of Chicago,
2015. Available online at http://papers.ssrn.com/sol3/papers.cfm?abstract id=2680296
CREAL, D. D.; L. A. ROBINSON; J. L. ROGERS; AND S. L. C. ZECHMAN. “The Multinational Advantage.” Working paper, Dartmouth College, 2014.
DESAI, M. A.; C. F. FOLEY; AND J. R. HINES JR. “The Demand for Tax Haven Operations.” Journal
of Public Economics 90 (2006): 513–31.
DUHIGG, C., AND D. KOCIENIEWSKI. “How Apple Sidesteps Billions in Taxes.” The
New York Times, April 29, 2012. Available online at http://www.nytimes.com/2012/
04/29/business/apples-tax-strategy-aims-at-low-tax-states-and-nations.html? r=;0
DYCK, A.; A. MORSE; AND L. ZINGALES. “Who Blows the Whistle on Corporate Fraud?” The
Journal of Finance 65 (2010): 2213–53.
DYCK, A.; N. VOLCHKOVA; AND L. ZINGALES. “The Corporate Governance Role of the Media:
Evidence from Russia.” The Journal of Finance 63 (2008): 1093–135.
DYRENG, S. D.; M. HANLON; AND E. L. MAYDEW. “Long-Run Corporate Tax Avoidance.” The
Accounting Review 83 (2008): 61–82.
DYRENG, S. D.; M. HANLON; AND E. L. MAYDEW. “The Effects of Executives on Corporate Tax
Avoidance.” The Accounting Review 85 (2010): 1163–89.
DYRENG, S.; M. HANLON; AND E. L. MAYDEW. “Rolling the Dice: When Does Tax Avoidance
Result in Tax Uncertainty?” Working paper, Duke University, 2014.
DYRENG, S. D., AND B. P. LINDSEY. “Using Financial Accounting Data to Examine the Effect
of Foreign Operations Located in Tax Havens and Other Countries on U.S. Multinational
Firms’ Tax Rates.” Journal of Accounting Research 47 (2009): 1283–316.
EDMANS, A.; M. HEINLE; AND C. HUANG. “The Real Costs of Disclosure.” Working paper, National Bureau of Economic Research, 2013.
ERICKSON, M.; M. HANLON; AND E. L. MAYDEW. “How Much Will Firms Pay for Earnings That
Do Not Exist? Evidence of Taxes Paid on Allegedly Fraudulent Earnings.” The Accounting
Review 79 (2004): 387–408.
ERNST & YOUNG. “Bridging the Divide: Highlights from the 2014 Tax Risk and Controversy Survey, 2014. Available online at http://www.ey.com/Publication/vwLUAssets/
EY-2014-tax-risk-and-controversy-survey-highlights/$FILE/EY-2014-tax-risk-and-controversysurvey-highlights.pdf
184
S. D. DYRENG, J. L. HOOPES, AND J. H. WILDE
ERTIMUR, Y.; F. FERRI; AND D. A. MABER. “Reputation Penalties for Poor Monitoring of Executive Pay: Evidence from Option Backdating.” Journal of Financial Economics 104 (2012):
118–44.
GALLEMORE, J.; E. L. MAYDEW; AND J. R. THORNOCK. “The Reputational Costs of Tax Avoidance.” Contemporary Accounting Research 31 (2014): 1103–33.
GIGLER, F.; C. KANODIA; H. SAPRA; AND R. VENUGOPALAN. “How Frequent Financial Reporting
Can Cause Managerial Short-Termism: An Analysis of the Costs and Benefits of Increasing
Reporting Frequency.” Journal of Accounting Research 52 (2014): 357–87.
GRAHAM, J. R.; M. HANLON; T. SHEVLIN; AND N. SHROFF. “Incentives for Tax Planning and
Avoidance: Evidence from the Field.” The Accounting Review 89 (2014): 991–1023.
GRANJA, J. “Disclosure Regulation in Commercial Banking: Lessons from the
National Banking Era.” Working paper, MIT, 2014. Available online at
http://papers.ssrn.com/sol3/papers.cfm?abstract id=2504783
HANLON, M., AND S. HEITZMAN. “A Review of Tax Research.” Journal of Accounting and Economics
50 (2010): 127–78.
HANLON, M.; J. L. HOOPES; AND N. SHROFF. “The Effect of Tax Authority Monitoring and
Enforcement on Financial Reporting Quality.” The Journal of the American Taxation Association
36 (2014): 137–70.
HANLON, M., AND J. SLEMROD. “What Does Tax Aggressiveness Signal? Evidence from Stock
Price Reactions to News About Tax Shelter Involvement.” Journal of Public Economics 93
(2009): 126–41.
HOLMS, L. “Cable Calls for Fines on Firms Failing to Disclose Tax Havens.” the
Daily Mail, March 5, 2011. Available online at https://www.pwc.com/gx/en/tax/
publications/assets/pwc-tax-strategy-and-corporate-reputation-a-tax-issue-a-business-issue.pdf
HOOPES, J. L.; D. MESCALL; AND J. A. PITTMAN. “Do IRS Audits Deter Corporate Tax Avoidance?” The Accounting Review 87 (2012): 1603–39.
HOPE, O.-K.; M. (SHUAI) MA; AND W. B. THOMAS. “Tax Avoidance and Geographic Earnings
Disclosure.” Journal of Accounting and Economics 56 (2013): 170–89.
JULIO, B., AND Y. YOOK. “Political Uncertainty and Corporate Investment Cycles.” The Journal of
Finance 67 (2012): 45–83.
KANODIA, C. “Accounting Disclosure and Real Effects.” Foundations and Trends in Accounting 1
(2007): 167–258.
KANODIA, C., AND H. SAPRA. “A Real Effects Perspective to Accounting Measurement and
Disclosure: Implications and Insights for Future Research.” Working paper, University of
Chicago, 2015.
LAWRENCE, F. “Brewer Accused of Depriving Poor Countries of Millions in Revenue.”
The Guardian, November 28, 2010. Available online at http://www.theguardian.com/
business/2010/nov/29/sabmiller-india-africa-actionaid-report
LEUZ, C. “Proprietary Versus Non-Proprietary Disclosures: Evidence from Germany,” in The
Economics and Politics of Accounting: International Perspectives on Trends, Policy, and Practice,
edited by D. Pfaff and A. G. Hopwood. Oxford University Press, 2004: 164–97.
MAYLIE, D. “SABMiller Under Scrutiny.” Wall Street Journal, June 17, 2011.
MCGUIRE, S. T.; D. WANG; AND R. J. WILSON. “Dual Class Ownership and Tax Avoidance.” The
Accounting Review 89 (2014): 1487–516.
MILLER, G. S. “The Press as a Watchdog for Accounting Fraud.” Journal of Accounting Research
44 (2006): 1001–33.
MILLS, L. F. “Book-Tax Differences and Internal Revenue Service Adjustments.” Journal of Accounting Research 36 (1998): 343–56.
MILLS, L. F.; S. E. NUTTER; AND C. M. SCHWAB. “The Effect of Political Sensitivity and Bargaining Power on Taxes: Evidence from Federal Contractors.” The Accounting Review 88 (2013):
977–1005.
MINTZ, J., AND B. VENKATACHALAM. “The Problem with the Low-Tax Backlash: Rethinking Corporate Tax Policies to Adjust for Uneven Reputational Risks.” Working paper, University of
Calgary, 2015.
PUBLIC PRESSURE AND CORPORATE TAX BEHAVIOR
185
NEUBIG, T. “Where’s the Applause? Why Most Corporations Prefer a Lower Rate. Tax Notes,
April 24, 2006.
PFANNER, E. “Starbucks Offers to Pay More British Tax Than Required.” The New
York Times, December 7, 2012. Available online at http://www.nytimes.com/
2012/12/07/business/global/07iht-uktax07.html B3
PRUITT, S. W., AND M. FRIEDMAN. “Determining the Effectiveness of Consumer Boycotts: A
Stock Price Analysis of Their Impact on Corporate Targets.” Journal of Consumer Policy 9
(1986): 375–87.
PWC. “Tax Strategy and Corporate Reputation: A Tax Issue.” A Business Issue, 2012. Available online at https://www.pwc.com/gx/en/tax/publications/assets/pwc-tax-strategy-andcorporate-reputation-a-tax-issue-a-business-issue.pdf
REGO, S. O. “Tax-Avoidance Activities of U.S. Multinational Corporations.” Contemporary Accounting Research 20 (2003): 805–33.
REGO, S. O., AND R. WILSON. “Equity Risk Incentives and Corporate Tax Aggressiveness.” Journal of Accounting Research 50 (2012): 775–810.
ROBERTS, M. R., AND T. M. WHITED. “Endogeneity in Empirical Corporate Finance.” Working
paper, University of Michigan, 2012.
SMITH, D. G. “Corporate Governance and Managerial Incompetence: Lessons from Kmart.”
North Carolina Law Review 74 (1995): 1037–140.
TITCOMB, J. “Starbucks to Pay Corporation Tax on Profits in the UK after HQ
Move,” 2014. Available online at http://www.telegraph.co.uk/finance/newsbysector/
retailandconsumer/10769497/Starbucks-to-pay-corporation-tax-on-profits-in-the-UK-afterHQ-move.html.
WESTPHAL, J. D., AND D. L. DEEPHOUSE. “Avoiding Bad Press: Interpersonal Influence in Relations Between CEOs and Journalists and the Consequences for Press Reporting About Firms
and Their Leadership.” Organization Science 22 (2010): 1061–86.
WILDE, J. “The Deterrence Effect of Employee Whistleblowing Allegations on Firms’ Subsequent Aggressive Financial Reporting and Tax Behavior.” Working paper, University of Iowa,
2015.
ZIMMERMAN, J. L. “Taxes and Firm Size.” Journal of Accounting and Economics 5 (1983): 119–49.
186