taxes and corporate accountability reporting: is … · importance of corporate social...
TRANSCRIPT
Taxes and Corporate Accountability Reporting: Is Paying Taxes Viewed as Socially Responsible?
Angela K. Davis, David A. Guenther*, Linda K. Krull, and Brian M. Williams
Lundquist College of Business, University of Oregon, Eugene, OR 97403 USA
January 2013
Abstract: Current corporate accountability (CA) reporting guidelines view a firm’s corporate tax payments as having a positive impact on social welfare. However, a detailed investigation of actual CA reports indicates that, while some firms view paying corporate taxes as socially responsible, many firms argue they enhance social welfare by lobbying for lower corporate taxes, and a third set of firms omit any mention of corporate taxes from their CA reports. This diversity in how corporate taxes are addressed in CA reports raises the question of whether managers and other stakeholders view the relation between corporate taxes and CA reporting in a manner consistent with that of the organizations that establish CA reporting guidelines. We examine the relation between the importance of CA reporting to a firm and the amount of corporate taxes the firm pays. We find that the importance of CA reporting is negatively related to five-year cash effective tax rates. Our evidence suggests that managers and other stakeholders of firms for which CA reporting is important do not view the payment of corporate taxes as socially responsible. JEL classification: G3; H2 Key Words: corporate accountability reporting; corporate social responsibility; tax avoidance *Corresponding author: Professor David A. Guenther Lundquist College of Business 1208 University of Oregon Eugene, OR 97403-1208 Phone: (541) 346-5384 Fax: (541) 346-3341 email: [email protected] Funding from the Finance and Securities Analysis Center at the Lundquist College of Business is gratefully acknowledged. The authors are also grateful for helpful comments from seminar participants at the University of Oregon.
1
Taxes and Corporate Accountability Reporting: Is Paying Taxes Viewed as Socially Responsible?
1. Introduction
As corporations have increased in size over time, they have attracted increasing attention
from outside groups that scrutinize how they affect the environment, treat their employees, and
impact the communities in which they operate. In response, corporations have begun to highlight
their social responsibility in corporate accountability reports (hereafter CA reports). In the past
two decades, the number of firms issuing CA reports worldwide has increased rapidly.
In response to this rapid increase in CA reports, several organizations have issued
guidelines for corporate accountability reporting, including the Global Reporting Initiative (GRI)
and the United Nations Global Compact (UN Global Compact). These organizations recognize
that one way in which public corporations make positive contributions to society is through the
payment of corporate taxes that support government programs created to improve social welfare.
For example, GRI asks firms to provide detailed information on tax payments because what is
“frequently desired by users of sustainability reports is the organization’s contribution to the
sustainability of a larger economic system (GRI 2011a, 25).”
However, while the current GRI guidelines suggest that users of CA reports view paying
more taxes as a positive contribution to social welfare, a detailed reading of CA reports reveals
stark differences in how firms (or their managers) view and report corporate taxes. Some firms
discuss taxes in a manner consistent with the GRI guidelines, i.e. paying more taxes improves
social welfare. In contrast, other firms argue that taxes are harmful to innovation, production, job
creation, and economic development, i.e. tax payments detract from social welfare. Many of
these firms go on to argue that they are being good corporate citizens by actively lobbying to
2
lower corporate taxes, an action they argue will lead to increased economic development.
Furthermore, some firms do not disclose any information about taxes in their CA reports, even
though GRI guidelines list taxes as a core economic indicator. For example, after analyzing CA
reports for a sample of forty U.S. public companies we find that nineteen (47.5%) of them do not
provide any information related to taxes paid. In some cases the reports refer the reader to the
firm’s Form 10-K for tax information, even though GRI disclosure guidelines for CA reporting
of taxes are different from those under U.S. GAAP.1
In addition to differences in the discussion and disclosure of taxes, firms also exhibit
differences in their assessment of the relative importance of taxes to their stakeholders. For
example, in assessing the importance of various items in their 2011 CA report, Intel rates “taxes”
as lowest in importance among sixteen items, whereas UPS rates “taxes paid” as higher in
importance than twenty-seven other items (including such items as “greenhouse gas policy and
advocacy”). Still other firms, such as The Coca-Cola Company, do not mention taxes at all in
their CA report, suggesting that these firms view taxes as being of low relative importance to
their stakeholders.2
The substantial differences across firms in their treatment of taxes in CA reports suggests
that the positive relation between (1) corporate tax payments and (2) the sustainability of a larger
economic system implicit in the GRI reporting guidelines is not necessarily consistent with how
corporate stakeholders view taxes. This difference in how corporate taxes are presented in CA
reports leads directly to our research question: Do corporate stakeholders (including managers)
1 For example, GRI guidelines include "all company taxes (corporate, income, property, etc.) and related penalties paid at the international, national, and local levels. This figure should not include deferred taxes because they may not be paid. For organizations operating in more than one country, report taxes paid by country (GRI 2011b, 5)." 2 Interestingly, while The Coca-Cola Company (a U.S. public company) does not mention taxes in its CA report, Coca-Cola HBC Switzerland Ltd (a Swiss subsidiary of a Greek bottling company) states in its 2011 sustainability report that "we contribute to society as an employer and tax payer."
3
for whom CA reporting is important view corporate tax payments as socially responsible? If
stakeholders view both CA reporting and corporate tax payments as socially responsible, we
expect a positive relation between a firm’s corporate tax payments and measures of the
importance of CA reporting to the firm. Under this view, managers of firms whose stakeholders
value CA reporting will dedicate resources to socially responsible activities and to reporting
these activities in their CA reports. To the extent a firm’s stakeholders view corporate tax
payments as increasing social welfare, managers of the firm will engage in less tax avoidance.3
However, existing evidence suggests either a positive or negative relation between the
importance of corporate social responsibility (CSR) and tax avoidance. Watson (2011) finds that
firms with low CSR indexes have larger reserves for unrecognized tax benefits under FIN48. He
argues this result is evidence that socially irresponsible firms are not as constrained by social
norms as other firms, and thus pursue aggressive tax strategies without fear of stakeholder
retribution. Lanis and Richardson (2011) find that higher levels of CSR disclosure are associated
with lower corporate tax aggressiveness for a sample of 408 Australian firms, suggesting that
firms deemed to be more socially responsible pay more taxes.
In contrast, Dhaliwal, Li, Tsang and Yang (2011) find a negative relation between CSR
and cost of capital, and Lev, Petrovits and Radhakrishnan (2010) find evidence of a positive
association between CSR activity and revenue growth. This evidence suggests that CSR can be
viewed as a mechanism to maximize firm value. If tax avoidance is also a mechanism to
maximize firm value, then we expect a positive correlation between CSR and tax avoidance,
3 At the extreme the managers may even pay more corporate taxes than legally required. See the Associated Press news story "Starbucks Promises to Pay More UK Tax" (December 6, 2012) in which Starbucks responded to criticism that its low UK taxes were "immoral" by agreeing to pay "somewhere in the range of 10 million pounds in each of the next two years" regardless of whether the company is profitable.
4
which suggests that firms that value CA reporting engage in more tax avoidance (i.e., have lower
tax payments).
To provide evidence on our research question, we examine the five-year cash effective
tax rates (ETRs) for a sample of U.S. public corporations that also have data on the MSCI ESG
STATS annual data set of environmental, social, and governance (ESG) ratings of publicly
traded companies for 2010. We use an indicator variable that captures the quality of a firm’s CA
reporting as a proxy for the importance of CA reporting to a firm’s stakeholders. As a second
proxy for the importance of CA reporting, we construct a CSR index for the period 2006-2010
following the approach in Kim, Park and Wier (2012). A positive association between ETRs and
the reporting quality measure would suggest that managers and other stakeholders view
corporate tax payments as socially responsible. An insignificant or negative association between
these variables would suggest that tax payments are not viewed as an important part of a
corporation’s CA reporting.
In our empirical tests we find that both the CA reporting quality measure and the CSR
index are negatively related to five-year cash ETRs. We find similar results using a modified
CSR index that only considers a firm’s activities with respect to “community,” which includes
such activities as charitable giving, innovative giving, community engagement and community
impact. In contrast to existing research, these results suggest a positive relation between CA
reporting and tax avoidance, i.e. firms that appear to value CA reporting pay lower corporate
taxes. This evidence is consistent with the conjecture that managers and other stakeholders do
not view the payment of corporate taxes as a positive indication of a corporation’s socially
responsible behavior.
5
In reviewing a sample of forty CA reports, we find that some firms state they actively
lobby for lower corporate taxes as a way of improving social welfare. Based on this finding we
also examine the tax lobbying behavior of firms for which CA reporting is important. In these
tests we use two measures of tax lobbying activity as the dependent variable, and find that both
the CA reporting quality measure and the CSR index are positively related to tax lobbying. This
result is consistent with our finding of a negative relation between cash ETRs and CA reporting
quality, and suggests that stakeholders do not necessarily view corporate taxes in the same way
as organizations that create CA reporting guidelines.
Our results make an important contribution to research on CA reporting in two ways.
First, we contribute to the public policy debate about the type of information that stakeholders
value in CA reports. This evidence should be useful to organizations such as the GRI that
produce CA reporting guidelines. Current GRI guidelines appear to view corporate tax payments
as contributing to the social welfare of the communities in which the corporation operates,
suggesting that paying more taxes is better from a social welfare standpoint. By incorporating
more information about the trade-off between corporate tax payments and other economic
activities that promote social welfare, CA reports may become a more important input to
stakeholders’ decision making process.
Second, our finding of a positive relation between CA reporting and tax avoidance is in
contrast to existing research that finds a negative relation between CSR and tax aggressiveness.
Watson (2011) finds that firms with very low CSR indexes are more tax aggressive than other
firms on average using a measure of tax aggressiveness that captures highly uncertain tax
positions. We instead focus on the full range of firms and document an overall positive relation
using a measure of legal tax avoidance. Thus, while Watson’s (2011) results suggest that firms
6
with the lowest CSR measures are more likely to take tax positions that would not be sustained
in court, our results suggest that, as CA reporting becomes more important, firms legally avoid
more taxes.
Overall our results highlight the potential offsetting effects of corporate taxes on social
welfare. While higher corporate tax payments provide governments with more revenue to spend
on social welfare, firms that focus on after-tax profitability are able to increase social welfare in
other ways, such as investment in infrastructure, R&D, and job creation. While our results do not
argue for any particular view of the relation between corporate tax payments and social welfare,
they do provide evidence that the relation is perhaps more complex than that reflected in the
current CA reporting guidelines.
2. Corporate Accountability Reporting and Taxes
We investigate how managers and other stakeholders for whom CA reporting is
important view corporate tax payments by examining the relation between measures of the
importance of CA reporting to a firm and the firm’s level of tax avoidance. This research
question is motivated by two observations about CA reports and corporate tax payments. First,
there appears to be a lack of any uniform treatment of information about corporate tax payments
in CA reports, despite GRI reporting guidelines that clearly emphasize the importance of tax
disclosures. Second, some firms for which CA reporting seems to be important, such as The
Coca-Cola Company, report no tax information at all.
As currently written, the GRI guidelines for CA reporting suggest that more socially
responsible firms pay more taxes. For example, the first core economic indicator suggested by
GRI is “direct economic value generated and distributed, including [...] payments to capital
7
providers and governments (GRI 2011a, 26).” To explain what it intends to capture with this
performance indicator, GRI (2011a, 25) states that what is “desired by users of sustainability
reports, is the organization’s contribution to the sustainability of a larger economic system.” It
further explains that disclosure of payments to governments should include, “all company taxes
(corporate, income, property, etc.) and related penalties paid at the international, national, and
local levels (GRI 2011b, 5).” Similarly, the UN Global Compact (UN 2009, 40) “encourages
participants to engage in partnership projects in support of broader UN goals.” One type of
partnership encouraged by the compact is a “core business partnership” in which partners
collaborate to, among other things, “generate tax revenues.”
Despite this positive relation between corporate tax payments and public welfare
suggested by CA reporting guidelines, cross-sectional differences exist in how firms view and
report taxes in their CA reports. Some firms such as Analog Devices report taxes in a manner
consistent with the view taken by the current guidelines and state that they contribute to the
community “through the ongoing operation of our business and taxes.” Other firms discuss taxes
in their CA reports as being harmful to innovation and economic development. For example,
Intel states in its 2011 Corporate Responsibility Report that it “believes in promoting tax policies
that encourage innovation and competition around the world. [...] the location of our facilities
can be substantially affected by the tax and economic development policies of potential host
countries.”
Some CA reports also discuss the fact that large public corporations affect tax policy
through their lobbying activities, which, in turn, affect social welfare.4 Many of these firms argue
that they are engaging in good corporate citizenship by actively lobbying to lower the tax
4 GRI Sustainability Reporting Guidelines version 3.1 includes lobbying activities as a core society performance indicator.
8
payments of businesses, a stance in direct contrast to how the issue of corporate tax payments is
framed by the current reporting guidelines. For example, in their 2011 CA report, 3M
Corporation states that their top objective in terms of public policy issues is “to make the case for
tax reform and lower corporate tax rates for a level global playing field.” Notably, in their 2011
CA report 3M rates tax policy as more important than either health or environmental concerns,
underscoring the importance with which they view the issue of lowering corporate tax payments.
3M further notes that their list of public policy issues “helps communicate 3M’s perspective and
guide its activities in the areas of lobbying […] as well as meetings with federal and state
government officials.”
Similarly, Texas Instruments’ 2011 CA report states that “We also joined other
companies in a multiyear effort to advocate for comprehensive tax reform that would lower the
corporate tax rate, embrace a territorial system of taxation in line with that of our foreign
competitors, and continue to provide incentives for investing in R&D in the United States.” If
firms that value CA reporting are more likely than other firms to lobby for tax policies that
promote welfare-maximizing activities and also lower corporate tax liabilities, this could lead to
a negative relation between CA reporting and taxes.
To understand how taxes and tax information are disclosed in CA reports, we examine
forty CA reports selected randomly from sample firms in the MSCI ESG STATS annual data set
from 2009 to 2010 that have high quality CA reports.5 We find that in nineteen (47.5%) of these
reports the firm either does not make any tax disclosures at all, or refers the reader to the firm’s
Form 10-K for tax information. We also observe wide variation in the amount and quality of tax
information that firms report.
5 We identify firms with high quality CA reports as those with Reporting Quality (CGOV_str_D) equal to one. We discuss this data in more detail in Section 3.1.
9
In Appendix A, we provide some examples of both positive and negative statements
firms make about taxes in the forty CA reports that we examined. Positive statements about taxes
focus on the importance of the firm’s tax payments to the social welfare of the community. As an
example of this type of statement, ConocoPhillips states “Our global operations contribute
substantially to social and economic development […] For example, our direct economic
contributions during 2011 included: Taxes – Our operations generated $14.9 billion in total tax
revenue to governments.” Negative statements about corporate taxes generally argue that high
tax rates discourage innovation and investment and harm job creation, and therefore hurt the
ability of the firm to contribute to social welfare. For example, 3M states “Tax reform is essential
to ensuring the long-term competitiveness of American businesses and workers. 3M believes
business tax reform should focus on a significant reduction of the corporate income tax rate.”
The fact that many companies do not provide tax information in their CA reports does not
necessarily mean the companies are not in compliance with reporting guidelines. GRI, for
example, only requires companies to report on items that are material. In discussing materiality
in the context of CA reporting, GRI states (2011a, 8): “Relevant topics and indicators are those
that may reasonably be considered important for reflecting the organization’s economic,
environmental, and social impacts, or influencing the decisions of stakeholders, and, therefore,
potentially merit inclusion in the report. Materiality is the threshold at which topics or indicators
become sufficiently important that they should be reported.” GRI gives an example of a two-
dimensional matrix on which to assess materiality, with the vertical axis reflecting “Influence on
Stakeholder Assessments and Decisions” and the horizontal axis reflecting “Significance of
Economic, Environmental, and Social Impacts.”
10
To illustrate the different ways that corporate taxes are viewed in this materiality matrix,
we present the matrixes from three public companies: Intel (Figure 1), UPS (Figure 2), and
Symantec (Figure 3). In the Intel matrix shown in Figure 1 “Taxes” is the lowest rated item on
the vertical axis labeled “Importance to Multiple Stakeholders.” Taxes is also rated to the left of
(i.e., below) the “medium” line on the horizontal axis labeled “Impact on Intel’s Business.” From
this figure one could conclude that corporate taxes are not very important to either Intel or its
stakeholders. Surprisingly, despite the low impact of taxes on Intel’s operations and
stakeholders, Intel provides a relatively large amount of information about taxes in its 2011
“Corporate Responsibility Report,” disclosing not only the amount of tax expense recorded, but
also the effective tax rate (ETR) for the year.6
Figure 2 shows the corresponding materiality matrix for UPS. Note that “Taxes Paid” is
quite high on the vertical axis indicating relatively high importance to stakeholders. However, it
rates to the far left (i.e., very low) on the horizontal axis, indicating low influence on UPS’s
business success. UPS also provides a relatively large amount of tax information in its CA
report. Figure 3 shows a similar materiality matrix for Symantec. Although Symantec does
provide information in their CA report about the amount of taxes paid, and even provides ETRs,
taxes are not shown on the materiality matrix. This may be because they would be to the far left
of the matrix (as with Intel and UPS) and are thus not considered material enough to place on the
matrix.
6 In addition, Intel provides a disclosure about the economic impact of Intel's corporate tax payments and other business activities on one of the communities in which it does business: "We periodically conduct local assessments to better understand Intel’s direct and indirect economic impact on the communities where we operate. For example, Intel commissioned ECONorthwest to prepare an economic impact assessment of our Oregon operations (our largest manufacturing site). Published in October 2011, the report found that 'total economic impacts attributed to Intel's operations, capital spending, contributions, and taxes amounted to almost $14.6 billion in economic activity, including $4.3 billion in personal income and 59,990 jobs in Washington County, Oregon.'"
11
=================== Insert Figures 1, 2, and 3
===================
From the above discussion it seems clear that for firms for which CA reporting is
important, corporate tax payments may or may not be an important (or material) item compared
with other CSR items in their reports. This view is consistent with the observation that many
corporations with CA reports make no mention of taxes in these reports. For example, the Coca-
Cola Company produces an extensive CA report prepared in accordance with GRI guidelines
(the 2012 report is 122 pages), yet the word “tax” does not appear once in the report. While this
evidence is merely anecdotal, it suggests that there is no clear relation between the importance of
CA reporting and corporate tax payments, despite the clear link between tax receipts and the
ability of governments to fund social services and the GRI guidelines stating that companies
should disclose their tax payments. This incongruence suggests that the relation between the
importance of CA reporting and tax payments is an empirical question.
2.1 Arguments suggesting a positive relation between CA reporting and taxes
Watson (2011) argues that the stakeholders of socially responsible firms would limit
managers’ ability to engage in tax aggressive strategies by establishing expectations, or social
norms, and imposing costs if firms deviate from these social norms. He suggests that
stakeholders of socially conscious firms would consider aggressive tax strategies as a deviation
from their social norm, which may lead them “to lose faith in the firm’s CSR reputation,
potentially causing them to sell the firm’s stock or to forgo purchasing it.” However, there are
two important differences between Watson’s study and ours. First, Watson focuses on firms with
low CSR index scores and argues that these firms will be more likely to engage in behavior that
12
deviates from social norms, whereas we are interested in those firms that seem to value CA
reporting, and whether this behavior is also associated with higher tax payments. Second,
Watson looks at the reserve for unrecognized tax benefits under FIN48 as his measure of tax
avoidance. FIN48 requires firms to record such a reserve when it is more likely than not that the
firm would not sustain the tax benefit in court. In our study we are interested in tax avoidance
behavior that managers and stakeholders would clearly consider to be “legal.” Since sustaining a
tax benefit in court may be viewed as a determination that the claimed benefit is “legal,”
recording a reserve under FIN48 may be considered to be taking a tax benefit that is likely to not
be legal.
The importance of this view that tax avoidance should be “legal” is evidenced by an
example from General Electric’s 2010 “Citizenship Report,” which states: “GE has taken
criticism recently regarding its U.S. tax obligations. Like any business or individual, we do like
to keep our tax rate low. But we fully comply with the law and there are no exceptions. GE acts
with integrity in relation to our tax obligations wherever we operate. At the same time, we have a
responsibility to our shareowners to reduce our tax costs as the law allows. Under any system,
GE will comply and pay what we owe.”
Anecdotal evidence suggests that public corporations are sensitive to how their tax
payments are perceived by stakeholders. First, firms communicate the impact of their tax
payments in their CA reports, as the example from Intel’s 2011 “Corporate Responsibility
Report” about Intel’s economic impact in Oregon illustrates. Second, public corporations also
take actions to make their tax payments appear larger than they really are. For example, a July
23, 2012 New York Times article reports “Apple Inc., already the world’s most valuable
company, understates its profits compared with other multinationals. It’s building up an
13
overlooked asset in the form of billions of dollars, tucked away for tax bills it may never pay.
Tax experts say the company could easily eliminate these phantom tax obligations. That would
boost Apple’s profits for the past three years by as much as $10.5 billion, according to
calculations by The Associated Press.”7
In another example, Apple responded to criticism that it’s taxes were too low by making
the following statement in an April 28, 2012 article in the New York Times: “In the first half of
fiscal year 2012 our U.S. operations have generated almost $5 billion in federal and state income
taxes, including income taxes withheld on employee stock gains, making us among the top
payers of U.S. income tax.” It is interesting to note that Apple includes in its tax payments the
amount of taxes paid by their employees on stock option compensation income. In the same
article Apple goes on to state: “Apple has been among the top creators of American jobs in the
past few years.”
Even legal tax avoidance has come under heavy criticism. Margaret Hodge, the chair of
the United Kingdom’s Parliament committee on public accounts, recently accused Starbucks,
Google, and Amazon of immoral behavior due to “using the letter of tax laws both nationally and
internationally to immorally minimize their tax obligations” (Lawless 2012). Even though
Margaret Hodge freely admits that these tax avoidance activities are clearly legal, she still
describes them as immoral. This suggests that, at least in some countries, legal tax avoidance
may not be viewed as a socially responsible activity. In response to this criticism, Starbucks
announced it would review its British tax practices in a bid to restore public trust (Lawless
7 Many U.S. multinational corporations designate unremitted foreign earnings as permanently reinvested earnings under APB 23, a designation that allows them to avoid recognizing the potential U.S. tax liability on these earnings. Apple, however, recognizes the potential U.S. tax expense in earnings, a practice that is considered conservative from an accounting perspective but also makes it appear that Apple pays more taxes than it does.
14
2012), providing further evidence that public corporations are sensitive to how their tax
payments are perceived by stakeholders.
2.2 Arguments suggesting a negative relation between CA reporting and taxes
While arguments in Watson (2011) and Lanis and Richardson (2012) are based on the
observation that tax payments increase the tax revenue that governments have at their disposal to
improve social welfare, other influences affect the relation between CA reporting and tax
payments. For example, as a firm’s after-tax profits increase, it impacts social welfare through
infrastructure investment and job creation, which both improves the income level of the
community (through employee salaries and payments to suppliers and contractors) and increases
other sources of tax revenue, including payroll taxes and employee taxes on salaries. The firm
can also use tax savings to directly invest in socially responsible activities. Moreover, some
activities that improve social welfare also decrease tax payments. For example, expenditures on
energy saving technology improve the environment while at the same time generating tax credits.
Each of these influences suggests a negative relation between CA reporting and tax payments.
Existing economics research on charitable payments also suggests either a negative or no
relation between CA reporting and tax payments. In investigating the relation between tax
payments and direct charitable payments that promote social welfare, economics research
identifies altruistic preferences and “warm glow” preferences as possible motives for voluntary
charitable contributions, which are difficult for economic models of selfish behavior to explain.
Altruistic preferences refer to individuals whose utility is increasing, not only in his or her own
payoff, but also in the total group payoff. To the extent altruistic preferences drive charitable
giving, individuals are indifferent between direct giving and the payment of taxes that provide an
15
equivalent increase in social welfare. Warm-glow preferences refer to individuals whose utility
increases by the act of giving, independent of how much the contribution increases group
payoffs. Harbaugh, Mayr, and Burghart (2007) study neural responses to tax-like transfers versus
voluntary transfers to a charitable organization and find evidence consistent with both the
altruistic and the warm-glow motives. To the extent managers and other stakeholders who value
CA reporting are subject to warm glow preferences, they prefer direct payments over tax
payments that increase social welfare, which suggests a negative relation between the importance
of CA reporting and taxes paid. To the extent managers and other stakeholders who value CA
reporting are subject to altruistic preferences, they are indifferent between direct payments or tax
payments that increase social welfare, which suggests no relation between the importance of CA
reporting and taxes paid.8
Another reason managers may not value tax payments as highly as direct expenditures on
social welfare is the fact that the direct expenditures are themselves tax deductible. For example,
if Walmart pays $1 of federal income taxes the cost to the company is $1, whereas if Walmart
spends $1 on other activities that promote social welfare the cost to the company is $1(1 – τ)
where τ is the company’s marginal tax rate. For a company facing the statutory 35% U.S.
corporate tax rate, spending on CSR activities effectively costs the company $0.65 for every
$1.00 spent to increase social welfare.9
Even if managers view the payment of taxes as an important element of a company’s
CSR activities, it may be the case that taxes are less important than other CSR activities, such as
environmental issues (e.g., greenhouse gases, pollution, and a company’s carbon footprint),
8 It may be the case that direct payments to improve social welfare result in a greater increase in social welfare than do tax payments of the same amount. In this case altruistic preferences would also suggest a negative relation between the importance of CA reporting and tax payments. 9 Note that this argument does not hold for all taxes. For example, state and local income taxes and property taxes are also deductible for federal income tax purposes.
16
product safety or sustainability, or employee welfare (e.g., treatment of workers in less
developed countries). Thus, these other issues could take more resources or manager attention
than tax payments. In fact, Figures 1 through 3 provide examples that taxes are less important (or
less material) to managers and other stakeholders than other CSR activities.
Based on the above analysis, we believe that there are important reasons for managers
and other stakeholders to view the payment of corporate taxes differently from other expenses
that a firm incurs to promote social welfare, and thus we argue that the link between the
importance of CA reporting and the amount of corporate taxes paid is an empirical question. We
provide empirical evidence to answer that question in the next section.
3. Empirical Tests
3.1 Sample
The sample used for our empirical tests initially consists of all firms for which data are
available on Compustat for the years 2002-2010. Data on CA reporting quality and the CSR
index, which we use as proxies for the importance of CA reporting to managers and other
stakeholders, come from the 2010 MSCI ESG STATS annual data set of environmental, social,
and governance ratings of publicly traded companies. According to the Wharton Research Data
Services web site (https://wrds-web.wharton.upenn.edu/wrds/), MSCI (which was formerly
known as KLD Research & Analytics, Inc.) is the leading source of data for institutional
investors conducting social research, and has provided high quality research, benchmarks,
compliance, and consulting services since 1988.
According to Kim et al. (2012), “KLD uses a combination of surveys, financial
statements, and articles in the popular press and academic journals, as well as government
17
reports, to assess social performance along dimensions such as corporate governance,
community, diversity, employee relations, environment, product, and exclusionary screen
categories including alcohol, gambling, military contracting, nuclear power, and tobacco.” To
assess a company’s level of social responsibility, KLD examines both positive indicators
(strengths) and negative indicators (concerns). However, only negative indicators are used to
evaluate exclusionary screen categories.
Numerous scholarly studies use the KLD data to operationalize the CSR construct (e.g.,
Turban and Greening 1997; Waddock and Graves 1997; Szwajkowski and Figlewicz 1999). In
describing the validity and usefulness of the data Kim et al (2012) go on to state “Deckop et al.
(2006) describe KLD as ‘the largest multidimensional corporate social performance database
available to the public.’ Szwajkowski and Figlewicz (1999) evaluate and assess the validity and
reliability of the KLD database. They find that KLD ratings have substantial and discernible
validity with especially strong internal discriminant validity. Some researchers maintain that the
KLD data are ‘the de facto research standard’ for measuring CSR in scholarly research (e.g.,
Waddock 2003, 369). Chatterji et al. (2009) contend that KLD’s social ratings are among the
most influential and the most widely accepted CSR measure used by academics. Mattingly and
Berman (2006) assert that the KLD dataset has become the standard for quantitative
measurement of corporate social actions.”
We use the MSCI (formerly KLD) measure of Reporting Quality Strength as our first
proxy for the importance of CA reporting to managers and other stakeholders. Reporting Quality
Strength is a binary variable (1 for high quality or 0 for low quality) and is explained by the
database as follows: “This indicator measures the quality of a firm’s reporting on its corporate
social responsibility (CSR)/sustainability efforts. Factors affecting this evaluation include, but
18
are not limited to, the completeness and specificity of a firm’s reporting, its setting of specific
goals for its CSR efforts, and quantitative measurement of progress towards these goals. This
indicator also measures whether a firm follows agreed-upon guidelines, such as those established
by the Global Reporting Initiative.”
An underlying assumption in using the importance of CA reporting to address our
research question is that it reflects the importance of corporate social responsibility (CSR) to a
firm’s managers and other stakeholders. In other words, if a firm values CSR and invests in
related activities, it will choose to report its efforts to stakeholders through the issuance of a high
quality CA report. One potential concern with this measure is that CA reporting may not be
consistent with a corporation’s commitment to CSR if managers are not truthful in their CA
reporting. To ensure that our reporting quality measure is identifying firms for which CA
reporting reflects an underlying concern with socially responsible activities, we also construct
CSR index scores as a second proxy for the importance of CA reporting. Following Kim et al.
(2012), we construct CSR index scores based on the following five KLD categories: community,
diversity, employee relations, environment, and product. We also follow Kim et al. in that we
exclude the category for corporate governance, since the link between corporate governance and
the value of corporate tax avoidance has been studied previously, and is different from the link
between CA reporting and taxes that is the focus of our study. We construct a CSR index score
by adding 1 for each identified “strength” and subtracting 1 for each identified “concern.” As in
Kim et al., we also do not adjust the CSR index score for activities in the exclusionary screen
categories: alcohol, gambling, military contracting, nuclear power, and tobacco.
To obtain a measure of a firm’s tax payments, we use the five-year cash ETR proposed
by Dyreng, Hanlon and Maydew (2008). For each sample firm we obtain from the Compustat
19
database both pretax income (Compustat data item PI) and the amount of cash paid for taxes
from the statement of cash flows (Compustat data item TXPD) for each year. The five-year cash
ETR is the sum of the taxes paid for a five year period divided by the sum of the pretax income
for the same five year period.
We obtain firms’ tax lobbying expenditures from the publicly available database
maintained by the Center for Responsive Politics. These data are made available under the
Lobbying Disclosure Act of 1995, which mandated reporting of lobbying activities. Our use of
this lobbying data is consistent with prior research (e.g., Hill, Kubick, Lockhart and Wan 2012).
The lobbying data identify amounts spent on different types of lobbying, and in line with our
research question we only use data on a firm’s tax lobbying.
Our final sample combines observations with data on all three databases for a given year:
Compustat, the MSCI ESG STATS database, and the lobbying database. We eliminate
observations for firms incorporated outside of the U.S., observations with one- or five-year cash
ETRs greater than 1 or less than 0, and observations with missing data for any of our regression
variables. This results in a final sample of 2,118 observations.
3.2 Research Design
Our empirical tests are versions of the following regression model:
DEPVAR = α + β1CA_Rpt + ΣβkCONTROLS + ε (1)
DEPVAR is one of three different variables:
Cash_ETR = five-year sum (from year t-4 to year t) of cash taxes paid (Compustat data item TXPD) divided by the five-year sum of pre-tax income (PI). To allow for meaningful interpretation firms are required to have a Cash ETR between 0 and 1.
20
Lobby(0/1) = an indicator variable equal to 1 if the firm reported lobbying expenditures for tax purposes in the current year and 0 otherwise.
Lobby_Exp = natural log of one plus the firm’s total reported lobbying expenditures for
tax purposes. CA_Rpt is one of three different variables:
Rpt_Quality = indicator variable equal to one if a firm is defined as having a strength in the governance category “Reporting Quality CGOV_str_D” of MSCI ESG STATS’ database. CSR_Index = total strengths minus total concerns in MSCI ESG STATS’ five social
rating categories: community, diversity, employee relations, environment, and product.
Community = total strengths minus total concerns in MSCI ESG STATS’ “Community” category. CONTROLS consist of two sets of control variables that differ depending on whether the
dependent variable is the Cash ETR or one of the lobbying measures. Cash ETR control
variables are:
Size = log of total assets (AT).
Leverage = long-term debt (DLTT) plus short-term debt (DLC) scaled by lagged total assets (AT).
Intang = intangible assets (INTAN) divided by lagged total assets (AT).
Tax_Bnft_Opt = tax benefit of stock options (TXCBO + TXBCOF) divided by lagged total assets (AT).
PTROA = pretax income (PI) divided by lagged total assets (AT).
MTB = price per share (PRCC_F) times total common shares outstanding (CSHO) over book value of equity (CEQ).
For_Inc = absolute value of pretax foreign income (PIFO) divided by the absolute
value of total pretax income (PI). Lobbying control variables (from Hill, Kelly, Lockhart, and Van Ness 2011) are Size, PTROA,
MTB, and:
21
Capital = an indicator variable equal to one if a firm is headquartered in the capital city of its state of incorporation, and zero else.
In addition, all regressions have year and industry fixed effects, where industry is based on the
Fama-French 17 industry classification. All regression statistical tests use standard errors that are
clustered by firm and year.10
Regressions using the five-year Cash ETR as the dependent variable use mean values of
all regression variables measured over the same five year period as the ETR. Regressions using
lobbying as the dependent variable use annual values for all variables. All continuous
independent variables are winsorized at the first and ninety-ninth percentiles. All variables are
defined in Appendix A.
3.3 Results
Table 1 presents descriptive statistics for the variables used in our tests. The 75th
percentile of Rpt_Quality is zero, indicating that less than 25% of the sample observations are
considered to have high quality supplemental CA reports for the sample period. Means and
medians of CSR_Index are negative, indicating that the average firm has more “concerns” than
“strengths” using the KLD ratings. Mean and median Cash_ETR is 29%, slightly lower than the
35% statutory U.S. corporate income tax rate. The mean ratio of foreign income to total income
(For_Inc) is 27%, while the median is only 4%, suggesting that most firms have a relatively
small amount of foreign income. On average only 6.6% of firms engage in tax lobbying activities
in any given year (Lobby(0/1)).
========== Insert Table 1 ==========
10 Results are robust to not clustering standard errors or to clustering by only firm or only year.
22
3.3.1 Cash_ETR Results
Table 2 presents the Pearson correlation coefficients between variables. The correlation
coefficient of the relation between either Rpt_Quality or CSR_Index and five-year cash ETRs is
essentially zero. Based on this result there is no evidence that firms with higher quality CSR
reports or higher CSR indexes, proxies for the importance of corporate accountability reporting
to managers and other stakeholders, pay more in corporate taxes. We next present multivariate
regression tests including control variables.
========== Insert Table 2 ==========
Table 3 presents results of estimating equation (1) when the dependent variable is the
five-year cash ETR. Since the ETRs are measured over five years, all other regression variables
are the mean values over the same five-year period in order to measure the same time period as
the ETRs. Columns (1) and (2) present results with only industry and year fixed effects, while
columns (3) and (4) contain all of the control variables. Columns (1) and (3) use Rpt_Quality as
an independent variable to capture the importance of corporate accountability reporting to
managers and other stakeholders, while columns (2) and (4) use CSR_Index. The coefficient on
Rpt_Quality is significantly negative with and without control variables, suggesting that firms
with higher CA report quality pay lower taxes. Furthermore, the coefficient on CSR_Index is also
significantly negative with and without control variables, indicating that firms with higher CSR
indexes also pay lower taxes than other firms.
========== Insert Table 3 ==========
23
To determine whether the negative relation between CSR_Index and Cash_ETR reported
in Table 3 is due to firms with low CSR indexes paying more taxes, or firms with high CSR
indexes paying less taxes (or both) we create dummy variables to indicate those firms in the
highest or lowest quintiles of CSR_Index and include these as independent variables in equation
(1) (replacing CSR_Index). Results of estimating this regression are presented in Table 4. The
results in columns (1) and (3) demonstrate that those observations in the highest quintile of
CSR_Index have significantly lower values of Cash_ETR than other firms. This result suggests
that it is the firms for which CSR and CA reporting are most important that drive the negative
relation between taxes and CSR indexes reported in Table 3.
========== Insert Table 4 ==========
3.3.2 Tax Lobbying Results
A detailed examination of CA reports indicates that some firms argue that they are
engaging in good corporate citizenship by lobbying for reduced tax payments on businesses.
While this evidence is in direct contrast to the view implicit in the current guidelines for CA
reporting, it is strictly anecdotal. We test whether firms whose managers and other stakeholders
place greater importance on CA reporting engage in more tax lobbying activities. If firms that
place a greater importance on CA reporting actively lobby for lower taxes, it suggests that these
firms (and their stakeholders) do not view the relation between corporate tax payments and CA
reporting in the same way as organizations like GRI that provide reporting guidelines.
Table 5 reports results of estimating equation (1) when the dependent variable is
Lobby(0/1), an indicator variable that is equal to 1 in the year a firm incurs expenditures on tax
24
lobbying and 0 otherwise. Since the dependent variable is a 0/1 dummy, we estimate equation (1)
as a logit model. Columns (1) and (2) contain only year and industry fixed effects, while columns
(3) and (4) report results with all control variables included. Columns (1) and (3) use
Rpt_Quality as an independent variable, while columns (2) and (4) use CSR_Index. Results are
similar in all of the regression specifications. We find a significant positive relation between tax
lobbying and both measures of CSR importance. This result suggests that those firms for which
CA reporting is most important are also more likely to engage in tax lobbying.
========== Insert Table 5 ==========
Table 6 reports results of estimating equation (1) when the dependent variable is
Lobby_Exp, the amount spent by the firm on tax lobbying activities in the year. Columns (1) and
(2) contain only year and industry fixed effects, while columns (3) and (4) report results with all
control variables included. Columns (1) and (3) use Rpt_Quality as an independent variable,
while columns (2) and (4) use CSR_Index. Results are similar in all of the regression
specifications, and are similar to those of Table 5 using the 0/1 dummy variable. We find a
significant positive relation between the amount spent on tax lobbying and both measures of
CSR importance. This result suggests that those firms for which CA reporting is most important
also spend more on tax lobbying.
Overall, the results in Tables 3 through 6 suggest that firms whose managers and other
stakeholders place greater importance on CA reporting pay less corporate taxes and engage in
more tax lobbying activities.
========== Insert Table 6 ==========
25
3.4 Sensitivity Tests
To investigate whether our results are sensitive to how we construct our CSR index we
conduct an additional test using an alternative CSR index measure. Because tax payments that
impact the social welfare of a corporation’s community are most similar to the MSCI ESG
STATS category of “Community,” we construct an alternative index using just the ratings from
the community category.
The community category consists of four Strengths and one Concern. Strengths are as
follows: (1) Charitable Giving, (2) Innovative Giving, (3) Community Engagement, and (4)
Other Strengths. Charitable Giving indicates that the company has given 1% or more of trailing
three-year net earnings before taxes to charity, or has otherwise been notably generous in its
giving. Innovative Giving indicates that the company donates 25% or more of its charitable
giving to support non-governmental organizations (NGOs) involved with affordable housing,
access to healthcare, K-12 education, and initiatives to relieve hunger and/or other services to
disadvantaged communities. Community Engagement indicates that the company has a notable
community engagement program concerning involvement of local communities in areas where
the firm has major operations. Other Strength indicates that the company has either an
exceptionally strong in-kind giving program or engages in other notably positive community
activities.
The one Concern included in the community category is Community Impact, which
measures the severity of controversies related to a firm’s interactions with communities in which
it does business. Factors affecting this evaluation include, but are not limited to, a history of
involvement in land use and/or development-related legal cases, widespread or egregious
26
community impacts due to company operations, and criticism by NGOs and/or other third-party
observers.
Table 7 presents results of estimating equation (1) when the dependent variable is the
five-year cash ETR. Since the ETR is measured over five years, all other regression variables are
the mean values over the same five-year period to match up with the ETRs. Column (1) presents
results with only industry and year fixed effects, while column (2) contains all of the control
variables. Both columns use Community as an independent variable to capture the importance of
community-related CSR activities to managers and/or stakeholders. The coefficient on
Community is significantly negative with and without control variables, suggesting that firms for
which community-related CSR activities are more important pay less corporate taxes. This result
is consistent with the result for CSR_Index reported in Table 3.
========== Insert Table 7 ==========
Table 8 reports results of estimating equation (1) when the dependent variable is
Lobby(0/1), an indicator variable that is equal to 1 in the year a firm incurs expenditures on tax
lobbying and 0 otherwise. Since the dependent variable is a 0/1 dummy, we estimate equation (1)
as a logit model. Column (1) contains only year and industry fixed effects, while column (2)
reports results with all control variables included. Both columns use Community as an
independent variable. Results are similar in both columns. We find a significant positive relation
between tax lobbying and both measures of CSR importance. However, the coefficient is only
significant at the 10% level (two-tail test) when the control variables are included. These results
suggest that those firms for which community related CSR is most important are also more likely
to engage in tax lobbying.
27
========== Insert Table 8 ==========
Table 9 reports results of estimating equation (1) when the dependent variable is
Lobby_Exp, the amount spent by the firm on tax lobbying activities in the year. Column (1)
contains only year and industry fixed effects, while column (2) reports results with all control
variables included. Both columns use Community as an independent variable. Results are similar
in both regression specifications, and are similar to those of Table 8 using the 0/1 dummy
variable. We find a significant positive relation between the amount spent on tax lobbying and
the measure of the importance of community related CSR to the firm. This result suggests that
those firms for which community related CSR is most important also spend more on tax
lobbying.
========== Insert Table 9 ==========
4. Conclusion
A great deal of inconsistency exists in the way public corporations address the issue of
taxes in their supplemental CA reports, and results from prior accounting research on the relation
between taxes and CSR are mixed. In addition, research from economics suggests that some
individuals derive more utility from making direct payments to charities than from the payment
of taxes that support the same charitable activity, suggesting that corporate tax payments may not
be viewed by managers and other stakeholders as substitutes for other types of expenditures on
CSR activities. In this study we investigate the relation between the apparent importance of CA
reporting to managers and other stakeholders of U.S. public corporations (as proxied for by both
28
a measure of the quality of a firm’s supplemental accountability reporting and a firm’s CSR
index score) and both the amount of taxes paid by the firm and the amount of tax lobbying
behavior the firm engages in.
We find that the CA reporting quality measure and the CSR index are both negatively
related to five-year cash ETRs, and are both positively related to tax lobbying activities. We find
similar results using a modified CSR index that only considers a firm’s activities with respect to
“community,” which includes such activities as charitable giving, innovative giving, community
engagement and community impact. Overall our evidence suggests that, on average, managers
and other stakeholders do not view the payment of corporate taxes as an important part of a
public corporation’s socially responsible behavior. Our evidence is also consistent with anecdotal
evidence that suggests that using legal means to reduce taxes is not considered to be a socially
undesirable activity.
Our results provide evidence that the relation between corporate tax payments and social
welfare is perhaps more complex than that reflected in the current CA reporting guidelines. As
such, our results provide important evidence that can be used in developing future guidelines.
29
Appendix A Examples of statements about taxes in CA reports
Positive Statements about Taxes Company Name Date Statement about Taxes ConocoPhillips 2012 Our global operations contribute substantially to social
and economic development in the communities in which we operate. For example, our direct economic contributions during 2011 included: Taxes – Our operations generated $14.9 billion in total tax revenue to governments, excluding royalties to government entities.
Walt Disney Company 2010 All Disney business units have an economic impact in the communities in which they operate [...] through direct and indirect employment, taxation and other factors.
Analog Devices 2010 ADI knows that our economic responsibility is not limited to our financial performance. IN addition to generating returns for investors, we create and support jobs in the communities in which we operate. We also contribute to those communities through the ongoing operation of our business, taxes, and voluntary employee activities. ADIs indirect economic impacts that go beyond the information provided above fall into two main categories… “secondary impacts from direct spending – these include taxes paid by employees…
Heinz 2010/2011 Heinz paid more than $293 million in income taxes over the two-year period in the countries where we manufacture and/or sell products, as well as real estate taxes and other applicable taxes.
Molson Coors Brewing 2010 2009 economic impact: Employed nearly 8,700 people, Paid $1.2 billion in wages and benefits, Contributed $1.3 billion in tax revenue, Spent $4.3 billion with its network of more than, 10,000 suppliers, Garnered total net revenue of $7.6 billion; ($8.9 billion pre-excise tax)
Negative Statements about Taxes Walt Disney Company 2010 ...the Company believes that active participation in the
political life of the countries and communities in which we do business is in the best interests of the Company and its shareholders. As a result, we participate in policy debates on many issues [...] including appropriate taxation.
30
Eastman Chemical 2011 Eastman officers and employees were active members and leaders of key committees within these chemical industry and business trade organizations during 2010, focusing on [...] tax issues...
3M 2011 Tax reform is essential to ensuring the long-term competitiveness of American businesses and workers. 3M believes business tax reform should focus on a significant reduction of the corporate income tax rate, making the research credit permanent, creating incentives for U.S.-based IP ownership and simplifying the tax code. 3M supports state government policies designed to incentivize capital investment and sustainable job creation. In particular, we believe state tax policy should reward both job creation and plant expansion and modernization investments.
Lilly (Eli) & Co. 2010 Lilly also supports efforts to decrease the final price of medicine to patients, such as through minimizing value-added taxes and markups applied in the supply chain.
31
Appendix B Variable Definitions
Cash_ETR = the 5 year sum (from year t-4 to year t) of cash taxes paid (TXPD)
divided by the 5-year sum of pre-tax income (PI). In order to allow for meaningful interpretation firms are required to have a Cash ETR between 0 and 1.
Size = log of total assets (AT)
MTB = price per share (PRCC_F) times total common shares outstanding (CSHO) over book value of equity (CEQ)
Leverage = long-term debt (DLTT) plus short-term debt (DLC) scaled by lagged total assets (AT)
Intang = intangible assets (INTAN) divided by lagged total assets (AT)
Tax_Bnft_Opt = tax benefit of stock options (TXCBO + TXBCOF) divided by lagged total assets (AT)
PTROA = pretax income (PI) divided by lagged total assets (AT)
For_Inc = absolute value of pretax foreign income (PIFO) divided by the absolute value of total pretax income (PI)
Rpt_Quality = an indicator variable equal to one if a firm is defined as having a strength in the transparency category of MSCI ESG STAT’s corporate governance database.
CSR_Index = total strengths minus total concerns in MSCI ESG STATS' five social rating categories: community, diversity, employee relations, environment, and product.
Community = total strengths minus total concerns in MSCI ESG STAT’s “community” category
Lobby(0/1) = an indicator variable equal to one if the firm reported lobbying expenditures for tax purposes in the current year and zero else.
Lobby_Exp = natural log of one plus the firm’s total reported lobbying expenditures for tax purposes.
Capital = an indicator variable equal to one if a firm is headquartered in the capital city of its state of incorporation, and zero else.
32
References
Chatterji, A., D. Levine, and M. Toffel. 2009. How well do social ratings actually measure corporate social responsibility? Journal of Economics and Management Strategy 18 (1): 125–169.
Deckop, J., K. Merriman, and S. Gupta. 2006. The effects of CEO pay structure on corporate
social performance. Journal of Management 32 (3): 329–342. Dhaliwal, D., O. Li, A. Tsang, and G. Y. Yang. 2011. Voluntary nonfinancial disclosure and the
cost of equity capital: The initiation of corporate social responsibility reporting. The Accounting Review 86 (1): 59–100.
Dyreng, S., M. Hanlon, and E. Maydew. 2008. Long-run corporate tax avoidance. The
Accounting Review 83 (1): 61-82. Gallemore, J., E. Maydew, and J. Thornock. 2012. The Reputational Costs of Tax Avoidance and
the Under-Sheltering Puzzle. Working Paper, University of North Carolina at Chapel Hill and University of Washington.
Global Reporting Initiative. 2011a. Sustainability Reporting Guidelines version 3.1.
https://www.globalreporting.org/resourcelibrary/G3.1-Sustainability-Reporting-Guidelines.pdf.
Global Reporting Initiative. 2011b. Indicator Protocols Set Economic (EC) version 3.1.
https://www.globalreporting.org/resourcelibrary/G3.1-Economic-Indicator-Protocols.pdf. Harbaugh, W., U. Mayr, and D. Burghart. 2007. Neural responses to taxation and voluntary
giving reveal motives for charitable contributions. Science 316: 1622-1625. Hill, M., T. Kubick, B. Lockhart, and H. Wan. 2012. The Effectiveness of Political Tax
Minimization. University of Nebraska Working Paper. Hill, M., G. Kelly, B. Lockhart, and R. Van Ness. 2011. Determinants and Effects of Corporate
Lobbying. University of Nebraska Working Paper. Kim, Y., M. Park, and B. Wier. 2012. Is Earnings Quality Associated with Corporate Social
Responsibility? The Accounting Review 87 (3): 761–796 Lawless, J. 2012. UK lawmakers seek end to corporate tax avoidance. Associated Press. Lanis, R. and G. Richardson. 2011. Corporate social responsibility and tax aggressiveness: An
empirical analysis. Journal of Accounting and Public Policy. doi: 10.1016/j.jaccpubpol.2011.10.006.
33
Lev, B., C. Petrovits, and S. Radhakrishnan. 2010. Is doing good good for you? How corporate charitable contributions enhance revenue growth. Strategic Management Journal (31): 182-200.
Mattingly, J. E., and S. Berman. 2006. Measurement of corporate social action: Discovering
taxonomy in the Kinder Lydenburg Domini ratings data. Business and Society 45 (1): 1–27.
Palfrey, T. and J. Prisbrey. 1997. Anomalous behavior in public goods experiments: How much
and why? The American Economic Review 87(5): 829-846. Szwajkowski, E., and R. Figlewicz. 1999. Evaluating corporate performance: A comparison of
the Fortune reputation survey and the Socrates social ranking database. Journal of Managerial Issues 11 (2): 137–154.
Turban, D. B., and D. W. Greening. 1997. Corporate social performance and organizational
attractiveness to prospective employees. Academy of Management Journal 40 (3): 658–672.
United Nations. 2009. The Practical Guide to the United Nations Global Compact. United
Nations Global Compact Office. http://www.unglobalcompact.org/docs/ communication_on_progress/Tools_and_Publications/Practical_Guide_2008.pdf.
Waddock, S. A. 2003. Stakeholder performance implications of corporate responsibility. Internal
Journal of Business Performance Management 5 (2/3): 113–124. Waddock, S., and S. Graves. 1997. The corporate social performance—Financial performance
link. Strategic Management Journal 18 (4): 303–319. Watson, L. 2011. Corporate social responsibility and tax aggressiveness: An examination of
unrecognized tax benefits. Working paper. The Pennsylvania State University.
34
Figure 1 Materiality assessment of taxes and other sustainability items from the 2011 Intel Corporate Responsibility Report
35
Figure 2 Materiality assessment of taxes paid and other sustainability items from the 2011 UPS Corporate Sustainability Report
36
Figure 3 Materiality assessment of sustainability items from the 2012 Symantec Corporate Responsibility Report
37
Table 1
Descriptive Statistics
Variablea N Mean Std Dev 25th Pctl 50th Pctl 75th Pctl
Rpt_Quality 2,118 0.076 0.264 0.000 0.000 0.000
CSR_Index 2,118 -0.041 2.596 -2.000 -1.000 1.000
Community 2,118 0.118 0.517 -1.000 0.000 2.000
Cash_ETR 2,118 0.285 0.158 0.198 0.288 0.359
Size 2,118 7.763 1.535 6.674 7.626 8.730
Leverage 2,118 0.218 0.212 0.041 0.176 0.323
Intang 2,118 0.210 0.220 0.020 0.131 0.350
Tax_Bnft_Opt 2,118 0.001 0.003 0.000 0.000 0.001
PTROA 2,118 0.099 0.096 0.035 0.082 0.142
MTB 2,118 2.692 2.043 1.347 2.043 3.134
For_Inc 2,118 0.268 0.406 0.000 0.043 0.436
Lobby(0/1) 2,118 0.066 0.248 0.000 0.000 0.000
Lobby_Exp 2,118 0.945 3.603 0.000 0.000 0.000
Capital 2,118 0.034 0.181 0.000 0.000 0.000 a All variables are as defined in Appendix A and all continuous independent variables are winsorized at the 1 and 99th percentiles.
38
Table 2
Pearson Correlations
I II III IV V VI VII VIII IX X XI XII XIII (I) Rpt_Quality
(II) CSR_Index 0.514
(III) Community 0.416 0.613
(IV) Cash_ETR 0.004 -0.021 0.014
(V) Size 0.359 0.325 0.239 -0.016
(VI) Leverage 0.040 -0.015 -0.018 -0.180 0.222
(VII) Intang 0.072 0.063 0.044 0.043 0.062 0.153 (VIII) Tax_Bnft_Opt -0.045 0.034 0.005 -0.029 -0.129 -0.087 0.009
(IX) PTROA 0.065 0.092 0.061 0.045 -0.177 -0.136 0.113 0.311
(X) MTB 0.091 0.139 0.072 -0.097 -0.099 -0.008 0.016 0.168 0.361
(XI) For_Inc 0.128 0.087 0.035 0.019 0.080 -0.046 0.123 -0.017 -0.006 0.015
(XII) Lobby(0/1) 0.314 0.240 0.191 -0.028 0.361 0.006 0.029 0.002 0.022 0.025 0.113
(XIII) Lobby_Exp 0.327 0.249 0.198 -0.025 0.378 0.003 0.031 -0.003 0.025 0.025 0.115 0.991
(XIV) Capital -0.024 0.009 0.038 0.027 0.005 0.001 -0.011 -0.025 0.018 -0.009 -0.037 -0.018 -0.019
39
Table 3 Relation between taxes and CA Reporting measures
(Dependent variable is the five-year cash effective tax rate)
Variablesa Predicted Sign (1) (2) (3) (4)
Rpt_Quality ? -0.020* -0.043***
(-1.95) (-2.73)
CSR_Index ? -0.003** -0.004*** (-2.34) (-2.82)
Size - 0.005 0.005
(1.48) (1.45)
Leverage - -0.171*** -0.173***
(-8.31) (-8.31)
Intang - 0.051*** 0.052*** (2.96) (3.19)
Tax_Bnft_Opt - -1.921** -1.699*
(-2.09) (-1.75)
PTROA + 0.082 0.075
(1.62) (1.53) MTB - -0.004*** -0.003***
(-3.31) (-3.07)
For_Inc - 0.026*** 0.027***
(2.74) (3.13)
Constant ? 0.254*** 0.252*** 0.290*** 0.284***
(14.95) (15.23) (12.79) (12.43) Industry Effects Yes Yes Yes Yes
Year Effects Yes Yes Yes Yes
Observations 2,118 2,118 2,118 2,118
R-squared 0.036 0.037 0.100 0.100 a All variables are as defined in Appendix A
*,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests). T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).
40
Table 4 Relation between taxes and high/low CA Reporting measure (Dependent variable is the five-year cash effective tax rate)
Variablesa Predicted Sign (1) (2) (3) (4)
CSR Highest Quintile - -0.012* -0.019***
(-1.83) (-2.64)
CSR Lowest Quintile + 0.004 0.007 (0.47) (0.67)
Size - 0.005 0.003
(1.55) (0.97)
Leverage - -0.172*** -0.171***
(-8.29) (-8.37)
Intang - 0.052*** 0.054*** (3.07) (3.32)
Tax_Bnft_Opt - -1.724* -1.814*
(-1.83) (-1.85)
PTROA + 0.079 0.072
(1.59) (1.46) MTB - -0.003*** -0.004***
(-3.19) (-3.33)
For_Inc - 0.025*** 0.025***
(2.70) (2.70)
Constant ? 0.255*** 0.247*** 0.288*** 0.291***
(15.19) (15.28) (12.94) (11.76) Industry Effects Yes Yes Yes Yes
Year Effects Yes Yes Yes Yes
Observations 2,118 2,118 2,118 2,118
R-squared 0.036 0.036 0.100 0.098 a All variables are as defined in Appendix A
*,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests). T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).
41
Table 5 Relation between tax lobbying activity and CA Reporting measures – Logit Model (Dependent variable is equal to one if the firm lobbied for tax purposes during the year)
Variablesa Predicted Sign (1) (2) (3) (4)
Rpt_Quality ? 2.696*** 0.867***
(14.48) (4.56)
CSR_Index ? 0.278*** 0.084** (13.99) (2.18)
Size + 0.983*** 1.019***
(15.94) (11.49)
PTROA + 2.826* 2.61
(1.77) (1.55)
MTB - 0.009 0.002 (0.44) (0.925)
Capital + -0.593 -0.595
(-0.69) (-0.84)
Constant ? -15.91*** -15.51*** -23.39*** -23.89***
(-19.32) (-11.43) (-16.09) (-8.22) Year Effects Yes Yes Yes Yes
Industry Effects Yes Yes Yes Yes
Observations 1,999 1,999 1,999 1,999 R-squared 0.145 0.114 0.323 0.320
a All variables are as defined in Appendix A *,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests).
T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).
42
Table 6 Relation between amount spent on tax lobbying and CA Reporting measures – OLS Model
(Dependent variable is tax lobbying expenditures)
Variablesa Predicted Sign (1) (2) (3) (4)
Rpt_Quality ? 4.607*** 2.751***
(12.24) (9.58)
CSR_Index ? 0.355*** 0.161***
(10.15) (3.63)
Size + 0.805*** 0.894*** (9.02) (13.30)
PTROA + 2.080** 2.139**
(2.17) (2.07)
MTB - 0.013 0.012
(0.76) (0.47)
Capital + -0.170 -0.310 (-0.42) (-0.92)
Constant ? -1.024** -1.067** -6.861*** -6.796***
(-2.47) (-2.01) (-9.40) (-10.53)
Year Effects Yes Yes Yes Yes
Industry Effects Yes Yes Yes Yes
Observations 2,118 2,118 2,118 2,118 R-squared 0.123 0.078 0.211 0.190
a All variables are as defined in Appendix A. *,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests). T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).
43
Table 7 Relation between taxes and CSR "community" measure
(Dependent variable is the five-year cash effective tax rate)
Variablesa Predicted Sign (1) (2)
Community ? -0.013** -0.021*** (-2.06) (-3.32) Size + 0.005* (1.65) Leverage - -0.173*** (-8.20) Intang + 0.052*** (3.19) Tax_Bnft_Opt - -1.760* (-1.87) PTROA + 0.0746 (1.53) MTB - -0.004*** (-3.22) For_Inc + 0.025*** (2.73) Constant ? 0.252*** 0.286*** (15.61) (12.82) Industry Effects Yes Yes Year Effects Yes Yes
Observations 2,118 2,118 R-squared 0.037 0.101
a All variables are as defined in Appendix A *,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests). T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).
44
Table 8 Relation between tax lobbying activity and CSR "community" measures – Logit Model (Dependent variable is equal to one if the firm lobbied for tax purposes during the year)
Variablesa Predicted Sign (1) (2)
Community ? 1.094*** 0.273*
(11.60) (1.81)
Size + 1.049***
(11.23)
PTROA + 2.747*
(1.73)
MTB - 0.011
(0.63)
Capital + -0.709
(-0.86)
Constant ? -14.014*** -23.99***
(-10.04) (-24.26)
Year Effects Yes Yes
Industry Effects Yes Yes
Observations 1,999 1,999
R-squared 0.082 0.317 a All variables are as defined in Appendix A *,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests). T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).
45
Table 9 Relation between amount spent on tax lobbying and CSR "community" measure – OLS Model
(Dependent variable is tax lobbying expenditures)
Variablesa Predicted Sign (1) (2)
Community ? 1.438*** 0.693***
(8.72) (3.76)
Size + 0.939***
(13.35)
PTROA + 2.260**
(2.30)
MTB - 0.02
(1.32)
Capital + -0.347
(-0.95)
Constant ? 0.265 -7.107***
(0.52) (-10.65)
Year Effects Yes Yes
Industry Effects Yes Yes
Observations 2,118 2,118
R-squared 0.057 0.188 a All variables are as defined in Appendix A *,**, and *** indicate statistical significance at the 0.10, 0.05 and 0.01 levels, respectively (two-tailed tests). T-Statistics are in parentheses and are calculated based on standard errors that are clustered by firm and year (Peterson 2009).