Financial Constraints and the Incentive for Tax Planning

Alexander Edwards
Rotman School of Management
University of Toronto
alex.edwards@rotman.utoronto.ca

Casey Schwab
Terry College of Business
University of Georgia
cschwab@terry.uga.edu

Terry Shevlin
The Paul Merage School of Business
University of California at Irvine
tshevlin@uci.edu

May 27, 2013

Keywords: Tax planning, Cash taxes, Financial constraints

JEL codes: E69, H25, H60

Data Availability: Data used in this study are available from public sources identified in the
paper.

We appreciate helpful comments and suggestions from the University of Iowa Tax Readings Group, Ben Ayers,
Andy Bauer, Bradley Blaylock, Andy Call, John Campbell, Anne Ehinger, Stacie Laplante, Scott Liao, Dan Lynch,
Devan Mescall, Wayne Nesbitt, Gord Richardson, and Steve Utke. We also thank seminar participants at the
University of Toronto and the 2013 ATA Midyear Meeting. We gratefully acknowledge financial support from the
Rotman School of Management, University of Toronto; the Terry College of Business, University of Georgia; and
the Paul Merage School of Business, University of California at Irvine.
Abstract

In this study, we investigate the association between financial constraints, at both the
macroeconomic and firm-specific level, and one potentially significant source of internal funds
available to firms – cash savings generated through tax planning. Based on capital structure
trade-off theory, as well as archival evidence on the value of cash, we predict that as external
funds become more costly (i.e., financial constraints increase) firms will take actions to increase
internally-generated funds via tax planning. Measuring financial constraints based on both
macroeconomic measures (change in GDP and bank lending tightening) and firm-specific
measures (the decile ranking of both the Altman Z-score and the Whited and Wu 2006 financial
constraint index), we find that firms facing financial constraints exhibit lower cash effective tax
rates. Understanding how financial constraints affect tax avoidance and the interplay between
macroeconomic forces and firm-level tax avoidance behavior is important as legislators look for
ways to reduce the federal deficit.
1. Introduction

In this study, we examine the effect of financial constraints, measured at both the

macroeconomic-level and the firm-level, on corporate tax avoidance behavior. While

understanding the tax effects of financial constraints at the firm-level is important as researchers

attempt to explain tax avoidance behavior (see Shackelford and Shevlin 2001 and Hanlon and

Heitzman 2010 for a comprehensive review), understanding the tax effects of macroeconomic

financial constraints is potentially more important because these constraints impact all firms in

the economy simultaneously. If firms simultaneously increase tax avoidance when faced with

macroeconomic financial constraints, this behavior has the potential to magnify the effect of an

economic contraction on government revenues. Government revenues are likely to decrease due

to both a reduction in the tax base (i.e., lower taxable incomes due to the contraction) and, to a

lesser extent, an increase in corporate tax avoidance activities. As the federal deficit becomes a

key factor in the future financial health of the country, it is important to understand the interplay

between macroeconomic forces and firm-level tax avoidance behavior as legislators look for

ways to reduce the federal deficit.

We define tax avoidance broadly as all actions taken by managers to reduce their cash

income tax liability and measure avoidance using current-year worldwide cash effective tax rates

(cash ETR). We define a firm as more financially constrained if it experiences an increased cost

of external financing or an increased difficulty in accessing desired levels of external funds.

Financial constraints can arise either from a firm being unable to fund all profitable investment

opportunities or a firm that is financially distressed needing funds to cover shortterm working

cash flow obligations.

Based on capital structure trade-off theory (Kraus and Litzenberger 1973; Bradley,

1
Jarrell, and Kim 1984; Fama and French 2002; Frank and Goyal 2008) as well as archival evidence

on the value of cash (Almeida, Campello, and Weisbach 2004; Faulkender and Wang 2006;

Pinkowitz and Williamson 2007), we predict that financially constrained firms will take actions

to increase internally-generated funds via cash tax planning. Trade-off theory suggests a firm sets

a target debt-to-value ratio based on a variety of factors including bankruptcy costs, the tax

benefit of debt, agency costs, and financing costs. Under trade-off theory, the costs (benefits)

associated with these factors increase (decrease) for financially distressed or constrained firms,

creating an increased preference for internally-generated funds over external sources of funds.

The literature on the value of cash holdings provides evidence of a higher valuation of cash and

higher retention of internally-generated funds by financially constrained firms, reinforcing the

increased importance of internally-generated cash for constrained firms.

We anticipate that constrained firms will engage in tax planning as a source of these

internal funds for several reasons. First, unlike many other cost-cutting techniques (e.g., reducing

research and development, advertising, and/or capital expenditures), reducing a firm’s cash taxes

is less likely to adversely impact a firm’s operations. Second, recent research provides evidence

that managers primarily focus on tax strategies that produce both a cash and financial reporting

benefit (i.e., tax strategies giving rise to permanent book tax differences) with only a secondary

interest in strategies that only produce a cash benefit (i.e., deferral strategies giving rise to

temporary book tax differences) (Armstrong, Blouin, and Larcker 2012; Graham, Hanlon,

Shevlin, and Shroff 2012), suggesting that firms likely have not exhausted their opportunities to

generate cash via deferral tax planning strategies. Third, anecdotal evidence suggests that

constrained firms are more likely to use tax planning as a source of cash. Specifically,

practitioners indicate that, during tough economic times, “cash is king, and a lot of companies

2
are receptive to focusing on opportunities in the tax area that they may not have been eager to

focus on in the past” (Leone 2008).

To test our prediction, we first investigate the association between cash tax avoidance and

various macroeconomic and firm-level financial constraints. When examining the effects of

financial constraints arising from macroeconomic economy-wide effects, we use change in GDP

and bank lending tightening as proxies for an increase in cost of external financing or an

increased difficulty in accessing external funds. Both proxies capture investment- and/or distress-

related financial constraints arising from macroeconomic contractions or a change in lending

practices. When examining the effects of financial constraints that vary in the cross-section for

individual firms, we use the decile rank of the Whited and Wu (2006) financial constraint index,

which captures cross-sectional variation in investment-related constraints, and the decile rank of

the Altman Z-score, which captures cross-sectional variation in distress-related constraints.

Consistent with our predictions, firms facing financial constraints arising from either

macroeconomic conditions or firm-specific financial constraints exhibit lower cash ETRs.

Moreover, the reduction in cash ETRs is economically significant. Our results indicate that, on

average, firms’ one year-ahead cash ETRs are approximately 1.5% lower following periods of

macroeconomic financial constraint relative to periods with no macroeconomic constraint and

approximately 2% lower for firms facing higher firm-level financial constraints.

We also investigate the relation between financial constraints and cash tax avoidance in

greater detail by separately analyzing cash tax avoidance attributable to permanent tax planning

strategies and tax deferral strategies. As stated above, recent research (Armstrong, Blouin, and

Larcker 2012; Graham, Hanlon, Shevlin, and Shroff 2012) and anecdotal evidence suggest that

managers tend to focus on tax avoidance that reduces financial statement tax expense (i.e.,

3
permanent strategies) with only a secondary interest in tax avoidance that enhances cash flows

but provides no current financial reporting benefit (i.e., deferral strategies). By analyzing

permanent and deferral strategies separately, we provide insight into whether "cash is king" for

financially constrained firms with their managers engaging in tax avoidance that increases after

tax cash flows regardless of the immediate impact on reported accounting earnings. We find that

financially constrained firms increase tax avoidance via both permanent- and deferral-based tax

avoidance strategies. The increase in deferral is particularly interesting because it indicates that

managers of constrained firms take actions to increase cash flows even if there is no current

financial reporting benefit.

In our analyses discussed above, we exclude loss firms to be consistent with prior

research examining tax planning. Because loss firms are likely an important subset of financially

constrained firms, we conduct supplemental analysis including firms with a loss in the current

year but cumulative profits over the available loss carryback period (i.e., five years for

observations with year ends from 2008 through 2010 and two years for all other observations).

These firms have an incentive to increase their taxable loss to maximize their current refund.

Results from these tests are consistent with our expectations and suggest that financially-

constrained firms with pretax profits or losses increase tax avoidance to generate additional cash

flows by reducing tax payments or increasing refunds, respectively.

Finally, we examine if firms subject to investment-related financial constraints that

engage in tax planning subsequently increase investment.1 Consistent with expectations, firms

facing investment-related financial constraints that increase tax planning (i.e., decrease their cash

1
We perform this analysis using our measure based on the Whited and Wu (2006) financial constraint index. We do
not necessarily expect increased investment using our other measures as firms are less likely to invest during time of
economic contraction and financially distressed firms are more likely to focus on core operations and less likely to
increase investment.

4
ETR) have higher levels of investment (i.e., capital expenditures, research and development, and

acquisitions) in the following period.

The findings of this study are important for several reasons. First, a reduction in taxes

paid during times of financial constraint could have significant consequences on the overall

economy. Specifically, if firms simultaneously increase their tax planning activities when faced

with financial constraints induced by macroeconomic conditions, this behavior has the potential

to magnify the effect of an economic contraction on government revenues. Government revenues

would decrease as a result of both the decrease in the corporate tax base caused directly by the

contraction and an increase in tax avoidance behavior.

The impact of the financial crisis of the late 2000s is consistent with the conjecture

presented above. In the United States, the federal deficit increased during the crisis to

unprecedented levels. Our analysis suggests an average decline of approximately 3% in cash

ETRs due to tax avoidance activities during the macroeconomic contraction that occurred during

2009. This estimated ETR effect equates to approximately a $23 billion decline in corporate tax

revenues and represents approximately 13% of the $182 billion total income taxes after credits

from all profitable C-corporations for 2009.

Second, this study contributes to the literature examining the effects of financial

constraints. Research examining the consequences of financial constraints has become of

increased interest in recent years as firms across the globe have experienced a significant credit

crisis. Examining firms’ responses through the tax account to financial constraints is

advantageous relative to examining responses to financial constraints through other behaviors

because financial accounting rules for income taxes require a relatively comprehensive set of

disclosures. In addition to the total accrued tax expense, firms must identify the portion of the

5
current fiscal period’s tax provision that is being deferred into subsequent periods as well as the

amount of total cash taxes paid during the period. Thus, we can measure with a high degree of

precision how financial constraints affect cash taxes paid by firms.

Finally, we contribute to the growing literature on the determinants of firms’ tax

avoidance behavior and the relative importance of avoiding cash payments of taxes versus

reducing financial accounting tax expense. We consider whether increased financial constraints

trigger an increase in tax avoidance behavior. It is of particular interest that this increase in

financial constraints can be firm-specific or economy-wide. This study is one of the first to

investigate whether macroeconomic forces affect firm-level tax avoidance behavior. This study

also sheds light on the debate related to the relative importance of tax planning activities

generating cash flow savings versus financial reporting benefits. Recent studies provide evidence

suggesting that managers focus primarily on tax planning strategies that reduce total tax expense

(which increases both net cash flows and reported earnings) with only a secondary interest in tax

planning strategies that produce a cash flow benefit but no financial statement benefit

(Armstrong, Blouin, and Larcker 2012; Graham, Hanlon, Shevlin, and Shroff 2012). Our study

provides evidence of specific conditions that increase managers’ focus on cash tax savings.

The remainder of this paper proceeds as follows. The next section develops our

hypotheses relating financial constraints to tax avoidance. Section 3 details the sample, Section 4

describes the research design, Section 5 presents univariate statistics, Section 6 presents our

findings and discusses the economic significance, and Section 7 concludes.

2. Hypothesis Development

In this study, we define both tax avoidance and financial constraint broadly. We define

tax avoidance as all actions taken by managers to reduce their total cash income tax liability. Tax

6
avoidance can include both legal planning strategies in full compliance with tax laws and more

aggressive strategies resulting from aggressive interpretations of ambiguous areas within the tax

laws. Our measures are discussed in greater detail in Section 4 and capture worldwide tax

avoidance. Our aim is not to specifically measure tax aggressiveness, tax risk, tax evasion, or tax

sheltering. Instead, we employ a broad measure that captures a firm’s propensity to reduce its

cash tax burden relative to its reported financial accounting pretax income. This definition is

consistent with prior research and stems primarily from Dyreng, Hanlon, and Maydew (2008).2

We take a broad view of financial constraint and define a firm as more financially

constrained if it experiences an increase in the cost of external financing or an increase in the

difficulty of accessing external funds (Whited and Wu 2006; Denis and Sibilkov 2010). The

financing need and the resulting constraint can arise from a variety of sources. We focus on two

major sources of the financing need: investment opportunities and financial distress. Firms with

unfunded investment opportunities require additional cash in order to finance those additional

investments. Firms in financial distress require additional cash in order to finance existing

operations and remain solvent. For both of these sources of the financing need, the financially

constrained firm needs additional financing through the least costly source available.

We predict that an increase in the external cost of funds (i.e., an increase in financial

constraints) incentivizes a firm to generate additional cash funds internally. Two different lines

of research – the capital structure literature and the value of cash literature – motivate this

prediction. We discuss each literature below and then discuss why we examine this prediction

using a tax planning setting.

2
Unlike Dyreng, Hanlon, and Maydew (2008) who focus on long-run tax avoidance (measured over 5 or 10 years),
we focus on one-year measures of tax avoidance to capture firms’ near-term responses to current financial
constraints. These near-term tax responses may or may not be reflected in long-run measures, which likely blend tax
avoidance during both constrained and unconstrained periods.

7
2.1. Financial Constraints and Capital Structure Theory

A firm facing financial constraints requires additional funds to finance operations. This

increased demand for financing could be satisfied through either internal or external sources.

Researchers have sought to explain how firms finance their operations, developing and testing

numerous theories of capital structure. One prominent theory is trade-off theory. Under trade-off

theory, a firm sets a target debt-to-value ratio and then gradually moves towards the target,

where the target is determined by a variety of factors including (1) the tax benefit of debt, (2)

bankruptcy costs, (3) agency costs, and (4) financing costs.

The tax benefit of debt results from the deductibility of corporate interest payments and

incentivizes firms to rely more heavily on debt. However, based on arguments outlined in

DeAngelo and Masulis (1980), the expected tax benefit of debt is lower for firms with lower

and/or more volatile earnings due to the asymmetric taxation of profits and losses (i.e., the

government does not subsidize losses as heavily as it taxes profits). Expected bankruptcy costs

increase when profitability declines or earnings become more volatile, resulting in firms relying

less on debt as a source of financing. Agency costs result from the misalignment of managers'

and shareholders' incentives, particularly when the firm has free cash flows which managers

prefer to spend on perquisite consumption (Jensen and Meckling 1976; Easterbrook 1984; Jensen

1986). To decrease these agency costs, more profitable firms will increase debt as a disciplining

mechanism to reduce a firm's free cash flows. Of course, issuing debt then introduces agency

costs between debt holders and shareholders. Information asymmetry between managers and

market participants (both debt and equity holders) increases financing costs (i.e., the cost of

obtaining external financing), with debt being more costly than internal cash flows and equity

8
being more costly than debt (Myers 1984; Myers and Majluf 1984).3 To reduce the likelihood of

having to rely on more costly external debt or equity financing or forego profitable investments,

firms with lower expected profits and more volatile cash flows have an increased demand for

internal sources of financing.4

In relation to our study, these factors, which are balanced to determine a firm’s optimal

capital structure, all suggest that a financially constrained or distressed firm is likely to rely more

heavily on internal funds and less on external funds (i.e., debt or equity). Specifically, financially

constrained firms who are generally less profitable and/or have lower free cash flows (1) benefit

less from the tax deductibility of corporate interest payments, (2) face greater bankruptcy costs,

(3) face greater external financing costs, but 4) inherently face lower concerns about free cash

flows (i.e., have lower agency costs). These factors suggest that a financially constrained firm,

which by definition has an increased demand for funds, faces an increased cost (and potentially a

decreased supply) of external funds. The net result is for financially constrained firms to rely

more heavily on internal funds to fund current and near-term future operations.

2.2. Financial Constraints and the Value of Cash

Our expectation that firms facing increased financial constraints will take action to

increase internally-generated cash can also be motivated by the literature examining the value of

cash. Although less theoretical, the argument is intuitive and based on empirical evidence.

3
Another prominent capital structure theory is pecking order theory (Myers and Majluf 1984). Due to information
asymmetry between managers and investors, pecking order suggests a financing hierarchy with internal funds as the
preferred source of financing, followed by debt, followed by equity. Unlike trade-off theory which considers
information asymmetry as one of the costs considered in the trade-off model, the pecking order financing hierarchy
is based on the notion that information asymmetry is the only cost that determines capital structure, with all other
costs playing a secondary role. Recent studies such as Leary and Roberts (2010) suggest that pecking order theory is
significantly more descriptive of firms' financing decisions when other costs (such as those suggested under trade-
off theory) are also considered. These findings are consistent with our focus on trade-off theory which considers
information asymmetry and a variety of other costs that impact firms' financing decisions.
4
See Fama and French (2002) and Frank and Goyal (2008) for excellent summaries of trade-off theory as well as
other theories of capital structure.

9
Studies within this literature generally argue that firms with high costs of external financing (i.e.,

financially constrained firms) can mitigate these costs by relying more on internal financial

resources: cash flows and cash holdings. Consistent with this argument, Almeida, Campello, and

Weisbach (2004) find that firms respond to financial constraints by retaining a higher percentage

of their cash flows to fund future investments. Faulkender and Wang (2006) and Pinkowitz and

Williamson (2007) examine the market valuation of cash for constrained firms and find that cash

holdings are more valuable for constrained firms relative to unconstrained firms. Denis and

Sibilkov (2010) examine why cash is more valuable to constrained firms and find that additional

cash allows investment-constrained firms to undertake value-increasing projects that might

otherwise be bypassed. Recent survey evidence by Campello, Graham, and Harvey (2010)

provides further support for the importance of internally-generated cash to financially

constrained firms. Campello, Graham, and Harvey (2010) find that, when external funding is

unavailable, constrained firms cancel or suspend planned investments, plan deeper cuts in tech

spending, employment levels, and capital spending, sell more assets to fund operations, and pass

on attractive investment opportunities.

Collectively, studies within the value of cash literature provide evidence that financially

constrained firms (1) have incentives to increase internally-generated cash, (2) take actions to

increase and retain internally-generated cash, (3) are rewarded with a higher market valuation of

cash, and (4) appear to use that cash to maximize the value of the firm. This literature, as well as

the capital structure literature, yields the same prediction: Financially constrained firms have an

incentive to increase internally-generated funds. In this study, we examine whether constrained

firms increase their internally-generated cash flows via tax avoidance.

2.3. Tax Planning and Financial Constraints

10
 Corporate income taxes are a non-discretionary expenditure imposed by the government

that all profitable firms must incur. Although income taxes are imposed on all firms at specified

statutory rates, a manager can implement various strategies to reduce the firm’s tax liability. A

growing stream of literature examines the determinants of a firm’s tax avoidance behavior (for a

review of the literature see Shackelford and Shevlin 2001 and Hanlon and Heitzman 2010). In

this study, we examine the impact of financial constraints, measured at both the macroeconomic

level and the firm level, on a firm's tax avoidance behavior. Based on trade-off theory as well as

archival evidence on the value of cash discussed above, we predict that financially constrained

firms will take actions to increase internally-generated funds via cash tax planning. We focus on

cash tax planning as a source of internal funds for the following reasons.

First, a financially constrained firm that attempts to generate additional internal funds

could reduce cash outflows through a variety of options. Most cost-cutting options, such as

reducing advertising, research and development, capital expenditures, administrative costs or

labor costs, often adversely impact the firm's long-term performance. Unlike these cost-cutting

techniques, reducing a firm's cash tax burden is unlikely to have negative long-term

consequences assuming the tax strategies are upheld upon audit. In addition, reducing costs via

tax planning is advantageous because one dollar in tax savings will increase after-tax cash flows

by one dollar whereas reducing most other expenditures by a dollar will only increase after-tax

cash flows by $0.65 ($1.00 x (1-35%)).5

Second, recent research suggests that managers primarily focus on strategies that produce

5
Alternatively stated, a firm can increase after-tax cash flows by $1 by reducing taxable income by $2.85 ($1 / 35%
tax rate) via income shifting or an alternative tax planning strategy as opposed to cutting expenses by $1.54 in
current period spending (i.e., $1 / [1 – 35% tax rate]). These calculations do not include costs associated with
implementing tax planning strategies. Although these costs are not trivial, prior research estimates an average return
of approximately $4 for every $1 invested in general tax planning (Mills, Erickson, and Maydew 1998).

11
both a cash and financial reporting benefit (i.e., permanent tax strategies) with only a secondary

interest in strategies that only produce a cash benefit (i.e., deferral strategies). For example,

Armstrong, Blouin, and Larcker (2012) find a strong negative association between tax director

incentive compensation and GAAP effective tax rates, but no relation between incentive

compensation and cash effective tax rates (i.e., cash tax savings).. Furthermore, Graham, Hanlon,

Shevlin, and Shroff (2012) survey corporate tax executives and report that 47 percent of the

managers in their sample of publicly-traded firms indicate that their firm’s GAAP effective tax

rate is the most important tax metric to top management, whereas only 15 percent state that cash

taxes paid is the most important tax metric. In general, these studies suggest that "cash is not

king" in terms of tax planning for most firms. Anecdotal evidence from practitioners supports

these recent academic findings. Practitioners from large international accounting firms indicate

that profitable public companies have not historically been interested in tax accounting issues

related to timing (i.e., deferral strategies) but instead focus on issues that impact earnings-per-

share (i.e., permanent strategies) (Leone 2008). Collectively, these studies and anecdotes suggest

that firms likely have not exhausted their opportunities to generate cash via tax planning.

Third, anecdotal evidence suggests that financially constrained firms are more likely to

use tax planning as a source of cash. For example, David Auclair, managing principal in Grant

Thornton's national tax office, indicates that during the recent financial crisis "a lot of companies

are receptive to focusing on opportunities in the tax area that they may not have been eager to

focus on in the past" (i.e., before the financial crisis). Mel Schwarz, Grant Thornton's director of

tax legislative affairs, states that "there is nothing like the current economic situation to spur

companies to generate cash." John Salza, a BDO Seidman tax partner and leader of the firm's

national tax accounting methods group, also adds that "with the economy sagging and cash

12
becoming more valuable, even publicly-traded companies are looking at what we call timing

items, cash benefits…even though it doesn't affect their effective tax rate."6 These statements

suggest that financial constraints impact not only a firm's focus on tax planning as a source of

cash but also the nature of the tax strategies that firms will consider (i.e., both permanent and

deferral-based tax strategies). Echoing the likelihood that constrained firms will turn to tax

planning to generate cash, Josephine Feehily, the Chairman of the Office of the Revenue

Commissioners in Ireland, suggested that there could be an increase in tax avoidance as the

economy experiences a recession stating, "In a downturn, money is tighter, so people may be

tempted to evade tax" (McBride 2009).

In sum, we predict that financially constrained firms will take actions to increase

internally-generated funds via cash tax planning because (1) firms prefer reducing tax costs

rather than cutting other operating costs, (2) firms likely have untapped opportunities to save

cash taxes, and (3) anecdotal evidence suggests that constrained firms are indeed more likely to

take advantage of those opportunities as a source of additional cash. Considering macroeconomic

constraints first, we anticipate that financially constrained firms will take actions to reduce cash

taxes paid. Stated formally, we propose the following hypothesis:

H1a: Near-term cash taxes paid are negatively associated with macroeconomic financial
constraints.

Based on the argument presented above, constrained firms have an increased incentive to

reduce cash taxes regardless of whether the source of the constraint is a macroeconomic shock or

a firm-specific event such as the loss of a major customer. As such, we also examine the impact

of firm-specific financial constraints on tax avoidance and we propose the following hypothesis:

6
These quotes are from Leone (2008). Anecdotes from conversations with tax partners at several of the Big 4 firms
present views similar to those discussed in Leone (2008).

13
 H1b: Near-term cash taxes paid are negatively associated with firm-specific financial
constraints.

Although we anticipate a negative association between both macroeconomic and firm-

level financial constraints and cash taxes, there are reasons why we might not observe our

hypothesized relation. For example, implementing tax avoidance strategies often requires an

upfront investment without an immediate benefit. If financially constrained firms simply cannot

afford to implement the tax avoidance strategies (even when the benefits are in the near future),

we will not observe a significant association between constraints and cash taxes. Also, if

potential future losses of constrained firms reduce the expected benefit of engaging in current

period tax avoidance, we might not observe our hypothesized relation. Further, if the financial

constraint is primarily at the parent company, constrained firms with international operations can

access foreign cash holdings to alleviate the constraint regardless of the tax consequences. If

constrained firms have foreign cash available in low tax jurisdictions, additional U.S. taxes could

be due upon repatriation of those foreign earnings. Taxes due upon repatriation would increase

the firm’s cash effective tax rate and reduce the likelihood of observing our hypothesized result.7

2.4. Permanent tax savings vs. tax deferral

To better understand the types of tax planning strategies that managers employ to reduce

cash taxes when facing financial constraints, we decompose tax savings into two components:

tax savings resulting from permanent planning strategies and tax savings from deferral-based

planning strategies. Although both permanent and deferral-based strategies produce cash tax

savings, permanent strategies are potentially more beneficial than deferral strategies because they

(1) result in higher earnings for financial reporting purposes (via a reduction in total tax expense)

7
In supplemental tests we repeat our analysis using only firms that do not report foreign earnings to ensure that any
confounding effect related to repatriations is not impacting our results. Inferences from the observed results under
this specification are unchanged from our main analysis.

14
and (2) produce, dollar for dollar, larger cash savings in present value terms because they do not

reverse in future years (unlike deferral-based strategies). Recent research (see the prior

discussion of Armstrong, Blouin, and Larcker 2012 and Graham, Hanlon, Shevlin, and Shroff

2012) provides evidence suggesting that managers on average focus primarily on tax planning

strategies that reduce both cash taxes and financial statement income tax expense (i.e., permanent

strategies) with only a secondary interest in tax planning strategies that only produce a cash flow

benefit (i.e., deferral strategies).

In our setting, we anticipate that managers of financially constrained firms likely still

have a preference for permanent tax planning strategies because financially constrained firms

have incentives to increase both cash flows and reported earnings. However, because financially

constrained firms have a greatly increased need for cash relative to unconstrained firms and firms

likely have untapped deferral-based tax planning opportunities, we anticipate that they are also

more likely to engage in deferral-based strategies despite the lack of a financial reporting benefit.

To better understand the types of tax planning strategies that managers employ to reduce cash

taxes when facing financial constraints, we propose the following hypotheses:

H2a: Tax savings resulting from permanent tax planning strategies are positively
associated with financial constraints.

H2b: Tax savings resulting from deferral-based tax planning strategies are positively
associated with financial constraints.

3. Sample

To investigate the effect of financial constraints on firms’ tax planning activities, we

begin with firms listed on Compustat from 1987, the first year following the most recent major

tax reform, to 2011. Consistent with prior research, we eliminate financial firms and utility firms.

Firms operating in regulated industries face different financial and tax incentives. Also for

15
consistency with prior research, we initially eliminate firm-year observations with missing or

negative pretax income. However, in supplemental analyses (see Section 6.3) we include loss

firms that can likely benefit from additional tax avoidance in the current period (i.e., firms that

can carry back the losses to obtain a refund of taxes paid in prior years) while continuing to

exclude loss firms that have to carryforward their losses (as they have no additional tax incentive

to increase their tax losses when facing financial constraints). Even after eliminating firms with

negative pretax income for our main analysis, we still observe large variation in our firm-specific

financial constraint measures. Finally, we eliminate firm-years without data to compute our firm-

specific tax planning measures, financial constraint measures, and control variables necessary in

our analyses. Table 1 summarizes our sample selection process.

To investigate the effect of macroeconomic financial constraints we require additional

data on GDP growth and economy-wide lending practices. To retain the largest sample possible,

we impose each of these restrictions separately resulting in samples of 44,328 (37,290) firm-year

observations when using GDP growth (tightening of lending standards) to proxy for

macroeconomic constraints.

To investigate the effect of firm-level financial constraints on firms’ tax planning

activities, we utilize the same primary sample of 44,328 firm-years with the firm-level financial

constraint measures (the decile rank of the Altman Z-score and the rank of the Whited and Wu

investment financial constraint index).

4. Research Design

To examine the relation between financial constraints and tax avoidance, we estimate the

following general regression:

Tax_Avoidancei,t = β0 + β1 Constrainti,t-1 + ΣβkControlsi,t + ε (1)

16
4.1. Tax Avoidance

When testing the relation between tax avoidance and macroeconomic (H1a) and firm-

level (H1b) financial constraints, Tax_Avoidance is set equal to a firm’s cash effective tax rate

(CashETR), measured as the ratio of cash taxes paid to pretax income adjusted for special items

[Compustat data items txpd / (pi – spi)].8,9 Cash taxes vary with firms profitability: more

profitable firms are expected to pay higher taxes. Thus we scale cash taxes by pretax book

income to reflect this relation, which is particularly important in our setting where we examine

macroeconomic downturns which lead to declines in firm profitability. If a firm’s CashETR is

above one (below negative one), CashETR is set equal to one (negative one).10 As discussed by

Hanlon and Heitzman (2010) in their review of the literature on tax research, it is important to

select a tax avoidance measure appropriate to the research question and not simply apply a

laundry list of measures. We utilize cash ETR because we are interested in how firms respond to

financial constraints that result in a need to generate additional cash. A firm’s cash ETR is the

most direct measure of a firm’s cash tax burden. Tax planning that decreases a firm’s cash tax

burden will have a direct impact on a firm’s cash ETR. Unlike prior studies that use long-run

ETR measures to avoid noise in the one-year measure (see Dyreng, Hanlon, and Maydew 2008),

we utilize a one-year cash ETR measure to capture timely responses to existing financial

constraints. There are a number of potential strategies that firms can implement in a relatively

8
Throughout the remainder of the manuscript all variable abbreviations detailing the construction of our measures
refer to Compustat data items unless specifically noted otherwise.
9
Following Dyreng, Hanlon, and Maydew (2008) we adjust pretax income for special items. However, due to
concerns that (1) negative special items may be more likely during a macroeconomic downturn or periods of
financial distress and (2) adjusting the denominator but not the numerator of CashETR may impact our results, we
conduct two robustness tests (untabulated). First, we repeat our analysis using unadjusted pretax income in the
denominator. Second, we repeat our analysis after deleting observations with special items. All results and
inferences hold under both of these alternatives.
10
Much of the prior literature resets CashETR at zero and one. To allow for refunds (i.e., negative cash taxes paid in
the numerator), we reset CashETR at negative one. Our results are qualitatively similar if we reset CashETR at zero
and one.

17
short time period. For example, firms can more aggressively expense instead of capitalize

expenditures, shift income to lower tax jurisdictions, take advantage of programs such as the

domestic manufacturing deduction, and engage in timing strategies which accelerate deductions

and delay income recognition.

We acknowledge a few limitations of this measure. First, CashETR captures non-

conforming tax planning strategies (i.e., strategies that reduce taxable income but not pretax

book income) but does not capture conforming strategies (i.e., strategies that reduce both taxable

income and pretax book income). Second, there is a potential timing mismatch between the

numerator and denominator (i.e., cash payments for taxes may not occur within the same fiscal

period as the income to which they relate). The potential mismatch can result in payments or

refunds from adjacent years being included in our year of interest, a concern that is heightened

for constrained firms if these firms delay tax payments to future years.11 In robustness tests we

use two additional accrual measures, CurrentETR and FederalETR, to help alleviate this

concern. CurrentETR equals current worldwide income tax expense over adjusted pretax income

[txc/(pi-spi)]. FederalETR equals current federal income tax expense over domestic pretax

income [txfed/pidom].

Although it is possible that constrained firms may be more willing to engage in more

aggressive tax planning strategies, a one-year cash ETR measure is still more appropriate than

common measures of aggressive tax planning such as tax sheltering or FIN 48 unrecognized tax

11
In addition to direct changes in tax avoidance, firms could attempt to reduce their estimated payments in order to
save cash. Although this would be a very short term savings and is not the primary mechanism we hypothesize, it
would reduce the cash ETR and increase net cash flows in difficult financial times. Another possible limitation of
our measure relates to tax refunds. Refunds could be associated with our financial constraint variables if firms make
estimated tax payments based on expected higher taxable income and then, following an unexpected macroeconomic
or firm-level shock, report lower taxable income and owe less taxes. Robustness test using the current tax expense
and federal tax expense, both accrual numbers, mitigate this concern. Further mitigating this concern, we find no
consistent correlation between financial constraints and either taxes payable or the presence of a refund
(untabulated).

18
benefit disclosures (UTBs) for several reasons. First, many aggressive tax transactions require

large upfront investments and then yield large tax benefits over a long time horizon. Because

financially constrained firms have immediate cash needs, they are unlikely to select a planning

strategy that requires current cash outflows and yields long-term cash inflows. Consistent with

aggressive transactions such as tax shelters not providing an immediate tax benefit, Wilson

(2009) does not observe an association between cash ETRs and actual sheltering cases. Second,

even ignoring whether constrained firms are likely to select aggressive transactions, both of these

measures are noisy proxies for aggressive tax transactions. Tax shelter prediction scores are

necessarily based on small samples and do not produce a clean measure of tax aggressiveness.

UTBs, while conceptually appealing, are affected by both tax and financial reporting

considerations, creating substantial uncertainty about whether these disclosures capture

aggressive tax planning. Third, regarding UTBs, these disclosures are unavailable during the

bulk of our sample period as FIN 48 is effective for years beginning after December 15, 2006.

For all of these reasons, cash ETR is the most comprehensive and appropriate measure of cash

tax avoidance for our research question.

When testing H2a and H2b, which examine whether any reduction in a firm’s cash ETR

is the result of permanent versus temporary tax planning strategies, Tax_Avoidance is measured

as Permanent or Deferral respectively.12 Permanent equals negative one multiplied by the

difference between the statutory tax rate and the ratio of total tax expense to pretax income

adjusted for special items [-1*(35% - txt / (pi – spi))]. Deferral equals negative one multiplied by

the ratio of deferred tax expense to pretax income adjusted for special items [-1*(txdfed + txdfo)

12
Permanent and Deferral are not meant to be a perfect decomposition of CashETR. Instead, Permanent and
Deferral are simply intended to shed light on two of the primary types of tax planning strategies that can be used to
decrease a firm's CashETR.

19
/ (pi – spi)].13 By multiplying each measure by negative one, lower values of Permanent and

Deferral indicate increased tax avoidance (just as lower values of CashETR indicate increased

tax avoidance). Hypothesis 2a (2b) predicts a negative coefficient on Constraint when

Permanent (Deferral) are used to measure tax avoidance.

4.2. Financial Constraints

The variable of interest in equation (1), Constraint, captures the level of either the

macroeconomic or firm-specific financial constraints faced by the firm. As discussed in the

hypothesis development section, we consider two possible drivers of the financial constraint:

investment opportunities and financial distress. Our measure, Constraint, equals the lagged value

of one of two macroeconomic financial constraint measures (GDP% and Tightening) or the

lagged value of one of two firm-level financial constraint measures (RankZ for distress related

financial constraint and RankWW for investment related financial constraint). We use lagged

values of financial constraints because tax avoidance strategies generally require time to plan and

implement. We examine the sensitivity of our results to this assumption in Section 6.6.

Our first macroeconomic financial constraint measure, GDP%, is based on data from the

Bureau of Economic Analysis. GDP% equals negative one times the percentage change in

inflation-adjusted GDP. We multiply the variable by negative one so that higher values of

GDP% correspond to higher levels of macroeconomic financial constraint. Higher (lower) values

of GDP% are associated with economic contractions (growth) during which external funding is

often more limited (more abundant) and firms are more (less) likely to face financial constraints.

Our second macroeconomic financial constraint measure, Tightening, is based on data from the

Federal Reserve Board’s Senior Loan Officer Opinion Survey. Tightening equals the average net

13
We substitute txdi for (txdfed + txdfo) when (txdfed + txdfo) is missing.

20
percentage of domestic respondents reporting tightening standards in year t-1 for commercial and

industrial loans. Positive (negative) values of Tightening correspond with a tightening

(loosening) of lending standards and an increase (decrease) in macroeconomic financial

constraints. The tightening of lending standards directly captures the ease with which firms can

obtain credit financing.14

Our first firm-specific measure, RankZ, is based on the firm’s Altman (1968) bankruptcy

prediction Z-score. Firms closer to bankruptcy face greater difficulty accessing external funds.

To reduce noise in this measure, we use the decile rank of the firm’s Z-score in year t-1 where

higher values indicate they are more likely to be financially distressed.15 RankWW equals the

decile ranking of a firm’s investment related financial constraint index developed in Whited and

Wu (2006). RankWW is coded so that higher values represent higher financial constraints.16 If

financial constraints increase tax planning behavior, we anticipate negative coefficients on

Constraint across our various specifications.

4.3. Controls

In addition to our explanatory variable of interest, Constraint, we include a battery of

control variables in our regression model. Many tax deductions, exclusions, and tax credits are

not perfectly correlated with economic activity of the firm but are instead fixed over a relevant

14
For brevity we focus on these two measures of macroeconomic financial constraint. In untabulated robustness
tests we repeat our analysis using a measure of economic contraction based on NBER data that equals the number of
months in year t-1 that the economy is contracting divided by 12, and a measure of equity market frothiness that
equals the inverse of IPO volume in year t-1 based on data from Jay Ritter’s website. Results are qualitatively and
quantitatively similar to those reported in the main analysis using these alternative proxies.
15
Many studies use the Z-score to construct a financial distress indicator equal to one when the Z-score is below 1.8
and zero otherwise. Our results are robust to replacing RankZ with the financial distress indicator.
16
There are a number of firm-specific financial constraint measures utilized in the literature. For brevity, we focus on
a subset of these measures but repeat our analysis with a number of alternative, commonly used measures.
Alternative measures include (1) an indicator variable equal to one for firms that pay no dividends in year t-1 and
zero otherwise, (2) an indicator variable equal to one for firms with a Standard & Poor's (S&P) domestic long-term
issuer credit rating in year t-1 consistent with non-investment grade debt and zero otherwise, and (3) the decile rank
of a firm’s likelihood of bankruptcy based on Shumway (2001). Results are qualitatively and quantitatively similar
to those reported in the main analysis using these alternative proxies.

21
range (Zimmerman 1983; Wilkie 1988). For example, a firm can often increase or decrease

production without purchasing additional property, plant and equipment such that the

depreciation tax deduction is fixed and does not vary directly with the activity. We refer to this

relation as the fixed tax shield effect. To address this “nonlinearity” we include both an intercept

and pretax return on assets, PROA (pi/at), as an additional explanatory variable. We expect a

positive coefficient on PROA because a decrease (increase) in PROA as a proxy for economic

activity will decrease (increase) the cash ETR due to the fixed tax shield effect. Additionally we

include industry fixed effects which allow the intercept to vary across industries and we identify

the association between tax avoidance and financial constraints by within industry variation

rather than across industry variation (since cross-industry variation, such as fixed asset intensity,

might reflect differences in the fixed tax shield effects).17

We control for firm size by including Size, measured as the natural log of total assets

[ln(at)]. Large firms are widely viewed as more sophisticated and can structure complex tax-

reduction transactions with the help of the best tax advisors (Mills, Erickson, and Maydew 1998;

Hanlon, Mills, and Slemrod 2007). If large firms conduct more successful tax planning, then Size

will be negatively related to tax costs. On the other hand, large, mature firms would have fewer

tax shields and hence higher ETRs as their capital investment slows. Because size can capture

many effects, we make no directional prediction.

17
We do not explicitly control for the various tax brackets in our main analysis. Although financially constrained
firms likely have lower taxable income than non-constrained firms, U.S. federal corporate taxes rates are at or near
35% for all income above $75,000 (the rate varies between 34% and 39% between $75,000 and $18,333,333 and is
35% for taxable income above $18,333,333) resulting in a relatively flat tax rate. In untabulated robustness tests, we
include indicator variables (together and separately) for pretax income below $75,000 and pretax income below
$18,333,333. Results from these specifications are qualitatively and quantitatively similar to those reported in the
paper. In a similar vein, it is unlikely our results are driven by overpayments or refunds that could result from large
changes in taxable income that could accompany financial constraint. We assume that firms, particularly firms
facing cash shortages during financial constraint, will adjust estimated payments in a timely matter and avoid
overpayments in order to conserve cash in the near term.

22
 An extensive literature establishes that corporate taxpayers respond to tax incentives to

place income in lower taxed jurisdictions (see Klassen and Laplante 2012 for recent evidence).

Because segment disclosures no longer require foreign asset disclosures, we use the absolute

value of the ratio of foreign pretax income to worldwide pretax income (pifo/pi) as a proxy for

foreign operations (Foreign). Following Mills and Newberry (2004), we set foreign pretax

income equal to zero for firm years where it is missing.

To control for the existing capital structure of the firm, we include Leverage which

equals the ratio of long-term debt to assets [(dltt+dlc)/at]. Prior literature argues that debt

provides an important tax shield (Graham 1996) and provides evidence of a negative relation

between leverage and marginal tax rates. However, interest deductions decrease income for both

tax and financial reporting purposes making it unclear whether leverage would necessarily

decrease effective tax rates. Further, additional debt would increase cash outflows in the form of

interest payments by a greater amount than the tax savings the interest expense would generate.

Newberry and Dhaliwal (2001) argue that multinationals can place debt in high-tax locations and

thus reduce their effective tax rates. Firms can also structure off-balance-sheet financing to

maximize interest deductions without decreasing book income (Mills and Newberry 2004) or can

structure debt to use foreign tax credits (Newberry 1998). Collectively, these studies suggest that

debt is negatively associated with effective tax rates. Finally, both of our firm-specific measures

of financial constraints have measures of leverage as inputs. The inclusion of leverage as a

control variable in our model helps ensure that the relation between financial constraints and tax

avoidance we document is incremental to the effect of leverage.

We include CapEx, the ratio of capital expenditures to total assets (capx/at), as a proxy

for tax planning opportunity. Governments often use tax policy to spur economic investment,

23
especially during economic downturns. Consistent with legislated tax shields, capital-intensive

firms have lower tax burdens (Gupta and Newberry 1997), higher book-tax differences (Mills

and Newberry 2001; Wilson 2009; Lisowsky 2010), and higher IRS deficiencies (Rice 1992;

Mills 1998). Furthermore, many tax shelters during the 1990s involved long-lived capital assets

(McGill and Outslay 2004). In sum, we expect that CapEx will be negatively related to cash

ETRs.18

We include R&D, the ratio of research and development expense to revenues (xrd/sale),

to control for intellectual property. Intellectual property, such as patents and brand intangibles,

increases opportunities to decrease taxes via income shifting. Additionally the R&D tax credit

reduces cash ETRs. As such, we expect R&D to be negatively related to cash ETRs.

To control for aggressive financial reporting practices, we include DiscAccruals.

DiscAccruals equals performance-matched discretionary accruals as measured in Kothari, Leone,

and Wasley (2005). If firms that exhibit more aggressive financial reporting practices are more

tax aggressive as suggested by Frank, Lynch, and Rego (2009), then we would expect a negative

association between DiscAccruals and tax rates.19

We include an indicator variable (NOL) to control for the presence of net operating loss

carryforwards (i.e., a non-zero value for tlcf). Consistent with Chen, Chen, Cheng, and Shevlin

(2010), we expect that NOL firms have lower tax rates because they are less profitable and are

able to utilize the loss carryforwards to reduce taxable income and thus cash taxes.20

We also include the book-to-market ratio, BM, to control for firm growth. Growth firms

often have substantial tax deferral opportunities and also often rely heavily on stock-based

18
Results are robust to defining CapEx as the ratio of net property, plant, and equipment to total assets (ppent/at).
19
Results are robust to the exclusion of the DiscAccruals control variable.
20
Also note that financially distressed firms are more likely to have NOL carryforwards. By including NOLs
separately, we are examining the effect of the financial distress constraint incremental to the effects of NOLs.

24
compensation, both of which decrease CashETRs. Finally, because tax subsidies are often unique

to particular industries, we include industry-fixed effects (Barth, Beaver, and Landsman 1998).21

5. Descriptive Statistics and Correlations

Table 2 presents descriptive statistics for the tax avoidance, financial constraint, and

control variables used in this study. Similar to evidence documented in prior studies (e.g.,

Dyreng, Hanlon, and Maydew 2008), cash effective tax rates are substantially lower than the

statutory rate of 35%. Interestingly, our sample firms report a reduction in taxes achieved

through permanent tax savings (Permanent) of approximately 3.9% on average and a reduction

in taxes achieved by deferring tax payments (Deferral) of approximately 1.6%.

Focusing first on the macroeconomic constraint measures, GDP grows at an average rate

of approximately 2.8% per year. The annual mean net tightening of lending standards

(Tightening) is positive 6.5%. Focusing on the firm-level constraint measures, we present

descriptive statistics for the underlying measures but use decile ranks in the regression models.

The interquartile range is 2.46 to 5.71 for the underlying Altman (1968) Z-score and -0.36 to -

0.22 for the underlying Whited and Wu (2006) financial constraint index. Collectively, these

descriptive statistics indicate substantial variation in financial constraints at both the

macroeconomic and firm level during our sample period.

Table 3 presents correlations between our tax and financial constraint measures, with

Pearson correlations above the diagonal and Spearman rank correlations below the diagonal. As

expected, CashETR is positively correlated with our Permanent and Deferral measures.

Consistent with H1a, CashETR is negatively correlated with GDP% and Tightening. Consistent

21
We do not include year fixed effects as our proxies for macroeconomic financial constraints are cross-sectional
(i.e., within year) constants. However, our results are robust to including year fixed effects when using firm-specific
measures of financial constraints (e.g., RankZ and RankWW).

25
with H1b, CashETR is also negatively correlated with RankZ and RankWW. These correlations

provide support for the assertion that financially constrained firms exhibit increased tax

avoidance. The macroeconomic financial constraints are positively correlated with one another.

It is interesting to note that RankWW is negatively correlated with RankZ (ρ = -0.09). This result

is not entirely surprising given the different aspects of financial constraint that these measures

are intended to capture (i.e., constraints due to financial distress and bankruptcy concerns versus

insufficient capital to fund investment opportunities).22

6. Multivariate Results

6.1. Financial Constraints and Cash ETRs

Panel A of Table 4 presents the results from estimating equation (1) using CashETR as

the dependent variable and macroeconomic measures of financial constraints. The negative and

significant coefficients on GDP% and Tightening are consistent with H1a. To assess the

economic significance of a macroeconomic increase in financial constraints, we focus on each

measure separately. An increase of one standard deviation in GDP% is accompanied by a 1.43%

decrease in a firm’s CashETR (-0.8251*0.0173), which represents an average reduction of $3.67

million in cash taxes and an increase of 1.33% of operating cash flows for sample firms. 23 An

increase of one standard deviation in Tightening is accompanied by a 1.44% decrease in a firm’s

CashETR (-0.0007*20.632), which represents an average reduction of $3.70 million in cash taxes

and an increase of 1.34% of operating cash flows for our sample firms. These reductions in

22
Insignificant or even negative correlations across measures of financial constraint are not uncommon. In fact,
Whited and Wu (2006, p.545) report that their measure is negatively correlated (ρ = -0.019) with another common
measure of financial constraint developed by Kaplan and Zingales (1997).
23
Calculations related to the cash tax reductions and the ratio of cash savings to operating cash flows for the
macroeconomic measures of financial constraint follow:
GDP%: $3.67 million cash tax benefit = -1.43%*$256.71 (mean adjusted pretax income). 1.33% of operating cash
flows = $3.67 million/$276.73 million (mean operating cash flows).
Tightening: $3.70 million cash tax benefit = -1.44%*$256.71 (mean adjusted pretax income). 1.34% of operating
cash flows = $3.70 million/$276.73 million (mean operating cash flows).

26
CashETR are economically significant, and are after controlling for the level of economic

activity on cash taxes by scaling cash taxes by pretax book income and including pretax ROA as

an additional explanatory variable.

Panel B of Table 4 presents the results from estimating equation (1) using CashETR as

the dependent variable and firm-specific measures of financial constraints. The coefficients on

both RankZ and RankWW are negative and significant. These results are consistent with H1b and

suggest that financially-constrained firms increase tax avoidance to generate additional cash. The

magnitude of the effect is also economically significant. For the ranked measure of financial

distress, RankZ, a one decile increase in the independent variable is associated with a decrease in

CashETR of 1.82, which represents an average reduction of $4.67 million in cash taxes and an

increase of 1.69% of operating cash flows for our sample firms.24 For the measure of financial

constraint from Whited and Wu (2006), RankWW, a one decile increase in the independent

variable is associated with a decrease in CashETR of 2.02%, which represents an average

reduction of $5.19 million in cash taxes and an increase of 1.87% of operating cash flows for our

sample firms. In the interests of brevity, we do not discuss in detail the results for the control

variables: many exhibit significant coefficients in directions consistent with prior literature.

6.2. Financial Constraints and Tax Planning Via Permanent and Deferral Based Strategies

In Table 5, we focus on whether the tax savings documented above result from reducing

tax expense via permanent tax savings and/or deferral. Column 1 and 2 of Panel A presents the

results examining the effect of macroeconomic financial constraints on firms’ permanent tax
24
Calculations related to the cash tax reductions and the ratio of cash savings to operating cash flows for the firm-
level measures of financial constraint follow:
RankZ: $4.67 million cash tax benefit = -1.82%*$256.71 (mean adjusted pretax income). 1.69% of operating cash
flows = $4.67 million/$276.73 million (mean operating cash flows).
RankWW: $5.19 million cash tax benefit = -2.02%*$256.71 (mean adjusted pretax income). 1.87% of operating cash
flows = $5.19 million/$276.73 million (mean operating cash flows).

27
savings (Permanent). Consistent with H2a, the coefficients on GDP% and Tightening are both

negative and significant (p-values of < 0.0001 and 0.0939 respectively). These results suggest

that financially constrained firms increase tax avoidance via permanent tax planning strategies.

Column 3 and 4 present the results examining the effect of macroeconomic financial constraints

on firm’s tax savings through deferral strategies (Deferral). Consistent with H2b, the coefficients

on GDP% and Tightening are both negative and highly significant (p-values of < 0.0001). These

results suggest that firms respond to macroeconomic financial constraints by also engaging in

deferral-based tax planning strategies.

Panel B repeats these analyses using firm-specific measures of financial constraints. In

column 1 and 2, consistent with H2a, the coefficients on RankZ and RankWW are negative and

highly significant (p-values < 0.0001), suggesting that firms also respond to firm-specific

financial constraints by engaging in tax planning strategies that result in permanent tax savings.

Column 3 and 4 present the results examining the effect of firm-specific financial constraints on

firm’s tax savings through deferral. Consistent with H2b, the coefficients on RankZ and

RankWW are negative and significant (p-values of < 0.0001 and 0.0010 respectively), suggesting

that financially constrained firms increase tax deferral strategies. These results suggest that firms

use a broad range of tax strategies to reduce cash taxes when faced with financial constraints.

To examine the relative importance of tax planning strategies that result in permanent tax

savings versus deferral-based tax planning strategies, we use seemingly unrelated regression

analysis (untabulated) to compare the magnitudes of the estimated coefficients on the various

financial constraints measures when Permanent is the dependent variable to the analogous

coefficients on the various financial constraint measures when Deferral is the dependent

variable. Focusing first on the macroeconomic measures of financial constraints, we observe

28
mixed evidence. The coefficient on GDP% is significantly more negative when the dependent

variable is Permanent, suggesting a greater increase in permanent tax strategies relative to

deferral tax strategies. The coefficient on Tightening is significantly more negative when

Deferral is the dependent variable, suggesting a greater increase in deferral tax strategies relative

to permanent tax strategies.

Focusing next on the firm-level measures of financial constraints, the coefficients on

RankZ and RankWW are significantly more negative when the dependent variable is Permanent

providing some evidence that managers facing firm-specific financial constraints focus more on

tax planning strategies that result in permanent tax savings.

Overall, it appears that for macroeconomic induced financial constraints, managers

implement tax planning strategies that achieve savings via both permanent and deferral tax

strategies without a strong propensity to implement one method over the other. For firm-level

induced financial constraints, managers achieve tax savings through both deferral and permanent

tax planning strategies but implement permanent tax saving strategies to a greater extent. Given

that financially constrained firms have incentives to increase cash flows as well as earnings for

financial reporting purposes, it is not entirely surprising that more often than not the evidence

suggests that firms attempt to increase tax avoidance via permanent tax planning strategies. In

light of prior research suggesting that managers generally favor tax avoidance strategies that

increase cash flows and lower ETRs (permanent strategies), the general increase in tax savings

via deferral strategies is nevertheless interesting.

6.3. Loss Firms

We exclude loss firms in our main analysis to be consistent with prior research examining

tax planning. Because loss firms are likely an important subset of financially constrained firms,

29
we conduct supplemental analysis including firms with a loss in the current year but cumulative

profits over the current and prior two years (prior five years for 2008-2010). In general, a firm is

able to carry a current year loss back to the prior two years and receive a refund of taxes

previously paid.25 As such, these firms have an incentive to increase their taxable loss to

maximize their current refund. Consistent with this notion, prior research has documented

intertemporal income shifting by firms to increase NOLs and cash refunds (Maydew 1997;

Erickson, Heitzman, and Zhang 2012). In this analysis, we re-estimate equation (1) after adding a

loss indicator (Loss) set equal to one when adjusted pretax income is negative (pi-spi<0) and an

interaction between Loss and Constraint. These additional variables allow us to examine the

relation between financial constraints and tax planning for loss firms.

There are four possible combinations of cash taxes and pretax book income when

including both profit and loss firms. We discuss profitable firms first. When a firm reports

positive pretax book income and positive cash taxes paid (i.e., CashETR>0), we expect a

negative coefficient on Constraint. This result would be consistent with our earlier hypotheses

and indicate that these constrained firms take actions to reduce their positive tax liability. When a

firm reports positive pretax book income and negative cash taxes (i.e., CashETR<0), we expect a

negative coefficient on Constraint. This result would be consistent with our earlier hypotheses

and indicate that these constrained firms take actions to increase their tax refund. Because the

coefficient on Constraint captures the relationship between financial constraints and cash ETRs

for profitable firms, these patterns suggest a negative coefficient on Constraint.

25
Firms with adjusted pretax book income less than zero must have cumulative pretax income over the prior two
years to be included. We extend this window from two to five years for 2008 through 2010 to correspond with the
legislated increase in the loss carryback period during that time. Firms with current losses that must be carried
forward cannot lower their current period cash taxes as they are already zero and cannot increase their cash refunds
because they cannot carry back. Thus we continue to exclude these firms.

30
 Next we consider loss firms. When a firm reports negative pretax book income and

positive cash taxes (i.e., CashETR<0 and Loss=1), we expect a positive coefficient on the

Loss*Constraint interaction. This result would indicate that these constrained firms take actions

to reduce their positive tax liability. When a firm reports negative pretax book income and

negative cash taxes (i.e., CashETR>0 and Loss=1), we expect a positive coefficient on the

Loss*Constraint interaction. This result would suggest that these constrained firms try to

increase their refund. These predictions imply an expected positive coefficient on the

Loss*Constraint term and on the sum of the coefficients on Constraint and Loss*Constraint (as

the sum represents the overall constraint coefficient estimate for the loss firms). The following

figure summarizes the predicted coefficients on Constraint and the Loss*Constraint interaction.

Expected Coefficient on Constraint when CashETR is the Dependent Variable

Positive cash taxes Negative cash taxes (refund)

Positive pretax book income CashETR>0 and Loss=0 CashETR<0 and Loss=0

NEGATIVE coefficient NEGATIVE coefficient

(constrained firm reduces cash (constrained firm increases
taxes paid) cash refund)

Negative pretax book income CashETR<0 and Loss=1 CashETR>0 and Loss=1

POSITIVE coefficient POSITIVE coefficient

(constrained firm reduces cash (constrained firm increases
taxes paid) cash refund)

Panel A of Table 6 presents the results from estimating equation (1) using the

macroeconomic measures of financial constraints when including loss firms and the variables

Loss and Loss*Constraint. The coefficients on Constraint and the Constraint*Loss interaction

are consistent with expectations regardless of whether GDP% or Tightening is used to measure

macroeconomic constraints. Specifically, we find negative and significant coefficients on

31
Constraint (p-values of < 0.0001) and positive and significant coefficients on the

Loss*Constraint interaction (p-values of < 0.0001). Also consistent with expectations, the sum of

the Constraint and Loss*Constraint coefficients are positive and significant.

Panel B of Table 6 presents the results from estimating equation (1) using the firm-

specific measures of financial constraints when including loss firms and the variables Loss and

Loss*Constraint. The coefficients on Constraint and the Constraint*Loss interaction are

consistent with expectations regardless of whether RankZ or RankWW is used to measure firm-

level constraints. Specifically, we find negative and significant coefficients on Constraint (p-

values of < 0.0001) and positive and significant coefficients on the Loss*Constraint interaction

(p-values of < 0.0001 and 0.0540 respectively). However, the sum of the Constraint and

Loss*Constraint coefficients are not significantly positive using the firm-level constraint

measures.

These results are generally consistent with our expectations and suggest that financially-

constrained firms with pretax profits or losses increase tax avoidance to generate additional cash

flows by reducing tax payments or increasing refunds. 26

6.4. Alternative Measures of Tax Planning

In this subsection we examine the impact of financial constraints on tax planning

activities using two alternative measures of tax avoidance, CurrentETR and FederalETR.

CurrentETR is defined as the ratio of current tax expense to pretax income adjusted for special

items [txc / (pi – spi)].27 FederalETR is defined as the ratio of current federal income taxes to

26
In untabulated robustness tests, we repeat this analysis including only loss firms to ensure that the substantial
portion of the sample with positive pretax income is not driving the result. The results are consistent with those
reported in the tables.
27
As with CashETR, for both CurrentETR and FederalETR, if a firm’s ETR is above one (below negative one), the
measure is set equal to one (negative one).

32
domestic pretax income (txfed/pidom). Although CashETR is a more direct measure of our

construct of interest, cash savings through tax planning, there is a potential timing mismatch

between the numerator and denominator (i.e., cash payments for taxes may not occur within the

same fiscal period as the income to which they relate). The potential mismatch can result in

payments or refunds from adjacent years being included in our year of interest, a concern that is

heightened for constrained firms if these firms delay tax payments to future years. By using the

accounting accrual (both current tax expense and federal tax expense) we are able to ensure that

our results are not driven by attempts to generate short term savings by manipulating the amount

and timing of estimated tax payments (although even this strategy could be viewed as a tax

avoidance strategy to reduce short term cash outflows).

Table 7, Panel A presents the results using CurrentETR to proxy for tax planning. The

coefficients on all constraint measures are negative and significant, consistent with our main

findings. These results confirm that our earlier results are not due to timing mismatches between

the recognition of pretax book income and the payment of cash taxes.

Although CashETR is a better proxy for a firm’s worldwide cash tax burden, FederalETR

is the best proxy available for a firm’s U.S. cash tax burden. As such, FederalETR can be used to

provide more direct evidence regarding domestic tax revenues. Table 7, Panel B presents the

results using FederalETR to proxy for tax planning. The coefficients on all constraint measures

are negative and significant, consistent with constrained firms engaging in greater levels of

federal tax planning. These results are consistent with our main findings and provide additional

evidence that tax planning during times of financial constraint reduces U.S. tax revenues.28

28
Some caution should be used when interpreting the revenue effect of the constraint coefficients in the FederalETR
regressions because FederalETR is an accrual number and does not capture cash flows directly.

33
6.5. Future Investment, Financial Constraint, and Tax Planning

Our results for firms subject to investment-related financial constraints (i.e., higher values

of RankWW) naturally leads to the following question: Does increased tax planning increase

future investment? To address these questions, we regress investment (i.e., capital expenditures,

research and development, and acquisitions in year t) on the lagged values of our Constraint

variable, the change in CashETR from t-1 to t, and the interaction of lagged Constraint and

change in CashETR. If increased tax avoidance relieves the constraint, we expect a negative

coefficient on the interaction term. Alternatively stated, we expect that decreases (increases) in

CashETR are associated with higher (lower) investment in the next period for constrained firms

if increased tax planning relaxes the constraint. We expect this relation to hold when we measure

Constraint using RankWW, which captures financing frictions.

We present the results from estimating this regression in column 1 of Table 8. Consistent

with expectations we observe a negative and significant coefficient (p-value = 0.0437) on the

LagConstraint*∆CashETR interaction. In column 2, we repeat this analysis including control

variables for other determinants of investment (discretionary accruals, size, leverage, book to

market, tangibility, operating cycle, Tobin’s Q, age, and a dividend indicator).29 The coefficient

on the LagConstraint*∆CashETR interaction remains negative and significant, although the

result is weaker (p-value = 0.0772).

In untabulated tests, we repeat our investment analysis using RankZ, GDP%, and

Tightening to measure Constraint. We do not necessarily expect increased investment for

distressed firms (i.e., high values of RankZ) that increase tax planning. These firms are struggling

to remain solvent and likely use the increased tax savings for many uses (e.g., to pay creditors)

29
See appendix A for variable definitions.

34
other than increasing investment. Similarly, we do not necessarily expect increased investment

for firms that increase tax planning during periods of economic contraction (GDP%) or

tightening of lending standards (Tightening). During macroeconomic economic contractions and

periods of tightening of lending standards (which also capture macroeconomic trends), firms

often retain additional cash in light of increased economic uncertainty. Evidence from this is

apparent from firms' record-high cash balances over the last several years during the financial

crisis (Pinkowitz, Stultz, and Williamson 2013). Not surprisingly, we do not observe significant

negative coefficients on the interaction of these lagged Constraint measures and change in

CashETR. These results demonstrate that the increase in investment only occurs within the

subsample of the population that is subject to investment constraints, not firms that are subject to

financial distress or macroeconomic constraints.

6.6. Additional Sensitivity Analyses

In this section we discuss a series of sensitivity tests (untabulated). Our first sensitivity

test examines the role that foreign operations play in our overall results. Multinational firms

(MNCs) face a potentially different set of tradeoffs when making decisions regarding financial

constraints and tax planning. MNCs can face constraints in the U.S., abroad, or globally (i.e.,

both in the U.S. and abroad). A MNC’s response to financial constraint will vary depending on

the source of the constraint. For example, a U.S. MNC could relieve a financial constraint at the

U.S. parent level by repatriating foreign profits held abroad. The repatriation would likely trigger

U.S. taxation on the foreign earnings (taxed at 35% less any available foreign tax credit) but

would result in a net inflow of cash to the U.S. parent. This type of behavior biases against us

finding a negative association between financial constraints and CashETR because repatriations

can result in an increase in cash taxes paid but not affect worldwide income, thereby increasing

35
the firm’s CashETR. Alternatively, it is possible that large MNCs are better able to increase their

tax avoidance behavior during periods of financial constraint through activities such as income

shifting. This behavior could result in drastically different outcomes depending upon the location

of the constraint. For U.S. firms constrained domestically, there is an incentive to shift income

into the U.S. from foreign locations. This could result in an increase in U.S. taxes paid (and

likely total taxes paid given the high statutory U.S. rate) but would alleviate the constraint. In

contrast, a firm facing financial constraints abroad could attempt to shift income out of the U.S.

into foreign locations resulting in a reduction in U.S. taxation but potentially an increase in

foreign taxes (and an uncertain impact on worldwide taxes). Due to data limitations it is difficult

to measure financial constraints and cash taxes at a country level. To ensure that our earlier

results are not solely driven by multinational firms, we perform several additional tests. First,

because our macroeconomic measures capture the economic conditions in the U.S., and our firm

specific measures capture worldwide constraints that are likely driven by constraints at the U.S.

parent company, we attempt to capture tax planning in the U.S. using the FederalETR measure

described in Section 6.5. As previously noted, results using FederalETR are consistent with our

hypotheses. Second, we repeat our analysis on a subsample of our firms that do not report any

foreign pretax income and, as a result, do not suffer from potential jurisdictional mismatching of

financial constraints and tax planning proxies. Inferences from the results using this sample of

domestic firms are largely unchanged from the main analysis.30

Our second sensitivity test examines the relation between cash tax avoidance and
30
One particular complication related to multinationals is the choice to designate foreign earnings as permanently
reinvested (PRE). When a firm designates foreign earnings as PRE, the firm does not need to record the associated
deferred tax expense related to repatriating those earnings. This designation would increase our Deferral measure
without actually reducing current period tax payments. In our supplemental analysis of firms that do not report any
foreign pretax income, we continue to observe a negative and significant coefficient on all four measures of
constraint when Deferral is the dependent variable. This gives comfort that the PRE designation is not driving our
main Deferral results.

36
financial constraints using a changes analysis. The changes analysis controls for any correlated

omitted variables that are stationary through time. In this analysis, we use the annual change in

CashETR as our dependent variable and include changes (rather than levels) of our control

variables. For the constraint measures we use the change in the firm specific measures, RankZ

and RankWW, as our test variables but continue to use the macroeconomic measures, GDP% and

Tightening, as previously defined since these are already measured in changes (i.e., the annual

percentage change in GDP and the change in lending standards). Consistent with our hypotheses,

we continue to observe significant negative coefficients on the constraint measures using this

specification.

We also perform a sensitivity test that examines the relation between cash tax avoidance

and both macroeconomic and firm-specific financial constraints simultaneously. To do this, we

include both a macroeconomic constraint measure and a firm-specific constraint measure in our

regression model concurrently. We find significant negative coefficients on both measures

indicating that firms’ cash tax avoidance behavior is incrementally related to both firm specific

and economy wide financial constraints.

Our next two sensitivity tests focus on when financial constraints are measured. Because

it is possible that tax planning strategies undertaken by financially constrained firms take longer

than a year to implement, we first repeat our analysis using the two year lag of our constraint

measures (i.e., we measure Constraint in year t-2) and observe results that are qualitatively and

quantitatively similar to those reported in the main analysis.

In our second timing sensitivity test, we measure Constraint in year t. Our original

research design, which measures Constraint in year t-1, implicitly assumes that managers do not

anticipate that their firms will become financially constrained. The insider trading literature,

37
however, provides evidence that managers do understand how profitable their firms will be for

several years into the future (see, for example, Piotroski and Roulstone 2004). If managers are

able to anticipate impending financial constraints, they will likely take actions to increase cash

flows in the periods of financial constraint. This foresight by managers is more likely to apply to

firm-specific financial constraints than macroeconomic constraints (as that would require

foresight into macroeconomic cycles which, as the financial crisis indicates, is difficult). We

repeat our analysis using both firm-specific constraints and CashETR in year t and observe

significant negative coefficients on our firm-specific measures of financial constraint. These

results are consistent with managerial foresight for firm-specific financial constraints. When we

repeat our analysis using both macroeconomic constraints and CashETR in year t, we observe

significant negative coefficients on only one of the two macroeconomic measures. These results

support the notion that managers have greater foresight for future firm conditions but less

foresight for macroeconomic conditions.

An additional potential concern with our research design is that our observed findings,

particularly in the case of the macroeconomic tests, are the result of government tax relief during

times of economic hardship. The most prominent example of such relief during our sample

period is the bonus depreciation provisions that were implemented during the 2000s.31 To control

for this possibility we create an indicator variable Bonus equal to one for years with bonus

depreciation and zero otherwise. We then interact Bonus with capital expenditures (as bonus

depreciation only applies to firms that make new capital expenditures during those time

31
The U.S. Government instituted bonus depreciation tax relief through various legislation covering assets
purchased from September 10, 2001 through December 31, 2004; and January 1, 2008 through December 31, 2010.

38
periods).32 This approach effectively allows the impact of capital expenditures on cash taxes to

vary in years with bonus depreciation in place. Regression results obtained after including the

Bonus*CapEx interaction term are qualitatively and quantitatively similar to those reported in the

main analysis. When performing this sensitivity test, we do not include the Bonus main effect.

Because bonus depreciation rules are enacted during periods of macroeconomic financial

constraint, including Bonus in equation (1) effectively results in two financial constraint

measures, one indicator (Bonus) and one continuous measure (Constraint), being simultaneously

included in the regression model. In fact, the correlations between our macroeconomic financial

constraint proxies and the bonus depreciation indicator are all greater than 0.35. As an alternative

test we drop Constraint and include the Bonus main effect as our measure of financial constraint.

Doing so results in a negative and significant coefficient on Bonus which is consistent with the

hypothesis that firms decrease CashETRs in times of financial constraint.

6.7. Impact of tax planning on corporate tax revenues during the financial crisis

In this section, we illustrate the impact of tax planning on government revenues during

2009, the height of the financial crisis. Based on our regression results, cash ETRs would be 3%

lower. Using the tax base from all profitable C-corporations in 2008 from the IRS Statistics of

Income Tax Stats website, this would represent a reduction in government revenue of $24 billion

or 13% of total income taxes after credits from all profitable C-corporations.33

32
Note, the Compustat data item capex does not capture assets purchases that are eligible for bonus depreciation
perfectly (for example, capital expenditures by foreign subsidiaries do not quality). Unfortunately we are not aware
of an alternative publicly available proxy that would capture this more precisely. We repeated this analysis
excluding firms with foreign operations to mitigate this concern. Results are consistent with those in the broader
sample.
33
Cash ETR change is calculated by multiplying the regression coefficient on GDP% by the change in real GDP
during 2009 (-0.8251 x -0.035 = 3%). Data in this subsection are from the IRS Statistics of Income Tax Stats
website (SOI), available at http://www.irs.gov/taxstats/article/0,,id=170728,00.html. The revenue reduction of $24
billion is calculated as 3% x $816 billion 2009 tax base. The 13% is calculated as $24 billion divided by $182 billion
income taxes after credits from all profitable C-corporations.

39
 The discussion and calculations above are based on the regression estimates obtained

from our tests. As a simple check on the reasonableness of our estimate, we obtain an alternative

estimate of the effect of tax planning using the raw SOI data from the IRS. From 2008 to 2009,

the tax base – total income subject to tax – for all profitable C-corporations fell at a rate of 10%

from $910 billion to $816 billion, a $94 billion decrease. At the same time total income tax after

credits fell at an even faster rate of 12% from $207 billion to $182 billion. The change in the tax

base of $94 billion is likely primarily due to the deterioration of macroeconomic conditions

during the financial crisis and, possibly to a lesser extent, due to an increase in tax planning by

firms. We calculate an effective tax rate using the SOI data by dividing “Total income tax after

credits” by “Income subject to tax.” In 2009, the ETR decreased by 1% from 23% ($207/910) in

2008 to 22% ($182/$816) in 2009. The change in effective tax rate of 1% is likely attributable to

an increase in tax planning by firms as financial resources became more scarce. In our analysis

we control for a number of other factors that could also impact the tax planning activities of

firms but the estimate we obtain from our regression models, 3%, is comparable to the

unadjusted rate reduction of 1% observed in the IRS data.

7. Conclusion

The extant academic literature examining the determinants of firms’ tax planning

activities focus almost exclusively on firm-level characteristics (for a review of the literature see

Shackelford and Shevlin 2001 and Hanlon and Heitzman 2010) and more recently on the impact

of specific executives (Dyreng, Hanlon, and Maydew 2010; Armstrong, Blouin, and Larcker

2012). In this study, we extend this literature and are one of the first to investigate whether

macroeconomic factors also affect firm-level tax avoidance behavior. The factor of interest in

this study, financial constraints, can impact a firm as a result of its own circumstances or as a

40
result of macroeconomic conditions.

In this study, we take a broad view of financial constraint and define a firm as more

financially constrained if it experiences an increase in the cost of external financing or an

increase in the difficulty of accessing external funds. The financing need and the resulting

constraint can arise from a variety of sources. We focus on two major sources of the financing

need: investment opportunities and financial distress. Firms with unfunded investment

opportunities require additional cash in order to finance those additional investments. Firms in

financial distress require additional cash in order to finance existing operations and remain

solvent. For both of these sources of financing need, the financially constrained firm needs

additional financing through the least costly source available. As such, financially constrained

firms are likely to search for new sources of internal funds. In this study, we investigate the

association between financial constraints, at both the macroeconomic and firm-specific level, and

a significant source of internal funds available to firms – cash tax savings, or more generally tax

avoidance.

We document results consistent with our predictions, firms facing either macroeconomic

financial constraints or firm-specific financial constraints exhibit lower near-term cash ETRs.

The reduction in cash ETR is also economically significant. Based on macroeconomic measures

of constraint (change in GDP and bank lending tightening) cash ETRs are approximately 2%

lower following periods of financial constraint. Based on firm-level measures of constraint (the

decile rank of both the Altman Z-score and Whited and Wu 2006 financial constraint index),

cash ETRs are approximately 2% lower for financially constrained firms.

In addition to documenting an association between both macroeconomic and firm-level

financial constraints and near-term cash ETRs, we examine whether the decrease in cash ETRs is

41
the result of permanent and/or deferral-based tax planning strategies. We find firms increase the

level of tax planning via both permanent and deferral-based strategies following periods of

financial constraint. Several recent studies have documented evidence consistent with managers

focusing on tax avoidance behavior that first and foremost reduces the tax expense accrued on

the financial statements with less concern on the actual cash flow savings (or deferral) generated

by the tax strategies (e.g., Graham, Hanlon, Shevlin, and Shroff 2012). Our study provides

evidence that there are some instances that increase managers’ focus on cash tax savings

regardless of the potential financial reporting benefits.

The findings of this study are potentially of interest for several reasons. First, we

contribute to the growing literature on the determinants of firms’ tax avoidance behavior.

Second, this study also contributes to the broad literature on financial constraints. Third, this

study informs the debate related to the relative importance of tax planning activities generating

cash flow savings versus financial reporting benefits. Finally, given the record high level of the

federal deficit and current economic conditions, it is important to understand the interplay

between macroeconomic forces and firm-level tax avoidance behavior as legislators look for

ways to reduce the federal deficit.

42
References

Altman, E. 1968. Financial Ratios, Discriminant Analysis and the Prediction of Corporate
Bankruptcy. Journal of Finance 23(4): 589–609.

Almeida, H., M. Campello, and M. Weisbach. 2004. The Cash Flow Sensitivity of Cash. Journal
of Finance 59(4): 1777-1804.

Armstrong, C., J. Blouin, and D. Larcker. 2012. The Incentives for Tax Planning. Journal of
Accounting and Economics 53(1-2): 391-411.

Barth, M. E., W. Beaver, and W. Landsman. 1998. Relative Valuation Roles of Equity Book
Value and Net Income as a Function of Financial Health. Journal of Accounting and Economics
25(1): 1-34.

Bradley, M., G.A. Jarrell and E. H. Kim, 1984. On the Existence of an Optimal Capital
Structure: Theory and Evidence. Journal of Finance 39: 857-877.

Campello, M., J. Graham, and C. Harvey. 2010. The Real Effects of Financial Constraints:
Evidence from a Financial Crisis. Journal of Financial Economics, 97(3): 470-487.

Chen, S., X. Chen, Q. Cheng, and T. Shevlin. 2010. Are Family Firms More Tax Aggressive
than Non-Family Firms? Journal of Financial Economics. 95(1): 41-61.

DeAngelo, H. and R. Masulis. 1980. Optimal Capital Structure under Corporate and Personal
Taxation. Journal of Financial Economics. 8: 3-29.

Denis, J. and V. Sibilkov. 2010. Financial Constraints, Investment, and the Value of Cash
Holdings. The Review of Financial Studies. 23: 247-269.

Dyreng, S., M. Hanlon, and E. Maydew. 2008. Long-run Corporate Tax Avoidance. The
Accounting Review. 83(1): 61-82.

Dyreng, S., M. Hanlon, and E. Maydew. 2010. The Effects of Executives on Corporate Tax
Avoidance. The Accounting Review. 85(4): 1163-1189.

Easterbrook, F. 1984. Two Agency-Cost Explanations of Dividends. The American Economic

Review. 74(4): 650-659.

Erickson, M., S. Heitzman, and F. Zhang. 2012. Tax-Motivated Loss Shifting. Working Paper,
University of Chicago.

Fama, E. and K. French. 2002. Testing Trade-off and Pecking Order Predictions about Dividends
and Debt. Review of Financial Studies 15(1): 1-33.

43
Faulkender, M. and R. Wang. 2006. Corporate Financial Policy and the Value of Cash. Journal
of Finance 61(4): 1957-1990.

Frank, M. and V. Goyal. 2008. Trade-off and Pecking Order Theories of Debt. Handbook of
Corporate Finance: Empirical Corporate Finance 2: 135-202.

Frank, M. M., L. Lynch, and S. Rego. 2009. Tax Reporting Aggressiveness and Its Relation to
Aggressive Financial Reporting. The Accounting Review 84(2): 467-496.

Graham, J., 1996. Proxies for the Corporate Marginal Tax Rate. Journal of Financial Economics.
42(2): 187-221.

Graham, J., M. Hanlon, T. Shevlin, and N. Shroff. 2012. Incentives for Tax Planning and
Avoidance: Evidence from the Field. Working Paper, MIT.

Gupta, S. and K. Newberry. 1997. Determinants of the Variability in Corporate Effective Tax
Rates: Evidence from Longitudinal Data. Journal of Accounting and Public Policy 16:1-34.

Hanlon, M. and S. Heitzman. 2010. A Review of Tax Research. Journal of Accounting and
Economics. 50(2-3):127-178

Hanlon, M., L. Mills, and J. Slemrod. 2007. An Empirical Examination of Corporate Tax
Noncompliance, in Taxing Corporate Income in the 21st Century, A. Auerbach, J.R. Hines Jr.,
and J. Slemrod (eds.). Cambridge: Cambridge University Press, pages 171-210.

Jensen, M. 1986. Agency Cost of Free Cash Flow, Corporate Finance, and Takeovers. American
Economic Review 76(2): 323-329.

Jensen M. and W. Meckling. 1976. Theory of the Firm: Managerial Behavior, Agency Costs and
Ownership Structure. Journal of Financial Economics 3(4): 305-360.

Kaplan, S. and L. Zingales. 1997. Do Investment-Cash Flow Sensitivities Provide Useful

Measures of Financing Constraints? Quarterly Journal of Economics, 112(1): 169-215.

Klassen K. and S. Laplante. 2012. Are U.S. Multinational Corporations Becoming More
Aggressive Income Shifters? Journal of Accounting Research, 50(5): 1245-1285.

Kraus, A., and R.H. Litzenberger, 1973. A State-Preference Model of Optimal Financial
Leverage, Journal of Finance 33, 911-922.

Kothari, S.P., A. Leone, and C. Wasley. 2005. Performance Matched Discretionary Accrual
Measures. Journal of Accounting and Economics 39(1): 163-197.

Leary, M. and M. Roberts. 2010. The Pecking Order, Debt Capacity, and Information
Asymmetry. Journal of Financial Economics 95(3): 332-355.

44
Leone, M. 2008. There’s a Cash Machine in Your Tax Department. CFO.com December 11,
2008.

Lisowsky, P. 2010. Seeking Shelter: Empirically Modeling Tax Shelters Using Financial
Statement Information. The Accounting Review 85(5): 1693-1720.

Maydew, E. 1997. Tax-Induced Earnings Management by Firms with Net Operating Losses.
Journal of Accounting Research. 35(1): 83-96.

McBride, L. 2009. Revenue Chief Fears Steep Rise in Tax Evasion. Independent.ie April 26,
2009.

McGill, G. and E. Outslay. 2004. Lost in Translation: Detecting Tax Shelter Activity in Financial
Statements. National Tax Journal 57(3). 739-756.

Mills, L. 1998. Book-Tax Differences and Internal Revenue Service Adjustments. Journal of
Accounting Research. 36(2): 343-356.

Mills, L., M. Erickson, and E. Maydew. 1998. Investments in Tax Planning. Journal of the
American Taxation Association. 20(1): 1-20.

Mills, L. and K. Newberry. 2001. The Influence of Tax and Non-Tax Costs on Book-Tax
Reporting Differences: Public and Private Firms. The Journal of the American Taxation
Association, 23(1): 1-19.

Mills, L. and K. Newberry. 2004. Do Foreign Multinationals’ Tax Incentives Influence Their
U.S. Income Reporting and Debt Policy? National Tax Journal 57(1): 89-107.

Myers, S. 1984. The Capital Structure Puzzle. Journal of Finance 39(3): 574-592.

Myers, S. and N. Majluf. 1984. Corporate Financing and Investment Decisions when Firms have
Information that Investors do not have. Journal of Financial Economics 13(2): 187-221.

Pinkowitz, L., R. Stulz., and R. Williamson. 2013. Is There a High Cash Holdings Puzzle after
the Financial Crisis? Working Paper, Georgetown University.

Pinkowitz, L. and R. Williamson. 2007. What is the Market Value of a Dollar of Corporate
Cash? Journal of Applied Corporate Finance 19(3): 74-81.

Piotroski J. and D. Roulstone. 2004. The Influence of Analysts, Institutional Investors, and
Insiders on the Incorporation of Market, Industry, and Firm-Specific Information into Stock
Prices. The Accounting Review. 15(2): 81-102.

Newberry, K. 1998. Foreign Tax Credit Limitations and Capital Structure Decisions. Journal of
Accounting Research 36(1): 157-66.

45
Newberry, K. and D. Dhaliwal. 2001. Cross-Jurisdictional Income Shifting by U.S.
Multinationals: Evidence from International Bond Offerings. Journal of Accounting Research 39
(3): 643-662.

Rice, E. 1992. The Corporate Tax Gap: Evidence on Tax Compliance by Small Corporations. In
Why People Pay Taxes, ed. J. Slemrod, 125-61. Ann Arbor: University of Michigan Press. 1992.

Shackelford, D. and T. Shevlin. 2001. Empirical Tax Research in Accounting. Journal of

Accounting and Economics. 31(1-3): 321–387.

Shumway, T. 2001. Forecasting Bankruptcy More Accurately: A Simple Hazard Model. The
Journal of Business. 74(1): 101-124.

Whited, T. and G. Wu. 2006. Financial Constraints Risk. Review of Financial Studies. 19(2):
531-559.

Wilkie, P. 1988. Corporate Average Effective Tax Rates and Inferences About Relative Tax
Preferences. The Journal of the American Taxation Association. 10(1): 75-88.

Wilson, R. 2009. An Examination of Corporate Tax Shelter Participants. The Accounting

Review 84(3): 969-999.

Zimmerman, J. 1983. Taxes and Firm Size. Journal of Accounting and Economics 5: 119-149.

46
 Appendix A
Variable definitions and construction
Variable Definition
Tax measures
CashETR the ratio of cash taxes paid to pretax income adjusted for special items (txpd/(pi – spi))
Permanent negative one multiplied by the U.S. statutory rate less the ratio of total tax expense to pretax
income adjusted for special items [-1*(35% - txt/(pi – spi))]
Deferral negative one multiplied by the ratio of deferred tax expense to pretax income adjusted for
special items [-1*(txdfed + txdfo)/(pi – spi); if missing (txdfed + txdfo), then (-1*txdi/(pi –
spi)]
CurrentETR the ratio of the current income taxes to pretax income adjusted for special items [txc/(pi -
spi)]
FederalETR the ratio of the current federal income taxes to domestic pretax income (txfed/pidom)

Financial constraints
GDP% negative one multiplied by the annual percentage change in GDP (measured in chained 2005
dollars) measured at the end of firm i’s fiscal year t. This data is from the Bureau of
Economic Analysis (http://www.bea.gov/national/index.htm#gdp)
Tightening the average net percentage of domestic respondents to the Federal Reserve Board’s Senior
Loan Officer Opinion Survey reporting tightening standards in year t-1 for commercial and
industrial loans. Tightening is the percentage reporting tightening minus the percentage
reporting loosening. Positive (negative) values indicate tightening (loosening) of standards
Z-Score the Altman (1968) Z-score calculated as 3.3*[(pi +xint)/at] + 1.2*(wcap/at) + (sale/at) +
1.4*(re/at) + 0.6*[(chso*prcc_f)/lt]
WW the Whited and Wu (2006) financial constraints index calculated as −0.091*cfo − 0.062*(1 if
dv>0, 0 otherwise) + 0.021 *(dltt/at) − 0.044[ln(at)] − 0.035*[(salet-salet-1)/salet-1)]
RankZ the decile rank of the firm’s Altman (1968) Z-score in year t-1
RankWW the decile rank of the firm’s Whited and Wu (2006) level of financial constraints in year t-1

Tax planning opportunities

PROA the ratio of pretax income adjusted for special items to total assets (pi-spi/at)
Size the natural logarithm of total assets (at)
Foreign the absolute value of the ratio of pretax foreign income to total pretax income (|pifo/pi|); if
missing pretax foreign income, foreign pretax income is set equal to zero
Leverage the ratio of total debt to total assets ((dltt+dlc)/at)
CapEx the ratio of capital expenditures to total assets (capx/at)
R&D the ratio of R&D expense to total revenues (xrd/sale)
DiscAccruals discretionary accruals computed at the end of the year estimated using a modified Jones'
model (Kothari, Leone, and Wasley 2005)
NOL an indicator variable if the firm reported a net operating loss (tlcf) in any of the last three
years and zero otherwise
BM the ratio of the book value of equity to the market value of equity (ceq/(csho*prcc_f))
Loss an indicator variable equal to one if adjusted pretax income is negative [(pi – spi)<0] and zero
otherwise
Bonus an indicator variable equal to one for years with bonus depreciation and zero otherwise. The
U.S. Government instituted bonus depreciation tax relief through various legislation covering
investment assets purchased between from September 10, 2001 through December 31, 2004;
and January 1, 2008 through December 31, 2010

Additional investment analysis control variables

MB the ratio of the market value of equity to the book value of equity ((csho*prcc_f)/ceq)
Tangibility the ratio of fixed assets to total assets (ppent/at)

47
OPC the natural logarithm of the operating cycle measured as the sum of receivables turnover and
inventory turnover divided by 360 (((rect/sale)+(invt/cogs))/360)
Q Tobin’s Q ((prcc_f*csho – ceq + at )/at)
Age the natural logarithm of the firm’s age (fyear-year1)
Dividend an indicator variable if the firm paid dividends during the year (dv>0) and zero otherwise

48
 Table 1
Sample Selection
Data Restrictions N
Starting Compustat sample of non-regulated firms reporting positive pretax income 93,348
Less firms without data necessary to compute tax measures:
Firms missing data necessary to compute tax measures -25,390
Less firms missing data necessary to compute control variables -12,402
Less firms missing data necessary to compute firm-level financial constraint measures: -11,228
Primary firm-level financial constraint sample 44,328

Samples with additional financial constraint restrictions:

Sample with GDP% 44,328
Sample with Tightening 37,290
Detailed variable definitions are presented in Appendix A.

49
 Table 2
Descriptive Statistics
Variable N Mean StdDev P10 Q1 Median Q3 P90
Tax Measures
CashETR 44,328 0.2613 0.2428 0.0160 0.1154 0.2610 0.3721 0.4950
Permanent 44,328 -0.0386 0.2042 -0.2578 -0.0858 0.0001 0.0399 0.0815
Deferral 44,328 -0.0160 0.2130 -0.2203 -0.0835 -0.0061 0.0349 0.1315

Macroeconomic Financial Constraint Measures

GDP% 44,328 -0.0275 0.0173 -0.0446 -0.0399 -0.0307 -0.0201 -0.0066
Tightening 37,290 6.476 20.632 -14.500 -7.975 -0.638 18.188 40.425

Firm-Level Financial Constraint Measures

Z-Score 44,328 5.3122 8.9165 1.5961 2.4612 3.7005 5.7177 9.6515
WW 44,328 -0.2948 0.3096 -0.4311 -0.3630 -0.2866 -0.2162 -0.1586

Control Variables
PROA 44,328 0.1060 0.0720 0.0272 0.0534 0.0909 0.1428 0.2047
Size 44,328 5.6816 1.9670 3.1587 4.2429 5.5923 7.0190 8.3351
Foreign 44,328 0.1543 0.2808 0.0000 0.0000 0.0000 0.1817 0.6126
Leverage 44,328 0.2088 0.1748 0.0000 0.0486 0.1905 0.3247 0.4476
CapEx 44,328 0.0636 0.0594 0.0129 0.0244 0.0458 0.0809 0.1341
R&D 44,328 0.0275 0.0508 0.0000 0.0000 0.0000 0.0305 0.0997
DiscAccr 44,328 -0.0029 0.1375 -0.1532 -0.0655 -0.0011 0.0636 0.1476
NOL 44,328 0.2812 0.4496 0.0000 0.0000 0.0000 1.0000 1.0000
BM 44,328 0.5958 0.4358 0.1804 0.3029 0.4907 0.7694 1.1326
Detailed variable definitions are presented in Appendix A. All continuous variables are winsorised at the 1 and 99 percent
levels.

50
 Table 3
Correlation Table
Variable (1) (2) (3) (4) (5) (6) (7)
(1) CashETR 1.0000 0.1899 0.2705 -0.0509 -0.0639 -0.1351 -0.0909
(<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001)

(2) Permanent 0.3245 1.0000 -0.4051 -0.0560 -0.0012 -0.0492 -0.0693

(<0.0001) (<0.0001) (<0.0001) (<0.0001) (0.8121) (<0.0001) (<0.0001)

(3) Deferral 0.3029 -0.2305 1.0000 -0.0054 -0.0432 -0.1009 0.0356

(<0.0001) (<0.0001) (<0.0001) (0.2450) (<0.0001) (<0.0001) (<0.0001)

(4) GDP% -0.0595 -0.0838 -0.0135 1.0000 0.6214 0.0486 -0.0763

(<0.0001) (<0.0001) (0.0035) (<0.0001) (<0.0001) (<0.0001) (<0.0001)

(5) Tightening -0.0634 -0.0128 -0.0649 0.4268 1.0000 0.0847 -0.0530

(<0.0001) (0.0113) (<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001)

(6) RankZ -0.2041 -0.0575 -0.1743 0.0323 0.0754 1.0000 -0.0919

(<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001)

(7) RankWW -0.1278 -0.0170 0.0595 -0.0612 -0.0359 -0.0948 1.0000

(<0.0001) (0.0002) (<0.0001) (<0.0001) (<0.0001) (<0.0001) (<0.0001)
Pearson (Spearman) correlations are presented above (below) the diagonal. Detailed variable definitions are
presented in Appendix A. All continuous variables are winsorized at the 1st and 99th percentiles.

51
 Table 4
The Relationship Between Financial Constraints and Tax Planning
Panel A: Macroeconomic Financial Constraints
CashETR
GDP% Tightening
Coeff. Coeff.
Variable Pred. (p-value) (p-value)
Intercept ? 0.2277 0.2343
(<0.0001) (<0.0001)
Constraint - -0.8251 -0.0007
(<0.0001) (<0.0001)
PROA + 0.2009 0.2395
(<0.0001) (<0.0001)
Size ? 0.0025 0.0028
(0.0071) (0.0057)
Foreign ? 0.0666 0.0588
(<0.0001) (<0.0001)
Leverage - -0.0611 -0.0623
(<0.0001) (<0.0001)
CapEx - -0.0105 -0.0168
(0.3400) (0.2741)
R&D - -0.3487 -0.3547
(<0.0001) (<0.0001)
DiscAccruals - 0.0507 0.0385
(<0.0001) (<0.0001)
NOL - -0.0736 -0.0652
(<0.0001) (<0.0001)
BM ? 0.0308 0.0330
(<0.0001) (<0.0001)

Industry FE Yes Yes

Adj R2 0.0530 0.0534

N 44,328 37,290
Detailed variable definitions are presented in Appendix A. All continuous variables are
winsorised at the 1 and 99 percent levels. Regression models include untabulated industry
fixed effects. Huber-White robust standard errors clustered by firm are used to control for
heteroscedasticity and serial correlation.

52
 Table 4 (continued)
The Relationship Between Financial Constraints and Tax Planning
Panel B: Firm-level Financial Constraints
CashETR
RankZ RankWW
Coeff. Coeff.
Variable Pred. (p-value) (p-value)
Intercept ? 0.3057 0.4856
(<0.0001) (<0.0001)
Constraint - -0.0182 -0.0202
(<0.0001) (<0.0001)
PROA + 0.0195 0.1514
(0.4380) (<0.0001)
Size ? 0.0037 -0.0242
(<0.0001) (<0.0001)
Foreign ? 0.0705 0.0642
(<0.0001) (<0.0001)
Leverage - 0.0775 -0.0459
(<0.0001) (<0.0001)
CapEx - -0.0546 -0.0176
(0.0153) (0.2455)
R&D - -0.4499 -0.2691
(<0.0001) (<0.0001)
DiscAccruals - 0.0274 0.0564
(0.0007) (<0.0001)
NOL - -0.0663 -0.0717
(<0.0001) (<0.0001)
BM ? 0.0469 0.0325
(<0.0001) (<0.0001)

Industry FE Yes Yes

Adj R2 0.0723 0.0600

N 44,328 44,328
Detailed variable definitions are presented in Appendix A. All continuous variables are
winsorised at the 1 and 99 percent levels. Regression models include untabulated industry
fixed effects. Huber-White robust standard errors clustered by firm are used to control for
heteroscedasticity and serial correlation.

53
 Table 5
The Relationship Between Financial Constraints and Tax Planning
Panel A: Macroeconomic Financial Constraints
Permanent Deferral
GDP% Tightening GDP% Tightening
Coeff. Coeff. Coeff. Coeff.
Variable Pred. (p-value) (p-value) (p-value) (p-value)
Intercept ? -0.1764 -0.1958 0.0856 0.1169
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Constraint - -0.5022 -0.0001 -0.2508 -0.0005
(<0.0001) (0.0939) (<0.0001) (<0.0001)
PROA + 0.5770 0.5993 -0.2397 -0.2311
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Size ? 0.0078 0.0102 -0.0084 -0.0087
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Foreign ? -0.0217 -0.0373 0.0826 0.0887
(0.0010) (<0.0001) (<0.0001) (<0.0001)
Leverage - 0.0684 0.0525 -0.1140 -0.1123
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
CapEx - 0.0043 0.0257 -0.2150 -0.2191
(0.4188) (0.1406) (<0.0001) (<0.0001)
R&D - -0.3415 -0.3473 0.1874 0.1935
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
DiscAccruals - -0.0349 -0.0404 -0.0016 0.0040
(<0.0001) (<0.0001) (0.4166) (0.3211)
NOL - -0.0400 -0.0406 -0.0205 -0.0189
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
BM ? 0.0582 0.0612 -0.0425 -0.0511
(<0.0001) (<0.0001) (<0.0001) (<0.0001)

Industry FE Yes Yes Yes Yes

Adj R2 0.0793 0.0829 0.0532 0.0606

N 44,328 37,290 44,328 37,290
Detailed variable definitions are presented in Appendix A. All continuous variables are winsorised at the 1
and 99 percent levels. Regression models include untabulated industry fixed effects. Huber-White robust
standard errors clustered by firm are used to control for heteroscedasticity and serial correlation.

54
 Table 5 (continued)
The Relationship Between Financial Constraints and Tax Planning
Panel B: Firm-level Financial Constraints
Permanent Deferral
RankZ RankWW RankZ RankWW
Coeff. Coeff. Coeff. Coeff.
Variable Pred. (p-value) (p-value) (p-value) (p-value)
Intercept ? -0.1444 -0.0650 0.1096 0.1267
(<0.0001) (0.0006) (<0.0001) (<0.0001)
Constraint - -0.0059 -0.0084 -0.0056 -0.0029
(<0.0001) (<0.0001) (<0.0001) (0.0010)
PROA + 0.5213 0.5583 -0.2961 -0.2454
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Size ? 0.0080 -0.0034 -0.0080 -0.0123
(<0.0001) (0.0203) (<0.0001) (<0.0001)
Foreign ? -0.0212 -0.0232 0.0839 0.0819
(0.0015) (0.0005) (<0.0001) (<0.0001)
Leverage - 0.1154 0.0763 -0.0711 -0.1105
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
CapEx - -0.0028 0.0062 -0.2288 -0.2121
(0.4469) (0.3840) (<0.0001) (<0.0001)
R&D - -0.3728 -0.3078 0.1560 0.1996
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
DiscAccruals - -0.0425 -0.0326 -0.0088 -0.0008
(<0.0001) (<0.0001) (0.1240) (0.4574)
NOL - -0.0385 -0.0398 -0.0182 -0.0207
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
BM ? 0.0637 0.0591 -0.0375 -0.0420
(<0.0001) (<0.0001) (<0.0001) (<0.0001)

Industry FE Yes Yes Yes Yes

Adj R2 0.0809 0.0801 0.0556 0.0531

N 44,328 44,328 44,328 44,328
Detailed variable definitions are presented in Appendix A. All continuous variables are winsorised at the 1
and 99 percent levels. Regression models include untabulated industry fixed effects. Huber-White robust
standard errors clustered by firm are used to control for heteroscedasticity and serial correlation.

55
 Table 6
The Relationship Between Financial Constraints and Tax Planning
Panel A: Macroeconomic Financial Constraints; Profit and Loss Firms Included
CashETR
GDP% Tightening
Coeff. Coeff.
Variable Pred. (p-value) (p-value)
Intercept ? 0.2460 0.2517
(<0.0001) (<0.0001)
Constraint - -0.8599 -0.0008
(<0.0001) (<0.0001)
Loss*Constraint + 2.4388 0.0035
(<0.0001) (<0.0001)
PROA + 0.0263 0.0825
(0.3009) (0.0025)
Loss ? -0.4115 -0.4944
(<0.0001) (<0.0001)
Size ? 0.0020 0.0022
(0.0268) (0.0240)
Foreign ? 0.0561 0.0496
(<0.0001) (<0.0001)
Leverage - -0.0771 -0.0766
(<0.0001) (<0.0001)
CapEx - -0.0051 -0.0072
(0.4209) (0.3967)
R&D - -0.2816 -0.2862
(<0.0001) (<0.0001)
DiscAccruals - 0.0404 0.0290
(<0.0001) (0.0017)
NOL - -0.0703 -0.0632
(<0.0001) (<0.0001)
BM ? 0.0175 0.0209
(<0.0001) (<0.0001)

Yes Yes

Adj R2 0.1966 0.1926

N 47,006 39,451
Detailed variable definitions are presented in Appendix A. All continuous variables are
winsorised at the 1 and 99 percent levels. Regression models include untabulated industry
fixed effects. Huber-White robust standard errors clustered by firm are used to control for
heteroscedasticity and serial correlation.

56
 Table 6 (continued)
The Relationship Between Financial Constraints and Tax Planning
Panel B:Firm-level Financial Constraints; Profit and Loss Firms Included
CashETR
RankZ RankWW
Coeff. Coeff.
Variable Pred. (p-value) (p-value)
Intercept ? 0.3426 0.5062
(<0.0001) (<0.0001)
Constraint - -0.0201 -0.0204
(<0.0001) (<0.0001)
Loss*Constraint + 0.0190 0.0059
(<0.0001) (0.0540)
PROA + -0.2242 -0.0274
(<0.0001) (0.2838)
Loss ? -0.6062 -0.5200
(<0.0001) (<0.0001)
Size ? 0.0034 -0.0244
(0.0001) (<0.0001)
Foreign ? 0.0607 0.0526
(<0.0001) (<0.0001)
Leverage - 0.0568 -0.0641
(<0.0001) (<0.0001)
CapEx - -0.0533 -0.0143
(0.0168) (0.2872)
R&D - -0.3707 -0.2065
(<0.0001) (<0.0001)
DiscAccruals - 0.0172 0.0464
(0.0279) (<0.0001)
NOL - -0.0631 -0.0684
(<0.0001) (<0.0001)
BM ? 0.0293 0.0188
(<0.0001) (<0.0001)

Yes Yes

Adj R2 0.2129 0.2012

N 47,006 47,006
Detailed variable definitions are presented in Appendix A. All continuous variables are
winsorised at the 1 and 99 percent levels. Regression models include untabulated industry
fixed effects. Huber-White robust standard errors clustered by firm are used to control for
heteroscedasticity and serial correlation.

57
 Table 7
The Relationship Between Financial Constraints and Tax Planning: Alternative Measures of Cash Tax
Planning
Panel A: Using Current ETR to proxy for tax planning
CurrentETR
GDP% Tightening RankZ RankWW
Coeff. Coeff. Coeff. Coeff.
Variable Pred. (p-value) (p-value) (p-value) (p-value)
Intercept ? 0.2373 0.2604 0.2956 0.3833
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Constraint - -0.8090 -0.0007 -0.0125 -0.0107
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
PROA + 0.4074 0.4188 0.2840 0.3850
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Size ? 0.0017 0.0019 0.0024 -0.0127
(0.0718) (0.0641) (0.0125) (<0.0001)
Foreign ? 0.0540 0.0470 0.0560 0.0515
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Leverage - -0.0664 -0.0744 0.0313 -0.0547
(<0.0001) (<0.0001) (0.0036) (<0.0001)
CapEx - -0.2077 -0.1974 -0.2289 -0.1994
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
R&D - -0.1939 -0.1924 -0.2625 -0.1498
(<0.0001) (<0.0001) (<0.0001) (0.0002)
DiscAccruals - -0.0360 -0.0348 -0.0515 -0.0331
(<0.0001) (0.0001) (<0.0001) (<0.0001)
NOL - -0.0645 -0.0596 -0.0603 -0.0649
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
BM ? 0.0084 0.0023 0.0196 0.0099
(0.0905) (0.6643) (<0.0001) (0.0450)
0.2373 0.2604 0.2956 0.3833
Industry Fixed Effects Yes Yes Yes Yes

Adj R2 0.0676 0.0696 0.0767 0.0672

N 40,313 34,008 40,313 40,313
Detailed variable definitions are presented in Appendix A. All continuous variables are winsorised at the 1 and
99 percent levels. Regression models include untabulated industry fixed effects. Huber-White robust standard
errors clustered by firm are used to control for heteroscedasticity and serial correlation.

58
 Table 7 (continued)
The Relationship Between Financial Constraints and Tax Planning: Alternative Measures of Cash Tax
Planning
Panel B: Using Federal ETR to proxy for tax planning
FederalETR
GDP% Tightening RankZ RankWW
Coeff. Coeff. Coeff. Coeff.
Variable Pred. (p-value) (p-value) (p-value) (p-value)
Intercept ? 0.1934 0.2192 0.2604 0.3243
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Constraint - -0.6863 -0.0007 -0.0175 -0.0097
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
PROA + 0.3228 0.3446 0.1444 0.3022
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Size ? 0.0114 0.0118 0.0127 -0.0016
(<0.0001) (<0.0001) (<0.0001) (0.4237)
Foreign ? -0.1087 -0.1131 -0.1049 -0.1105
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
Leverage - -0.1007 -0.1050 0.0310 -0.0908
(<0.0001) (<0.0001) (0.0073) (<0.0001)
CapEx - -0.1478 -0.1368 -0.1897 -0.1415
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
R&D - -0.1655 -0.1552 -0.2626 -0.1253
(0.0007) (0.0023) (<0.0001) (0.0079)
DiscAccruals - -0.0266 -0.0249 -0.0474 -0.0237
(0.0031) (0.0104) (<0.0001) (0.0075)
NOL - -0.0765 -0.0707 -0.0692 -0.0768
(<0.0001) (<0.0001) (<0.0001) (<0.0001)
BM ? -0.0111 -0.0184 0.0040 -0.0098
(0.0332) (0.0013) (0.4391) (0.0601)

Industry Fixed Effects Yes Yes Yes Yes

Adj R2 0.0774 0.0750 0.0802 0.0782

N 36,010 33,744 30,175 36,010
Detailed variable definitions are presented in Appendix A. All continuous variables are winsorised at the 1
and 99 percent levels. Regression models include untabulated industry fixed effects. Huber-White robust
standard errors clustered by firm are used to control for heteroscedasticity and serial correlation.

59
 Table 8
The Relationship Between Financial Constraints, Tax Planning, and Future Investment
RankWW
Coeff. Coeff.
Variable Pred. (p-value) (p-value)
Intercept ? 0.1087 0.0779
(<0.0001) (0.0008)
LagConstraint ? 0.0022 0.0089
(<0.0001) (<0.0001)
∆CashETR ? 0.0041 0.0044
(0.4607) (0.4175)
LagConstraint*∆CashETR - -0.0016 -0.0013
(0.0437) (0.0772)
DiscAccruals ? 0.0492
(<0.0001)
Size ? 0.0154
(<0.0001)
Leverage ? 0.0830
(<0.0001)
MB + 0.0000
(<0.0001)
Tangibility + 0.1495
(<0.0001)
OPC ? 0.0116
(<0.0001)
Q + 0.0231
(<0.0001)
Age - -0.0309
(<0.0001)
Dividend - -0.0063
(0.0168)
Industry Fixed Effects Yes Yes

Adj R2 0.0568 0.1658

N 35,538 35,538
Detailed variable definitions are presented in Appendix A. All continuous variables are winsorised at
the 1 and 99 percent levels. Regression models include untabulated industry fixed effects. Huber-White
robust standard errors clustered by firm are used to control for heteroscedasticity and serial correlation.

60