Identifying Different Types of Tax Avoidance: Implications for Empirical Research

David A. Guenther, Linda K. Krull, and Brian M. Williams

In this study, we provide a theoretical framework for how measures of tax avoidance can reflect tax avoidance related to, or unrelated to, earnings management. Pretax income is often used as a benchmark from which to measure tax payments. Tax avoidance can thus come from two mechanisms: either decreasing tax payments holding pretax income constant (numerator effect), as is commonly assumed in the literature, or by increasing pretax income holding taxes paid constant (denominator effect). We argue that a measure of tax avoidance that has been used in prior research—the ratio of cash taxes paid to pretax operating cash flows—captures tax avoidance that is unrelated to earnings management. Dependent on the research question being addressed, researchers may wish to use a pretax income-based measure, a cash flow-based measure, or both.

We use comparative statics to compare four different cases of tax avoidance to provide conceptual evidence on how different forms of tax avoidance can reflect numerator or denominator effects. Two of these cases reflect denominator effects that are motivated primarily by financial statement earnings management, and two reflect numerator effects that are motivated primarily by the desire to pay less tax. We demonstrate how differences in these four cases affect taxes paid and pretax income. We also show that cash flow from operations can serve as a useful alternative benchmark to help identify what type of tax avoidance is occurring. We believe the identification of the underlying type of tax avoidance (in terms of numerator or denominator effects) is a useful implication of our study. For example, assume that a researcher uses an accounting income-based measure of tax avoidance and finds evidence of tax avoidance. If the researcher also uses an alternative cash flow-based measure and fails to find evidence of tax avoidance, the conclusion would not be that the firm did not avoid tax, but rather that the firm avoided tax by increasing earnings in a nonconforming way, reflecting the denominator effect. On the other hand, if the cash flow-based measure also suggested tax avoidance, the conclusion would be that the firm decreased tax payments, reflecting the numerator effect.

To confirm our conceptual analysis, we also empirically test the relation between various measures of tax avoidance and earnings management. We find that measures of tax avoidance that use pretax income as the benchmark are generally related to earnings management, while measures of tax avoidance that use pretax operating cash flows as the benchmark are generally less related to earnings management. We find that the influence of earnings management on measures of tax avoidance is strongest for one-year measures, and that aggregating pretax income over a longer period of time such as five years somewhat mitigates, but does not completely eliminate, the relation with earnings management.

We note that the potential link between measures of tax avoidance and earnings management has been recognized by researchers in prior studies, and cash flow-based benchmarks have been proposed as a way to deal with that problem. An important insight of our study is that a failure to find consistent results when using a cash flow-based benchmark does not necessarily suggest that there is no tax avoidance, but rather provides information on the type of tax avoidance that is occurring (i.e., a numerator or denominator effect). Our study contributes to this literature by providing a theoretical framework bolstered by comparative statics, as well as empirical evidence on the extent of the relation between earnings management and measures of tax avoidance.

SALTy Citizens: Which State and Local Taxes Contribute to State-to-State Migration?

Amy M. Hageman, Sean W. G. Robb, Jason M. Schwebke

State and local governments depend heavily on revenue generated from state and local taxes (SALTs). Migration between states occurs for many non-tax reasons, but prior research suggests tax-related factors may play an important role. Even so, the specific SALTs motivating state-to-state migration remain unclear given mixed and inconclusive results of prior studies. This study utilizes a larger sample population than prior research, examines state and local tax variables individually and then jointly in one model, and controls for important economic and noneconomic factors to determine which SALTs matter to residents when making relocation decisions. Using data collected from multiple sources for 2008 to 2015, results indicate overall state and local tax burden, individual income taxes, select sales taxes, and property taxes all are significantly associated with taxpayer migration. Further, select sales taxes and property taxes seem to be most significantly associated with migration. We also provide some intuition related to the economic impact associated with net migration.

We contribute to the tax literature in several important ways. First, we demonstrate that certain state and local taxes are associated with taxpayer migration patterns. Second, we extend prior research by examining the tax variables categorically to test individual associations, and then collecting the variables into one model to determine which SALTs have the strongest association with migration. Finally, we contribute to the tax policy literature by quantifying specific economic impacts that are associated with state and local tax policy. As state policymakers battle to keep residents in their state and try to attract new residents from other states, knowledge of the impact that changes to certain taxes have on taxpayer migration is vitally important.

The Concave Association between Tax Reserves and Equity Value

Rebekah D. Moore

This study investigates whether the association between tax reserves and firm value is concave. Recent accounting research documents an average positive equity investor valuation of tax reserves reported according to FASB Interpretation No. 48 (FIN 48) (Drake, Lusch, and Stekelberg 2019; Frischmann, Shevlin, and Wilson 2008; Koester 2011; Koester, Lim, and Vigeland 2015; Robinson and Schmidt 2013). This research suggests tax reserves reflect good managerial stewardship and potential future tax savings (Koester et al. 2015). However, tax reserves are a contingent liability representing potential future cash outflows and signals about the costs associated with tax planning activities, which should be viewed negatively by investors.

A concave association between tax reserves and firm value allows for both views of tax reserves depending on the relative amount of tax reserves. Investors perceive both positive and negative signals from tax reserves. In particular, a concave association suggests the expected costs will be perceived as greater than the expected benefits at higher levels of tax reserves.

Using a modified Ohlson (1995) stock price level estimation, I find firm value is increasing in tax reserves at a diminishing rate until an optimal level is reached, after which firm value is decreasing in tax reserves. This concave association is consistent with investors having a firm-specific, optimal level of tax reserves where the expected benefits equal the expected costs.

I also examine the effect of conservative financial reporting on the association between tax reserves and firm value. Firms that exhibit conservative financial reporting are more likely to record higher levels of tax reserves for the same underlying level of tax risk. I find that conditional conservatism moderates the concave association between tax reserves and firm value such that conservative firms experience a muted equity price sensitivity to tax reserves. This result suggests that investors of conservative financial reporting firms recognize the conservatism and discount the signals about both the expected benefits and costs associated with the underlying tax activities. I do not find any evidence that unconditional conservatism moderates the association between tax reserves and firm value.

I also investigate whether the concave association between tax reserves and firm value is driven by a numerator (expectations about future cash flows) or denominator (cost of capital) effect. I find a consistent concave numerator impact across several specifications. I find mixed evidence regarding the association between tax reserves and the cost of capital. Analyzed together, the results suggest the numerator is the primary driver of the concave association between tax reserves and firm value, consistent with investors incorporating their perceptions about tax reserves into expectations about future cash flows.

Understanding how investors value tax reserves at different levels represents an important extension to research about the valuation of tax reserves because managers and policymakers rely on information about investor preferences to inform their decision making.

Is Corporate Social Responsibility Related to Corporate Tax Avoidance? Evidence from a Natural Experiment

Michael A. Mayberry and Luke Watson

We investigate the relation between corporate social responsibility (CSR) and corporate tax avoidance. Theoretically, three plausible hypotheses pertain to this relation. First, the transparent reporting hypothesis suggests tax avoidance undermines the credibility of firms' CSR investments and policies, leading to a negative relation. Second, the opportunistic reporting hypothesis suggests firms use CSR as an insurance-like mechanism to hedge against the risks of tax avoidance, leading to a positive relation. Third, the decoupling hypothesis suggests firms fail to integrate CSR into their tax policies, leading to no relation. Firms decouple when facing competing objectives and lacking a clear consensus as to the optimal strategy. Firms could decouple CSR from tax avoidance because of the conflicting objectives and preferences between responsibly paying their “fair share” of tax and reducing their tax burden.

We employ state constituency statutes as a natural experiment in which directors are, for the first time, legally permitted to consider non-shareholders' implicit claims on firm resources in their business decisions (Flammer and Kacperczyk 2016). This exogenous reduction in the expected marginal cost of CSR activities (Luoma and Goodstein 1999) yields a setting for an intention-to-treat analysis (Lazuka 2020; Rouse 1998). Critically, constituency statutes only affect firms incorporated in a given state rather than headquartered there. Using this identification strategy, we examine whether CSR is associated with corporate tax avoidance.

Our primary analyses use a difference-in-differences specification with firm and headquarters state-by-year fixed effects, thereby isolating the firm-level change in corporate tax avoidance around the enactment of constituency statutes while extracting the effects of local economic, social, and political conditions within the firm's headquarters state and year. We fail to find a statistically significant association between firms' CSR environment and tax avoidance, measured using book and cash effective tax rates, discretionary permanent book-tax differences (Frank, Lynch, and Rego 2009), and tax shelter probability (Rego and Wilson 2012). Therefore, our results are consistent with firms decoupling CSR from tax policy.

Relatively narrow confidence intervals, economic insignificance, in-sample validation of treatment effects, and results of other studies indicate that a lack of power does not drive our lack of results. We employ entropy balancing, alternate fixed effects structures, and alternate treatment events to ensure our inferences are robust. We investigate changes in tax avoidance in subsamples of firms most likely to increase CSR activities because of constituency statutes. We perform instrumental variable analysis. All along, we fail to find evidence of a relation between CSR and tax avoidance.

The relation between CSR and tax avoidance has generated significant interest in accounting research, with dramatically inconsistent prior results despite similar research designs. Huang and Watson (2015) recommend using alternative identification strategies to address CSR-related research questions. We respond by using exogenous variation in CSR to investigate the association between CSR and tax avoidance, thereby advancing the literature on CSR and tax avoidance. We contribute broadly to the accounting literature on CSR by applying a natural experiment that has been used in other business disciplines to an accounting-related question. Finally, we contribute to the vast literature on corporate tax avoidance by identifying a rare finding of no association amid a bevy of studies identifying determinants.

Proprietary Costs and the Reporting of Segment-Level Tax Expense

Zero Deng, Fabio B. Gaertner, Daniel P. Lynch, and Logan B. Steele

This study investigates how tax decisions affect disclosure policy by examining whether these proprietary costs of tax avoidance reduce the likelihood that a firm reports segment-level tax expense. Segment-level information can provide valuable information to investors about the drivers of segment-level performance (AIMR 1993; Bens, Berger, and Monahan 2011). Current accounting rules for segment-level reporting afford managers significant discretion in what line items to report (Paul and Largay 2005; Bens et al. 2011). Since segment tax information could be used by tax authorities and competitors to identify tax planning strategies, firms engaging in greater levels of tax avoidance likely face greater proprietary costs of disclosure.

ASC 280 outlines the management approach and states that segment-level tax expense should be disclosed if reviewed by the chief operating decision maker in assessing segment performance. However, the standard also exempts firms from disclosure of segment tax information when the information is not already computed. The apparent discretion in the decision to report segment-level tax expense, combined with the higher proprietary costs of disclosure faced by firms engaging in tax avoidance, leads to the prediction that tax avoidance is negatively associated with disclosure of segment-level tax expense. Consistent with this prediction, we find firms with lower GAAP effective tax rates (ETRs) are less likely to report segment-level tax expense, on average. The economic magnitude suggests that a one standard deviation decrease in ETR corresponds to a 5.0 percent decrease in segment tax reporting.

According to current guidance, managers define business segments along either geographic or non-geographic lines. Since geographic disclosure reveals tax jurisdiction-specific information, we predict that tax avoidance is more negatively associated with disclosure of segment-level tax expense for firms that disclose geographic segments. Consistent with this prediction, we only observe a negative relation between tax avoidance and disclosure incidence within the geographic segment group. For firms that define segments on a geographic basis, we find that a one standard deviation decrease in ETR is associated with a 21.7 percent decrease in the probability of reporting segment-level tax expense.

We also examine alternative measures of tax-related proprietary costs. Given proprietary costs of disclosure are concentrated in geographic segments, we conjecture that one specific tax planning strategy that could be adversely affected by the disclosure of segment-level tax expense is international income shifting. We find firms with income shifting incentives are less likely to report segment-level tax expense, especially within the subsample of firms that report on a geographic basis. We also find that firms with more aggressive tax positions, which we measure using the level of unrecognized tax benefits, are less likely to disclose segment-level tax expense. Taken together, our findings suggest firms facing the highest proprietary costs of disclosure potentially use the discretion afforded to them under segment reporting standards to avoid reporting segment-level tax expense. These results should be of interest to standard setters, as incomplete and biased disclosures weaken the faithful representation of accounting numbers.

Regulation and Tax Preparer Qualifications

Matthew Reidenbach, Trevor L. Sorensen, and John S. Treu

Tax preparer regulation is a controversial topic among policymakers, tax preparers, and certifying bodies, with much of the debate surrounding whether regulations increase or decrease the qualifications of tax preparers and the quality of tax returns. This study primarily focuses on whether tax preparer regulations are associated with more or fewer highly qualified tax preparers (i.e., enrolled agents, CPAs, and attorneys). Regulation is often implemented to function as a screening mechanism to reduce the number of non-qualified preparers. However, to the extent that regulation creates market confusion about tax preparer qualifications and competencies, such regulation may actually reduce the number of highly qualified tax preparers.

Concerns over the quality of tax preparation services motivated the IRS to regulate the profession starting in 2012 with the Registered Tax Return Preparer program. Following the longstanding regulations in Oregon and California, the federal regulation imposed a minimum competency exam and continuing education requirements for any tax preparer not otherwise qualified as an enrolled agent, CPA, or attorney. However, the program was unexpectedly invalidated in just its second year after a judge ruled that the IRS lacked Congressional authorization to impose such requirements.

The changes to the federal regulation, combined with the mix of regulated and non-regulated states, provide an ideal setting that we analyze to better understand the impact of these regulations on tax preparer qualifications. Based on our empirical analysis of all IRS-registered tax preparers in the U.S. between 2012 and 2016, we find evidence that the regulation of tax preparers is associated with higher proportions and numbers of highly qualified tax preparers. We also find evidence that CPAs and enrolled agents are more likely to enter the tax preparation market, and less likely to exit, if tax preparer regulations are in place.

These findings suggest that regulations both screen non-qualified tax preparers from the market and increase the visibility of certifications to motivate greater participation by highly qualified preparers, consistent with signaling theory. In additional analysis, we find some evidence that tax preparer regulations are associated with increased tax preparation fees and improved tax return quality, suggesting there is a tradeoff between tax preparer qualifications and the cost of tax preparation services.