Reference dependent preferences, hedonic adaptation and tax evasion: Does the tax burden matter?
Highlights
► People’s decisions to evade depend on how they adapt the level of fiscal pressure. ► With adaptation, people respond to changes of the tax rates, not to their levels. ► Moreover, people may adapt asymmetrically to tax increases versus tax cuts. ► Experimental evidence supports the prediction of the model with adaptation.
Introduction
How does the tax burden affect tax compliance? Is the level of the tax burden important to explain cross-country differences in tax compliance? How does a change in the tax rate influence the level of tax evasion?
Although these are fundamental questions to design effective fiscal policies, they have not been considered with the same attention. In this respect, the benchmark analysis is the classical expected utility model developed by Allingham and Sandmo (1972). The authors consider the problem faced by a taxpayer who chooses how much of an exogenous income to report for the payment of a proportional (income) tax. With a given probability, the choice of the taxpayer is audited and, in the case of ascertained evasion, she is convicted to pay (in addition to the taxes) a sanction that is proportional to the evaded income. Under these assumptions, an increase in the tax rate exerts two opposed effects on the choice of the taxpayer. On the one hand, it lowers the after-tax income from full compliance. If the taxpayer has a decreasing absolute risk aversion utility function, then the fiscal shock makes her less likely to evade. On the other hand, an increase in the tax burden implies a higher return to evading, while the potential penalty remains unchanged. Thus, the net effect of the fiscal change on the level of tax evasion is ambiguous and depends on the specific form of the utility function of the taxpayer. However, as shown by Yitzhaki (1974), if the penalty is proportional to the amount of tax evaded, then the model predicts, under decreasing absolute risk aversion, a positive relationship between the tax rate and the level of income reported by the taxpayer.
The comparative static predictions of the traditional model are not intuitive and have been criticized by those who believe that a rise in the tax burden tends to stimulate (rather than discourage) evasion. Unfortunately, the empirical evidence is not conclusive and, to some extent, seems to depend on the specific features of the econometric strategy employed. Time-series econometric studies using either aggregate or average measures of noncompliance generally find a negative relationship between the tax rate and the level of tax compliance (see Crane and Nourzad, 1987, Poterba, 1987, Pommerehne and Weck-Hannemann, 1996).1 Cross-sectional studies using data on individual audited tax returns report mixed results. While Clotfelter (1983) finds that the level of the tax burden is positively associated with tax evasion, other contributions document either a nonsignificant correlation (among others, Cox, 1984, Slemrod, 1985, Kamdar, 1995) or (even) a negative relationship (Feinstein, 1991).
Also experimental studies generate ambiguous evidence. Two of the earliest within-subject experiments (Friedland et al., 1978, Baldry, 1987) find that a reduction in the tax rate is associated with a higher level of tax compliance. However, the robustness of this result has been questioned by between-subject designs. Indeed, while in a study conducted on US students, Alm et al., 1992, Alm et al., 1995 observe a positive relationship between the tax rate and the amount evaded, in a subsequent experiment involving Spanish subjects, Alm et al. (1995) document contradictory results.
A possible explanation for this empirical impasse is that tax evasion decisions are more complex than what the classical model postulates. For instance, tax compliance can depend on frames and references, which in turn adapt to circumstances and events.
Cumulative Prospect Theory (Tversky and Kahneman, 1992) represents the leading approach to incorporate the notion of reference dependent preferences in behavioral economic models. According to this theory, rather than being measured in absolute levels, economic outcomes are evaluated as gains or losses relative to a reference. Applications of Prospect Theory to tax evasion have been recently proposed by various researchers (e.g., Shepanski and Shearer, 1995, Bernasconi and Zanardi, 2004, Kirchler, 2007, Dhami and al-Nowaihi, 2007, Rablen, 2010). These contributions have investigated specific aspects of the theory, such as the subjective weighting of probabilities, the diminishing sensitivity of gains and losses, the property of loss-aversion.2
While studies have introduced reasonable assumptions on the formulation of references in tax evasion decisions, less attention has been devoted to analyze how taxpayers react and adapt to changes in fiscal conditions over time. In this paper, we extend the existing analysis by considering how the adaptive process of the reference used by the taxpayers in taking their decisions influences the level of tax evasion.
Psychologists use the term hedonic adaptation to refer to “processes that attenuate the long-term emotional or hedonic impact of favorable and unfavorable circumstances” (Frederick and Loewenstein, 1999, p. 302). By hedonic adaptation, reference dependent preferences tend to adjust to changes in pre-existing conditions such that the behavioral response to repeated stimuli can be limited in time.
Recently, hedonic adaptation has been extensively used by behavioral economists to analyze consumers’ habit formation, the response to changes in health conditions and investment decisions (Frederick and Loewenstein, 1999). Moreover, there is substantial evidence showing that reference points adapt to favorable and unfavorable events at a different speed (Arkes et al., 2008, Arkes et al., 2010, Lyubomirsky, 2011).
Our analysis offers new insights on how hedonic adaptation affects taxpayers’ perception of fiscal variables and their decision to evade. First, we study how taxpayers (instantaneously) react to a change in the tax rate. In particular, we find that an increase in the tax rate discourages taxpayers’ compliance. Second, we show that, once taxpayers completely adapt their reference to the new fiscal conditions, the level of tax evasion is independent of the tax burden, a prediction that is also consistent with the old adage in public finance stating that “an old tax is no tax”(Bastable, 1892).3 We find supporting experimental evidence in favor of the predictions of our model. Moreover, in our experiment, we also find that subjects tend to adjust faster to tax cuts than to an increase in the tax rate.
The rest of the paper is structured as follows. In Section 2, we briefly review the literature analyzing the relationship between tax evasion and the tax burden. In Section 3, we present a theoretical framework that introduces reference dependent preferences and hedonic adaptation in a model of tax evasion decisions. In Section 4, we describe the experiment aimed at testing the main theoretical predictions and we present the results. We conclude in Section 5 with a discussion of the policy implications of our analysis.
Section snippets
Classical expected utility approaches
The benchmark model of tax evasion is based on two seminal contributions: Allingham and Sandmo, 1972, Yitzhaki, 1974. A taxpayer with gross income Y > 0 is required to pay taxes according to a flat tax rate τ ∈ (0, 1). The tax authority does not observe the gross income of the taxpayer, directly. However, with probability p ∈ (0, 1), the tax authority conducts an audit procedure on the decision of the taxpayer. If audited and found to have concealed part of her income, the taxpayer is convicted to
Tax evasion with reference dependent preference and hedonic adaptation
Under reference dependent preferences (as entailed by the classical version of Cumulative Prospect Theory), the objective function of the taxpayer can be written as follows:where r is the ‘reference’ income that is used by the taxpayer to take her tax decision and v(·) is the reference dependent value function, which is concave in ‘gains’ and convex in ‘losses’. In particular, we assume the original specification of v(·):9
Design
Our experimental design is aimed at testing the theoretical predictions implied by the model of tax evasion with reference dependent preferences and hedonic adaptation. Each session consisted of two consecutive phases of 12 periods and involved twenty subjects.19 In every period of the experiment, the computer randomly assigned to each subject a gross income included between 120 and 180 tokens. On the basis of her gross income, each
Concluding remarks: policy implications
The analysis presented in this paper has several policy implications. It is generally thought that the problem of tax evasion has been exacerbated by the general increase in the tax burden that has occured in several (developed) countries in the second half of the last century (Tanzi and Schuknecht, 1997). A fundamental question suggested by this empirical observation concerns the possibility of using a tax cut as an instrument to reduce tax evasion (Clotfelter, 1983). In contrast to the
Acknowledgements
We thank Johannes Abeler, Armenak Antinyan, Thomas Bassetti, Ennio Bilancini, Christopher Cotton, participants at the seminar at the University of Modena and Reggio Emilia, the 2011 Shadow Economy, Tax Evasion and Money Laundering Conference in Münster, the 2011 GRASS meeting in Venice, the 2010 ESA World Meeting in Copenhagen and two anonymous referees for useful comments. Financial support for this research from the National Research Project “New approaches to tax evasion” (COFIN 2007), Grant
References (55)
- et al.
Income tax evasion: A theoretical analysis
Journal of Public Economics
(1972) - et al.
Why do people pay taxes?
Journal of Public Economics
(1992) Why free ride? Strategies and learning in public goods experiments
Journal of Public Economics
(1988)- et al.
True positives and false alarms in the detection of tax evasion
Journal of Economic Psychology
(1995) - et al.
A cross-cultural study of reference point adaptation: Evidence from China, Korea and the US
Organizational Behavior and Human Decision Process
(2010) - et al.
Reference point adaptation: Tests in the domain of security trading
Organizational Behavior and Human Decision Process
(2008) Tax evasion and orders of risk aversion
Journal of Public Economics
(1998)- et al.
Loss aversion in a consumption-savings model
Journal of Economic Behavior and Organization
(1999) - et al.
Why do people pay taxes? Prospect theory versus expected utility theory
Journal of Economic Behavior & Organization
(2007) - et al.
Influencing the prospects of tax evasion
Journal of Economic Psychology
(1997)