THEORIES OF TAX EVASION AND THE HIDDEN ECONOMY Nordic Workshop on Tax Evasion AGNAR SANDMO Norwegian School of Economics (NHH)
TAX EVASION: AN OVERVIEW Point of departure: The expected utility theory of personal income tax evasion, its assumptions and predictions. Extensions, e.g. interaction with labour supply, indirect tax evasion, optimal taxation. Relating the literature to the broader concerns of public finance.
A BRIEF HISTORY OF THOUGHT The older economists (Smith 1776, Pigou 1928) acknowledged the presence of tax evasion but paid little attention to it. Musgrave s treatise (1959) does not mention it. It entered the literature on public economics in earnest around 1970. Theory: Allingham and Sandmo (1972). Empirical work: A number of studies based on the cash balance approach.
INCOME TAX EVASION: THE PORTFOLIO APPROACH Basic assumption: The risk averse taxpayer chooses the amount of income to report with the objective of maximizing expected utility, V. V = (1-p)U(Y) + pu(z), where p is the (subjective) probability of detection and Y and Z after tax income in the states of nondiscovery and discovery.
AFTER TAX INCOME AND EVASION Let W be true income and e the amount evaded. Then Y = W t(w-e) = (1-t)W +te. Z = W t(w-e) θe = (1-t)W (θ-t)e. Here t is the tax rate and θ the penalty rate. θ >t.
INTERIOR SOLUTION: e > 0 e/ θ < 0. e/ p < 0. e/ t = -[(W - e)/(1 t)]( e/ W) + S. Income effect (< 0) + substitution effect (> 0).
THE EXTENSIVE MARGIN When is it optimal to choose e > 0? Answer: When t > pθ. Critics of the model have argued that since this inequality is likely to hold for the large majority of taxpayers, the model predicts too much tax evasion.
DOES THE MODEL PREDICT TOO MUCH EVASION? But note: When p = 1, e = 0. For a large number of taxpayers their income will be reported separately by the source of their income, so that it is known to the tax authority. In this case, no rational taxpayer will attempt to evade. In other words, part of W is non-evadable.
FURTHER WORDS OF CAUTION Taxpayers cannot be assumed to have exact, objective knowledge of the magnitude of p and θ, and they may be likely to overestimate them. Alternatively, the taxpayers may have more complex preferences than assumed in the standard model.
ALTERNATIVE VIEWS OF TAXPAYER PREFERENCES The standard model contains a view of taxpayers as calculating and cynical. Many of them probably are but some are not! Social stigma: V = (1-p)U(Y) + pu(z, s). Bad conscience: V = (1-p)U(Y) + pu(z) B(e).
IMPLICATIONS FOR THE EXTENSIVE MARGIN In the bad conscience case, the condition for e > 0 to be optimal becomes t > pθ + B (0)/U (W(1-t)). The second term on the RHS can be interpreted as a conscience tax.
ETHICAL INFLUENCES ON TAX EVASION These two cases represent flexible ethics. One could also envisage different kinds of ethical influences, like Kant s categorical imperative: Act according to the maxim that you would wish to become a universal law. In this case the taxpayer might adopt the rule e = 0 as a constraint on his behaviour.
THE ROLE OF PUBLIC EXPENDITURE Taxpayer preferences presumably also include public expenditure, G, so that V = (1-p)U(Y, G) + pu(z, G). As long as taxpayers take no account of the government s budget constraint, this does not change the predictions of the model.
DISCARDING EXPECTED UTILITY Some have argued that expected utility makes unrealistic demands on taxpayer rationality and have explored non-expected utility alternatives. One alternative is the Arrow-Debreu state preference theory. V = Φ(Y, Z) where Φ is an ordinal utility function.
IMPLICATIONS OF STATE PREFERENCE THEORY Comparative static implications are similar to the expected utility approach (for e>0). But a weakness of the approach is that it does not distinguish between attitudes to risk and subjective probabilities in its concept of ordinal preference.
INTERACTIONS WITH LABOUR SUPPLY The portfolio model takes pre-tax income as exogenous. To capture the interaction with black labour markets one has to extend the model to one of consumption-leisure preferences and variable labour supply. Time constraint: L + H + h = T
VARIABLE LABOUR SUPPLY Y = (1-t)WH + wh Z = (1-t)WH + (1-θ)wh The comparative static properties are more complex than in the portfolio model, mainly because of the opportunities for time substitution.
PARTICIPATION CONDITION When is it optimal to do some work in the hidden economy? In the simplest model the condition is W(1-t) < w(1-pθ). This generalizes the previous condition.
LABOUR MARKET EQUILIBRIUM The study of market equilibrium should analyze the determination of the relative wage w/w. This will depend both on tax policy parameters and the extent of collusion between employers and employees. Since w/w is likely to differ between industries, tax evasion and black labour markets will have implications for industrial structure.
TAX EVASION BY FIRMS Indirect taxes: Opportunities differ by types of tax (VAT vs. environmental taxes). Can be studied via the portfolio approach, although with risk neutral firms (Marrelli 1984). Corporate income tax: Essential to study the relationship between managers and owners (Crocker and Slemrod 2005). Requires a contract theoretic approach.
INTERACTION WITH OTHER TYPES OF CRIME Tax evasion often forms part of other criminal activities, e.g. violation of labour and immigration acts, laws that regulate financial transactions etc. Smuggling (Bhagwati and Hansen 1973) is to some extent a special case of tax evasion but also concerns import of goods where consumption is prohibited, e.g. drugs.
SOCIAL INTERACTION To understand the extent of tax evasion and black market activity one needs to study the interaction between taxpayers (as in Schelling 1978). The perceived gains from evading taxes may depend on how common evasion is. E.g., the extent of evasion may determine perceptions of the probability of detection or the social stigma associated with detection. This may lead to multiple equilibria that could explain large differences in compliance behaviour between countries.
CAN TAX EVASION BE IN THE PUBLIC INTEREST? Efficiency: By lowering effective tax rates, tax evasion may but need not! lead to smaller distortions and efficiency losses. Equity: By relieving the tax burden on lowincome workers, tax evasion may but need not! lead to a more equitable distribution of after tax income.
OPTIMAL TAXATION Efficiency considerations may lead to lower taxes on goods (including labour) for which evasion is important. Intuition: Evasion implies a larger elasticity of the tax base. But revenue losses must be compensated by increases in other taxes involving distortions and efficiency losses.
OPTIMAL TAXATION (CONT.) However, the choice of tax rates must be considered together with the allocation of resources to ensure compliance. These considerations together with distributional concerns lead to a complicated exercise in the theory of optimal taxation.
FURTHER PROBLEMS FOR OPTIMAL TAX THEORY Many would argue that tax evasion is morally objectionable. This raises doubts about the welfarist framework of optimal tax analysis. Judgements concerning the fairness of income redistribution may relate both to the outcome of the redistribution process and to its moral acceptability.
INTERNATIONAL DIMENSIONS In recent decades the emergence of tax havens has become important for the understanding of tax evasion, particularly related to capital income (Schjelderup et al. 2009). This has implications for national and international distribution of income as well as international efficiency of resource allocation.
THE TAX ADMINISTRATION PERSPECTIVE In the public finance literature the costs of taxation have mostly been related to the costs of price distortions while administrative costs have been neglected. The theory of tax evasion can be seen as an attempt to include these costs, thereby contributing to a more balanced assessment of the social costs of taxation (Slemrod).
THE WELFARE STATE Tax evasion makes it more difficult to target instruments to objectives in welfare state policy. E.g., widespread income tax evasion makes the progressive income tax less effective as an instrument of redistribution. Tax evasion may also reduce the social legitimacy of the welfare state if some citizens are perceived as escaping from paying their fair share.