Tax Avoidance: Rational Ignorance of Law

The past decade has witnessed a growing interest in researching tax avoidance in a range of disciplines outside the traditional accounting and law —— from anthropology; to psychology and sociology; and, not least, in economics. The growth of interest was stimulated by politicians, mass media, and the public focussing on tax avoidance, often conflated with tax evasion, in many countries after the global financial and economic crisis put a squeeze on private and public finances.

Changing definitions, controversies over the interpretation of concepts, rules, and policies appear to have created inconsistent assessments by legislators, tax authorities, tax advisors and taxpayers over the legality of tax avoidance schemes. The disagreement suggests that parties sometimes are not fully aware of tax law, or the law is ambiguous, or both. Most importantly, there has been a shift in the view of whether exploiting the ambiguities in law for reducing tax liabilities is acceptable.

Uncertainty or ambiguity in the tax law, and the ensuing issue of intended or unintended non-compliance or over-compliance, has been a topic of debate in the tax law literature for some time. An increasing complexity of the economic and financial environment has led to the emergence of new types of transactions and instruments whose tax treatment does not fall into existing categories.

Within existing tax systems, legislative efforts aspiring to cover every contingency in detail cause the law to become extremely complicated. The resulting rules may be prohibitively costly both to implement and to learn. As an alternative, a set of general rules allows for flexibility in covering the existing and evolving transactions, but is prone to wide interpretation and misinterpretation. Tax law is thus rendered uncertain, practically in the former case and fundamentally in the latter.

Individuals and businesses aiming to exploit preferential tax treatment of certain activities to their benefit face the possibility that their interpretation of the rules will be deemed unacceptable by the revenue authority. Disputes sometimes end up in courts, where judges make decisions based on the relative strength of legal arguments and evidence submitted by the parties. Litigation is costly and, on the grounds of the cost-benefit comparison, either party may find it impractical to reduce uncertainty beyond some point.

A theoretical framework of tax avoidance game

What can a standard framework of economic analysis say about the behaviour of taxpayers and tax authority in an uncertain legal environment? First, one must recognize that a taxpayer no longer takes tax law as a given; rather, a taxpayer holds a belief about the interpretation of their tax planning activity. Second, tax authority is unable to pre-commit to its interpretation of tax planning activities of a taxpayer and may reverse its decision. Most importantly, the tax authority’s decision is not final: the interpretation of the tax law remains uncertain and may swing either way with positive probability.

One possible approach is to model this situation as a dynamic game of strategy. In what follows I describe the basic setup, the results, and some possible extensions.

In the first stage of the game a taxpayer enters a tax-saving scheme without being fully certain of its interpretation, and the tax authority can either disallow the scheme and demand repayment of tax, or allow the scheme, but is unable to commit to its ruling. The taxpayer can also disagree with the tax authority’s decision.

In the second stage, a dispute can be taken to the court. The outcome depends on the relative weight of evidence submitted to the judge and on the “bias” in the perception of the taxpayer’s case. This framework is used to address the following questions:

(1) What influences the ruling of a tax authority on a tax avoidance scheme and the decision of a taxpayer on the involvement in a scheme?

(2) Does more uncertainty in tax law encourage more avoidance?

To defend their position, the parties acquire costly information about the law (law discovery) and the specific case (fact discovery), or hire tax specialists for this purpose. A tax specialist interprets the law and evaluates the chance of success. Further costs can be associated with developing legal arguments as well as the administrative procedures. Each party compares the cost and the expected benefit of the outcome, given the prevailing bias.

When the bias is against the taxpayer, the revenue authority declares the scheme unacceptable, or abusive, and the taxpayer accepts the ruling. Conversely, when the bias is sufficiently strong in the favour of the taxpayer the revenue authority rules the scheme acceptable. In an intermediate case, an equilibrium is possible where a tax authority can declare a tax avoidance scheme acceptable or unacceptable at random, and both the taxpayer and the tax authority choose to be only partly informed about the tax law. Ignorance of law is economically rational.

The bias described in the model captures social perceptions of tax planning activities which differ across countries and change over time. Thus, tax planning by an individual investor may be less likely to be deemed abusive than a scheme pitched to a group; a society with stronger institutions and higher trust in the government may be less tolerant to aggressive tax behaviour than the one with weaker institutions and lower trust; in the wake of a financial crisis, perception may turn against large corporate taxpayers seen as exploiting tax planning opportunities not available to ordinary individuals and small businesses, and thus not paying their “fair share” of taxes.

Possible extensions to investigate positive and normative issues

The baseline framework is intentionally simple and allows various extensions to investigate further positive and normative issues. One interesting extension would be to introduce risk aversion or loss aversion to describe an individual taxpayer.

Loss aversion reduces the expected payoff from taking the dispute to the court. When bias is in the favour of taxpayer and the loss due to unsuccessful tax planning is substantial a loss-averse taxpayer acquires more information in equilibrium to increase their chance of winning in court, relative to the case without loss aversion. In response, the tax authority chooses to acquire less information. When loss is relatively small, the outcome depends on the combination of the degree of loss aversion and the magnitudes of the payoffs. Another possibility is to model the evolution of bias in a repeated setting; a repeated game will also allow introducing accumulation of information. The extent of the ambiguity in law could also be quantified, and the issue of optimal uncertainty could be investigated.

Rational ignorance and the value of information can be further analysed in a modified framework, with richer assumptions on the distribution of payoffs. In the simple framework, the ex-ante payoffs have effectively a binary distribution. Many disputes, however, are settled by negotiation, with the value of settlement up to the full amount of saved tax, with or without interest charge. The objective of the taxpayer and the tax authority can then be modelled as the choice of the distribution of the settlement conditional on the information —— to maximize the expected payoff under the capacity constraint on the amount of information. This capacity constraint reflects both the material cost of acquiring information and the limitation of the power of the players to process available information.

Choice under uncertainty is widely researched in economic literature in various contexts. Study of tax avoidance, with combined expertise of lawyers and economists, can offer new insights in behaviour of individuals and businesses interacting with uncertain institutions, where the behaviour itself influences the degree of uncertainty.

This paper is a work in progress. Please do not cite without permission. Thanks are due to Yuliya Epifantseva, PhD, J.D., for valuable suggestions and comments on the draft.