Journal Issue

Do Tax Amnesties Work?

International Monetary Fund. External Relations Dept.
Published Date:
January 1989
  • ShareShare
Show Summary Details

In theory perhaps. But in practice, they have limited advantages

During the 1980s, over half of the states in the United States have enacted tax amnesty legislation. A number of countries including Belgium, France, Ireland, and Italy have offered amnesty during the same period, while Argentina, Bolivia, Chile, Colombia, Ecuador, Panama, Peru, Mexico, and the Philippines have all had tax moratoria or amnesties more than once in the recent past. This recent surge in tax amnesty programs has engendered a lively debate over their effectiveness and ultimate ramifications.

The primary motivation for most amnesties has been to raise revenue, either through the funds immediately collected or through an increase in the tax base. The latter may be achieved through additions to the tax rolls or through increased reporting of certain types of income. Another motivation—that of the French and Belgian amnesties—has been the desire to facilitate the repatriation of capital illegally transferred abroad. Still other amnesties have been justified on moral grounds to allow individuals who did not pay taxes under previous regimes the chance to pay the current government without penalty, as in the Philippine amnesty of 1986.

Amnesties hold an alluring—though often illusory—prospect for governments; namely, they give the appearance of a relatively painless way to increase revenue. In reality, however, amnesties carry with them serious dangers that may have long-lasting and far-reaching effects. This article highlights important distinctions between various types of amnesties and points out that certain fundamental aspects of tax administration lead one to conclude that—apart from rather exceptional circumstances—the risks in granting amnesty are likely to exceed the potential benefits.

For a more comprehensive review, see “An Economic Analysis of Tax Amnesties,” IMF Working Paper (WP/89/42), available from the author.

Types of amnesty

At this stage, there is no agreement on whether tax amnesties succeed, in large part because of the lack of a clear definition of amnesty. In general, they offer a reduced penalty to those individuals or corporations who voluntarily regularize their situation with the tax authorities. But that is where the consensus ends. At one extreme, the possibility exists that the taxpayer would still have to pay interest and penalties, albeit at a reduced rate for the latter; and at the other, that all penalties and interest would be waived and the tax rate lowered ex post. Moreover, the forms of the amnesties themselves frequently differ in several critical respects: the magnitude of the penalty reduction, eligibility, and duration.

Magnitude. A key question is whether the amnesty participant ends up paying less than would have been the case had the evasion not occurred. If, taking into account interest charges and any penalty imposed, the evader winds up paying less in real terms than his or her original obligation, positive reinforcement is given to evasion. In addition, such an amnesty would provide a strong signal to the rest of the taxpaying population that evasion can be profitable. This brings up an important consideration—the impact of an amnesty on future tax compliance. Whatever the amount of revenue collected during an amnesty, the negative impact of a heightened expectation of future amnesties on tax compliance must be taken seriously. The expectation of a tax amnesty has been widely cited as a contributing factor to the fall in tax revenues in Argentina during the first half of 1989; this is probably because Argentina has a long history of tax amnesties.

Eligibility. Another distinction is the scope of the amnesty—that is, what sort of tax evader is eligible. An awareness of the requirements for eligibility is essential, both for interpreting the significance of the revenue collected and for making a judgment with respect to the ultimate impact of the amnesty on future tax compliance. Some governments have allowed those who are under investigation, or even those with identified tax arrears, to participate. Of course, in such cases, although revenue will surely flow to the government from these individuals, most of such revenue should have been received by the government in any event. Consequently, it is important to recognize that the revenue received during an amnesty may be a very misleading guide to the additional revenue generated by the amnesty. It is notable that studies by the US Internal Revenue Service and others suggest that the US state amnesties that generated the greatest response were, in fact, those that offered amnesty to individuals with known overdue tax obligations. One must question to what extent these “successful” cases increased revenues, and what effect such a policy is likely to have on the ability of these governments to collect tax arrears in the future.

In practice, it is difficult to estimate the amount of revenue that should be generated under an amnesty. In some countries, the authorities have been pleasantly surprised by the actual amounts collected. For example, during the Irish amnesty of 1988, approximately $770 million was collected, far greater than the expected $45 million. But in others, the participation rate and yield hardly seem to have even warranted the effort of designing the laws. In the 1982 French program, there was a general amnesty and a special provision for repatriated capital that had been illegally sent abroad, yet fewer than 3,000 taxpayers took advantage of the general amnesty, paying only about $19 million, and fewer than 300 individuals chose to repatriate capital. The 1982 Italian amnesty, although aimed particularly at taxpayers with tax appeals pending, generated less than $1 million, a fraction of the several billion dollars predicted by some officials.

When the government offers an amnesty to individuals with known tax arrears, it signals that it is willing to forgive the tax evader, even after much effort has been invested in detecting and perhaps prosecuting the evader. Although this may be tempting in situations where the government is in a revenue bind, the impact of such a policy on individuals’ perceptions of the revenue authority should not be ignored. If the revenue authority is willing to grant amnesty to those whom it has already caught evading taxes, can it have a great incentive to look for evasion in the first place? These considerations would reduce the potential evader’s fear that he will be prosecuted to the full extent of the law even if caught. Problems with judicial case backlogs and deficiencies in tax systems are fundamental and clearly cannot be solved by granting arbitrary reductions in penalties on tax evasion.

Duration. Amnesties also differ markedly in terms of whether they are temporary or standing offers. As the names imply, a temporary offer is one that is explicitly valid for a limited period of time only, while a standing amnesty is valid for an indefinite period of time. Temporary amnesties often emphasize their “once in a lifetime” nature while standing amnesties form part of the basic structure of the tax system. Recent amnesties that have attracted attention have been of a temporary kind.

…the risks in granting amnesty are likely to exceed the potential benefits.

Problems of administration

The very need for a tax amnesty presupposes evasion, a phenomenon that certainly is not new to the world. Since ancient times, tax collectors have occupied a rather unenviable position in the social order—a dislike of taxes appears at times to be almost part of the human genetic makeup. It is quite likely that even honest individuals will generally not volunteer to pay the full amount of their taxes, as they derive only a fraction of the benefit from their contribution to the funding of social expenditures. As a result, even popular governments (i.e., governments that spend according to the wishes of the people) have a need for tax administrators.

The challenge of successful tax administration rests on encouraging voluntary compliance as an adversarial atmosphere could prove quite costly to both parties. For example, theory tells us that an individual has an incentive to spend $0.99 to save $1 in taxes, whereas the government should not be willing to spend more than a fraction of this amount in order to collect $1 of revenue. The reason is that the government should value $1 taken from the individual as a loss of $1 to society and, therefore, the net social gain—if, for example, the social value of the expenditure is $1.20—would be only $.20. The government thus should not spend more than $.20 to collect $1, and under these circumstances, it is clear that voluntary compliance that minimizes both the effort spent on evasion and collection would be beneficial for society as a whole.

But how about those individuals who have already failed to comply voluntarily? To combat evasion, governments have established auditing and penalty mechanisms. It is important to realize, however, that the purpose of auditing and penalties is not only to punish tax evaders; it also forms an important part of the strategy to encourage voluntary compliance. Penalties are important since, if there is no penalty, no one will care whether or not he or she is caught. Similarly, auditing is important as people will not care what the penalty is if they know they will never be discovered cheating. Here, naturally, one must think in terms of effective penalties and not statutory ones. Although penalties may exist in the tax code, unless the legal system works efficiently to enforce the law, the penalties will not serve as a deterrent. One may envision the system as a set of three interlocking links forming a chain. If any one of the links—auditing, prosecution, and penalties—is broken, the others, no matter how strong individually, cease to be very effective.

Auditing benefits the revenue authority both directly and indirectly. The former covers the amount of revenue, penalties, and interest received from those who are audited and found to owe taxes. The latter—certainly more difficult to measure, but probably greater in magnitude—derives from the revenue paid by taxpayers who fear the very idea of being audited. A revenue authority judging audit performance in an extremely superficial manner might simply compare the revenue directly generated from auditing with the costs involved in that process, but an ideal revenue authority would consider any direct revenue from audits as a failure to induce 100 percent compliance!

The existence of both direct and indirect revenue effects places the revenue authority in a curious position. On the one hand, it would like everyone to believe it is auditing at a high rate as this would be useful in inducing compliance. But, on the other hand, it would actually like to audit at a low rate, based on the microeconomic consideration of the relative costs and benefits of individual audits. To invert a famous saying of a US President, the revenue authority strives to “speak loudly but carry a small stick.” A problem arises if individuals become aware of this, however, making them naturally skeptical of government claims that audit rates are high or that tax enforcement is going to be improved. This explains, perhaps, why the refrain “improvements in tax administration” is so often heard but so seldom believed. Moreover, this rational skepticism carries important implications as most temporary amnesties are coupled with claims that tax enforcement is on the verge of being enhanced.

Let us stop for a moment and ask why this is so. Although it might be supposed that lowering the penalty rates for those volunteering to report tax evasion would encourage individuals to come forward, the reason why evasion was undertaken in the first place must be considered. If the individual made a rational decision to evade taxes in the past, he or she must have weighed the benefits of evasion against the possibility of being caught, prosecuted, and fined. If it was sensible to evade in the past, why would it not make sense to continue to conceal this evasion, especially since the likelihood of being caught no doubt falls over time? It would seem that, in the absence of a credible threat to strengthen enforcement, individuals would be unlikely to freely admit evasion. In addition, individuals would remain skeptical about the implementation of tax enforcement measures in the short run—which is precisely the time frame of most amnesties. Thus, unless the revenue authority can make convincing changes in the short run, expectations will be slow to change and the temporary amnesty will fail.

Reasons for amnesty

The flip side of this argument, however, is that enhancing enforcement without any sort of an amnesty may have some drawbacks as well. For example, some individuals may be locked into evasion. After all, if it is just as bad to be caught evading as to admit voluntarily that one is an evader, few will volunteer. Individuals who have been habitual tax evaders may be deterred from changing their ways in the face of enhanced enforcement, fearing that a marked increase in reported tax liabilities would increase the probability of their being investigated for tax arrears during previous years. An amnesty would lessen such disincentives to reform their behavior. Both Mexico and Colombia have programs under which individuals showing a specified increase in tax declared for a given year are guaranteed they will not be audited for prior years if the current return is accurate. The drawback to this type of scheme is that individuals, by making a proper return for one year, are ensured immunity from prosecution for all previous evasions. Ideally, it would be better to assure the taxpayer that an increase in reported tax liabilities will not lead to a greater likelihood of being audited for past evasions; however, this should not mean that there is no chance of being audited.

Clearly, it is sensible to allow individuals who voluntarily reveal evasion some incentive to do so. From an efficiency standpoint, voluntary payment should strongly be encouraged as society would be able to channel the resources formerly used for enforcement and evasion to more productive areas. As a consequence, there is logic in a reduced penalty for those who voluntarily reveal tax evasion. But this logic would support a permanent, rather than a temporary, amnesty as the former would encourage voluntary payment of tax obligations. The credibility problems inherent in the temporary amnesty would be obviated and the taxpayer would be presented with a stable set of rules. Such amnesties are in place in the Federal Republic of Germany, Norway, Sweden, and the Netherlands.

While in principle, permanent amnesty sounds quite sensible, in practice a critical problem surfaces. If the policy is meant to apply only to those individuals who are not already under investigation for tax-related questions, one runs the risk of giving a large incentive to individuals working in the revenue authority to give advance warnings—in exchange for bribes—to those about to undergo investigation. Similarly, how should business associates of individuals who are under investigation be treated if they apply for amnesty after their associate is notified of an investigation that may implicate them? Administrative problems of this sort led the US Internal Revenue Service to drop, in 1952, its formal standing amnesty.


A temporary amnesty, even if the government claims that it is a “one time only” measure, may engender expectations of future amnesties. This will lead individuals to increase tax evasion and delay payment of taxes owed—particularly if accounts receivable are included in the program. The potential for harm could be quite substantial. In the US Internal Revenue Service study cited earlier, for example, it was shown that less than 1 percent of all taxpayers participated in the amnesties offered by the nine states for which such statistics were presented. Thus, a small deterioration in the tax compliance of the rest of the population as a result of such amnesties would far outweigh the benefits obtained from such a small minority of tax evaders. Moreover, when judging the effects of tax amnesties, it is important to recognize that the revenue collected during an amnesty does not represent its true yield. The collections made from individuals who are already known to owe back taxes should be offset by the present discounted value of the revenue, penalties, and interest that would have been collected in any event.

In spite of the drawbacks of amnesties, however, there may exist circumstances under which an amnesty would prove useful in improving tax compliance. If the government is seriously implementing significant improvements in tax enforcement, long-term evaders may wish to change their behavior but may be reluctant to do so as a dramatic change in reported tax liabilities would reveal their prior evasion. Hence, in the absence of an amnesty, the desired change in behavior might not occur. But in situations where the preconditions for a successful improvement in tax enforcement are lacking, temporary amnesties would appear to have little to offer and may even have an adverse impact on the administration’s credibility as well as future tax compliance.

A final caveat: with reference to what has been perceived by some as the successful experience in a number of the US states, it must be pointed out that the opportunities for improving tax enforcement that existed in these states prior to the introduction of temporary amnesties were unique and not likely to be found in many developing countries or even at the central government level in developed countries. In particular, relatively low-cost ways of improving auditing and enforcement procedures were available—including cross-checking data with the federal government—and budgetary allocations were often made to improve the functioning of the tax service. Yet even in these instances, the value of the amnesty, perse, does not appear to have been very high.

Other Resources Citing This Publication