Tax administration plays a crucial role In determining a country’s real (or effective) tax system. Unfortunately, tax administrations in many countries do not function optimally and distort the intention of tax laws. In order for taxation to have its intended effect on the allocation of resources, the distribution of income, and macroeconomic stability and growth, the tax administration must function effectively and efficiently. This paper addresses tax administration reform by describing briefly the causes for inefficient tax administration, identifying the essential elements of successful reform, and presenting measures to improve the tax administration. The common thread of a reform strategy is to focus administrative resources on problems in the tax collection system through promotion of voluntary taxpayer compliance and adoption of a logical sequence of procedures for efficiently identifying and handling instances of noncompliance.


Tax administration plays a crucial role In determining a country’s real (or effective) tax system. Unfortunately, tax administrations in many countries do not function optimally and distort the intention of tax laws. In order for taxation to have its intended effect on the allocation of resources, the distribution of income, and macroeconomic stability and growth, the tax administration must function effectively and efficiently. This paper addresses tax administration reform by describing briefly the causes for inefficient tax administration, identifying the essential elements of successful reform, and presenting measures to improve the tax administration. The common thread of a reform strategy is to focus administrative resources on problems in the tax collection system through promotion of voluntary taxpayer compliance and adoption of a logical sequence of procedures for efficiently identifying and handling instances of noncompliance.

I. Introduction

The background information for authors of papers sent by the organizers of the IRIS Conference stated that “the IRIS project is based on the idea that a principal determinant of a country’s economic success is the quality of the country’s institutions and policies.” Few, today, would question this statement. Because institutions are the vehicles that carry policies, the existence of good institutions is a necessary, if not a sufficient, condition for the implementation of good policies. This is true in many areas of economic policy, but is especially true in taxation. In this area the distinction between policy and administration may, at times, become artificial. In some cases, tax administration becomes, in a way, tax “policy” because it can distort and change the objectives that the policymakers want to pursue through tax policy and can, thus, greatly influence the structure of taxes and lead to unintended results.

In spite of the obvious importance of the topic, there is a remarkably limited literature on tax administration. The reason for this may be that those who work in the tax administrations often do not have the time or the interest to write about it. Furthermore, when they do write, they normally limit themselves to an evaluation of the tax administration of the country in which they operate. In other words, their writing tends to be country-specific and, as a consequence, it lacks universality. On the other hand, those who venture to write in a more general way often lack the detailed knowledge necessary to make their writing useful. The reason for this is that very few individuals have had the opportunity to observe closely the workings of several tax administrations. A sizeable proportion of these individuals works for the Fiscal Affairs Department of the IMF. These individuals have been engaged over many years with the task of assisting many countries with the reform of their tax administrations. The paper draws a lot from their experiences and from the various (confidential) reports, prepared at the request of ministers of finance, over many years.

A country’s tax system can be described in, at least, three ways: a statistical, a statutory, and an effective way. The statistical description allocates actual tax revenue to notional tax bases and provides an impression of the revenue importance of various tax bases and tax rates; the statutory description is the one found in the tax codes, tax laws, and tax regulations; the effective description is the one based on economic reality, taking into account all the distortions brought about by tax evasion, tax avoidance, misapplication of laws, abuses on the part of tax officials, and so forth. It is the description that would be needed for the kind of tax analysis undertaken by economists. In a country where the tax administration functioned optimally, and where tax laws and regulations were unambiguous, there would not be significant distinction between statutory and effective tax systems. In such a country a tax on income would in fact be a tax on a correctly defined, or measured, concept of income and the tax law would provide a good indication of the tax effectively collected from income. 1/ The function of the tax administration would not distort the intention of the tax law. Equally, the statistical description would provide a faithful account of the initial impact of the tax system on particular tax bases. A sales tax would in fact be a tax on sales; an income tax would be one on income.

Unfortunately, tax administrations do not function optimally. In some cases, they can be so inefficient as to distort completely the intention expressed by the tax laws. Inefficient tax administration: (a) reduces tax revenue; (b) creates unintended distortions or non-neutralities in the tax system; this means that taxes affect markets in ways not intended by the legislators; (c) introduces various kinds of inequities through the tax system as for example between honest citizens and tax evaders. In other words, a poor tax administration will chaise the way in which taxation affects the traditional objectives of government policy, namely, allocation of resources, redistribution of income, and stabilization and growth. In this case and in terms of its effect on the economy tax administration become indistinguishable from tax policy, but it becomes a tax policy that is arbitrary and capricious.

The causes for inefficient tax administrations can be several:

They can originate in the tax laws themselves. These laws can be exceedingly complex, opaque, and requiring from the taxpayers the kind of information and attention that is difficult to provide. Tax laws have often become legislative labyrinths, represented by exceedingly complex tax codes, in which only a few can find their way. 2/ When interpretation can vary between those who administer the taxes (or even among them), and those who pay them, administrative problems are bound to arise. 3/ At times, the laws may be clear but the incentives for the taxpayers may be wrong. For example, in some countries, the penalties for delaying the payment of taxes are so ridiculously low that taxpayers simply do not bother to send the tax payments at the time they fall due, thus creating tax arrears and administrative problems. For them borrowing from the government becomes the cheapest source of credit.

They can originate in the political system. At times those who make the laws write them in ways that will protect themselves, or those associated with them, against high tax liabilities. 1/ In our work with countries (at the IMF) we have found that those who propose the tax laws and those who have to approve them may not be able to separate their role as taxpayers from their role as policymakers. The view that policymakers will pursue the public interest regardless of the impact that the laws will have on them as taxpayers is often naive, especially for taxation. Public choice considerations are especially relevant in this area.

They can originate in the tax administrations themselves. Lack of resources, lack of the necessary degree of professionalism, and lack of a clear strategy for improving efficiency all play a role.

Lack of resources is often blamed for poor tax administration. One often hears that in order to improve its effectiveness, a tax administration needs some significant increase in its budget and, particularly, in its personnel. There is no question that there are cases where this is true. However, tax administrations are notoriously inefficient in the way they use their available resources (both personnel and money). Personnel are often assigned to tasks that have very low productivity while important functions go unattended. It will be one of the main objectives of this paper to argue that a reallocation of the resources available to the tax administrations, away from traditional tasks and toward more productive ones, can result in a much improved administration even when total resources available do not change. However, this requires a clear strategy for reform and the flexibility to make the necessary changes.

Lack of professionalism often results from a tradition where (a) there are no rewards for taking initiative and for doing a good job; (b) where routine determines the tasks carried out; (c) where salaries are low and undifferentiated between good and bad workers; (d) where tax inspectors are not above accepting favors or even bribes from taxpayers in exchange for preferential treatment; and (e) where political influence in the work of the tax administration is not rare. When the salaries of the tax inspectors are low, the temptation on their part to take bribes or favors from taxpayers become high and society becomes more predisposed to condone this behavior. This is particularly so when decisions within the tax administration are seen as responding to political pressures. At times, cynicism lowers the resistance to accepting bribes.

Motivating tax inspectors whose salaries are low to fulfill their public function in an ethical manner is a common problem. However, an appropriate, effective incentive structure for rewarding tax inspectors whose efforts can improve tax collection and compliance has proven elusive. 1/ Some countries have adopted schemes in which a portion of the revenue from the collection of delinquent taxes and penalties is awarded as a bonus to the tax inspector responsible for collection. However, in some cases, this has produced perverse results. Specifically, tax inspectors have purposely not informed taxpayers of their tax obligations so that violations take place, penalties are collected, and bonuses paid. 2/ Sometimes, taxpayers pay as much in penalties as in taxes, creating the suspicion of collusion between the tax inspector and the taxpayer to raise the tax inspector’s income through penalty bonuses and lower the taxpayer’s total payment of taxes and penalties. The design of incentive structures to improve tax inspectors’ job performance is a constant challenge in tax administration.

Lack of a clear strategy for reform is probably the most important reason why many tax administrations continue with their inefficient practices, although the lack of a strategy may be symptomatic of a lack of political commitment to reform. Such a strategy must be simple, must consist of few basic steps, and must not require unavailable resources. This strategy rarely consists of copying what other, and often different, countries are doing. This issue is taken up more directly in the next section, along with other essential elements of tax administration reform.

To a certain extent the causes for inefficient tax administrations can originate in the economy itself or in the culture of the country. For example, the American system of administering taxes depends much on the mail system. Such a system would have difficulties in a country where the postal service is unreliable. A country with an atomistic structure of production will have greater difficulty in establishing a good tax administration than a country where industries are more concentrated. 3/ And, to some extent, the history of a country, and especially the way taxpayers view the government, may have some effect on their propensity to evade taxes and, thus, on the tax administration. However, in our view, this aspect is often exaggerated. Given the right institutions and incentives, taxpayers tend to behave in the same way regardless of where they reside. For example, following many years of poor tax administration and low compliance, Argentina, after having reformed its tax structure to make it more efficient and fair, turned to overhauling its tax administration. The reform of the tax administration encompassed a massive campaign to inform the public of their tax obligations and the consequences of noncompliance as well as higher salaries but less secure jobs for tax administrators. The ratio of tax revenue to GDP rose in Argentina from 14.7 percent in 1989 to 20 percent in 1993 in spite of the elimination of almost four percent of GDP in highly distortive taxes. Similar experiences are found in Chile, Peru, Mexico, Uganda, Ghana, and other countries. These experiences indicate that given the right incentives and institutions and the necessary political support, the efficiency of tax administration can be improved. 1/

II. Essential Elements of Tax Administration Reforms

Reforms of tax administration do not just happen. They must be made to happen. To be successful and durable, they require several important elements of which six could be considered essential. These are:

  • An explicit and sustained political commitment.

  • A team of capable, hard-working officials dedicated full-time to tax administration reform.

  • A well-defined and appropriate strategy.

  • Relevant training for staff.

  • Additional resources to the tax administration or at least some reallocation of existing resources.

  • Changes in incentives for both taxpayers and tax administrators.

All these elements must be present for the reform to be successful. In the rest of this section, we will briefly discuss them.

Political Commitment

Reforms of tax administrations, and tax reforms in general, require time as well as the explicit commitment of the country’s government. Here, when we talk about “the government”, we are really talking about one, or perhaps a few, powerful political figures who have the stamina, the interest, and the vision to push the reform through and to stay with it over a period that is likely to extend over several years. These individuals become the mentors of the reform and must be willing to assume responsibility for failure and must also be able to claim some “property rights” in the success of the reform. Usually, they must be willing to spend some of their political capital to launch the reforms because the benefits from the reform are typically widespread and delayed in time, while the costs are concentrated and occur early. These costs rise when the losing groups are politically powerful and well organized. There is, thus, always a risk to the government and especially to the individuals who sponsor and are associated with the reform.

There are two important implications that follow from the above. One is that such a process will be started and continued only when a government has a long time horizon in front of it. The rate of discount that the government will use to make the benefit-cost evaluation of the reform will often be determined by how long it expects to remain in power. A government that is on the way out is unlikely to show much interest in this activity. The second implication is that a government that remains in power but whose members keep changing will have less chance of seeing a major reform accomplished. Successful reforms in tax administrations are always characterized by the involvement over an extended period of time of a powerful political mentor who provides the reform with a face and with a kind of trademark. Examples of strong political leadership in successful reforms can be found in Argentina, Bolivia, Chile, Ghana, Mexico, Peru, Taiwan Province of China, and Singapore. For example, in Argentina, both President Menem and Minister Cavallo adopted the reform and provided the needed political support to Carlos Tacchi and Ricardo Cossio who were directly responsible for carrying the reform through. In Bolivia the mentor was Ramiro Cabezas, who, as Minister in Charge of Revenue, created a new tax department while reforming the tax system in a revolutionary way. In Peru, Minister Boloñas provided the needed political support.

Planning Team

Major administrative reforms cannot be a part-time or an occasional activity for those responsible for them. They require the full-time commitment of a working group that must plan the strategy, determine the sequence of the needed changes, and follow, on a daily basis, the steps to be taken. These individuals cannot have other major operational distractions. They must be fully committed and totally immersed in a process that is likely to be time consuming. This group is responsible for keeping the relevant minister or ministers fully informed through regular briefings. 1/

Effective teams have comprised individuals who know the workings of the current administration but are not an operational part of it. When this is not the case, they tend to be too concerned about the shifts in status and power that the reform inevitably brings within the current administrative structure. Within the tax administration, there will be winners and losers as a consequence of the reform. Furthermore, the team must be ruthless enough to get rid of, or, at least, to isolate individuals who are likely to interfere with the new mechanism that is being put into place, or who have been associated with corrupt practices. In all the countries where reforms have been successful, there has been a substantial pruning of the work force of the tax administration to get rid of corrupt or inefficient individuals. This pruning requires strong political support and relaxation or reform of civil service rules to be carried out. In extreme cases, where the existing administrative structure was so corrupt or so inefficient as to be considered beyond reform, a parallel tax administration has been put into place.


The reform of the tax administration will not be successful unless it conforms to a well-defined plan or strategy. Reforms that have failed have done so either because of lack of political commitment, or because no team had been set up with the responsibility to push it through, or because they lacked a clearly defined strategy. For example, a recent report dealing with the tax administration of a large developing country listed literally hundreds of faults with the existing tax administration without providing a sense of which are important and which should be eliminated first. In other words, it provided no sense of a strategy and no indication of the sequence of the steps to be undertaken. No wonder that the tax administrators of the country were baffled on what to do. 1/ There are also examples in which foreign advisers essentially suggested a strategy that was no less than implementing the tax administration procedures of the home countries of these advisers. Local conditions, particular characteristics of the statutory tax systems, local customs, and other limitations were ignored.

There is no strategy that is mechanically implementable in all countries. In each case, the strategy has to be adjusted to local conditions. Whichever strategy is chosen, it must be simple, must specify priorities and sequencing, and must be flexible enough to allow changes to deal with unanticipated events. In some cases the strategy can be gradual; in others it must be more drastic. Which is preferable depends on the country’s specific economic and other conditions. At times, the administrative reform must be accompanied by the reform of the tax system mainly because of the need to simplify the latter. In fact, reforms of the tax system have often preceded or accompanied reforms of the tax administration.

The possible approaches are essentially two: extensive and targeted. The extensive approach is the one implicit in the report on the major country referred to above. It attempts to improve the tax administration by changing everything at the same time. The targeted approach starts with improvements in a particular and well-defined area, which could be seen as a pilot project, and progressively enlarges the successful area. The discussion in the third section of this paper suggests the main elements of this targeted approach. Of course, the specific application has to be fitted to the conditions of the particular country. It is this approach that the International Monetary Fund has found more promising and more likely to succeed.


One area where large resources have been wasted has been training. To be effective, training must be aimed at achieving a well-specified and precise objective. If one wants to win a marathon, one must train specifically for it. General athletic training, while of some benefit, will not help much with the specific objective of winning the marathon. The same is true for tax administration. In this area, training must be linked to the adopted strategy. One must train to develop the skills necessary to implement the well-defined strategy. General training in tax administration is just too general to be of much use. For this reason, the benefit that comes from exposing the tax administrators of countries to those of more advanced countries is very limited unless this training conforms with the strategy chosen by the reforming country. Faced with a different environment and conditions, those who are exposed to general training soon go back to their old habits.

Thus, training is an integral part of the strategy for the reform of the tax administration but it must follow the decision on the strategy to be adopted and must be closely linked to it. In other words, the tax administrators must learn the tools that will be required by the chosen strategy and the training must teach or sharpen those tools.


A prevalent view is that the reform of the tax administration of a country requires considerable additional resources which often are not available to the country. These additional resources are in the form of personnel and materials and especially computers. One is often told that without a large spending for computers and without large additions to personnel, no administrative improvements are possible. The refrain is often heard that, with more personnel, there would be more auditing and more controls in general and thus a better tax administration.

The introduction of computer technology is, in fact, often a basic ingredient for substantive administrative reform. However, one mistake often made is to think that simply buying computers and making them available to the tax administrators is sufficient to improve tax administration. The offices of many of the tax administrations of the world are full of expensive computers which have accumulated dust because no one has the vaguest idea on what to do with them. Alternatively, when these computers are used, they are often used to perform electronically many of the same useless or low productivity tasks previously done by hand. In other words the computers are not used to introduce a new strategy. Unfortunately, donor countries, and even international institutions, have occasionally pushed particular countries to spend large amounts of resources on tools that proved to be of limited, if any, value. The basic principle must be the same as that for training. The purchase of the computers must follow, not precede, the development of a specific strategy. The computers, and the right kind of computers, must be acquired to serve specific and well-defined objectives and to perform useful tasks.

In the area of personnel, similar considerations apply. Traditional tax administrations generally misuse their personnel and resources on a truly large scale. Very often, a large share of the personnel is occupied with trivial, mechanical, and highly unproductive tasks, while useful tasks are starved for resources. 1/ The reason for this is that in many cases today’s allocation reflects yesterday’s needs. 2/ For example, in many cases a large proportion of the available personnel is busy receiving payments when banks could perform this function much more efficiently and cheaply. Often administrative or union rules prevent any reallocation of personnel. 3/ Furthermore, because of low salaries, it is difficult to attract the kind of personnel that can bring a new vision to the tax administration.

The reform of the tax administration must relax some of these constraints. It must allow firing. It must allow hiring of new personnel. It must allow salary differentiation and the provision of adequate salaries. It must allow the transfer of individuals from low productive to more productive functions or even from location to location. The tax administration must be seen as a firm that, given a budget and subject to some constraints, must maximize a product that in this case is revenue. This maximization may require a massive reallocation of the resources available to the administration. A good example is the transformation of the tax administration from an organization based on type of tax and taxpayer to an organization based on the key functions of tax administration. With enough flexibility and with a proper strategy, the reform may not require new resources. Before new resources are budgeted, it is important to ascertain that the resources available are being put to the best uses. 1/ For example, the shifting to the banking system of the function of receiving the tax payments is likely to release personnel resources that can be used to strengthen other functions such as auditing.


Successful reforms of tax administrations have, almost always, included changes in incentives aimed at encouraging better compliance on the part of the taxpayers and more efficient and more honest performance on the part of the employees of the tax administrations. Taxpayers who complied with the tax laws were assured a more courteous and less harassing relationship with the tax administration. Those who did not, faced more certain and heavier penalties. More efficient and honest tax employees of administrations were compensated with higher salaries and faster promotions. Less efficient or less honest employees faced dismissal or more severe punishments. For tax evaders penalties such as the temporary closing of their shops or places of business have proved very successful in several countries including Argentina and Peru. These penalties could be applied administratively and, thus, quickly. For corrupt officials or large scale tax evaders jail sentences have been applied in some countries for the first time. In some countries (e.g., Mexico) the application of these measures created the need to protect the main official responsible for the reform from the threat of affected citizens. 2/

Section III of this paper outlines the main elements of the targeted strategy mentioned earlier.

III. Reforming Tax Administration

A country’s tax administration represents an institutional arrangement in which citizens fulfill their tax obligations to the Government. Tax administrations should strive to be both effective and efficient. A tax administration is effective when a high level of compliance is attained among citizens. Efficiency is achieved when the administrative cost per dollar of tax revenue is minimal. Governments must be careful not to pursue efficiency at the expense of effectiveness. Achieving the balance between effectiveness and efficiency, or at least moving toward it, is the goal of tax administration reform. Measures to improve effectiveness and efficiency in a systematic manner are presented in the following two main subsections.

1. Interventions to improve effectiveness

Goal of voluntary compliance

The primary goal of tax administration in both developed and developing countries should be to achieve the highest possible voluntary tax compliance among taxpayers. Toward this end, the tax administration should present to citizens obligated to pay taxes procedures for minimizing the cost of complying with these obligations. It should also be clear that compliance costs, at least in terms of time spent interacting with the tax administration, will increase if taxpayers do not fulfill their obligations properly. The procedures should reflect a logical sequence of actions undertaken by the tax administration to employ its scarce resources in the most efficient manner to process collections, identify instances of noncompliance, and take appropriate action.

Adoption of the principle of self-assessment

The tax administration resources devoted to routine collection of revenue can be minimized by empowering taxpayers to assess and pay their own tax liabilities according to prescribed procedures. A strategy to reform tax administration should take advantage of the fact that, given clear instructions, simple procedures, and sufficient encouragement, taxpayers can calculate and pay their tax liabilities on their own. The principle of self-assessment (and self-payment) is the foundation of a modern tax administration. Self-assessment permits tax administrations throughout the world to effectively administer collection of taxes with high ratios of taxpayers per tax official, and should be a basic principle guiding tax administration reform.

Fundamental need for taxpayer education

The principle of self-assessment implies that a fundamental requirement for tax administration to increase voluntary compliance is taxpayer education. Taxpayers must receive clear and concise information describing what is taxable, when taxes apply, how to calculate their tax liabilities, and exactly what they are required to do to pay their taxes. Also, forms and procedures for calculating and paying taxes should be as simple as possible. Given this, taxpayers who take full advantage of the opportunity to self-assess their tax liabilities and accurately follow instructions and procedures for paying these liabilities will employ minimal resources of their own and of the tax administration. An example of the importance of taxpayer education can be found in countries that have successfully introduced value added taxes in recent years, for example, New Zealand, Spain, Thailand, and Trinidad and Tobago. Most of these countries have allocated a large amount of resources, including the time of senior government officials, to the task of explaining the tax and the obligations of the taxpayers. 1/ The countries that have not done so have had much less success with this tax.

Prompt detection of problems with tax filing and paying

A critical step in tax administration is the prompt detection of taxpayers who fail to submit tax returns (stopfilers) and taxpayers who fail to calculate correctly their tax liability on their tax return or make timely payment (delinquent taxpayers). A key role of tax administration is to ensure that if the taxpayer fails to comply, something is going to happen--a notice will be sent, significant interest and penalties will be applied, and other actions taken. It is important to convey the impression that, in this area, big brother is watching.

Given the need for an accurate up-to-date registry in order to detect stopfilers and delinquent taxpayers, tax administration reform should address whether there is a problem with nonregistration. This can include surveys in which tax officials go door-to-door within a district and check all businesses to see if they are properly registered. Another option is to examine records of registered taxpayers in order to detect suppliers or customers who are not registered taxpayers.

Tax administration reform should be guided by the need for adequate resources to detect stopfilers and delinquent taxpayers. This encompasses the development and implementation of a system for tax returns processing and accounting that quickly and accurately records taxpayers’ transactions with the tax administration and identifies stopfilers and delinquent taxpayers. Whether this system is based on manual ledgers, computers, or other means depends on many factors, including the customary business practices and relative costs of labor and capital, in a specific country. The nature of the system is not as important as the fact that it constitutes an efficient system for processing returns and detecting noncompliance.

The Fiscal Affairs Department (FAD) of the International Monetary Fund has delved into great detail on the development of such systems in its technical assistance on tax administration in many countries. These details are not pertinent here except to say that they constitute one of the building blocks of the institutional relationship between taxpayers and the tax administration.

Improvement of audit coverage

Noncompliance extends beyond stopfiling and delinquency to encompass fraudulent actions by taxpayers to understate their tax liability in ways that cannot be detected in the collection process. In a modern tax system it is critical for taxpayers to believe that if they engage in such fraudulent actions there is a reasonable chance they will be caught and appropriately penalized. Also, it is important that taxpayers be confident that their competitors are paying the same taxes so they may compete fairly. This requires that an effective audit program be carried out by the tax administration.

Taxpayer compliance depends to a significant degree on their belief in the effectiveness of the audit program. Further, the chance of being audited will appear greater, and compliance will be higher, if the actual details of the audit program are not known. It is important that taxpayers believe that a larger percentage of taxpayers is audited than is actually the case. This suggests a rationale for the tax administration not to disclose the information regarding its audit selection strategy and the number of taxpayers to be audited in a given period of time. Instead, the tax administration should promote the view that a variety of factors may lead to an audit. Toward this end, the tax administration should show taxpayers that it has both the information and the capacity to detect violations.

An audit plan represents a strategy to detect violations as efficiently as possible using the different types of audit with the resources available to the tax administration. The plan should specify how many of the different types of audit should be conducted each year in each local tax office, with the proportions varying based on the registration of new taxpayers, business developments, and the results of previous audits. The plan should reflect an analytic approach to increasing the probability of detection. Specific strategies would likely vary from year to year. For example, one year it may be decided to conduct wide audit of taxpayers in the same or similar business, say restaurants. In another year, more audits might be conducted than usual on wholesalers. 1/

Application of adequate penalties

The tax system can be expected to function smoothly and yield anticipated revenues only if adequate penalties are imposed for violations that strike at the heart of the tax system, such as failure to file returns and to pay tax on time. It is important that interest provisions for late tax payment compensate the government for the time that the taxpayer has use of the government’s revenue. The total interest cost for late tax payment should exceed the interest rate for borrowing money--it should be less costly for the taxpayer to borrow to pay taxes rather than to delay paying taxes as a way to obtain cheap financing from the government. In addition, if registration and other requirements critical for a smooth functioning tax system are adopted, adequate penalties should be applied for violations of these requirements as well.

The theoretical literature on tax evasion has often concluded that if penalties are high enough, audit coverage can be limited. See Cowell (1990) and Allingham and Sandmo (1972). In a way, the literature has suggested that there is a trade off between administrative expenditure and size of penalties. However, when the penalties are very high and audits are rare, the penalties are unlikely to be fully enforced by the magistrates. 1/ Furthermore, the taxpayers can appeal these sentences and delay for years the application of the penalties. Thus, the penalties lose their deterrent effect. In many countries, it has proven more effective to apply smaller but highly visible penalties, such as the closing of shops for a few days or a few weeks as was done very effectively in Argentina, Chile, and Peru, than high penalties; see Tanzi and Shome (1993).

2. Measures to improve efficiency of tax administration

Along with an overall strategy outlined above for enhancing effectiveness, tax administrations can adopt a number of measures to focus their scarce resources in the most efficient manner for revenue collection and enforcement. The following measures should be regarded as options as it is unlikely that it would be useful or possible to pursue all of them. The appropriate options for a particular country depend on its circumstances. For example, and at the risk of getting ahead of the presentation, where the value-added tax (VAT) is a major source of revenue and a threshold is adopted for VAT registration, a large taxpayers unit and a VAT unit may be virtually the same.

Establishment of a large taxpayers unit

The establishment of a large taxpayers unit can be an effective initial step toward a modern tax administration. It represents a targeted strategy for improving tax administration. This special unit, possibly made up of an elite group of better trained and less corruptible officials, monitors collection of taxes from large taxpayers who, although not large in number, account for the major part, sometimes as high as 90 percent, of tax revenue from VAT and corporate income taxes. 2/ Such a unit can virtually guarantee the timely collection of most tax revenue and the prompt identification of stopfilers and delinquent taxpayers so that appropriate action, including the imposition of penalties, can be taken. A large taxpayer unit can significantly reduce errors in tax declarations and minimize delays in the collection of substantial amounts of revenue due to errors, late filing, and other violations. Such a unit properly focuses initial efforts in tax administration reform on the adoption of more efficient procedures for the more important taxpayers, rather than on small taxpayers whose revenue generation is not substantial.

A large taxpayers unit serves as a model for the adoption and expansion of modern tax administration practices for other taxpayers. Some form of monitoring large taxpayers has been adopted in over 25 countries, both industrial and developing; see Table 1 for a list of these countries and a brief description of large taxpayer monitoring in each. In many countries, large taxpayer units were introduced as pilot projects, thus allowing the tax administration to try out new methods and technologies. The successful implementation of such pilots demonstrates to the tax administration, which as a bureaucracy tends to be resistant to change, that it is possible and advantageous to modify procedures and systems. In many countries once these modern practices were tested by the large taxpayers unit they were applied to progressively large groups. For example, Argentina developed these new practices for a group of less than 1,000. It has expanded the group to more than 200,000.

Table 1.

Countries With Special Systems to Monitor Large Taxpayers

article image
Source: dos Santos, Paulo, “Administration of Large Taxpayers,” IMF, October 1994.

Only for the VAT.

Offices supervising large taxpayers perform desk audits only to correct inconsistencies detected in the returns.

A large taxpayers unit may include not just the collection but also the audit function for several reasons. The detection of evasion is important for the same reasons as the monitoring of compliance with filing and paying requirements for large taxpayers. In addition, large taxpayers usually carry out a high volume of transactions and have complex operations. They frequently have several branches of operations and sometimes branches outside the country. They have highly skilled accountants and lawyers, and may use forms of evasion that are difficult to detect, such as transfer pricing. These characteristics of large taxpayers often make it necessary to establish special monitoring units with highly specialized staff.

Adoption of a threshold for tax registration

In a modern economy, it is not feasible to collect all major taxes from every economic enterprise. The accurate computation of tax liability by a taxpayer, and its control by the tax administration, requires that the taxpayer be able to maintain books and records of a certain minimum standard. Furthermore, for the system as a whole to operate reasonably effectively it is not necessary that every enterprise should be liable to pay all taxes that apply to enterprises. For example, the objective of the VAT is to impose a tax on the value of goods and services consumed by individuals. This tax is collected from enterprises on the basis of their value added (and also on imports), as a matter of convenience. In effect, although such enterprises are specified as taxpayers in the VAT law, they are acting as “tax collectors”. The operation of such a system is not undermined significantly if small scale enterprises--which, even though their numbers may be very large, account for small proportion of the value of goods supplied to final consumers--are exempted from liability to pay the tax.

Almost all VAT systems exempt small scale enterptises from liability to pay the tax. 1/ For this purpose, the most satisfactory measure of the size of the enterprise is its annual turnover: those enterprises with turnover below a certain threshold are not required to register as VAT taxpayers (though they may be liable to pay other taxes, such as the corporate income tax). This provision is necessary to allow the tax administration to concentrate its limited resources for controlling and monitoring the VAT on those taxpayers which account for the major part of value added. The appropriate level for this threshold should be calculated in such a way that potential VAT revenues are reduced by only a small percentage.

Imposition of some form of alternative tax on small taxpayers

With the threshold proposed above, most small enterprises engaging in cash transactions would not be liable, in principle, to pay VAT on their sales, and would not receive any credit for VAT already paid on their purchases. Every country finds it difficult to tax the income (or value added) of small traders in an appropriate manner. To assess the true level of income reasonably accurately, when most transactions are not recorded, can absorb considerable resources and the revenues generated by this use of the tax administration’s limited resources are likely to be small.

Three main alternative solutions to this dilemma have been adopted in different countries. The first is to levy an annual license fee on small traders, as a rough substitute for liability to VAT, income tax, or both. Different license fees are often set for different types of business. Such fees need to be set at a level high enough to act as a reasonably fair substitute for income tax or VAT, but not so high as to discourage enterprise.

The second approach is a “presumptive tax”, under which the taxable income of the small business is computed indirectly on the basis of certain objective indicators--such as the value of goods for sale, the number of employees, total wages paid, and the amount and type of physical assets. Estimating taxable income on a presumptive basis can, however, require considerable effort on the part of the tax administration. 2/

The third general approach to taxing the incomes of small traders engaged primarily on cash transactions relies on the withholding of tax on their purchases. In many countries, imports by traders who are not registered for the VAT, and sales of goods to such traders by registered enterprises, are subject to a special withholding tax--which is typically around two percent of the value of the purchase. Traders who believe that the amounts withheld by their suppliers exceed the liability on their value-added have the option of registering for the VAT and submitting returns showing their actual VAT liability.

Use of final withholding

Withholding taxes on income when it is earned is a mechanism widely used in developed and developing countries for collecting revenue with minimal use of administration resources. Tax withholding is a very important mechanism given the large volume of revenue collected through withholding in many countries. A popular mechanism is wage withholding, commonly known as pay-as-you-earn (PAYE), in which employers are withholding agents for personal income tax on the earnings of their employees. 1/ This shifts the tax administration’s monitoring responsibilities from a large number of individual taxpayers to a much smaller number of employers, thus achieving a more efficient focusing of administrative resources.

Taxes can be withheld on sources of income other than wages by collecting from the payer rather than the payee with the same efficiency gains. This includes payments of interest by banks, dividend payments by companies to stockholders, interest and dividend payments by mutual funds, and other like payments. This process is facilitated when the taxes withheld become final taxes as in the case when a schedular approach is followed for the taxation of personal incomes.

Use of banks

Banks have a natural role in the administration of taxes as they constitute an institution well suited for receiving tax payments. Banks are accustomed to receiving payments and handling money, and are well equipped for this task. The main tax administration services provided by banks entail the receipt of payment forms, the receipt of money from taxpayers in payment of taxes, the batching of payment forms and balancing with monies received, the deposit of these monies into the proper tax accounts, and the submission of payment forms to the tax administration. In some countries, banks are also responsible for processing tax returns combined with payment forms; this system has substantially improved the efficiency of collection. The responsibility for maintaining an accurate record of taxpayers’ transactions and enforcing the collection of taxes remains, however, with the tax administration.

3. Organizational considerations

The measures discussed above for improving the effectiveness and efficiency of tax administration strongly influence the organization of the tax administration. Once the tax administration determines its key functions, an organizational structure should be adopted to support these functions. The discussion in the previous two sections suggests the following five principal functions:

  • Taxpayer education. Staff in this function are responsible for educating and assisting taxpayers in tax matters. They distribute printed material that explains tax regulations. They respond to inquiries from taxpayers. This is a very important function that in the past has often not receive the attention that it deserves. As mentioned earlier, Canada, New Zealand, and Trinidad and Tobago provide good examples of successful taxpayer education.

  • Registration, accounting, and returns processing. Staff in this function are responsible for data entry and the processing of declarations and payments processing. For instance, if a taxpayer uses the wrong tax rate, a computer can detect the error and automatically send a letter asking for payment at the correct rate. After the data have been verified, they are stored in a master file, which is a central record of all taxpayer transactions with the tax administration.

  • Collection enforcement. The emphasis of this function is on collection of taxes from stopfilers and delinquent taxpayers. Basically, all compliance control except inspection is centralized in this function. Much of the information that is needed is provided by the accounting and processing department.

  • Audit. The audit function is devoted to the detection of underreported taxes. Staff in this function usually perform audits of taxpayers’ files and records at tax administration offices or audits of business activities and records at the taxpayer’s premises. It is important to note that routine checks have already been performed by the accounting and processing department.

  • Legal services and appeals. The staff in this function take cases to court, defend tax authorities against legal appeals in the courts, and provide advice as to which actions are legal and which are not.

With a functional organization, the work flow within the tax administration is more specialized in order to best support a self-assessment system. The main objective is to identify the exceptions, that is, the instances of noncompliance. The major part of the staff concentrates on the exceptions, as well as on audit to check compliance. Such an organization entails a strategic element of self-checking among staff whereby the work in one function serves as a control on another. In other words, the organizational structure itself becomes a self-enforcing measure.

4. Concluding perspectives on measures to improve effectiveness and efficiency

Tax administration reform should represent a systematic effort to improve tax compliance and increase administrative efficiency. The common thread of a reform strategy should be to focus administrative resources on problems in the tax collection system through promotion of voluntary compliance and the adoption of a logical sequence of procedures for efficiently identifying and handling instances of noncompliance.

The focusing of administrative resources on successive problems in the tax collection system is illustrated by the need to control the four major tax gaps. The first is the difference between the potential taxpayers and registered taxpayers. The tax administration must make efforts to identify nonregistered taxpayers. The second is the difference between registered taxpayers and those who file returns. This corresponds to stopfilers. If the tax administration can quickly identify stopfilers, it can reduce the probability that large tax debts will accumulate. The third gap is the difference between the tax reported as due and the tax actually paid. This gap corresponds to delinquent taxpayers. The fourth gap is the difference between the full tax liability that taxpayers should pay and the tax actually reported by the taxpayer. This gap corresponds to evaders. To improve the effectiveness of tax administration, all four gaps must be controlled. The goal of tax administration reform is to accomplish this as efficiently as possible with scarce administrative resources.

IV. Concluding Remarks

This paper has discussed issues related to the reform of tax administrations. It has started with a brief survey of the factors that cause the tax administrations to be inefficient. It has then proceeded to a discussion of the essential requirements for a successful reform of tax administrations. It has argued that given the right conditions, much progress can be achieved in a relatively short period of time. However, tax administrations cannot be reformed in months but, instead, require years. In Section III, the paper has outlined the main ingredients for a flexible strategy that has proved successful in a growing number of countries. This is a strategy in which foreign advisers can be useful provided that they have clear ideas and relevant multi-country experience.

The experience of several countries in recent years proves that tax administrations can be reformed and made more efficient. However, these successful experiences almost always occurred in countries which met several or all of the conditions outlined in Section II but which, additionally, were making concrete efforts to reform their economies. Tax administration cannot be an institutional island that reforms itself independently from the rest of the economy. Its reform will be facilitated or indeed it will be made possible by a general change of attitude on the part of policymakers which welcomes a greater reliance on market forces while, at the same time, attempts to stamp out corruption and cronyism and to modernize the workings of the public sector.

If the rest of the economy is not being reformed and if attitudes about the role of the public sector and of the market are not changing, it will be difficult to reallocate resources within the administration or to create incentives for both taxpayers and tax administrations to improve compliance. If populism remains strong, if union rules and other constraints continue to restrict policy actions, some of the essential elements of the reform will not be there.


  • Allingham, M.C. and A. Sandmo, “Income Tax Evasion: A Theoretical Analysis,” Journal of Public Economics, Vol. 1, No. 314 (November 1972) pp. 32338.

    • Search Google Scholar
    • Export Citation
  • Bird, Richard M. and Milka Casanegra de Jantscher, Improving Tax Administration in Developing Countries, IMF, 1992.

  • Casanegra de Jantscher, Milka, “Types of Tax Noncompliance”, IMF, FAD/82/4, August 17, 1982.

  • Cowell, Frank A., Cheating the Government: The Economics of Evasion (Cambridge: MIT Press, 1990).

  • dos Santos, Paulo, “Administration of Large Taxpayers”, IMF, Mimeo, October 1994.

  • Silvani, Carlos and Richard Fulford, “A Strategy for Administering the VAT in the Russian Federation”, IMF, March 1993.

  • Tait, Alan A., “Value-Added Tax: Administrative and Policy Issues”, IMF, Occasional Paper No. 88, October 1991.

  • Tanzi, Vito, “Potential Income as a Tax Base in Theory and in Practice”, IMF, DM/80/84, December 1980, published as Chapter 13 of V. Tanzi, Public Finance in Developing Countries (Aldershot, England: Edward Elgar, 1991).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito and Milka Casanegra de Jantscher, “The Use of Presumptive Income in Modern Tax Systems,” in Changes in Revenue Structures, Proceedings of the 42nd Congress of the International Institute of Public Finance, 1986, pp. 3251 (Wayne State University Press).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito and Parthasarathi Shome, “A Primer on Tax Evasion” in IMF Staff Papers, Vol. 40, No. 4, December 1993, pp. 807828. Also published in Bulletin for International Fiscal Documentation, 1994, Vol. 4, Number 617, pp. 328337.

    • Search Google Scholar
    • Export Citation
  • Van der Heeden, Koenraad, “The Pay-As-You-Earn Tax on Wages--Options for Developing Countries and Countries in Transition,” IMF, WP/94/105, September 1994.

    • Search Google Scholar
    • Export Citation

Prepared for the “IRIS Conference on Economic and Political Institutions for Sustainable Development: Implications for Assistance”, October 24-25, 1994, New State Building, Washington, D.C. Forthcoming in Institutions and Economic Development, Christopher Clague, editor. The views expressed are strictly those of the authors. They should not be interpreted as official IMF positions. This paper has been much influenced by discussions with Mrs. Milka Casanegra de Jantscher to whom the authors owe a great debt. Comments received on an earlier draft from Prof. Christopher Clague were much appreciated.


This discussion ignores the issue of tax shifting.


In many countries the tax laws may not even have been compiled in a code.


For example, at this moment in Italy, there are more than three million cases of litigation between taxpayers and tax administration concerning the proper tax liability. Each case must be resolved by commissions. This will require many years. The Italian situation is far from unique.


In some cases a small clause in the drafting of the particular law which may not attract attention at the time the law is passed is all that is needed to create a particular and well targeted loophole.


However, maj or tax administration reforms have often been accompanied by higher salaries for the personnel of tax administrations.


In some of these countries it is very difficult or almost impossible for a taxpayer to obtain a copy of the tax laws and, especially, the tax regulations.


The atomistic structure of the Italian economy has at times been blamed for the poorer tax administration and the higher tax evasion of Italy as compared to, say. the United Kingdom where the productive structure is much more concentrated.


In Argentina and Peru, Presidents Menem and Fujimori provided strong support to the fight against tax evasion. President Menem made several Saturday speeches to condemn tax evasion and to urge Argentine taxpayers to comply with their tax obligations.


Obviously, ministers of finance, with their many commitments and strong pressures on their time, cannot provide to the reform process the day-to-day attention that it needs.


In that country the current attempt at reform is being carried out by the existing bureaucracy without the creation of a planning team. It is unlikely that this attempt will be successful in carrying through major, as compared with marginal, changes.


In other cases the misallocation of resources may be on a regional basis.


At times they reflect past political pressures to provide employment for some individuals.


The minister of finance of a large country once remarked to one of the authors of this paper that he did not even have the power to move a security guard from one entrance to another entrance of the ministry. It is not surprising that in that particular country problems of tax administration remain important.


However, the need to replace “dead wood” or “corrupt wood” with new personnel may require some new resources. As indicated earlier, in some of the major tax policy and tax administration reforms it became necessary to get rid of existing organizations and replace them with new ones.


There is almost no economic literature that has analyzed the incentives to be provided to the tax administrators. This should be a fertile area for research.


In Thailand, the Permanent Secretary in the Ministry of Finance and other senior personnel spent many months meeting with large groups of taxpayers to explain how the tax worked and the taxpayers obligations


Such a strategy has recently been announced by the Italian tax administration.


In this case the penalty system violates one of the fundamental criteria of taxation, namely horizontal equity. This occurs because the tax evaders who get caught are highly punished while those who are not caught escape the penalty.


These large taxpayers also account for a sizeable share of the income taxes withheld at source.


This exemption covers the value-added (the margin) of small enterprises, but not the VAT on their purchases, for which they remain liable.


For a discussion of presumptive taxes, see Tanzi (1980) and Tanzi and Casanegra (1987).