The Art of Tax Policies

Tax policies are a good starting point for seeing the policymaking process in action. First and foremost, a good number of public policy decisions are related to taxation, and taxation touches almost every aspect of the economy and society. The size of the State, the amount of redistribution from the rich to the poor, and decisions to consume and invest are all related to this fundamental area of policy. Given its potentially large effects on efficiency and equity, tax policy is perhaps the area of public policy where the most interests are at stake.

Thus the policymaking process (PMP) for taxes tends to be a good reflection of the broader (“global”) PMP: that is to say, the process through which a host of interests, both public and private, work their way through the wheels of political negotiation in the making of public policies. As a general rule, actors that play key roles in the broader (“global”) PMP are also active players in the process of discussing, enacting, and implementing tax policies. This applies not only to actors that are central in the PMP in every country—notably the executive (whose powers and constraints strongly influence the quality of tax policies) and the legislature—but also to other actors whose influence is more specific to certain countries, such as regional authorities in Argentina and Brazil, the judiciary in Colombia after the 1991 Constitution, or business groups in Guatemala.

Despite the far-reaching economic impact of taxes, economic considerations do not go very far in explaining the features of tax policies. Countries with similar levels of income, income distribution, or sectoral composition of output have very different tax structures. There is clearly no economic model that explains tax policy outcomes.

Moreover, the tax structures and policies that countries put in place are often far from ideal. Some countries have tax revenues that are too low or too high, even when most of the main actors and observers acknowledge that this is a problem (for example, Guatemala and Brazil). There are countries that rely heavily on inefficient or distortionary tax revenues (Colombia) or have tax systems that are full of exemptions (Costa Rica and Paraguay), although simpler and more efficient systems would be technically preferable. Then there are those countries that approve reform after reform because each gets watered down in the approval process (Colombia). These outcomes cannot be justified by economic considerations alone.

Furthermore, each and every country in Latin America has faced the same external shock to its economy in the past 15 years: the trend toward globalization. This has increased the international mobility of goods, investments, and financial capital around the world, thus limiting the possibilities of taxing them. Globalization trends have forced countries to slash import tariff rates and trim tax rates on business. In response to this common shock, countries throughout the region have come up with a wide range of tax structures and tax policies. This variation across countries exemplifies how, even in response to a common shock, different political institutions and political actors can result in different policy outcomes.

This chapter examines four very different country cases—Brazil, Colombia, Guatemala, and Paraguay—and discusses the general features of the tax policymaking process. The analysis reveals that to a large extent, differences in taxation reflect differences in political institutions and structure, which interact with some important features of tax policies.

Profound Changes in Tax Regimes

At first glance, the landscape of tax policies in Latin America is hardly encouraging. Tax revenues of the typical central government in Latin America reached 13.2 percent of GDP on average in the first years of the current decade, a decrease from the average of 14 percent in the late 1980s, and less than half the current average of 30 percent in developed countries. Only a handful of countries in the region, including Brazil, Bolivia, and the Dominican Republic, succeeded in raising tax collections by more than 3 percentage points of GDP over that period. But such a superficial evaluation turns out to be mistaken. Behind this apparent stagnation in taxes are hidden very profound changes in tax regimes that countries have had to make to respond to globalization.

In the face of the challenges of globalization, taxation policy has been a very active area of reform. Every Latin American country has undertaken important reforms in this area since 1990—an average in the region of 4.2 such reforms—and 11 Latin American countries have overhauled their tax systems since that time. Taxes on the incomes of businesses and individuals have risen from 2.7 percent of GDP in the 1980s to 3.9 percent at present, although they remain very low in comparison to averages worldwide or in developed countries (12 percent of GDP). The loss of revenues from the reduction of import tariffs has been largely compensated for by the revenues produced by the value-added tax (VAT), which was introduced in most countries between the mid-1980s and the mid-1990s and now generates more than a third (37 percent) of all tax revenues of the region’s central governments: equivalent to 5.5 percent of GDP, on average.

Why, given this surge of activity, have the policies adopted often been less than ideal? While tax policymaking proceeds against the backdrop of each country’s broader (“global”) PMP, specific features of tax policies constrain the possibilities for policy reform. These constraints go a long way toward explaining the peculiarities of the tax policies discussed, enacted, and ultimately implemented and enforced.

Among the main features of tax policies that must be kept in mind, four are especially important. First, taxes have wide-reaching effects throughout the economy. This means that many players will be active in the policymaking process. If they perceive that their interests are directly or immediately affected, they are likely to be more intensely involved. The policymaking game in taxation is therefore more likely to have a large number of players.

Second, taxation is subject to severe common-pool problems. That is, taxation may be viewed as a game in which each individual player wins or loses depending on his ability to extract more than he contributes to a common pool of resources. Each individual has incentives to minimize contributions to the pool: for instance, by being exempted from paying taxes or by taking advantage of a loophole to pay less than others. Each individual also has incentives to extract as much as possible from the common pool by earmarking some taxes or assigning some share of the tax pool for specific uses or regions he favors, as is the case in Argentina, Brazil, or Colombia.

Third, the effective implementation of tax policies relies heavily on the capabilities of the tax administration office. The experience, resources, and administrative capabilities, and even the preferences and biases of the tax administrators, determine what is put into practice. Thus, as is often said, “In developing countries, tax administration is tax policy.”

Fourth, once tax systems are structured in a given way, they are very hard to change. In the parlance of social scientists, they are strongly path-dependent. What makes tax systems so remarkably hard to improve? In principle, any common-pool problem can be solved by a cooperative agreement among the participants—provided the costs and uncertainty of the negotiation are low compared with the expected benefits, and enforceable mechanisms can be put in place to avoid opportunistic behavior and ensure compliance.

Unfortunately, these conditions rarely exist in taxation issues, for several reasons. Most fundamentally, since economic power is usually very concentrated and powerful elites are able to influence the political system through a variety of channels, they tend to be protected from the vagaries of the political system: richer taxpayers are often able to prevent reforms that would affect them negatively. This argument can be labeled the elite resistance hypothesis.[2]

Further complicating matters, the revenue and distributional effects of structural changes to the tax system cannot be easily predicted. Given that many Latin American countries face a precarious fiscal situation, fiscal authorities are often deterred from pursuing deep reforms in the face of uncertainty about revenues in the short and the medium run, even when there might be confidence in the long-term benefits of reform. Tax administrators are intrinsically very conservative, as implied in the adage “Good taxes are old taxes.” This explanation can be called the uncertainty hypothesis.

Moreover, if a country is beset by political instability or polarization, a government or a political party in power may prefer not to pursue a reform that makes the tax system more efficient and productive for fear that, while the political costs of the reform will fall upon the government or party itself, the benefits will be showered upon the next government. This is the strategic argument.[3]

Finally, there is the enforcement argument, which states that welfare-enhancing reforms can be implemented only if an “enforcement technology” can be put in place to ensure compliance and punish defectors. The superiority of self-compliance over punishment has been well documented in the literature on tax compliance. But a culture of self-compliance is hard to build and easy to destroy, because it rests on trust in public institutions, the legitimacy of those in power, and a feeling of fairness, transparency, and reciprocity in the collection and use of public revenues. In the absence of such a culture, some external enforcement mechanism is necessary to facilitate effective reform. The IMF has played this role in some countries: the probability of tax reform is higher during a fiscal adjustment program undertaken in the context of an IMF loan agreement.[4] The exact channel is not clear, but it may have something to do with greater consistency of tax policies or control over expenditures, or it may simply be that the IMF shifts the political costs away from the government. An effective and transparent tax administration can also be seen as an additional enforcement mechanism, as will be discussed below.

The Immediacy, Pervasiveness, and Complexity of Taxation Effects

The effects of taxation vary in their immediacy, pervasiveness, and complexity, complicating both policymaking and the analysis of tax policymaking. Some taxes work many of their effects through the economic system in a short period of time. Others have long-term consequences that are far more important (think of taxes that affect saving and investment decisions). Another challenge is disentangling the direct or partial equilibrium effects of taxation, as opposed to the indirect or general equilibrium effects. For example, some taxes displace productive activities in important ways, therefore reducing the base for collection (an increasingly relevant factor in a globalized economy). General equilibrium considerations are of particular importance in relation to the burden of many taxes, which may be transferred from those that initially face the obligation to other firms or individuals. In general, the burden of direct taxes, such as income and property taxes, is less easily transferable than that of indirect taxes, such as sales or value-added taxes, which tend to be shifted to the final consumers. Thus the visibility of the burden varies from tax to tax, and in many cases depends heavily on technical details usually beyond the grasp of those ultimately affected. Finally, the effects of taxation are in many cases highly uncertain, introducing another complexity into the PMP.

There are two important implications of the immediacy, pervasiveness, and complexity of taxation effects. First, as noted, since taxes have such far-reaching economic and policy impacts, the PMP for tax policy tends to mirror a nation’s broader (“global”) PMP. Second, tax systems should be expected to have multiple bases, varied rate structures, and a myriad of special provisions as a result of the policymaking process.[5] Since it is reasonable to assume that support for any political party depends on how that party affects the interests of its current and potential supporters, every party will try to tailor the tax mix in order to best suit those interests. As a result, a variety of tax bases, rate structures, and special provisions should be expected. Even a one-party system will be interested in introducing some complexity into the tax system because, by doing so, it may cater to a variety of its supporters.

However, in any political system, the forces that push toward making the tax system more complex are constrained by at least two factors. First, each party’s supporters may also benefit from the expenditures that are financed by the common pool of taxes. Thus for some specific aspects, the players may find it more beneficial to shift the policymaking game toward the budget process. Second, tax collection costs will increase as the tax system becomes more complex. This will divert resources from other uses that may produce higher political benefits and will reduce total tax revenues. Thus there is pressure to reduce costs and keep the system simpler if this would lead to higher tax revenues, which can potentially improve the outcomes for some or all players. Tax systems can be seen as the result of these conflicting forces.

Since meddling with the many features of tax systems is essentially a political game, it is not surprising that tax systems are not only complex but also subject to continuous change. However, most changes are gradual, rather than radical. Countries with tax revenues that are too low remain in that situation for long periods; tax systems that are clearly inefficient are improved only slightly year after year; and so on.

The last feature of tax policies that needs to be stressed is the central role that the tax administration office plays in the implementation of tax decisions.[6] In the tax policy process, there is a clear separation between the stages of policy enactment and policy implementation. Although the tax administration office usually contributes to drafting tax reform proposals, it plays an entirely passive role during the process of discussion in congress. Often, congress introduces changes into the draft that may look inconvenient or impracticable to the tax administrators, without consulting them. However, once the reform is approved, the tax administration office has wide latitude in deciding how, when, and where the new tax code will be implemented.

Tax administrators are usually under pressure from the minister of finance to increase effectiveness in the collection of tax revenues. Faced with this pressure, tax administrators prefer to concentrate collection efforts on the largest contributors through large-taxpayer units. Over the last two decades, to facilitate compliance among small firms and to reduce the costs of monitoring and collecting revenues, tax administration offices have introduced simplified taxation schemes for small firms (whereby a single combined payment is made for income taxes, VATs, and excise taxes). Although usually effective in reducing collection costs, these systems may introduce important horizontal inequities and displace the burden of some taxes in ways that lawmakers did not intend.

A Look at Four Countries

Colombia: Changes in the Balance of Power

Colombia has been a very active reformer since the early 1990s, but the reforms have come out of congress substantially diluted and altered. The changes in the PMP introduced by the 1991 Constitution, discussed in detail in Chapter 7, are largely responsible for these difficulties, as they weakened the executive vis-à-vis the legislature and the judiciary.
Since 1991, the size of the State and its mandate has grown considerably. Until 1991, the public sector was relatively small by regional standards. Between 1990 and 2003, aggregate public expenditures jumped from 21.2 to 33.7 percent of GDP, reflecting a deliberate intention in both the constitution and the political system to increase the size of the State and use fiscal expenditures for redistribution. The fact that total revenues grew from 20.6 to 29.7 percent of GDP in the same period suggests that the decision to raise government expenditures was accompanied by an important effort on the revenue side—although an insufficient one. The central government’s deficit has been close to 6 percent of GDP since the late 1990s.

The executive’s major objective in the tax reform process has been to increase tax revenues as a means of reestablishing fiscal balances. However, the draft tax reform projects submitted to congress have also given importance to the structure of the tax system, an area in which successive administrations have been only partially successful. Reliance on the VAT has been increasing, and the VAT rate has increased from 10 to 16 percent through the various reforms since 1990. However, the most recent attempts to widen the VAT base or raise the rate have had very limited success, in part because the legislature assigns greater priority to the progressivity of the tax system than to its efficiency—regardless of the progressivity of the expenditure structure. This has severely limited attempts by the executive to reduce the dispersion of VAT rates and the number of exemptions.

Colombia is characterized by very high income tax rates, and is one of the few countries (along with Argentina and Bolivia) that has increased corporate and personal income tax rates since 1990. Currently, the corporate tax rate is 38.5 percent, the highest in Latin America, and the personal income tax rate is 35 percent. The effectiveness of these rates, however, is severely undermined by a host of exemptions and loopholes. Consequently, very few pay direct taxes, but those who do pay bear an excessive share of the burden. Similarly, although the average VAT rate has increased substantially, the tendency has been to increase the number of rates, with the declared intention of making the system more progressive (by assigning lower VAT rates to basic goods and higher rates for luxury goods), thus distorting it considerably. As a result of the high tax rates and their low effectiveness, the tax system as a whole is considered one of the least neutral in Latin America.

As a consequence of congress’s reluctance to widen the income and value-added tax bases, the executive has introduced new—and highly distortionary—tax sources. In 1998, a temporary 0.2 percent financial transactions tax was adopted through an emergency decree. This was raised to 0.3 percent and made permanent in the 2000 tax reform, and raised to 0.4 percent in the 2003 reform. In the same vein, through an extraordinary “internal commotion” decree, the Uribe government adopted a temporary net wealth tax earmarked for the strengthening of democratic security in Colombia: the proceeds go to financing the military and social operations necessary for ending the guerrilla and terrorist insurgency. This tax was extended for three additional years in 2003. Thus revenue pressures have led to decisions that disregard the basic principles of an equitable and efficient tax structure. For example, as a concession to the business community for enacting this tax, a temporary tax holiday was enacted simultaneously, consisting of a 30 percent deduction for reinvested earnings in capital, which according to the ministry of finance ended up costing the Treasury more than what the net wealth tax yielded in 2004.

The inability of the political system to deliver more efficient tax reforms is explained in part by the changes introduced into the PMP in the Constitution of 1991. While the constitution preserved the prerogatives that make the Colombian president the main agenda-setter in most policy areas, it reduced presidential powers in a number of dimensions. It deliberately curbed the legislative powers of the president by limiting to 90 days the declaration of either a state of internal commotion or a state of economic emergency, and by establishing that the decrees issued remain in force after the emergency only if congress enacts them in regular sessions. In this way, the constitution severely curtailed a method that had been used several times to enact major tax reforms. For example, the government used special legislative powers to enact the 1974 tax reform—which incorporated many of the recommendations of international experts. In 1997, the Samper administration attempted to use economic emergency powers to tax capital inflows, but the Constitutional Court declared this unconstitutional because it did not consider conditions pressing enough to justify an emergency. Nonetheless, against the backdrop of two economic emergencies (during the 1998–2002 Pastrana administration) and one internal commotion (during the 2002–06 Uribe administration), various administrations have been able to introduce new temporary taxes, which were extended by regular legislation with the approval of the Constitutional Court. Accustomed to these practices, Colombians say that “nothing is more permanent than a temporary tax.”

Since 1991, the Constitutional Court has been a key player in the policymaking process. (It is more active and independent than its predecessor, the Supreme Court.) The main reason is that many policy issues, particularly tax reform issues, were elevated to constitutional status by the Constitution of 1991. Around 10 percent of the total legal claims on economic matters handled by the Constitutional Court since 1991 have been tax issues. In 1999, a ruling on the financial transactions tax limited the executive’s scope in the use of resources. More recently, the Constitutional Court denied the approval of the generalization of the VAT and the taxation of specific activities.

Another trend that has been reinforced by the constitution—the increase in the number of political parties and the factionalization of existing parties—has made the normal passage of legislation through congress more difficult for the executive. Electoral rules traditionally used in Colombia (the “Hare” quota system, whereby the majority of seats end up being allocated to the largest remainder of candidates) generate incentives for parties to fragment into factions, presenting multiple lists of candidates for congressional elections in each district.[7] The result is that parties have increased the number of lists over time, maximizing their share of seats, while enhancing decentralization and factionalization. While the constitution did not change the electoral rules, the trend toward party fragmentation was reinforced around the time of the constitutional reform with the introduction of a system of direct public funding for congressional and presidential campaigns, where the political movement rather than the party is the recipient of the funds. These reforms lowered the costs of challenging party hierarchies and created room for small party factions to influence the legislative process. To some extent these centrifugal forces are contained in congress, where the main committees, such as the budget and tax committees, are controlled by recognized party leaders with deep knowledge of fiscal affairs, who lead the debate and who are influential in the legislative outcome. However, to ensure the support of the larger number of parties, the executive resorts to nominating a larger number of sponsors for each bill—which implies that it must deliver a larger amount of “pork” to get the bills passed. In addition to increasing the cost of passing tax bills, the legislative process also reduces the benefits by watering down government proposals. This cost-benefit analysis led the Uribe government to withdraw its latest tax proposal in December 2004. The government proposed a VAT reform, but legislators wanted to raise wealth taxation instead. The government opted for keeping the status quo.

In sum, the Colombian experience highlights how changes in the balance of power between the executive and other branches influence the enactment of tax reforms. In earlier decades, congress largely rubber-stamped the tax reforms the executive submitted, oftentimes through emergency legislation. Since the early 1990s, however, legislative involvement in the design of tax packages has been increasing. Congress has passed eight tax reforms since 1990, but it tends to water down the proposals during debate, not only in terms of revenues, but also, more importantly, in terms of the quality of the reforms. Likewise, in earlier decades, the judiciary played no significant role in the process of approval of reforms. However, the greater independence and extended powers it received from the revised constitution to oversee the enactment of laws has severely limited the room for maneuver of both the executive and the legislature in tax issues.

Brazil: Inaction in the Face of Uncertainty

Brazil is a puzzling case. It has a cumbersome and inefficient tax system that has resisted a badly needed overhaul. Yet over the last two decades, total revenues have increased and the productivity of the major taxes compares very well with that in other Latin American countries. Brazil now has the highest tax burden in Latin America and one of the highest in the developing world. As a federal country, Brazil is one of the most fiscally decentralized countries in Latin America­—so much so that its main source of fiscal revenue, the VAT, is collected mainly at the state level, a feature seldom observed in the developing world. As in other decentralized countries, vertical imbalances across states are partly compensated for by transfers of national tax revenues to the states. The possibility that the states can resort to VAT revenues reduces the probability of states running large deficits that can sometimes be associated with large vertical imbalances, at the cost of creating inefficiencies and problems of coordination and competition among states. How has Brazil managed to weather the coordination and enforcement problems of its complex tax system? As will be argued, much of the answer lies in the combination of a powerful executive and a strong tax administration office.

Gross federal government tax revenues increased from an average of 16.5 percent of GDP in 1985–89 to 24.4 percent in 2000–02. The total national tax burden reached 34 percent of GDP in 2000–02, up from 24 percent in 1985–89, as state and provincial taxes also have increased. It is predicted to rise to 38 percent of GDP by 2005: a rate roughly similar to that of Great Britain in the 1990s.

The centerpiece of the tax system is Brazil’s version of the VAT, the ICMS, which is collected by the states and represents about a third of all tax revenues (excluding social security). As a result of changes in the Constitution of 1988, which gave governors a central role as part of the transition to democracy, states were allowed to set different rates for the ICMS. The constitution also deepened the process of fiscal decentralization by increasing the mandated transfers to municipalities and states of the main federal tax revenues, namely, the income and industrial products (IP) taxes. These transfers currently represent 3 percent of GDP, or about half the federal revenues of those taxes.

The constitution imposed other rigidities in the use of fiscal resources, reducing the flexibility of the executive. This has led the federal government to resort to taxes that are not shared with the municipalities and states, contributing to an increasingly inefficient tax system. The two major sources of additional revenues are the financial transactions tax (CPMF), which was introduced at the end of 1993 and has been abolished and reintroduced several times since then; and the Social Security Financing Contribution (COFINS), which was associated with increases in tax rates and a series of court rulings favoring the federal government. Since the mid-1990s, temporary (so-called “extraordinary”) revenues have become commonplace and widely used. Extraordinary revenues peaked at over 3 percent of GDP in 1999 and contributed 2.5 percent of GDP in 2000–02, on average, suggesting that temporary taxes were a success in terms of revenues. However, as a result, the structure of the tax system has deteriorated.

As in Colombia, the new constitution introduced fiscal rigidities that weakened the central government and reduced the discretion of the executive in tax issues. However, unlike in Colombia, the executive was able to regain its capacity to impose its fiscal preferences. While governors were central during the democratic transition, their power lessened with the passage of time (as noted in Chapter 4), partly because of the executive’s significant powers to shape the legislative agenda and to build support for the enactment of legislation. In fiscal issues, the governors derived their power from their substantial tax powers, and the prerogative of the states to own banks and public enterprises. However, after the monetary stabilization of the Plan Real (1994), the fiscal situation of the states deteriorated. As a condition for extending federal bailouts, the federal government was able to impose privatizations of banks and public enterprises, as well as other conditions. This culminated with the enactment of the Fiscal Responsibility Law of 2000, which acts as an enforcement mechanism that improves the effectiveness of the subnational tax system and mitigates the common-pool problem of the federal tax system (Box 8.1).

An effective tax administration has also been instrumental in maintaining the productivity of the tax system, despite all its complexities. Since its inception in 1969, the Brazilian Internal Revenue Service (Secretaria da Receita Federal, SRF) has received the support of the government and society at large, partly due to a long tradition of strong public administration that goes back to the process of formation of key bureaucracies during the period of the monarchy in the 19th century. Brazil was one of the first countries in the world to introduce a comprehensive VAT, generating revenues equal to 26 percent of GDP by 1971. The more than 13,000 tax auditors at the federal level were among the best-paid career civil servants in Brazil throughout the 1980s and 1990s. Meritocratic recruitment and low turnover have been permanent features of the SRF, except under the Collor administration (1990–92), which politicized the agency and curtailed its administrative and functional autonomy.[8] Since then, modernization has continued unabated and the SRF’s enforcement capabilities have been strengthened. Information on tax collections, tax legislation, and taxpayer services is provided on the Internet. About 90 percent of personal income tax returns are filed through the Internet, and all corporate income taxes are filed electronically. A 2001 law allows bank secrecy to be broken for tax enforcement purposes. Paradoxically, the effectiveness of the tax administration office may have eased the pressure to overhaul the cumbersome tax system.

As noted, the power of the executive has been instrumental in introducing discipline in the tax relations between the national and the subnational governments and also in strengthening the tax enforcement capabilities of the tax administration office. However, the federal government has not been able to implement comprehensive tax reform. This failure illustrates the great difficulty in moving to a new equilibrium, especially when the system has become more complex. In late 1997, and not for the first time, the idea of a deep tax reform began to circulate. The government considered some proposals that were quite radical. The main proposal was to discard turnover and cascading taxes,[9] as well as state VATs, and replace them with three new taxes: a consistent broad-based nationally managed VAT; a new federal excise tax on a small number of goods and services; and a local retail sales tax. After the Asian and Russian crises and election to a second term, President Cardoso decided to raise the issue in public in 1999. Over the next 18 months, tax reform dominated the political debate. Eventually, it was impossible to coordinate such a move; only opposition and stalemate resulted. While it may have superficially looked like a situation where the executive did not push hard enough on reform, once one considers the complexity of the change expected, it is not surprising. There was uncertainty about the revenues that the new tax system would raise, uncertainty as to whether the intergovernmental compensation rules would continue, and uncertainty about whether the distribution across states would be preserved.[10]

Thus, although everyone agrees that the indirect tax system in Brazil needs to be reformed to remove cascading taxes, no agreement has been reached on how to get there. While the president is typically able to pass his own agenda, the overhaul of the tax system has been prevented mainly by the uncertain fiscal effects of such reform, not only on aggregate tax revenues, but also on each state’s tax revenues and composition. For instance, São Paulo would have lost revenues from the introduction of a general consumption VAT, but would have benefited from the elimination of the VAT on exports; the opposite would have happened to Paraná. The difficulty in measuring whether the gains would offset the losses, combined with an aversion to losing revenues, meant that it was hard to build a winning coalition that would support the passage of the reforms.

In 2003 President Lula’s government once again made a concerted effort to arrive at a consensus on tax reform. Originally the main pillars of the reform were unifying the VAT, putting the 27 state codes under a single national value-added tax, and converting the main cascading taxes into a non-cumulative tax. Although the reform was planned to be revenue-neutral at every level of government, congress did not support it. The government decided to concentrate on measures that had the best chance of being approved, and postponed or shelved the most controversial aspects of the reform. The unification of the VAT was thus postponed to 2005, and the number of tax rates allowed was reduced to five (as opposed to one). The proposed reform of the cascading taxes was watered down. Moreover, tax relief for exports and capital goods was granted. The federal government enjoyed partial success with the replacement of the payroll tax with a turnover tax and the extension of two temporary taxes. But the price was a transition to lower-quality taxes. Despite all the progress in the negotiation process, politics took over. With so many interests and players in such a complex issue, the reform process was temporarily shelved in 2004. The biggest problem, once again, was uncertainty created by the multidimensional nature of the reform.

While overhauling the tax system has been impossible, the government has been able to pass piecemeal tax reforms through ordinary legislation several times in the last few decades. Some important reforms have resulted, including the restructuring of the corporate taxation system, the introduction of norms to curb transfer pricing, and the creation of an entirely new simplified system for taxing small business.

In sum, the Brazilian case clearly shows that although a powerful executive and a capable and well-respected tax administration are instrumental in enforcing the tax code and inducing cooperative behavior, both from the subnational levels of government and the taxpayers at large, they may not be enough to introduce deep reforms into the tax system. Path-dependence is difficult to break when there is a high number of players and multidimensional reforms are needed to overhaul the tax system, thus creating uncertainty about the revenue effects for each individual player.

Guatemala: Exploring the Political Underpinnings of Low Taxation

There is broad social and political consensus in favor of increasing taxation in Guatemala, yet attempts to raise the tax burden substantially have failed repeatedly. A key reason why has been the lack of an effective political counterweight to the strength of the business sector.

Tax revenues in Guatemala have traditionally been among the lowest in Latin America. The tax burden was 7.9 percent of GDP, on average, in the second half of the 1980s, and it declined further, to just 7.4 percent of GDP, in the first half of the 1990s. Following the stabilization of the economy and a series of tax reforms, tax revenues increased significantly, to 8.9 percent of GDP in the second half of the 1990s, and to 10.2 percent of GDP in 2000–2002.

Instrumental to these improvements were the Peace Accords signed in 1996. To finance the social and infrastructure expenditures identified as priorities to cement peace, a 12 percent tax burden was widely agreed upon as a target that should be met (originally in 2000). Most of the increase in tax revenue was produced not by taxes on business profits or income taxes, but by the VAT, which jumped from 2.5 to 4.4 percent of GDP between the early 1990s and 2000–2002. Income taxes increased, but only from 1.6 to 2.4 percent of GDP, through the introduction of a new tax on the income of agricultural and business firms (known as IEMA, for its Spanish acronym).[11] However, even this incomplete achievement proved to be short lived. In 2004 the new tax was overturned by the Constitutional Court in response to legal action by some organizations in the private sector, and tax revenue remained at 10.3 percent of GDP in 2004, the same figure as in 2003. New attempts by the administration of Óscar Berger to reach the 12 percent target have faced opposition not only from some business groups but, surprisingly, also from some popular organizations.[12]

The history of Guatemala is replete with instances of failed attempts to increase taxation. During the administration of Julio César Méndez Montenegro (1966–70), the minister of finance, Alberto Fuentes Mohr, was removed from office after pushing for a tax reform that was opposed by certain business interests. In 1982, under the de facto administration of Efraín Ríos Montt, the entire economic cabinet was dismissed after attempting to introduce a package of ambitious tax reforms to curtail the fiscal deficit. In 1984, the new military government of Óscar Humberto Mejía Víctores decreed several new export and consumption taxes and raised others, prompting opposition from some business organizations. These organizations retaliated by leaving the goods affected by the taxes in customs and successfully pressed for repeal of the measures.[13]

The influence of business interests in tax matters left its mark on the Constitution of 1985, which limits the power of the State to levy tax revenues through “constitutional locks” that prevent any type of “double taxation.” This legal provision has been invoked on several occasions, apart from the recent repeal of the IEMA. For instance, in 1987–88, during the administration of Vinicio Cerezo, the Program for National Reorganization lost its economic base when taxes that had been previously agreed upon with the business sectors were later challenged on the basis of provisions in the new constitution.

As this brief history of taxation reveals, the ability of some private interests to block or reverse tax measures is deeply entrenched in Guatemala. Business and business organizations around the world pursue their interests in a variety of ways, depending on the perceived opportunities for influence offered by the political system. As discussed in Chapter 5, business is most likely to mobilize against the adoption of a certain policy when its costs are immediate and certain. This is the case with several taxes, such as direct taxes on business profits. It is not surprising, then, that some business organizations in Guatemala oppose attempts to raise taxation. What needs to be explained is why they are so effective in achieving their objectives. Part of the explanation is that the business sector has traditionally been under the control of a small number of families and concentrated in a few sectors of economic activity, especially agriculture and commerce. These two features, according to the discussion of the role of business in Chapter 5, would mitigate problems of collective action, and lead to more intense and effective participation by the business community. But this is hardly the whole explanation. The effectiveness of the business sector in influencing policies essentially rests on the relative institutional weakness of the Guatemalan State.

As the Association of Research and Social Studies of Guatemala, a prominent policy research group, explains, “a government that starts almost from scratch every four years, political parties that do not know if they will survive legally after governing, and a diffuse social movement, without significant representation and with little capacity to formulate and advance policy proposals, do not create an adequate context for the emergence of an organized business participation with a focus on the long-term. It is rather an appropriate context for the development of narrow and short-term business interests that prevail in the exercise of their disproportionate influence.”[14]

A vivid example of the effectiveness of the business sector in the face of weak institutions is the recent history of the Fiscal Pact. When it became clear that it would be impossible to meet the target of a 12 percent tax burden in 2000, as agreed in the Peace Accords, the Peace Accord Accompaniment Commission, at government request, agreed in 1998 to reprogram the target for 2002. Along with the postponement, a series of additional actions were taken: extension of a temporary tax (IEMA), the elimination of the deductions related to the VAT and income tax; changes to the Free Zone Law; resumption of the IUSI property tax; the contracting of activity verifiers for foreign trade; and, most importantly, the promotion of a consultation process on a Fiscal Pact, under the auspices of the ministry of public finance and the Accompaniment Commission itself.

A committee of prominent professionals prepared a draft Fiscal Pact in consultation with various economic and political groups. After the public presentation of the source document, the issues were widely publicized and discussed in a rich and very participatory process including representatives from universities, study centers, labor unions, campesino organizations, women’s organizations, chambers of commerce, departmental consultation groups, mixed and non-mixed committees, peace institutions, nongovernmental organizations (NGOs), and cooperative federations. Political parties participated only to a limited extent, although party participation increased appreciably at the time of the signing of the pact. The process ended successfully with the signing of the Fiscal Pact by the heads of the three branches of government, together with the Accompaniment Commission and representatives of the participating social, political, and economic groups.

In the political climate at the moment, which was not particularly auspicious to the business sector, leading private representatives participated in and supported this process, adopting an attitude of dialogue and cooperation, even on issues that directly touched their interests, such as tax breaks and exemptions (although not in relation to the “constitutional locks”).[15] However, a short time later, the balance of power moved back in favor of the business sector—and the Fiscal Pact was never fully implemented. Although the highest authorities of the State had formally backed the process and various political parties had endorsed the document at the last moment, the executive and legislative branches ultimately did not support or respect the pact for various political reasons.

The new tax reform proposed in 2004 by President Óscar Berger also failed to obtain political backing. Paradoxically, on this occasion, the strongest opposition came not from the business sector but from a group of popular organizations and NGOs. Despite the attempt to incorporate progressivity and control of evasion into the bill, they decided to reject the changes, viewing the changes as originating with a government which they perceived as being too closely allied with the business sector.

The result was the reinstatement of the status quo: one in which lack of confidence in the capacity of the State, in the context of a system in which parties are weakly disciplined and relatively unprogrammatic, confers great de facto power on groups that tend to oppose attempts at reform.

Paraguay: Laying the Groundwork for Major Change

In sharp contrast with Brazil, Paraguay does not have a tradition of strong public administration that commands the respect of the public and facilitates the enforcement of the tax code. However, unlike the tax system of Brazil, that of Paraguay is free from many of the complexities of decentralization and the diversity of interests and fears that have prevented Brazil from overhauling its tax system. A recent tax reform became possible when a progressive party sided with the executive to respond to public demands for transparency and effectiveness.

The tax burden in Paraguay is one of the lowest in Latin America. The tax revenues of the central government barely increased from 8.5 percent to 9.7 percent of GDP between the late 1980s and 2000–2002. There is no personal income tax, other tax rates are low, exemptions are plentiful, and evasion is pervasive,[16] the latter reflecting a very weak and poorly financed tax administration and a very high level of informality in the economy.

Since 1990, when Paraguay moved toward democracy and witnessed the reduction of tariff barriers of its large neighbors, Argentina and Brazil, a 50-year-old paradigm of smuggling across borders has begun to break down. Among the early reforms to address the high level of informality was tax reform legislation, Law 125/91. The original law was biased against the industrial sector, while the agricultural sector was barely taxed. There was no personal income tax, and it was not particularly hard for small businesses to be exempt from the VAT. Initially the law had two features: it afforded considerable flexibility to the executive, which could decide which articles to apply, and it even allowed some latitude for the executive to decide the rate and base of various taxes. Various aspects of the law and its regulations were ambiguous and subject to interpretation, thus opening up the measure to a series of changes. Taking advantage of its authority to initiate bills on tax issues and bills targeted to the private sector, congress passed some 42 reform laws over the following decade,[17] adding several exemptions, particularly for the industrial sector. The largest was a five-year holiday on all taxes for firms that presented reasonable investment projects, even if they were never carried out. This clearly allowed any firm to be eligible for the tax holiday, and to continue reapplying.

Although the tax rate on profits was kept at 30 percent, firms on average effectively paid a tenth of that as a result of exemptions and quasi-legal alternatives. Moreover, politically and economically powerful agricultural interests, well represented in the government and congress, prevented the updating of land values for tax purposes. Partly as a consequence, the tax base value of agricultural land in Paraguay averages only 5.6 percent of its market price.[18] The agricultural sector accounts for more than 20 percent of GDP, but directly contributes less than 2 percent of total tax revenue, as extremely low land taxation makes it very cheap to preserve idle land in Paraguay. A personal income tax was established only recently and is not expected to be fully operational for years to come. The VAT rate of 10 percent (the lowest in South America) also had numerous exemptions, many introduced through amendments in the 1990s. (Most of these exemptions were eliminated by a law passed in 2004.) Another important indirect tax is the excise tax on diesel fuel (used by 85 percent of the vehicles in Paraguay). The price of diesel is regulated, and so the tax rate has suffered various modifications, the result of negotiations between the state oil company, Petropar, the central government, and the agricultural sector.

Paraguay has historically had a two-party system. The Colorados retained power the longest, even through non-democratic means, until the early 1990s. Attempts at overhauling the tax system began in 2001, but repeated efforts never acquired enough steam to get through congress, particularly in the face of a progressively weaker government with low credibility, a highly factionalized Colorado Party, and determined opposition spearheaded by the Liberal Party in congress. A new reformist party, Patria Querida, gained considerable popularity during the presidential election race in 2003, as it ran on a platform of change and opposing corruption, an endemic problem in Paraguay. The party was led by a group of progressive technocrats from various sectors of society, particularly from the business sector. As a result, the two-party hegemony broke down. Although the Colorado Party candidate, Nicanor Duarte Frutos, won the election, his party did not win a majority in congress. The seats were more or less equally divided among the three parties: Colorado, Liberal, and Patria Querida. Moreover, Duarte came from the reformist faction of the Colorado Party, and while the rhetoric was populist and appealed to the rural less-well-off constituents, in practice the policies were pragmatic and reformist and thus met with some resistance from the traditional faction of his own Colorado Party.

Duarte was elected on a promise to attack corruption at the highest levels. Immediately after taking office, he replaced the top levels of government with individuals who were renowned for their integrity and professionalism and were independent of the traditional Colorado Party apparatus. This included replacing the heads of the tax and customs administrations (which were perceived as the most corrupt institutions in Paraguay) and providing full support to a technocratic group of policymakers. The government then proposed its own strong reform agenda. One of the most important aspects was tax reform. The draft tax law in its purest version essentially eliminated tax exemptions and equalized the rates of both direct and indirect taxes at 10 percent, strengthened tax administration by adding independence and accountability, and increased financing by earmarking a percentage of tax revenues for the tax administration authority. The government made very strong efforts to include as broad a set of groups as possible in the dialogue on tax reform, including civil society, business groups, and trade union groups. To create additional external pressure, the administration secured a commitment for funding from multilateral and bilateral development organizations, conditional on an IMF program. In the end, the ruling Colorado Party supported the law, the Liberal Party opposed it, and Patria Querida was left in a very strong bargaining position to support the law on the condition that certain features opposed by its special interests were removed. It thus became the main veto player. In the end, the law was watered down. The share of taxes to be paid by the agricultural sector increased, but remained relatively small. Nonetheless, through a significant broadening of the tax base and the elimination of some egregious exemptions, the law was estimated to increase the tax revenue yield by 1.3 percent of GDP in the long run.[19]

Although the new law (Ley de Reordenamiento Administrativo y de Adecuación Fiscal) clearly moved in a desirable direction, in practice the traditional pressure groups continue to lobby for the postponement of some aspects of the law. (The executive has the power to decide the timing of implementation of the articles of the law.) Remarkably, even before the law went into effect, the new government was able to increase tax revenues by 40 percent between August 2002 and 2003, in part through some efficiency measures, but mostly as a result of better compliance and a reduction of internal corruption in the tax administration.

The experience of Paraguay suggests that major change is possible in tax systems that are seriously underperforming when the political circumstances shift to enable the executive and public-minded politicians to introduce sweeping changes. Of course, this story has yet to unfold fully. Two main challenges persist. First, the tax and customs administrations are not sufficiently financed or independent to perform their mandates adequately. The recent successes in tax collection may not be sustainable unless the resources and powers of the tax administration are increased and maintained. Second, the trend toward eroding the law through amendments may repeat itself. Congress can initiate reforms that benefit particular interests that are very strong and have influence in the political parties and with individual politicians. As the other case studies have highlighted, the executive must exert its power to prevent those trends from eroding a very promising tax reform.

Conclusion

Why do some countries have higher tax revenues relative to their GDPs than others? Why do some countries opt for indirect taxes, while others prefer direct ones? Why do some countries have simple tax regimes, while others prefer complex structures with large numbers of exemptions? Why is taxation so hard to change?

To answer these questions, this chapter has explored some political and institutional aspects of the tax policymaking process. Drawing closely from the experiences of Brazil, Colombia, Guatemala, and Paraguay, the analysis has found that, to a large extent, differences in taxation reflect differences in political institutions and structures, which interact with some important features of tax policies. The wide-reaching effects of taxation throughout the economy make cooperative solutions difficult when the number of players is large. The common-pool nature of taxation revenues creates incentives to renege on cooperative agreements. It also necessitates enforcement mechanisms to implement tax policies effectively. These problems may be acute in fiscally decentralized countries, especially if subnational authorities have strong powers vis-à-vis the executive and, in general, in countries where political power is fragmented. Taxation systems are very highly path-dependent because of the resistance of elites; uncertainty about the revenue consequences of the reforms; the strategic interests of weak governments or parties in polarized political systems; and the difficulty of putting enforcement mechanisms in place. Path-dependence is more likely to be overcome in more centralized systems, especially when the balance of power favors the executive and leadership is strong.