Taxation in developing countries: some recent support and challenges to the conventional view.

The general advice given by international institutions such as the International Monetary Fund (IMF) and the World Bank to developing countries over the past few decades has been to replace trade taxes with domestic consumption taxes, particularly value-added taxes (VAT), and to maintain relatively high corporate income tax rates. This article reviews recent literature that supports and challenges this conventional view.

I. INTRODUCTION

The seminal work on the subject of taxation and development was done by Burgess and Stern, who reviewed previous literature and presented what is still, thirteen years later, regarded as one of the most important works in the field. (1)

According to their view, developing countries should have an indirect tax system based on the following elements:

(i) a VAT with one or two rates and some exemptions; (ii) excises on

alcohol, tobacco, petroleum products, and some luxury goods; and (iii)

direct support for certain groups, possibly through subsidized

rations.... [Such a system] may be supplemented by temporary tariffs

to maintain revenue or where infant industry arguments have genuine

empirical support.... (2)

Burgess and Stern strongly support the now common movement away from trade taxes to sales tax in general, and to value-added taxes (VAT) in particular. (3)

The role of direct taxation in developing countries is much more limited. In contrast to developed countries where taxation on personal income and social security contributions raises two-thirds of the total tax revenue, a narrow tax base and high enforcement costs in developing countries render direct taxation impractical. (4) The income tax base in developing countries is mostly comprised of the wages of public sector employees because most other taxpayers are self-employed or small businesses who evade paying all, or most, of the income tax. (5) In addition, taxation of personal capital income is easily evaded. (6)

Taxing the income of corporations, on the other hand, provides developing countries with a large portion of their total tax revenue (estimated in 1993 to be close to one-third), compared to only a small portion (less than one-tenth) in developed countries. (7) Taxing large corporations does not involve significant administrative and compliance costs because the companies are already forced to comply with statutory accounting requirements. (8)

Following this conventional wisdom, the general advice given to developing countries over the past few decades by international institutions such as the International Monetary Fund (IMF) and the World Bank has been to replace trade taxes with domestic consumption taxes, particularly VATs, (9) and to maintain relatively high corporate income tax rates. (10)

Some recent literature challenges this conventional view, arguing that the relatively large informal sector in developing countries may justify a different tax policy design. (11) In a previous article, Margalioth suggests that maintaining high corporate income tax rates may come at a high cost in terms of economic growth as corporate tax rates affect foreign direct investment (FDI) location, and may cause significant spillover effects. (12)

This article is structured as follows. Part II provides some general background information. Part III reviews some recent papers that support the conventional wisdom. Part IV reviews some recent criticisms of the conventional wisdom. Finally, Part V draws upon this body of recent research to raise a few questions and offer suggestions for future study.

II. BACKGROUND

Taxes are necessary to overcome the free riding inherent in the financing of public goods, to control market imperfections, and to achieve social justice through redistribution. Economic growth (efficiency) is promoted by the first set of goals, whereas social justice (equity) is promoted by redistribution and the provision of public and merit goods, most notably health and education.

Literature on the subject generally assumes that the goals of promoting economic growth and social justice are shared by developed and developing countries; however, a number of major differences between developed and developing countries may call for different tax designs. These differences include variations in industry type (primarily the relatively high shares of agricultural and small businesses in developing countries), in the size of administrative and compliance costs, in the levels of corruption, in the levels of monetization in the economy, in political constraints, and in the relative size of the informal economy.

The overall proportion of government expenditure of the gross domestic product (GDP) is higher in developed than developing countries. (13) This situation is not necessarily optimal, because a greater need exists for government intervention in developing countries (e.g., for building infrastructure and education) than in developed countries. On the other hand, the costs of corruption, administration, and compliance are much greater in developing countries, making the outcome of this trade-off unclear.

The portion of total revenue comprised of non-tax revenue is, on average, larger for developing countries than developed countries. (14) Nevertheless, the main source of government revenue in developing countries, taken as a whole, is the tax system. (15)

The structure of taxation in developing countries is radically different from the structure of taxation in developed countries. About two-thirds of the tax revenue in developed countries is obtained from direct taxes, mostly personal income tax and social security contributions. (16) The remaining one-third comes primarily from the domestic sales tax. (17) The situation is exactly reversed in developing countries, in which about two-thirds of the tax revenue comes from indirect taxes. (18) These indirect taxes include the VAT, the sales tax, and excises and taxes on trade. (19) The remaining one-third comes primarily from the corporate income tax. (20)

Since the 1980s, developing countries have undergone frequent tax reforms, gradually replacing trade taxes with domestic consumption taxes, particularly the VAT. (21) These reforms were part of two world-wide trends that affected developed countries as well. The first trend was economic liberalization and adherence to the World Trade Organization (WTO) requirements, which called for the elimination of all barriers to free trade. (22) The second trend was the rapidly increasing popularity of the VAT all over the world. (23)

The purpose of replacing trade taxes with domestic consumption taxes was principally to improve macroeconomic stability and to introduce the benefits of free trade to developing economies. (24) Export taxes are seen as inefficient because they put the local producers who export their goods at a disadvantage compared with foreign producers. The VAT was viewed as more efficient than import taxes because it does not discriminate between domestic and imported goods. By eliminating import taxes, local consumers benefit from the lower prices created by the competition between domestic and foreign producers. (25) Additionally, eliminating import taxes forces local producers to become more efficient and concentrate their efforts on their comparative advantage. (26)

Equity considerations, namely reducing poverty and inequality, have been of secondary importance, when considered at all, in the tax reforms of developing countries. Focusing only on efficiency can result in the adoption of regressive tax policies. (27) For example, taxes on goods with low price elasticities of demand, such as some cereals and domestic fuel, are efficient in that they do not distort behavior. (28) However, since the poor consume these goods disproportionately, equity considerations will weigh against the adoption of such regressive tax policies. (29)

III. RECENT LITERATURE SUPPORTING THE CONVENTIONAL WISDOM

Gemmell and Morrissey have analyzed the distributional impact of tax reforms in developing countries. (30) They conclude that "[t]he available evidence suggests that sales taxes are slightly more progressive, or less regressive, than taxes on imports." (31) Additionally, Gemmell and Morrissey have found that, in most developing countries, "export taxes were regressive, typically incident on smallholder agricultural producers (who, if not actually poor, were relatively low income). The removal of such taxes, combined with the reduction of other implicit taxes on agriculture, should have had a favourable impact on distribution and the poor." (32) As a result, Gemmell and Morrissey conclude further that "it seems likely that the reforms will not have worsened the effects of the tax structure on distribution and the poor." (33)

Gemmell and Morrissey add to the conventional wisdom an emphasis on the distributive effects of consumption taxes. They argue that no conclusive evidence exists "regarding the impact of generally replacing tariffs with sales taxes, largely because we do not know enough about economic incidence and the implications of a large informal sector," such as those prevalent in developing countries. (34)

They cite the example of excise taxes on fuel as a particularly troublesome manifestation of this problem. (35) Gemmell and Morrissey assert that a tax on kerosene (or paraffin) may have high social costs (36):

Kerosene (or paraffin) is often important within poor household[s,

where it is used for heating, lighting, and cooking] but is not widely

used by the rich. Thus, not only will kerosene taxes be harmful to the

poor, but [kerosene can be exempted] from more general fuel taxes to

improve equity without encouraging inefficient substitutions between

fuel types. (37)

On the other hand, it is not entirely clear that taxing gas for cars has a low social cost. (38) Gemmell and Morrissey argue that "taxes on intermediates such as fuel are often thought to be regressive because they affect transport costs (thus increasing prices of goods consumed by the poor)." (39) They suggest that "[t]he important implication for tax policy is that, on distribution and poverty grounds, taxes on goods that are most important in the consumption bundles of the poor should be kept as low as possible." (40)

The distributive consequences of conventional reliance on consumption taxes may seem especially grim when considering Latin America in general, especially Central America, where poverty and inequality rates are exceptionally high. (41) Bird acknowledges such inequality is "[t]he central social and economic problem in many Latin American countries." (42) He argues, however, that this is primarily a political problem. (43) The rich elite in Latin America have a great deal of political power, but they refuse to reduce inequality. (44) Bird notes that a positive correlation exists between inequality in Latin America and the extremely unequal distribution of land ownership. (45) He therefore suggests increasing the very low property tax rates now in force and improving tax administration through "more comprehensive coverage, better assessments, more frequent assessment revaluations, and enforced penalties for late payment." (46)

Bird goes on to advocate "[r]eforms that link taxes and benefits more tightly ..., such as decentralization and more reliance on user charges." (47) He argues that, "contrary to popular rhetoric, most user charges are progressive in their incidence. The property tax, and some local business taxation, may be considered to be 'generalized user charges' if properly designed and implemented." (48)

In addition, Bird argues that pro-poor spending programs are more effective in reducing poverty than the tax system. Nonetheless, at the same time, it is important to "untax" the poor by "setting higher thresholds for certain taxes or charges (e.g. lifeline [utilities]) or granting certain exemptions from the VAT." (49)

In addition to taxing land, Bird suggests imposing taxes on estates. Imposing a wealth tax on the top 1-2% of society, even if badly administered, "may sometimes be worthwhile not only in symbolic but also in actual terms." (50) For example, a 1% annual wealth tax for property yielding a 5% average annual return has the same effect as taxing that return at 20%. (51)

Bird supports the conventional view of tax reform in developing countries by stressing that, despite the extreme inequality in Latin America countries, "the best tax system [for them] is the one that produces the most revenue in the least costly and distorting way." (52) Such a system is "[a] broad-based VAT, and not a steeply progressive income tax." (53) Bird does note, however, that this system should be supplemented with taxes on land and other property, good user charges, and taxes on motor vehicles and fuel. (54)

These ideas are further developed by Bird and Zolt. (55) Bird and Zolt stress that, in developing countries, expenditure policy is much more important for redistribution purposes than income tax policy and consumption taxes can be progressive and should be supplemented with user charges. They further emphasize that "[g]reater fiscal decentralization (moving tax and expenditure authority to lower levels of governments) may allow for better matching of those who benefit and those who pay for government activity," Bird and Zolt then add a series of suggestions on methods to improve income tax enforcement. (56) In addition to the standard administrative advice detailed below, they suggest greater reliance on presumptive taxation and the adoption of the Nordic dual income tax system. (57) They also propose that "[t]ax authorities could try to increase the number and types of individuals subject to withholding [taxes] on labor income, for example, by expanding the definition of employee for tax purposes beyond [that required by] employment law." (58)

Bird and Zolt further recommend "heavier reliance on withholding (for example, by banks) and [on] third-party information." (59) Taxpayers who fail to withhold or report information on payments they have made should not be allowed to deduct those payments for income tax purposes. Bird and Zolt describe a range of options for tax authorities, including "using taxpayer identification numbers, outsourcing routine data processing, adopting case-tracking systems, and improving and expanding audit systems." (60)

Bird and Zolt contend that the use of presumptive taxes "is surprisingly widespread in taxation around the world, though [such taxes appear] under many names." (61) Taxable income is estimated by the authorities on the basis of "coefficients for different factors applied to specific taxpayers, specific types of taxpayers [for example, the location as well as the number of chairs in a restaurant], or in some cases on more aggregate indicators, such as industry and region, or external indicators of income." (62) The authors indicate that "[s]uch taxes are intended to capture at least some minimum level of tax from entities, regardless of either their reported or their true net income." (63) Usually, the taxpayer is allowed to rebut the presumption by proving his or her true income.

A dual income tax system imposes a flat tax on income from capital. In developing countries, this has the advantage of including in the tax base capital income that was previously exempt, and improving enforcement and compliance by allowing the fixed withholding tax rate to be the final tax. (64)

The income tax proposed by Bird and Zolt is designed to be modest reform (justified in whole or in part on symbolic grounds) designed to complement "a broad-based VAT, appropriate excise taxes, more use of local and benefit financing, and above all, an improved expenditure policy." (65)

IV. RECENT LITERATURE CRITICIZING THE CONVENTIONAL WISDOM

In a 2005 article, Emran and Stiglitz challenged "the current consensus that favors a reduction and eventual elimination of trade taxes, and almost exclusively relies on VAT as the instrument of indirect taxation in developing countries." (66) In their 2005 article, they argue that the consensus "is built on fragile results derived from a partial model that ignores the existence of an informal sector." (67) Instead, they contend that "the results from a more complete model demonstrate[] [that replacing trade taxes with VAT] can reduce welfare under plausible assumptions" (68) and conclude that "[t]he results raise serious doubts about the wisdom of the indirect tax reform policies pursued by a large number of developing countries." (69)

In an earlier version of the article, Emran and Stiglitz asserted "that while a radial (across the board) uniform reduction in trade taxes reduces the production distortions and the distortions between tradable and nontradable sectors, a revenue-neutral radial increase in VAT increases the inter-sectoral distortions between formal and informal sectors." (70) That is, goods may be produced and sold in both the formal and informal sectors, but the consumption tax is only paid by the formal sector and creates a distortion between formal and informal sectors, resulting in a potential reduction in aggregate welfare.

In their 2005 article, Emran and Stiglitz "extend[ed their] analysis to the case of a selective reform of trade tax and VAT in an economy with an informal sector." (71) The term selective reform refers to tax changes that apply only to a subset of the commodities falling under the tax net. (72) In such a context, they state:

Michael et al .... show that, in a tradables-only economy with no

informal sector, a reduction in the import tariff on the commodity

bearing the highest tariff and also the highest total indirect tax

burden increases welfare under suitable assumptions of

substitutability, when the lost revenue is compensated for by an

increase in the consumption tax on the commodity bearing the lowest

indirect tax burden. (73)

In their view, however, "[t]he extant literature ... completely ignores the implications of an informal economy for the efficiency of consumption tax (VAT) as an instrument of revenue-raising, which can be especially important in the developing countries." (74)

According to Emran and Stiglitz, for the existing results on "revenue-neutral selective reform of tariffs and consumption taxes" to be valid and applicable, it is necessary to make the assumption "that it is feasible to impose and collect consumption tax (VAT) on the commodity bearing the lowest indirect tax on consumption." (75) This assumption is problematic, for "[w]hile [it] is automatically satisfied when an economy consists of only the formal sector, it is not a plausible assumption in the presence of a large informal segment in the economy that, by definition, escapes VAT coverage." (76)

Instead, they assert that "[i]n an economy with both formal and informal sectors, the best one can do is to select the commodity enjoying the lowest indirect tax burden among the subset of formal commodities as the candidate for VAT increase." (77) In light of this restriction, they then prove "that there are plausible (sufficient) conditions under which such a selective reform of VAT and import tariff reduces welfare." (78) They also provide plausible, sufficient conditions for "worsening of welfare from a reduction in import tariff with a revenue-neutral VAT base broadening." (79)

Emran and Stiglitz also criticize the fact that "the extant literature exclusively deals with the coordinated reform of import tariffs and consumption taxes, and ignores the case of a coordinated reform of export taxes and consumption taxes, although such reforms are frequently prescribed by the policy advisors." (80) They argue:

Our results on export tax reform in the absence of an informal sector

show that the conditions required for a welfare improvement from the

reduction in export tax on one commodity combined with a revenue

neutral increase in VAT on another are much more stringent than the

case of an import tariff reform. Unlike the case of an import tariff

reform, the selective revenue-neutral reform of VAT and export tax can

reduce welfare in an economy without an informal sector, even when all

commodities are ... substitutable. The results of this paper thus

complement and strengthen the conclusions reached by [our earlier

article]. (81)

In addition, Emran and Stiglitz argue that "trade taxes enjoy a clear advantage over VAT" due to administrative costs, which is "the usual explanation for the pervasive use of trade taxes in early stages of development." (82) This primarily stems from the fact that "[t]he informational and compliance costs of VAT are likely to be high, especially in developing countries, because of high rates of illiteracy and scant written record-keeping." (83)

Lastly, they argue that trade taxes are not more vulnerable to smuggling than VAT. They assert that:

[A]n increase in import taxes increases the returns to both domestic

production and smuggling, so that the extent of smuggling is

constrained by the higher domestic supply of a commodity. A higher

VAT, on the other hand, increases the consumer price but leaves the

returns to the domestic producers unchanged. This implies a higher

return to smuggling relative to domestic production, assuming that the

commodity in question is an importable. (84)

Other authors have also questioned the wisdom of eliminating trade taxes. Baunsgaard and Keen analyzed panel data for 111 countries over 25 years--from 1975 to 2000. (85) They show that developing countries find it very difficult to replace the revenue lost by trade liberalization with revenue from domestic sources. This reality is especially troubling since "revenue recovery has been extremely weak in low-income countries (which are those most dependent on trade tax revenues)." (86) These countries have recovered, at best, no more than about 30 cents of each lost dollar. (87) Moreover the presence of a VAT has not in itself made it easier to cope with the revenue effects of trade liberalization. (88)

Baunsgaard and Keen seem to present another challenge to the conventional wisdom that eliminating trade taxes is necessarily good. They also argue, however, that "it is perfectly possible for trade reform to be socially beneficial even when accompanied by a reduction in total revenue...." (89) There is no support for this argument in their paper, though, as the paper focuses on income, not welfare measurement, and there is no explanation as to why Emran and Stiglitz, who reach the opposite conclusion, (90) may be wrong.

Two other authors, Gordon and Li, criticize the conventional wisdom indirectly by suggesting a rationale for the distinct structure of tax systems in developing countries that is fundamentally different from the rationale offered by the IMF and World Bank staff. (91)

Gordon and Li describe the characteristics of the tax systems in developing countries in the following way:

[R]evenue/GDP is surprisingly small compared with that in developed

economies. Taxes on labor income play a minor role. Taxes on

consumption are important, but effective tax rates vary dramatically

by firm, with many firms avoiding taxes entirely by operating through

cash in the informal economy and others facing very high

liabilities. (92)

In this model, corporate tax is also significant, as are tariffs and seignorage (printing money). (93) This description runs counter to mainstream theoretical literature analyzing tax policy in developed countries. (94)

The description suggests that all of these aspects of policy (high corporate tax rates, tariffs, and seignorage) may be explained as a reaction to major tax enforcement difficulties. (95) The key assumption in this theory is that firms in developing countries can evade taxes completely "by shifting entirely to cash transactions and not using the financial sector...." (96)

Gordon and Li's model provides that:

When firms make use of the financial sector ... the government can

gain access to their bank records and use this information in

enforcing the tax law. Firms then have to choose whether the economic

benefits from use of the financial sector are greater or less than the

resulting tax liabilities. Poorer countries differ from richer

countries under [this] hypothesis simply because the value firms

receive from using the financial sector is much more modest. When the

value from using the financial sector is low, the government needs to

worry about possible disintermediation and the resulting loss of its

tax base when choosing its tax structure. This threat of

disintermediation not only can keep tax rates low, but can also have

important effects on the design of the tax structure, and on

government policies more generally. (97)

Taxes can most easily be collected from the firms most dependent on the financial sector, presumably capital-intensive firms. (98) Gordon and Li state that "[r]elying more heavily on corporate income taxes is one means of focusing tax collection on the firms that are most dependent on the financial sector." (99)

Given the resulting differential tax rates by sector, other policies would sensibly be used to offset these tax distortions. Tariff protection for capital-intensive firms makes sense since the government wishes to protect its tax base. (100) Another policy is inflation, which is relatively prevalent in developing countries and enhanced by seignorage. Firms that use the financial sector are largely protected from inflation, since in equilibrium their bank deposits should earn a higher nominal interest rate, reflecting expected inflation, whereas those firms that rely on cash transactions so as to evade tax are thereby vulnerable to inflation. (101)

Gordon and Li argue that "[p]olicies may sensibly encourage or hinder investments by multinationals, depending on the government's ability to tax multinationals vs. competing domestic firms. There may even be an efficiency gain from introducing red tape hindering activity in the untaxed sector." (102)

In the tax scheme proposed by Gordon and Li, the financial sector has an essential role in ensuring a functioning tax structure. (103) The government must be able to act to ensure access to bank records on each firm in order to use this information in enforcing the tax law. (104) Gordon and Li contend that "[s]tate ownership of the banks is one extreme policy that can in principle assure that banks make information available to the government. Another approach is use of bank regulations, whereby any bank that refuses to cooperate with the tax authorities loses its license to function as a bank." (105) Entry of foreign banks will be particularly discouraged in this model, given the ease with which foreign banks can facilitate tax evasion by domestic firms. (106)

The policy implications of accepting the hypothesis suggested by Gordon and Li are quite different from those offered by the conventional wisdom analysis. (107) For example, reducing tariffs or inflation may not be optimal, as doing so would reduce tax revenue. The key policy focus in this model would be reform of the domestic financial sector. (108) This system provides that:

Any policies that raise the value of the services provided by

financial intermediaries will increase the usage of the financial

sector, raising efficiency and allowing the government to collect more

revenue. Conversely, anything that undercuts the perceived value of

the services provided by the financial sector, e.g. a bank failure,

can undermine the fraction of GDP collected in tax revenue, in

addition to any direct effects on GDP through loss of financial

intermediation. (109)

The conventional wisdom's reliance on corporate taxes was challenged in a recent article by Margalioth that builds on the conventional wisdom's emphasis on the optimality of consumption taxes in developing countries but criticizes its reliance on corporate income tax. (110) Margalioth argues that developing countries should adopt policies that attract foreign direct investment (FDI). (111) Case studies suggest that substantial technological diffusion takes place in domestically-owned firms, and, according to new growth theories, such spillover effects are the engine of economic growth. (112)

He also argues that as former FDI barriers, like tariffs, currency exchange controls, and the costs of spreading production processes over multiple countries, are reduced or gone, taxes became a more decisive factor, hence offering tax incentives or having a uniform low corporate tax rate may be an effective policy to attract FDI. (113)

Margalioth suggests that:

[T]ax incentives, like any other market intervention, are justified if

they correct market inefficiencies or generate positive

externalities.... FDI generates positive externalities in the form of

productivity spillovers. As with any positive externality, the amount

of FDI absent government intervention is socially sub-optimal because

foreign investors cannot capture the full gains of their

investments. (114)

He describes tax incentives as:

[T]ax provisions that deviate from baseline provisions. If the

baseline is the standard international or regional tax rate, or even

the individual tax rate, then a low corporate tax rate qualifies as a

'tax incentive.' If the motivation behind the low tax rate is

attracting investments, then it is even more appropriate to classify

it as a tax incentive. If the baseline is the corporate tax itself,

then a low corporate tax rate [that does not distinguish between

foreign and domestic investors] is not an incentive since by

definition tax incentives are targeted at specific types of investors

or investments. (115)

He goes on to argue that targeted tax incentives are much more powerful and cost effective policy tools than low across-the-board corporate income tax rates, at least where policymakers know what types of investments involve the greatest positive spillovers at the lowest administrative cost. (116) Therefore, policymakers might prefer to identify good potential investments on a case-by-case basis. (117)

Margalioth points out that there are, however, many disadvantages to such a regime. First, it is harder to make potential investors aware of this type of incentive. Case-by-case review works best with large investments where investors are more likely to shop around. Second, the advantage over the general tax incentives mentioned above is limited. There is no guarantee policymakers can correctly assess potential spillovers, even when they are examining a specific investment. The most acute disadvantage of discretionary tax incentives, especially in developing countries, is that they are susceptible to corruption. In many countries, discretionary application of tax incentives is one of the most important contributors to corruption. (118)

Margalioth maintains that a general corporate income tax rate reduction is a viable possibility to attract growth-promoting FDI. (119) By far the most troubling aspect of using tax incentives or a low corporate tax rate to attract FDI is surrendering the ability to impose corporate income tax on domestic taxpayers. This frequently is a consequence of tax incentives because limiting tax incentives to foreign investors is administratively infeasible. (120)

However, he notes that losing the ability to impose regular corporate income tax rates on domestic taxpayers might not be too harmful, for the following reason. Shifting from an income to a consumption tax usually is considered efficient but regressive. (121) However, as is now the conventional wisdom (which was not that clear at the time the paper was written), consumption taxes are not necessarily regressive. The regressivity is offsetable through a more progressive use of the tax revenue generated from other sources, mostly through the expenditure side of the national budget. (122) Another consideration is the potential difference in tax incidences between developed and developing countries. (123) Margalioth notes that according to Shah and Whalley, the incidence of corporate income tax in developing countries makes it somewhat regressive. (124) Replacing a regressive corporate income tax with greater reliance on a regressive consumption tax probably adds little, if any, to the overall regressivity of the tax system--even without adjusting the expenditure side of the budget. Hence, a shift to a consumption tax is justified if the tax incentives attract FDI that results in growth-promoting spillovers. (125)

Margalioth further suggests that developed countries should replace some of their foreign direct aid with an equity-based tax expenditure policy. (126) They should allow residents who invest in developing countries to fully benefit from the tax incentives offered by exempting or otherwise sparing this foreign income. (127) His proposal transfers revenue from the treasuries of developed countries to developing countries, but it does so indirectly and with targeted money. (128) It is equivalent to giving the governments of developing countries money that can be used only to attract FDI. (129) The underlying assumption is that governments of developing countries lack the capacity to run large industrial and commercial enterprises, so promoting growth through multinationals' activity is more efficient. (130)

According to Margalioth, these revenue transfers would increase if limits were placed on the ability of rich countries, such as Ireland, to engage in tax competition with poor countries. (131) He argues that a special form of tax harmonization with a sharp division based on per-capita GDP is justified as part of an international vertical equity regime for transferring wealth to developing countries. (132)

Margalioth concludes that since developing countries may also engage in harmful tax competition, we can apply similar anti-tax competition rules to them. (133) This leads to the establishment of two different harmonized tax levels, one for developed countries and the other for developing countries. (134)

V. CONCLUDING REMARKS AND SUGGESTIONS FOR FUTURE RESEARCH

The Washington Consensus, articulated in 1990, was meant to synthesize the reforms that most economists in the World Bank, the IMF, the U.S. Treasury, and some of Washington's think tanks believed were necessary for sustained economic growth. The experience of the last two decades has proved the Consensus wrong. (135)

There is no reason to think that things are different in the tax arena. Common sense advice, based on the experience of advisers educated in developed countries, may not necessarily make sense in any specific developing country. A good tax system is one that fits both the social institutions as well as other specific determinants of distribution and economic growth in each country. Searching for one optimal tax system for countries grouped together by a definition based on GDP per capita is problematic. The vast differences, for example, between Latin America, Sub-Saharan Africa, China, and India, should make it impossible to design a generalized tax system, or even to offer useful guidelines, unless we first study each country separately. Therefore, our initial suggestion for future research is the allocation of developing countries into new categories that will better enable some generalization of tax policy advice.

Based on current literature, we tried to portray the conventional wisdom in the field. There seems to be a general agreement that consumption taxation is superior to income taxation in developing countries in terms of both efficiency and redistribution but that corporate income taxes should be withheld.

Moreover, according to conventional wisdom, developing countries should tax as efficiently as possible and rely on the expenditure policy to take care of inequality and poverty. But if that is the advice, then why use consumption tax? Why not impose a head tax, which is by definition even more efficient? The answer is that a head tax plus an expenditure policy is not necessarily easier to implement than income or consumption taxes. But, exactly how and why is yet to be explored.

If the criterion for welfare is income, then there is no difference between an income tax and an expenditure program, which is in fact a negative income tax. If, on the other hand, we use other proxies for ability, targeting individuals according to non-income characteristics, there is a difference between expenditure policy and an income tax, and the former could be easier to implement.

If we believe in universality (providing public and merit goods of decent quality and possibly providing a cash or in-kind transfer to everyone), expenditure policy could be much simpler and less distortive than an income tax (as it imposes zero marginal tax rate).

The conventional wisdom is further challenged by Emran and Stiglitz who argue that trade taxes may be superior to VAT. They suggest that there is a need for empirical work that explicitly incorporates the role of the informal economy to examine the question of whether the proposed move away from trade taxes to domestic consumption taxes is welfare enhancing or not. (136)

In addition, it is possible that the underlying explanation for a whole set of policies exercised by developing countries is their desire to cultivate a certain type of tax abiding firm. This hypothesis, suggested by Gordon and Li, calls for a close examination of questions such as who is taxed in developing countries and what information flows from banks to tax authorities in developing countries. (137) Lastly, it is still an open question whether offering tax incentives to attract foreign investments is warranted or not.

Reuven Avi-Yonah* and Yoram Margalioth**

* University of Michigan Law School.

** Tel Aviv University Faculty of Law. This paper was presented at the first OECD International Network for Tax Research (INTR) Conference, which took place in November 2006 at the University of Michigan Law School. We would like to thank the participants for their valuable comments. Yoram Margalioth would like to thank the Cegla Center for Interdisciplinary Research of Law for financial support.

(1) Robin Burgess & Nicholas Stern, Taxation and Development, 31 J. ECON. LIT. 762 (1993).

(2) Id. at 821.

(3) Id.

(4) See id. at 819-20.

(5) See id. at 799.

(6) Id.

(7) Id. at 773 tbl.5, 777.

(8) Id. at 777.

(9) See M. Shahe Emran & Joseph E. Stiglitz, On Selective Indirect Tax Reform in Developing Countries, 89 J. PUB. ECON. 599, 599-600 (2005). Emran and Stiglitz state:

A reduction in the trade tax with a compensating or revenue-enhancing

increase in value-added tax (henceforth VAT) has been the center-piece

of such a reform, and it has been implemented in a large number of

developing countries under the structural adjustment and stabilization

policy conditionalities of the IMF and the World Bank. Id.

(10) See, e.g., Alex Easson, Tax Incentives for Foreign Direct Investment Part 1: Recent Trends and Countertrends, 55 BULL. FOR INT'L FISCAL DOCUMENTATION 266, 266 (2001).

(11) See Emran & Stiglitz, supra note 9, at 599; see also Roger Gordon & Wei Li, Tax Structure in Developing Countries: Many Puzzles and a Possible Explanation (Nat'l Bureau of Econ. Research, Working Paper No. 11267, 2005).

(12) Yoram Margalioth, Tax Competition, Foreign Direct Investments, and Growth: Using the Tax System to Promote Developing Countries, 23 VA. TAX REV. 161, 188 (2003).

(13) Government expenditure of the gross domestic product (GDP) is 31.5% in developed countries, as opposed to 25.4% in developing countries. Burgess & Stern, supra note 1, at 765.

(14) Non-tax revenue in developing countries comprised about 21% of GDP compared to 10% in developed countries. These are aggregate figures and substantial variation exists across countries. Id. at 782.

(15) Id. at 770.

(16) Id. at 772 tbl.4, 773 tbl.5, 775.

(17) Id.

(18) Id.

(19) Id.

(20) Id. "[T]here appears to be a pattern of evolution of tax structure [in developing countries]." Norman Gemmell & Oliver Morrissey, Tax Structure and the Incidence on the Poor in Developing Countries 5 (Centre for Res. in Econ. Dev. and Int'l Trade, Univ. of Nottingham, No. 03/18, 2003), available at http://ssrn.com/abstract=503101. "At low levels of income, trade taxes are relatively important and income taxes relatively less important." Id. (citation omitted). A shift occurs, from trade to domestic sales taxes, as income increases. "As incomes rise again, trade taxes become unimportant and various income taxes become most important. These averages, however, conceal wide disparities between countries." Id.

(21) See Emran & Stiglitz, supra note 9.

(22) See, e.g., The WTO in Brief, http://www.wto.org/english/thewto_e/whatis_e/inbrief_e/inbr00_e.htm.

(23) See Richard M. Bird, Value-Added Taxes in Developing and Transitional Countries: Lessons and Questions (Joseph L. Rotman Sch. of Mgmt., Univ. of Toronto, ITP Paper 0505, 2005), available at http://www.rotman.utoronto.ca/iib/ITP0505.pdf.

(24) See Emran & Stiglitz, supra note 9, at 600.

(25) See id. at 600.

(26) The theory of comparative advantage was first introduced by David Ricardo. See DAVID RICARDO, ON THE PRINCIPLES OF POLITICAL ECONOMY AND TAXATION (3d ed. 1821), available at http://www.econlib.org/library/Ricardo/ricP.html.

(27) Gemmell & Morrissey, supra note 20.

(28) Id.

(29) Id. at 22.

(30) Id.

(31) Id. at abstract.

(32) Id. at 29.

(33) Id. at abstract.

(34) Id. at 29.

(35) Id.

(36) Id. at 22.

(37) Id. at 27-28.

(38) Id. at 29.

(39) Id. at 28.

(40) Id. at 29.

(41) Richard M. Bird, Taxation in Latin America: Reflections on Sustainability and the Balance between Equity and Efficiency 7 (Joseph L. Rotman Sch. of Mgmt., Univ. of Toronto, ITP Paper 0306, 2003), available at http://ideas.repec.org/p/ttp/itpwps/0306.html.

(42) Id. at 44.

(43) See, e.g., id.

(44) Id. at 41; see also Richard M. Bird, Jorge Martinez-Vazquez & Benno Torgler, Societal Institutions and Tax Effort in Developing Countries 13-14 (Joseph L. Rotman Sch. of Mgmt., Univ. of Toronto, ITP Paper 04011, 2004), available at http://ssrn.com/abstract=662081 (describing Latin America as a region in which the combination of the dominant policy ideas and the dominant economic and social interests combine with the key political and economic institutions (democracy, decentralization, and budgetary; and free trade, protectionism, macroeconomic policy, and market structure) to produce a generally low tax level and an uneven tax structure).

(45) See Bird, supra note 41, at 40.

(46) Id.

(47) Id. at 43.

(48) Id. at 43 n.58.

(49) Id. at 43.

(50) Id.

(51) Id.

(52) Id. at 47.

(53) Id.

(54) Id.

(55) Richard M. Bird & Eric M. Zolt, Rethinking Redistribution: Tax Policy in an Era of Rising Inequality: Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries, 52 UCLA L. REV. 1627 (2005).

(56) Id. at 1630.

(57) Id. at 1688-90.

(58) Id. at 1683.

(59) Id. at 1684.

(60) Id.

(61) Id. at 1685.

(62) Id.

(63) Id. at 1686.

(64) Id. at 1690.

(65) Id. at 1695.

(66) Emran & Stiglitz, supra note 9, at 618.

(67) Id. at 602.

(68) Id.

(69) Id. at abstract.

(70) Id. at 600 (describing M.S. Emran & J.E. Stiglitz, VAT Versus Trade Taxes: The (In)efficiency of Indirect Tax Reform in Developing Countries (mimeo, Stanford Univ. & Brookings Inst., Washington, D.C., 2000)).

(71) Id.

(72) See id. at 600 & n.4.

(73) Id. at 602 (citing M. Michael et al., Integrated Reforms of Tariffs and Consumption Taxes, 52 J. PUB. ECON. 417 (1993)).

(74) Id.

(75) Id.

(76) Id.

(77) Id.

(78) Id.

(79) Id.

(80) Id.

(81) Id. at 602 (footnote omitted).

(82) Id. at 620 (citation omitted).

(83) Id.

(84) Id. at 621.

(85) Thomas Baunsgaard & Michael Keen, Tax Revenue and (or?) Trade Liberalization (Int'l Monetary Fund, Working Paper No. 50/112, 2005).

(86) Id.

(87) Id. at 14.

(88) Id.

(89) Id. at 3.

(90) Emran & Stiglitz, supra note 9, at 600.

(91) Gordon & Li, supra note 11.

(92) Id. at abstract.

(93) Id. at 2.

(94) Id.

(95) Id. at 3.

(96) Id.

(97) Id. at 3-4.

(98) Id.

(99) Id.

(100) Id.

(101) Id. Gordon and Li state:

The optimal inflation tax is limited, though, by the possibility that

dollars or some other foreign currency replace the local currency, in

order to avoid the inflation tax. Since this currency substitution

provides no further shift in resources towards the taxed sector but

leads to a discrete fall in seignorage revenue, the optimal inflation

rate is capped due to this treat of currency substitution.

Id. at 21.

(102) Id. at 4.

(103) Id. at 26.

(104) Id.

(105) Id. at 26.

(106) Id. at 27.

(107) Compare id. (suggesting the imposition of tariffs, printing money, state-ownership of banks, discouraging entrance of foreign banks thereby limiting the competition in the banking sector, and introducing red tape in certain sectors of the economy), with Burgess & Stern, supra note 1 (suggesting a switch from relying on tariffs to general consumption taxes, preferably VAT, avoiding money printing, eliminating red tape, and increasing competition).

(108) Gordon & Li, supra note 11.

(109) Id. at 30.

(110) Margalioth, supra note 12, at 169; see also Burgess & Stern, supra note 1, at 821 (predicting that developing countries would not change their reliance on corporate taxation).

(111) Margalioth, supra note 12, at 168-69.

(112) Id. at 167, 177.

(113) Id. at 183.

(114) Id. at 184-85.

(115) Id. at 187.

(116) Id. Margalioth states:

We generally assume that plant and equipment investments in medium and

high tech industries promote growth. Various tax code provisions, such

as faster-than-economic cost recovery, investment allowances and tax

credits attract these desirable investments. Tax incentives for

increasing investment in human capital range from allowing taxpayers

to deduct or expense education and training costs to providing

employers with tax credits for the same. Id. (footnote omitted).

(117) Id. at 188.

(118) Id.

(119) Id.

(120) Id. at 190. "Of course, extending benefits to domestic corporate taxpayers may be desirable so that they are not at a competitive disadvantage relative to the FDI in local markets. Tax incentives also help place domestic corporations in a position to absorb spillovers with respect to export activity." Id.

(121) Id.

(122) Id. at 191.

(123) Id.

(124) Id. (citing Anwar Shah & John Whalley, The Redistributive Impact of Taxation in Developing Countries, in TAX POLICY IN DEVELOPING COUNTRIES 166, 172 (Javad Khalilzadeh-Shirazi & Anwar Shah eds., 1991)).

(125) Id.

(126) Id. at 201.

(127) Id.

(128) Id. at 201-02.

(129) Id. at 202.

(130) Id.

(131) Id.

(132) Id.

(133) Id.

(134) Id.

(135) See, e.g., Dani Rodrik, Goodbye Washington Consensus, Hello Washington Confusion?: A Review of the World Bank's Economic Growth in the 1990's: Learning from a Decade of Reform, 44 J. ECON. LIT. 973 (2006).

(136) Emran & Stiglitz, supra note 9.

(137) Gordon & Li, supra note 11.