Tax Competition, Foreign Direct Investments and Growth: Using the Tax System to Promote Developing Countries


The world population is approximately six billion people. 1.2 billion, or one-fifth, are rich while 4.8 billion, or four-fifths, are poor.[1]  The world is more unequal today than at any time in world history. (2)  Only 240 years ago, at the beginning of the industrial revolution, most everybody was poor. Author Lant Pritchett explains that very poor nations today are just barely above the subsistence level as measured in income per person.[3]Since subsistence is defined as not starving to death, the poorest nations today have about the same income in real terms today as they did two centuries ago.[4] It could not be less, because that would mean they were below subsistence level, which, by definition, cannot be true.[5] The period of economic growth is a fairly recent one, but it has been a period of extreme divergence in economic performance.[6] A relatively small part of the world achieved what economists call modern economic growth.[7] Most of these countries sustained increases in the growth rate year after year. Accumulated over two centuries, this leads to a twenty-fold increase or more in the standard of living measured in material terms.[8] The effect of sustained growth rates on the standards of living is comparable to that of compounding interest on the wealth of savers. Nobel Prize winner Robert Lucas introduced a rule of thumb to development economics that is often used by investors. Lucas found that a country growing at g percent per year will double its per capita income every 70/g years.[9] With the 1.8% the per capita growth rate that has persisted in the U.S. over most of the last 200 years, we see that the average American is two or three times as rich as his or her grandparents. A few other countries, including Japan and the so- called “Asian Tigers” (Singapore, Hong-Kong, South Korea and Taiwan) experienced high growth rates beginning in the 1950s and continuing well into the 1990s.[10] Most other parts of the world did not approach the kind of achievement seen in the United States, Western Europe and South East Asia, though some improved their standard of living. Some poor countries grew very slowly, while others, such as most African nations, did not grow at all.[11] This created a huge gap between the countries that enjoyed sustained growth and the rest of the world. Developing countries are not only relatively poorer, they are worse off by absolute measures as well. The median per capita growth in developing countries between 1980 and 1999 was zero.[12]

Sustained growth is the key to improving the situation in poor countries.[13] Of course, there are important factors besides growth in measuring quality of life. Gross Domestic Product (GDP) measures only market transactions and fails to account for environmental costs of production (e.g., pollution), equality of wealth (income) distribution, or for non-economic factors such as freedom, education and health.[14] Despite the shortcomings mentioned above, there is strong positive correlation between GDP per capita and most other factors – including health, education and income distribution. The residents of rich countries benefit from better health,[15] more education (probably of higher quality, but this was not measured),[16] and greater income equality[17] than the residents of poor countries.

Finally, as William Easterly notes,[18]a rise in incomes in developing countries generates additional benefits, such as lower infant and child mortality rates,[19] more food to eat,[20] better basic medical treatment,[21] and freedom from oppression.[22]

Part II below deals with the neoclassical growth theories and the lack of success of the past 50 years in using them to promote growth in developing countries. Growth theory has evolved substantially in the relatively recent past. Some great economists, including Solow, Kaldor, Mirlees and Arrow, developed it extensively during the 1960s. Growth theory went out of fashion during the 1970s, but came back in full force about fifteen years ago with the development of the new growth theories. The traditional growth models stressed the importance of savings as a means of accumulating capital because economists believed at the time it was the key to growth. Developing countries were considered too poor to afford the required levels of savings, so developed countries lent them money to fill the temporary financing gap. This was supposed to lead developing countries to self- sustained growth, making it possible for them to later repay the loans. Transfers to developing countries over the past fifty years total more than a trillion dollars, but most of that money was lost.[23] Currently, the World Bank and the International Monetary Fund (IMF) transfer $54 billion annually, and more is transferred through private institutions.[24]

Most of the money transfers were meant to achieve growth, but they had very little success. The World Bank and the IMF pursued an ambitious plan to make developing countries improve their institutions and policies by conditioning loans on implementing reforms.[25] These “adjustment loans” were not successful because most countries did not comply. They used the money to finance current consumption instead of promoting growth, and were later unable to service the debts. Where did the money go? Some of it addressed immediate needs, like famine relief. Most of it financed corruption or unnecessary construction projects, such as new capital cities and dams. [26] By September 2002, $37.2 billion of debt had been forgiven, and the total amount of debt forgiveness is planned at more than $50 billion.[27]

How can developed countries help promote sustained growth in developing countries? The answer lies in our modern understanding of what causes growth – sound institutions and technological progress. In this paper, I build on these new theories by proposing to promote growth in developing countries through the use of their tax systems and the international tax regime.

In the second half of Part II, I address the new growth theories. Like neoclassical theories, new growth theories postulate that technology is the engine of growth. Unlike neoclassical theories, which assume that policy cannot affect technological progress, that it is exogenous to the model, the new theories recognize that policies like promoting research and development can increase the rate of technological progress. In other words, policy can increase the production of ideas. New ideas, in turn, allow a given bundle of inputs to produce more or better output. There are plenty of examples of growth-promoting ideas: Henry Ford’s assembly lines and mass production techniques, software, drive-through windows at fast-food restaurants and the double-entry method for bookkeeping are just a few.

Ideas are non-rivalrous, meaning the use of an idea by one person does not preclude its use by another. Ideas also tend to be only partially excludable. The degree to which a good is excludable is the degree to which the owner of the good can charge a fee for its use. The marginal cost of most ideas is so low that it is properly described as zero. For example, once new software is invented, additional copies can be created at very little cost. These increasing returns to scale make ideas a powerful engine of growth.

Of course, the government must intervene to allow inventors to cover their costs and capture some of the benefits their ideas produce. Otherwise, there would be little incentive to invent. The development of intellectual property rights, a cumulative process that occurred over centuries, probably played a critical role in sparking the Industrial Revolution and is largely responsible for modern economic growth.[28]

Usually, inventors do not capture the full value of their inventions and society benefits from positive externalities, also known as spillover effects. In other words, the social rates of return far exceed the private returns to the investor.[29]

A classic example is the solving of the longitude problem. Up until the middle of the eighteenth century, sea travelers could not determine the longitude of their location.[30] A generous prize of £20,000 offered by the British Board of Longitude drove a clockmaker named John Harrison to spend many years building a chronometer that solved this navigation problem. The prize money clearly did not reflect the huge benefit realized by society from the ability to navigate accurately.[31]

Most of the world’s technological progress takes place in about twenty countries. It is then transferred to other parts of the world through international trade, cross-border education, and foreign direct investments (FDI). FDI is an important avenue of transfer, especially for knowledge such as management techniques and blueprints.[32]

Only a small fraction of the world’s FDI is located in developing countries,[33] and multinational corporations (MNCs) are the main carriers of FDI.[34] They bring host economies valuable tangible and intangible assets such as capital, technology, market access, and management skills. FDI in general, and FDI by MNCs in particular, can strengthen or even build economic markets and institutions. Developing countries benefit from the positive externalities created by the existence and activities of MNCs, and these externalities may help create market discipline.

The thesis offered in this article is based on a three-level proposal. The proposal is modular, with the first stage forming the basis and the two other levels serving as additions. Part III describes the first stage, which outlines what developing countries can do on their own as a form of self-help. I argue that they can attract foreign direct investment by offering tax incentives. This is widely considered bad policy, so a main goal of this article is to explain why this is a good policy for developing countries to adopt.

I propose that developing nations attract FDI by offering zero or very low corporate tax rates. Activities involving economic rent, such as mining or retail activities are, however, not ideal targets for tax incentives. Low tax rates are appropriate only for corporations in sectors generating spillover effects. Eliminating most or all of the corporate tax requires that a country rely on alternative revenue sources. The most likely alternative revenue source is consumption taxes. Consequently, the greater marginal attraction of FDI with spillover effects justifies a move from taxing corporate income to greater reliance on consumption taxes.

Eliminating the corporate tax rate and increasing consumption taxes does not necessarily make the income distribution more regressive. It is an open question in developing and developed nations alike what the incidence, or economic burden, of the corporate tax is. But, the general assumption is that in developing countries, fifty percent of the corporate tax is shifted to consumers and fifty percent is borne by capital owners.[35] Hence, at least in developing countries, corporate tax is not as progressive as an individual income tax and replacing it with consumption tax would make the tax system significantly more regressive. At the same time, consumption taxes, like the “value-added tax” found in most developing countries, are usually considered more efficient than income taxation.

Tax incentives or low corporate tax rates fail to attract investments to the extent that developed countries tax residents on income generated in developing countries. The result is a net transfer from the developing country to the treasury of the developed country. As detailed in Part IV, this is not a significant phenomenon. Insofar as it exists, the second and more ambitious level of my proposal applies, which requires the active support of developed countries.

I do not propose an increase in the amounts transferred to developing countries.[36] I suggest changing the mixture of these contributions. Instead of contributing only money, I propose that some aid take the form of foregone tax revenues. This foregone revenue goes not to the governments of developing countries, but instead to those who invest in developing countries. This might seem like a counter-intuitive argument since direct grants are usually considered superior to tax subsidies,[37] but a closer examination reveals the difference in the identity of the recipient. Instead of giving all the foreign aid money to the government, I propose to give some of it to private market players whose profit-seeking activity in the developing country is key to that country’s economic growth.

The current international tax regime[38] determines the proper division of the tax base between countries in cross-border transactions according to principles of entitlement and administrative ease.[39] The wealth of the countries sharing the tax revenue is not taken into account.[40] On the other hand, domestic tax systems usually redistribute income from richer to poorer taxpayers, and federal states often use their tax systems to redistribute tax revenue from richer localities to poorer ones.[41] Using the international tax regime to redistribute revenues from richer to poorer countries is another way to fight poverty and promote greater equality among the nations of the world.[42] Moreover, as I explain in this paper, replacing some of the direct transfers to poor countries with indirect transfers carried through the international tax laws (i.e., replacing money contributions with tax revenue losses incurred by the rich countries) is a more efficient way to promote economic growth in developing countries than the current provision of grants.

Part IV describes the third stage to my proposal. I propose limiting the ability of developed countries to engage in tax competition over FDI with developing countries. I also suggest limiting the ability of developing countries to engage in tax competition with each other. This Part relates to what is known in the legal tax literature as the “tax competition debate.” On one side of the debate are those concerned that tax competition increases the regressivity of tax systems while simultaneously reducing the social safety net countries offer to their residents.[43] On the other side are those who view tax competition as free market price competition that generally leads to the most efficient result.[44]

My proposal is neutral with respect to this debate because my proposal is the same whether tax competition is harmful or not.[45]

Either way, I argue that developing countries should be able to engage in tax competition while developed countries should not. I justify this on an inter-nation equity basis, since tax competition redistributes wealth from rich to poor countries. To the extent that tax competition between developing countries is harmful, an international body can help developing countries commit to a bottom threshold below which they will not compete. Defecting developing countries could be sanctioned by the repeal of exemption or tax sparing treatment given to investors in that country by the foreign investors’ country. Allowing developing countries to coordinate tax rates lets them extract cartel profits by improving their bargaining position against the MNCs. Of course, they still retain the ability to offer tax advantages to FDI that developed countries are prohibited from offering.

Both sides to the tax competition debate agree that tax competition on portfolio investment is a harmful problem of tax evasion. My proposal deals with FDI, not portfolio investment, but since developing countries are especially harmed by this type of tax evasion I devote a short part of this paper, Part V, to analyzing proposals to curb this type of tax competition.

The ideology behind my proposal has two guiding principles. The first is that close economic contact between the industrial core and the developing periphery is the best way to accelerate the transfer of technology. Technology transfer is the indispensable condition for making poor economies rich. Thus all barriers to international trade must be eliminated as quickly as possible,[46] and developing countries must attract FDI. The second guiding principle is that governments generally, and developing ones in particular, lack the capacity to run large industrial and commercial enterprises. Thus, governments should shrink and privatize in all but their core missions. Core missions include things like income distribution, social insurance, public-good infrastructure, and administration of justice..

Attracting FDI to developing countries clearly serves an equity interest. It is also efficient from the world’s perspective, because even developed countries benefit. Spreading knowledge and sharing it with the world population leads to greater technological progress since ideas serve as a platform for the generation of new ideas. All nations stand to benefit from the policy proposed here.


A. The Neoclassical Growth Models

The first post-Keynesian growth theory is known as the Harrod- Domar model.[47] According to this model, all the government has to do to secure growth is invest in machines. More machines meant greater output and thus greater GDP. The model excludes labor and assumes an unlimited supply of workers because high unemployment rates are taken as a given.

A later model, developed by Noble laureate Robert Solow in 1956-57, points out that the Harrod-Domar model cannot explain sustained growth. Solow showed that as capital per worker increases, the marginal productivity of capital declines until the capital-labor ratio approaches a steady-state level. At that point, savings (assumed in the model to be a constant fraction of income) are just sufficient to replace worn out machines and equip new workers (assuming population growth), so productivity growth is zero.[48]

To explain the long-run sustained growth of the United States economy, Solow added an exogenous term called “technological progress.” Under Solow’s formulation, technology is a pure public good that is freely accessible to all. Solow’s model assumes perfect competition, so technology (knowledge) was not only non-rivalous, as knowledge usually is,[49] but also non-excludable. Extracting monopoly rents through patents, for example, did not exist in Solow’s model. Technological progress increases the supply of “effective” labor[50] and keeps the capital-labor ratio going, preventing the decline of marginal returns to capital.

Solow never applied his model beyond the United States. Subsequent economists have, revealing a major problem with the assumption that technology/knowledge is freely and equally accessible to everyone.[51] Countries are not converging to a common level of per- capita income the way this model predicts, since it assumes that all countries share similar savings behavior and have the same access to technology.

Robert Lucas illustrated the difficulty with the neoclassical model’s assumptions.[52 ] He showed that if production per worker differs between two countries, it must be because they have different levels of capital per worker. The model’s assumptions rule out all other possible differences. Thus, argued Lucas, the law of diminishing returns implies that the marginal product of capital is higher in the less productive or poorer economy. If so, capital should flow to the poorer country until the returns on capital in both countries are equalized. Comparing production per worker in the United States with that in India (taking into account the average of U.S. and Indian capital shares), Lucas found that the model stipulated marginal returns on capital in India fifty-eight times the marginal returns on capital in the United States.[53] Investors should have been flocking to India if marginal returns were indeed that high. Investors were not, so Lucas concluded that Solow’s assumptions about technology and trade conditions must be drastically wrong.[54]

B. The New Growth Theories

The main new growth theory is one of intentional endogenous technological progress. It assumes that a separate technology sector exists in the economy, supplying other sectors with new technologies. Producers pay the technology sector for the right to use the technology. These are monopoly profits because information has no opportunity cost. A basic assumption is that technology producers have some control over access to the information. Producers take advantage of this imperfect competition, charging a price above marginal cost for what they produce. Otherwise, technology producers would not generate enough income to cover their costs, including the initial investment in new technology.

In addition to the private proprietary component, innovation has a positive externality that increases the productivity of all subsequent innovation projects.[55] This externality offsets the tendency toward decreasing productivity of new investments in innovative activity, allowing innovation production – and growth – to go on. In these models, the rate of growth depends on (1) the amount of resources devoted to innovation (e.g., research and development); (2) the degree to which new technology can be privately appropriated; and (3) the time horizon of investors. High-growth also implies high- growth in physical capital. This is a change from previous models, where it is a cause of technological progress. Here, it is a result. Even though innovation activity is risky and only a small percentage of R&D projects succeed, actual success is not an exogenous event. The more time and money firms invest to achieve technological progress, the more technological progress occurs.

The scope of technological progress in these models is quite comprehensive, raising many fascinating legal and economic issues. Paul Romer notes that this model allows policy-makers to do something more insightful than the standard neo-classical prescription: more saving and more schooling.[56] The examples he provides include tax subsidies for private research, antitrust exemptions for research joint ventures, the effect of government procurement, the feedback between trade policy and innovation, the scope of protection for intellectual property rights, the links between private firms and universities, the mechanisms for selecting the research areas that receive public support, the cost and benefits of an explicit government led technology policy, and the activities of multinational enterprises. This last item is where my article fits in. I suggest using these profit-maximizers to stimulate technological processes in, and transfer technological knowledge to, developing countries. My definition of “technology process” is as broad as possible. Technology process includes imitation, adaptation of Western technology to unique national circumstances,[57] and innovation. The types of knowledge involved include modes of organization, blueprints, patents, manufacturing methods, marketing, product design, and any other knowledge that promotes growth.

One of the major channels for technology diffusion to the developing world is FDI by multinational corporations. They are among the most technologically advanced firms in the world, accounting for a substantial part of the world’s research and development investment. Most studies confirm the assumption that foreign-owned firms are more productive than their domestically owned competitors.[58] Since domestic firms have better knowledge and access to domestic markets, a foreign firm entering the market must overcome its disadvantage through lower costs and greater productive efficiency.[59] In the case of developing countries in particular, the higher efficiency of FDI usually results from a combination of advanced management skills and more modern technology. FDI is therefore an important way to transfer advanced technology to developing countries.

Case studies suggest that substantial technological diffusion takes place in domestically-owned firms.[60] Domestic firms in sectors with a relatively high presence of MNCs also tend to be more productive.[61]

There is a difficult question of causality in these cases because it is possible that FDI is attracted only to sectors or firms with greater productivity potential. Yet it is also plausible to assume that FDI inflows raise the levels of technology in the host economy and contribute to higher productivity. One plausible explanation for this is that FDI inflows increase the variety of intermediate products and types of capital equipment in the host economy.[62] As such, FDI inflows increase productivity in the host economy. Another important way in which FDI affects growth is through learning. FDI inflows disseminate knowledge about production methods, product design, and new organizational and managerial techniques. Seen this way, imitation is a crucial element. Consequently, FDI also raises the productivity of domestic research and development activities.

A different kind of productivity spillover takes place when domestic firms learn from the export activities of MNCs or their host subsidiaries through information externalities. Exporting involves fixed costs like establishing distribution networks, marketing, and regulatory compliance. MNCs, by definition, have considerable knowledge and experience with international transactions. Domestic firms can piggyback on some of the MNC’s investments, such as those in infrastructure, and may also benefit from marketing the MNC does in the developing country. These domestic firms also stand to benefit from technological information and management techniques that leak through local clients, suppliers and MNC workers.

Spillovers generally take the form of imitation. Products, processes, managerial techniques and organizational methods can all be reverse engineered if complicated or imitated if simple. Native MNC employees are another important source of spillover information when the MNC supplies its local employees with general training – any knowledge useful to employees outside the MNC’s workplace. When employees leave the MNC for domestic firms, or leave to start their own firms, there is a knowledge transfer from the MNC to the domestic market. Over time, the MNC faces fierce competition from domestic firms employing its own methods and unprotected technology. This is an implicit tax paid by an MNC when it comes to a developing country and employs local workers. Some argue that this is the most important channel for spillovers.[63]

An example of the tremendous potential of foreign direct investment for developing countries is the story of Desh Garments Ltd. Desh Garments was a small company established in December 1977 in Bangladesh.[64] Its joint-venture with Daewoo, a major world textile producer based in South-Korea, is directly responsible for Bangladesh’s thriving garment industry. In 1978, Bangladesh had no textile industry. Today, it exports $4.5 billion worth of textile goods annually, approximately 75% of all Bangladeshi exports.[65] This remarkable transformation began when Desh Garments sent 130 workers and management trainees for training at Daewoo’s state-of- the-art factories in South Korea. After six months of training, they returned to Bangladesh to start a shirt factory. The most valuable part of their training was not learning how to operate textile machinery; it was learning a clever way to get around the protectionist Bangladeshi trading system and strict foreign exchange controls. This method required obtaining back-to-back letters of credit and the government’s approval of duty-free imports for exporters like Desh. Daewoo acquired this knowledge by dealing with similar issues in Korea, transferring it to Desh as part of running the business. After little more than a year of successful production, Desh canceled the collaboration agreement with Daewoo. Of the 130 Desh workers trained by Daewoo, 115 of them left during the 1980s to set up their own garment export firms. This explosion of garment companies started by ex-Desh workers gave Bangladesh its textile industry, currently responsible for three-quarters of its exports.[66]

This story demonstrates the value of knowledge leaks. Daewoo made a small profit, though in the unusual circumstances of this specific example, perhaps an unusually small profit. The founder of Desh Garments made a small profit as well. Most of the gains were shared by the Bangladeshi society as a whole.

Another example of the importance of knowledge transfer as a catalyst for growth is illustrated by the difference in growth rates between California’s Silicon Valley and Massachusetts’ Route 128. One explanation for Silicon Valley’s higher rate of growth is that covenants not to compete are less strictly enforced in California than in Massachusetts. This, among other things, contributed to more knowledge leaks in Silicon Valley, and and thus a higher growth rate.[67]

FDI-created spillovers are not limited to the everyday meaning of the term “technology.” I refer to FDI spillovers that create markets where none yet exist, like the textile industry in Bangladesh. I also envision spillovers that improve the quality of existing markets and strengthen national institutions necessary to growth.[68] The dual listing of stock on the Israeli stock exchange is an example supporting the latter point. Israel sought to improve market discipline and trading norms by attracting larger companies into the local market. To this end, the Israeli legislature enacted a pioneering unilateral recognition arrangement that allowed issuers who comply with reporting requirements under certain foreign laws to list their stocks in Israel as well.[69] Israeli institutions hoped to benefit from the spillover effects of listing large corporations on the Israeli stock exchange.

Foreign direct investment may also help developing countries escape the lock-in situation created when the country has a comparative advantage in traditional industries. These countries risk being permanently worse off relative to countries with a comparative advantage in more advanced industries.[70] Escaping a lock-in situation is only possible when the host country’s government facilitates knowledge transfer. In other words, developing countries must do more than just attract FDI in order to benefit from the FDI spillovers. They must also provide a framework that increases the chances of knowledge absorption. This is possible through tax incentives to locals and subsidized or free training. Developing countries may also require foreign investors to employ and train local employees.

Lastly, hosting multinationals’ activity can help developing countries prevent competent residents from emigrating to developed countries. This widespread phenomenon is known as the “brain drain.”[71] Educated citizens are attracted to developed countries where they benefit from higher standards of living and earn more than they could in their country of origin. Multinationals allow developing countries to retain talent by providing opportunity at home.

Foreign direct investment is not the only way to transfer technology to the developing world. As mentioned above, trade and education are important alternative routes. Fulbright scholarships, which finance the studies of individuals from developing countries at leading universities in developed countries, condition that support on the students’ promise return to their home countries. So why not run this type of program on a much larger scale in an effort to promote growth in developing countries? Because FDI achieves better results. FDI offers better potential for spillover effects and avoids the risk of scholarship recipients who emigrate to developed countries a year or two after returning home. Moreover, scholarship programs like the Fulbright require the goodwill and action of developed countries.

While I argue that developed nations can play a key role in helping developing nations attract FDI, on a fundamental level it is something developing countries can do on their own.


A. Standard arguments against the use of tax incentives

Conventional wisdom weighs against using tax incentives to attract investment in general and foreign direct investment in particular. International organizations including the United Nations, World Bank, IMF, OECD and the EU have unanimously opposed the use of tax incentives to attract investments.[72]

The main arguments against the use of tax incentives to attract FDI are the following:

(a) Tax incentives distort behavior and are therefore inefficient.[73]

(b) They are ineffective and harmful, and have only minor effect on FDI decisions.[74]

(c) It is better to deal directly with the problems that make developing countries less attractive for investments instead of trying to compensate for the disadvantages by offering tax incentives.

These arguments are addressed in turn.

1. Tax Incentives Distort Behavior

The argument that tax incentives are inefficient because they distort behavior is flawed because tax incentives are meant to distort behavior. They attract investments that would not otherwise take place. This does not mean they are inefficient. If investments are not taking place because of certain market inefficiencies (e.g., asymmetric information between potential investors in developed countries and entrepreneurs in developing countries, or incomplete world capital markets), or if the FDI attracted by tax incentives entails positive externalities,[75] tax incentives that “distort behavior” are nevertheless efficient because they increase social welfare. From an efficiency point of view, taxes on capital distort investment behavior. Therefore tax incentives, so long as they do not result in negative tax rates, reduce economic distortions.

2. Tax Incentives Are Ineffective and Harmful

The second argument, that tax incentives are ineffective and harmful because their cost in forgone revenue exceeds their benefits, is also problematic. If tax incentives are as ineffective as alleged, then no harm is done. Investors are not attracted, behavior is not distorted, and tax revenue is not forgone. Tax incentives cannot be harmful and ineffective at the same time unless taxpayers can take advantage of the tax incentives without actually investing, or unless investors who would have made the investment even without the tax incentives benefit from them.[76] However, this is an issue of design and implementation relevant to any form of government intervention.[77]

Most of the literature on tax incentives deals with various specific problems they cause, ultimately concluding that they are harmful.[78]

Nevertheless, building on this literature to support the conventional wisdom against tax incentives is unwarranted. A better approach, suggested in this paper, is to consider tax incentives as another instrument available to a government trying to attract investments. They should be adopted only if it is cost-effective, taking into account their specific design and all associated costs.

Answering the question raised in the second half of the argument against the use of tax incentives – whether tax incentives can attract FDI – is crucial to my proposal. Until ten years ago, a consensus existed in the literature that tax considerations have only a minor effect on FDI decisions.[79] Determinants like consumer market size, labor skills, infrastructure, trade policies and political and macroeconomic stability dominated decisions regarding investment location. Globalization has dramatically reduced the importance of these factors, and elevated the role tax incentives play. Former FDI barriers like tariffs and currency exchange controls are reduced or gone, making taxes a more decisive factor.[80] Moreover, advances on a number of fronts allow multinationals to spread production processes over multiple countries.81 Theoretical arguments and empirical studies support the claim that tax incentives are now an important determinant of FDI location. [82]

3. Tax Incentives Should Not Be Used to Compensate for an Unattractive Investment Environment

Opponents of using tax incentives to attract FDI argue that using them to compensate for corruption, political instability or lack of infrastructure fails to address these problems directly. Though this is an appealing argument, I suggest a more sophisticated approach.

If developing countries could create good infrastructure, a highly skilled labor force, zero inflation, a progressive tax and transfer system,[83] political stability, and a functioning judicial system, they would not be developing countries; they would be the United States. One must be realistic. Developing countries lack most of these qualities, which is precisely why they are less attractive to investors than developed countries. Total World FDI inflows in 2001 were $735 billion. Less than $110 billion went to poor countries. China alone accounted for $46 billion, nearly half of this amount.[84] Tax incentives are not used to attract FDI instead of adopting sound policies and building good institutions, but in addition to such efforts. The lack of progress on this front has little to do with offering tax incentives. These are difficult goals to achieve. Complicating the problem is that in many cases, government officials lack the correct set of incentives for reform.

Tax incentives, like any other market intervention, are justified if they correct market inefficiencies or generate positive externalities. My core assumption in this paper is that 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.

According to one extreme but well-known model, the small open economies of developing countries cannot tax FDI at all. Since capital is mobile, the tax burden is shifted to immobile factors like land and labor because their marginal product falls when the capital stock declines. In this scenario, it is more efficient to exempt FDI from taxation, except for taxation of economic rents, and impose taxes on the immobile factors directly.[85] Taking this last point to the extreme, the real issue raised by the third argument is one of subsidy, not tax. The question is how much the developing country should pay in order to attract FDI. The compensation is formally done through the tax system, as the country imposes little or no tax. In substance, the foreign investor receives the subsidy through the free use of infrastructure financed by domestic taxpayers.

This debate is not a question of theory, but of fact.[86] Tax incentives cannot be ruled out on theoretical grounds alone. Their use or not is a question of marginal cost/benefit analysis. Policymakers must compare the benefits of using the marginal dollar to build institutions directly versus using it to attract FDI. Both avenues are important, and I suspect that it is a rare case indeed where the marginal efficiency of direct government actions is so high that no money should be spent on attracting FDI. Measuring the exact marginal costs and benefits of these two options is difficult, but I assume both avenues involve diminishing marginal returns. As such, some type of tax incentives are justified in addition to those coming at no cost. The amount of incentives is a complicated question requiring careful analysis of each case.

One of the basic claims I make in this paper is that building sound institutions is very taxing on a government. The experience of the past 50 years teaches us that governments in many developing countries are not having much success. The gap in growth rates and absolute living standards is widening. Another advantage of tax incentives is that they require less from the developing countries’ governments. As I have mentioned already, spillovers are not automatic. The developing country must invest in education and build the right set of incentives for individuals to succeed. Even so, tax incentives are far less demanding than institution building in isolation from the world.

The theoretical case for using tax revenue to attract FDI is undiminished by the fact that multinationals will benefit. That benefit does not come at the expense of developing countries. Multinationals are not in a position to exploit developing countries unless there is an inherent difference in their bargaining positions. As long as there are enough players on both sides, multinationals cannot dissipate the entire surplus generated by the transaction by forcing developing countries to pay for the spillovers they gain.[87]

In some cases, tax incentives are always good,[88] and no cost/benefit analysis is required. This happens when tax incentives satisfy two conditions:

(1) They attract FDI that otherwise would not have been made; and

(2) They do not involve a transfer of tax revenue from domestic taxpayers to foreign investors. Such transfers are avoided if foreign investors pay taxes covering infrastructure and other costs incurred by the developing country as a result of investment activities. The foreign investor need not pay any taxes if it uses only pure public goods that impose no marginal costs on the developing country. Tax incentives attracting such marginal FDI are always justified because the spillovers they generate come at no cost to the local taxpayer.

The use of tax incentives may be justified even if some of the investments would have taken place anyway. This is true when the administrative costs of using separate tax rules for different investors are high and the overall outcome is justified under a cost/benefit analysis.

B. Tax Incentives versus a Low Corporate Income Tax Rate

Tax incentives are tax 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.[89] If the baseline is the corporate tax itself, then a low corporate tax rate is not an incentive since by definition tax incentives are targeted at specific types of investors or investments.

This implicates the general and well-known tradeoff between targeted and universal tax provisions. 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. 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.[90] 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.[91]

The disadvantage of targeting is that policymakers often cannot estimate which types of investments have the greatest positive spillover effects. As explained above,[92] spillovers take a variety of forms. Many spillovers are subtle advances that are easy to miss when tax incentives are restricted to investments in plant, equipment and job training. Conventional tax incentives, targeted at investments in machines, are based on rationales like the Harrod-Domar formula for capital formation or technology transfers. FDI in almost any economic sector enables the host economy to absorb and adopt international best practices in terms of governance and other institutional structures. Still, optimally tailoring tax incentives to lure FDI with the greatest potential spillover effects is a very complicated task.

Though policymakers might prefer to identify good potential investments on a case-by-case basis, there are many disadvantages to such a regime.[93] 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.[94]

I maintain that a general corporate income tax rate reduction is a viable possibility to attract growth-promoting FDI.[95] If FDI spillovers are large enough, and generated by a wide range of activities that are difficult for policymakers to define, then perhaps they are not such a bad form of tax incentive after all.

Tax holidays are usually targeted at production activities, since they are more likely to produce spillovers. Moreover, retail enterprises locate in consumer markets even without tax incentives. Targeting tax holidays to production therefore makes sense, even though it is likely to generate litigation by taxpayers trying to broaden the definition of the terms “production” or “manufacturing” by means of court interpretation. These litigation costs are justifiable in light of the increased tax revenue from corporations locating in the developing country to sell to local consumers.

Tax holidays must be carefully managed to avoid unintended pitfalls. If tax holidays are restricted to foreign investors, they either create a competitive disadvantage for domestic firms or invite abuse by domestic taxpayers forming foreign corporations and investing as “foreigners” in their own country, a practice known as “round tripping”. Three ways to mitigate these problems are: (1) apply the tax holiday to domestic taxpayers as well; (2) restrict benefited foreign enterprises from selling their products in the domestic market (“ring fencing”); or (3) put effective anti-avoidance measures in place.

Time limits are another potential issue with tax holidays. Since I oppose limiting their duration, what I am proposing is not a “tax holiday” in the traditional sense. Often, time limitations are imposed because policymakers desire a politically easy way to terminate an unsuccessful program. Designing the holiday so that FDI is not less attractive than the regular tax system eliminates the need for these time limits. One way to do this is to prohibit foreign taxpayers from deducting depreciation when the tax holiday is over. Depreciation deductions are only necessary for start-up FDI, where net earnings during the tax holiday were likely low or negative because of depreciation.[96]

Tax holidays result in greater revenue loss than more targeted tax incentives because they are available to investors who would invest even without the incentives. The revenue loss is reducible by limiting the tax holiday in time or by restricting the amount of revenue that is tax-exempt. Since I generally oppose time limitations on tax holidays, I think the more elegant solution is to give the investor a large fixed- sum tax credit, determined in size when the investment is made. Under the usual assumption of a uniform corporate income tax rate, this type of tax credit is equivalent to a tax rate reduction – possibly to zero if the credit is greater than the taxable income. Any unused portion of the tax credit rolls over to the following tax year, offsetting that year’s taxable income. The taxpayer computes tax liability every year under the regular tax system, offsetting it with whatever tax credit amount is left. Policymakers wishing to further limit the credit can stipulate that only a portion of it is usable each year, thus extending the effective period of the tax incentive. Once the tax credit is depleted, the investment is taxed at the normal rate.

Combating revenue loss from tax holidays with tax credits might be too administratively demanding for some developing countries. It also makes the investment less attractive to investors, who know that they eventually face regular tax rates. Furthermore, tax credits risk attracting investments that are easily relocated at the end of the incentive period, an undesirable phenomenon.[97] Offering a non- discriminating tax incentive system with the associated revenue loss may still make more sense than a targeted system that is difficult and expensive for a developing country to administer. A better solution in this scenario is an across-the-board rate reduction.[98] Activities generating economic rent, like mining and domestic retail sales, should not benefit from this rate cut unless the cost of differentiation is too high. A zero tax rate is undesirable because it eliminates the need to report to authorities, making it difficult to gage the effectiveness of the tax incentive regime.

C. Surrendering the Ability to Impose Corporate Income Tax on Domestic Taxpayers

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. 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.

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. Capital tends to be concentrated with the rich and exempting capital income from tax[99] obviously benefits them. Nevertheless, 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.[100]

Another consideration is the potential difference in tax incidences between developed and developing countries. According to Shah and Whalley, the incidence of corporate income tax in developing countries makes it somewhat regressive.[101] 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. Indeed, one of the current differences between tax systems in the developed and developing world is that rich countries rely on individual income taxes while developing countries rely on corporate income taxes.[102] Reducing corporate income tax rates may force developing countries to exert greater efforts to tax individual income, making their tax systems more like those of developed countries. Since we usually assume that developed countries adopt better policies, this is probably a desirable outcome.

A possible harmful effect of tax incentives is greater corruption and rent-seeking behavior on the part of government officials. We address this problem by limiting the discretion officials have by legislating the tax incentive, or by using a low universal corporate tax rate. In any case tax incentives, as a policy to promote growth in developing countries, are no more susceptible to corruption and rent- seeking behavior than growth policies like the provision of grants or loans. Tax incentives, if carefully designed, are probably less vulnerable to corruption than most other available options.


One of the current disadvantages of tax incentives is that their benefits may not accrue to the investor, depending upon his country of residence. Countries like the United States, the United Kingdom, or Japan impose taxes on residents’ worldwide income. These countries avoid double taxation by providing a tax credit for foreign tax paid (“FTC countries”).[103] The developing country’s tax incentives reduce the foreign tax paid by the investor, but they also reduce the foreign tax credit given to the investor by his country of residence. In this situation, tax incentives are ineffective in attracting FDI because the investor does not benefit from the tax incentive.

This result, harsh from both the investor’s and the host country’s point of view, is avoidable if the investor incorporates in the host country or channels the FDI through an entity incorporated in a third country that does exempt foreign income from tax.[104] Most FTC countries tax the income of foreign subsidiaries only when it is repatriated to the resident parent company in the form of dividends, interest, royalties, or capital gains.

Moreover, FTC countries often allow resident corporations to pool their foreign income and apply a foreign tax credit to all of it. The amount of foreign tax a corporation may credit against the tax liability in the resident country is limited by that country’s tax rate. Therefore, corporations with foreign-sourced income subject to foreign taxes at rates higher than the rate imposed by the resident country cannot credit the full foreign tax paid against the resident country’s tax. These corporations are in an “excess credit position.” Situated in an excess credit position, corporations benefit from tax incentives offered by the host countries in which they invest, since the income they generate there is added to the foreign-sourced income taxable in their country of residence. Tax incentives by host countries reduce the total foreign tax rate, which is the average of the tax rates of the high and low tax countries. The incentives enable the corporation to fully credit the foreign taxes paid against its resident country tax.

Empirical studies support the assumption that U.S. resident corporations do not pay U.S. effective tax rates on their foreign source income, although they are formally taxed by the United States on their worldwide income. In other words, tax incentives benefit investors, not the investors’ country of residence.[105] The actual division of the host country’s tax concessions between investors and their home (FTC) countries is unknown. Benefit accrues to the home country to the extent any tax is paid. This happens when foreign subsidiaries repatriate profits to the parent company, or when FDI is carried through a branch and the investor is in an “excess limit position” where no averaging with investments in high tax countries is possible.

To benefit from tax incentives, investors incur some transaction costs. If they operate through foreign branches, they are limited in their investment choices, since they must invest at a certain ratio of high to low tax countries to avoid excess credit or excess limit positions.

If they operate through foreign subsidiaries, MNCs suffer a cash flow constraint because they cannot repatriate foreign profits to the parent corporation without paying the home country tax. Operating through a foreign subsidiary may also result in accepting less desirable corporate and securities laws. A host of bureaucratic inconveniences are implicated as well, like long waiting periods for registration, fees, and greater risk of government appropriation. Even operating through an entity incorporated in a third country runs the risk of anti- avoidance provisions bearing various legal and accounting costs.

I therefore propose that FTC countries that are home to investors making FDI in developing countries either exempt or apply tax- sparing provisions to income generated in developing countries. This will make FDI more attractive in developing countries offering tax incentives for two reasons. First, it saves the costs that currently are incurred to avoid the home country tax. Second, to the extent that investors pay tax to their home country on foreign-sourced income from developing countries, this proposal eliminates the tax and directs the tax incentive benefits to the investor.

My proposal transfers revenue from the treasuries of developed countries to their resident multinationals making FDI in developing countries. It is an indirect transfer to developing countries, as it increases the attractiveness of their tax incentives. It is equivalent to giving governments of developing countries targeted aid that they can only use to attract FDI.[106] An alternative is for developed countries to give resident MNCs direct grants instead of a tax exemption to the extent these MNCs invest in developing countries. These two alternatives are essentially equivalent. The difference is in the administrative details.

Direct grants promote FDI to an unlimited extent, while operating through the tax system limits the size of the grant to the difference in tax rates. Of course, a direct grant is only as high as the developed country makes it. The advantage of operating through the tax system is saving the transaction costs of operating through branches or a subsidiary.[107] Direct grants to MNCs give them greater ability to divert the funds to their operations in developed countries, resulting in only minimal investment in developing countries. Establishing workable rules to link the grants to FDI in developing countries is an entirely separate challenge. Exempting the tax, on the other hand, establishes a direct link between the FDI in developing countries and the subsidy – only the income generated by FDI benefits from the tax break.

Transfers from rich to poor countries further benefit by imposing limitations on rich countries’ abilities to engage in tax competition with poor countries.[108] Tax harmonization is justifiable as part of an international equity regime transferring wealth to developing countries.[109]

For purposes of my proposal, I expand the definition of tax competition to include low corporate tax rates, even where those rates apply to domestic and foreign taxpayers alike. My expanded definition means a change in the status quo for some developed countries with low corporate tax rates. For example, Ireland would be required to raise its current corporate tax rate of 12.5% in order to decrease the relative disadvantage that developing countries have when competing with Ireland for FDI. Ireland can preserve its relative tax advantage over other developed countries by maintaining a lower tax rate than they do, but its rate should be significantly higher than the tax rates offered by developing countries, which may be zero.[110] This naturally involves a transfer of some FDI from low tax developed countries, such as Ireland, to developing countries. The size of the transfer may not be large since developed countries, such as Ireland, compete mostly against similarly situated countries. Ireland’s geographic location and EU membership mean that its low tax rates give it an advantage over other European countries. Since I favor allowing it to keep that relative advantage, the impact on countries like it should be relatively small. Relative advantage is the most crucial aspect of the investment attraction game.[111]

The potential for harmful tax competition exists among developing countries as well. With the application of anti-tax competition rules to developing countries, we establish two different harmonized tax levels – one for developed countries and the other for developing countries. I recommend distinguishing between developed and developing countries for these purposes on the basis of per capita GDP. In contrast to domestic tax and transfer systems that induce low-income people to quit work in order to qualify for subsidies, countries will not reduce their GDP per capita in order to benefit from the proposed transfer system. Of course, some arbitrary and perhaps unfair cut off must be established, as gradual rate implementation appears unworkable.

Unfortunately, complicated rules are required to prevent channeling of investments in developed countries through developing countries in order to take advantage of the tax exemption or tax sparing provisions. Other detailed rules are needed to assess the effective tax rates, taking into account statutory tax rates, the tax base (which can be eroded by measures such as accelerated depreciation deductions), and non-tax subsidies to investments. The various harmonization proposals deal with many of these issues, and some helpful provisions already exist under current laws.

My proposal could be implemented gradually through a bilateral treaty network between developed and developing countries, assuming that a significant number of the strongest economies will take the lead in signing. Multilateral agreement, a type of GAAT for taxes, would be much more suitable. The establishment of an organization equivalent to the WTO to monitor the agreement is desirable and potentially quite beneficial.[112]

A. Limits Imposed by International Organizations

International organizations such as the World Trade Organization (WTO), the IMF, and the EU often use their powers to force countries to give up tax incentives. When the affected country is a developed country, as in the case of EU members, this practice conforms with my proposal. I suggest that international organizations should disallow developed countries to engage in tax competition with developing countries. The EU, which is concerned with tax competition among its members, would embrace this type of redistribution policy.[113]

My proposal calls for a change in the IMF’s policy of conditioning loans to developing countries on their repeal of tax incentives.[114]I suggest that the IMF and World Bank experts who advise developing countries on fiscal issues abandon their policy against tax incentives and help governments in developing countries perform the cost/benefit analysis explained above.[115] Often this will lead to designing and implementing tax incentives to attract FDI that creates maximum spillover effects. These experts can also assist developing countries in taking the non-fiscal actions necessary for the country to absorb the FDI spillover effects.

I do not suggest limiting the WTO’s ability to ensure free trade. For example, I believe that developing countries should generally not be allowed to subsidize exports, because it is unlikely to give them an economic advantage over developed countries. Developed countries can benefit from such subsidies or trade taxes, because those subsidies change the terms of commerce to their advantage. Developing countries are unlikely to have that leverage, and will generally be worse off by imposing trade taxes or providing export subsidies. Those measures isolate their industries from efficiency-enhancing competition.[116] The WTO should not, however, prevent developing countries from offering tax incentives that attract FDI through non- export related tax concessions.[117] A recent WTO ruling that Thailand must end some of its investment incentives by 2004 is an intriguing example of a borderline case.[118]


I do not intend this paper to promote tax competition in general. Although I propose allowing developing countries to engage in tax competition in order to lure FDI away from developed countries, I also suggest forbidding developed countries, such as Ireland, from offering low tax rates that may undercut efforts by the developing countries. I further propose to limit tax competition between developing countries. While developing countries stand to benefit from tax competition that attracts FDI, they are severely harmed by tax competition that attracts portfolio investments.[119]

Under the current international tax regime, the country of residence has priority in taxing passive income. The problem is that it is not administratively feasible to tax portfolio income based on residence, because even fiscal authorities of developed countries lack the means to discern what income their residents have earned abroad. First, there is no uniform, worldwide system of tax identification numbers. Second, investors can channel their portfolio investments through a tax haven with bank secrecy. Consequently, even if the payer in the host country is willing to disclose information to the tax authorities of the investor’s country of residence, it will not have any information to disclose.

Developing countries are unlikely to attract portfolio investments. Portfolio investors tend to invest in the large capital markets in developed countries, primarily the United States, Western Europe and Japan. The few rich residents of developing countries usually invest their money outside of the country and do not report that investment income to their tax authorities. As Professor Avi-Yonah describes:

Latin American countries provide a prime example: after the enactment of the portfolio interest exemption, about $300 billion fled from Latin American countries to bank accounts and other forms of portfolio investment in the United States. (footnote omitted) Most of these funds were channeled though tax-haven corporations and therefore escaped taxation in the country of residence. Estimates of the capital flight from all developing countries to the United States in the 1980s range as high as $148 billion in a single year.[120]

Eliminating this form of tax competition, which is really a problem of tax evasion, is especially beneficial to developing countries.[121] There are several proposals to do so. A European Union proposal would allow countries to choose between levying a 20% withholding tax on interest payments to individual residents of another member state or providing a tax information report to the investor’s country of residence.[122] This proposal does not help developing countries because it is limited to interest payments to residents of EU member states.[123] Even if it were not so limited, the proposal allows host countries to keep the withheld taxes for themselves. Most would likely choose the tax withholding option over the reporting option.

A more sensible proposal is the one offered by Professor Avi- Yonah. He suggests a mandatory 40% withholding tax, refundable upon the taxpayer’s provision of proof that his income has been reported to his country of residence.[124] Special rules could address cases of multiple-residence taxpayers. This proposal is likely to help developing countries, because investors prefer to report the income to their countries of residence and pay a lower tax rather than pay the high withholding tax. Channeling investments through tax havens with bank secrecy laws would not help taxpayers avoid the withholding tax, because proof of residency would be required.

Professor Roin raises a concern that low-tax countries, or countries that do not tax their residents’ foreign portfolio investment income, might “sell” their residence status, thus allowing portfolio investors to avoid the withholding tax without paying tax in their true country of residence.[125] I do not have data to evaluate the relevance of this claim, but I assume that it is much more difficult and expensive to buy a residency status than to open a bank account in a tax haven. Therefore, few rich investors would use this route of evasion. As Professor Roin suggests, to combat such schemes “real” residence countries must be able to obtain taxpayer information from source states. This requires a high degree of international information sharing, a potentially difficult achievement given the current political situation. Professor Roin suggests a comprehensive solution to problems of tax evasion in cross-border transactions by fighting secrecy laws, and by “distinguishing between situations where secrecy can be maintained as a form of protection against political (or racial or ethnic) persecution and those where it serves less laudable goals.”[126]

I have a limited solution, addressing only the problem of false residency that creates a loophole in Professor Avi-Yonah’s proposal. Source countries where the portfolio investments take place can impose a 40% withholding tax, refundable upon the taxpayer’s provision of proof that such income has been reported to the country of residence. The source country keeps 10% of the tax withheld each time a false residency is exposed. The “real” country of residence collects the tax and either gives the taxpayer a tax credit for the 10% of tax withheld by the source country, or uses this amount as a civil penalty for attempted tax evasion. This mechanism provides source countries with an incentive to share information with residence countries in order to expose fraudulent residencies. Otherwise, they lack the incentive to share this information in order to attract investments. If a 10% withholding tax is too low to counteract the incentive problem, a higher rate of withholding tax is a possible solution.

The difficulty with my proposal is that it requires the participation of most, if not all, developed countries. Without uniform or nearly uniform participation, most countries would refuse to commit for fear of driving away portfolio investments.


Over the last 50 years, developed countries have transferred more than one trillion U.S. dollars to developing countries. Even so, the world is more unequal today than at any time in world history. A large share of the population lives at or below subsistence level. The key to raising living standards and reducing poverty around the world is increasing productivity growth in the poorest countries. About 15 years ago, economic growth theory experienced a revival with the development of the new growth theories. I build on these new understandings by taking an unpopular stand regarding tax incentives. I suggest that developing countries should use tax incentives to attract multinational enterprises to achieve market discipline, gain access to technology and benefit from productivity spillovers. Foreign direct investments in almost any economic sector enable the host economies to absorb and adopt international best practices. In addition to promoting sustained growth, multinationals’ activities help developing countries fight the “brain drain” phenomenon and escape from a permanent comparative advantage in traditional low-tech industries.

Unfortunately, the spillovers are not automatic. In addition to offering well-designed tax incentives (which may be a simple uniform low or zero tax rate), the developing countries must take the non- fiscal actions necessary to allow the country to absorb the FDI spillover effects. The developing country must invest in education and build the right set of incentives for individuals to succeed. Even so, these challenges are far less demanding than building these good institutions in isolation from the world. Trade with other countries and FDI are the ways in which the developing country can benefit from spillovers.

I further suggest that rich countries should replace some of their foreign direct aid with an equity-based tax expenditure policy. 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. My proposal transfers revenue from the treasuries of developed countries to developing countries, but it does so indirectly and with targeted money. It is equivalent to giving the governments of developing countries money that can be used only to attract FDI. 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.

The transfers from rich to poor countries can be further increased by imposing limitations on rich countries’ ability to engage in tax competition with poor countries. I argue 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.

My proposal defines tax competition in broad terms, extending it to include low corporate tax rates. This is true even if the low corporate tax rates are applied to both domestic and foreign taxpayers. Since developing countries may also engage in harmful tax competition, we can apply similar anti-tax competition rules to them. This leads to the establishment of two different harmonized tax levels, one for developed countries and the other for developing countries.

In the meantime, international organizations such as the WTO, the IMF, and the EU should refrain from using their powers to force developing countries into giving up tax incentives. They should also make it as difficult as possible for developed countries to engage in tax competition with developing countries. The IMF should change its policy of conditioning loans to developing countries on the repeal of any tax incentives. Instead, the IMF should help governments in developing countries to perform the cost/benefit analysis mentioned in this paper. The IMF should then aid these countries in designing and implementing tax incentives to attract FDI with maximum spillover effects, based on the specific country’s situation.

Finally, I offer a small contribution to Professor Avi-Yonah’s proposal to eliminate tax competition over portfolio investments. If enacted, my proposal would be especially beneficial to developing countries.

The proposal presented in this paper is in four separate parts. The first part establishes the case for the use of tax incentives. The second suggests that developed countries should give exemptions or tax sparing to foreign-source income generated by their residents through FDI in developing countries. The third imposes limits on the abilities of developed countries to engage in tax competition with developing countries. The fourth limits tax competition among developing countries. Each of the parts stands alone, but the adoption of each additional part increases the likelihood that the proposal will succeed in promoting growth in developing countries.

As a closing note, I draw the reader’s attention to the fact that there is one country whose recent history supports each and every part of my basic model: (1) that tax incentives attract FDI; (2) that FDI promotes economic growth; (3) that tax incentives can be limited to foreign investors; and (4) that if incentives are offered to domestic taxpayers as well, the lost corporate tax revenue can be picked up by other components in the tax system. The country is Ireland.

In the 1950s Ireland was the poorest country in Europe. Its GDP per capita was half that of the United Kingdom. Today, its GDP per capita is greater than the U.K.’s. Ireland has enjoyed 15 years of uninterrupted growth and was Europe’s fastest growing economy in the 1990s, with an average growth rate of over 9% per year from 1995 to 2001.[127] The key to Ireland’s economic success was attracting FDI by offering foreign investors a low corporate tax rate of 10%, so long as those investors agreed to engage in manufacturing activity within Ireland.[128] American high-tech corporations, attracted by the low corporate tax rates, established subsidiaries in Ireland. Soon after, spillover effects materialized. Ireland became the European leader in software exports with 500 of its own newly formed corporations. In response to pressure from the European Union, Ireland changed its tax law beginning in 2003. The new corporate tax rate is 12.5% across the board. Corporations that had already invested were grandfathered into the 10% tax rate until 2010.

Tax policy is not the only factor that attracted FDI to Ireland, but it is the dominant one. Without it, these dramatic changes would not have happened. The people in Ireland were not exceptionally educated or equipped for high-tech industry as compared to other European countries. The nature of the foreign direct investments Ireland attracted, combined with the spillover effects, made it the European leader in software exports.

The story of Ireland demonstrates the great potential and optimism at the core of the new growth theories. The key to growth is innovation. Only the latest technology counts, and countries that are behind can catch up by adopting the latest technology without incurring the burden of writing off investments in old technologies. Once the new technology is in place, through FDI, workers in developing countries can achieve higher productivity than workers in developed countries using older technologies. The technology is improvable by adapting it to the circumstances in the developing country. Accordingly, growth depends significantly on intelligent people being exposed to current technology. These spillover effects give great hope for developing countries.

Yoram Margalioth[*][*] Deputy Director, International Tax Program, Harvard Law School; Assistant Professor of Law, IDC. I am grateful for comments from Reuven Avi-Yonah, Richard Bird, Graeme Cooper, John Donohue, Dan Halperin, Elhanan Helpman, Assaf Jacov, Louis Kaplow, Saul Levmore, Doug Lichtman, Ollie Oldman, Eric Posner, Sharon Rabin-Margalioth, Mark Ramseyer, Diane Ring, Julie Roin, Dan Shaviro, Lior Strahilevitz, Alan Sykes, Richard Vann, David Weisbach and Al Warren. A preliminary draft of this article was presented at Chicago Law School’s Work-in- Progress Workshop, the IDC Law faculty seminar, and the Harvard Tax Lunch Group. I would like to thank the participants in these events for their helpful comments and suggestions. Finally, I would like to thank the editors of this Journal for their excellent editing work.

[1] “An estimated 1.2 billion people subsist on under $1 per day, and the majority of the developing world’s population lives on less than $2 per day.” IAN GOLDIN ET AL., THE ROLE AND EFFECTIVENESS OF DEVELOPMENT ASSISTANCE: LESSONS FROM WORLD BANK EXPERIENCE, at ix (2002), available at

13080_Development_Effectiveness.pdf [hereinafter WORLD BANK REPORT].

[2] According to a World Bank study, per capita income in the richest countries was eleven times greater than in the poorest countries in 1870, thirty-eight times greater in 1960, and fifty-two times greater in 1985. In the early 1990s, of $23 trillion in global GDP, only $5 trillion – less than twenty percent – was generated in developing countries – even though these countries accounted for about eighty percent of the world’s population. TATYANA P. SOUBBOTINA & KATHERINE SHERAM, BEYOND ECONOMIC GROWTH 23 (2000).

[3] See LantPritchett,Divergence, Big Time,11J.ECON.PERSP. 3,7(1997).

[4] Unlike measures of poverty that vary across time or cultures, the concept of subsistence cannot change. Incomes cannot get much lower than $1 per day. Below this level, widespread starvation and death set in. For measurements of subsistence and the relationship between caloric intake, average height and longevity, see Robert Fogel, New Findings on Secular Trends in Nutrition and Mortality: Some Implications for Population Theory, in THE HANDBOOK OF POPULATION AND FAMILY ECONOMICS (Rosenzweig & Stark, eds. 1997).

[5 ] See Pritchett, supra note 4, at 7.

[6] Between 1950 and 1990, world per capita GDP grew at a rate of 2.2 percent per year. Between 1850 and 1950 the annual growth rate was only 0.88 percent. Prior to 1850, it was less than 0.2 percent. Over most of the world’s history growth rates were close to zero. See ANGUS MADDISON, MONITORING THE WORLD ECONOMY

1820-1992 20 (1995) (finding that between 500 to 1500 A.D. the growth rate was zero).

[7] Only twenty-six countries are considered high-income countries. The cutoff in

2002 was $9,266 per capita or more. The combined population of 0.9 billion in these

26 countries is less than one-sixth of the world’s population. High-income country, World Bank DEPweb Glossary, at modules/glossary.html#high-income (2001).

[8 ] The UnitedStates’real GDP per capita grewfrom about $1,450in1776to

$36,300 in 2002. This means that U.S. living standards are twenty-five times higher today than they were 226 years ago. See United States, CIA World Factbook 2002, at

[9 ] RobertE. Lucas, Jr., On the Mechanics of Economic Development,22 J.

MONETARY ECON. 3, 4 (1988).

[10] Per capita income in the Asian Tigers increased from eighteen percent of the developed countries’ average in 1965 to sixty-six percent in 1995. SUOBBOTINA & SHERAM, supra note 2, at 25. After the Second World War Japan’s GDP per capita was about one-sixth that of the United States. Now it is about seventy-five percent. The average citizen of those countries is twenty times as rich as her grandparents.

[11] The average per capita income in African countries equaled fourteen percent of the developed countries’ level in 1965 and only seven percent in 1995. Id. at 25, 26.

[12] See Alan Heston et al., Penn World Table Version 6.1, CENTER FOR



(arguing that “(g)rowth has been considered the major and in many cases the only objective of economic policy in developing countries for much of the period following the end of World War II. . .[T]he overwhelming objective of tax and budgetary policy in developing countries must be the acceleration of economic development.”); see also Kevin Davis, Ethnically Homogeneous Commercial Elites in Developing Countries, 32

LAW & POL’Y INT’L BUS. 331, 359 (2001) (suggesting “redistributive policies that operate outside of, and alongside, markets rather than redistributive policies designed to displace or severely constrain market functions,” and citing empirical studies finding that “market-oriented developing states generate substantially higher levels of economic growth than developing states that assign markets a highly constrained role.”).

[14] GDP isdivided bythenumberofpeople (GDPpercapita) orworkers (GDP

per worker) to account for differences in populations. GDP is translated into dollars using Purchasing Power Parity (PPP) rather than prevailing exchange rates so that the actual ability to purchase products and services is measured.

[15] This raises a questionof causality, since a healthier workforceis a more productive one. It is difficult to say which of the two came first.

[16] The average formal education in the least developed countries is three years, compared with eleven in the most developed countries.

[17 ] SeeWORLDBANK REPORT,supranote 1,atxiii;David Dollar &Art Kraay,

Growth is Good for the Poor (2001), available at

research/growth (finding a one-to-one relationship, on average, between the income

growth rate of the poor and the growth rate of per capita income); see generally Philippe Aghion et al., Inequality and Economic Growth: The Perspective of New Growth Theories, 37 J. ECON. LITERATURE 1615 (Dec. 1999) (reviewing the empirical literature and suggesting theoretical explanations for the counterintuitive positive correlation between growth and equality).


[hereinafter THE QUEST FOR GROWTH].

[19] According to Easterly, infant mortality in the richest fifth of countries is four out of every 1,000 births; in the poorest fifth of countries it is 200 out of every 1,000 births; see also Lant Pritchett & Lawrence H. Summers, Wealthier is Healthier, 31 J. HUM. RESOURCES 841, 863-65 (1996) (underpinning Easterly’s calculation that the deaths of about half a million children in 1990 would have been averted if Africa’s growth rate in the 1980s had been 1.5 percentage points higher).

[20 ] Inthepoorestofnations, nearly halfofallchildren under theageofthree are

abnormally short because of nutritional deficiency.

[21] Millions of children die every year from easily curable diseases, such as bacterial pneumonia and dehydration from diarrhea. Bacterial pneumonia, for example, can be cured by a five-day course of antibiotics that costs about twenty-five cents.

[22] Aparticularlycruel burdenfallsonthe children, asparentstake out meager

loans in exchange for consigning or selling a child to a factory or plantation owner. An estimated 20 million, and perhaps as many as 40 million, girls and boys in South Asia toil in this debt servitude, hunched over looms, making bricks, or rolling cigarettes by hand. Countless others spend their childhood and adolescence in domestic servitude, sweeping floors and scrubbing pots and pans. CAROL BELLAMY, THE STATE OF THE WORLD’S CHILDREN 2000 20 (2000). Even more horrifying is the evidence of widespread child prostitution in developing countries, as well as the use of 200,000 child soldiers ranging from six to sixteen years of age to fight wars in poor African countries.

[23] See THE QUEST FOR GROWTH, supra note 22, at 33 (stating that $1 trillion, measured in 1985 dollars, was transferred between 1950 and 1995, and that the financing-gap approach was one of the largest policy experiments ever based on a single economic theory).

[24] WORLDBANK REPORT,supra note2, atxv.

[25] The World Bank’s main goal is to fight poverty in the world, and the IMF’s purpose is to relieve countries from short-term financial distress. Both institutions employ dedicated, hard working economists and lawyers who travel around the world and advise governments on fiscal and macroeconomic matters, mostly in an effort to help those countries achieve sustained economic growth.

[26] Brasilia, thecapital ofBrazil,andtheAkosomboDam ontheVolta River in

Ghana are two such examples.

[27] Heavily Indebted Poor Countries Unit & External Affairs Dep’t, World Bank, Frequently Asked Questions (2002), at faq.html (stating that approximately $55 billion in debt forgiveness is planned); IMF and Int’l Dev. Ass’n, Heavily Indebted Poor Countries (HIPC) Initiative: Status of Implementation 9 (2002), available at date/progress-to-date.html (reporting that $37.2 billion in debt already has been forgiven).


167 (1981).

[29] See Zvi Griliches, The Search for R&D Spillovers, 94 SCAND. J. ECON. 29 (Supp. 1992) (evaluating calculations of the social rates of return for research and development).

[30] Thisis whyColumbusmistakenlythoughthefoundanewroute toIndia.

[31] In fact, John Harrison was viciously denied the prize for many years after the creation of his fabulous invention. He finally received it by a special act of Parliament. HEATHER & MERVYN HOBDEN, JOHN HARRISON AND THE PROBLEM OF LONGITUDE (2002).

[32] FDIisdefined as“ownershipofassetsinonecountry byforeign residentsfor

purposes of controlling the use of those assets.” EDWARD M. GRAHAM & PAUL R. KRUGMAN, FOREIGN DIRECT INVESTMENT IN THE UNITED STATES 7 (Institute for International Economics 1989) [hereinafter GRAHAM & KRUGMAN]. The aspect of control “distinguishes direct investment from portfolio investment, which is simply the establishment of a claim on asset for the purpose of realizing some return.” Id. at

8. Direct investment is defined by the U.S. Department of Commerce as ownership or control of ten percent or more of an enterprise’s voting securities, or the equivalent, by a single person (including legal entities). BUREAU OF ECON. ANALYSIS, U.S. DEP’T OF COMMERCE, FOREIGN DIRECT INVESTMENT IN THE UNITED STATES 39 (2001).

[33] In1990,countriesthat werenotOECDmembers(i.e.,averybroad definition

of developing countries) received approximately fifteen percent of the $200 billion in world FDI. James R. Hines, Jr., “Tax Sparing” and Direct Investment in Developing Countries, in INTERNATIONAL TAXATION AND MULTINATIONAL ACTIVITY 39 (James R. Hines, Jr. ed., 2001).

[34 ] Recent estimatessuggest there are about 65,000MNCs today, with about

850,000 foreign affiliates across the globe. Their economic impact can be measured in different ways. In 2001, foreign affiliates accounted for about 54 million employees, compared to 24 million in 1990; their sales of almost $19 trillion were more than twice as high as world exports in 2001, compared to 1990 when both were roughly equal; and the stock of outward foreign direct investment (FDI), increased from $1.7 trillion to $6.6 trillion over the same period. Foreign affiliates now account for one-tenth of world GDP and one-third of world exports. UNITED NATIONS CONFERENCE ON TRADE AND DEV., WORLD INVESTMENT REPORT 2002: TRANSNATIONAL CORPORATIONS AND EXPORT COMPETITIVENESS xv (2002).

[35] 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)

[36] Though an increase could be helpful assuming the monies are well spent. The amounts transferred have been decreasing in real (as opposed to nominal) terms over the years. Official flows (ODA) are now significantly lower than during the 1980s. See


[37] See STANLEY S. SURREY & PAUL R. MCDANIEL, TAX EXPENDITURES 103-108 (1985) (listing negative aspects of tax expenditures).

[38] See Reuven S. Avi-Yonah, The Structure of International Taxation: A Proposal for Simplification, 74 TEX. L. REV. 1301, 1303 (1996) (finding that “a

coherent international tax regime exists that enjoys nearly universal support . . . . embodied both in the model tax treaties developed by the Organization for Economic Co-operation and Development (OECD) and the United Nations and in the multitude of bilateral treaties that are based on those models.”).

[39] Active business income istaxed in the country in which it originates (the

source country). Passive income, such as dividends, interest, royalties and capital gains, is taxed in the country in which the recipient resides (the residence country). Countries either exempt their residents’ active foreign- sourced income, or give them a tax credit for foreign taxes paid if they tax worldwide income. The source (location) of active business income is defined in tax treaties as income generated by a “permanent establishment,” or, in U.S. terms, “effectively connected with the conduct of a trade or business in the United States.” There are important exceptions and flaws to these rules, but they are not relevant to the argument made in this paper.

[40] With the minor exceptionthat the impositionof the burdenof alleviating

double taxation on the residence country, in the case of active business income, was

meant to give an advantage to poor countries. They are more likely to be the source countries, while investors will typically come from rich countries. See Reuven S. Avi- Yonah, Globalization, Tax Competition, and the Fiscal Crisis of the Welfare State, 113 HARV. L. REV. 1573, 1649 (2000) [hereinafter Avi-Yonah, Globalization].

[41] See, e.g., Albert J. Rädler, General Description: Germany, in COMPARATIVE


[42] The use of the international tax system to transfer tax revenue to the governments of poor countries was proposed thirty years ago by Peggy and Richard Musgrave. Their proposal offered a uniform rate schedule for corporate and withholding tax, agreed upon by international convention (equivalent to GAAT) wherein tax rates related inversely to per capita income in the host country (the developing country) and directly to per capita income in the home country (the developed country). See RICHARD A. MUSGRAVE & PEGGY B. MUSGRAVE, Inter- nation Equity, in MODERN FISCAL ISSUES 63, 74 (Richard M. Bird & John G. Head eds., 1972). My proposal has a different focus. I do not suggest replacing direct money transfers with transfer of tax revenue to the governments of developing countries. I propose to use the international tax regime to pay multinationals with some of the tax revenue of rich countries, thereby inducing foreign direct investments in developing countries.

[43] The leading legalpaper isAvi-Yonah,Globalization,supranote 41;seealso

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT, HARMFUL TAX COMPETITION: AN EMERGING GLOBAL ISSUE (1998) [hereinafter OECD Report] (arguing in favor of limiting global tax competition).

[44] The leading legal paper is Julie Roin, Competition and Evasion: Another

Perspective on International Tax Competition, 89 GEO. L.J. 543 (2001).

[45] The question of whether or not global taxes should be harmonized is still open in the economic literature as well. Economic analysis crucially depends on modeling assumptions, and there is not yet consensus on one model. For a relatively simple classic stylized model, see JACOB A. FRENKEL ET AL., International Tax Competition and Gains from Harmonization, in INTERNATIONAL TAXATION IN AN INTEGRATED WORLD 197 (1991).

[46] Awellknown, widelyrecognized,andimportantpolicythat isnotthesubject of this paper.

[47 ] See R.F. Harrod, An Essay in Dynamic Theory, 49 ECON. J. 14 (1939); see also

Evsey D. Domar, Capital Expansion, Rate of Growth, and Employment, 14

ECONOMETRICA 137 (1946).

[48] See Robert M. Solow, A Contribution to the Theory of Economic Growth, 70

Q. J. ECON. 65 (1956); Robert M. Solow, Technical Change and the Aggregate

Production Function, 39 REV. ECON. & STAT. 312 (1957).

[49] Ordinary goods are rival goods, but information is non-rivalrous. For example, use of the double entry bookkeeping accounting method by one accountant

does not affect another’s ability to use it.

[50] The number of workers does not increase because labor is a fixed factor, but technology makes each worker more effective. This gives the same result as an increase in the number of workers.



[52] See Lucas, supra note 10.

[53] Lucas also used this example to show that under the model, each American worker should have about 900 times more capital than each Indian worker. The actual American advantage was closer to twenty times more capital. Paul Romer advanced a similar critique of the neoclassical model, comparing productivity in the United States and the Philippines. In order to justify the model’s predictions for differences in labor productivity between the two countries, the United States’ share of investment would have to be thirty to 100 times the Philippines’ share. In fact, the United States’ share was only twice as large. See Paul M. Romer, The Origins of Endogenous Growth, 8 J. ECON. PERSP. 3, 6 (1994). For similar criticism, see also Paul M. Romer, Increasing Returns and Long-Run Growth, 94 J. POL. ECON. 1002 (1986). The latter paper and Lucas’s paper, supra note 10, pioneered the new endogenous growth theories.

[54] However, at least one influential paper defends the neoclassical model. See N. Gregory Mankiw et al., A Contribution to the Empirics of Economic Growth, 107 Q. J. ECON. 407 (1992) (arguing that if saving rates vary across countries, reflecting differences in tastes or culture, and the model includes human capital as well as physical capital, the evidence on the international disparity in levels of per capita income and rates of growth is consistent with a standard Solow model.) But see Gene M. Grossman & Elhanan Helpman, Endogenous Innovation in the Theory of Growth, 8 J. ECON. PERSP. 23, 29 (1994) (answering that the assumption of a common rate of technological progress in all ninety-eight countries included in the study over twenty- five years “is simply indefensible,” and that “[t]he rate at which producers in Japan have acquired new technologies, be they technologies that were new to the global economy or those that were new only to the local economy or the individual firm, has been markedly different from the rate of technology acquisition in Chad, for example.”). Most of the literature mentioning this debate sides with Grossman and Helpman, but views Mankiw’s paper as a respectable excuse for those who cling to the neoclassical model.

[55] Aclassicexample isIsaacNewton’s famously modest saying:“IfIhavebeen

able to see further, it was only because I stood on the shoulders of giants,” referring to the work done by previous scientists such as Copernicus, Galileo and Kepler.

[56] Romer, The Origins of Economic Growth, supra note 56, at 20-21.

[57] In developing countries, technology acquisition often amounts to adapting existing methods to local circumstances. See Robert E. Evenson & Larry E. Westphal, Technological Change and Technology Strategy, in 3 HANDBOOK OF DEVELOPMENT ECONOMICS 2209 (Jere Behrman & T.N. Srinivasan eds., 1995).

[58] SeeMonaHaddad&Ann Harrison,AreTherePositiveSpillovers fromDirect

Foreign Investment? Evidence from Panel Data for Morocco, 42 J. DEV. ECON. 51 (1993); see also Rajeeva Sinha, Foreign Participation and Technical Efficiency in Indian Industry, 25 APPLIED ECON. 583 (1993).

[59] See GRAHAM & KRUGMAN, supra note 38, at 36.


AFFECTS HOST COUNTRIES (World Bank, Policy Research Working Paper No. 1745,


[61] See Brian J. Aitken & Ann E. Harrison, Do Domestic Firms Benefit from Direct Foreign Investment? Evidence from Venezuela, 89 AM. ECON. REV. 605 (1999); see also Ari Kokko et al., Local Technological Capability and Productivity Spillovers from FDI in the Uruguayan Manufacturing Sector, 32 J. DEV. STUD. 602 (1996).

[62] See E. Borensztein et al., How Does Foreign Direct Investment Affect

Economic Growth?, 45 J. INT’L ECON. 115, 116-118 (1998).

[63] See generally Massimo Motta et al., FOREIGN DIRECT INVESTMENT AND

SPILLOVERS THROUGH WORKERS’ MOBILITY (Center for Economic Policy Research


[64] See

[65] See CIA World Factbook 2002, supra note 8; see also Anisur Sina, President’s Message, Bangladesh Garment Manufacturers and Exporters Association, at



AND ROUTE 128 105-13 (Harvard University Press 1994).

[68] See Mancur Olson, Jr., Big Bills Left on the Sidewalk: Why Some Nations are Rich, and Others Poor, 10 J. ECON. PERSPECTIVES 3, 19-20 (1996) (stressing the importance of institutions).

[69] SeeAmir N.Licht,David’sDilemma: ACaseStudy ofSecuritiesRegulation in a Small Open Market, 2 THEORETICAL INQ. L. 673, 675-79 (2001).

[70] It is interesting to note that the United States has the greatest comparative advantage possible in agriculture. Yet only 2 percent of its economy is devoted to it. Most developing countries, including those ill-suited to agriculture, base most of their economy on it. See CIA World Factbook 2002, supra note 8.

[71] The economic damage suffered by developing countries when skilled

individuals emigrate is estimated at many billions of dollars. This exceeds the amount of foreign aid given to developing countries. For an ingenious proposal to compensate developing countries for this loss by taxing individual emigrants and their employers, see TAXING THE BRAIN DRAIN 46-47, 115-19 (Jagdish N. Bhagwati & Martin Partington eds., 1976).


1995); TAX POLICY HANDBOOK 165 (Parthasarathi Shome ed., 1995).

[73] See Alex Easson, Tax Incentives for Foreign Direct Investment Part I: Recent Trends and Countertrends, 55 BULL. INT’L FOREIGN DIRECT INVESTMENT 266, 266 (2001); id., (Summarizing the literature and stating that tax incentives “are bad in theory principally because they cause distortions: investment decisions are made that would not have been made without the inducement of special tax concessions.”).

[74 ] Id. (reviewing the literatureand concluding that tax incentives “are bad in

practice, being both ineffective and inefficient. They are ineffective in that tax considerations are only rarely a major determinant in FDI decisions; they are inefficient because their cost, in terms of revenue forgone, often far exceeds any benefits they might produce.”)

[75] See supra Part II.B.

[76] See Boris Bittker, Equity, Efficiency and Income Tax Theory: Do Misallocations Drive Out Inequities?, 16 SAN DIEGO L. REV. 735 (1979) (Explaining an efficiency and equity tradeoff in the effect of tax provisions).


DEVELOPING COUNTRIES 201 (Perry et al. eds., 2000).

[78] For a good description of advantages and disadvantages of various tax incentives, see Howell Zee et al., Tax Incentives for Business Investment: A Primer for Tax Policy Makers in Developing Countries, 30 WORLD DEVELOPMENT 1497 (2002).

[79] See, e.g., 2 TAX LAW DESIGN AND DRAFTING 986, 998, 1009 (Victor Thuronyi ed., 1998); Janet Stotsky, Summary of IMF Tax Policy Advice, in TAX POLICY HANDBOOK 279, 282 (IMF 1995).

[80 ] SeeEasson, supranote80,at272.

[81 ] See Avi-Yonah, Globalization, supra note 47, at 1589 (“An Intel chip developed at a design center in Oregon might be manufactured at a wafer fabrication facility in Ireland, packaged and tested in Malaysia, and then sold to a customer in Australia. Another chip might be designed in Japan, fabricated in Israel, packaged and tested in Arizona, and sold in China.”).

[82] For acommon-sensetheoreticalargumentseePerry etal.,supranote 84,at

204 (“After all, if investment is assumed to respond to changes in the cost of capital, and incentives change the cost of capital, then incentives must affect investment.”). For an empirical study that supports this assumption see FISCAL POLICY: LESSONS FROM ECONOMIC RESEARCH 339 (Alan J. Auerbach ed., 1997); but see Austan Goolsbee, Investment Tax Incentives and the Supply of Capital Goods, 113 Q. J. ECON.

121 (1998) (finding that much of the benefit of tax incentives is captured by the suppliers of capital goods through higher prices); Cf. Kevin A. Hassett & R. Glenn Hubbard, Are Investment Incentives Blunted by Changes in Prices of Capital Goods?,

1 I’NTL FIN. 103 (1998) (suggesting regression problems in Goolsbee’s econometrics). Other empirical studies supporting this argument include G. Peter Wilson, The Role of Taxes in Location and Sourcing Decisions, in STUDIES IN INTERNATIONAL TAXATION 195, 228-31 (Alberto Giovannini et al. eds., 1993); INTERNATIONAL TAXATION AND MULTINATIONAL ACTIVITY 9 (James R. Hines, Jr. ed., 2001); FISCAL POLICY: LESSONS FROM ECONOMIC RESEARCH 401 (Alan J. Aucherbauch ed., 1997); Michael Devereux & Rachel Griffith, Taxes and the Location of Production: Evidence from a Panel of Multinationals, 68 J. PUB. ECON. 335 (1998); James R. Hines, Jr., Lessons from Behavioral Responses to International Taxation, 52 NAT’L TAX J. 305 (1999); James R. Hines, Jr. & Mihir Desai, “Basket Cases”: Tax Incentives and International Joint Venture Participation by American Multinational Firms, 71 J. PUB. ECON. 379 (1999); Steven Clark, Tax Incentives for Foreign Direct Investment:

Empirical Evidence on Effects and Alternative Policy Options, 48 CAN. TAX J. 1139 (2000); Christopher Taylor, The Impact of Host Country Government Policy on U.S. Mutlinational Investment Decisions, 23 WORLD ECON. 635 (2000); Harry Grubert & John Mutti, Do Taxes Influence Where U.S. Corporations Invest?, 53 NAT’L TAX J.

825 (2000). China provided an interesting test case when it repealed some of its tax incentives for foreign investors in 1996. Within a few months, the level of FDI dropped. China largely restored the incentives by the end of 1997. See Easson, supra note 73, at 268. MNC executives now readily admit that incentives play an important role in their investment decisions. See id. at 272.

[83] Evenunder myproposal,taxincentives shouldnotbeusedtocompensatefor

deficiencies in the way tax provisions are drafted. The time spent designing, implementing, and monitoring the tax incentive provisions could instead be spent on improving the design, implementation and monitoring of the regular tax provisions.

[84] Overall, the category “developingcountries” (which includes Hong-Kong,

Singapore, Taiwan and South-Korea) benefited from $225 billion of FDI inflows. The countries that attracted the most FDI offered significant tax incentives. See UNCTAD WORLD INVESTMENT REPORT (2002), available at Templates/Page.asp?intItemID=1894&lang=1.

[85] The only exception is when the investor’s home country taxes his worldwide income, providing a tax credit for foreign taxes paid – assuming the investor is in a position to use the credit (in an ‘excess limit position’). See Part III infra; see also Joel Slemrod, Tax Policy Toward Foreign Direct Investment in Developing Countries in Light of Recent International Tax Changes, in FISCAL INCENTIVES FOR INVESTMENT AND INNOVATION 289, 291 (Anwar Shah ed., 1995).

[86] This istrue even ifwedo not assume that any tax imposed bydeveloping countries is shifted to labor and land income.

[87] Part IV of this paper proposes a mechanism to allow developing countries to form a cartel. This gives them a bargaining position advantage over multinationals, assuming that both parties are otherwise in equal bargaining positions. If we assume that multinationals have an advantage due to the relatively small number of players, then the proposed mechanism will offset this advantage.

[88 ] Assuming theyarewelldesigned toavoidmanipulation.

[89] Ireland’s current corporate tax rate is a good example. It used to have a 10%

corporate tax rate for foreigners engaged in manufacturing activity within Ireland. In

2003, it changed to a 12.5% across the board rate in response to pressure from the

European Union.

[90 ] See Zee et al., supra note 78, at 1504-05.

[91] The conventional position that accelerated depreciation is the only good tax incentive is based on an old rationale, the Harrod-Domar formula. Tax policy must account for the fact that the old asset accumulation theories are no longer mainstream in the developmental economics literature.

[92] SeesupraPartII.B.

[93] Identifying investments on a case-by-case basis allows government officials of the host country to verify whether the investor is eligible for a foreign tax credit in his home country. When the investor is eligible for such a credit, it is in the host country’s interest to impose a tax that is offset by the foreign tax credit.

[94] SeeAndrewGoudie&David Stasavage, Corruption:The Issues,122OECD


[95] Tax holidays are typically considered the worst type of tax incentive, yet they are the most prevalent form found in developing countries. As Professor Bird asked in the context of tax incentives generally, “If everyone is doing it, can it really be wrong?” Richard Bird, Tax Incentives for Investment in Developing Countries, in FISCAL REFORM AND STRUCTURAL CHANGE IN DEVELOPING COUNTRIES 201 (Perry et al. eds., 2000).

[96 ] See Jack Mintz, Tax Holidays and Investment, in FISCAL INCENTIVES FOR

INVESTMENT AND INNOVATION 165, 190 (Anwar Shah ed., 1995).

[97] See Charles McLure, Tax Holidays and Investment Incentives: A Comparative

Analysis, 53 BULL. FOR INT’L FISCAL DOCUMENTATION 326, 329 (1999).

[98] Domestic policymakers often operate under a similar rationale. When they want to provide poor children with free or subsidized milk in schools, for example, the solution is to offer the milk to most if not all schoolchildren. This is wasteful in the sense that the wealthy receive the subsidy as well as the poor. However, targeting poor children exclusively is more expensive because it involves a variety of costs. For an excellent demonstration of the costs of targeting and the advantages of universality see Daniel N. Shaviro, Effective Marginal Tax Rates on Low Income Households, 84 TAX NOTES 1191 (1999).

[99] That is, the difference between income and consumption tax bases.

[100] An essential corollary is preventing individuals from transforming their labor income into corporate retained earnings or from receiving their income as exempt capital gains. One possible solution is a corporate cash flow tax levying zero tax on marginal investment in net present value terms but collecting some tax on economic rents. See Zee et al., supra note 85, at 1500

[101] Shah&Whalley,supranote36,at172.

[102] See Robin Burgess & Nicholas Stern, Taxation and Development, 31 J. ECON. LITERATURE 762, 776 (1993) (“For rich countries . . . personal income taxes, to a substantial extent, replace taxes on corporations.”). Developing countries obtain most of their tax revenues from equal shares of consumption taxes, trade taxes and income taxes (mostly corporate income tax). Developed countries, on the other hand, obtain

36% of their tax revenue from income taxes (mainly on individuals), 29% from consumption taxes and about 29% from social security contributions. Id. at 775.

[103] Other countries such as France and the Netherlands exempt foreign income generated by their residents.

[104 ] Provideditisdone soastoavoidtheapplicationofthehome country’s anti- avoidance provisions, such as CFC rules.

[105] See, e.g., Rosanne Altshuler & T. Scott Newlon, The Effects of U.S. Tax Policy on the Income Repatriation Patterns of U.S. Multinational Corporations, in STUDIES IN INTERNATIONAL TAXATION, supra note 83, at 91 (finding that U.S. parent corporations paid an average U.S. tax rate of 3.4% on their foreign source income).

[106] Despite the prevalence of this type of targeted aid (e.g., food stamps) there is a consensus against it. However, the relative incompetence of developing countries in promoting growth (and possibly the incompetence of any government when compared to the operations of a free market) over the past 50 years leads me to suggest this proposal, which may be termed ‘paternalistic.’

[107] Operatingthroughthe tax system in this case has the same administrative elegance of a negative income tax system or a refundable credit. Instead of imposing a full tax and giving a cash grant, which involves two separate administratively costly transactions, there is only one transaction, a tax break.

[108] Administrativebodies advocatingtaxharmonization,suchastheEU andthe OECD, would welcome the proposal because their main concern is tax competition among developed countries.

[109] These proposalsofferacarrot toMNCs forthe purposeofdirecting FDI to developing countries. The tax system becomes an instrument for increasing the rate of return on investments in developing countries. To the extent that these proposals require cooperation and goodwill from developed countries, similar results are achievable using a stick instead of a carrot. If developed countries agreed to regulate their residents’ research and development activities, some of those activities could be transferred to developing countries. Though this plan would be difficult to execute because it requires the cooperation of all or nearly all developed countries, it could be more effective than the tax avenue.

[110 ] Ireland’s low corporatetax rate isnot consideredharmful tax competition according to EU and OECD definitions because it is not limited to foreign investors.

See OECD REPORT, supra note 50, at 26-28.

[111] Switzerland and Luxemburg understood this fact when they cleverly abstained from voting on, instead of vetoing, the OECD report suggesting the elimination of preferential tax regimes. By abstaining, they did not prevent the adoption of the report by other OECD member countries. The adoption by the other member countries increases the relative attractiveness of these two countries. See OECD REPORT, supra note 50, at 78.

[112] For proposals to establish multilateral tax agreement and a world tax organization, see e.g., VITO TANZI, TAXATION IN AN INTEGRATING WORLD 140 (The Brookings Institution 1995) (“There is no world institution with the responsibility to establish desirable rules for taxation and with enough clout to induce countries to follow those rules. Perhaps the time has come to establish one.”).

[113] Asapreconditiontotheir accession intotheEuropeanUnion, eachofthe10

countries that became full EU members in May 2004 committed to EU tax standards. See Chuck Gnaedinger, 10 Accession Countries Poised for EU Membership, 28 TAX NOTES INT’L 1183 (Dec. 13, 2002); see also Balazs Bekes, EU Pact Raises Questions About Foreign Direct Investment, 29 TAX NOTES INT’L 28, 29 (Dec. 24, 2002) (Hungary was required, as a preliminary condition to its acceptance to the EU, to give up its corporate tax breaks for foreign investment. Investors that may be adversely affected by the change represent approximately 40% of Hungary’s total exports).

[114] See Easson, supra note 80, at 270 (giving examples of IMF pressure on Indonesia, the Philippines, Romania, Tanzania and Uganda); see also Bulgarian News Digest, IMF Nixes Budget Proposals, 28 TAX NOTES INT’L 23 (Oct. 7, 2002) (stating that the IMF did not approve Bulgaria’s proposed zero tax on reinvested profits in its 2003 budget).

[115] See supra Part III.A.3.

[116] The U.S.-Africa Growth and Opportunity Act (AGOA) provides thirty-eight reforming African countries with the most liberal access to the U.S. market available to any country or region with which the United States does not have a free trade agreement. AGOA was signed into law on May 18, 2000 as Title 1 of The Trade and Development Act of 2000, and amended on August 6, 2002 as Sec. 3108 of the Trade Act of 2002. See Trade and Development Act of 2000, 114 Stat. 251 (2000), amended by Trade Act of 2002, 116 Stat. 933 (2002). It is possible to analogize between my proposal and AGOA. AGOA, however, deals with tariffs, which are a well-known barrier to free trade and hence to worldwide growth. Taxes, on the other hand, should not be eliminated as a means to promote growth. Further analysis of this analogy requires an in-depth discussion of the tax competition debate, a topic beyond the scope of this paper.

[117] Thisrequiresdrawing afineline,acomplicatedtask necessitatingadetailed explanation outside the scope of this article.

[118 ] In the last three years, the Thai government has relied heavily on tax incentives to lure many large foreign firms and investment projects to Thailand. The WTO ruling may affect nearly 1,400 such projects, and some of them are likely to move elsewhere. See Jonathan Rickman, WTO Orders Thailand to Phase Out Foreign Investment Tax Breaks, 29 TAX NOTES INT’L 39 (Jan. 6, 2003).

[119] Portfolioinvestmentsare claimsonassets forthe purposeofrealizing some return. The investor does not have control over the firm itself. It is this control aspect that distinguishes portfolio investments from FDI. See GRAHAM & KRUGMAN, supra note 32, at 26-28.

[120 ] See Avi-Yonah, Globalization, supra note 47, at 1585-85.

[121] The tax competition over FDI is not a problem of tax evasion. Countries offer tax incentives to attract foreign direct investments by investors such as MNCs. MNCs do not conceal the existence of this FDI from the tax authorities in their countries of residence.

[122] See Proposal for a Council Directive to Ensurea Minimum of Effective Taxation of Savings Income in the Form of Interest Payments Within the Community, 1998 O.J. (C 212). As of the date of writing, this proposal has not been implemented because it must be accepted unanimously. Luxembourg, Belgium and Austria refuse to approve it as long as non-EU countries such as Switzerland and the United States refuse to take equivalent measures. They fear losing investments to those countries. Switzerland refuses to abandon its banking secrecy, and the United States is unwilling to support the EU directive. See Cordia Scott, White House Signals Lack of Support for EU Savings Tax Directive, 28 TAX NOTES INT’L 421 (Nov. 4, 2002).

[123] This addresses the fear,on the part of EU member states that adoption of the proposal will drive capital away from EU members to other countries. This limitation is not a perfect solution, as investors from EU member countries may still opt to invest in non-EU member countries. Both Luxembourg and the United Kingdom have opposed the directive for this reason. See Conclusions of the ECOFIN Council Meeting on 1 December 1997 Concerning Taxation Policy, 1998 O.J. (C 2); see also Mark Thompson, Luxembourg Says No Agreement Reached on EU Savings Tax, 17

TAX NOTES INT’L 1039 (Oct. 5, 1998).

[124] See Avi-Yonah, Globalization, supra note 47, at 1669.

[125] See Roin, supra note 51, at 596.

[126] Id. at 599.

[127] See The CIA World Factbook 2002, supra note 8 (cross analysis).

[128] See Meredith Coleman, Comment, The Republic of Ireland’s Economic Boom: Can the Emerald Isle Sustain Its Exponential Growth?, 21 U. PA. J. INT’L ECON. L. 833, 851 (2000) (“The favorable corporate tax policy of Ireland and the relative ease with which a company was able paint itself as performing a manufacturing function were the magnets that drew foreign corporations to Ireland.”)




A. The Neoclassical Growth Models

B. The New Growth Theories


A. Standard Arguments Against the Use of Tax Incentives

1. Tax Incentives Distort Behavior

2. Tax Incentives Are Ineffective and Harmful

3. Tax Incentives Should Not Be Used to Compensate for an Unattractive Investment Environment

B. Tax Incentives versus a Low Corporate Income Tax Rate

C. Surrendering theAbilitytoImpose CorporateIncome Tax on Domestic Taxpayers


A. Limits Imposed by International Organizations



Professor, Tel Aviv University - Buchmann Faculty of Law, Israel