Tax Policy in a Global Economy
Paper for the American Institute of Contemporary German Studies
© Peterson Institute for International Economics
The hallmark of a global economy is greater mobility of economic transactions and economic agents. Firms and households enjoy far more geographic choice for purchases of goods and services, for locations to invest, and for places to work and retire. These choices often spill across national borders, as firms and households seek to buy at the lowest price, to sell at the highest price, to invest at the preferred combination of risk and return, and to live in communities that best meet their needs. In the long run, the outcome of greater choice is more efficient economies and better living standards. Along the way, however, are numerous social challenges. Among these challenges is the threat to customary systems of taxation.1
In the "old days"-say before 1970-international commerce was largely confined to merchandise trade; multinational enterprises (MNEs) were a modest part of the world economy; and most individuals consumed and invested at home. In this setting, legislative bodies could set the tax rates on different types of transactions and agents without worrying too much about a disappearing tax base, or the balance between benefits and burdens for different categories of firms and households.
The old days are gone forever. The enhanced mobility we have already seen between 1970 and 2000 only previews greater mobility in the next thirty years.2 The chart below offers a qualitative Summary of mobility past, mobility present, and mobility future, as it affects major elements of the tax structure.
|Tax Base Item||Mobility in 1970||Mobility in 2000||Mobility in 2030|
|Wage & salary income||Low||Low||Moderate|
|Consumption of goods||Low||Moderate||Moderate|
|Consumption of services||Low||Low||Moderate|
The technologies underlying greater mobility are familiar, and require only brief description. Wage and salary income will acquire greater mobility in the next thirty years because any work that can be performed on the computer will in time be capable of remote performance: an individual in Bombay can sell her engineering services in Berlin. Physical migration may well remain tightly controlled, perhaps limited to professionals, family reunification, and some guest workers, but mental migration will be practically unlimited. Electronic commerce (e-commerce) will greatly increase the mobility of goods consumption as shoppers search websites for bargains far away, and as goods are delivered by private shippers such as FedEx and UPS. E-commerce will also enhance the mobility of services consumption as households buy education, entertainment, insurance, legal, and accounting services from distant suppliers. Falling airfares will enable households to spend more of their tourist and health dollars in distant locations. Investment income will become highly mobile as pension funds and brokerage firms develop worldwide networks and households seek to diversify their portfolios. Likewise, corporate profits will become even more mobile as dense intra-firm networks of purchases and sales enable multinational enterprises to shift production and distribution to locations with the highest returns, and maintain their financial accounts with a view to minimizing their taxes. Among the top 100 multinational firms, it is commonplace for more than half of sales, assets, and employees to be located outside the home country.3 The extent of international operations will only grow.
Faced with these challenges, the taxman's first reaction is to devise clever technical fixes, implemented on a national basis, designed to prevent tax base erosion. His next reaction is to think about international tax agreements that might enable member countries to shore up their systems. Further down the taxman's list are responses that would adjust the tax system to the new realities of the international economy. In the sections that follow, I explore these alternatives.
National Technical Fixes
The basic weaknesses of national technical fixes are not their design flaws. Finance ministries and legislative committees are staffed with capable experts who can fashion good technical solutions to the problem of tax base erosion. The basic problems are political and administrative feasibility.
Consider first the problem of political feasibility. Jean-Babtiste Colbert, treasurer to Louis XIV, famously observed, "The art of taxation consists in so plucking the goose as to obtain the largest possible amount of feathers with the smallest possible amount of hissing."4 A modern extension is that an old tax is a good tax, a new tax is a bad tax. Most technical fixes, alas, entail measures that will be perceived as new taxes-after all, they respond to new technologies. And since new taxes are bad taxes, they cause a lot of hissing. A few examples illustrate this basic problem.
In recent years, the notion of a "bit tax" on e-commerce was raised and rejected in Europe. Meanwhile, the U.S. Congress called for a temporary standstill on new state taxes on e-commerce. The Congress created an Advisory Commission on Electronic Commerce, chaired by Governor James S. Gillmore III of Virginia. Governor Gillmore and some of his colleagues now propose a a five-year tax moratorium. During the five-year period, the states would supposedly harmonize and simplify their sales tax systems, which in turn would enable uniform taxation of e-commerce.5 The e-commerce tax debate has historical antecedents in retail sales taxation. Since the 1920s, many states have tried to tax mail order retail sales, with very limited success.
Other examples can be cited. About a decade ago, the Internal Revenue Service proposed a system of withholding at source, to cope with enormous underreporting of interest and dividend income. The proposal never made it through Congress, and instead was replaced by a system of backup withholding for proven tax cheats. Finally, since the 1960s, the U.S. Treasury Department has sought to end "deferral," the compromise under which the active foreign income of U.S.-controlled subsidiary firms incorporated and operating abroad is not taxed until repatriated as dividends to the U.S. parent firm. The U.S. Treasury has made inroads on "deferral", but the goal remains elusive.6
These examples have two common political threads. In some cases, the technical fixes would increase the reporting and tax burdens on a very large number of individuals. In those cases, the agitated opposition of literally millions of taxpayers has doomed the proposed solutions. This outcome would befall any scheme that taxed e-commerce purchases at the household level. In other cases, the technical fixes would put important firms at a severe competitive disadvantage, relative to competing firms not subject to the same tax rules. This was especially true of proposals to end deferral, but it has also doomed efforts at taxing mail order sales within the United States. Telecommunications giants effectively opposed bit tax concepts; in the future, parcel delivery firms will stoutly resist efforts to assess value-added or sales taxes on their cross-border deliveries.
Next consider the problem of administrative feasibility. Extending the tax net to reach mobile transactions depends either on voluntary reporting or the cooperation of foreign tax authorities. (For more on international cooperation, see the next section.) Voluntary reporting in the United States and Germany may be high, compared to Italy or Spain. Nevertheless, "taxpayer morale" is probably declining everywhere, and without the cooperation of foreign tax authorities, more and more transactions will escape the tax net. Again, some examples are illustrative. It is thought that about three million civilian Americans live outside the United States, but only about 300,000 submit tax returns.7 On a worldwide basis, interest, dividends and capital gains are substantially underreported and account for a major portion of the tens of billions of dollars of "errors and omissions" in national balance of payments accounts. Some states have tried to impose "use taxes" on resident purchasers of expensive items (such as autos and electronics) from out-of-state retailers-but use taxes have flopped as a source of revenue.8
This recitation of political and administrative obstacles does not mean that all technical solutions devised at a national level are doomed to fail. But it does mean rising frustration awaits the taxman as mobility comes to characterize wider swaths of the economy.
Limits to International Cooperation
The difficulties of preventing tax erosion through purely national measures will inevitably prompt a search for cooperative solutions at the international level. Sober words of caution are thus in order.
The international problem is not a shortage of technical solutions. Ever since the Europe 1992 program was devised, the European Commission has offered numerous remedies for the tax anomalies between member states. In particular, the EC has tried to narrow national differences in value-added tax rates (by setting a band of 15-25 percent) and thereby solve the thorny problem of border adjustments. The EC has also tried to reconcile different approaches to withholding on interest and dividend income.9 Neither of these EC ventures has been fully successful. In the United States, the Multi-State Tax Commission has tried for decades to create a higher degree of tax cooperation between several states. In a volume published in 1992, Joanna van Rooij and I put forth several recommendations for improving collections on international portfolio income.10 Few of these efforts have done much to cope with the tax realities of a global economy. Why not?
The central reason is that international cooperation on a significant scale confronts a combination of problems arising from zero-sum arithmetic and federalism. To be sure, almost everyone is happy when a notorious tax cheat is nailed. But the big money does not come from a handful of high-profile tax cheats. It comes from thousands, even millions, of transactions that mix avoidance and evasion in varying proportions.
The first point to emphasize has to do with tax convergence between national systems. H. David Rosenbloom rightly emphasizes that there is very little tendency towards tax convergence, much less tax harmonization, even among OECD countries.11 To be sure, at a gross level, there are similarities. Most OECD countries have cut their top marginal corporate rates to the vicinity of 35 percent and their top marginal personal rates to the vicinity of 50 percent. And most (notably excluding the United States) have introduced some form of broad-based consumption tax (a value-added tax or a similar levy), alongside personal income taxes, corporate income taxes, and social security taxes. But similarities progressively vanish the closer one examines the critical details of national tax systems. Countries differ on personal deductions such as mortgage interest and child care, special provisions for retirement savings, taxation of fringe benefits, and so forth. They differ on rates of value-added taxation and social security levies. They differ enormously in the details of corporate taxation: investment tax credits, depreciation allowances, classic vs. integrated systems, territorial vs. worldwide systems, etc. No country has a unique claim to the "right" tax system; in fact, every national system reflects the continually evolving mix of forces in a democratic society. Hence, in the year 2030, national tax systems will likely remain as different from one another as they are today.
In terms of meaningful tax cooperation, these essential differences imply that one country or the other must surrender some important tax advantage when it enforces the system applied by its partner. To be sure, both countries may collect more revenue; but in the process, each country will anger important constituents who are asked to accommodate the tax rules of a foreign power. Examples will illustrate this dilemma.
Ireland has dramatically narrowed its per capita income gap with the United Kingdom and continental Europe since joining the Economic Community in 1973. A key ingredient of Ireland's success has been its system of tax holidays and low tax rates for foreign-owned subsidiaries that establish local operations. Pressure from the European Union on Ireland to adopt the European norms of corporate taxation has been stoutly resisted in Dublin. While certain local subsidies have been challenged by Brussels, Ireland's current 10 percent rate on manufacturing firms, and its plans for an across-the-board corporate rate of 12.5 percent starting in 2003, are both permitted under existing EU rules.12 Similarly, Puerto Rico has long enjoyed tax advantages under the U.S. Internal Revenue Code for pharmaceutical and other high-tech assembly operations. San Juan has vigorously, if not always successfully, resisted attempts by the U.S. Treasury to narrow these advantages. Banks in London and Luxembourg thrive on managing nonresident funds, free from the tax scrutiny of other European countries. Accordingly they resist any European-wide reporting or withholding system.13
As a technical matter it may be possible to devise revenue-neutral solutions to national differences of the kind illustrated-i.e., solutions that leave all national finance ministries equally well off (or even better off). But revenue-neutral solutions will still create losing constituencies, and in a democratic system losing constituencies can often block proposed changes in the tax law.
The other structural reason why tax convergence will remain a distant goal finds its basis in the federal structure of the United States, Germany, the European Union itself, as well as important countries such as Australia, Brazil, Canada, and India. Briefly, federal governments find it difficult or impossible to agree at an international level on tax rules fashioned to constrain their states or provinces. Meanwhile, subfederal units are playing a larger role in economic life.14 Often the result is subfederal incentive schemes, operating through tax relief and investment subsidies, designed to attract national and multinational firms.
The response rate to these incentives may be quite high. Between countries, the response coefficient may have increased from about 1.5 in the 1980s to 3.0 in the 1990s. A response coefficient of 3.0 implies that a one percentage point decrease in the effective corporate tax rate (or an equivalent increase in the effective subsidy may induce three percent more investment than would have otherwise occurred.15 Within the United States, response coefficients may be as high as ten.16 In other words, response coefficients are high and probably rising. Hence it is difficult for national finance ministries to lecture their state and provincial colleagues over the "foolishness" of tax and subsidy "giveaways." It is even more difficult for national finance ministries to curtail the economic powers of subfederal units through international agreements.
Scope of International Cooperation
While these cautions severely limit the scope of practical international tax cooperation, they leave room for modest initiatives to cope with the challenges of a global economy. In the remaining sections of this paper, I outline what I regard as the outer limits of feasible tax cooperation between the United States and the European Union in the next decade.
E-commerce: Catalyst for Reform
One of the great revolutions of our time is the explosion of e-commerce. With e-commerce, households as well as firms can buy goods and services from a much larger field of suppliers than otherwise. The result will be an approximation to the model of perfect competition for goods and services ranging from wine and shoes to electricity to entertainment and education. Widening the market will bring substantial efficiencies in terms of larger sales by low cost suppliers; it will bring even greater efficiencies in terms of the more rapid diffusion of technologies for new products, new production methods, and better distribution systems. In order to realize these gains, national tax systems will need to accommodate cross-border production of goods and services sold direct to households. For goods, this implies a huge volume of small parcels; for services it means an enormous expansion of internet deliveries.
If tax systems were being designed on a blank slate, the most practical way to accommodate e-commerce would entail adoption of the origin principle of border tax adjustments for retail sales, value-added and similar consumption taxes. Under the origin principle, taxes are not imposed on imports of goods and services, but they are imposed on exports. From an administrative standpoint, it is much easier to collect taxes on firms at the location of production than on households at the place of delivery. This is surely true for small parcels arriving in large volumes in the customs sheds of the world. It is even more true for services arriving by internet at the computers of millions of individual buyers.
However, our tax systems were designed long before anyone dreamed of the internet or e-commerce. Tax systems have enormous inertia, because they decisively affect capital values, location decisions, and political interests. E-commerce will adapt to inherited tax rules more than the other way around. Nonetheless, e-commerce can provide an important catalyst for reform.
Under longstanding GATT and now WTO rules, sales, value-added and similar taxes are typically adjusted at the border according to the destination principle: they are imposed on imports but not on exports (the exact reverse of the origin principle). In general equilibrium analysis, with flexible exchange rates, it makes no difference to a nation's current account balance whether it adopts the origin or destination principle for border tax adjustments. But the fact that two principles have an equivalent effect on the current account in general equilibrium does not mean they have the same effect on individual economic sectors, nor does it mean that switching from one rule to another can be done without significant transition costs.17 Firms and sectors that are highly taxed will vociferously object if an origin tax system is adopted and their foreign-based competitors can sell in the domestic market without paying comparable taxes. And the same firms and sectors will strongly object if they have to pay high domestic taxes on their export sales.
In short, the change from existing border tax adjustments under the destination principle to a system of no adjustments under the origin principle would be politically unpopular. As a practical matter, e-commerce will have to adapt to the destination principle, rather than the other way around. However, for the destination principle to operate in a world of e-commerce, important administrative simplifications need to be adopted. These simplifications can act as a catalyst for reforms that have been too long delayed in the manner the current destination rules discriminate between corporate taxes and value added taxes.
- First, for small parcel goods sales made directly to households (e.g., with values up to $1,000), the United States and Europe should reach an agreement on a standard tax rate to be imposed in lieu of all retail sales taxes, value added taxes, and customs duties. This tax would be paid by the seller and deposited into a clearinghouse, with receipts distributed by formula according to the final destination of sales.18 In other words, the seller and the clearinghouse-not UPS, FedEx or the customs authorities-would be responsible for the administrative burden. To carry out the spirit of the destination system, the selling firm should be exempt from local sales taxes and value added taxes.
- Second, for service sales made directly to households over the internet (e.g., entertainment, financial, education, legal, accounting, etc.), the United States and Europe should likewise reach an agreement that the sellers and the clearinghouse are responsible for tax administration. In the United States, it is rare that services are taxed. About a decade ago, an attempt by Florida to tax legal, accounting a few other services met with outrage, and was quickly dropped. By contrast, in Europe, services are commonly subject to value added taxation. Hence, under the system outlined, U.S. firms selling services via e-commerce to European households would generally collect taxes (at agreed standard rates), but European firms selling services to American households would seldom collect taxes. A tax imbalance would result in favor of Europe. In mercantilist terms, this imbalance would be seen as a U.S. "concession" to Europe.
- Third, the European negotiating "quid" for the U.S. "quo" on e-services sold households would be modification of the existing GATT and GATS destination rules on business purchases of goods and services from providers in the other country.19 This is where e-commerce can act as a catalyst for tax reform. The new basic rule would disallow deductions both for corporate income tax purposes and for value added tax purposes of purchases of goods and services from suppliers based in the other area. Disallowance of deductions would effectively implement the destination principle for business purchases of goods and services. Purchasing firm would bear the tax burden of their imports. This already happens with value added taxes, either through the mechanism of tax collection at the border by the customs authorities, or by disallowing deductions. Under current GATT and GATS rules, no comparable destination principle border adjustment is allowed for corporate taxes. The proposal made here would put corporate taxes and value added taxes on the same footing. To complement this change, all sales by business firms to foreign purchasers (whether households or business) would be excluded from the corporate tax base-just as they are now excluded from the value added tax base.
If the United States and Europe could reach agreement along these lines, other members of the World Trade Organization would probably accept the new framework. Eventually, in twenty years, these destination rules might give way to an origin-based system. The shift would be relatively easy for sales made direct to households in the other region. Through reciprocal agreements, sector-by-sector, selling firms would no longer collect taxes on direct household purchases for payment to the other region. The shift would be harder for sales made to business firms in the other region. The reason is that the combination of value added and corporate taxes in Europe are much higher (as a percent of sales) than corporate taxes in the United States. In mercantilist terms-looking only at the short-run and ignoring general equilibrium price and exchange rate effects-an origin system would appear to disadvantage European firms. For this reason, the transition to an origin system for business purchases and sales might await other developments that led to a closer convergence of combined tax rates.
Personal Income Taxes
In the global economy of 2030, households will derive more of their income from sources outside their country of citizenship or residence. Interest, dividends and capital gains from abroad are already important and will become more so as pension funds continue to diversify their portfolios internationally.20 In addition, more individuals will derive service income (accounting, legal, engineering, design, consulting, etc.) from foreign clients, and they will earn rents on foreign properties. Finally, with the aging of Americans and Europeans, many more people will choose to spend their retirement years abroad, while drawing social security and other pension benefits from the country where they once worked. These incidents of mobility all play havoc with personal income tax systems-especially the U.S. system which claims to tax the worldwide income of all citizens, regardless where they reside. (In fact, the United States even claims to tax the income and estates of ex-citizens, for a period of ten years after they renounce their U.S. passports.21)
Between the United States and the European Union, it might be possible to bring a degree of order to the tax chaos by agreements that accomplished three objectives. The first objective is to devise a system of comprehensive reporting of interest, dividends, capital gains, rents and personal service income paid to residents of the other country. The system would need to be reinforced by backup withholding taxes on payments outside the net of treaty countries; otherwise fraudulent addresses would explode. The first step is to create a system of reporting and backup withholding taxes within Europe; and, as already noted, this initial step faces stiff resistance from Luxembourg and the U.K. banks. The second step is to negotiate an EU-U.S. accord; and, since member states retain competence over most tax matters, they would first need to agree on a common line among themselves.
The second objective is to reach an EU-U.S. agreement that personal income taxes will be levied on the basis of residence, not citizenship. Residence would be determined by a factual test (e.g., country of citizenship unless the person spent more than 270 days in the tax year in the other country, meaning, in the case of Europe, a single member state). In other words, the great majority of expatriates would pay personal income tax to only one country, and there would be no withholding tax or foreign tax credits for income derived from sources in the other country. To reduce "gaming" between tax rules and health care provisions, residents would be eligible for publicly supported health care (e.g., Medicare) only in the residence country-the same place they pay personal income taxes. Again, an accord with these various dimensions presupposes that the member states can reach a common line among themselves-not an easy task.
Corporate Taxes on Multinational Enterprises
Most international tax law concerns the activities of firms, especially multinational enterprises (MNE) with operations in multiple countries. Multinational firms take investment decisions that affect hundreds of billions of dollars and shape millions of lives. Many MNEs derive more than half their earnings from sources outside the home country. Not surprisingly, given the importance of MNEs and the density of their transactions, today's international tax regimes are deeply entrenched both in the law and the politics of OECD countries. The regimes will not be easily altered.
Elsewhere, I have argued that the United States would do itself a national favor by adopting the key elements of territorial taxation.22 The National Foreign Trade Council recently issued a detailed report that reached similar conclusions.23 In practice, if not in theory, most European countries practice territorial taxation of their multinational enterprises-through a combination of statutes, treaties, and flexible administration of tax rules.
The territorial approach would focus attention on measuring and taxing corporate income earned in the United States, rather than the worldwide income earned by U.S. corporations and their controlled subsidiaries. Recent analysis shows that the United States would actually gain about $5 billion revenue annually by exempting the foreign source dividends of manufacturing firms from the U.S. tax net.24 This revenue gain could be used to extend the territorial system to other kinds of active foreign source income. But however strong the logic of reform, there is little chance that the U.S. Congress will rationalize the antiquated U.S. system of worldwide corporate taxation without "something else" in the bargain.
The "something else" was furnished by the December 1999 breakdown of WTO talks in Seattle.25 An important cause of the breakdown was bitter disagreement between the United States and nearly all developing countries over the inclusion of labor and environmental standards in the Millennium Round of WTO talks. Developing countries hold well-grounded fears that, once these subjects become part of WTO negotiations, the next step will be the design of formidable trade barriers modeled after the U.S. and EU anti-dumping statutes. Exports from developing countries would, they fear, be severely penalized.
As one alternative to the pursuit of labor and environmental trade sanctions in the WTO, the United States and Europe could broadly, but conditionally, exempt the foreign income of their MNEs from domestic taxation. The exemption would be conditioned on certification that the U.S. or EU parent corporation, all its foreign subsidiaries, and its principal suppliers all exceed core labor and environmental standards in their operations. Independent commissions (not national tax authorities) would set certification standards.26
Under this approach, when the exemption applied, each country would tax corporate income earned at home, but would exempt the income earned by subsidiary firms incorporated and operating abroad.
Inherent in this recommendation is the fundamental idea that each country would tolerate (within agreed limits) incentives that other countries or their subfederal units might offer to attract investment and jobs-provided they respect core labor and environmental standards. To ensure this degree of toleration, the European Union, its member states, and the United States (and its fifty states) would need to reach agreement on much tighter rules setting maximum subsidy levels than now exist within the WTO. If the United States and Europe can reach agreement on permissible subsidies, it should be possible to draw other adherents in the context of the Millennium Round. However, with this suggestion I am near the limits of practical politics-if not completely over the boundary.
Conclusion: Adjusting Public Finance to Private Mobility
Taxes are compulsory payments to government for which the taxpayer receives no specific benefit. But the taxpayer is entitled to the general benefits of governance; and if the taxpayer is dissatisfied with the balance between public benefits and tax burdens, he may consider moving to another jurisdiction-another city, another state, another country. This fundamental observation was made in 1956 by Charles Tiebout.27 Forty years later, the costs of moving are rapidly falling and the opportunities for moving are greatly increasing, both for firms and households. The Tiebout Hypothesis applies not only locally but also globally.28 Jurisdictions that want to avoid a shrinking tax base and dwindling revenues must accordingly deliver quality government services at competitive costs. They must also pay more attention to the benefit/burden calculation for groups of firms and households, not just for society as a whole.
Let me conclude this essay with brief comments on the benefit/burden calculation. In the decades ahead, enhanced mobility means that most jurisdictions will be left with fewer "cash cows"-individuals or firms that can be taxed heavily without prompting their move to greener pastures. Instead, tax systems will necessarily be redesigned to attract mobile individuals with desirable skills and mobile firms that offer good jobs. Public benefits will need to show up in quality education and other public services that appeal to skilled employees. Otherwise, the individuals, the firms and the jobs will move, both physically and via the internet.
These benefit/burden considerations will compel a number of micro adjustments in tax systems and public expenditure profiles. However, they will also force a major social challenge. Europe and the U.S. will find it increasingly difficult to extract large sums of money from the working population to pay generous social security and health benefits to vast numbers of retired citizens. In the decades ahead, the core redistribution component of current tax and expenditure systems-redistribution between working and retired citizens-will be severely challenged by the realities of the global economy.
Peter G. Peterson explores this theme in a new book.29 In the year 2030, about 35 percent of Americans over the age of 15 will in fact be older than 65; the comparable proportion in Europe will be 40 to 45 percent. The large cohorts of politically powerful elder citizens will stress the fiscal systems of OECD nations. Until Europe and the United States convert their current pay-as-you-go social security and old-age health systems into fully funded plans (where each age cohort pays for its own retirement and health benefits), working people will find themselves paying stiff taxes to care for their elders. In the meantime, a combination of three unpleasant outcomes will emerge: (1) working Americans and Europeans must come to accept a significant drop in their after-tax incomes; and (2) public old-age benefits must be curtailed; or (3) firms and jobs will migrate to younger countries. Over the next two decades, these unpleasant realities will become the dominant challenge of public finance on both sides of the Atlantic.
1. For a perceptive analysis of trends, see Vito Tanzi, "Globalization, Tax Competition and the Future of Tax Systems", in Gerold Krause-Junk, editor, Steuersysteme der Zukunft, Dunker & Humblot, Berlin, 1996.
2. The impact of globalization on tax structure has caught the attention of Vito Tanzi and his colleagues at the IMF Fiscal Affairs Division. See IMF Survey, July 6, 1998. Also see The Economist, "A Survey of Globalisation and Tax", January 29, 2000.
13. For an explanation, see Edmond Israel Foundation, The Proposed Directive on Harmonisation of Taxation of Interest Payments, Luxembourg, 1999. For a spirited denuciation of the proposed withholding taxes, see The Economist, November 27, 1999, p. 17.
16. For a review of the literature (especially the pioneering work of James R. Hines and Dean Kemsley, see Gary C. Hufbauer and Dean A. DeRosa, ACosts and Benefits of the Export Source Rule," Tax Notes International, vol. 14, no. 20, May 19, 1997.
18. A similar system needs to be established for the collection of retail sales taxes within the United States. See Mines, Op. Cit., and Henry J. Aaron, "The Internet Tax Dodge", Washington Post, December 3, 1999, p. A41. For the opposing view, see Wall Street Journal, "The E-Grinch", November 29, 1999, p. A28.
19. The GATT stands for General Agreement on Tariffs and Trade. The original GATT (GATT-1947) was part of the postwar Bretton Woods system. As a result of the Uruguay Round of Multilateral Trade Negotiations, GATT-1947 was replaced by GATT-1994, one of the components of the new World Trade Organization (WTO). Another component is the new GATS, General Agreement on Trade in Services. For an analysis of the tax rules in the GATT and GATS, see Hufbauer and Gabyzon, Op. Cit., 1996.
20. According to the OECD, The World in 2020: Towards a New Global Age, Paris, 1997, p. 87, the pension funds in major OECD countries now invest between 5 and 35 percent of their assets internationally; these proportions could double in the next twenty years.
25. Gary Clyde Hufbauer, "World Trade After Seattle: Implications for the United States", Washington DC: Institute for International Economics, International Economics Policy Briefs Number 99-10, December 1999.