The economies of the United States and Canada are closely linked-trade between the two countries is substantial, their capital markets are highly integrated, and even movement of individual workers between the two countries is non eligible. The U.S. and Canada have now agreed to eliminate all remaining tariff barriers between the two countries during the next few years. To what degree does this increasing economic integration create pressure on the two countries to change their tax systems? Which aspects of the systems will be most affected? Will their tax systems inevitably become more alike, as each country finds it in its economic interest to choose tax provisions resembling those in the other country? When will explicit coordination and harmonization of tax provisions be called for? Addressing these questions is the objective of this paper.
The effects of mobility of goods and factors between jurisdictions on their fiscal systems has been explored at length in the local public finance literature, stimulated by Tiebout (1956). These models assume that everything and everyone is mobile without cost-implicitly, even community boundaries can adjust. The basic conclusion of this literature is that competition among communities drives the tax system toward one in which each individual’s or firm’s tax payment closely matches the cost of the services received from the community (or the costs imposed on the community, from pollution or congestion for example), a tax system known in the literature as benefit taxation. With any other tax structure, decisions in one community clearly affect welfare in other communities, creating the potential for mutually beneficial coordination of fiscal policies. 
The existing tax systems in the U.S. and Canada differ substantially from A benefit-tax structure. To what extent can these differences continue to survive, given the increasing openness of the two economies? Of course, economic mobility between the U.S. and Canada, while substantial, is hardly costless. In this paper, I attempt to assess the pressures created by current and prospective levels of mobility between the two countries. In section 2.1, the implications of existing levels of capital mobility are examined: What implications does capital mobility alone have on domestic tax policies and the need for tax harmonization? Section 2.2 examines the further pressures created by free trade in the full range of outputs. Finally, in section 2.3, the pressures created by labor mobility, to the extent that it exists between the U.S. and Canada, are assessed.
There are three key conclusions of the discussion. First, taxes on capital income are unlikely to survive for long, even under existing levels of international capital mobility, without explicit coordination of capital income tax policies among all major countries; an agreement between just the U.S. and Canada would accomplish little. While existing double-taxation conventions may have led in the past to implicit coordination of capital income taxes, the in creasing complexity of international capital markets will make any coordination in the future much more difficult. Second, unrestricted trade between the two countries will force them either to “level the playing field,” by eliminating any tax or regulatory distortions to the relative prices of traded goods, or else to agree on a common set of tax distortions (e.g., agricultural price supports) with a common set of trade barriers to support these internal price distortions. Finally, to the extent that labor mobility is allowed, redistribution through the tax system becomes more difficult. Under existing levels of mobility, however, this pressure is not yet very important.
Inevitably, the analysis is somewhat abstract, attempting to forecast the broad direction of change in the tax structures in each country. Boadway and Bruce (ch. 1 in this volume) examine in much more detail how these pressures are currently being felt and the likely short-term responses to them.
2.1 Tax Implications of Capital Mobility
How does capital mobility affect the design of tax policy? To explore this question, assume for simplicity that only one good is traded among countries. Trade therefore simply takes the form of some of this good being imported now, in return for an acceptable amount of this good being exported back as return payment in a later period. To shorten the discussion, I will ignore the implications of risk or inflation.
Without taxes, capital would flow between countries until the rate of return from investing in each country is the same. Let ij represent the rate of re turn on asset a in country j. Without taxes and uncertainty, the return on all assets would be equalized in equilibrium, so that ij = it for any asset a in country j and asset b in country k. Given this, investors would be indifferent between investing in domestic or foreign capital and between investing in different types of financial securities.
The equilibrium ownership structure of securities, and the equilibrium al location of capital, can be affected in many ways by taxes. The existing tax treatment of capital income is quite complex. To begin with, a corporation’s income is directly subject to tax in the country in which it is located, under the corporate tax. If the owners of a corporation reside in the same country, then they are taxed as well on the income they receive from the investment under the personal income tax. For foreign owners of the firm, however, the tax treatment is more complicated. Payments may first be subject to a with holding tax in the source country. If the owner is an individual, the pre-with holding-tax income is then taxable in the home country, but with a credit for any withholding tax. If the owner is a corporation, the pre-corporate-tax in come underlying the payments is subject to tax in the home country, but with a credit for any corporate income and withholding taxes already paid on this income. Finally, payouts to the ultimate individual owners are also taxed.
What pressures does capital mobility create, given the existing tax system?
To simplify the discussion, I will initially assume that income from capital is subject to corporate taxation only in the source country, and that income to individuals from capital is taxable only in the country where the individuals reside. In effect, these assumptions ignore withholding taxes and corporate surtaxes on repatriated income. The discussion will start by examining the implications of capital mobility for residence-based taxes, such as the personal income tax, then will turn to source-based taxes, such as the corporate tax. Finally, the discussion will return to explore the implications of capital mobility for withholding taxes and corporate surtaxes and to explore the implications of Canada’s dividend-credit scheme.
2.1.1 Capital Taxation under the Residence Principle
In principle, under a residence-based tax each country taxes the capital in come of its own residents at accrual, regardless of where this income is earned, but does not tax the income of nonresidents, even when they invest in local securities or in local real capital. Let the effective tax rate for residents of country j on income from asset a in country k be tajk Then, equilibrium for investors residing in country j requires that iaj(1 – tajj) = ibk(1 – tbjk),,while equilibrium for investors residing in country k requires that iaj(1 – tbjk), As emphasized in Slemrod (1988), these two equilibrium conditions cannot hold simultaneously unless for all assets a and b. Given equation (l), investors will again be indifferent between investing in any of the available financial securities. Firms will then seek the cheapest form of financing, given the resulting pretax rates of return on different financial securities.
If equation (I) does not hold for all asset pairs, however, then tax arbitrage possibilities exist enabling investors to rearrange their portfolio holdings to reduce tax payments. Each investor has the incentive to reduce his holdings of assets that are taxed relatively heavily in his country and increase his holdings of assets that are taxed relatively lightly. In the process, investors save on taxes. If investors can own negative amounts of some assets and can deduct the required payments, then this rearrangement of portfolios can in principle continue without limit though risk considerations presumably limit the ex tent of this arbitrage.
Of course, similar arbitrage possibilities can arise even in a closed economy.” In fact, Gordon and Slemrod (1988) found that in 1983 in the U.S., as a result of such arbitrage, the attempt to tax the return to saving and investment resulted in a slight net loss in tax revenue; interest deductions more than offset the taxable income generated by both real and financial investments. Countries in practice seem to recognize arbitrage opportunities gradually, and then attempt to eliminate them case by case. For example, in the U.S. individuals are not allowed to deduct interest when they borrow for the purpose of buying a tax-exempt bond. But this is just an example of a wide variety of possible forms of tax arbitrage, and enforcement of even this restriction is very difficult. Under the 1986 tax reform in the U.S., a broader attempt was made to limit arbitrage possibilities by restricting interest deductions, except for businesses, and restricting taxpayers’ ability to deduct losses more generally. When this arbitrage takes place across borders, detecting and dealing with it is that much more difficult.
If investors in each country can “go short” in the appropriate asset, then capital income taxation collects significant revenue only if these arbitrage possibilities are closed off, which requires that equation (1) be satisfied for all pairs of assets. An agreement between the two countries on relative tax rates could occur implicitly as well as explicitly. Neither country would want to deviate from a common set of relative tax rates, since doing so would open up arbitrage opportunities for investors in both countries-any set of relative tax rates would be a Nash equilibrium. However, both countries may gain by jointly agreeing on a particular set of relative tax rates. The normal presumption has been that a “neutral” tax system, under which income from all assets is taxed at the same rate, is the most attractive.
If no deductions are allowed for payments on debt or other “short” positions, then a country would never lose revenue from taxing capital income, even without agreeing with the other country on the relative tax rates on different assets. Equilibrium portfolio holdings in each country would still depend on the tax policies in both countries, however, making welfare in the two countries interdependent. Coordination of relative tax rates would still in principle be justified. However, as shown formally in Gordon (1986), each country acting in isolation would have the incentive to set its tax rates so that its residents invest in the security paying the largest amount pretax. This is accomplished simply by equating the tax rates on all assets, so that tajj; = tbjk regardless of the tax policy chosen in the other country. Therefore, a “neutral” tax system may well be the Nash equilibrium as well as the optimal policy chosen after full coordination.
Given any agreement on relative tax rates on different assets, each country could then choose independently the absolute level of its tax rates without opening up arbitrage opportunities. Each country’s policies affect the welfare in the other country only through any resulting changes in the market interest rate. If each country is small relative to the world capital market, then these changes will be small, implying no important externalities when choosing the absolute level of residence-based capital income tax rates. The U.S., however, is not plausibly small relative to the world capital market. It has the incentive to reduce its borrowing from foreigners in order to reduce the market interest rate, thereby reducing the interest payments on its existing debt. Given that Canada is a net debtor in the world capital market, a reduction in the market interest rate would be a benefit for Canada, a benefit ignored by the U.S. in designing its own policies.  In particular, under optimal policies the U.S. would be indifferent to borrowing still less, but Canada would gain from the resulting fall in the interest rate. This creates the potential for mutually beneficial agreements on tax policy.
Use of a residence principle for capital income taxation leads to a major problem with tax enforcement, however. Within a country, firms and institutions that pay dividends and interest can be required to report the names of the recipients, and how much they receive, to the local tax authorities. A country has no direct way to require foreign firms and institutions to make such re ports. But if the tax authority receives no information directly about the capital income received by its residents from foreign sources, then it will find it extremely difficult to enforce the taxes due on this income. Reporting income from assets owned abroad in effect becomes voluntary, and normally investors do not knowingly make voluntary tax payments. If in practice savings in vested abroad are tax free, then all savings become tax free, because investors can invest through a foreign financial intermediary in all assets,including domestic assets. In fact, they may be able to borrow domestically, deduct the interest, then invest the funds abroad tax free.
Can this enforcement problem be solved through suitable cooperation between the two countries? The countries could, for example, agree to share information provided by firms and institutions regarding the names of recipients of capital income. Any such agreement would allow each country to tax the capital income of its residents and so would appear to be mutually beneficial. However, given the disparity in the sizes of the U.S. and Canada, Canada might have an incentive to refuse to cooperate. Without the agreement, the relatively huge number of U.S. investors could flock to Canada hoping thereby to evade U.S. taxes. The resulting gains to the Canadian economy, whether or not the gains were taxed, might well more than offset the losses to Canada from not being able to tax the capital income of Canadian residents. If so, the U.S. would need to compensate Canada in order to secure any such agreement.
Such an agreement would be futile, in any case, given that a third country (e.g., Switzerland) could agree to facilitate the tax evasion of U.S. or Canadian investors. Such a country could open its own financial intermediaries to foreign depositors and refuse to share information with other countries. The income to foreign investors working through these financial intermediaries would again be exempt in practice from residence-based taxes. This third country could tax away some of the gain that investors receive from evading their domestic taxes and still attract funds. By refusing to cooperate with other countries, it might not be able to tax the capital income of its own residents, but if the country were small enough that would be a minor consideration.
Taxation at Repatriation
Even if a country cannot independently detect capital income earned abroad by its residents, it may be able to detect income as it is repatriated, through monitoring all deposits in domestic financial intermediaries or through auditing individuals whose expenditures clearly exceed their cash flow. What hap pens if a country simply taxes capital income at repatriation? If repatriated income is taxed at the same rate, regardless of the date of repatriation, then the effective tax rate on capital income is reduced the longer repatriation is postponed; if repatriation can be postponed indefinitely, then the effective tax rate goes to zero. Economic repatriation may even be possible without triggering the repatriation tax. For example, the investor may be able to borrow at home, possibly using the foreign assets as collateral. The borrowed funds could be used to finance any desired expenditures at home and in fact could lead to further tax savings through interest deductions. The U.S. tax law has evolved over time, trying to close off such devices for avoiding the repatriation tax, but doing so is very difficult.
2.1.2 Capital Taxation under the Source Principle
Under a source-based tax, each country would tax the return to real capital located within its borders, with rates perhaps varying by type of real capital. 
If in country j the returns to asset a are taxed at rate ti, then in equilibrium iaj( 1 – taj) = ibk( 1 – tbk) for all assets a and b. Since this condition is the same for investors in each country, allowing for capital mobility does not create additional complications when characterizing the equilibrium.
What can be said about the optimal source-based capital income tax rate? Diamond and Mirrlees (1971) argued that when all excise taxes can be used flexibly, and when there are no pure profits, then the optimal tax system will lead to efficient production. In particular, if a country is a price taker in the world capital market, then efficient production means that investment occurs until the marginal rate of return equals that prevailing on the world market. Therefore, in such a setting, the optimal source-based capital income tax rate should be zero. The intuition underlying this result is very simple: In a small open economy, a source-based tax on capital cannot be borne by capital,since capital owners will not invest in the country unless they earn the same return as they earn elsewhere. Therefore, the tax ultimately must be paid by immobile factors, presumably land and labor. But in that case, a direct taxon these factors would dominate, since it would have the same incidence yet not distort the international flow of capital.
What if firms can earn a rate of return above the world rate?
Within a closed economy, a tax on pure profits, as occurs under a cash-flow tax, does not distort allocations, and so is attractive on efficiency grounds. In an open economy, however, pure profits may also be mobile. For example, if the profits are tied to technology rather than to location, then the firm will locate production based on economic conditions in the available countries. An open economy would then be able to extract rents from the firm only to the extent to which the country provides locational advantages greater than exist else where. A small country presumably provides at best small advantages; the implication, based on the same reasoning as before, is that the optimal tax on these pure profits is close to zero. 
The Diamond-Mirrlees argument also implies, however, that a large open economy will wish to equate the domestic marginal product of capital to the marginal cost to the country of extra funds on the world market. As a result, a country such as the U.S., which is large relative to the world capital market, has the incentive to take advantage of this market power by restricting net capital flows. Given that the U.S. has recently been a net borrower in the world capital market, this would imply taxing investment in order to reduce net borrowing. Before the 1980s, when the U.S. was a capital exporter, the incentives would instead have been to subsidize investment to restrict capital exports. Canada does not plausibly have market power in the world capital markets, so should not attempt to change investment incentives. 
Except as a means to take advantage of monopoly power, are there any other ways of explaining the continued though relatively minor role of corporate taxes in the U.S. and Canada? One traditional rationalization for the corporate income tax is that it is necessary to prevent wholesale avoidance of a residence-based tax on equity income, given the favorable treatment of accruing capital gains under existing tax systems. This argument is appropriate only in a closed economy, however, where domestic shareholders can be taxed indirectly on their accruing capital gains through imposing a corporate tax on domestic corporations. In a small open economy, the rate of return earned by domestic residents on their savings, before personal taxes, is set by the world market and so is unchanged by a domestic source-based tax.
A related argument is that the corporate tax prevents avoidance of the domestic tax on labor income, at least in closely held corporations. Without the corporate tax, shareholder-employees in such firms have the incentive to leave their labor earnings in the firm, thereby allowing their shares to increase in value. When they need cash, they can simply sell some of their shares in the firm, paying tax on the accumulated gains at the more favorable capital gains rate. A cash-flow tax on corporate income at the same rate as the labor income tax would eliminate this opportunity, though it might discourage firms earning pure profits from locating in the country. A better alternative, at least in theory, would be to shift from a labor income tax to a consumption tax. Given the appropriate treatment of bequests, both have the same lifetime incidence, but the consumption tax is not vulnerable to the above evasion strategy.
Certainly if a firm imposes costs on the public sector through use of public services and facilities, then user fees would be appropriate, even in the Diamond-Mirrlees setting. It is difficult to justify a tax on capital income based on this reasoning, however.
These arguments together suggest that at least a small open economy should not make use of source-based capital income taxes. However, source based capital income taxation in one country imposes clear externalities on other countries, suggesting that countries may gain by jointly agreeing to use source-based taxes. In particular, when one country raises its source-based capital income tax, capital flows to other countries, raising wage rates in the other countries and raising tax revenues if these countries also use source based capital income taxes. In fact, a uniform capital income tax at source is equivalent to a uniform tax based on residence. While a residence-based tax is very difficult to enforce, given the government’s lack of independent information about capital income earned abroad by its residents, enforcement of a source-based tax in theory should be much easier, since any activity within the country can be monitored by the tax authorities. Therefore, countries may well find it attractive to jointly tax capital income at source as a means of taxing indirectly the capital income earned by their residents. The U.S.and Canada together, however, are not much larger relative to the world capital market than the U.S. alone, suggesting that the room for Pareto-improving gains between these two countries alone may be quite limited.
Furthermore, source-based taxation, at least of multinational firms, has its own enforcement problems. There are many ways in which a multinational can shift accounting profits toward the country with the lowest statutory tax rate, even without changing the location of real activity. The easiest approach is probably through manipulation of the transfer prices assigned to goods and services moving between firms within the multinational. Similarly, the multi national can locate patents for new technology in the country with the lowest tax rate. Yet another approach is to do the bulk of the debt financing for the multinational in the country with the highest tax rate, using perhaps as collateral the assets located in other countries. Governments have little ability to monitor the diverse nature of transactions within a firm and can effectively challenge only a small fraction of these schemes.
Given that multinationals can quickly and easily shift taxable income to ward those countries with the lowest statutory tax rate, each country has a strong incentive to cut its statutory tax rate in order to benefit from this process.Tax competition then drives statutory tax rates towards zero, even if the location of real activity is not very sensitive to relative tax rates.
The above discussion of optimal tax policy assumes that capital is fully mobile in response to differences in rates of return. Once capital is invested in a country, however, it is difficult to move even in response to high tax rates. Therefore, while the amount of new investment may be very sensitive to tax rates, the amount of existing capital may be virtually fixed. As a result, at any date a country has an incentive to seize any existing capital but then to promise never to do so again, in order not to discourage new investment.  Assuming it could make such a binding promise, then by the above arguments it would choose never to tax new investment. But governments have no way to pre commit their future tax policy. If no commitment has been made, then once new investment occurs and the capital has become immobile, the country again has the incentive to seize the capital. This is known as the”time consistency” problem. Perhaps reputation effects inhibit even the initial seizure of capital. Alternatively, the country can subsidize initial investments to compensate for the taxes that inevitably will be collected from these investments at a later date, regardless of what may have been promised.
2.1.3 Capital Income Taxation under the Current Law
So far, we have ignored the incentives created by existing double-taxation conventions. Given these conventions, how does the forecasted behavior of each government change? What joint tax structure would be forecasted to arise? Comparing this tax structure with that which arises without this convention, can we explain why countries choose to adopt it? Existing double taxation conventions affect the taxation in the home country of both portfolio income earned abroad by domestic investors and corporate income repatriated from foreign subsidiaries by a domestically based multinational. In each case, the home country allows a tax credit for particular taxes paid abroad,whereas our previous discussion assumed that foreign tax payments were deductible.
One further complication ignored in the previous discussion was the use of a dividend-credit scheme in Canada. The incentives created by this scheme are complicated enough that they merit a separate discussion.
Withholding Taxes on Portfolio Income
Let us begin by examining the equilibrium use of withholding taxes on portfolio income earned by foreign investors. Existing double-taxation conventions allow a tax credit in the home country for withholding taxes paid on portfolio income accruing in the host country, with a maximum credit equal to the taxes due on the income in the home country. If tax treaties did not also specify the rate of withholding tax, how would each country respond?
Assume first that tax evasion is not a problem, so that each country can effectively tax income earned by domestic residents from foreign portfolio holdings. Consider first the incentives faced by a small host country. Since a withholding tax does not affect the net-of-tax earnings of foreign investors, as long as the withholding tax rate remains below the domestic tax rate faced by these foreign investors on their portfolio income, the tax produces revenue without any loss to domestic residents. Therefore, the host country should choose to raise this tax rate at least up to the foreign tax rate.  If the tax rate is raised further, however, it does discourage capital inflows, and the Diamond-Mirrlees reasoning still implies that a small open economy would not choose to impose such distortions.
How would the home government behave, given this foreign withholding tax rate? In Gordon (1992), I find that the home country, taking the foreign withholding tax rate as given, would never choose a tax rate on the portfolio income of domestic residents equal to this foreign withholding tax rate; the optimal tax rate could in principle be either higher than this rate or zero. 
When the tax rate is below this point, raising the tax rate affects the net-of-tax rate of return only of domestic investments, and so acts like a source-based tax. Within this range, a tax increase is therefore undesirable. When the tax rate is above the foreign withholding tax rate, however, a tax increase affects foreign and domestic holdings equally, making tax increases just above the foreign rate more attractive than tax increases just below the foreign rate. This implies that there is no Nash equilibrium set of tax rates.
Without tax evasion as a problem, I argued above that residence-based taxes should create few externalities, implying little gain from coordination. There fore, even if the treaty led to a clear outcome, it would be very unlikely that both countries would prefer this outcome to the situation without the treaty.
What if investors can evade domestic taxes without any cost or risk by in vesting through foreign financial intermediaries? Then domestic investors can always avoid tax by investing abroad, so a withholding tax is simply a source based tax, and by the same arguments used above we conclude that a small open economy should not impose a source-based tax.
If evasion is costly enough, however, then Gordon (1992) shows that host countries will again impose a withholding tax at a rate equal to the home country rate, rather than act as a tax haven. If the home country acts as a Stackelberg leader, then an equilibrium exists under the double-taxation convention. From each nation’s perspective, this equilibrium Pareto-dominates the equilibrium without the double-taxation convention, in which there are no capital income taxes. In principle, further welfare gains should be possible through explicit coordination of withholding tax rates; capital income tax rates are lower in the above equilibrium than would be jointly optimal. However, successful coordination must be done on a worldwide basis; if Canada and the U.S. alone increased their tax rates, then their investors would simply invest elsewhere, where tax rates remained low.
Another possible explanation for why countries impose withholding taxes is that each country’s equity is a unique asset, if only because it provides risk diversification not available elsewhere. In that case, each country has the incentive to take advantage of its monopoly power by, for example, imposing a withholding tax on payments to foreign equity holders. What if those paying this tax include domestic investors buying through a foreign intermediary to evade domestic taxes? Imposing a withholding tax remains attractive as long as taxing capital income is part of the desired tax system.
If the only motivation toward imposing withholding taxes is to take advantage of market power, then there would be a joint efficiency gain from reducing these trade distortions by jointly setting a low ceiling for withholding tax rates. In theory, Canada should gain more from such an agreement, given the much greater market power of the U.S. Note that the agreement in this case reduces withholding tax rates from those chosen in the Nash equilibrium, whereas the treaty would jointly raise withholding tax rates when its objective is to stem tax evasion.
Corporate Taxation of Repatriated Earnings
Previously, I argued that source-based tax rates would be at or near zero in a Nash equilibrium. How do existing double-taxation conventions dealing with repatriated corporate earnings change the equilibrium behavior of the two governments? Let us ignore initially the effects of taxation at repatriation rather than at accrual and assume that all capital flows are direct investments by multinational corporations.
Under the provisions of the tax treaty, a source-based tax assessed just on foreign direct investment does not affect investment incentives as long as the source-based tax rate is below the corporate tax rate in the multinational’s home country. Therefore, each country has the incentive to set its source based corporate tax rate on foreign direct investment equal to the corporate tax rate prevailing in the other country.
The incentives faced by the host country do not end here, however. Since a multinational pays the same tax rate regardless of the location of an investment, the before-tax rate of return on investments in the two countries would be equated. However, when one country acquires funds from the other country, it pays the net-of-tax rate of return on these funds, as a result of the source based taxes. A small open economy would therefore want to equate the value of the marginal product of capital with this net-of-tax rate of return paid for funds acquired from abroad, or earned on funds invested abroad. In order to induce firms to equate the marginal product of capital with the net-of-tax cost of funds, the government could provide a suitable direct subsidy to new in vestment. This subsidy produces the desired result as long as it is treated as extra income, rather than as a reduction in the creditable tax payment under the tax treaty.
What are the incentives faced by the home country? As before, given the tax rate imposed in the host country, the home country would set its corporate tax rate either to zero or to some rate above the host country’s tax rate. When its rate is below the host country’s tax rate, the tax is simply a source based tax and so is undesirable. When the rate is higher than the source country’s tax rate, then the tax at the margin is a residence-based tax and so is potentially desirable. The fact that the host country receives some of the revenue makes the tax more attractive at the margin, since less is lost from any drop in investment due to a tax increase.
Given that multinationals are based in both countries, each country is both a home country and a host country for some investment. A Nash equilibrium mayor may not exist. If not, one country, the U.S. for example, could act as a Stackelberg leader. As before, the resulting equilibrium should provide higher welfare in each country than exists in the equilibrium without the double-taxation convention, in which capital income taxes are not used. Further gains from worldwide coordination exist, but not much can be accomplished on a bilateral basis.
How do the results change if we take into account that investments made through a foreign subsidiary are taxed only at repatriation rather than at accrual? As Jun (1987) shows, postponement of realization drives the effective tax rate on the initial equity investment down toward (and in the limit equal to) the tax rate in the host country, while Hartman (1985) argues that for in vestments financed by retained earnings, the effective tax rate is simply the host country tax rate. But if the effective tax rate is the host country tax rate, then a small open host country would not choose to impose such a tax.
What happens if capital flows to foreign firms can take the form of portfolio investments, rather than just direct investment by foreign subsidiaries? Funds can flow from the home country to the host country either by direct investment by a multinational, which itself is owned by home country individuals, or else by purchase of equity in host country firms by home country individuals. In either case, the same host country corporate taxes are paid, and home country individuals owe tax at the same rate on the net income they receive. The key difference is that with direct investment by a multinational, supplementary taxes might be owed to the home country. If so, portfolio investment is preferred for tax reasons. If there are no nontax reasons favoring direct investment, then supplementary taxes would never be paid at repatriation, and a small open host country would therefore not impose a source-based tax.
What incentives are created by the presence of the dividend-credit scheme in Canada? Most of the discussion of the effects of such a scheme assume a closed economy. But as Boadway and Bruce (1989) emphasize, the effects of the scheme are very different in an open economy. For simplicity, assume that the scheme provides full integration of the corporate and personal tax systems, and assume to begin with that the Canadian corporate tax rate is below the U.S. corporate rate.
When Canadian corporations invest in the U.S., they must pay U.S. corporate taxes on their foreign earnings. When the earnings are repatriated, no corporate surtax is due, but shareholders still receive a dividend credit based on the difference between their personal tax rate and the Canadian corporate tax rate. If the corporate tax rate exceeds the personal tax rate, then on net
Canada provides a subsidy for direct investment by Canadian multinationals in the U.S. and thereby raises the return to savings in Canada above the return available in the world market. This subsidy is not available when Canadian individuals buy shares in U.S. corporations, so this scheme favors direct in vestment over portfolio investment.
How does its presence affect the equilibrium corporate tax rates? As a home country, when Canada raises its corporate tax rate it increases the subsidy it gives to investments by Canadian investors in the U.S., making it more likely that the optimal tax rate is zero rather than above the U.S. rate. As a host country, however, Canada would still wish to set its corporate tax rate equal to the U.S. rate. Therefore, on net Canada would more likely prefer a tax rate below the U.S. rate.
The U.S., as Stackelberg leader, would now have to take into account that Canada would more likely keep its corporate tax rate below that of the U.S. This would certainly affect the optimal tax rate in the U.S.
2.2 Tax Policy, Given Free Trade
Based on recent agreements, all tariffs and most nontariff barriers to trade between the U.S. and Canada will be eliminated by 1998. What implications will this policy change have for the domestic tax structure in each country?
Assume to begin with that each country is free to use tariffs, but that trade is not otherwise restricted. As noted above, Diamond and Mirrlees (1971) showed that a small open economy that imposes excise taxes on all goods would choose to produce efficiently under an optimal tax system. This implies that it would choose not to distort trade patterns. As emphasized in Gordon and Levinsohn (1990), however, this does not necessarily imply that a country would avoid use of tariffs. In particular, the effects of an excise tax on a particular good can be duplicated by a production tax and an import tariff or export subsidy on that good. It may be that it is easier to administer a combination of a tariff and a production tax, at equal rates, on some goods than to administer an excise tax on these goods. Both have the same economic effects, and neither distorts trade patterns.  To the degree that tax or other policies distort relative output prices, then optimal policy would involve undoing these distortions at the border through suitable export taxes and subsidies.
Of course, if a country does have market power in a particular good, then it will want to take advantage of this market power, as shown in the optimal tariff literature. However, doing so does not require use of explicit tariffs, since again the combination of a production tax and a consumption subsidy has the same effects.
One of the main source-based taxes in the U.S. and Canada is the corporate income tax. This tax raises the prices of corporate relative to noncorporate goods, and alters relative corporate prices due to differences in capital-output ratios and due to idiosyncracies in depreciation and other detailed provisions in the tax law. Under optimal tariff policy in a small country, these distortions would be offset at the border.
Domestic regulations may also distort relative prices, creating the incentive to use tariffs to offset these distortions. For example, agricultural price sup ports lead food prices to be above marginal costs, justifying export subsidies on these products. Similarly, the U.S. lumber industry may face a below market price for use of the National Forests, leading lumber prices to be below marginal costs and thereby justifying export taxes.
The Canadian sales tax also creates nontrivial distortions to the relative prices of imported and domestically produced goods, as reported in Dodge and Sargent (1987). Since the tax is imposed at the wholesale rather than the retail level, the amount of tax collected on a finished product depends on the number of transactions that occur between firms at the wholesale stage. To some degree, industries can change how they organize their production in order to minimize the total sales tax payments that are incurred, but doing so has its own costs.
What will be the implications for source-based taxes, and for regulatory distortions, of the free-trade agreement between the U.S. and Canada? This agreement will have no economic effect if each country can costlessly offset the change through a suitable modification to its domestic tax structure as it applies specifically to income flows between the two countries. To compensate for the drop in tariff or nontariff barriers for a particular good, each country could compensate by cutting the domestic tax (increasing the subsidy) on production of that good and increasing the sales tax rate on consumption of that good. To avoid any economic changes, tariffs between the U.S.or Canada and third countries would need to be suitably readjusted.
These compensating adjustments in the domestic tax system, to neutralize the effects of the free-trade agreement between the U.S. and Canada, are substantial and awkward. If they cannot be made, then source-based taxes on production will become more costly from each country’s perspective, since the resulting distortions to the trade pattern between the U.S. and Canada could no longer be neutralized by suitable border distortions. Given the large volume of trade between the two countries, these distortions will be important, creating significant pressure to cut distortions to the relative prices of different goods to maintain an efficient composition of trade. Similarly,regulations that distort relative output prices become much more costly.
Various responses to these pressures are possible. For one, policy distortions to the relative prices of domestic output can be reduced by “leveling the playing field” by eliminating differences in the effective tax rates on different industries. If the tax system raises the prices of all domestically produced goods by the same percentage, then the exchange rate between the Canadian and the U.S. dollars would simply readjust, leaving trading incentives un changed. Such a shift in the tax system in each country has been occurring recently in any case, whether or not connected to the U.S.-Canada free-trade agreement. To the extent that regulations create price distortions (e.g.,various agricultural programs designed to raise crop prices artificially), then each country would face competitive pressure to redesign these regulations to “level the playing field” between affected sectors and other domestic industries.
From the perspective of the two countries together, however, as long as any particular industry is equally favored or disfavored by the tax and regulatory system in both countries, no policy distortion to trade patterns is created.In some cases, harmonizing the relative tax rates on different industries in the two countries may be easier or more desirable than allowing competitive pressures to undermine rate differences across industries, which each country might in principle have desired. Tariffs between the U.S. or Canada and third countries can then be used to neutralize these distortions.
Coordination between the two countries can also affect the size of tariff barriers with respect to the rest of the world. Together the two countries have more market power than either country has in isolation, particularly in goods such as lumber or wheat. Coordination would therefore lead to increased restrictions on their combined trade with the rest of the world.
Coordination of policies between the two countries, whether concerning policies with respect to each other or with respect to the rest of the world, does not require a written treaty. A country that deviates from an implicit agreement could incur “punishment” from the other country in some form. As long as the threat of “punishment” is a sufficient deterrent, the implicit agreement will be sustainable. This use of threats to enforce an implicit agreement is commonly seen with regard to tariff policy, and may well occur with regard to tax policy as well .
2.3 Tax Policy, Given Mobility of Individuals
So far, the discussion has ignored tax and expenditure pressures created by the movement of individuals across borders. Yet travel between the two countries is extensive, taking the form of tourism and business trips as well as changes in the location of employment or even of citizenship. The U.S. Canada free-trade agreement reduces some restrictions on the movement of individuals across the border. What pressures are created by such movement of individuals?
This is in many ways the key question examined in the local public finance literature. As argued by Buchanan and Goetz (1972) and many others, mobility of individuals imposes externalities on jurisdictions, which depend on the degree to which an individual pays an amount in taxes that differs from the costs the individual imposes on the jurisdiction, whether in the form of in creased costs of public services or increased congestion. This applies in both the sending and the receiving jurisdiction. When a community gains on net from the presence of an individual, because tax payments exceed the costs the individual imposes on the community, the community has an incentive to encourage immigration; the converse is also true. This competitive pressure pushes the tax system toward a benefit-tax structure in which the net gain to the jurisdiction from acquiring or losing an extra individual is competed down to zero. At that point, individuals simply pay for the costs they impose on the community.
What implications does this story have for tax policy at the national level?
Consider first the pressures created by temporary migration, such as tourism. Through such migration, countries trade in services as well as in goods. Therefore, tax and regulatory policies potentially can distort the relative trade in services versus goods, as well as distort the composition of trade in goods that physically cross the border. Countries would face competitive pressure to reduce or eliminate policy distortions to the composition of trade,which now includes services as well as “tradables.”
When individuals cross the border, however, they also normally increase the costs of public services, since they make use of roads, police protection, and other services. A country would want to encourage immigration if tax payments exceed the net costs imposed by immigration, as well as the con verse. By the Diamond-Mirrlees reasoning, a small open economy would simply charge for the services obtained, whether directly or indirectly, and so would design tax policy to increase the cost of goods used by migrants above the cost of goods that are physically exported. The playing field would be intentionally “tilted” to compensate for the costs imposed on the public sector when certain goods are purchased. Countries with relative market power in, for example, tourist-related services, would attempt to charge even more to take advantage of this market power.
Individuals who change their country of employment create more extensive changes in the tax revenue and public service costs in each country. Presumably, the relocation of higher-paid individuals creates larger relative gains, since their tax payments are relatively large compared to the cost of the public services they require. Similarly, each country has an incentive to discourage the immigration of those who impose net fiscal costs, be they the poor,the sick, or the elderly.
Competition for individuals who provide a net fiscal gain to the jurisdiction therefore reduces the degree to which the fiscal system redistributes from rich to poor and again pushes toward a benefit-tax system. But the resulting tax structure cannot simply equate benefits and tax payments in present value over the lifetime, since individuals can remain in the country during those periods when they gain on net, and leave when they lose on net. Therefore,even the timing of taxes would be pushed to coincide with the timing of benefits. As a result, national debt would be discouraged, since it creates the incentive to emigrate during those periods when the debt is repaid. Similarly,redistributive policies such as social security would come under pressure, since those who work for a short period under existing law gain substantially from the system. Since non workers do not pay labor income taxes but do make use of public services, and often more extensively than workers do, even labor in come taxes may not easily be sustainable. A country that relies heavily on a labor income tax would become a haven for non workers, such as students or the retired. Public services such as subsidized college education or free medical care would attract residents of the other country who hope to take advantage of these subsidies.  All of these are examples of pressures towards a benefit-tax structure. Which tax system most resembles a benefit-tax structure depends on the composition of public expenditures. If consumption of public services roughly corresponds with consumption of private goods, then a consumption tax or a VAT may most closely approximate a benefit tax. User fees certainly approximate a benefit tax.
This evolution towards a benefit-tax structure would occur even if both countries desired a redistributive fiscal policy. Based on this reasoning, the conventional wisdom in public finance has always been that redistribution should be done at the national rather than at the local level. Retranslated to this context, the analysis suggests that policies regarding redistribution should be coordinated between the two countries. Of course, coordination must cover both tax and expenditure policies, otherwise each country can make use of its remaining flexibility to attract those who pay more than they receive.
The Tiebout literature argues, however, that expenditures financed by benefit taxes should not need coordination; competition among jurisdictions pushes them to offer an efficient composition and level of public services.
How can jurisdictions choose the composition and level of services financed by benefit taxes without coordination, however, and yet agree to restrictions on their expenditure policies to prevent undermining of inter jurisdictional agreements on redistribution? In a federal system, these contradictory pressures are avoided by having redistribution done at the national level. Competition among communities then leads to a local benefit-tax structure. In this case there is no fiscal gain to a community from attracting those who pay relatively more to the national government. Without a federal structure, how ever, coordination of at least some expenditure policies may be necessary to preserve redistributive policies, even though this coordination undermines the ability of each government to provide the composition or level of public expenditures desired by its citizenry.
At this point, these various pressures will be much more important within the EC, where all restrictions to migration are being eliminated, than in the U.S.-Canadian context. The pressures are still there, however, and will surely increase over time.
When analyzing the fiscal implications of unrestricted mobility among jurisdictions, the local public finance literature concludes that the fiscal system will be driven toward a benefit-tax structure, in which people pay in taxes an amount appropriate to cover the costs they impose on the public sector. Yet existing national tax systems in the U.S. and Canada differ substantially from benefit-tax structures. As a result, the increasing mobility of output,capital, and even labor, between the two economies will create a variety of pressures pushing the tax system towards a benefit-tax structure.
Where this pressure will be strongest depends on the degree of mobility of particular types of goods, services, and people across the border. This paper explored in turn the types of pressures created by mobility of capital,unrestricted trade in all outputs, and mobility of people.
Even though tax competition will push each country’s fiscal structure to ward that of a benefit tax, such a tax system may not be mutually advantageous; in fact, both countries may well gain through explicit or implicit coordination of fiscal policies. In many cases the appropriate form of coordination involves equalization of tax rates. Where, for example, the local public finance literature calls for the national government to handle redistribution, given the degree to which individual mobility undermines anyone community’s efforts at redistribution, the same logic calls here for coordinating redistributive policies between the two countries. A number of other examples of fruitful areas for policy coordination are discussed.
Ultimately, the implications of the increasing interdependence of the two economies for their national fiscal structures should be substantial. Fortunately, the two countries can watch the European experience after 1992 to learn better how to redesign the existing fiscal systems in the two countries.
Boadway, Robin, and Neil Bruce. 1989. Problems with Integrating Corporate and Personal Income Taxes in an Open Economy. Queen’s University. Manuscript.
Bond, Eric w., and Larry Samuelson. 1989. Strategic Behavior and the Rules for International Taxation of Capital. Economic Journal 99 (398): 1099-1111.
Buchanan, James, and C. Goetz. 1972. Efficiency Limits of Fiscal Mobility: An Assessment of the Tiebout Model. Journal of Public Economics 1 (1):25-43.
Diamond, Peter, and James Mirrlees. 1971. Optimal Taxation and Public Production: II-Tax Rules. American Economic Review 61, no. 3, pt. 1:261-78.
Dodge, David A., and John H. Sargent. 1987. Canada. In World Tax Reform: A Progress Report, ed. Joseph Pechman. Washington, D.C.: Brookings.
Findlay, Christopher. 1986. Optimal Taxation of International Income Flows. Economic Record 62 (June):208-14.
Giovannini, Alberto, and James R. Hines, Jr. 1990. Capital Flight and Tax Competi tion: Are There Viable Solutions to Both Problems? Columbia University. Manu script (March).
Gordon, Roger H. 1992. Can Capital Income Taxes Survive in Open Economies? Journal of Finance 47(3).
—. 1988. Comment on “Tax Policy and International Competitiveness,” by Law rence H. Summers. In International Aspects of Fiscal Policies, ed. Jacob A. Fren kel. Chicago: University of Chicago Press.
—. 1986. Taxation of Investment and Savings in a World Economy. American Economic Review 76 (5): 1086-1102.
—. 1983. An Optimal Taxation Approach to Fiscal Federalism. Quarterly Journal of Economics 98 (4):567-86.
Gordon, Roger H., and James Levinsohn. 1990. The Linkage Between Domestic Taxes and Border Taxes. In Taxation in the Global Economy, ed. Assaf Razin and Joel Slemrod. Chicago: University of Chicago Press.
Gordon, Roger H., and Joel Slemrod. 1988. Do We Collect any Revenue from Taxing Capital Income? In Tax Policy and the Economy, vol. 2, ed. Lawrence H. Summers, 89-130. Cambridge: MIT Press.
Gordon, Roger H., and Hal Varian. 1989. Taxation of Asset Income in the Presence of a World Securities Market. Journal of International Economics 26 (314):205-26.
Hartman, David. 1985. Tax Policy and Foreign Direct Investment. Journal of Public Economics 26:107-21.
Hines, James R., Jr., and R. Glenn Hubbard. 1990. Coming Home to America: Dividend Repatriation Decisions of U.S. Multinationals. In Taxation in the Global Econ omy, ed. Assaf Razin and Joel Slemrod, 161-200. Chicago: University of Chicago Press.
Jun, Joosung. 1987. Taxation, International Investment and Financing Sources. Yale University. Manuscript.
Mirrlees, James A. 1972. On Producer Taxation. Review of Economic Studies 39 (I), no. 117:105-11.
Slemrod, Joel. 1988. Effect of Taxation with International Capital Mobility. In Uneasy Compromise: Problems of Hybrid Income-Consumption Tax, ed. Henry Aaron, Harvey Galper, and Joseph A. Pechman, 115-48. Washington, D.C.: Brookings.
Stiglitz, Joseph. 1985. The General Theory of Tax Avoidance. National Tax Journal 38 (3):325-37.
—. 1983. The Theory of Local Public Goods Twenty-fiveYears after Tiebout: A Perspective. In Local Provision of Public Services: The Tiebout Model after Twenty five Years. New York: Academic Press.
Tiebout, Charles M. 1956. A Pure Theory of Local Expenditure. Journal of Political Economy 64 (5):416-24.
Roger H. Gordon is professor of economics at the University of Michigan and a research associate of the National Bureau of Economic Research.
Much of this paper was written while the author was visiting the Universidade Nova de Lisboa. He would like to thank the other participants in the Canada-U.S. Tax Comparisons project, and especially John Shoven and John Whalley, for comments on a previous draft.
[1.] By “coordination,” I will mean negotiation to internalize fiscal externalities. “Harmonization,” in contrast, will refer to equalization of tax rates and tax bases, whether this occurs through agreement or as a result of market forces.
[2.] To the degree to which a community’s taxes deviate from benefit taxes, other communities have the incentive to bid to attract those individuals or firms who on net pay more than they impose in costs on the community. For further discussion, see Buchanan and Goetz (1972).
[3.] See Gordon (1983) for an exploration of the various possible sources of externalities.
[4.] See Boadway and Bruce (ch. I in this volume) for a detailed description of the existing tax structures in the two countries.
[5.] For simplicity, the discussion ignores non corporate firms.
[6.] In Canada, there is a dividend credit, which reduces the extent of the double taxation inherent in this tax structure.
[7.] In all cases, the credit is not refundable, so is limited to the amount of taxes due in the home country on that income. For further detail on U.S. and Canadian provisions, see Boadway and Bruce (ch. I in this volume).
[8.] This last assumption may not be that unreasonable. Hines and Hubbard (1990) provide evidence that U.S. multinationals, at least, pay little or no U.S. taxes on their repatriated earnings.
[9.] For example, borrowing implies a negative holding of bonds, and interest payments are normally deductible,
[10.] Technically, this requires that each investor be able to “go short” in at least one asset, deducting the payments from taxable income, and that each investor be taxed relatively more heavily on the asset he goes short in, II.See Stiglitz (1985) for a number of examples,
[12.] One apparent response has been an increase in corporate borrowing, since corporate interest deductions are still allowed.
[13.] Changes in the market interest rate have further effects on efficiency to the degree to which choices were not efficient initially due to distorting taxes. In particular, if income from savings is taxed, then any resulting decrease in savings reduces welfare. A change in the market interest rate also has distributional consequences which may be of concern to the government.
[14.] Competition among such countries would drive any tax down to zero, however.
[15.] See below, however, for a discussion of use of source-based taxes to help enforce residence-based taxes.
[16.] Tax rates might also depend on the form of the financial claim to the real capital income.
[17.] See, for example, Mirrlees (1972).
[18.] If the profits arise from control of a patent, then the patent right itself can be relocated to a tax-free country, and the pure profits paid in the form of a tax-deductible license fee to this coun try, with no change in the location of production.
[19.] See Gordon and Varian (1989), however, for an argument that even a small country may have market power with respect to equity issued in the country, due to its idiosyncratic risk. See also Gordon (1988), who argues that when each country produces a distinct good, each country has market power and the optimal use of this market power will lead it to restrict net capital flows.
[20.] See Giovannini and Hines (1990) for a discussion of how transfers might be made between governments so that the allocation of revenue among countries would be equivalent to that arising under a residence-based tax.
[21.] The same incentives can exist even with taxation based on residence. For example, if foreigners have large holdings of domestically issued bonds, then a government has the incentive to inflate the currency unexpectedly,thereby wiping out its debt to foreigners.
[22.] With a diversity of foreign tax rates, the story becomes a bit more complicated, since the country is no longer a price-taker in the world capital market.
[23.] As noted before, a large open economy would set its taxes to take advantage of this market power.
[24.] Bond and Samuelson (1989) find, under different assumptions that allow each country to tax domestic and foreign income at different rates, that the optimal tax rate must be higher.
[25.] Gordon (1992) shows that there will be a Stackelberg equilibrium, however
[26.] For further discussion, see Gordon and Varian (1989).
[27.] The withholding tax also reduces the incentive to use foreign financial intermediaries, leading to potential efficiency gains if domestic intermediaries are more efficient at handling do mestic investments.
[28.] Such an agreement could require coordination of withholding tax rates with respect to third countries. Otherwise, the optimal Canadian withholding tax rates toward third countries would presumably be low, given the limits on its market power. As a result, third parties could then purchase U.S. equity through Canadian financial intermediaries, paying two rounds of low withholding tax rates as the funds traveled from the U.S. to Canada and then to the third country, rather than paying the higher U.S. withholding tax rate that applied to that third country.
[29.] I assume here that if the resulting tax rate on domestic investors is higher than is desired, then the government can rebate the excess through, for example, a dividend-credit scheme.
[30.] The appropriate subsidy rate would be tf’/(I – t), where t is the residence-based tax rate, andf’ is the marginal product of capital.
[3 1.] For further discussion, see Findlay (1986).
[32.] This ignores the dividend-credit scheme available in Canada for income from domestic corporations. See below for further discussion.
[33.] Specifically, under the dividend-credit scheme, shareholders receiving a dollar of dividends are credited with earning I/(l – T) in pre-corporate-tax profits, where T, is the corporate tax rate. On this income, they owe personal taxes at rate t” but receive a credit for corporate taxes already paid, implying a net tax liability of (t, – T)/(l – T,). By assumption, t, < T,. If income Y is repatriated from foreign earnings and generates no corporate surtax at repatriation, then net tax payments in Canada on these earnings are (t, – T,)/(l – T,)Y < 0, implying a subsidy to foreign investment.
[34.] Trade distortions are present in this argument when the tax law favors purchases of goods produced in a particular location.
[35.] There is a large literature in cooperative game theory on the sustainability of such a cooperative outcome.
[36.] Since tariffs can be duplicated through use of a suitable set of domestic taxes, the two policy areas are not really distinct in any case.
[37.] Migration is not even needed for trade in financial services.
[38.] Of course, countries may impose residency requirements for these benefits, to some de gree lessening the pressures.
[39. ] The recent shift from a property tax to a head tax in financing public services in the United Kingdom could be interpreted as a response to this type of pressure.
[40.] Given that a VAT should roughly correspond to a benefit tax, at least relative to other taxes used at the national level, it is ironic that the EC has focused its tax coordination efforts on this particular tax. For discussion of various limitations of the Tiebout argument, see, for example, Stiglitz (1983).
Roger H. Gordon