The real exchange rate under various systems of international capital taxation.

Hatzipanayotou, Panos

I. Introduction

International organizations, such as the Organization for Economic Co-Operation and Development (OECD) and the European Community (EC), have promoted, to varying degrees, the liberalization of capital movements among their member countries.(1) The "Code of Liberalization of Capital Movements," an agreement drawn in 1982 by the OECD, sets different degrees of capital liberalization among its member countries, depending on the nature of capital transactions (e.g., direct or portfolio investment). The Treaty of Rome (1957) recommended, among other things, the liberalization of capital movements among the EC countries (Chapter IV, articles 67-72). By and large, capital movements in the EC have been liberalized since 1990. The remaining restrictions in the capital markets of some member countries, such as Greece, Ireland, Portugal, and Spain, were to be eliminated by the end of 1992.(2) These developments in the world capital markets have prompted a renewed interest among trade theorists regarding the response of key economic variables (e.g., national income, employment and welfare) to the liberalization of capital movements.

With capital becoming more mobile internationally, capital flows in a country depend, among other things, on the domestic and the foreign capital tax rate, and on the prevailing system of taxing net repatriated capital earnings in the capital-exporting countries. Frequently, capital-exporting countries tax the net repatriated capital earnings according to a system (i) of tax credits, under which tax payments by foreign capital invested in a capital-importing country are credited toward its tax liability to the source capital-exporting country,(3) (ii) of tax deductions, under which payments by foreign capital invested in a capital-importing country are deducted as a cost in calculating its tax liability to the source capital-exporting country, and (iii) of untaxed repatriated net capital earnings by the capital-exporting country. Recently, Bond and Samuelson |4~ using a game theoretic approach in a one-commodity two-country model, examine the welfare effects of capital taxes under a system of capital tax credits and of capital tax deductions. Slemrod |17~ using aggregate data from seven countries (i.e., Canada, France, Italy, Japan, the Netherlands, the U.K., and the U.S.) examines how the U.S. tax system in conjunction with the tax system of a capital-exporting country affects the FDI in the U.S. Slemrod's results support a negative relationship between effective U.S. capital tax rates and FDI into the U.S. Disaggregating the data by capital-exporting country, he does not find systematic differences in the responsiveness of FDI, with respect to changes in U.S. tax rates, from capital-exporting countries with a tax credit system (i.e., Italy, Japan, and the U.K.), and the remaining capital-exporting countries with a system of untaxed repatriated capital earnings.(4) Hatzipanayotou and Michael |11~ in a model of a small capital-importing and a small capital-exporting country, assuming free trade and variable labor supply due to endogenous supply adjustments examine, among other things, the employment and welfare effects of capital taxes under a system of capital tax credits, tax deductions, and untaxed repatriated net capital earnings.

The response of another key variable, the real exchange rate, to the liberalization of capital movements is yet to be examined in a trade theoretic context, despite its important analytical and policy implications. Trade theoretic studies of the real exchange rate, ignoring international capital mobility, have primarily examined its response to trade restrictions (e.g., a higher tariff), and to exogenous shocks (e.g., income transfers, and an improvement in the terms of trade). In the context of a small open economy, it has been concluded that a higher tariff or an improvement in the terms of trade leads to a real exchange rate appreciation |2; 6; 7; 13~.(5) Recent studies, however, have shown that these conventional results may not necessarily hold in more general models, such as the three-good (exported, imported and nontraded goods) two-factor Heckscher-Ohlin model, or the three-good specific factor model |5; 8~.(6) In such general models (e.g., the three-good specific factor model), the higher tariff exerts two effects, which work in opposite directions, on the real exchange rate. A substitution effect, due to the induced production and consumption distortions, and an income effect whenever the initial tariff is non-zero. As a result, whether a higher tariff leads to a real exchange rate appreciation or depreciation depends, among other things, on the relative size of these two effects.

This paper examines the real exchange rate effect of liberalizing capital movements, through reduced capital taxes under the three alternative systems of taxing repatriated capital earnings, for a small capital-importing and a small capital-exporting country. Two results emerge from the present analysis. First, whether liberalization of capital movements affects the real exchange rate depends on whether the country is a net capital-exporter or a net capital-importer, and on the prevailing capital tax system. Second, when changes in capital taxes affect the real exchange rate, it is, as in the case of a higher tariff, via a substitution and an income effect. But, contrary to the case of a higher tariff, the substitution effect is only due to a production distortion, while the income effect is due to changes in capital tax revenue that is distributed to households as a lump-sum. The analysis also stresses under which tax systems the real exchange rate effect of a higher capital tax depends, among other things, on the foreign capital tax rate.

Section II examines the real exchange rate effect of liberalizing capital movements under the three alternative tax systems for a small capital-importing country, and section III for a small capital-exporting country. The last section offers some concluding remarks.

II. A Model of a Small Capital-importing Country

Consider a small capital-importing country that produces many traded and nontraded goods, using an internationally mobile factor of production, capital, and some nontraded factors. The country's endowment of capital and of the nontraded factors are fixed. Commodity trade is free and the domestic prices of the traded goods, which equal the world prices, are given by the vector p. The relative prices of the nontraded goods, given by the vector q, are endogenously determined in order to equate domestic demand and supply of the nontraded goods.

Let r and r* denote the domestic and world rate of return to capital. Similarly, positive (negative) values for |Rho~ and |Rho~* denote the capital tax (subsidy) by the home capital-importing and the foreign capital-exporting country. Three alternative systems of taxing the repatriated earnings of the internationally mobile capital are considered.

First, a system of capital tax credits under which the capital-exporting country offers capital tax credits that reduce the domestic tax liability of its capital invested abroad by the amount of tax payments made in the capital-importing country. That is, a unit of capital invested in the capital-importing country pays an amount r|Rho~ in taxes in the host country, and an amount max((|Rho~* - |Rho~)r, 0) to the foreign capital-exporting country. Thus, a unit of capital earns a net rate of return equal to r(l - max(|Rho~*,|Rho~)), when invested in the home capital-importing country, and equal to r*(1 - |Rho~*), when invested in the foreign capital-exporting country. Equilibrium in the world capital market requires that

r(1 - max(|Rho~*, |Rho~)) = r*(1 - |Rho~*). (1)

Given the small country assumption (i.e., dp= dr* =d|Rho~* = 0), dr = 0 under a system of capital tax credits when p |is less than~ |Rho~* since, in this case, equilibrium in the world capital market requires that r = r*. That is, when |Rho~ |is less than~ |Rho~* under a system of capital tax credits, reducing the domestic capital tax simply transfers capital tax revenue from the home capital-importing to the foreign capital-exporting country without affecting the domestic rate of return to capital.(7) When p |is greater than~ |Rho~* under tax credits, equation (1) shows that dr = |r/(1 - |Rho~)~d|Rho~. In this case, changing the domestic capital tax affects the domestic after-tax rate of return to capital.

The second system of international taxation of repatriated capital earnings is that of tax deductions. Under this system, tax payments by foreign capital in the capital-importing country are deducted as a cost in calculating its gross tax liability in the capital-exporting country. A unit of capital earns a net rate of return r(1 - |Rho~)(1 - |Rho~*), when invested in the capital-importing country, and r*(1 -|Rho~*), when invested in the foreign capital-exporting country. Equilibrium in the world capital market requires that

r(1 - |Rho~) = r*. (2)

The last tax system considered is one of untaxed repatriated capital earnings, under which equilibrium in the world capital market requires that

r(1 - p) = r*(1 - |Rho~*). (3)

Utilizing the small country assumption under the last two capital tax systems, dr = |r/(1 - |Rho~)~d|Rho~.

Utility depends on the consumption of the traded and nontraded goods, all of which are normal. The aggregate expenditure function E(p, q, u) captures the minimum expenditure required to achieve utility u at price vectors p and q. The expenditure function is increasing and strictly concave in prices. Its derivatives with respect to p (i.e., |E.sub.p~ = |Delta~E/|Delta~p), and with respect to q (i.e., |E.sub.q~ = |Delta~E/|Delta~q) are the vectors of compensated demand functions for the traded and the nontraded goods. By choice of units, |E.sub.u~(=|Delta~E/|Delta~u) = 1.

The domestic supply of capital in the capital-importing country is |Mathematical Expression Omitted~, where |Mathematical Expression Omitted~ is the domestic fixed endowment of capital, and K* (|is greater than~ 0) is the stock of foreign capital invested domestically. Since the country's capital endowment is fixed, dK = dK*. For the rest of the analysis, nontraded factors are omitted since they do not affect the results. Let R(p, q, K) be the country's gross domestic product (GDP) function, showing the value of the maximum output produced at price vectors p and q, and supply of capital K, given the endowment of the nontraded factors. The GDP function is strictly convex in prices and concave in K (i.e., |R.sub.KK~ |is less than~ 0).(8) Its derivatives with respect to p (i.e., |R.sub.p~), and with respect to q (i.e., |R.sub.q~) are the vectors of supply functions of the traded and the nontraded goods, and with respect to K (i.e., |R.sub.K~) is the value of the marginal product of capital. The domestic rate of return to capital is equated to the value of the marginal product of capital in every domestic use. That is,

r = |R.sub.K~(p, q, K). (4)

Differentiating equation (4), holding p constant, gives the changes in the domestic supply of capital due to changes in r and q as follows:

|Mathematical Expression Omitted~.

In equilibrium, the country's expenditure equals revenue from production minus net payments to foreign capital invested at home. The expenditure-income identity is

E(p, q, u) = R(p, q, K) - r(1 - |Rho~)K*. (6)

Equilibrium in the nontraded good markets requires that demand for the

nontraded goods equals supply of the nontraded goods. That is,

|Z.sub.q~(p, q, u, K) = Eq(p, q, u) - Rq(p, q, K) = 0, (7)

where |Z.sub.q~ denotes the vector of excess demands for the nontraded goods.

Equations (4), (6) and (7) constitute a specialized long-run steady state trade model, consistent with a complete intertemporal equilibrium and zero long-run growth. Specifically, since the domestic rate of return to capital equals the value of its marginal product (i.e., equation (4)), capital markets are in long-run equilibrium, further capital formation does not take place (i.e., dK = dK* = 0). Then equation (6), which is the current account long-run equilibrium, indicates that the merchandise trade deficit (i.e., E(p, q, u) - R(p, q)) must equal the net payments to foreign capital invested at home, (i.e., -r (1 - |Rho~)K*).(9) Lastly, within a general equilibrium context, equation (7) must hold at all times to determine the endogenous nontraded goods prices.

Differentiating equations (6) and (7), and using (5), gives the changes in the endogenous variables u and q due to changes in the domestic rate of return to capital and the capital tax rate as follows:

|Mathematical Expression Omitted~,

where, since goods are normal in consumption, |E.sub.qu~ (=|Z.sub.qu~) is a vector with all positive elements denoting the changes in the compensated demands for the nontraded goods due to changes in income. The determinant of the left-hand side coefficient matrix, |Mathematical Expression Omitted~) is negative if nontraded good markets are Walras stable.(10)

The Real Exchange Rate Effect of Liberalizing Capital Movements

In this section, assuming that commodity trade is free, and that the rate of return to the internationally mobile capital is taxed according to various rules of international capital taxation, we establish sufficient conditions under which liberalization of capital movements, through the reduction of the tax on the domestic rate of return to capital, by the small capital-importing country (i.e., d|Rho~ |is less than~ 0) leads to an appreciation or depreciation of its real exchange rate.

In a multi-commodity (i.e., many traded and nontraded goods) model, the country's real exchange rate (|Pi~) can be defined as a fixed-weight index number of the relative prices of nontraded goods (i.e., x|prime~q) to a price index of traded goods (i.e., |p.sub.T~) |14~. That is, |Pi~ = x|prime~q/|p.sub.T~, where x is the vector of base-period production and consumption of nontraded goods. Since prices of traded goods are constant, changes in |Pi~ are due to changes in the fixed-weight index number of the relative prices of nontraded goods. That is,

d|Pi~ = x|prime~dq/|p.sub.T~.(9)

Equation (9) indicates that an increase (decrease) in the fixed-weight index number of the relative prices of the nontraded goods |i.e., x|prime~dq |is greater than~ 0(|is less than~ 0)~, leads to an appreciation (depreciation) of the real exchange rate.

Consider the real exchange rate effect of a lower capital tax under a system of capital tax credits when |Rho~ |is less than~ |Rho~*. Since changes in the real exchange rate are due to changes in the fixed-weight index number of the relative prices of nontraded goods, equations (8) can give the effect of the lower capital tax on the relative prices of nontraded goods as follows:

|Delta~dq = - rK*|E.sub.qu~d|Rho~. (10)

In this case, the lower capital tax transfers capital tax revenue from the home capital-importing to the foreign capital-exporting country. Since all goods are normal, the induced lower capital tax revenue reduces the domestic demand and relative prices of the nontraded goods, and causes a depreciation of the real exchange rate.

Next, consider the real exchange rate effect of a lower capital tax under a system of capital tax credits when |Rho~ |is greater than~ |Rho~*, or under the other two systems of taxing the repatriated earnings of the internationally mobile capital considered by this analysis. In these cases, the lower capital tax raises the after-tax domestic rate of return to capital, its domestic use, and capital tax revenue for the capital-importing country. Since dr = |r/(1 - |Rho~)~d|Rho~, equations (8) give the effect of the lower capital tax on the relative prices of nontraded goods as follows:

|Mathematical Expression Omitted~.

Equation (11) shows that the lower capital tax, which increases the domestic use of capital, affects the relative prices of nontraded goods directly through changes in production (i.e., |Mathematical Expression Omitted~), and indirectly through changes in consumption (i.e., |Mathematical Expression Omitted~) due to changes in capital tax revenue. The lower capital tax by increasing the domestic use of capital increases capital tax revenue and through its indirect consumption effect has a positive effect on the demands and relative prices of the nontraded goods. The lower capital tax through its production effect increases (decreases) the production of the capital (non-capital) intensive nontraded goods.(11)

To get better intuition about this result, consider the case of a single nontraded good. The lower capital tax, which increases the domestic use of capital and the capital tax revenue, raises the consumption of the nontraded good, exerting a positive effect on its relative price. If also the nontraded good is non-capital intensive (i.e., |R.sub.qK~ |is less than~ 0), then, since by assumption |Mathematical Expression Omitted~, the lower capital tax reduces its production, causing an additional increase in its relative price and an appreciation of the real exchange rate (i.e., |Mathematical Expression Omitted~). But, if the nontraded good is capital intensive (i.e., |R.sub.qK~ |is greater than~ 0), the lower capital tax increases its supply, exerting a negative effect on its relative price. Thus, the relative price of the nontraded good increases (may decrease) and the real exchange rate appreciates (may depreciate), if the nontraded good is non-capital (capital) intensive.

This ambiguity in the effect of a lower capital tax on the real exchange rate, due to the substitution effect in production and the negative income effect in consumption, resembles the recent results of trade theoretic studies regarding the effects of a higher tariff on the real exchange rate. For example, Edwards and van Wijnbergen |8~ in a three-good (exported, imported and nontraded goods) specific factor (i.e., capital) model demonstrate that a higher tariff produces a substitution and an income effect, which work in opposite directions, on the relative price of the nontraded good and on the real exchange rate. The substitution effect, which draws resources out of the production of the nontraded good and increases its consumption, puts upward pressure on the relative price of the nontraded good. The income effect, which comes into play assuming initially a non-zero tariff, reduces welfare and expenditure, and puts downward pressure on the relative price of the nontraded good. As a result, the real exchange rate, defined as the domestic price index of the traded to the nontraded good prices, depreciates (may appreciate) if the income effect dominates (is smaller than) the substitution effect. As a result, the real exchange rate appreciates if the substitution effect dominates the income effect, that is, if initially the economy is not very distorted (e.g., a small departure from free trade). Similarly, a higher capital tax (i.e., d|Rho~ |is greater than~ 0), leads to a real exchange rate appreciation if the nontraded good is capital intensive, and the substitution effect dominates the income effect.

PROPOSITION 1. Consider a multi-commodity model of a small capital-importing country, and assume that (i) nontraded good markets are Walras stable, and (ii) goods are normal in consumption. Then, under a system of capital tax credits when |Rho~ |is less than~ |Rho~*, liberalization of capital movements, through a lower capital tax by the small capital-importing country, induces a real exchange rate depreciation. Under a system of capital tax credits when |Rho~ |is greater than~ |Rho~*, of capital tax deductions, or of untaxed repatriated capital earnings, the real exchange rate effect of a higher capital tax depends on the factor intensity of the nontraded goods. Specifically, in the case of a single nontraded good, the lower capital tax raises (may reduce) its relative price and causes an appreciation (possible depreciation) of the real exchange rate, if the nontraded good is non-capital (capital) intensive.

Substituting equation (11) into (9) shows that under a system of tax deductions, or of untaxed repatriated capital earnings, the effect of a higher capital tax on the real exchange rate depends, among other things, on the domestic capital tax rate but, it is independent of the foreign tax rate. Under a system of capital tax credits, the effect of a higher capital tax on the real exchange rate, in addition to the domestic tax rate, it also depends indirectly on the foreign capital tax.

III. A Model of a Small Capital-exporting Country

In this section, we establish sufficient conditions under which liberalization of capital movements, through a lower capital tax by a small capital-exporting country, leads to a real exchange rate appreciation, assuming that commodity trade is free, and that the rate of return to the internationally mobile capital is taxed according to the various rules of international taxation discussed in the previous section.

Under all three capital tax systems, the capital-exporting country taxes its capital uniformly whether it is invested at home or abroad.(12) When a unit of capital is invested domestically, it earns a net rate of return r(1 - |Rho~). When it is invested abroad, it earns a net rate of return |Rho~*(1 - max(|Rho~*, |Rho~)) under a system of capital tax credits, r*(1 - |Rho~*)(1 - |Rho~) under a system of capital tax deductions, and r*(1 - |Rho~*) under a system of untaxed repatriated capital earnings. Equilibrium in the world capital market under a system of capital tax credits requires that

r(1 - |Rho~) = r*(1 - max(|Rho~,|Rho~*)). (12)

Under a system of capital tax deductions, it requires that

r = r*(1 - |Rho~*), (13)

and under a system of untaxed repatriated capital earnings, it requires that

r(1 - |Rho~) = r*(1 - |Rho~*). (14)

Since the country is small in world markets, then under a system of capital tax credits when |Rho~ |is greater than~ |Rho~*, or under a system of capital tax deductions, equations (12) and (13) show that a higher capital tax does not affect the domestic rate of return to capital (i.e., dr = 0). Under a system of capital tax credits when |Rho~ |is less than~ |Rho~*, or when repatriated capital earnings are untaxed by the capital-exporting country, equations (13) and (14) show that changes in the capital tax affect the domestic after-tax rate of return to capital (i.e., dr = |r/(1 - |Rho~)~d|Rho~).

In equilibrium, the country's expenditure-income identity is

E(p, q, u) = R(p, q, K) - r*(1 - |Rho~*)K*, (15)

where K*(|is less than~ 0) is the country's capital exports, and r*(1 - |Rho~*)K* denotes the net payments to domestic capital invested in the foreign capital-importing country. Equilibrium in the country's nontraded good markets, requiring that the demand equals supply of nontraded goods, is again given by equation (7). Equation (9) captures the changes in the capital-exporting country's real exchange rate.

Using equation (5), the differentiation of equations (15) and (7) captures the changes in the country's utility, and relative prices of the nontraded goods due to changes in the domestic rate of return to capital. That is,

|Mathematical Expression Omitted~.

The determinant (|Delta~|prime~) of the left-hand side coefficient matrix is negative, if the nontraded good markets are Walras stable.

The Real Exchange Rate Effect of Liberalizing Capital Movements

Since dr = 0 under a system of capital tax credits when |Rho~ |is greater than~ |Rho~*, or under a system of capital tax deductions, a decrease in the domestic capital tax does not affect the gross rate of return to capital, its domestic use, and the country's repatriated net earnings of home capital invested abroad. Income and consumption in the capital-exporting country remain unchanged, and so do the relative prices of nontraded goods and the real exchange rate (i.e., |Mathematical Expression Omitted~).

Under a system of capital tax credits when |Rho~ |is less than~ |Rho~*, or under a system of untaxed repatriated capital earnings, dr = |r/(1 - |Rho~)~d|Rho~. Also, using the equilibrium condition for the world capital market (i.e., equations (12) and (14), respectively), r - r*(1 - |Rho~*) = r|Rho~. Substituting this result into equations (16), the effect of the lower capital tax, under these two capital tax systems, on the relative prices of the nontraded goods is given as follows:

|Mathematical Expression Omitted~.

The right-hand side of equation (17) is identical to that of equation (11) in the case of a small capital-importing country under a system of tax credits when |Rho~ |is greater than~ |Rho~*, or under the other two systems of international capital taxation. Similarly, the change in the country's real exchange rate due to the higher capital tax, is given after substituting (17) into (9). The example of a single nontraded good can be used again for interpreting the effect of a higher capital tax on the small capital-exporting country's real exchange rate.

PROPOSITION 2. Consider a multi-commodity model of a small capital-exporting country, and assume that (i) nontraded good markets are Walras stable, and (ii) goods are normal in consumption. Then, liberalization of capital movements, through a lower tax on repatriated capital earnings by the small capital-exporting country, has no effect on the real exchange rate under a system of capital tax credits when |Rho~ |is greater than~ |Rho~*, or under a system of capital tax deductions. Under a system of capital tax credits when |Rho~ |is less than~ |Rho~*, or of untaxed repatriated capital earnings, the real exchange rate effect of a lower capital tax depends on the factor intensity of the nontraded good. Specifically, in the case of a single nontraded good, a lower capital tax raises (may reduce) its relative price and causes an appreciation (possible depreciation) of the real exchange rate if the nontraded good is non-capital (capital) intensive.

Similar to the case of a small capital-importing country, the effect of a lower capital tax on the relative prices of nontraded goods and the real exchange rate depends indirectly on the foreign capital tax under a system of capital tax credits. It is independent of the foreign capital tax under a system of capital tax deductions, or of untaxed repatriated capital earnings.

IV. Conclusions

Prompted by recent developments in the world capital markets, a voluminous trade theoretic literature has examined the effects of capital taxes on key economic variables such as employment and welfare. This paper, in order to examine the real exchange rate effects of liberalizing capital movements, develops a many traded and nontraded goods general equilibrium model of a small capital-importing and a small capital-exporting country. Commodity trade is free, the supply of nontraded factors is fixed, and the supply of another factor, capital, is variable due to international mobility. The rate of return to capital is taxed according to a system of tax credits, of tax deductions, or of untaxed repatriated capital earnings. The analysis shows that whether liberalization of capital movements, through reduced capital taxes under the three tax systems, affects the real exchange rate depends on whether the country is a net capital-exporter or a net capital-importer, and on the prevailing capital tax system. For the cases where the lower capital tax affects the real exchange rate, the analysis establishes sufficient conditions leading to a real appreciation or real depreciation.

For a small capital-exporting country, liberalization of capital movements through a lower capital tax has no effect on the real exchange rate under a system of capital tax deductions, and of tax credits when the domestic capital tax exceeds the foreign tax rate. Assuming that (i) non-traded good markets are Walras stable, and (ii) all goods are normal, then a real exchange rate depreciation due to the lower capital tax occurs in the small capital-importing country when under a system of tax credits the domestic capital tax is smaller than the foreign tax rate. Under assumptions (i) and (ii), the lower capital tax may lead to a real exchange rate appreciation, depending on the factor intensity of the nontraded goods, under a system of untaxed repatriated capital earnings by either country, under a system of capital tax deductions by the small capital-importing country, and under a system of tax credits when the domestic capital tax by the small capital-importing (capital-exporting) country exceeds (is smaller than) the foreign tax rate. Under these capital tax systems, in the case of a single non-capital (capital) intensive nontraded good, a lower capital tax leads to an appreciation (possible depreciation) of the real exchange rate.

Finally, in either the small capital-importing or capital-exporting country, the effect of liberalizing capital movements through a lower capital tax, on the real exchange rate indirectly depends on the foreign capital tax under a system of capital tax credits. Under the other two systems of international capital taxation, the effect of a higher capital tax on the real exchange rate is independent of the foreign tax rate.

1. Out-(-in)ward foreign direct investment (FDI) has also surged dramatically to (from) the U.S. in recent years. In 1988, FDI into the U.S. was $57.00 billion compared to an average of $18.5 billion during 1980-85. Outward FDI from the U.S. was $54.00 billion in 1987 compared to an average of $10.00 billion during 1977-1984.

2. Greece and Portugal are exempted, if needed, to eliminate the restrictions on capital markets by the end of 1995.

3. It is a long standing policy of the U.S., on equity grounds, to avoid the double taxation of foreign earned income by U.S. residents, and, on efficiency grounds, to maintain "capital-export neutrality" which requires that capital taxes should not influence the country location of capital. Thus, U.S. residents are allowed to credit capital tax payments made in the foreign (i.e., host) country against their U.S. tax liability according to the fraction: ((foreign-source taxable income) / (worldwide taxable income)) (U.S. tax liability). The U.S., however, retains the right to tax foreign earned income if the foreign tax rate is lower than the U.S. tax rate |18~.

4. Slemrod |17~ offers two alternative explanations for his second result. Either inaccurate data which cannot pick-up important differences in FDI behavior that do in fact exist, or sophisticated financial strategies by firms in foreign tax credit capital-exporting countries, through which they can defer indefinitely the home country taxation of their repatriated capital earnings.

5. These and other authors make explicit reference to either one or to both results.

6. Hazari et al. |12~ using a Harris-Todaro model of a small developing economy, and Edwards and Ostry |9~ using a model with taxes on international lending and borrowing, also demonstrate that the conventional results may not always hold.

7. Capital taxes imposed by a capital-importing country to capture capital tax revenue otherwise collected by the source capital-exporting country are referred to as "soak-up" taxes |3~.

8. In models with international capital mobility, free trade may lead to complete specialization in production and trade. Strict concavity of the GDP function with respect to K is assumed in order to avoid this problem.

9. In the literature of international finance, the portfolio balance (PB) approach for the determination of the exchange rate has widely considered the interaction between capital flows and the real exchange rate in multi-period models through adjustments in assets and product markets, and through the intertemporal allocation of savings, investment, and labor supply |1; 10~. The present model is consistent with the long-run equilibrium of the PB models, where wealth adjustments have been completed through assets (de)accumulation, the real exchange rate has moved to its long-run equilibrium level, and the current account is balanced.

10. If the right-hand side of equation (7) is replaced by |Mathematical Expression Omitted~, the vector of excess demands for nontraded goods, then the sign of |Mathematical Expression Omitted~ is derived by combining equations (6) and the modified (7). Walrasian stability states that |Mathematical Expression Omitted~ must be a negative semi-definite matrix, which requires |Delta~ to be negative.

11. We define the nontraded good to be capital (non-capital) intensive if an inflow of capital increases (decreases) its production, i.e., |R.sub.qk~ |is greater than~ 0(|is less than~ 0).

12. This implies that the capital-exporting country can fully tax the foreign-source income of domestic capital invested abroad. Razin and Sadka |16~ examine the optimum rules of taxing capital in a capital-exporting country in the presence of free international capital mobility and difficulties in taxing foreign-source capital income.

References

1. Allen, Polly and Peter Kenen. Assets Markets and Exchange Rates: Modeling an Open Economy. Cambridge: Cambridge University Press, 1980.

2. Balassa, Bela. Development in Semi-industrial Economies. Oxford: Oxford University Press, 1982.

3. Bond, Eric, "Optimal Tax and Tariff Policies with Tax Credits." Journal of International Economics, May 1991, 317-29.

4.----- and Larry Samuelson, "Strategic Behavior and the Rules of International Taxation of Capital." Economic Journal, December 1989, 1099-11.

5. Clague, Christopher, "Tariffs, Transfers and the Real Exchange Rate." Southern Economic Journal, July 1986, 155-69.

6. Diaz-Alejandro, Carlos. "Exchange Rate and Terms of Trade in the Argentine Republic: 1913-76," in Trade, Stability, Technology and Equity in Latin America, edited by M. Syrquin and S. Teitel. New York: Academic Press, 1983.

7. Dornbusch, Rudinger. Open Economy Macroeconomics. New York: Basic Books, 1980.

8. Edwards, Sebastian and Swender van Wijnbergen, "Tariffs, the Real Exchange Rate and the Terms of Trade: On Two Popular Propositions in International Economics." Oxford Economic Papers, September 1987, 458-64.

9. Edwards, Sebastian and Jonathan Ostry, "Terms of Trade Disturbances, Real Exchange Rates, and Welfare: The Role of Capital Controls and Labor Market Distortions." Oxford Economic Papers, January 1992, 20-34.

10. Hallwood, Paul and Ronald MacDonald. International Money: Theory, Evidence and Institutions. Oxford: Blackwell, 1986.

11. Hatzipanayotou, Panos and Michael Michael. "Variable Labor Supply, Employment, and Welfare Under Various Rules of International Taxation of Capital." Working Paper No. 93-123, The University of Connecticut, March 1993.

12. Hazari, Bharat, Sisira Jayasuriya and Pasquale Sgro, "Tariffs, Terms of Trade, Unemployment and the Real Exchange Rate." Southern Economic Journal, January 1992, 721-31.

13. Johnson, Harry, "A Model of Protection and the Real Exchange Rate," Economica, May 1969, 119-38.

14. Neary, Peter, "Determinants of the Equilibrium Real Exchange Rate." American Economic Review, March 1988, 210-15.

15. Organization for Economic Co-Operation and Development. Code of Liberalization of Capital Movements. Paris, 1982.

16. Razin, Assaf and Efraim Sadka, "Efficient Investment Incentives in the Presence of Capital Flight." Journal of International Economics, August 1991, 171-81.

17. Slemrod, Joel. "Tax Effects on Foreign Direct Investment in the United States: Evidence from a Cross-country Comparison," in Taxation in the Global Economy, edited by A. Razin and J. Slemrod. Chicago: The University of Chicago Press, 1990.

18. U.S. Congress. Tax Reform Act of 1986. Washington, D.C., 1987.