Economic theory suggests that high tax rates on income from capital are likely to discourage capital formation and that differences in tax burdens across countries can be expected to induce an inefficient allocation of capital. As discussed in Chapter III, this issue is particularly relevant to the EC because the drive to remove all barriers to the free flow of goods, labor, and capital under the Single European Act will fully expose investment decisions to differences in tax burdens across member countries.

Economic theory suggests that high tax rates on income from capital are likely to discourage capital formation and that differences in tax burdens across countries can be expected to induce an inefficient allocation of capital. As discussed in Chapter III, this issue is particularly relevant to the EC because the drive to remove all barriers to the free flow of goods, labor, and capital under the Single European Act will fully expose investment decisions to differences in tax burdens across member countries.

Attempts to quantify the importance of these distortions have been complicated by the intricacy of modern tax systems, wherein the effective marginal tax rate on income from capital depends in a complex way on the statutory tax rates, depreciation allowances, and other rules underlying the computation of the tax base, as well as on macroeconomic variables such as the rate of interest and the expected rate of inflation. Earlier studies used average tax rates computed from actual tax collections as approximations for the theoretically more relevant marginal tax rates.1 More recently, following a study by King and Fullerton (1984), researchers have constructed standardized measures of the marginal tax burden on capital directly from the tax code.2 In this approach, effective tax rates are based, more or less explicitly, on the formula for the user cost of capital derived by Jorgenson (1967) from the neoclassical theory of the firm. The user cost of capital provides a comprehensive measure of the real marginal cost of capital to the firm, inclusive of taxes and economic depreciation. By subtracting depreciation and the rate of return on the underlying financial asset from the user cost, one obtains a measure of the effective tax on the marginal unit of capital in the form of a wedge—driven by taxes—between the net marginal product of capital and the return to the owners of the firm. The advantage of this approach is that it allows comparisons of tax wedges across countries, sectors, or hypothetical changes in the tax code.

The first section of this chapter discusses the conditions under which differences in corporate taxation (as opposed to personal taxation) of income from capital provide a sufficient description of potential tax distortions in the allocation of real assets. In the second section, we construct measures of effective tax burdens on corporate investment income for the 12 EC member countries under the tax rules in effect in 1990 (including preannounced changes) and five alternative harmonization proposals. The dispersion of effective tax rates across countries serves to illustrate the potential allocative distortions of the tax systems. The analysis is based on the assumption that international capital movements take the form of portfolio flows; under certain conditions, however, the same results would apply to the case of direct investment.

In the third section of the chapter, the estimated effective tax rates are used in the context of a simplified general equilibrium model to assess the long-run allocative effects of the different tax harmonization scenarios. Given the simplistic assumptions underlying the model, the exercise serves merely to indicate the direction and broad order of magnitude of the effects of the tax changes under scrutiny. Concluding remarks are contained in the last section of the chapter. Differences in the taxation of income from capital also distort saving behavior and affect the location of financial intermediation, but these issues remain outside the scope of this analysis.

Taxation of Capital in an Open Economy

Corporate and Personal Tax Wedges

Taxes on the income from capital drive a wedge between the gross rate of return on real assets (p) and the net rate of return received by households on their financial claims on those assets (s).3 The distortionary effects of taxation can, in turn, be related to this wedge: its size affects the degree of capital accumulation,4 whereas differences across countries distort the international allocation of capital and of savings as well as the location of financial intermediation.

Because income from capital is taxed at both the corporate and personal levels—with some degree of integration provided in certain countries—the total wedge can be broken down into two components: the corporate tax wedge, measured as the difference between the before-tax rate of return on the real asset (p) and the market rate of return on the underlying financial asset (r); and the personal tax wedge, measured as the difference between the market rate (r) and the after-tax rate of return on the financial asset (s) from the point of view of the final investor.

The decomposition of the overall tax wedge on investment income into a corporate and a personal component is analytically convenient in addressing allocative issues among small open economies when international capital movements take the form of portfolio flows—that is, transactions in foreign financial assets normally not involving controlling ownership. The integration of financial markets implies that r, although affected by the average level of taxation, is determined in world markets. The personal and corporate tax wedges of each country will then have separate and specific effects only on the after-tax rate of return on domestic financial assets (s) and on the gross rate of return on domestic real assets (p), respectively. A tax on capital income at the personal level reduces s and only affects saving behavior. Differences in personal tax wedges across countries thus induce an inefficient allocation of saving and distort the pattern of ownership of capital. A tax on capital income at the corporate level raises, instead, the level of p necessary to cover r, the cost of financing, and thereby reduces the size of the desired capital stock. Differential rates of taxation of capital income at the company level thus prevent equalization of the marginal rate of return from capital and induce an inefficient allocation of capital.5

The integration of personal and corporate tax systems, intended to alleviate the problem of double taxation, reduces the size of the overall wedge (ps) but also complicates the breakdown into its two components. The reduction in the size of the overall wedge can either translate into a rise in s or a reduction in p, with very different implications for saving and investment. In a small open economy, the effect depends on the mode of application of integration. If the method of integration extends to foreign shareholders, it will effectively lower the required rate of return on the underlying financial asset, and the benefit will be entirely captured by domestic corporations in the form of a lower cost of capital (p). If the method of integration does not reach the foreign shareholder, then integration of the two tax systems has no effect on the rate of return on the financial asset, and the benefit is fully captured by resident households in the form of a higher after-tax return (s) (Boadway and Bruce (1988)).

Implications for the EC

It was argued above that, with integrated financial markets and small open economies, the source of tax distortions to the allocation of capital can be reduced to differences in corporate tax wedges.6 The specific conditions under which this simplification holds for the EC are examined herein. The assumption of integrated financial markets appears to be broadly consistent with the abolition of capital controls and is, in any case, consistent with the goal of the Single European Act.

The small-economy assumption is clearly more questionable, since it cannot be supposed that all EC member countries are price takers in capital markets. In larger countries, domestic demand and supply conditions could affect the rate of return on financial assets, and the personal tax system could, consequently, distort the relative cost of funding investment in domestic and foreign markets. Under these circumstances, the potential for allocative distortions could not be traced to differences in corporate tax wedges alone.

The integration of credit markets in the EC provides an alternative condition that ensures the irrelevance of personal income taxation to the allocation of real capital. If domestic firms have access to foreign and offshore financial markets on the same terms as foreign enterprises, they can effectively avoid absorbing the gross-up effect of local personal and withholding taxes on financing costs. The lowest-cost financial market would then become the marginal source of financing for all enterprises—the Eurobond market being a case in point—and taxes at the personal level would affect inframarginal savers and investors as well as the location of intermediation.

Access to foreign and offshore markets cannot, however, neutralize the allocative distortions of taxes on financial investment income when the tax treatment of domestic assets is more favorable than that of offshore or foreign assets. Such is the situation of dividend taxation when integration of personal and corporate tax systems is limited to resident shareholders, locally earned profits, or both.7 In this case, through integration, the larger economies can effectively reduce the cost of equity financing of domestic investments. The problem cannot be easily resolved analytically, and the small-economy assumption is retained in the calculations below; that is, any form of integration that does not reach the international investor is assumed to be passed on to domestic households in the form of higher after-tax returns, rather than to firms in the form of lower financing costs.

Another condition under which taxes on financial investment income would not affect the allocation of capital even among large countries is if the residence principle of taxation were enforced.8 Under this principle, taxes do not discriminate among assets but only on the basis of the residence of the investor. This guarantees the convergence of the rates of return on all financial assets and, hence, of the cost of capital to enterprises residing in different countries.

Although investors in all EC countries are, in principle, taxed on this basis, the residence principle fails to hold for several reasons. First, as discussed in Chapter III, uneven enforcement implies that financial investment income is often taxed at source only, and differences in source taxation can therefore be reflected in asset prices. A second reason for the failure of the residence principle is that foreign assets may be subject to a heavier tax burden than domestic assets. This possibility arises when foreign withholding taxes are not fully creditable in the country of residence, a problem that applies mostly to tax-exempt institutions. A third reason is that, with differences in inflation rates and compensatory exchange rate adjustments, the effective taxation of foreign and domestic assets will diverge, even under the residence principle, if exchange rate gains and losses are taxed at a different rate than that for ordinary interest income, or are simply taxed on a realized rather than accrual basis. Finally, the integration of personal and corporate income taxes, when available at the national level, is usually not extended to nonresidents or to domestic recipients of foreign-source dividends, thus creating another form of discrimination among assets, as discussed above.

To account for these complications, the allocative issue could be examined by constructing a matrix of tax wedges (or effective tax rates), inclusive of both personal and corporate taxes, associated with investment flows to and from each of the 12 EC member countries. This approach was adopted by Bovenberg and others (1990), who construct bilateral tax wedges for portfolio investment flows between the United States and Japan. The inclusion of personal income taxation in the tax wedges may, however, be misleading for our purposes. First, widespread tax evasion at the personal level, the proliferation of tax avoidance schemes (such as pension accounts and nondistributing mutual funds), and the uneven tax treatment of individuals and institutional investors make it difficult to construct comparable effective rates of personal taxation. Second, as discussed above, the availability of offshore financing limits the adverse effect of personal taxes and withholding taxes on the cost of capital at the margin.

Our analysis supposes that foreign investment takes the form of portfolio flows, but under specific circumstances the results would extend to investment channeled through financial intermediaries or in the form of corporate direct investment. In the case of foreign investment undertaken by financial institutions, the same allocative implications would obtain if these institutions were competitive and if intermediated capital income were subject to the same tax treatment as personal portfolio income.9 The analysis of foreign direct investment can be subsumed by the present analysis if foreign direct investment income is taxed only in the source country. This effectively requires that the country of residence not exercise any claim on that income. Although practices vary widely, there are two cases in which this condition is met: first, if foreign source income is exempt outright in the country of residence and, second, if taxes in the country of residence can be postponed indefinitely by deferring the repatriation of foreign profits. The first condition is applied less frequently than the second, which normally can be used for subsidiary income. In general, however, the complexity of tax practices with regard to foreign direct investment income would require a separate analysis. Again, a matrix of effective tax rates would have to be constructed for flows to and from each of the EC member countries.10

Corporate Tax Wedges


The methodology used in this exercise is based on the concept of the required rate of return or user cost of capital.11 A profit-maximizing firm requires a gross return on the last unit invested that allows the firm to pay market returns on internal and borrowed funds after covering depreciation and corporate taxes. (A detailed derivation is presented in the appendix to this chapter.) By subtracting depreciation from the user cost (divided by the price of capital goods), a measure of the before-tax rate of return on capital is obtained:


where u denotes the statutory corporate income tax rate, z is the present value of the depreciation allowance, k is the present value of investment grants, ρ is the firm’s marginal cost of capital, δ is the rate of economic depreciation, and π is the expected rate of inflation. The discount factor (p) depends in turn on the market rate of return on the underlying financial instrument.

If we denote by r the market rate of return on the financial instruments sold by the firm to raise capital, the corporate wedge can be defined as


This expression can be converted into an effective marginal corporate income tax rate, expressed in terms of the before-tax rate of return:


In practice, there are as many “market” rates as there are sources of financing available to the firm. With debt and equity as the two broad forms of financing, a single composite market rate reflecting the financing mix of the enterprise can be constructed, and a single wedge derived from it. The alternative, followed here, is to compute separate tax wedges for debt- and equity-financed projects and then calculate a weighted average.12

In the case of debt (subscript d), the appropriate measure of the real market return is clearly the nominal interest rate less the expected rate of inflation:


Because interest payments are deductible as expenses for tax purposes, the (nominal) marginal cost of funds is given by


and the company wedge on a debt-financed investment project is given by


For equity (subscript e), the computation is complicated because the required market return (re) cannot be directly observed. A convenient way to derive re is to impose an arbitrage condition requiring that the net, risk-adjusted returns on debt and equity be equal from the perspective of a representative investor who is willing to hold both instruments.13 This condition can be expressed as


where h is an exogenous risk premium. Consistent with the assumption of an internationally integrated capital market discussed in the previous part of this section, the arbitrage condition (7) must hold for the same representative investor for the whole of the EC. To make matters tractable, the representative investor is chosen to be an international (institutional) investor with a nonresident status in each EC member country. Because nonresident investors are usually exempt from withholding taxes on interest income, the net return from a debt instrument is then simply the real interest rate, rd, as in equation (4).14

We can now work back from re to determine the marginal cost of equity finance to the firm (ρe). We assume that a fixed fraction (v) of the firm”s real yield on equity (ρe − π) is distributed as dividends.15 The real return paid out by the firm (ρe − π) and that received by the representative investor (re) can differ for several reasons. First, dividends paid to nonresidents are usually subject to a withholding tax (wt) (Table 21 in Chapter III).16 As a result, part of the firm”s payout does not reach the investor, and for a given required return re and payout ratio v, this tends to raise the firm”s cost of funding. Second, because of partial or full integration of the personal and corporate tax systems, dividends may receive a preferential tax treatment over retained earnings in the form of a dividend tax credit (or deduction) or a split rate system (Table 19 in Chapter III).17 In that case, each dollar distributed by the firm may be worth more than one dollar to the investor. This effect would tend to reduce the cost of equity financing. The degree of integration is measured by the integration variable θ, defined as the opportunity cost of retained earnings in terms of gross dividends forgone. Finally, the undistributed portion of earnings is presumably capitalized in the price of the stock and can be taxed, in principle, at the investor’s level through a tax on capital gains. In practice, such taxes are virtually nil, either by statutory treatment or because they are levied on a realized rather than accrual basis. Considering all these factors, the firm’s marginal cost (ρe) of providing the required return to the marginal shareholder can be expressed as


The before-tax rate of return on equity-financed capital (ρe) can then be computed by substituting equation (8) into equation (1), and the tax wedge on equity is then derived as


where re is defined as (i + h − ρe).

The derived tax wedges provide a comprehensive measure of the effective tax burden on capital income but cannot, obviously, capture differences in the degree of enforcement of tax collection and in the scope for tax avoidance through financial transactions across member countries. Moreover, the tax wedges cannot account for differences in the tax treatment of losses.


Tax wedges can be measured to illustrate differences in the effective taxation of income from capital under current systems and alternative scenarios. The degree of dispersion of tax wedges reflects the potential for distortions in the allocation of capital across countries, by type of asset and by form of financing. Corporate tax wedges for domestic investment have been calculated under six scenarios for all EC member countries. Under each scenario, tax wedges were computed separately for investment in buildings and machinery, financed with either debt or equity. The real interest rate is assumed to be constant, at 5 percent, in combination with two alternative inflation rate assumptions:18 a common inflation rate of 2 percent and different inflation rates for three country groups (2 percent for Belgium, France, Germany, Ireland, Luxembourg, and the Netherlands; 5 percent for Denmark, Italy, Spain, and the United Kingdom; and 10 percent for Greece and Portugal).19

The six tax scenarios are summarized in Table 23. Scenario 1 is based on tax systems effective in 1990, qualified by proposed tax reforms (see Table 19).20 Scenario 2 assumes adoption of the following rules for the determination of taxable profits of enterprises by EC member countries:21 full first-year convention for depreciation, allowing enterprises to claim the full amount of depreciation the first tax year, irrespective of when in the year the investment actually takes place; reduction of the depreciable base by the amount of the subsidy received through investment tax credits and deductions; elimination of accelerated depreciation; elimination of depreciation of capital not yet in use (advance depreciation); elimination of indexation of the depreciable base; and straight-line or declining-balance methods of depreciation (assumed at 2.5 times the existing straight-line rate) allowed for both buildings and machinery, with switchover from declining balance to straight-line depreciation during the life of the asset (Table 24).22 Scenario 3 includes equalization of corporate income tax rates at the weighted EC average rate of 43 percent and elimination of local income taxes,23 in addition to the assumptions under scenario 2 (Table 25). Scenario 4 includes the elimination of taxes levied on the value of assets or net worth in France, Luxembourg, and Germany,24 in addition to assumptions under scenario 3. It also assumes the adoption of a common imputation system, consisting of credit on dividends equivalent to 50 percent of the corporate income tax extended to both residents and nonresidents, and a common 15 percent with-holding tax on dividends paid to nonresidents (Table 25); only differences in depreciation rates and investment grants remain. Scenario 5 assumes the equalization of tax rates and imputation systems and the elimination of capital-based taxes and investment grants but maintains current differences in depreciation rates and in the definition of the depreciable base. Scenario 6 assumes complete equalization of company income tax systems; only inflation rates differ among the three groups of countries.

Table 23.

Summary of Corporate Tax Harmonization Scenarios

article image
Table 24.

Corporate Tax Base Harmonization

article image
Note: Scenario 2 relative to current systems (scenario I; see Chapter III, Table 19); + = increase in effective tax rate, — = decrease in effective tax rate, … = no effect or no change relative to current practice.

DB = declining balance; SL = straight line.

In the SL case, Denmark is assumed to move from a two-rate (0.06, 0.02) system to a single rate of 0.05 over the life of the asset.

The rate of 0.10 is in lieu of a system of three rates (0.10, 0.05, 0.02) over the life of the asset.

Table 25.

Corporate Tax Rate Harmonization

article image
Note: Harmonization of corporate tax rates (scenarios 3–6) and of dividend withholding and dividend credit rates (scenarios 4–6) relative to current systems (scenario I; see Chapter III, Table 19); — = decrease in effective tax rate; … = no effect or no change relative to current practice.

To common mrate of 43 percent.

The payout rate is defined as the share of profits reaching the shareholder after corporate and dividend taxation (inclusive of dividend credit); the common rate of 73 percent is derived assuming a common 15 percent withholding tax rate and a 50 percent dividend tax credit.

National and local income taxes combined.

For each scenario, 48 different tax wedges (12 countries, 2 types of assets, 2 sources of finance) were computed. Effective tax rates, calculated from the average wedges over both sources of finance and asset types, provide a normalized measure of the overall tax burden on capital income in each country (Tables 26 and 27).25 Standard deviations capture the degree of dispersion in tax burdens across countries. Differences in the tax wedges across sources of financing (Table 28) and asset types (Table 29) reflect the biases of the tax systems.

Table 26.

Average Corporate Tax Wedges

(Percentage-point difference between gross and net-of-tax real rates of return)

article image
Note: Weighted average over buildings (0.4) and machinery (0.6) and over debt (0.4) and equity (0.6).

Weights correspond to countries’ shares of EC capital stock.

Table 27.

Effective Corporate Tax Rates

(In percent of after-tax rate of return)

article image
Note: Weighted average over buildings (0.4) and machinery (0.6) and over debt (0.4) and equity (0.6).

Weights correspond to countries’ shares of EC capital stock.

Table 28.

Corporate Tax Wedges by Source of Financing

(Percentage-point difference between gross and net-of-tax real rates of return)

article image
Note: Weighted average over buildings (0.4) and machinery (0.6).

Weights correspond to countries’ shares of EC capital stock.

Table 29.

Corporate Tax Wedges by Type of Asset

(Percentage-point difference between gross and net-of-tax real rates of return)

article image
Note: Weighted average over equity (0.6) and debt (0.4); common 2 percent annual inflation rate.

Weights correspond to countries’ shares of EC capital stock.


Under the tax practices that are likely to prevail in the absence of any concerted harmonization, on average, Germany appears to be the country with the highest corporate tax burden (Tables 26 and 27). In the low-tax range, starting from the lowest wedge, are Ireland and Luxembourg, followed by Denmark, Belgium, and the United Kingdom (the ordering of the last three depends on the underlying assumption about inflation). The other countries’ wedges all lie well within one standard deviation of the mean. The ordering of countries according to the average wedge is quite sensitive to the weights assigned to the two types of assets (Table 29) and sources of financing (Table 28).

As also can be seen from Tables 28 and 29, tax systems discriminate in favor of machinery and in favor of debt financing. The relative tax advantage accorded to investment in machinery reflects more generous tax depreciation allowances, although the results also depend strongly on the assumed rates of economic depreciation.26 On the financing side, the corporate tax systems uniformly result in a subsidy at the margin for debt-financed investments. In large part the subsidy comes from the deductibility of interest payments from the corporate tax, which lowers the discount rate below the market rate of interest. Inflation raises the advantage of debt financing because of the deductibility of nominal rather than real interest payments. In contrast, inflation reduces the present value of depreciation allowances, based on historical cost, under both debt and equity financing, except for Denmark where the depreciable base is indexed for changes in the price level.27 The difference between the debt and equity cases also depends on the chosen arbitrage assumption, which excludes any effect on the cost of capital of integration systems that are not extended to nonresidents. Hence, the wedge under equity financing—and thus the distortion in favor of debt financing—would be smaller for Germany and Italy under a purely domestic arbitrage assumption. The average wedges do not change significantly under alternative assumptions about inflation. In countries with a higher inflation rate, the advantages of the increased nominal interest deductibility appear to be offset by the reduced real value of depreciation allowances, although the distortion in favor of debt increases.

Under tax base harmonization (scenario 2), most country wedges would fall, in large part because of more liberal depreciation allowances—specifically, more than double declining balance for buildings and full first-year convention for all assets. Only Denmark, because of the elimination of advance depreciation and of indexation of the depreciable base, and Luxembourg, because of the reduction of the depreciable base by the value of the investment tax credit, would experience a rise in their average wedges. However, base harmonization would on balance contribute minimally to the convergence of effective tax rates, as reflected in the virtually unchanged standard deviation of country wedges. Germany would remain the highest-tax country; Ireland would remain the lowest-tax country, followed by the United Kingdom, Luxembourg, and France.

The equalization of income tax rates at 43 percent (scenario 3)—superimposed on base harmonization but without equalizing depreciation rates—produces a more significant contribution to the convergence of country wedges, yielding a reduction of the weighted standard deviation from 1.1 in scenario 1 to 0.7. The remaining dispersion is due to country differences with respect to depreciation rates and, more significant, to the degree of integration, investment tax credits, and wealth and net worth taxes. Italy replaces Germany as the highest-tax country, and the United Kingdom, Luxembourg, and France assume low-tax positions.

The added elimination of wealth and net worth taxes and the equalization of the method and degree of integration (scenario 4) cause a significant drop in both the average wedge and in the standard deviation. The decrease in the average wedge is a direct result of the arbitrage assumption, whereby the extension of the dividend credit to nonresidents is entirely passed on to firms in the form of a reduction in the cost of equity financing. Italy and Luxembourg appear as the two outlying countries at the high and low ends of the spectrum, respectively.

The harmonization of tax rates and methods of integration—superimposed, however, on differing definitions of the tax base (scenario 5)—produces slightly less convergence of tax wedges than the previous case (the standard deviation falls to 0.5 percent from 0.3 percent in scenario 4). Moreover, the benefits arising from greater transparency of the tax systems and lower compliance costs for enterprises operating throughout the EC would be lost as long as national definitions of the tax base continue to vary as they do now.

The distortionary effect of inflation differentials is illustrated with full harmonization of corporate tax systems. As mentioned above, inflation non-neutralities arising from the deductibility of nominal interest payments and from historical-cost depreciation allowances tend to work in opposite directions, and the resultant net effect of inflation on tax wedges is nonlinear. The tax wedge rises from 0.4 to 0.7 as the inflation rate is increased from 2 percent to 5 percent, but it falls to 0.6 as inflation is increased from 5 percent to 10 percent.

In general, the results do not change markedly between the two inflation variants of the simulations, suggesting that the harmonization measures considered do not exacerbate the inflation non-neutralities inherent in the corporate tax systems. Also, the standard deviation of wedges under current or proposed systems (scenario 1) is virtually the same under the two inflationary assumptions, indicating that present differences in tax systems do not, in an average sense, compensate for or enhance the distortionary effects of inflation differentials on effective tax rates.

Simulation of Allocative Effects

The Model

A simple computable general-equilibrium model is derived in this section to explore the potential effects of differential effective tax burdens and of changes in tax systems on the allocation of capital in the EC.28 The model is based on fixed and immobile labor endowments and profit-maximizing competitive firms and uses the wedge calculations described above to simulate the allocation of a perfectly mobile capital stock under various harmonization scenarios. Simulations are conducted under two different assumptions about the supply of capital. The first assumption is that of a fixed but mobile capital stock within the EC. This assumption allows us to isolate and quantify, albeit in a very simplified way, the purely allocative and efficiency implications of harmonization, leaving aside the effects of tax changes on the total capital stock.29 In this framework, the interest rate—common to the whole EC—adjusts endogenously to satisfy capital market equilibrium.30 Within each country the wage rate adjusts to clear the labor market, and the level of output of each country is endogenously determined. Given fixed factor supplies, changes in the level of output of the EC as a whole provide a convenient measure of the efficiency gains and losses of alternative scenarios in comparative static terms.

The second assumption is that capital is fixed but mobile worldwide—the world consisting of the EC, Japan, and the United States. In this context, changes in effective company tax rates can alter both the total EC stock of capital and its allocation among member countries. The two effects cannot be isolated, but this exercise provides a gauge of the pressures placed on the rest of the world by changes in the level of taxation in the EC.

Within each country we consider a representative firm that operates under competitive conditions. There is a single good produced under constant returns to scale and with production function F(K, L), where K and L are the capital and labor inputs. The representative firm of country i maximizes profits


where gi, is the gross real wage (inclusive of any tax on labor use or income) and ci is the user cost of capital,31 where


r is the real market rate of return on the underlying financial instrument, δ is the rate of economic depreciation, wi is the corporate tax wedge, and q is the real price of capital goods in country i. The model is solved by using a Cobb-Douglas production function:32


with α estimated from labor’s share in GDP and the scale parameter A set to reproduce the real interest rate assumed to prevail under current tax systems (scenario 1).33 A capital demand function can then be explicitly derived from profit maximization:


At the aggregate level, the model is closed by requiring that factor and goods markets clear. With fixed and immobile labor endowments (Ni) in each country, and fixed but mobile global supply of capital (S), the factor market equilibrium conditions reduce to


Given a fixed supply of capital, whether in the EC or in the world, the goods market clears implicitly in the absence of net saving or capital accumulation.

The equilibrium values of the interest rate and each country’s wage rate, capital stock, and output level can then be expressed as functions of factor endowments and tax parameters. The country corporate tax wedges presented earlier—averaged over the two types of financing and assets—are substituted for wi in equation (11) and then in equation (13), so as to solve the model, subject to conditions (14) and (15).34

By construction, a uniform change in wedges (Δwi = Δw) has no effect on the allocation of the capital stock across countries (ΔKi= 0), and capital market equilibrium obtains through a one-for-one offsetting change in the interest rate (Δr = −Δw) so as to leave the cost of capital unchanged (Δc = 0). Only relative changes in tax wedges have real effects. Specifically, the relative change of a country’s capital stock attributable to a change in its tax wedge is inversely related to the country’s size. A tax rate increase in a large country will reduce significantly both the world demand for capital and the world interest rate. The fall in the interest rate compensates partly for the rise in the cost of capital due to the tax change, with a consequently smaller proportional decline in the country’s equilibrium capital stock. Meanwhile, the other countries benefit in terms of lower interest rates and higher capital stocks. In the case of a small country, the induced effect of a tax rate increase on the world demand for capital—and, by implication, on the interest rate—is negligible. The loss of own capital is proportionately larger, and the spillover effect on other countries is insignificant.

There are several inherent limitations in this exercise. First, the magnitude of the allocative effects of differential tax burdens depends entirely on the assumed technological parameters. Although the assumed form and the parameter values of the production function are broadly consistent with available estimates, more sophisticated specifications could alter the magnitude of the effects. Second, the assumption of a fixed capital stock ignores entirely the effect of taxation on capital formation, savings, and growth. Thus, the results obtained are of a static general equilibrium nature. Third, the analysis does not provide an explicit treatment of the government sector; in particular, it is assumed that budgetary measures compensating for changes in corporate taxation—which account for a relatively small share of general government revenue in most EC countries—do not affect the production function or relative labor costs.35 Fourth, international capital flows take the form of portfolio claims on foreign capital, and only corporate taxation matters—a reasonable assumption in view of the large role of institutional investors and widely held corporations. Finally, the absence of country-specific risk is being approximated by the creation of a single EC market.


The first column of Table 30 shows the equilibrium solution after the equalization of after-tax rates of return on capital. The resulting capital stocks and corresponding changes in potential output are shown relative to their estimated current (stylized “autarky”) values. Both are necessary as a benchmark for the harmonization scenarios.36 The reallocation of capital predicted by the model is considerable, with Portugal and Ireland more than doubling their capital stocks and Germany losing nearly 40 percent in the new steady state. The allocative results of the model are driven by the assumption that corporate taxes confer no associated benefits and by the fact that taxes and factor proportions are the only determinants of differences in the rates of return on capital across countries. If taxes were benefit charges, tax rate differentials would have no allocative effects. Moreover, if appropriate consideration were given to all other nontax factors affecting returns, greater uniformity in after-tax rates of return adjusted for such factors would be found in the “autarky” case.

Table 30.

Allocative Effects of Corporate Tax Harmonization

article image
Note: Common 2 percent annual inflation rate.

Weights are calculated from the capital stock distributions for each scenario. Average wedges differ from those of Table 26 because the weights and the interest rate are recalculated for each equilibrium solution.

Alternatively, the result may be explained by the fact that capital has not yet fully adjusted to equalize expected rates of return, in which case capital tends to move toward countries with a higher than average after-tax rate of return. Higher than average after-tax rates of return may be the result, other things being equal, of either lower than average taxes—in which case the reallocation of capital would cause an efficiency loss—or low capital-labor ratios—in which case an efficiency gain would ensue. The fact that overall EC output rises by nearly 2 percent indicates, under this interpretation, that the efficiency gain from the equalization of capital-labor ratios would outweigh the efficiency loss from tax-induced distortions. An additional explanation for cross-country differences in the deviations shown in the first column is the likely measurement error inherent in the underlying factor proportions.

The allocative effects of harmonization are measured relative to the equilibrium results of column 1—that is, after equalization of after-tax rates of return. The indices thus reflect solely the effect of changes in effective tax rates under selected harmonization scenarios. Because factor supplies are assumed to be fixed, albeit mobile in the case of capital, the aggregate output index provides a static measure of the efficiency gain from harmonization. As discussed earlier, harmonization of the tax base (scenario 2) does not significantly reduce the dispersion of effective tax rates and consequently produces no efficiency gain; that is, the aggregate output index remains unchanged. In the EC context (the EC isolated from the rest of the world), the reallocation of capital reflects the change in the country ordering, with the largest proportional loss for Denmark and Luxembourg (8 percent) and the largest proportional gain for Greece, France, and Portugal (nearly 4 percent). Luxembourg, Belgium, Italy, Ireland, and Spain would also lose capital. The reduction in the dispersion of effective tax rates under the added harmonization of statutory income tax rates (scenario 3) would not be sufficient to produce a noticeable change in aggregate output. Tax rate harmonization would produce a relatively large gain in capital for Germany (9 percent), followed by France (3 percent), at the expense of all other countries; at the other end of the spectrum, Ireland would experience the largest proportional loss (16 percent). A small gain in overall EC output (0.1 percent) appears under complete harmonization (scenario 6). Relative to base and rate harmonization (scenario 3), most changes in capital stock would result on account of the elimination of investment tax credits (Luxembourg and Spain), the elimination of capital and net worth taxes (France, Germany, and Luxembourg), and the reduced cost of equity financing attributable to integration. This last change would affect all countries except France and the United Kingdom, where integration already applies under current conditions and the chosen arbitrage assumption.

By construction, the average wedge declines as harmonization is broadened from base to rates, with a consequent rise in the demand for capital. In the EC context, capital market equilibrium is restored through a rise in the interest rate. In a “worldwide” context, however, the rise in the interest rate spills over to the United States and Japan, forcing a reallocation of capital towards the EC. The net inflow of capital to the EC reduces the size of the capital losses—which in some instances turn into a gain—and raises the magnitude of the gains of individual countries.

This effect derives purely from the specific (average) effective tax rate toward which the EC would converge. The large loss of capital for Japan and the United States from complete harmonization (scenario 6), relative to base and rate harmonization (scenario 3), can be explained largely by the effect of integration on the cost of capital in the EC under the chosen arbitrage assumption. By assumption, the extension of corporate-personal tax integration to all EC residents (but not necessarily to non-EC residents) is fully passed on to the firm. How much the extension of integration to all EC residents can in practice contribute to a reduction in the cost of equity financing remains an open empirical question. The answer depends on the relative size of each country’s capital market relative to the EC and the rest of the world and on the degree of integration among equity markets—factors that are not treated explicitly in our simplified arbitrage condition. This illustrates the important role of the arbitrage assumption and the associated empirical difficulties.


Differences in the taxation of income from capital are likely to become a determining factor in the allocation of capital as the remaining barriers to the free flow of goods, labor, and capital are progressively removed within the EC. With the complete liberalization of financial capital movements accomplished by mid-1990, differences in the taxation of financial investment income across countries may lead to a misallocation of financial assets but, because of the parallel integration of credit markets, would not interfere significantly with the allocation of real assets. Differences in the tax treatment of corporate income could thus become the primary source of allocative distortion in fixed capital formation; consequently, the harmonization of corporate tax systems could go a long way toward establishing allocative neutrality in the EC.

Three important empirical policy questions have been addressed here: first, the magnitude of existing differences in effective corporate tax rates; second, the degree of harmonization that would be necessary to reduce the dispersion of effective tax rates; third, the allocative and efficiency implications of tax harmonization (or nonharmonization). In answer to the first question, our calculations have shown that there is considerable divergence in effective corporate income tax rates among EC member countries, with effective tax rates varying from 40 percent for Germany to 5 percent for Ireland. Second, it was shown that only the harmonization of both the statutory tax rate and the tax base results in a meaningful convergence of effective tax rates. Harmonization of the statutory tax rate without base harmonization yields a lower degree of convergence and diminished transparency. Harmonization of the tax base alone, leaving statutory tax rates and investment grants in place, increases transparency but does not produce any significant reduction in the dispersion of effective tax rates.

As regards the third question, the allocative implications of differential effective tax rates—and, conversely, of harmonization—depend on the elasticity of capital formation with respect to differential tax rates. In principle, investment flows should respond to tax-induced differences in the after-tax rate of return as they do to differences caused by other components of the rental cost of capital. The empirical evidence is, however, mixed, and many authors have found only small effects.37 Using a simple computable general-equilibrium model, where capital is allocated to equalize after-tax rates of return, we derived steady-state allocative effects of company tax harmonization in the EC. The results show a modest efficiency gain from harmonization but very significant effects for individual countries. When the overall EC capital stock is held unchanged, Germany’s share rises by 15 percent, and Italy’s by 5 percent, relative to the equilibrium capital stock under the present tax systems. Ireland and Luxembourg are the largest losers, with a 17 percent and 15 percent contraction in their respective shares.

The case for concerted harmonization of effective corporate tax rates has relied crucially on efficiency considerations. These results suggest that the overall static efficiency gains from equalization of tax burdens may in fact be quite small, although tax equalization may imply substantial adjustments in some countries. In the event, the debate over company tax harmonization may have to rely primarily on fiscal considerations.

The case for concerted harmonization seems strongest as regards definition of the tax base (capital cost recovery allowances, loss carryover, and so forth). Although the results of our analysis indicate that harmonization of the tax base does not produce significant convergence of effective tax rates, base harmonization may have very important efficiency implications within the EC by increasing transparency and reducing the compliance costs for multinational enterprises. Harmonization of the tax base would not in any way constrain the determination of effective tax rates, because national authorities would continue to exercise sovereignty over statutory tax rates and explicit fiscal incentives for investment (tax credits, cash grants).38 In practice, however, with the advent of the single market, EC member countries are likely to be under considerable pressure to engage in spontaneous, downward harmonization of statutory rates, a trend that has already been under way in most industrial countries since the early 1980s.

Appendix: Cost of Capital and the Corporate Tax Wedge

This Appendix develops the model of the firm underlying the derivation of the tax wedge. We consider a competitive firm operating in a stationary environment in a one-good neoclassical world with no adjustment costs and derive a demand for capital function from the necessary conditions for the maximization of the firm’s market value. In a steady state, the optimal capital-labor ratio is a decreasing function of the user cost of capital, itself a function of the rate of economic depreciation, corporate tax parameters, the rate of inflation, and the market rate of return on the financial asset (r). For a firm with a mixed financial structure, r is given by a weighted average of the interest rate and the return on equity.

Firm Optimization

The production technology is described by a production function F(K, L) where K is the firm’s capital stock and L its level of labor input. The function F() is assumed to be concave and twice differentiable, with positive and decreasing marginal products and FKL > 0. All firms produce a single homogeneous output that can also be used as capital in production and can be installed or dismantled instantaneously and at no cost. The evolution of a firm’s capital stock over time is described by the equation


where K’ is the time derivative of K, 5 is the rate of economic depreciation, and I is the level of real investment expenditures.

We consider a competitive firm operating in a stationary environment in which input and output prices increase at a constant rate of inflation n (alternatively, the firm has static expectations). At each point in time the firm chooses levels of labor input and investment expenditures and may raise financial capital by selling a single type of security defined as bonds (B). Output is sold at a price p, and after-tax profits plus the proceeds from the sale of new bonds are either paid out to security holders or used to finance investment:


where W is the real wage rate, i is the nominal rate of interest on bonds, and Tc denotes real corporate tax payments by the firm. Dividing both sides of this expression by p and rearranging gives the firm’s cash flow constraint in real terms:39


where b = Blp.

Corporate income is taxed at a flat rate tc. However, the government provides a grant or investment tax credit at a rate g per unit of new capital purchased and allows the deduction of depreciation of the (nominal) book value of the firm’s capital at a rate d. Finally, payments to security holders may also be subject to a tax or a subsidy at source. Real corporate tax payments are given by


where φ is the unit tax or subsidy on payments to security holders40 and V is the deflated book value of the firm’s capital stock, which evolves over time according to the equation


Substituting equation (18) into equation (17) yields


Assume that the firm behaves so as to maximize the market value of its initial securities, b(0).41 Integrating equation (21) allows the firm’s problem to be written as


subject to

K′ = I − δK

V′ = I − (d + π)V,

where ρ = i + φ is the cost of financial capital to the firm. The current-value Hamiltonian for this problem is given by

H = (1 − tc) [F(K, L) − WL] + tcdV − (1 − g)I + q(I − δK) + ν[I − (d + π)V],

where q and v are the costate variables or multipliers associated with the dynamic constraints. The necessary conditions for an optimal solution to this problem are given by

H/∂L = (1 − tc) [FL (K, L) − W] = 0,


Here r = i − π is the real interest rate, τc = (tc, d, g, φ) is the vector of corporate tax parameters, and the user’s cost of capital is defined as


where v is the present value of the tax savings from the depreciation of one unit of capital, given by

v = tcd/(ρ + d).

Demand for Capital

If the production function exhibits decreasing returns to scale, the necessary conditions (22) and (23) can be solved for the optimal capital stock K* and the optimal level of employment as functions of factor prices and tax parameters (that is, these equations characterize the firm’s factor demand functions). From a partial equilibrium perspective (and in the absence of installation costs), there is nothing to prevent an individual firm from achieving its optimal input level at market-determined prices, even if that involves a discrete jump in the capital stock. Hence, I will be set so as to reach and maintain K*.

Under the assumption of constant returns to scale, the firm’s size is indeterminate, but conditions (22) and (23) characterize the optimal capital-labor ratio k* as a function of factor prices and tax parameters. At the national level, the size of the corporate sector will be determined by the size of the labor force, N, which we take as given. In that case K* = Nk* in a steady state, investment is set accordingly, and condition (23) evaluated with L = N determines the equilibrium wage. If there is population growth, on a balanced growth path investment will be set so as to maintain a constant stock of capital per worker.

To derive the per capita demand for capital function, it will be convenient to work with the per capita production function, defined as


where k = K/L is the capital-labor ratio.

This allows us to rewrite the marginal productivity conditions in the form


Equation (26a) implicitly defines the per capita demand for capital as a function of the user cost of capital, which is in turn a function of corporate tax parameters and the rates of inflation and interest; that is


Differentiating equation (26a) implicitly, we find that42


Thus, we may think of the firm as renting capital services from itself and the imputed rental on capital as being equal to the user’s cost.43 At any rate, the firm’s optimal capital stock per worker is a decreasing function of the cost of capital, and inflation, interest rates, and tax parameters affect the demand for capital only through their effect on c.

Capital Market Equilibrium and the Arbitrage Condition

A complication arises because the rate of return on the financial assets (r) is actually a function of the financing mix of the firm. Consider an economy in which firms and households trade two types of securities, debt (d) and equity (e). Income from each type of security (n = d, e) is taxed at both the corporate and personal levels. Thus, taxes on financial assets levied at source (φn) are subtracted from the firm’s marginal cost of funds (ρn) to arrive at a market return (in); from this, personal taxes on security income (ηn) are deducted to obtain the net return (sn) to the security holder:

ρn = sn + φn + ηn,

where n = d, e.

We ask whether a systematic relationship can be expected to exist between the equilibrium returns on the two assets. One possible answer is based on arbitrage considerations: if in equilibrium both securities are held (or issued) by the same agent, their return (or cost to the issuer) must be the same at the margin.44

A problem arises because existing tax systems typically treat income from equity and debt differently. The theory suggests that firms will use both sources of finance only if their marginal cost is the same (ρe = ρd) and that households will hold both types of securities only if their net-of-personal-tax returns are the same (se = sd). In general, however, φd + ηd ≠ φe + ηe, and differences in the tax treatment of debt and equity make it impossible for both conditions to hold at once.

On theoretical grounds, then, there does not seem to be a clear-cut case for a specific arbitrage assumption. In practice, however, which particular assumption is chosen makes a big difference because tax wedges are quite sensitive to the discount rate. In this analysis we have taken a “middle ground” and have assumed that the arbitrage is done by an international (institutional) tax-exempt investor that requires a risk premium (h) on equity.

Cost of Capital for a Firm with Mixed Financial Structure

The overall cost of capital of a firm with a mixed financial structure can be computed in one of two ways. It can be taken as a (weighted) average of the pure equity and the pure debt cases, each derived separately, or it can be derived directly by assuming that the firm finances itself by issuing a composite security, with weights corresponding to the financing mix of the firm.45 The two methods are not strictly equivalent. In the derivation that follows, a single composite security is used, but an average of the pure equity and pure debt cases—the approach followed in the computation of tax wedges—provides an adequate approximation.

Think of the firm’s “bonds” as composite securities, part equity and part debt, with exogenously determined weights (ξe, 1 − ξe). Then, the nominal “market” return to the holders of the firm’s securities is given by


where id is the nominal return on debt (the interest rate) and ie is the nominal return on equity. The cost of financial capital to the firm can then be written as


where φ = (1 − ξe) φd + ξe φe

is the tax-induced wedge between the firm’s payout on one unit of the composite security and the return to its holders (before personal taxes). The parameters φe and φd depend on withholding taxes on interest and dividends and on the system of integration of personal and corporate taxes. Using the arbitrage condition to write ie as a function of id, leaves a single rate of return to be determined endogenously. From the arbitrage condition, the real return to holders of the composite security is given by


In conclusion, the discount rate for a firm with a mixed financial structure (with weights ξe for equity and 1 − ξe for debt) can be written as


where ω = [νθ(1 − td) + (1 − ν)], ν is the share of real earnings distributed as dividends, and 0 is the integration variable.

We are now in a position to derive the partials of the cost of capital function c(r, π, tc) with respect to the various tax parameters:




The cost of capital is a complex, nonlinear function of the market interest rate, the rate of inflation, and company tax parameters [τc = (g, d, tc, ω)]. The partial derivatives of this function are shown in Table 31.

Table 31.

Partial Derivatives of the Cost of Capital Function

article image

As we would expect, increases in the investment tax credit and the rate of tax depreciation lower the cost of capital, and increases in the real interest rate increase the cost of capital. In contrast, the effects of inflation and of the corporate tax rate are in principle ambiguous.

The direct effect of an increase in the corporate tax rate on c is clearly positive, reflecting the de-crease in the after-tax marginal product of capital. An increase in tc, however, lowers the real discount rate (through the interest deductibility provision) and increases the present value of the tax savings from depreciation (as a result of both the lower discount rate and the increased value of the tax savings associated with any given d).

In the case of inflation, the situation is more or less analogous. Because nominal interest payments are tax deductible, an increase in n reduces the real discount rate and, other things being equal, tends to lower the cost of capital. In the absence of indexation of the depreciable base, however, the increase in inflation reduces the present value of the tax savings from depreciation (even at the lower discount rate).

The corporate tax wedge is a natural by-product of the user cost of capital. We may think of the cost of capital as the sum of three components: economic depreciation, a return to the owners of the firm’s securities (r), and the tax burden on the marginal unit of capital. This last component (cr − δ) is what we refer to as the corporate tax wedge.

The concept of the wedge is useful in that it provides a summary measure of the corporate tax rate on the marginal unit of capital. It corresponds to the size of the tax-induced shift in the demand for capital relative to the net marginal product schedule and can therefore be interpreted as an indication of the incentive effect of corporate taxation.


See, for example, Harberger (1966) and, for cross-country comparisons, Kyrouz (1975).


Alternatively, the tax can be expressed in terms of a rate by dividing the absolute value of the wedge (ps) by any of the two rates of return (p or s).


The welfare implications of this effect depend on whether, in the absence of taxation, private behavior leads to a socially optimal level of capital accumulation. In an overlapping-generations model, life-cycle savers may in fact overaccumulate capital relative to the Pareto optimum and taxation can be Pareto improving.


In a closed economy and for a given total tax wedge, the breakdown between the personal and the corporate tax wedges determines only the level of r that clears the capital market and has no effect on p and s.


In addition, domestic and foreign capital must be fully substitutable from the point of view of the saver. Failing this condition, personal taxation will also affect the allocation of the capital stock. Otherwise, differences in the taxation of income from capital at the personal level only affect saving and the pattern of ownership of the capital stock.


This is one of the reasons why offshore or foreign equity markets rarely provide a cheaper source of funds than domestic equity markets. Under a proposed EC directive for the harmonization of company taxation, a common withholding tax on dividends and a common system of integration extended to all EC residents would have eliminated differences in the cost of equity financing across markets in the EC. The proposed directive was, however, formally withdrawn in 1990. See EC Commission (1975a).


Under the residence principle, a country exercises a tax claim on all income earned by residents and taxes it at a uniform rate.


This form of tax transparency fails to hold when financial institutions cannot credit withholding taxes paid abroad against their domestic tax liability—for instance, if the institutions are tax-exempt.


For some estimates of tax wedges on foreign direct investment income see Crooks and others (1989).


For a derivation from the objective function of the firm, see Hall and Jorgenson (1967). An international comparison of the required rate of return based on this methodology is given in Kopits (1982).


The financing mix depends on the relative tax treatment of debt and equity financing. The analysis abstracts from this form of endogeneity, and common fixed weights are attributed to debt and equityfinancingfor all 12 EC countries on the basis of source of funds data for private nonfinancial corporations in OECD (various years).


Some studies have approximated the marginal return to equity by the ex post return on the stock market; see Auerbach (1983). Another possibility is to impose the arbitrage condition at the level of the firm, thus equating the marginal cost of debt and equity financing.


Even where nonresidents are not exempt from withholding tax on interest from corporate bonds, it is assumed that such taxes do not alter the cost of debt because of the possibility of borrowing from banks or through the Euromarket.


Based on average data for the EC stock markets, the parameter v is set at 50 percent. Following the traditional view of dividend taxation, taxes and credits on dividends and other imputation measures are assumed to affect the cost of capital in proportion to the share of earnings distributed as dividends, rather than in proportion to the share of new equity issues in equity financing (“new view”). See Poterba (1987) for a discussion of the two views.


For the institutional investors the tax is also a final tax. The rate varies and is usually lower under tax treaty. The values chosen for each country correspond to the most favorable rate generally applicable to nonresidents.


Only the forms of integration that reach the representative investor are considered.


A full pass-through of inflation rates in nominal interest rates is assumed. Our calculations correspond to the fixed r case discussed in King and Fullerton (1984).


In principle, the existence of the EMS, coupled with agreement on phase one of the Delors Committee’s proposal for monetary integration (see EC Council of Ministers (1989)), including the removal of capital controls, should result in the convergence of inflation rates for the EC members. With the deutsche mark continuing to play the role of nominal anchor for the system, such convergence would presumably be toward the lower end of the spectrum. A common inflation rate of 2 percent is therefore our benchmark assumption. It is plausible, however, that such convergence may take time, with considerable inflation differentials prevailing during a transition period, making the second inflation scenario a reasonable alternative.


For Belgium, we use the corporate income tax rate announced for 1992, or 39 percent. For Denmark, the newly introduced tax rate of 40 percent is used.


See the description in Kuiper (1988) of a draft proposal considered earlier by the EC Commission.


For some countries the declining-balance method is not currently allowed. In that case the declining-balance rate is derived as a multiple (2.5) of the current straight-line rate.


Weights are derived from national capital stocks. The German rate on distributed profits is kept at 36 percent.


Local property taxes are not included in the analysis.


The weights assigned (0.6 to machinery and equipment and 0.4 to buildings; 0.6 to equity and 0.4 to debt) are based on national accounting averages of financial flows and on the composition of fixed capital formation.


The rates are 15 percent for machinery and 7 percent for buildings. In the case of Luxembourg, machinery investment also benefits from an investment tax credit not applicable in the case of buildings.


For a discussion and estimates of the sensitivity of the required rate of return to inflation under various forms of indexation in industrial countries, see Kopits (1983).


Despite excellent analytical work in this area by Sinn (1987) and Bovenberg (1986), among others, there have been few attempts to integrate estimates of tax wedges into a general equilibrium model to provide numerical estimates of the effect of tax changes on the allocation of capital. An exercise similar to ours was carried out by Fukao and Hanazaki (1987).


The introduction of capital accumulation would not eliminate the allocative distortions inherent to differential tax wedges, although the dynamics of capital accumulation could delay significantly convergence toward a steady state.


An alternative would be to take the interest rate as given and allow the overall capital stock to change in response to changes in the average level of taxation in the EC. This would correspond to the assumption of a small open economy and seems unrealistic.


This can be shown to be consistent with expressions (1) and (2), insofar as p = c/q – δ.


The unit elasticity of substitution between factors implied by the Cobb-Douglas production function is broadly consistent with estimates in Kopits (1982).


Estimates of the capital stock of each country for 1975 are obtained from Leamer (1984), updated to 1985. The labor force is adjusted by using the share of professional workers in the labor force as an index of quality; see United Nations (1988).


Because the tax wedges are themselves functions of the interest rate, the solution process requires an iterative procedure: initial values of the tax wedges are used to derive an equilibrium interest rate. The values of the tax wedges are then recalculated, using the interest rate, until convergence is achieved.


This would be generally true if corporate taxes were not benefit charges or, more specifically, if a constant expenditure level were maintained through compensating changes in labor income taxation and fully absorbed in lower after-tax wages, with no consequences for labor supply and employment.


Tax wedges for Japan and the United States are set to reproduce the equilibrium solution in scenario 1 and are held fixed thereafter.


See Snoy (1975), Kopits (1976), Caves (1982), Alworth (1988), and Slemrod (1989) for the effects of tax differentials on foreign direct investment, and Papke and Papke (1986) and Papke (1989) for the effects of cross-state differences in corporate taxation in the United States on business locational decisions.


This is the practice in Canada, where different rates at the provincial level are imposed on a commonly defined tax base.


By definition, b = B/p; hence b’ = d(B/p)/dt = (pB′ − Bp′)/p2 = (B′/p) – πb, and B′ = p(b′ + πb).


For example, interest payments on debt are considered a deductible expense. In that case ø = −tci; the tax-induced wedge between what the firm pays out and what the security holder gets is negative.


The general solution to this differential equation can be written in the form


article image

where c is an arbitrary constant. Setting c = 0 is equivalent to the assumption that the value of the securities is determined by “fundamentals” (that is, reflects the underlying cash flow of the firm).


The cost of capital is a function of the whole time path of expected future interest rates. Thus, the demand for capital function derived here must be interpreted either as a steady-state construct or as describing the behavior of a firm that myopically expects the current interest rate to remain constant over time.


It is tempting to go one step further and think of the firm as renting capital from households, who retain title to it. This is somewhat misleading, since what firms and households trade is not capital but bonds, and in general the relative price of the bonds can differ from unity. This procedure has been used, however, because it simplifies the analysis by hiding financial variables and defining the equilibrium directly in terms of the demand for capital.


In a stochastic environment, returns are defined in terms of expected utility, and differences in asset return can be ascribed to differences in risk. In a certainty setting, differences in equilibrium rates of return can be explained by differences in the rates of taxation on interest income, dividends, and capital gains among investors, with some investors favoring one type of asset over another. In this kind of segmented equilibrium, the returns on debt and equity need not be equalized at the personal level because no investor needs to hold them both.


See Boadway, Bruce, and Mintz (1987) for a discussion on the two approaches.