Tax Incentives and Investment in the Eastern Caribbean

Contributor Notes

Author(s) E-Mail Address: ssosar@ucla.edu

Tax incentives have been used extensively in the countries of the Eastern Caribbean Currency Union (ECCU) to promote investment. The associated revenue losses are large, and benefits in terms of new investment have been limited, raising doubts about the cost effectiveness of the tax incentive schemes. This paper examines the effects of incentives using the marginal effective tax rate approach (METR), adapting this methodology to the case of a small open economy where the marginal investor is a nonresident. The results show that METRs are high in the region; that there is a large dispersion in the size of METRs across financing source; and that METRs on investment are larger than the overall distortion on capital, with a substantial subsidy to domestic saving. In the presence of tax holidays-the most common incentive scheme in the region-the distortion on capital basically vanishes.

Abstract

Tax incentives have been used extensively in the countries of the Eastern Caribbean Currency Union (ECCU) to promote investment. The associated revenue losses are large, and benefits in terms of new investment have been limited, raising doubts about the cost effectiveness of the tax incentive schemes. This paper examines the effects of incentives using the marginal effective tax rate approach (METR), adapting this methodology to the case of a small open economy where the marginal investor is a nonresident. The results show that METRs are high in the region; that there is a large dispersion in the size of METRs across financing source; and that METRs on investment are larger than the overall distortion on capital, with a substantial subsidy to domestic saving. In the presence of tax holidays-the most common incentive scheme in the region-the distortion on capital basically vanishes.

I. Introduction

Tax incentives have been widely used in the countries of the Eastern Caribbean Currency Union (ECCU) as an instrument to promote investment. These incentives have typically been granted in the form of tax holidays, exempting certain firms from paying both import duties and corporate income taxes.

Recently, doubts have been raised about the efficacy and cost effectiveness of tax concession schemes in the region. On the one hand, revenue losses from them are large and have increased over the last decade. Chai and Goyal (2005) estimate forgone tax revenues in the ECCU countries to range between 10 and 16 percent of GDP. The high cost in terms of loss of revenues is particularly relevant given the fiscal and financial vulnerabilities of the region. Tax incentives in the ECCU have other less obvious costs. In some cases tax holidays are granted with a considerable degree of discretion, relying on case-by-case evaluations.2 These types of concessions are not only difficult to administer effectively, but are also nontransparent. In addition, they are in general non-neutral, distorting the allocation of resources. On the other hand, the benefits of tax incentives in terms of increased foreign direct investment (FDI) appear to have been limited. In fact, FDI flows to the ECCU countries relative to the rest of the world declined dramatically during the second half of the 1990s.

The main objective of the paper is to examine the incentives for investment in the corporate sector offered by the tax systems in the ECCU using the marginal effective tax rate (METR) approach. This approach has been widely used to examine tax incentives and the distortions generated in the capital market by the tax system. However, no previous work has been done applying this methodology to the case of the ECCU.

The METR is a forward-looking measure that summarizes the incentives to invest in a particular asset as provided by complex tax laws. It can be interpreted as the tax rate that bears on revenue from marginal investment, hence constituting the adequate measure of incentives to additional investment. The METR, rather than the statutory or average effective tax rate (AETR), is the tax rate that really matters in capital allocation.

The contributions of the paper are twofold. First, a methodology is developed based on King and Fullerton (1984), extending their approach to the case of a small open economy where the marginal investor is nonresident. As will be discussed later, the corresponding arbitrage assumption implies that the saving and investment sides of the domestic capital market are effectively segmented. An important implication is that the METR on capital decisions can be decomposed into that corresponding to the investment side and that corresponding to the saving side. Indirect taxes—which are typically excluded from the METR framework—are also incorporated into the analysis.

The second contribution of the paper is the application of the methodology to the countries of the ECCU. METRs are computed for all the countries in the region in order to get a quantitative estimation of the distortions generated in the capital market by the tax system. Moreover, the total distortion on capital is broken into two components by estimating separately the distortions affecting the investment side and those affecting the saving side. The effectiveness of alternative tax schemes is analyzed by performing the following exercises: (i) simulation of METRs for a scenario with tax holiday, in order to analyze the effective incentive provided by this type of concession, which is the most commonly used in the region; and (ii) simulation of the effects on METRs of eliminating tax holidays and lowering the general corporate income tax rate. Finally, in addition to the intertemporal distortion—the extent of which is measured by the size of the METR—other types of distortions caused by the tax system on capital are estimated; namely the distortions across sector, asset, and financing instrument.

The paper proceeds as follows. The next section summarizes the related literature. Section III presents the conceptual framework and methodology. Section IV describes the main components of the tax system in ECCU countries, in particular the different incentive provisions. Section V undertakes the application of the methodology to these countries, and presents the results of the simulations for a set of alternative scenarios. Section VI contains some concluding remarks.

II. Related Literature

The effect of tax incentives in developing countries has been examined using different methodologies in recent empirical studies. Several country studies have used the METR approach: Estache and Gaspar (1995) analyze the Brazilian case; Boadway, Chua, and Flatters (1995) the case of Malaysia; and Mintz and Tsiopoulos (1995) the case of Central and Eastern European countries. Another branch of the literature studies the effect of tax systems on FDI in developing countries. Shah and Slemrod (1995) use time series data to examine the sensitivity of FDI to taxes in Mexico, incorporating METRs and other measures for tax rates into the analysis. Altshuler, Grubert, and Newlon (1998) use cross-section data to study the effects of taxation on FDI in 58 countries in 1984 and 1992. Using panel data, Wei (2000) analyzes the effect of taxes and corruption on FDI for a sample of 45 developed and developing countries. Based on a dynamic model of production, Bernstein and Shah (1995) provide an empirical framework for assessing the case of Mexico, Pakistan, and Turkey. Another case study is provided by Chalk (2001), who analyzes tax incentives in the Philippines from a regional perspective. Surveys of investors have also frequently been used to evaluate the effectiveness of tax incentives in developing countries. Examples of this approach are provided by Halvorsen (1995) for Thailand; and OECD (1995) for a group of transition economies.

Previous studies on tax incentives in the ECCU have approached the issue from different perspectives. Bain (1995) focuses on revenue losses associated with tax concessions; Chai and Goyal (2005) compare these costs with the benefits in terms of increased FDI; Andrews and Williams (1999) analyze the administrative aspects of tax concessions; and Lecraw (2003) focuses on coordination and harmonization issues. However, as was mentioned before, no previous work has been done using the METR approach to the case of the ECCU countries.

III. Methodology

The METR approach is used, following the seminal contribution by King and Fullerton (1984), and extending their methodology to the case of a small open economy where the marginal investor is nonresident.

The METR approach has been widely used to examine tax incentives and distortions in the capital market induced by the tax system. The METR measures the tax burden on a marginal investment in a given type of asset; that is, how the marginal rate of return on such investment is affected by tax provisions. The focus is on marginal effective tax rates because for allocation decisions they are more relevant than average ones. The effect of alternative investment incentives can be estimated by computing how the METR is affected by the incentive scheme.

The METR is defined as the wedge between the expected pre-tax real rate of return on a new marginal investment project, net of true economic depreciation (p) and the after-tax real return to the saver who supplied the funds for the investment (s), typically expressed as a percentage of the pre-tax rate of return:

METR=psp.(1)

In a scenario without of taxes, a saver investing his funds in a marginal project would earn a real rate of return equal to the rate of return of the project itself. On the other hand, with distortionary taxes and concessions, the two rates of return can differ. The size of the wedge depends on such factors as: specific corporate tax and incentives scheme, the interaction of taxes with inflation, the tax treatment of depreciation and inventories, the personal tax regime, the treatment of capital gains, dividends and other forms of income, the type of asset, the economic sector of the project, and the source of finance.

A. Derivation of the Required Pre-Tax Real Rate of Return (p): A Simple Model of Investment

The required pre-tax real rate of return (p) is typically inferred by measuring the user cost of capital. We use the standard theory of the firm, following the Jorgenson (1963) and Hall and Jorgenson (1967) neoclassical models of investment, in order to derive an expression for the user cost of capital.

Consider a firm that produces output according to the following production function with the standard properties:

Yt=F(Kt,Lt)

where Yt, Kt, and Lt are output, capital stock and labor in period t.

The firm’s problem is to maximize the net present value of the cash flow discounted by the nominal cost of capital (r).3 Hence, firm’s optimization problem is:4

Max0[(PtF(Kt,Lt)WtLt)(1τ)(1ϕ)(1Z)QtIt]erts.t.Kt=ItδKt;K0 given

where It stands for investment in period t, Pt is the price of output, Qt is the price of investment goods, Wt is the nominal wage, τ is the corporate income tax rate, ϕ is the investment tax credit, δ is the economic depreciation rate of capital, and Z stands for the present value of the future tax savings from depreciation allowances per dollar of gross investment.

The current value Hamiltonian for this problem is:

H=[PtF(Kt,Lt)WtLt](1τ)(1ϕ)(1Z)QtIt+λt[ItδKt]

where λ is the co-state variable representing the shadow price of capital. The first order conditions for the optimal value of the capital path, using Lt and It as control variables and Kt as the state variable are:

HIt=(1ϕ)(1Z)Qt+λt=0(2)
HLt=[PtFL(Kt,Lt)Wt](1τ)=0(3)
λt=rλtHKt=rλtPtFK(Kt,Lt)(1τ)+δλt(4)
Hλt=Kt=ItδKt.(5)

First order condition (3) implies the standard equilibrium condition, FL=WtPt=w that is, labor is employed until its marginal product is equal to the real wage.

From (2) we infer that λt=(1ϕ)(1Z)Qt, thus λt=(1ϕ)(1Z)Qt. Hence, using (4) we get:

PtFK(Kt,Lt)=(r+δQtQt)1τ(1ϕ)(1Z)Qt.(6)

Deflating Pt and Qt by e-πt, where π is the inflation rate and p* and q are then real prices, we get:5

pt*FK(Kt,Lt)qt=(rπ+δqtqt)1τ(1ϕ)(1Z).(7)

The right-hand side of equation (7) constitutes an expression for the user cost of capital incorporating tax issues. In equilibrium this user cost of capital is equal to the marginal product of capital gross of taxes. In order to convert it to a rate of return, the economic depreciation rate is subtracted. The required real rate of return, gross of taxes but net of depreciation (p), is then:

pt=(rπ+δqtqt)1τ(1ϕ)(1Z)(δqtqt).(8)

The present value of the future tax savings from depreciation allowances per dollar of gross investment (Z) depends on the details of the tax system, in particular the depreciation schedules for tax purposes. The expressions for Z for alternative tax depreciation provisions are presented in Appendix I.

B. After-Tax Real Rate of Return to Savers (s) and the Nominal Cost of Funds (r)

The after-tax real rate of return to savers (s) depends on the source of financing and can be measured as a weighted average of the rates on those sources: corporate bonds, retained earnings, and new equity issues. Hence:

s=βi(1mi)+(1β)[α(1md+τ)ρ+(1α)ρ]π(9)

where i is the corporate borrowing rate, mi and md are the individual’s personal tax rate on interest and dividends respectively, β is the debt-asset ratio of the firm, α is the proportion of new shares in total equity finance, and ρ is the firm’s cost of equity. Intuitively, a saver holding corporate debt receives a rate of return equal to i(1- mi). The after-tax nominal rate of return on new shares is (1–md + τ)ρ, because dividends are taxed at the personal level and can be credited for corporate taxes paid. Finally, the rate of return to savers on retained earnings is equal to ρ, since capital gains are in general nontaxable.

The nominal cost of funds, r, i.e. the one that the firm uses for discounting its cash flow, is a weighted average of its after-tax borrowing costs and the cost of raising equity:

r=βi(1τ)+(1β)ρ.(10)

The cost of debt financing is i(1-τ), due to the fact that interest payments are tax deductible. Note that it is assumed that the cost of equity finance from new issues is equal to that from retained earnings. This is a correct assumption if there is no tax on capital gains and if there is full imputation of corporate taxes, which ensures that dividends remain un-taxed at the personal level.

C. The Arbitrage Assumption

A key issue in the use of the METR approach is the selection of an arbitrage assumption. The arbitrage assumption determines which rates of return are taken as given, and consequently which rates are computed given the true values of the parameters. In the seminal work by Fullerton and King (1984) two alternative arbitrage assumptions are adopted: the so called fixed-p and fixed-s cases. The former implies that investments be compared with the same pre-tax rate of return by taking p as given. The latter, on the other hand, takes the after-tax return to savers (s) as given.

In this paper an alternative, small open economy assumption is used—the financing costs facing a country are determined by international capital markets—and it is assumed that the marginal investor is a nonresident. Therefore, the after-tax rate of return to foreign investors is exogenous. Using this arbitrage assumption both i and ρ can be computed, which in turn are used to calculate s, r and p.

More specifically, nominal rates of return are based on the world real rate of return, adjusted for taxes on nonresidents:6

i=r*+π1τd(11)
ρ=r*+π1τe(12)

where r* is the world real interest rate, and τd and τe stand for the noncreditable tax rate on nonresidents’ interest and dividend income, respectively.

Given our arbitrage assumption, the required rate of return is exogenously given by international capital markets. Therefore, the saving and investment sides of the domestic capital market are effectively segmented; in any given year domestic saving and investment need not necessarily be equal. An important implication is that the METR on a capital decision can be broken into two components, one corresponding to the investment side and the other corresponding to the saving side, as illustrated in Figure 1.7

Figure 1.
Figure 1.

Marginal Effective Taxes in a Small Open Economy

Citation: IMF Working Papers 2006, 023; 10.5089/9781451862836.001.A001

METR=pspMETR(I)=pr*pMETR(S)=r*ss

The effects of each country’s tax system on incentives to invest are measured by the marginal effective tax rate on investment, computed as the percentage difference between the pre-tax real rate of return under each system (p), and the real rate of return if income from capital were untaxed—that is, the world real rate of return, r*. The marginal effective tax rate on saving is measured by the percentage difference between the after-tax real return to domestic savers (s) and the prevailing rate of return available in international capital markets (r*), which would be the return to investors in the absence of any taxes. Therefore, the marginal effective tax rates on investment and saving can be expressed as:

METR(I)=pr*p(13)
METR(S)=r*ss.(14)

D. Incorporation of Indirect Taxes

The METR methodology is usually used to estimate the distortion caused only by direct taxes. In this paper, some indirect taxes—in particular import duties and tariffs—are also incorporated into the analysis. This is particularly relevant in our case study, since in the ECCU countries tax concessions typically include exemptions from this type of tax. In fact, one of the most common incentives in the region is the exemption from duties of imports of plant and equipment. This type of tax affects the price of investment, and the expression for p becomes:

pt=(rπ+δqtqt)1τ(1ϕ)(1+ψ)(1Z)(δqtqt)(15)

where ψ represents the import tariff.

E. METR in the Presence of Tax Holidays

As mentioned earlier, the tax holiday is the most common incentive scheme in the ECCU. The METR approach for the case of a firm which operates under a tax holiday is now described, following Boadway, Chua, and Flatters (1995).

Suppose that the length of the holiday period is equal to T years, and no taxes are paid by the firm during that lapse of time. The computation of p in the presence of tax holidays is more complicated, and can be expressed as:

pt=(rtπ+δqtqt)1τt(1+ψt)(1Zt)+[1+rtπqtqt](Zt)1τt(δqtqt).(16)

The tax holiday provisions make τt, ψt, rt, it, ρt, and Zt vary over time. Statutory tax rates are equal to zero between period 0 and T, and τ and ψ afterwards. The cost of funds for the firm is now given by rt = βit(1–τt) + (1–β)ρt. Since statutory taxes take only two values depending on whether the firm is operating during or after the tax holiday, rt has two values as well: r0 during the holiday and r1 afterwards.

It is assumed that there is an initial depreciation allowance of γ on gross investment, that the remaining (1- γ) can be depreciated according to a declining balance schedule at a rate equal to η, and that depreciation allowances cannot be deferred until the end of the tax holiday. Under these assumptions, Zt is given by:8

Zt=τ(1γ)(1+r1)(ηη+r1)(1md)[(1γ1+r0)]T1(17)

F. Limitations of the Methodology

The METR approach is used extensively as an indicator of the incentive effects of the tax structure. Nonetheless, this methodology has some limitations. First, the METR measures only the distortion on capital decisions by quantifying the effect of taxes and concessions on the rate of return; it does not measure the responsiveness of investment, nor it does show the effects on government tax revenues. Second, the computation of METRs typically implies some strong assumptions. All markets are assumed to be competitive. Firms operate in a risk-free environment or at least one in which they maximize only expected returns; in fact firms act as if the future cash flows and tax rules were known with certainty. The financial structure of the firm is taken as given; that is, the incentive effects of the tax provisions on the financial structure are not endogenous. Finally, the same tax structure, tax rates, and inflation rate are expected to remain constant over the entire life of the project; hence METRs computations estimate capital decisions in a world in which economic agents expect no changes.

IV. Tax Incentives in the Eccu

Countries in the ECCU grant a wide range of tax concessions, currently provided by various laws. The most important are the Hotel Aid Ordinances, which apply to hotels, the Fiscal Incentives Ordinances, which apply to the manufacturing sector, and the Import Duty Act, which refers to import duties. Each ECCU country issues its own version of this legislation. In general, tax concessions have to be approved by Cabinet, but some authority is delegated to Ministries of Industries, Development Corporations, Directors of Finance, and Customs authorities. Even though most concessions are granted under the existing legislation, there is a significant degree of discretionary power available to Ministers and the Cabinet. The incentive schemes typically exempt specific firms from paying corporate income taxes for a tax holiday period varying from 5 to 20 years, and may also allow the firms to carry forward net losses during the holiday period. On the indirect tax side, tax incentives comprise in general exemption from import duties on raw materials and equipment. Appendix II summarizes the most relevant issues of the tax system in each country of the ECCU, describing in particular the tax incentive scheme in each of them.

V. Results

This section presents the estimation of the distortions on capital decisions resulting from the ECCU countries’ tax systems. METRs are computed for each of the six countries in the region, considering different sectors (tourism and hotels, and manufacturing), assets (machinery and buildings), and sources of financing (debt, equity, and a weighted average of these two sources).9

The main objectives of this section are as follows. First, METRs for each country of the ECCU are computed in order to quantify the magnitude of the intertemporal distortions affecting capital markets caused by the tax system. Second, indirect taxes are incorporated in the estimation of METRs. Third, other types of distortions caused by the tax system on capital are estimated, measured by the dispersion of the METRs across sectors, across assets and across financing instruments. Fourth, the total distortion on capital is decomposed by estimating separately the distortions affecting the investment side and those affecting the saving side. Finally, simulations of the effectiveness of alternative tax schemes are undertaken: (i) simulation of the METRs in every country of the ECCU for a scenario with tax holidays, the most widely used concession scheme in the region, in order to analyze the effective incentive provided thereby; and (ii) simulation of the effects on METRs of eliminating tax holidays and lowering the general corporate income tax rate.

A. Magnitude of the Distortions

We begin by computing METRs for a base case, in which the sector considered is tourism and hotels, the asset considered is machinery, and the financing source considered is a weighted average of debt (35 percent) and equity (65 percent) financing. The first column of Table 1 presents the results for each country in the region for a scenario without tax holidays. The size of METRs is high in the region, 23.7 percent on average, and METRs range from 16.9 percent in Grenada to 33.7 percent in Antigua and Barbuda. Intuitively, for every dollar of income from the marginal investment, approximately 24 cents are paid in taxes and 76 cents are earned by the saver who supplied the funds for the investment. The magnitude of these distortions is fairly similar to those of the rest of the Caribbean, where the METR is 24.7 percent on average.10 The size of METRs in the ECCU is higher if we include indirect taxes into the analysis. As the second column of the table shows, the average METR increases to 41.1 percent, varying from 34.1 percent in the case of Grenada to 47.8 percent in the case of Antigua and Barbuda.

Table 1.

METRs in the ECCU Region

(Tourism and hotels; machinery; 35 percent debt -65 percent equity)

article image
Source: Author’s calculations.

B. The Case of Tax Holidays

As was mentioned earlier, the tax holiday is the most common incentive scheme within the region. The third column of Table 1 presents the results of the simulations for the base case assuming now that the investment is undertaken by a firm enjoying a tax holiday. By comparing METRs for cases with and without a tax holiday, the effective incentive to investment provided by such concession scheme can be estimated. In the presence of tax holidays there is a dramatic reduction in METRs and the distortion on capital caused by taxes basically vanishes, and in some cases—Antigua and Barbuda, and Grenada—it becomes a subsidy. To the extent that tax holidays are in some cases granted on a discretionary basis, this substantial difference on METRs implies a large distortion discriminating among investments which are granted a tax holiday and those which are not, a distortion that may imply an important misallocation of resources.

C. Distortions Across Sectors, Assets, and Financing Sources

Tables 25 present the results of the simulations for each country, for two different sectors— tourism and hotels, and manufacturing—for two types of assets—machinery and buildings— and for three alternative financing sources—100 percent debt, 100 percent equity, and a weighted average of both. There is a small dispersion of the size of METRs across sectors and across assets, which implies that both the intersectoral and cross-asset distortions are not of substantial magnitude. However, there is a large dispersion in the size of METRs across financing instruments. In fact, METRs for investments financed through debt are substantially lower, and in some cases—when indirect taxes are not considered—are negative. This different pattern of distortions across method of finance is not related to any clear economic goal; therefore, the unintended consequences of these distortions are likely to cause an inefficient use of resources.

Table 2.

METRs in the ECCU Region

(Tourism and hotels; machinery)

article image
Source: Author’s calculations.
Table 3.

METRs in the ECCU Region

(Tourism and Hotels; buildings)

article image
Source: Author’s calculations.
Table 4.

METRs in the ECCU Region

(Manufacturing; machinery)

article image
Source: Author’s calculations.
Table 5.

METRs in the ECCU Region

(Manufacturing; buildings)

article image
Source: Author’s calculations.

D. Distortions on Investment and on Saving

The selected arbitrage assumption discussed in Section III allows us to decompose the METR on a capital decision into two components, one corresponding to the investment side and the other corresponding to the saving side. As illustrated in Table 6, METRs on investment for our base case11 are larger than the overall distortion on capital; in fact, the average METR(I) is equal to 53.1 percent. As in the case of the overall METRs, the presence of a tax holiday implies a large reduction of the METR(I), to an average equal to 11.0 percent. The fact that METR(I)s are higher than the overall distortions, indicates that METRs on domestic saving are negative. Table 7 illustrates the magnitude of the subsidies imposed by that the tax system on domestic saving. These subsidies are essentially explained by the lower burden imposed by personal income taxation on domestic savers relative to foreign savers. As in the case of the investment side, when tax holidays are considered the size of the distortion on saving is reduced.

Table 6.

METRs on Investment in the ECCU Region

(Tourism and hotels; machinery; 35 percent debt - 65 percent equity)

article image
Source: Author’s calculations.
Table 7.

METRs on Saving in the ECCU Region

(Tourism and hotels; machinery; 35 percent debt - 65 percent equity)

article image
Source: Author’s calculations.

E. An Alternative Incentive Scheme

As stated earlier, the presence of tax holidays causes a sharp reduction of METRs. A large reduction can also be obtained by considering an alternative incentive scheme, consisting of a lower corporate income tax rate and an initial depreciation allowance with further allowances of the undepreciated asset according to the true depreciation.12 The results of the simulations for this alternative incentive scheme are presented in Table 8. The average METR in this scenario is equal to 6.3 percent. This alternative incentive scheme would not be granted on a discretionary basis; therefore, it would not cause the misallocation of resources caused by tax holidays (due to the distortions across investments benefited and those not benefited from the concession). Tax holidays have other disadvantages compared to this alternative scheme which have been highlighted in the related literature. First, the revenue losses associated to tax holidays are typically larger, without providing additional benefits in terms of lower METRs. Second, tax holidays represent a less transparent, less simple and less easy to administer incentive scheme. Finally, as has been shown by Harberger (1980), while the tax holiday system is non-neutral, an incentive scheme with a low corporate income tax rate and an initial depreciation allowance with further depreciation according to the true deprecation is neutral.13 As a non-neutral scheme, tax holidays subsidize not only the net return of the investment but also depreciation, hence discriminating between projects with different lifetimes.

Table 8.

METRs in the ECCU Region: An Alternative Incentive Scheme

Corporate income tax rate = 30 percent, initial depreciation allowance = 60 percent

(Tourism and hotels; machinery; 35 percent debt - 65 percent equity)

article image
Source: Author’s calculations.

VI. Concluding Remarks

In this paper, the METR approach is used to analyze the distortions on capital decisions created by the tax system in the ECCU member countries. This methodology is adapted to the case of a small open economy with a foreign marginal investor, incorporating both direct and indirect taxes into the analysis.

The main results of the paper are summarized as follows. METRs are high in the region, in particular when we include indirect taxes into the analysis. There is a small dispersion of the size of METRs across sectors and across assets, hence neither the intersectoral nor the cross-asset distortions appear to be large. In contrast, there is a large dispersion in the size of METRs across financing sources. METRs for debt-financed investments are substantially lower, and in some cases are negative. METRs on investment are larger than the overall distortion on capital, whereas there is a large subsidy to domestic saving explained by the lower burden on domestic savers relative to foreign savers imposed by the tax system at the personal level.

Tax holidays are the most common incentive scheme in all ECCU countries. METRs in the case of firms granted a tax holiday are substantially lower than those in the case without such a concession. In fact, with a tax holiday the distortion on capital caused by taxes basically disappears, and in same cases it becomes a subsidy. To the extent that in some cases tax holidays are granted with a considerable degree of discretion, this difference on METRs implies a substantial distortion discriminating among investments which are granted a tax holiday and those which are not, which may imply an important misallocation of resources.

A similar reduction in the METRs can also be obtained by considering an alternative incentive scheme, consisting in a lower corporate income tax rate and an initial depreciation allowance with further allowances of the undepreciated asset based on the true depreciation. In contrast to the tax holiday, this scheme would not be granted on a discretionary basis and therefore it would not cause the inefficient allocation of resources due to distortions across investments benefited and those not benefited from the concession.

APPENDIX I

Alternative Expressions for Z

The present value of the future tax savings from depreciation allowances per dollar of gross investment (Z) depends on the details of the tax system, in particular the depreciation schedules for tax purposes. In this Appendix we present the expressions for Z for alternative tax depreciation provisions, following King and Fullerton (1984).

First, consider a case in which tax depreciation is granted at an exponential rate equal to η (which is the continuous time version of declining balance depreciation) and tax depreciation allowances are computed at historic cost. The expression for Z in this case is:

Z=0τηe(η+r)tdt=τηη+r.

Second, consider the case in which the tax system provides straight line depreciation. A tax lifetime, L, is specified for each asset; and the asset may be depreciated for tax purposes by 1/L per unit in each year. In this case, Z can be expressed as:

Z=0Lτ(1L)ertdt=τ(1erL)rL.

Finally, suppose that a proportion γ of an asset’s cost can be immediately expensed, and the remaining (1- γ) can be depreciated according to a declining balance schedule on a historic cost basis. In this case Z becomes:

Z=τγ+(1γ)0ητe(r+η)tdt=τ(rγ+η)rη.

APPENDIX II

Summary of the Tax Provisions and Fiscal Incentives in the ECCU

article image
article image
Sources: International Monetary Fund; and country authorities.

Corporate income tax rate.

APPENDIX III

Data and Assumptions

Most of the data used for each country are described in Section IV and in Appendix II, in particular the corporate and personal income tax rates, the nonresident withholding tax rates, the depreciation schedule provided by law, the loss carry-forward provisions, and the details regarding tax holiday schemes. In this Appendix, we describe additional data and assumptions used for the empirical analysis.

The projected inflation rate for 2005 for the ECCU—2.1 percent—is used as the inflation rate for all the countries in the region. Following previous work in the related literature, the real world interest rate (r*) is assumed to be equal to 10 percent. We perform a sensitivity analysis assuming r* to be 5 percent, and the results do not vary significantly.

The economic lifetime of the asset is assumed to be 20 years in the case of machinery and 50 years in the case of buildings. When not specified by law, we assumed that lifetime and depreciation rates are equal to the true ones. Finally, in the simulations corresponding to the tax holiday cases, it is assumed that depreciation allowances cannot be deferred until the end of the tax holiday.

References

  • Andrews, Lennox, and Karen Williams, 1999, “Improving the Administration for Granting Tax Concessions in the OECS,” mimeo, Eastern Caribbean Central Bank.

    • Search Google Scholar
    • Export Citation
  • Altshuler, R., H. Grubert, and T. S. Newlon, 1998, “Has U.S. Investment Abroad Become More Sensitive to Tax Rates?,” NBER Working Paper No. 6383, Cambridge MA: National Bureau of Economic Research.

    • Search Google Scholar
    • Export Citation
  • Bain, Laurel, 1995, “Tax Concessions and Their Impact on the Revenue Base of the ECCB Territories,” mimeo, Eastern Caribbean Central Bank.

    • Search Google Scholar
    • Export Citation
  • Bernstein, Jeffrey I., and Anwar Shah, 1995, “Corporate Tax Structure and Production”, in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 503543). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Boadway, Robin W., Dale Chua, and Frank Flatters, 1995, “Investment Incentives and the Corporate Tax System in Malaysia,” in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 341373). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Chai, Jingqing, and Rishi Goyal, 2005, “Tax Concessions and Foreign Direct Investment in the ECCU,” IMF Country Report No. 05/305, International Monetary Fund, pp. 7392.

    • Search Google Scholar
    • Export Citation
  • Chalk, Nigel A., 2001, “Tax Incentives in The Philippines: A Regional Perspective.” IMF Working Paper WP/01/181, Washington D.C.: International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Estache, Antonio, and Vitor Gaspar, 1995, “Why Tax Incentives Do Not Promote Investment in Brazil”, in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 309340). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Hall, Robert, and Dale W. Jorgenson, 1967, “Tax Policy and Investment Behavior.” American Economic Review 57 (June 1967), pp. 391414.

    • Search Google Scholar
    • Export Citation
  • Halvorsen, Robert, 1995, “Fiscal Incentives for Investment in Thailand,” in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 399436). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Harberger, Arnold C., 1980, “Tax Neutrality in Investment Incentives”, in Aaron, Henry J., and Boskin, Michael J. (Ed.), The Economics of Taxation, pp. 299316. Washington, D.C., The Brooking Institution.

    • Search Google Scholar
    • Export Citation
  • Hansson, Ingemar, and Charles Stuart, 1986, “The Fisher Hypothesis and International Capital Markets,” Journal of Political Economy, 94(6): pp. 133037.

    • Search Google Scholar
    • Export Citation
  • Jorgenson, Dale W., 1963, “Capital Theory and Investment Behavior,” American Economic Review 53 (1): pp. 24759.

  • King, Mervyn, and Don Fullerton, 1984, The Taxation of Income from Capital: A Comparative Study of the U.S., U.K., Sweden and West Germany. Chicago, Ill.: University of Chicago Press.

    • Search Google Scholar
    • Export Citation
  • Lecraw, Donald, 2003, “Investment Incentives and Harmonization,” mimeo.

  • Mintz, Jack M., 1995, “Tax Holidays and Investment,” in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 165194). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Mintz, Jack M., and Thomas Tsiopoulos, 1995, “Corporate Income Taxation and Foreign Direct Investment in Central and Eastern Europe,” in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 455480). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • OECD, 1995, “Taxation and Foreign Direct Investment: the Experience of the Economies in Transition.” Paris: OECD.

  • Shah, Anwar, 1995, “Overview”, in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 130). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Shah, Anwar, and Joel Slemrod, 1995, “Do Taxes Matter for Foreign Direct Investment?,” in Shah, A. (Ed.), Fiscal Incentives for Investment and Innovation (pp. 481501). New York: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Wei, Shang Jin, 2000, “How Taxing is Corruption on International Investors?,” Review of Economics and Statistics, Vol. 82, pp. 111.

    • Search Google Scholar
    • Export Citation
  • Zee, Howell H., Janet G. Stotsky, and Eduardo Ley, 2002, “Tax Incentives for Business Investment: A Primer for Policy Makers in Developing Countries,” World Development Vol. 30, No. 9, pp. 14971516.

    • Search Google Scholar
    • Export Citation
1

This paper was prepared while the author was a summer intern in the Caribbean I Division of the Western Hemisphere Department. The author is grateful to Paul Cashin, Jingqing Chai, Rishi Goyal, Guillermo Tolosa, and seminar participants at the IMF, for useful suggestions and comments.

2

Even though most concessions are granted under the existing legislation (Hotel Aid Ordinances, Fiscal Incentives Ordinances, Import Duty Acts, etc.), some others are granted by special decisions by Ministers or the Cabinet, with a significant degree of administrative discretion.

3

This optimization problem is equivalent to maximize the equity or market value of the firm, since the latter is proportional to the net present value of its cash flow.

4

We set up the model in continuous time only for convenience.

5

In the literature, most applications of this methodology assume qtqt=0, either because estimates of real capital gains are not reliable or because the tax system is indexed to inflation.

6

This formulation is based on Hansson and Stuart (1986).

7

The graph illustrates a “normal” situation in which METRs, both on investment and saving, are positive. Many tax systems, including the ECCU as will be shown later, have the effect of subsidizing either marginal investments or savings.

8

For a detailed derivation of Zt in the case of tax holidays see Mintz (1995).

9

The data used for each country’s estimation is summarized in Appendix II. Additional data and assumptions used in the empirical analysis are described in Appendix III.

10

The average for the rest of the Caribbean includes The Bahamas, Barbados, Belize, Dominican Republic, Guyana, Haiti, Jamaica, Suriname, and Trinidad and Tobago.

11

We present only the results of the simulations of the METR(I) and METR(S) for the base case; we undertook the simulation for all the combinations of sectors, assets, and financing instruments and the results do not vary significantly.

12

In the following simulations we consider a corporate income tax rate equal to 30 percent and an initial depreciation allowance equal to 60 percent of the value of the asset. Similar results can be obtained with a lower corporate income tax rate—20 percent—and a less generous initial depreciation allowance.

13

A non-neutral scheme is defined as one that promotes projects with low pre-tax rate of return while other projects with higher pre-tax rates of return are not promoted.

14

Enclave enterprises are defined as 100 percent export-oriented projects, according to the 1974 CARICOM Agreement.