Credit and deferral as international investment incentives

JOURNAL, OF

PUBLIC

ECONOMICS

EISEVIEX

Journal

of Public

Economics

55 (1994)

323-347

Credit and deferral as international incentives * James John F. Kennedy

R. Hines,

Jr.*

School of Government, Harvard lJniver&y, Cambridge, MA 02138, USA

Received

September

1992, final version

investment

received

79 John F. Kennedy

March

Street,

1994

Abstract Many governments tax the foreign income of their firms, using a system that grants credits for foreign taxes paid and permits tax deferral for unrepatriated income. This paper shows that such tax systems encourage firms to restrict their equity stakes in new foreign investments, and to finance their new investments with considerable debt. These incentives exist even when transfer price regulation effectively limits the profit rates foreign subsidiaries can earn. The behavior of the foreign subsidiaries of US multinationals in 1984 appears to be consistent with these incentives. Key

words:

JEL

Classification:

Foreign tax credit; Foreign direct investment; H87;

H25;

Multinational

firms

F23.

1. Introduction

Foreign direct investment plays an important part in the lives of modem economies. Tax policy has the ability to exert considerable influence over the level of new foreign investment, and the behavior of established foreign+ 1 617 496 5960; email: * Corresponding author. Tel. : + 1 617 495 1340; fax: [email protected]. * This is a revised version of a paper presented at the NBER Summer Institute in August 1989 and at the Munich TAPES conference in June 1992. I thank Rosanne Altshuler, Anthony Atkinson, David Bradford, Roger Gordon, Kai Konrad, Hans-Werner Sinn, Joel Slemrod, and two anonymous referees for helpful comments. 0047-2727/94/$07.00 0 1994 Elsevier SSDI 0047-2727(94)05008-6

Science

B.V. All rights

reserved

324

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

owned firms. As a consequence, governments devote considerable effort to tailor their tax systems in order to the meet their own revenue needs while not excessively discouraging foreign investors. Naturally, in order to design effective tax systems, it is necessary to understand the incentives that foreign investors face. The purpose of this paper is to analyze the effect of international tax systems on foreign direct investment, and in particular, to analyze the incentives built into worldwide tax systems that grant foreign tax credits and permit tax deferral for outbound investments. The United States uses such a system to tax the foreign earnings of subsidiaries of US corporations, and a number of other capital-exporting countries use similar systems. The second section of this paper outlines the US system of taxing international income, and reviews other work on the impact of tax law on foreign investment decisions. Section 3 considers the impact of credit and deferral systems on the dynamic investment strategies US multinationals use in making foreign investments. Home-country tax systems can encourage firms to finance their foreign investments with limited amounts of parent equity. The reason is that, by keeping initial investment levels down, firms provide themselves profitable opportunities to reinvest their earnings as they accumulate, thereby creating greater opportunities for tax deferral than otherwise would be available. Firms that restrict their initial foreign investment levels may then find it profitable to make short-term loans to their foreign subsidiaries in the early years of their investments. Since the underinvestment phenomenon is likely to be most pronounced in low-tax foreign countries, the paradoxical result is that multinational parent firms located in high-tax countries may find themselves making loans to their subsidiaries in low-tax countries. One difficulty that arises in analyzing the dynamic path of a subsidiary’s growth is that production functions often imply that subsidiaries earn pure profits. The source of these profits is, presumably, either serendipity or the presence of intangible assets that the parent firm contributes to its subsidiary. The tax systems of most capital-exporting countries like the US do not permit their own firms to exploit their (home-country) patents, trademarks, and other intangibles to earn pure profits in foreign jurisdictions without somehow attributing the pure profit components to the homecountry’s taxing jurisdiction. One way in which this restriction applies in practice is that governments look critically at the profit rates of foreign affiliates, particularly those in low-tax foreign countries. Section 4 reconsiders the incentives facing firms making foreign investments, taking as its premise that tax rules cap the allowable profit rates that foreign subsidiaries can earn. Under this assumption, the home-country parent corporation, and not the foreign subsidiary, becomes the residual claimant on the excess profits of a subsidiary’s marginal investments. The

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

325

section, along with some extensions in Appendix A, shows that the standard view (Hartman, 1985) of the investment incentives facing mature subsidiaries need no longer hold if tax laws are enforced in this way, since the marginal investment decision turns, in part, on its effect on the allocation of profits between parent and subsidiary. This interpretation of the tax rules also carries important implications for the dynamic path of a subsidiary’s capital accumulation, and indeed, appears to simplify the analysis relative to the more standard treatment of section 3. The theoretical sections of this paper complement a number of other studies that analyze the effect of home-country taxation on the dynamic behavior of foreign subsidiaries. Sinn (1990, 1991) analyzes the incentives parent firms have to underinvest equity in their foreign subsidiaries. Newlon (1987) also analyzes this incentive, and observes that it carries the implication that parent firms will finance their foreign subsidiaries with debt. This paper looks closely at the effects of foreign country tax rates on incentives to finance foreign subsidiaries with debt, finding that the fraction of initial investment financed with debt can be zero (or 100%) and may be inversely related to local tax rates. Other studies do not consider the impact of transfer price regulation on the growth path of foreign subsidiaries (analyzed here in section 4), or the steady-state effects of impure foreign tax credit systems (analyzed in the Appendix A). The Appendix also considers the effect of differences between home and host country tax definitions of subsidiary profits. Hines (1988) and Leechor and Mintz (1990) examine the steady-state effects of these differences, but the Appendix shows that steady states may not exist when the differences are important. Section 5 examines whether the behavior of the foreign subsidiaries of US firms is consistent with the model sketched in section 4. Evidence from 1984 suggests that interest payments (and, by implication, the debt levels of US subsidiaries) exhibit a nonlinear sensitivity to local tax conditions, which the model implies. Section 6 is the conclusion.

2. The tax system, its interpretation

and application

This section outlines the US system of taxing international corporate income, and reviews some earlier theoretical work on its implications. The focus in this section, as in the work that is reviewed, is on systems of tax credits and tax deferral for foreign subsidiaries. 2.1. The US system

of taxing international

income

The United States taxes income on a residence basis, meaning that American corporations and individuals owe taxes to the US government on

326

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

all of their worldwide income, whether earned in the United States or not. In order to avoid subjecting American multinationals to double taxation, US law provides a foreign tax credit for income taxes (and related taxes) paid to foreign governments. The US corporate tax rate is currently 35%. With the foreign tax credit, a US corporation that earns $100 in a foreign country with a 15% tax rate pays a tax of $15 to the foreign government and $20 to the US government, since its US corporate tax liability of $35 (35% of $100) is reduced to $20 by the foreign tax credit of $15. The foreign tax credit is limited to US tax liability on foreign income. In the example, if the foreign tax rate were 50%, then the firm would pay $50 to the foreign government, and no taxes to the US government. The US firm would not receive a tax rebate from the US government, because its US foreign tax credit would be limited to $35. Hence a US firm receives full tax credits for its foreign taxes paid only when it is in a ‘deficit credit’ position, i.e. when its average foreign tax rate is less than its domestic tax rate. A firm has ‘excess credits’ if its available foreign tax credits exceed US tax liability on its foreign income.’ Deferral of US taxation of certain foreign earnings is another important feature of the US international tax system. A US parent firm is taxed on its subsidiaries’ foreign income only when returned (‘repatriated’) to the parent corporation. This type of deferral is available only to foreign operations that are separately incorporated in foreign countries (‘subsidiaries’ of the parent) and not to consolidated (‘branch’) operations.2 The US government taxes the profits of foreign branches of US companies as they are earned.3

1 Furthermore, income is broken into different functional ‘baskets’ in the calculation of applicable credits and limits. In order to qualify for the foreign tax credit, firms must own at least 10% of a foreign affiliate and only those taxes that qualify as income taxes are creditable. In addition, there are some complications in the calculation of tax credits that are discussed further in the Appendix. ’ The nomenclature is rather involved. All foreign operations take place through affiliates; those that are separately incorporated are subsidiaries. 3The deferral of US taxation may create incentives for US firms to delay repatriating dividends from their foreign subsidiaries. In 1962, Congress enacted the Subpart F provisions in part to prevent indefinite deferral of US tax liability on income earned abroad that is continually reinvested merely in order to escape US taxes. Subpart F rules apply to controlled foreign corporations (CFCs), which are foreign corporations owned at least 50% by US persons holding stakes of at least 10% each. The Subpart F rules include provisions that treat a CFC’s passive income (and income invested in US property) as if that income were distributed to its American owners, so it is subject to immediate US taxation. CFCs that reinvest their earnings in active foreign businesses avoid the Subpart F restrictions and can continue to defer US tax liability on those earnings. The Tax Reform Act of 1986 further expands the coverage of Subpart F, and also makes currently taxable the income of American investors in passive foreign investment companies that do not qualify as CFCs because they do not meet the 50% ownership rule.

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

327

2.2. Incentives created by credit and deferral systems

One of the surprising aspects of an American-style credit and deferral system is that it need not discourage foreign investment by firms in certain situations. Hartman (1985) shows that home-country taxation of dividendpaying subsidiaries in low-tax locations does not affect their steady-state investment behaviorP Newlon (1987) and Sinn (1990, 1991) confirm the Hartman result, but find that home-country taxation can exert an important effect on the time profile of capital accumulation for a subsidiary prior to the steady state.

3. Investment

incentives

over the life of a subsidiary

This section examines the effect of tax credits and deferral on the time pattern of investments by a multinational firm’s foreign subisidiary. The subsidiary is assumed to be located in a foreign country with a tax rate (T*) lower than the home-country tax rate (7). Furthermore, internal economies are assumed to require the parent firm and its subsidiary to be owned jointly, so parents cannot avoid taxes by selling profitable subsidiaries to investors who are not subject to home-country taxation.5 Finally, moral hazard considerations are taken to imply that any debt used to finance start-up subsidiaries must come from their parent firms. The object of the models in this section and in section 4 is to identify the effect of taxation on the incentive to finance subsidiaries with debt. Sinn (1990, 1991) does not analyze the impact of tax rules on borrowing by subsidiaries. An earlier study by Newlon (1987) considers debt finance, but does not examine the impact of foreign tax rates on the magnitude of initial borrowing. 3.1. The time profile of a projitable subsidiary Consider the incentives facing a firm that produces output with a concave 4 See Appendix A for a formalization and extension of Hartman’s model, and an analysis of the importance of tax base definitional differences to the steady-state behavior of dividendpaying subsidiaries. Stewart (1986) and Altshuler and Fulghieri (1990) analyze the incentives subsidiaries face in other circumstances. 5 This incentive exists even though the proceeds of such a sale would be treated as a dividend for tax purposes. The reason that such a sale would be tax favored is that part of the sale proceeds would be classified as return of paid-in equity, and would be untaxed by the US. An alternative way to get around home-country taxation is for the parent corporation to move its site of legal residence, but this is not easily done; see Hines (1991) for a discussion of the issues involved.

328

J.R. Hines,

Jr. I Journal of Public Economics

55 (1994) 323-347

production function Q(K: ), in which K: is the capital stock employed by the subsidiary in year t, and the Q( * ) function subsumes profit-maximizing choices of labor and other inputs. For the moment, take Q to be output net of capital depreciation, and assume that home and host countries’ tax systems apply true economic depreciation for tax purposes. Assume that the output is sold locally at an unchanging price, taken to be unity and parametric to the firm; units are defined so that the exchange rate between domestic and foreign currencies is one, and is constant over time. The parent firm chooses the real and financial policies of its subsidiary to maximize the present value of the parent’s after-tax cash flow. Let A represent the factor used to discount after-tax cash flows (in the hands of the parent corporation). Denote by D, the dividend payment from the subsidiary to the parent in period t; by definition, D, >O. Home-country taxation of foreign-source income reduces the after-tax value of a dividend payment of D, to Dl( 1 - r)/( 1 - T*) (see Appendix A for a derivation). Denote by E, the flow of equity funds from the parent to the subsidiary in period t. It is possible that E,< 0, but only in those cases in which the subsidiary has already paid out all of its after-tax foreign profits, since tax laws do not permit equity repatriation until profit repatriation is complete. Home countries do not treat the repatriation of initial equity as taxable incomeP Let S, denote the stock of accumulated parent equity in the subsidiary. Introduce a new state variable K defined as the net worth of the subsidiary: K, equals S, plus accumulated reinvested profits. Let B, denote the level of borrowing by the subsidiary, and r* the interest rate the subsidiary faces on this borrowing. Calculations in Appendix A indicate that the firm maximizes the present value of the stream: [D,(l - r)/( 1 - T*) - E,]exp( - At)

subject to the constraints

(1)

that

D,,B,,S,,K,sO

(2)

Either (i) E,aO, or (ii) (S, - K,)20, or (iii) both E,20 and (S, - K,)aO

(3)

dK,ldt = {[Q(K:)- r*B,](l - r*) - D, + EJ 6 This is the basis of Jun’s (1989) explanation of some curious features of the financial flows between US parent companies and their foreign affiliates. For an alternative explanation, see Hines and Hubbard (1990).

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

329

dS,ldt = E,

(5)

K; = K, + B,

(6)

The current-value hamiltonian subject to (2)-(6) is

corresponding

to the maximization

of (1)

H=[D,(l-7)/(1-T*)-EJ + Al{[Q(K, + B,) - r*B,](l -T*) - D, + E,} + h,E, + A,D, + A,B, + A& + A,K, + A,E,[l -A&

- K,)] (7)

in which the multipliers A, and A, correspond to the laws of motion of the state variables K, and S,, respectively, and the multipliers A, - A, reflect the inequality constraints in (2) and (3). The first-order conditions that characterize the maximum of (7), assuming the appropriate continuity conditions to hold, are: aHlaD, = [(l - r)/(l

-T*)]

-A, + A, = 0

aH/aE, = - 1 + A, + A, + A,[1 - A,(,.$ - K,)] = 0 aHlaB, = A,(1 - T*)[Q’(K;) while the costate equations A,Q’(K:)(l

-r*)

-

r*] + A, = 0

(8) (9) (10)

are:

- AA, + A, + A,A,E, = - dA,ldt

A, - AA, - A,A,E, = - dA,ldt

(11) (12)

3.2. Implications There are several aspects of subsidiary behavior implied by conditions (8)-(12). The first is standard in the analysis of foreign tax credits and deferral: that subsidiaries in low-tax locations do not simultaneously receive equity transfers and pay dividends. For mature firms with positive net equity, A, = 0, A, = 0, and A, = 0. Then (8) and (9) imply [(I-+(I-r*)]+Ag=l-A7

(13)

A, + A, = [(r - r*)/(l - r*)] > 0

(14)

or

The multiplier A, corresponds to the constraint that dividends are nonnegative, while A, corresponds to the constraint that equity infusions are non-negative. Since the two multipliers sum to a positive number, at least one of the constraints must bind, and the subsidiary does not simultaneously pay dividends and receive equity transfers.

330

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

A second lesson of (8)-(12) is that, from (lo), subsidiaries borrow funds from their parent firms (or from third parties) as long as Q’(KF) 2 I*. A third implication of (8)-(12) is that the financing arrangements of the subsidiary are discontinuous in time. Consider first the steady state; the steady state of the subsidiary is, as Hartman (1985) showed, achieved when Q’(K,*)(l - T*) = A. Equation (11) illustrates this condition: for a mature subsidiary paying dividends (and with positive net worth), the value of K* is unchanging, and so, therefore, is Q’(K:). Since E, = 0 for such a subsidiary, and A, = 0, it must be the case that - dA,ldt = - A,[A - Q(K:)(l -r*)]. Since the subsidiary pays dividends in the steady state, A, = 0, and from (S), A, = [(l - r)/(l - r*)] # 0; consequently, the steady state value of Q’(K:)(l -T*) = A. Prior to the steady state, the value of A, steadily declines over time, since - dh, ldt = Al[Q(K:)(l - r*) - A] > 0. Differentiating (9), and imposing that A, = A, = 0 yields that dh, ldt = - dh, ldt > 0. The implication is that the parent transfers equity to its subsidiary (so that E, > 0 and A, = 0) either in the initial period or never. The parent cannot transfer equity to its subsidiary in more than one instance, since a continuously declining A, cannot take the value zero more than once. Similarly, differentiating (8) yields that dA,ldt = dA,ldt < 0 prior to the steady state. Hence A, takes the value zero, and dividends are positive, only in the steady state. Finally, it should be noted that the (mathematical) complications that surround the repatriation of initial equity, and the associated values of A,, A,, and A,, are not important, since the subsidiary stays in the steady state (with Q’(K:)(l - r*) = A and dh, ldt = 0) once capital accumulation proceeds far enough to drive down the marginal product of capital sufficiently. 3.3. The role of debt Fig. 1 illustrates the pattern just described. The idea is that a subsidiary may be financed in the initial period with some equity and some debt. There are several possible regimes; Fig. 1 pictures the standard case. In this case, the subsidiary is funded initially with E, in equity. Let r( - ) be the inverse function of the marginal product of capital, so that r[Q’(K:)] = KT. A subsidiary borrows to finance its investments as long as Q’(K:) > r*. In a frictionless capital market, subsidiary borrowing guarantees that this condition is met with equality, and that the physical capital employed by the subsidiary never falls below r(r*). If E,
J.R.

TIMl~*)l

Hines, Jr.

I Journal of Public Economics

331

55 (1994) 323-347

..._____L.-.........-----.S K’ 1’ t

-Subsidiary(physical)capital

Dividends SubsidiaryNet Worth

D, I I I

c

0 t1

t2

Time

Fig. 1.

During this phase, the physical capital employed by the subsidiary does not change. At a certain point (tl in Fig. l), the subsidiary accumulates enough net worth that S, 3 r(r*), and there is no need (or profitable capacity) to borrow any more. From that point on, the subsidiary grows through reinvesting its earnings, until the marginal product of capital is driven down to the value identified by Hartman: Q’(K<*) = A/(1 - T*). At this point (t2 in Fig. l), the subsidiary’s capital stock equals T[A/(l - 7*)], and the subsidiary stops accumulating; thereafter, it pays annual dividends equal to its net output, maintaining the same capital stock indefinitely. It might appear paradoxical that a firm chooses to finance its foreign subsidiary partly with debt and partly with parent equity, since one generally expects the cheaper alternative to dominate the other. The reason that firms typically choose both debt and equity is that the marginal cost of equity finance rises as more equity is used; the marginal cost of equity includes not only the opportunity cost of funds but also the effect of additional equity on total returns. Equity investments in subsidiaries are valuable not only because subsidiaries generate profits, but also because subsidiary profits can be reinvested abroad where the returns are lightly taxed (until repatriated). The value of deferral depends on the length of time between initial The deferral interval shortens as investment and ultimate repatriation. additional equity is invested in the subsidiary, and consequently, the cost of marginal equity includes the cost of reducing deferral opportunities for the

332

J.R. Hines,

Jr. I Journal

of Public Economics

5.5 (1994) 323-347

profits earned by inframarginal equity. This cost rises as the total amount of equity increases. It should be noted that the first-order conditions do not guarantee that E, > 0, or that E, < T(r*). Consequently, it is possible that the subsidiary is financed with no parent equity, or alternatively, with no borrowing.7 The model does not predict a clear sign of the effect of T* on the level of E,. Sinn (1990, 1991), who analyzes a model of the same type but in which subsidiaries are assumed not to have access to debt finance,8 also finds the effect of r* on E, to be ambiguous. Low foreign tax rates imply that the subsidiary grows quickly through reinvesting its after-tax profits, since lower tax rates correspond to higher after-tax rates of return. Low tax rates also make deferral more attractive. Both of these considerations suggest that lower tax rates would be associated with low values of E,. However, the size of the steady-state capital stock (T[A/(l - r*)]) is inversely related to the foreign tax rate, so a low foreign tax rate may stimulate greater initial investment. There are sporadic proposals in the United States to eliminate deferral in order to reduce the volume of US capital exportedP But the comparison between the amounts of capital invested abroad under accrual and deferral systems turns on complicated considerations. Under accrual taxation, an equity-financed subsidiary invests up to the point at which Q’(K:)(l - r) = A, which corresponds to a capital stock that is smaller than that of a dividend-paying mature subsidiary. It is in this sense that proposals to eliminate deferral are consistent with the stated objective. However, the time profile depicted in Fig. 1 illustrates the possibility that firms respond to deferral by so underinvesting in the initial period that the effect of deferral might be to reduce the amount of capital invested abroad. Sinn (1990) derives some (weak) sufficient conditions for deferral to reduce intial investment in a model in which subsidiaries do not have access to debt finance. If, however, A = r*(l - r), then the initial capital stock (inclusive of the debt-financed portion) of a subsidiary whose home country defers tax liability until repatriation satisfies: Q’(K:) = A/(1 - T), the same condition satisfied by subsidiaries not eligible for deferral. The condition A = r*(l - T) ’ This assumes, of course, that local tax authorities would permit (in the first case) such thin financing of a subsidiary. Newlon (1987) offers a proof that debt is used before the period of internal growth of the subsidiary, but the proof misses the possibility that E, > T(r*). No doubt responsibility for this oversight must be borne in part by the members of his Ph.D. committee. * It is possible to reinterpret the production functions in Sinn (1990, 1991) as generally subsuming the availability of debt finance. The model analyzed here makes explicit the role of debt finance, thereby restricting the net-of-debt production function in a particular way, but one that does not resolve the ambiguity about the effect of T* on E,,. Newlon (1987) is the first study to examine explicity the use of debt by foreign subsidiaries. 9 See, for example, McIntyre (1989).

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

333

characterizes a world capital market in which the (before-tax) rate of return r* is available to all investors, and in which parent firms discount after-tax cash flows at the after-tax rate of interest. Consequently, the availability of debt finance makes it unlikely that deferral reduces investments by subsidiaries, even in the early stages of investments. 3.4. Alternative repatriation strategies The preceding analysis relies on the assumption that foreign subsidiaries have only two alternative uses for their after-tax profits: profits can remitted to the parent as dividends or they can be reinvested. The attractiveness of the latter use is ultimately limited by the (assumed) concave shape of the subsidiary’s production function. Firms have a third alternative, since they are able to defer home-country taxation of profits earned by foreign subsidiaries if the profits are reinvested in other active foreign businesses. Information asymmetries or the absence of internal economies may, however, make it difficult for foreign-based multinational firms to find a sufficient volume of attractive investment outlets to absorb their accumulated profits. If so, then the concavity of the production function Q( *) accurately reflects the subsidiary’s alternatives inclusive of outside investments. Firms have a fourth alternative of passively investing accumulated subsidiary profits in world capital markets. Home countries such as the United States tax the returns of passive investments as earned, but the passive investment strategy may offer the advantage of deferring homecountry taxation of subsidiary profits that constitute the investment principle. Hines and Rice (1994) and Scholes and Wolfson (1992) examine the conditions under which subsidiaries maximize after-tax returns by passively investing accumulated profits. There are, however, two difficulties that such a strategy can encounter, The first is that the available rate of return may not be sufficient to warrant making passive investments, in spite of their tax advantages. This problem arises when the pre-tax component of the firm’s discount rate [A/( 1 - T)] is higher than the available rate of return on passive investments in world capital markets. This scenario requires that access to capital markets is difficult or costly for the affected firms.” The second difficulty with the passive investment strategy is that some

lo Possible sources of capital market imperfection include asymmetric information, moral hazard, and the costs incurred in bankruptcy. Recent evidence that internal cash flow exerts an independent effect on firm investment (Fazarri et al., 1988; Blanchard et al., 1993) is suggestive of certain types of capital market imperfections. Of course, if firms have perfect access to bond markets, then the corporate income tax should exert no influence over their investment decisions (Stiglitz, 1973).

334

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

home-country tax systems penalize subsidiaries that accumulate passive assets. A passive investment strategy implies that the fraction of a subsidiary’s net worth invested in passive assets increases over time, since its level of active investment does not change, and it uses all of its after-tax earnings from active investments to purchase passive assets. American tax law treats subsidiaries as ‘Passive foreign investment companies’ (PFICs) if more than half of their assets produce passive income or if more than 75% of their income is passive. All income (passive and active) earned by PFICs is taxed as though immediately repatriated, and firms are not eligible to claim foreign tax credits for foreign income taxes paid by certain PFICs. This represents a serious tax cost for passive investment, particularly since a subsidiary that once is deemed a PFIC always remains a PFIC. Recent (1993) US legislation adds a separate requirement that subsidiaries with greater than 25% of their assets producing passive income treat the passively invested principal as if immediately repatriated. Firms that can avoid home-country taxation of dividend receipts from foreign subsidiaries need not select investment profiles like that depicted in Fig. 1. In the extreme case, firms that freely borrow and lend at world interest rates equal to [A/( 1 - r)], and are not subject to special homecountry tax liabilities on passive income, have no incentives to underinvest in their subsidiaries. Historically, US subsidiaries report little in the way of passive income subject to Subpart F (see the data reported in Hines and Hubbard, 1990), and, for the remainder of the paper, I assume that the combination of low rates of return and special tax treatment make investment in passive assets an unattractive alternative to dividend repatriations.

4. Investment

incentives

with regulated

returns

One difficulty with the preceding analysis is that subsidiaries earn pure economic profits of unidentified source. In what follows, I assume that the profitability of the subsidiary stems from home-country activity on the part of its parent firm. Examples of such activity might include advertising in the home country, spending money to develop patents, or strategic investments designed to limit competition from other firms. Under US law, and the laws of most other developed, capital-exporting countries, it is incumbent on a foreign subsidiary to remit rents or royalties to its parent company in return for the use of any intangible assets that generate pure profits for the subsidiary. These royalties are taxable in the home country, and can be deducted from taxable income in (most) host countries. Royalties should, in principle, reflect the true economic value provided by the parent to the subsidiary. For the purpose of the following analysis, I will assume that the royalty is

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

335

chosen to afford the subsidiary a normal (before-tax) return on capital invested. This assumption represents a reasonable interpretation of the impact of Section 482 and related legislation designed to prevent firms from transferring profits to tax-favored foreign locations. Indeed, rates of return typically are used by the courts in determining the appropriateness of royalties and transfer prices.” The goal of this section is to identify the degree to which such a system of tax enforcement changes the implications of the analysis described in section 3. Regulated subsidiary profit rates simplify the analysis of growing subsidiaries, since subsidiary growth rates do not depend on the precise forms of their production functions. Let p denote the maximum rate of return that home-country tax authorities permit the subsidiary to earn on its net worth. I consider cases in which subsidiaries earn sufficient inframarginal profits that the maximum rate of return constraint always binds. Any difference between (before-tax) subsidiary profit and pK, must be paid to the parent firm as a royalty. In period t, the parent firm receives from its subsidiary a dividend of D, and a royalty equal to {Q(K:) - r*B, - pK,}. Royalties represent taxable income in the home country. The system of royalty payments changes the equation of motion for the subsidiary’s capital stock, which now becomes: dK,ldt = [pK,(l - r*) - D, + E,]. Hence the currentvalue hamiltonian that corresponds to maximizing the parent firm’s after-tax profits is H = [D,(l - ~)/(l -T*) - E,] + (1 - T)[Q(K, + B,) - r*B, - pK,] + h,{[pK,(l

-T*) - D, + E,} + A,E, + A@,

+ h4B, + ASS, + A,K, + A,E,[l - A&

- K,)]

(15)

The first-order conditions that characterize the maximum of (15) are identical to Eqs. (8)-(10) that characterize the maximum of (7). The only difference between the two solutions appears in one of the costate equations (1 - T)[Q’(K,*)

-p]

+ A+(1 - r*) -AA, + A, + A,A,E,

= - dA,ldt

(16) which replaces Eq. (11) for the case of a subsidiary with regulated profits. Since the first-order conditions that correspond to maximizing (15) are the same as the first-order conditions that correspond to maximizing (7), the first two implications derived earlier - that subsidiaries in low-tax locations do not simultaneously receive equity transfers and pay dividends, and that subsidiaries borrow as long as Q’(K:) 2 r* - are implied by (15). The ‘I See Berry et al. (1992) for a critical discussion of recent US court decisions. Courts use rates of return as a concession to the vast complexity that surrounds the analysis of separate profitability.

336

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

steady state behavior of a dividend-paying subsidiary is characterized by dh, ldt = 0, A, = A, = 0, and A, = [(l - r)/(l - r*)]. Substituting these values into (16) yields the familiar condition that Q’(K:)(l - r*) = A. As before, it is useful to consider the case in which parent firms discount after-tax cash flows at the after-home-tax world interest rate, so A = r*(l 7). If tax authorities require subsidiaries to earn (before-tax) profits at a rate equal to the before-tax interest rate, then p = r* and A = ~(1 - r). Substituting the last equation into (16), and imposing that A, = A, = 0, yields: - dhildt = (1 - r)[Q’(K:) - p] + A,p(r - G-*). It is straightforward to verify that - dh, ldt > 0 prior to the steady state, since A, takes an initial value of 1, its steady-state value is [(l - r)/(l - T*)], and dh, ldt < 0 for any value of A, in the interval ([(l- r)/(l -T*)], 1). H ence the subsidiary’s financial policy exhibits the intertemporal discontinuity pictured in Fig. 1: the parent transfers equity to its subsidiary either in the initial period or never, and the subsidiary pays positive dividends to the parent only upon reaching the steady state. It is possible to solve analytically only for certain aspects of the subsidiary’s behavior. Appendix A analyzes the steady-state behavior of subsidiaries of parent firms located in countries that tax and regulate subsidiary profits while providing something other than pure foreign tax credits. The dynamic analysis presented in this section indicates that the pattern of Fig. 1 continues to describe the time pattern of subsidiaries even when returns are regulated. Analytical details of a subsidiary’s investment pattern continue to depend on the form of the subsidiary’s production function, since total output of the subsidiar,y determines the amount of royalties it pays. Nevertheless, a clear pattern emerges in which subsidiaries in low-tax countries borrow in the early stages of their development and pay dividends only upon maturity. An important remaining question is whether the behavior of subsidiaries conforms to this prediction.

5. Evidence

from the behavior

of US multinationals

This section evaluates the dynamic models presented in sections 3 and 4 by considering evidence on the behavior of the foreign subsidiaries of US multinationals. The goal is to analyze the dynamics of behavior rather than the determinants of levels of international investment. Two important pieces of evidence appear to be consistent with the pattern predicted by the models. The first is that only a small number of subsidiaries remit dividends to their parent firms each year, which is consistent with a model in which the majority of US-owned foreign subsidiaries are still in the growth phase. The second piece of evidence is that interest payments by US-owned foreign subsidiaries to their American parent firms bear the type of nonlinear relationship to host-country tax rates that the dynamic models imply.

J.R.

Hines, Jr.

I Journal of Public Economics

5.1 Studies of the determinants of foreign

55 (1994) 323-347

337

direct investment

A number of studies apply the Hartman model to estimate the impact of taxes on international investment. Since the Hartman framework implies that home-country taxation does not affect investments by subsidiaries out of their retained earnings, it is common to test the model by regressing outbound foreign direct investment on home-country tax rates. US balance of payments data on foreign direct investment conveniently distinguish between investments out of retained earnings and investments financed by new equity transfers. The Hartman model is taken to imply that investments out of retained earnings will be unaffected by home-country taxation. By contrast, it is usually assumed that investments financed by new equity transfers are discouraged by high rates of home-country taxation of repatriated profits. Studies usually find that the after-foreign-tax rate of return in host countries affects foreign investment out of retained earnings, but that tax variables perform poorly in explaining equity transfer components of foreign direct investment.” These empirical investigations are difficult to interpret for two reasons that the dynamic models illustrate. First, the Hartman model implies that, for any given mature subsidiary, investment out of retained earnings is unaffected by its home-country tax rate. This does not imply that aggregate investments out of retained earnings are unaffected. Since host country tax rates affect the level of initial equity transfers, and hence the scale and number of ongoing operations by multinational subsidiaries, taxes should also affect the aggregate level of reinvested earnings, even if the last are simply constant functions of subsidiary size. Second, the Hartman model applies only to mature subsidiaries that repatriate dividends, and these may represent a distinct subset of the subsidiaries in any particular country.

5.2. Evidence concerning subsidiary dynamics There is’ already considerable support for. the view that subsidiaries experience a period of internally-funded growth during their development. Hines and Hubbard (1990) ,report that 84% of US controlled foreign corporations (representing 67% of the total assets of all US controlled foreign corporations) made zero dividend payments to their American parent firms in 1984. Altshuler and Newlon (1993) report similar findings for US controlled foreign corporations in 1986. A second type of evidence is available in the reported financial transactions between US parent firms and their foreign subsidiaries. The models of I2 This literature is summarized in Slemrod (1990). Studies include Hartman (1981, 1984), Boskin and Gale (1987)) Newlon (1987), Slemrod (1990), and Jun (1989).

338

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

sections 3 and 4 imply that a subsidiary is financed with debt in the early stages of its development, and that this debt is repaid as soon as possible with profits earned by the subsidiary. US parent firms are required to charge market interest rates on loans to their foreign subsidiaries. Taking this rate to be the same for all subsidiaries, the model implies that subsidiaries differ in their interest payments to US parents insofar as they have different capital stocks, receive differing amounts of initial equity from their parents, and are in different stages of development. The models described in sections 3 and 4 do not offer closed-form solutions indicating the impact of host-country tax rates on the indebtedness of local subsidiaries, but one pattern is clear. If the local tax rate is sufficiently high (T* close to T), then the subsidiary experiences only a very short period of internal growth prior to reaching its steady state. There exists a wide region over which E, > Qr*) and subsidiaries use no debt finance from their parent firms. For firms in this region, small changes in host-country tax rates (T*) will affect E, without affecting borrowing from parent firms, since subsidiaries are not close to the margin of borrowing from their parents. As the discussion in section 3 suggests, theory does not predict unambiguously the effect of local tax rates on E,. Initial borrowing is inversely related to E, for those subsidiaries that borrow from their parents. The preceding discussion makes it clear, however, that any impact that local tax rates have on borrowing must be nonlinear, since as local tax rates rise subsidiaries eventually stop borrowing from their parents, and any further tax increases obviously have no effect. This process influences the aggregate crosssectional picture of subsidiary borrowing, and suggests that debt owed to parent firms should be inversely related to local tax rates, though with a nonlinear effect that vanishes as the foreign tax rate approaches the home tax rate. The idea of the empirical work is to examine the effect of tax rates on aggregate interest obligations, scaled by measures of the size of activities of US-owned foreign subsidiaries in differing countries. In the models of sections 3 and 4, the ratio [Interest payment/(Profits + Interest payment)] reflects simply the interest rate used and the degree to which a subsidiary is financed with debt from the parent. Table 2 reports regressions of this ratio on tax rates faced by US controlled foreign corporations located in various foreign countries in 1984.13 The regressions use data on the aggregate behavior of all US firms in 1984, as reported by Bradford (1990); the data suffer from the usual problems of aggregation, mixing excess foreign tax I3 Table 1 reports summary statistics for the financial variables. Local tax rates are calculated as the reported ratios of local income taxes paid to local (pre-tax) earnings and profits. Countries were excluded from the sample if, in the aggregate, US firms had negative earnings and profits in 1984.

J.R. Table

Hines, Jr.

I Journal of Public Economics

339

55 (1994) 3.23-347

1

Variable

means

and standard

deviations,

US controlled

foreign

corporations,

Variable

Mean

Standard

Interest/(profits + interest) Interest/[(profits + interest) - D/(1 -T*)] log[interest/dividends] log[interest(l - r*)/dividends] r* r*2

0.1440 0.2244 -0.6616 -1.1821 0.3794 0.1747

0.1282 0.1937 1.4282 1.4979 0.1769 0.1349

1984

deviation

No. obs. 57 57 57 57 57 57

Note. Variables describe the country-by-country aggregate behavior of US controlled foreign corporations (CFCs) in 1984. The variable ‘Interest’ is the sum of interest, rent, and royalty payments from CFCs to their US parent companies. ‘Profits’ is the before-tax earnings and profits (net of interest payments) of US CFCs. ‘Dividends’ (also abbreviated ‘D’) is dividend payments by CFCs to their US parent companies in 1984. The variable r* is the average foreign tax rate paid by US CFCs in each country in 1984.

credit firms with deficit foreign tax credit firms, and the difficulty of measuring the appropriate foreign tax rate. The striking feature of the regressions reported in the first two columns of Table 2 is that the tax effect, which, when entered linearly, is zero, exhibits considerable curvature. When tax rates are low, higher tax rates encourage encourage parent firms to lend to their subsidiaries, while this effect disappears as the foreign tax rate approaches 40%. Another implication of the models in sections 3 and 4 is that, in low-tax Table 2 Foreign subsidiary

interest

payments

Dependent

and foreign

variable

Interest/ (profits + interest) Constant r*

Interest/ [profits + interest

-D/(1

- r*)]

0.1387 (0.0330) 0.1388 (0.7877)

0.0329 (0.0414) 0.7584 (0.2796) -1.0112 (0.3654)

0.1908 (0.0517) 0.0886 (0.1342)

0.0315 (0.0628) 1.2094 (0.3831) -1.5223 (0.4840)

0.0004 0.129 57

0.078 0.125 57

0.007 0.195 57

0.083 0.189 57

r*2

Adj. R2 & II

tax rates

Note. Coefficients are estimated based on the country-by-country aggregate behavior of US controlled foreign corporations (CFCs) in 1984. The variable ‘Interest’ is the sum of interest, rent, and royalty payments from CFCs to their US parent companies. ‘Profits’ is the before-tax earnings and profits (net of interest payments) of US CFCs. The variable Dl( 1 - r*) represents dividend payments by CFCs to their US parents divided by (1 - r*), in which r* is the average foreign tax rate paid by US CFCs in each country in 1984. Values in parentheses are heteroskedasticity-corrected standard errors.

340

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

jurisdictions, dividend-paying firms are not the same firms that borrow. Consequently, the propensity to borrow to finance investments by immature firms can be estimated by removing dividends from before-tax profits. In the next set of regressions, the dependent variable is adjusted by subtracting from the denominator dividends paid (divided by [l - T*], so that the dividend measure corresponds to pre-tax profits from which the dividends are paid). These regressions effectively restrict attention to subsidiaries in the growth phase, and measure interest payments relative to their profits. The results, reported in columns 3 and 4 of Table 2, are similar to those obtained without removing dividends. Table 3 reports estimated coefficients from a logarithmic specification of the same interest payment equations. The results are quite similar to those reported in Table 2.

6. Conclusion Domestic and foreign tax policies influence the investment behavior of multinational firms, and do so in important ways. Tax laws often encourage firms to defer dividend payments from foreign subdidiaries to their parent companies; knowing that, firms have incentives to limit the initial capitalization of their subsidiaries in order to permit profits to accumulate for lengthy periods of time before repatriation. This effect is present even when tax authorities effectively regulate the profits that subsidiaries can earn. Hence Table 3 Interest payments,

dividends, Dependent

Constant r*

& n

variable

log[interest/dividends]

log[interest(l

-0.4719 (0.4377) -0.4999 (1.1711)

-0.3472 (0.4517) -2.2006 (1.2223)

r*2

Adj . RZ

and foreign tax rates

0.004 1.438 57

-1.6714 (0.5884) 7.9392 (3.9732) -11.4626 (5.7326) 0.084 1.392 57

0.068 1.460 57

- r*)/dividends] -1.7196 (0.5928) 7.4556 (4.0186) -13.1158 (5.8138) 0.163 1.400 57

Note. Coefficients are estimated based on the country-by-country aggregate behavior of US controlled foreign corporations (CFCs) in 1984. The variable ‘Interest’ is the sum of interest, rent, and royalty payments from CFCs to their US parent companies. The variable ‘Dividend’ represents dividend payments by CFCs to their US parents in 1984. The variable r* is the average foreign tax rate paid by US CFCs in each country in 1984. Values in parentheses are heteroskedasticity-corrected standard errors.

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

341

one may not always observe large transfers of equity investments from parent companies to their subsidiaries located in the most attractive investment locations. Evidence from the behavior of the foreign subsidiaries of American firms in 1984 appears to reflect the tax incentives they face, since most subsidiaries pay no dividends at all to their parent companies, and the borrowing behavior of subsidiaries exhibits a nonlinear sensitivity to that tax policy exerts an local tax conditions. It appears, therefore, important effect on the financial policy that accompanies foreign direct investment, which is one step toward determining the influence of taxation on the level and direction of the real activities of multinational firms.

Appendix

A: Investment

tax incentives

and mature

subsidiaries

This appendix considers two complications to the investment incentives faced by mature foreign subsidiaries of firms based in countries that operate worldwide tax systems. The first complication is taxation on the basis of something other than a pure foreign tax credit system. The operation of the US foreign tax credit system, with its various limitations to available credits, occasionally takes this form. The second complication is the distinction between the definition of host-country tax base for the purpose of income taxation and the definition of home-country tax base for the purpose of calculating the foreign tax credit. Impure foreign

tax credit systems

Consider a subsidiary that earns before-tax profits of T in the foreign country, pays T*T in taxes to the foreign government, and remits a dividend of D to its parent company. In a pure foreign tax credit system, the subsidiary receives a home-country credit for (T*,T){D/[T(~ - T*)]} of foreign taxes paid. In an impure foreign tax credit system, the subsidiary might receive a credit for LY(T*~F){D/[T(~ - T*)]} of foreign taxes paid, with (Y not equal to unity.i4 Home country tax liability is the difference between the home country tax rate multiplied by ‘grossed up’ foreign income (the sum of the dividend and the foreign tax credit) and the foreign tax credit: i4 If a = 0 then this system is a deduction and deferral system; if, instead, 0 < (I < 1, then this system offers partial foreign tax credits. ((I = [r(l - T*)]/[T*(~ - T)] corresponds to exempting foreign income from home-country taxation.) The shadow value of foreign tax credits in practice often lies between zero and one. Under US law, any excess foreign tax credits may be carried forward five years (with no adjustment for inflation), and the ex ante uncertainty of their ultimate use drives the value of Q (this year) below one. See Scholes and Wolfson (1992) for a discussion of the shadow value of foreign tax credits.

342

J.R.

Hines, Jr.

I Journal of Public Economics

Tax = r[D + u(r*7r){Dl[7r(l

-r*)]}]

= D[T - cX7*+ (1 - (Y)rT*]/(l The after-home-tax which equals After-Tax

dividend

received

55 (1994) 323-347

- CX(T*Z-){Dl[~(l -r*)]}

r*>

(Al)

by the parent

firm is D - Tax,

Dividend = D(1 - 7)[1 - (1 - a)~*]/(1 - r*)

(A2)

and it is the present value of (A2) that the firm maximizes. Clearly, actions that maximize the present value of D also maximize the present value of (A2). Hartman (1985) shows that foreign tax credit systems, as well as home-country tax systems that tax dividends without providing foreign tax credits, give firms incentives to maximize the present value of dividends from subsidiaries. Firms respond by setting Q’(K:)(l -T*) = A. Eq. (A2) indicates that Hartman’s conclusion applies not only to foreign tax credit systems (a = l), and to deduction systems ((Y= 0), but to all other impure foreign tax credit systems. The implication of Hartman’s model is not quite so robust in settings of the type analyzed in section 4, in which the subsidiary’s rate of return for tax purposes is regulated by the home country. A subsidiary that pays dividends in a steady state must be indifferent between repatriating $1 of after-foreigntax profit as a dividend, with after-tax value given by (A2), and reinvesting the $1 for an additional period before repatriating. Deferring the dividend payment one year permits the subsidiary to pay a dividend of [ 1 + p( 1 - r *)] the following year, where p is the regulated rate of return for the subsidiary. Applying (A2), the after-tax value of this dividend is [l + ~(1 - ~*)][l T*(l - a)](1 - r)/(l - 7*>. I n addition, the parent firm’s royalty receipt from its subsidiary changes by an amount equal to the (after-tax) difference between p and the marginal product of the dollar reinvested: (lr)[Q’(K,*) - p]. Equating the present values of immediate repatriation and one-period delay implies Q’(K:)(l

- r*) = A - ~*(l - a)[A -p(l

-T*)]

(A3)

Eq. (A3) indicates that the Hartman result holds either if (Y= 1 (the home country grants full foreign tax credits for foreign taxes paid), or if p( lT*) = A. It is striking that the Hartman result for pure foreign tax credit systems is resilient to such a change in the interpretation of the tax rules. If, however, the firm receives only partial foreign tax credits, then home country taxation influences the size of the subsidiary’s capital stock in the steady state. Higher values of p, the regulated rate of return, correspond to higher values of Q’(K:), and lower steady-state values of K,* (as long as (Y< 1). It is noteworthy, however, that home-country taxation influences capital accumulation by a subsidiary through (Y and p, and not through the domestic tax rate 7.

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

343

Tax base differences

Hartman’s analysis of the investment behavior of dividend-paying subsidiaries assumes that home and host governments define taxable income in the same way. In practice, this is often not the case. When there are differences between home-country and host-country definitions of taxable income, Hartman’s conclusion that the investment behavior of a dividendpaying subsidiary is unaffected by home-country taxation may no longer be valid. The reason is that the subsidiary’s dividend decision no longer turns just on the comparison between the after-tax value of a dividend this year and the dividend that could be paid (by reinvesting the money) the following year. If tax base definitions differ, any marginal investment that a subsidiary makes influences the foreign tax credits its parent company can claim for the inframarginal dividends the subsidiary is already paying. This effect arises because marginal investements change the average foreign tax rate paid by the subsidiary, as calculated using the home-country tax base definintion. Hines (1988) estimates of the impact of tax base differences on effective tax rates facing mature US subsidiaries located in various foreign countries. Leechor and Mintz (1990) make a similar point, and extend it to include differences in tax law definitions of interest deductions. These studies examine conditions that describe the steady states of subsidiaries. The function of this section of the Appendix is to show that subsidiaries may not exhibit steady-state behavior in cases in which tax bases differ significantly. Consider a firm facing the standard credit and deferral system described by Eq. (Al) and (A2), for the case in which LY= 1. Let 7rf represent the subsidiary’s foreign taxable income as defined by the foreign host country, and 7~,, its taxable income as defined by the home country. Let /3 represent the ratio of home-defined foreign-source income to foreign-defined taxable income:

The home government grants foreign tax credits based on the fraction of after-tax foreign income repatriated by subsidiaries as dividends. Hence the foreign tax credit is D[rrr*l(7r,, - r*nr)]. Homecountry tax liability is then

home-definition

Tax = rD{l + [r*+(rus

- T*T&]} - D[~~r*l(r,,

- race)]

= D(p7 - T*)l(p -r*) The after-home-tax After-Tax

(A5)

dividend received by the parent firm is

Dividend = Dp( 1 - T) /( /3 - r*)

(A@

344

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

And it is the present discounted value of (A6) that a firm maximizes, not the present value described in (A2) (for the case (Y= 1). Naturally, the two problems yield the same solution if /3 is constant. As a general matter, /3 is a function both of the dividends paid by the foreign subsidiary, and of the capital stock (and earnings) of the subsidiary, and can be written p(D,, K,). Other variables, such as previous years’ investments and dividends, may also be important. One example of a system in which /3 is a function of current dividends and current capital is the case in which foreign and home governments use first-year capital recovery allowances in place of depreciation schedules.r5 With this scheme, a foreign subsidiary investing Z, in year t gets an immediate tax deduction of Z,z against its home-definition foreign income and Z,z* against its taxable income in the foreign country. The effective foreign tax rate on foreign investments is then a function of z* as well as the foreign statutory tax rate. Denote this effective foreign tax rate as r*. Until 1986, the US tax system used the average (US-definition) foreign tax rate in the most recent applicable year to calculate foreign tax credits that accompany dividend repatriations.r6 These depreciation allowances were considerably less generous than those granted by foreign governments that provided accelerated depreciation in order to attract new investment. Replacing D,(l - r)/(l -r*) in (1) with the expression for after-tax dividends in (A6), the firm’s current-value hamiltonian becomes H = [D,(l - r)pl(p

- r*> - E,]

+ AI{[Q(K, + Z?,) - r*B,](l -T*) - D, + E,} + h,E, + A,D, + A,B, + A& + A,K, + A,E,[l - A& - K,)] (A7) the first-order conditions of which are identical to those obtained in section 3, with two exceptions. Eq. (8) is now replaced by aHlaD, = [(l - r)pl(p

-T*)]

- r*D,(l - ~)(aplao,)l(p

-7*)’

- A, + A, = 0 = [(l - +I(/3

- ~*)]{l

- T*T/(~ -T*)}

- A, + A, = 0

(A8)

is As proposed by Auerbach and Jorgenson (1980) in the domestic context. Their idea was to minimize investor risk arising from inflation-rate uncertainty. It is used here to simplify some of the analysis by restricting the number of state variables, though it may exaggerate the importance.of actual differences in tax base definitions. 16U.S. tax law changed in 1986 to introduce multi-year ‘pooling’ of foreign earnings and dividends to prevent perceived manipulation of the foreign tax credit system by multinational firms. The ‘pooling’ method still produces foreign tax credits that are sensitive to the issues described in this section, albeit in a more complicated way than was true before passage of the 1986 act.

J.R.

Hines, Jr.

in which n = D(ap/aD)lp, (11) is replaced by h,Q’(K:)(l -Ah,

-r*)

I Journal of Public Economics

55 (1994) 323-347

345

the elasticity of p with respect to D. In addition,

- r*D,(l - r)(ap/aK,)l(p

- T*)~

+ A, + A,A,E, = - dhildt

(A9)

From (A9), it is clear that Q’(K:)(l - r*) = A no longer describes the steady state of the subsidiary, since now there is an additional term, r*D,(l - T)(aplaK,)l(p - r*)2, in the equation of motion for A,. If (apl aK,) is nonzero, then A, = A, = 0 is not sufficient for the Hartman condition to hold. It is worth considering the likely signs of (a/3/X,) and (ap/aD,). If a foreign country offers attractive investment incentives, so that z* > z, then marginal investments are likely to raise the value of p, since marginal effective tax rates for countries with investment incentives are typically well below average effective tax rates.17 Consequently, (apl X,) > 0 and (apl aD,) < 0. From (A9), the steady-state condition requires that Q’(Kf )( 1 r*) > A. Hence, foreign investment incentives would not encourage subsidiaries to invest as much as the Hartman condition indicates. Another complication that appears when tax base definitions differ is that the steady-state condition (dhildt = 0 in (A9)) may not be consistent with the second-order condition corresponding to the maximum described by (A8). Recall that the second-order condition for maximizing firm value is that t12HlilD2 s 0. Differentiating (A8) with respect to D yields d2HldD2 = [/3(1- +*l(P

- T*)~]

x {%/aD - (aplaD)[l

- 27*77/(p -r*)]}

(AlO)

which, if (aplao) < 0 and is sufficiently large in absolute value, may imply that d*HldD* >O when the steady-state condition of (A9) is satisfied. Consequently, it may often be inappropriate to consider the steady-state behavior of subsidiaries facing investment tax incentives.‘* The reason is that a subsidiary, by paying out more today and less tomorrow, may raise the creditable foreign tax rate on many of the repatriated dividends. r’ This is as much a statement about the longevity of attractive investment incentives as it is about accounting definitions of profits. In some steady states, average and marginal tax rates are equal. I8 Leechor and Mintz (1990) examine whether a subsidiary will settle in a steady state configuration over its debt/equity choices, but go on to consider the steady-state behavior of firms with investment incentives, assuming that such a steady state exists and represents the optimal choice on the part of the firm. Altshuler and Fulghieri (1990), and, in truth, Hines (1988) also consider steady states without confirming that they exist.

346

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

References Altshuler, R. and P. Fulghieri, 1990, Incentive effects of foreign tax credits on multinationals, Columbia University Department of Economics Discussion Paper No. 478. Altshuler, R. and T.S. Newlon, 1993, The effects of U.S. tax policy on the income repatriation patterns of U.S. multinational corporations, in: A. Giovannini, R.G. Hubbard, and J. Slemrod, eds., Studies in international taxation (University of Chicago Press, Chicago) 77-115. Auerbach, A.J. and D.W. Jorgenson, 1980, Inflation-proof depreciation of assets, Harvard Business Review 58, 113-118. pricing: some economic Berry, C.H., D.F. Bradford, and J.R. Hines Jr., 1992, Arm’s_length perspectives, Tax Notes 54, 731-740. Blanchard, O.J., F. Lopez-de-Silanes, and A. Shleifer, 1993, What do firms do with cash windfalls? NBER Working Paper No. 4258. Boskin, M.J. and W.G. Gale, 1987, New results on the effects of tax policy on the international location of investment, in: M. Feldstein, ed., The effects of taxation on capital accumulation (University of Chicago Press, Chicago) 201-219. Bradford, J.J., 1990, Controlled foreign corporations, 1984: a geographic focus, Statistics Of Income Bulletin 9, 115-134. Fazzari, S., R.G. Hubbard, and B. Peterson, 1988, Financing constraints and corporate investment, Brookings Papers on Economic Activity 1, 141-195. Hartman, D.G., 1981, Domestic tax policy and foreign investment: some evidence, NBER Working Paper No. 784. Hartman, D.G., 1984, Tax policy and foreign direct investment in the United States, National Tax Journal 37, 475-488. Hartman, D.G., 1985, Tax policy and foreign direct investment, Journal of Public Economics 26, 107-121. Hines, J.R., Jr., 1988, Taxation and U.S. multinational investment, in: L.H. Summers, ed., Tax policy and the economy, vol. 2 (MIT Press, Cambridge) 33-61. Hines, J.R., Jr., 1991, The flight paths of migratory corporations, Journal of Accounting, Auditing, and Finance 6, 447-479. Hines, J.R., Jr. and R.G. Hubbard, 1990, Coming home to America: Dividend repatriations by U.S. multinationals, in: A. Razin and J. Slemrod, eds., Taxation in the global economy (University of Chicago Press, Chicago) 161-200. Hines, J.R., Jr. and E.M. Rice, 1994, Fiscal paradise: foreign tax havens and American business, Quarterly Journal of Economics 109, 149-182. Jun, J., 1989, What is the marginal source of funds for foreign investment? NBER Working Paper No. 3064. Leechor, C. and J. Mintz, 1990, On the taxation of multinational corporate investment when the deferral method is used by the capital exporting country, mimeograph, University of Toronto. McIntyre, R.S., 1989, Tax Americana, The New Republic, March 27, 18-20. Newlon, T.S., 1987, Tax policy and the multinational firm’s financial policy and investment decisions, unpublished Ph.D. dissertation, Princeton University. Scholes, M.S. and M.A. Wolfson, 1992, Taxes and business strategy: a planning approach (Prentice Hall, Englewood Cliffs, NJ). Sinn, H.-W., 1990, Taxation and the birth of foreign subsidiaries, NBER Working Paper No. 3519. Sinn, H.-W., 1991, The vanishing Harberger triangle, Journal of Public Economics 45,271-300.

J.R.

Hines, Jr.

I Journal of Public Economics

55 (1994) 323-347

347

Slemrod, J., 1990, Tax effects on foreign direct investment in the United States: Evidence from a cross-country comparison, in: A. Razin and J. Slemrod, eds., Taxation in the global economy (University of Chicago Press, Chicago) 79-117. Stewart, M.B., 1986, U.S. tax policy, intrafirm transfers, and the allocative efficiency of transnational corporations, Public Finance 41, 350-371. Stiglitz, J.E., 1973, Taxation, corporate financial policy, and the cost of capital, Journal of Public Economics 2, l-34.