Forum on Mobility and Tax Analysis Theories of Tax Competition

Abstract - A central message of the tax competition literature is that independent governments engage in wasteful competition for scarce capital through reductions in tax rates and public expendi- ture levels. This paper discusses many of the contributions to this literature, ranging from early demonstrations of wasteful tax com- petition to more recent contributions that identify efficiency- enhancing roles for competition among governments. Such roles involve considerations not present in earlier models, including im- perfectly-competitive market structures, government commitment problems, and political economy considerations.

INTRODUCTION

he modern literature on tax competition began with an Tattempt to understand the potential efficiency problems associated with competition for capital by local governments. Oates (1972, p. 143) describes this problem as follows: “The result of tax competition may well be a tendency toward less than efficient levels of output of local services. In an at- tempt to keep taxes low to attract business investment, local officials may hold spending below those levels for which mar- ginal benefits equal marginal costs, particularly for those pro- grams that do not offer direct benefits to local business.” In other words, local officials will supplement the conven- tional measures of marginal costs with those costs arising from the negative impact of taxation on business investment. These additional costs might include lower wages and em- ployment levels, capital losses on homes or other assets, and reduced tax bases. Their presence will reduce public spend- ing and taxes to levels where the marginal benefits equal the higher marginal costs. Oates’s conclusion that this behavior is inefficient rests on the idea that when all governments be- have this way, none gain a competitive advantage, and con- sequently communities are all worse off than they would have been if local officials had simply used the conventional mea- sures of marginal costs in their decision rules. John Douglas Since the mid-1980s, there has been an outpouring of aca- Wilson demic research on tax competition, and this research contin- Department of ues unabated. Interest in this area has been stimulated by Economics, highly publicized instances where U.S. states and localities Michigan State do seem to have engaged in tax competition, including the University, East many cases where they have offered large subsidies to for- Lansing, MI 48824 eign and domestic automobile companies in an attempt to 269 NATIONAL TAX JOURNAL influence plant location decisions. In ad- to be unsuitable for private markets” (p. dition, researchers and policymakers have 248). Through a series of examples, Sinn found that Oates’s (1972) description of tax essentially argues that if private markets competition can be applied more broadly fail to efficiently provide particular goods to a host of important policy concerns, such and services, then introducing competi- as competition for investment through tion among governments that seek to pro- weaker environmental standards or reduc- vide them will generate similar problems. tions in welfare payments by states trying Given the central importance of tax com- to avoid attracting poor households. petition in these ongoing policy debates, More generally, the view that intergov- the time seems ripe to assess what we have ernmental competition is wasteful has learned from the large and growing theo- raised fundamental questions about the retical literature on this phenomenon. This appropriate roles of the central govern- is the objective of the present paper. ment and lower level governments. In The scope of my discussion is broader particular, this view runs counter to the than Oates’s (1972) original description of highly influential “Tiebout Hypothesis.” tax competition. Thus, I discuss models Tiebout (1956) argues that competition for in which the governments may be inter- mobile households is welfare enhancing, preted as local, state, or provincial gov- and subsequent work has applied similar ernments within a country, or as countries ideas to competition for mobile firms, in within the world economy. I often use the direct conflict with much of the tax com- term “region” to encompass these various petition literature. This conflict is reflected interpretations. A common feature of most in the ongoing “devolution debate” over of the models considered here is that the desirability of “devolving” federal re- each government independently (or sponsibilities to lower levels of govern- “noncooperatively”) chooses its tax or ment within the United States.1 subsidy policies to maximize the welfare Moreover, this conflict has surfaced in of residents within the region, and its debates concerning the appropriate de- choice affects the size of the tax bases gree of economic integration between available to other governments. These tax countries. In Europe, integration has bases often consist of mobile capital, but I proceeded to the point where barriers to consider alternatives, such as competition labor and capital mobility have been for cross-border shoppers. While the main officially dismantled (although cultural focus of my discussion is on the nonco- barriers remain). The Tiebout literature operative choice of taxes or subsidies, I supports policies that allow free factor also discuss cases in which governments mobility and enable national governments compete by using nontax instruments to function independently in most policy such as expenditure or regulatory policies. areas, and the Treaty on the European My survey shows that the initial formal Union reflects a presumption in favor of models of tax competition largely confirm this independence. In contrast, the tax Oates’s (1972) conjecture, since they stick competition literature draws attention to most closely to the situation he envi- the potentially adverse effects of this in- sioned. Subsequent extensions of these dependence. Similar concerns about inte- models produce a variety of inefficiencies gration surface in Sinn’s (1997) discussion that differ from the original conclusion of the “selection principle,” which identi- that taxes and expenditures are ineffi- fies a “fundamental selection bias towards ciently low. Toward the end of the paper, [government] activities that have proved I discuss recent work that identifies some

1 See Tannenwald (1998) for an overview of this debate. 270 Forum on Mobility and Tax Analysis beneficial effects of competition among models. As Huang (1992) argues, unem- governments that seem not to have been ployment may provide an additional in- previously recognized, in part because they centive for wasteful tax competition, since represent departures from both the stan- governments now benefit from the em- dard tax competition models and Tiebout ployment generated by additional capital.2 models. My concluding remarks call for My discussion of competition for capital more work on the potentially important is also restricted to “industrial capital,” trade-offs between the good and bad as- rather than “residential capital.” The lat- pects of intergovernmental competition. ter type of competition has been investi- This paper is organized as follows. I be- gated using what Mieszkowski and gin in the next section by discussing the Zodrow (1989) refer to as “metropolitan idealized Tiebout world and how tax com- models.” Such models contain residential petition models typically depart from it. housing and assume interjurisdictionally The third section then surveys a large num- mobile residents, making them less appli- ber of models in which governments use cable to tax competition between countries. tax policies to compete for mobile capital. Although I briefly discuss multinationals, Throughout this discussion, firms are as- interested readers should consult more sumed to be competitive, and only taxes on specialized surveys on this topic, such as mobile capital are considered. The fourth Gresik (1998) and Hines (1997, 1999). Fi- section discusses several attempts to ex- nally, my focus on theories of tax competi- pand the basic model to include multiple tion precludes a detailed examination of tax instruments, including models where the empirical work on its existence; see both capital and labor are mobile. In the Brueckner and Saavedra (1998) for some fifth section, I change the focus to “com- empirical evidence and related references. modity tax competition,” where regions compete for cross-border shoppers, a situ- THE TIEBOUT MODEL ation that is relevant not only for regions, states, or provinces within a single coun- Tiebout’s (1956) theory of local public try, but also for countries with few border good provision also provides a theory of controls, such as the European Union. The efficient tax competition. In modern for- sixth section briefly discusses competition mulations of the theory, it is often as- involving nontax instruments. In the sev- sumed that each region’s government is enth section, I turn to two areas of research controlled by its landowners, who seek to where the concern has been that taxes are maximize the after-tax value of the too high, not too low: competition between region’s land by attracting individuals to different levels of government (“vertical tax reside on this land. To do so, the govern- competition”) and the use of double taxa- ment offers public goods that are financed tion conventions. The eighth section then by local taxes. A critical assumption is that discuses several lines of research that have there are many “utility-taking” regions, in identified some possible efficiency-enhanc- the sense that no single region can alter ing roles for tax competition. The ninth sec- the utilities that must be offered to indi- tion provides conclusions. viduals to induce them to reside there. To keep this survey manageable, I re- Thus, the model is quite similar to mod- strict its focus in several respects. In keep- els of competitive markets for private ing with common practice in the literature, goods, with land rents serving as profits I limit the discussion to full employment and utilities serving as prices. Therefore,

2 Haaparanta (1996) also examines a model with unemployment, using the common agency approach dis- cussed below. 271 NATIONAL TAX JOURNAL it is not surprising that the equilibria for actions that one region’s government such models are found to be efficient un- takes to increase the welfare of its own der the usual definition of efficiency: a residents leads to reductions in the wel- central authority cannot feasibly reallocate fare of residents in other regions. In the goods and resources in a way that makes tax competition literature, this external- some individuals better off without mak- ity is often described as a “fiscal external- ing anyone worse off. There is tax compe- ity,” which occurs through the effects of tition here in the sense that a region’s taxes one region’s public policies on the gov- must be kept low enough to induce indi- ernment budgets in another region viduals to reside in the region, given the (Wildasin, 1989). For example, when a re- public goods that are being provided. gion lowers its tax rate on mobile capital, These taxes are collected from residents it gains capital at the expense of other re- in the form of efficient “head taxes,” and gions, causing their tax bases to fall and, they are chosen so that each resident’s tax hence, their tax revenues to decline. Such payment equals the cost of providing him externalities are often present in tax com- with the chosen levels of public goods and petition models because governments are services. This marginal-cost-pricing rule assumed not to possess unlimited taxing results in efficient migration decisions. powers. In particular, the assumption of There is now a vast literature extending efficient taxes on residents or firms is these efficiency results in various direc- clearly violated in practice.3 tions. As originally suggested by Fischel Another type of externality is the “pe- (1975) and White (1975), the theory may cuniary externality” that exists when be easily extended to include mobile firms. regions are large enough to affect the These firms are modeled in effectively the product or factor prices confronting other same way as mobile residents by assum- regions. I show in the “Large Regions” ing that they are in infinitely elastic sup- section that these externalities lead to in- ply to any given region and that the re- efficient policy differences across regions, gion supplies them with various “public causing a misallocation of factors. They inputs.” In equilibrium, each firm is as- do not, however, allow us to conclude that sessed a tax equal to the marginal cost of taxes and public spending are too low in supplying it with these public inputs, and some overall sense. the equilibrium is efficient. See Richter and Other interregional externalities arise as Wellisch (1996) for a detailed development a result of inefficiencies in private markets, of this extension. Wellisch (forthcoming) coupled with the failure of governments discusses this and several other exten- to correct them. In particular, this study sions. Note, in particular, that the model briefly addresses the issue of tax compe- may be extended to include not only mo- tition under imperfect competition. bile firms, but also mobility of both the Finally, regional governments may labor and capital that they employ. make policy choices that are not in the best Wasteful tax competition involves some interests of their residents. Such behavior type of departure from the idealized set- may give rise to various interregional ex- tings of “Tiebout models.” The main ternalities. However, we cannot say a source of departure is the existence of priori whether the interregional mobility “interregional externalities,” whereby the of factors improves or worsens the behav-

3 Hamilton (1975, 1976) argues that the zoning policies employed by local governments can turn property taxes into efficient “user fees” for local public goods, but the availability and use of efficient zoning policies is open to question. See Mieszkowski and Zodrow (1989) and Ross and Yinger (1998) for critical assessments of Hamilton’s view, and see Hoyt (1991a), Krelove (1993), and Wilson (1997) for analyses of the residential property tax as an “inefficient user fee.” 272 Forum on Mobility and Tax Analysis ior of these governments. The “Political forms into a public good. Income distri- Economy” section discusses a model in bution issues are ignored by assuming which government behavior is made more either that each region’s residents are efficient by the presence of free factor mo- identical or that their aggregate welfare bility. can be depicted by the preferences of a “representative consumer.” In particular, CAPITAL TAX COMPETITION these preferences are represented by a well-behaved utility function, U(C, G), Following Oates’s (1972) discussion of where C is private consumption and G is tax competition, it was not until the mid- consumption of the public good. The rep- 1980s that economists began to build for- resentative consumer finances C with the mal models based on his ideas. Two of the wage and capital income from her endow- earliest papers were Zodrow and ments of labor and capital. Summing the Mieszkowski (1986) and Wilson (1986).4 capital endowments across the residents Since the production structure in Zodrow in all regions gives the fixed supply of and Mieszkowski’s model is simpler than capital in the “world economy.” Wilson’s structure, I let their model serve A critical assumption in this model is as the “basic tax competition model” in that each region’s public good supply is this survey. The model is next presented financed by a tax on the capital employed in its simplest form, and then it is ex- within its borders. One simple justifica- tended in various directions. tion for this assumption is that the gov- ernment finds it administratively conve- Competitive Regions nient to tax both capital and land at the same rate, as in the case of local property Consider a system of many regions, taxation in the United States. I later dis- which may be variously interpreted as cit- cuss a three-factor model that explicitly ies, states, provinces, or countries. Within includes a uniform property tax on mo- each region, competitive firms produce a bile capital and immobile land, and I also single output, using two factors of produc- discuss models where governments tion: mobile “capital” and an immobile choose among multiple tax instruments. factor that I will call “labor.”5 The immo- But for the “basic model,” I follow bile factor is inelastically supplied by the Zodrow and Mieszkowski (1986) by as- region’s residents, who are themselves im- suming that only capital is taxed, although mobile (an assumption that I drop in the each region’s residents would be better off “Labor Mobility” section). These residents paying a head tax. also own fixed endowments of capital, The problem confronting a region’s and the assumption of perfect capital government is to choose the unit tax rate mobility means that they are free to in- on capital, t, to maximize the representa- vest their capital anywhere.6 Once invest- tive consumer’s utility, U(C, G), subject to ment and production take place, the a budget constraint requiring that tax rev- output is sold to residents as a final con- enue equal public good expenditures: sumption good, and to the government as an intermediate good, which it then trans- [1] tK(r + t) = G

4 See Beck (1983) for another early, but less general, treatment of tax competition. 5 Zodrow and Mieszkowsk (1986) interpret the immobile factor as land and assume no absentee ownership, an assumption that I address below. They also allow regions to possess market power, an extension I leave to the “Large Regions” section. 6 The tax competition literature has devoted little attention to the intermediate case of “imperfect” capital mobility. For an exception, see Lee (1997). 273 NATIONAL TAX JOURNAL where r is the after-tax return on capital, demand for capital with respect to the cost which is equated across regions by capital of capital to rewrite equation 2 as follows:8 mobility but treated as fixed by any small K r t MB MC region; and ( + ) is a function relating [3] = τε the demand for capital in the region to the 1 – cost of capital, r + t.7 As more capital is where ε denotes this demand elasticity invested in the region, its marginal prod- (measured positively) and τ is the ad valo- uct falls and the marginal product of labor rem tax rate, τ = t/(r + t). rises. Firms invest to the point where the To conclude, both of these rules dem- marginal product of capital equals r + t. onstrate that the marginal benefit of Consider now the government’s G exceeds the marginal resource cost to optimality condition for G. To finance a compensate for the tax-induced capital unit rise in G, the government must raise outflow. In other words, we have a “modi- t. As a result, the cost of capital rises, caus- fied” Samuelson rule for public good pro- ing the demand for capital to change by vision, since the actual Samuelson rule for some negative amount, ∆K. Since r is fixed efficient public good provision would re- from the region’s viewpoint, however, the quire that MB = MC. higher tax rate does not reduce the resi- The tax rate t in these rules has an im- dents’ capital incomes. Instead, they in- portant interpretation. It represents the dis- directly pay the tax through a decline in crepancy between the social value of an their wages. In particular, if a unit rise in additional unit of capital and the social G requires that t be raised by ∆t to bal- opportunity cost of this unit, measured ance the budget, then wage income will from the single region’s viewpoint. This fall by the resulting rise in the cost of capi- social value is the marginal product of capi- ∆ tal, K t, to prevent firms from earning tal, MPK, which firms equate to negative profits. Unfortunately, this tax r + t when they choose their profit-maxi- increase must be high enough to not only mizing investment levels. In contrast, the pay for the “marginal resource cost” of G, social opportunity cost is only r, since the denoted MC, but also to offset the nega- tax provides revenue for the government tive impact of the capital outflow on tax and is therefore not a social cost. Thus, ∆ revenue, t K. Thus, the resulting fall in t = MPK – r. It is this discrepancy between the residents’ wage income will exceed the value and opportunity cost of capital MC by the positive amount, –t∆K. At the at the margin that implies that the region region’s optimal level of G, the sum of the benefits from a capital inflow and is residents’ marginal willingness to pay for harmed by a capital outflow. In a more gen- another unit of G, or “marginal benefit” eral model, we would recognize that capi- of G, is equated to this wage reduction: tal investments impose various burdens on the public sector, such as the increased de- [2] MB = MC – t∆K mands for public infrastructure and vari- ous public services. If capital were taxed Alternatively, it is possible to use the de- to cover the marginal costs of these public pendence of ∆K on the elasticity of the goods and services, then the region would

7 For simplicity, the literature often specifies capital taxes as tax rates on each unit of capital. The distinction between unit and ad valorem taxes is irrelevant for the current model. When regions with market power are later considered, the equilibrium can depend on the form of taxation, but its qualitative properties do not change. 8 Using calculus notation, ∆K = (dK/dt)(dt/dG), where dt/dG is the rise in t needed to finance a unit increase in G. As noted in the text, wages fall by K(dt/dG) to offset the higher capital costs, and this wage reduction is set equal to MB at the optimum. Thus, we also have ∆K = (dK/dt)(MB/K). Substituting this relation into equation 2 and rearranging gives equation 3. 274 Forum on Mobility and Tax Analysis be indifferent about a marginal inflow or cient public good provision, MB = MC. outflow of capital, and so the terms involv- Starting from the equilibrium level of pub- ing t in equations 2 and 3 would disappear. lic good provision, where MB > MC, we But if capital is not taxed efficiently and we can satisfy the efficiency condition by in- view the term t more generally as the dif- creasing all regions’ tax rates and public ference between the social value and so- good levels by identical amounts. These cial opportunity cost of capital at the mar- changes raise welfare in all regions. gin, then equations 2 and 3 remain valid. Another way to satisfy the efficiency The subsequent analysis therefore can be condition would be to close each region’s given this more general interpretation.9 borders to factor mobility, so that the Having described a single region’s op- region’s firms could use only the capital timal public good supply, let us now con- supplied by its residents. In this case, the sider whether it is efficient for the system interregional externalities described above of regions as a whole. A critical insight would disappear, and the capital tax from the tax competition literature is that would become equivalent to an efficient a rise in one region’s tax rate benefits other lump-sum tax. Thus, each government regions by increasing their capital supplies. would independently choose to set its Under the assumed fixity of the total capi- public good supply where MB = MC. With tal stock, the tax-induced outflow of capi- the borders open, however, no single gov- tal from the given region represents an in- ernment has an incentive to raise G to the flow of capital to other regions, and the point where MB = MC, given the costs as- value of this inflow will depend on the tax sociated with the resulting capital outflow. rates being set in these regions. In particu- There must be some type of coordination lar, another region j benefits by the amount among regions, suggesting a role for a cen- ∆ ∆ tj Kj, where tj is its tax rate and Kj is the tral authority. This role is discussed below capital inflow that it experiences.10 In other in the context of nonidentical regions. words, a rise in the given region’s tax rate Finally, two considerations may miti- creates a positive externality. The region’s gate the tendency of regions to under- government fails to account for such ex- provide public goods. First, suppose ternal benefits because it is concerned only that the model is modified to include a with the welfare of its own residents. Con- variable supply of capital for the system sequently, it sets its tax rates and public of regions as a whole, due to the savings good levels at inefficiently low levels. behavior of residents. If a subset of regions The tax competition literature has fo- increases its tax rates, then total savings cused heavily on the case where all regions may decline, dampening the amount of are identical and therefore choose the same capital that is redirected to other regions. tax rates. This case nicely isolates ineffi- The interregional externalities remain, ciencies in the overall level of public good but their importance is somewhat reduced. provision from the efficiency and equity We may conclude that allowing a variable issues concerning differences in tax rates supply of capital reduces, but does not and public good levels across regions. In eliminate, the tax competition problem. this case, the cost of a capital outflow from Second, extending the basic model one region is exactly offset by the benefits to include absentee ownership of the from the accompanying capital inflows to immobile factor, reinterpreted as “land,” other regions. Hence, ∆K disappears from introduces “tax exporting” into the analy- rule 2 and we are left with the rule for effi- sis. The capital tax is now capitalized into

9 This discussion assumes an exogenous level of public infrastructure. See footnote 27 on this issue. 10 I use subscripts to identify regions only in cases where they are needed to avoid confusion. 275 NATIONAL TAX JOURNAL

the return on land, passing part of its [4] Si = ai + si(tK) burden on to nonresidents. This form of tax exporting counteracts the effects of where the ai’s are positive or negative, re- tax competition, thereby raising the sup- flecting the interregional transfers, and the 11 ply of public goods. As an empirical mat- si’s denote the corrective subsidies. These ter, it seems unlikely that households subsidies have been suggested by Wildasin come close to owning the diversified port- (1989) and analyzed further by DePater folio of land needed to fully eliminate the and Myers (1994). For a system of N underprovision problem in the basic tax regions, the ai’s and si’s represent 2N vari- competition model. ables that are chosen to satisfy the 2N con- ditions for an efficient allocation: the N – 1 Corrective Policies equations requiring that all regions choose the same tax rates, the N equations requir- Regions may choose different tax rates ing that each region’s public good level be because they possess different production set where MB = MC, and the requirement technologies or their residents possess dif- that the central authority’s budget balance. ferent preferences or factor endowments. In equilibrium, the corrective subsidies If the tax rates do differ, then two types of will generally differ across regions to in- inefficiencies exist in the economy. First, duce them to choose the same tax rates. public good levels are set inefficiently, be- But the critical point here is that all correc- cause regions fail to account for the inter- tive subsidies should be positive in regional externalities discussed above. equilbrium. Only then will regions be com- Second, capital is misallocated across re- pensated for the positive externalities as- gions, so that the marginal product of capi- sociated with increases in their tax rates.12 tal is relatively high in high-tax regions. A In practice, however, it is doubtful that fully efficient allocation cannot be the policy intervention described by equa- achieved if tax rates differ across regions, tion 4 would be feasible, since it would and identical tax rates are usually not con- require that the grant functions be tailored sistent with efficient differences in public to the individual characteristics of regions, good levels across regions, unless a cen- which might be difficult for a central au- tral authority also redistributes revenue thority to observe. A recent paper by across the government treasuries. Bucovetsky, Marchand, and Pestieau The central authority can achieve this (1998) designs an alternative form of efficient allocation by providing each re- policy intervention that takes into account gion with a “corrective subsidy” on the the central authority’s information prob- revenue it raises, while at the same time lem. Their model contains two types of engaging in lump-sum transfers of income regions, distinguished only by the resi- between the regional governments. In dents’ demands for public goods.13 The symbols, a region-i government faces a central authority has two policy instru- grant schedule of the form ments: a “national capital tax,” T, and a

11 See Lee (1998) for a fuller treatment of the case of absentee landowners. Burbidge and Myers (1994) provide limited efficiency results for a three-factor model where capital and labor are mobile between two nonidentical regions and each individual owns equal amounts of capital, labor, and land in each region. But Bucovetsky (1995) emphasizes the inefficiencies that emerge from this setup when factor endowments are allowed to differ. 12 By focusing solely on efficiency issues, this analysis ignores issues involving the interregional distribution of income. Given the limited policy instruments available to a central authority, it might want to deviate from an efficient allocation to reduce inequities in the income distribution. 13 Following the basic tax competition model, Bucovetsky et al. (1998) assume many regions of each type. For the difficult problem of regulating large regions that possess informational advantages, see Dhillon, Perroni, and Scharf (forthcoming). 276 Forum on Mobility and Tax Analysis single grant schedule, which determines tries can cause the central authority to the total grant provided to a region as a “overcorrect” the tax competition problem function of the region’s chosen tax rate. In by inducing some regions to go from symbols, the grant schedule may be de- undersupplying public goods to oversup- scribed by a function, Si = f(ti). The critical plying them. difference between this type of policy and To advance our understanding of “op- the one given by equation 4 is that both timal fiscal federalism,” future research types of regions must face the same func- will increasingly need to take information tion f, in contrast to the different linear problems into account. However, the functions in equation 4. Given this restric- information-based approach has two tion, we should not expect the optimal limitations. First, we do not have a good grant function to induce the two types of understanding of how information asym- regions to choose the same tax rates, since metries occur between different levels of they would then receive the same grants, government, and what exact form these despite their preference differences. asymmetries take. Rather, we have vague The problem confronting the central ideas, such as the understanding that lo- authority is to choose T and the grant cal officials know more because they are schedule to maximize the total welfare of “closer to the people.” Formal models that residents, measured by summing the wel- try to capture such ideas must usually fare levels of residents across all regions. assume such a simple economic environ- Using solution techniques from principal- ment that it seems difficult to justify why agent theory, Bucovetsky, Marchand, and the central authority cannot easily obtain Pestieau (1998) demonstrate that the op- the information that is assumed absent. timal grant function induces the high- In Bucovetsky, Marchand, and Pestieau demand regions to choose a higher tax (1998), for example, it would seem to be a rate than low-demand regions. Thus, the simple matter to check whether the de- capital market is necessarily distorted at mand for the single public good is higher the optimum as a means of inducing the in one region than in another. Future re- different types of regions to select differ- search should try to “endogenize” this ent grant levels. Moreover, the regions re- lack of information, rather than simply spond to the grant function by choosing assume it. inefficient public good levels. For the spe- Second, this information approach to cial case where public good demands are fiscal federalism is largely a normative independent of income levels (quasi-lin- exercise, in the sense that government of- ear preferences), Bucovetsky, Marchand, ficials at all levels of government seek only and Pestieau show that the high-demand to raise the welfare of those individuals regions underprovide the public good they represent. This ignores self-interested (i.e., MB > MC), but the low-demand re- behavior and the constraints imposed by gions are induced to overprovide the pub- existing political institutions, both of lic good (i.e., MB < MC). This makes sense. which would be useful to model. In order to lessen the capital market dis- At the international level, there do not tortions caused by raising the high- exist strong institutions for coordinating demand tax rate above the low-demand the activities of sovereign nations. The tax rate, the central authority accepts poli- need to induce cooperation among such cies where the interregional difference be- nations severely restricts the set of feasible tween public good levels is less than it policies. Fortunately, the tax competition would be under an efficient allocation literature often identifies forms of inter- (where MB = MC in all regions). It is strik- vention that leave all regions better off, ing, however, that information asymme- suggesting some scope for cooperation. 277 NATIONAL TAX JOURNAL

Large Regions Thus, market power is actually beneficial from the viewpoint of welfare in the Suppose now that regions are large entire system of regions, since it lowers enough to influence the equilibrium after- the perceived marginal cost of public good tax return on mobile capital. For this large- provision, thereby stimulating public region case, a Nash equilibrium is the good provision. Hoyt (1991b) carries this accepted equilibrium concept. Specifi- idea further by showing that public good cally, the economy is in equilibrium when levels and tax rates rise as the number of each region’s strategy maximizes its ob- competing regions drops. Of course, there jective function, given the strategies pur- is no longer any tax competition problem sued by the other regions. The literature once the economy contains only one re- typically treats tax rates as the strategy gion. In this case, there are no capital out- variables, in which case public good lev- flows, and a unit rise in t is fully capital- els adjust to satisfy each region’s govern- i ized into r, so that there is no change in ment budget constraint once all of the tax the cost of capital, r + t . rates have been chosen. A single region i New considerations arise when regions chooses its tax rate, treating as fixed the differ in size. Bucovetsky (1991) and Wil- tax rates chosen by other regions.14 It then son (1991) analyze “asymmetric tax com- recognizes that the equilibrium after-tax petition” between a “large” region and a return on capital depends on its chosen “small” region, as distinguished by the tax rate and that of every other region: number of residents, each possessing the r = r(t , . . . t ), where t is the tax rate cho- 1 N i same endowments of capital and labor. sen by region i in a system of N regions. Since the large region is the relatively large In particular, if region i raises its tax rate a demander in the capital market, an in- unit, then its cost of capital, r + t , will rise i crease in its tax rate depresses the after- by 1 + ∆r, which is less than one, because tax return on capital, r, by a relatively large the resulting change in r, ∆r, must be nega- amount. Thus, the cost of capital, r + t, is tive to clear the capital market. As a re- less sensitive to tax changes in the large sult, K(r + t ) is less sensitive to changes in i region than in the small region. This con- t than in the case of many small regions, i sideration suggests that the large region where r is fixed from a single region’s will compete less vigorously for capital viewpoint. An increase in one region’s tax through tax rate reductions and therefore rate continues to create a positive exter- end up with the higher tax rate. nality through a capital outflow, but now Bucovetsky and Wilson demonstrate that this outflow is less severe, due to the par- this is indeed the case. tial capitalization of higher tax rates into the after-tax return on capital. This finding leads to interesting conclu- Consider first the simpler case of iden- sions about the advantage of “smallness” tical regions. Rule 2 remains valid, recog- in a tax competition model. Because the nizing that the capital outflow, ∆K, is small region possesses the lower cost of lower than before, and rule 3 is modified capital in equilibrium, firms there employ to account for the capitalization effect: more capital per unit of labor and there- fore offer higher wage rates than in the MC [5] MB = . large region. As a result, the residents of 1 – τε(1 + ∆r) the small region can be shown to be bet-

14 Wildasin (1988, 1991) analyzes the alternative formulation in which the public good levels serve as the strat- egy variables, with the tax rates adjusting to satisfy the government budget constraints. For a system of identical regions, this change in strategy variables reduces equilibrium public good levels further below the efficient level. 278 Forum on Mobility and Tax Analysis ter off than the residents of the large re- overprovide the public good relative to gion. In fact, Wilson (1991) demonstrates the rule for efficient provision. Differences that if the difference between regional in regional size, production technologies, sizes is sufficiently large, then the small or consumer preferences can be expected region will be better off than it would be to cause one region to export capital to the in the absence of tax competition, where other, and the capital-importing region the capital tax is replaced with head taxes has an incentive to restrict such imports, on residents. Thus, we find that although thereby driving down the required after- tax competition is inefficient, it can actu- tax return on capital. The tax-induced ally benefit some regions. change in r, again denoted ∆r < 0, now This comparison between small and enters the rule for equilibrium public good large regions shows that tax competition provision as follows: is quite different from the analysis of “tar- MC{[1 – (k/k)]∆r + 1} iff wars.” In the latter case, it is generally [6] MB = . τε ∆ believed that sufficiently large countries 1 – (1 + r) win tariff wars in the sense that they are where k is the capital owned by the better off than they would be under free region’s residents and k is the capital used trade. The basic idea is that they have by the region’s firms, making k – k im- more ability to manipulate the terms of ports of capital. Comparing this rule with trade through their use of tariffs and will the rule given by equation 5 for the sym- therefore employ higher tariffs than small metric case, we see that the new term, countries. Kennan and Riezman (1988) [1 – (k/k)]∆r, acts to reduce the marginal present a formal analysis of a tariff war cost of the public good for capital-import- between two countries, modeled as a ing regions. In other words, such regions Nash equilibrium in tariff rates, and they have an extra incentive to increase the tax find that the larger country does “win” if rate in order to achieve desirable “terms- the size difference is sufficiently great. The of-trade effects.” Of course, such effects reason for this difference in results can be come at the expense of capital-exporting traced back to interregional externalities. regions, which are directly harmed by the In the case of tax competition, we have drop in r. DePater and Myers (1994) refer seen that one region’s tax creates a posi- to these terms-of-trade effects as pecuni- tive externality through the flow of capi- ary externalities, in contrast to the fiscal tal to other regions. Since the smaller re- externality associated with the capital gion has the lower tax rate, it is therefore elasticity in equation 6. It is the existence the net beneficiary of these interregional of a pecuniary externality that may lead externalities. In the case of tariff wars, to overprovision of the public good in however, a country’s tariff creates a nega- capital-importing regions. On the other tive externality by changing the terms of hand, its presence works in the opposite trade in an unfavorable way from the direction in capital-exporting regions, other country’s viewpoint. This difference thereby aggravating the underprovision in the sign of the externalities is respon- problem. sible for the different welfare results.15 This overprovision possibility has im- Another interesting implication of plications for the optimal form of central asymmetric tax competition between two authority intervention, an issue that regions is that one region may choose to DePater and Myers (1994) also address.

15 In a paper that combines elements of trade and tax competition, Haufler and Wooton (1999) consider competi- tion between two countries for a foreign-owned monopolist and conclude the large country wins the competi- tion, because the monopolist benefits from a larger market, due to the assumed existence of transport costs. 279 NATIONAL TAX JOURNAL

Once again, the first-best allocation can be differences in comparative advantage, dif- achieved with a subsidy function of the ferent regions choose different tax rates form given by equation 4. But now the and trade goods with each other. In fact, corrective subsidy may be negative in the low-tax regions produce only the capi- capital-importing regions, to offset the tal-intensive good, and the high-tax re- pecuniary externality. Thus, we can no gions produce only the labor-intensive longer say that all regions engage in good, with the share of regions produc- wasteful tax competition, in the sense that ing each good determined by the require- they set their tax rates inefficiently low to ment that demand equal supply in the attract capital. On the other hand, the mis- goods markets. To see this, observe that if allocation of capital resulting from the a region produced both goods, then both differing incentives faced by capital-im- industries would earn zero economic prof- porting and capital-exporting regions is a its, as required for a competitive equilib- new and potentially important type of rium. But then a tiny reduction in the tax inefficiency. rate would lower the cost of capital and Finally, the issue of large versus small raise the wage rate so that only the capi- regions raises interesting questions about tal-intensive industry could break even, the incentives that groups of regions thereby driving labor-intensive firms out might possess to form into coalitions (or of the region and leaving the region spe- “federations”) for the purpose of compet- cialized in the capital-intensive good. The ing with other regions for scarce capital. result would be a large jump in the Burbidge et al. (1997) analyze this issue region’s capital stock, which would raise by combining the basic tax competition tax revenue and thereby provide residents model with a model of coalition forma- with more public good provision. Thus, tion. If the number of regions exceeds two, it can never be optimal for a region to pro- then the equilibrium coalition structure duce both goods. can involve more than one independent Stated differently, some regions choose coalition. Thus, the tax competition prob- to compete vigorously for capital, thereby lem does not necessarily disappear when ending up with capital-intensive firms there exists endogenous coalition forma- and high wages, but low public good lev- tion. els, whereas others forego vigorous com- petition and settle for labor-intensive Trade firms and low wages, but high public good levels. Any single region is indiffer- We have seen how different regions ent between the two tax policies, since all may choose different tax rates, creating a regions are innately identical. Since the misallocation of capital. A natural ques- model assumes that all individuals are tion is whether the pattern of interregional identical, it is inefficient for the residents trade in private goods is also distorted by of different regions to consume different taxation. The basic tax competition model bundles of private and public goods. does not address this issue, because pri- Thus, the analysis identifies another po- vate consumption is aggregated into a tential type of inefficiency in tax competi- single good. Wilson (1987) considers in- tion models. stead a system of many regions with two private goods, one labor intensive and the MULTIPLE TAX INSTRUMENTS other capital intensive. The implications of this change are surprising: even if there There would not exist a tax competition exist no innate differences between re- problem if the regional governments gions, such as the usual trade-creating could utilize head taxes or other forms of 280 Forum on Mobility and Tax Analysis lump-sum taxation, i.e., taxes that do not provision is found to hold in a case where distort private sector behavior because governments choose to tax capital. Taken they are collected in fixed amounts, inde- together, these results demonstrate that the pendent of consumption and factor sup- underprovision of public goods is not tied ply decisions. For example, a tax on labor to the taxation of mobile capital when other income would be a lump-sum tax in the tax instruments are available. In a final basic tax competition model, since the subsection, I discuss the “common-agency supply of labor is fixed. As such, it would approach,” which explicitly models the be definitely preferred to the capital tax. information problems that governments The modeling practice of much of the tax face when they design their tax policies. competition literature is to exclude such taxes in order to analyze taxes that gov- Optimal Commodity Taxation ernments typically levy on mobile factors. The original tax competition models fo- The analysis of alternative tax instru- cused on capital taxes, because such taxes ments must confront a fundamental result are a component of property taxes in the on optimal taxation in an open economy: United States, given the common practice if a government can satisfy its revenue of not distinguishing between the values requirement using a system of optimal of capital and land at a given site. commodity taxes, then it should not use One reason given for not relying heavily tax instruments that distort the pattern of on lump-sum taxes is that administra- goods trade or factor trade with other re- tively feasible forms of lump-sum taxation gions. In particular, it should not use the would not be equitable or politically fea- type of “source-based” capital tax em- sible. For example, Margaret Thatcher’s ployed in the basic tax competition model, implementation of a poll tax in Great Brit- which is levied on only the capital income ain is widely viewed as having helped earned within a region’s borders. Instead, drive her from office. If it is not possible it should use a “residence-based” capital or desirable to generate tax revenue tax, which is levied on each resident’s through the use of lump-sum taxes, we worldwide capital income. This result are still left with the question of whether applies to regions that are small in the the inefficiencies from tax competition in sense that they cannot manipulate the the basic model remain if capital taxation terms of trade, including the required af- is supplemented with other forms of taxa- ter-tax return on capital, and it also as- tion that distort consumer or producer sumes the absence of untaxed profits. decisions. This section discusses a few at- Gordon (1986) provides a proof within the tempts to address this issue. context of a two-period model of a single The next two subsections discuss some region, in which residents choose how alternative assumptions about available much labor to supply to competitive firms taxes, using extensions of the basic tax in the first period and how much to save competition model. I then amend the for consumption in the second period. The model in a fundamental way by allowing residence-based tax on capital is basically both labor and capital to be mobile. In both a tax on the residents’ income from sav- cases, I identify assumptions under which ings, which is a tax on future consump- public goods are underprovided, but some tion. By also taxing labor income, the of these assumptions imply that only la- government implements an optimal com- bor is taxed, not capital. In contrast, the modity tax system, leaving no room for a Samuelson rule for efficient public good beneficial source-based tax.16 Although

16 Current consumption serves the role of an untaxed numeraire commodity in this model. 281 NATIONAL TAX JOURNAL

Gordon assumes away income distribu- employ the two-period setup used by tion problems, the desirability of com- Gordon (1986) to examine a single region’s modity taxation extends to regions with tax policy, and they allow the number of heterogeneous populations. regions to be either small or large. If these This negative conclusion about the use regions have access to both source- and of source-based capital taxation essen- residence-based capital taxes, then the tially extends a theorem by Diamond and equilibrium is efficient, given the available Mirrlees (1971) to an open economy con- tax instruments. Thus, the tax competition text. They show that the government problem disappears when a residence- should keep the economy on the frontier based capital tax is available.17 of its aggregate production possibility set In practice, it is quite difficult to tax if it can employ an optimal commodity tax capital income on a residence basis, inde- system (where factors such as labor are pendently of where it is owned. The ad- among the commodities). For a small re- ministrative and tax compliance problems gion, “international trade” in goods or involved in taxing foreign-source income factors may be viewed as another produc- are much more severe than those associ- tion sector, where exports are used to ated with taxing domestic income.18 As a “produce” imports via a linear technology. result, researchers have investigated mod- Viewed this way, aggregate production els where residence-based taxation is ei- efficiency requires that government poli- ther limited or not available.19 cies not distort trade (Dixit and Norman, 1980). In particular, source-based capital taxes should not be employed. In the ab- Taxes on Labor and Capital sence of income distribution problems, this particular conclusion is quite intui- The absence of a residence-based capi- tive. A small region faces an infinitely elas- tal tax does not justify taxing capital at tic supply of capital at the after-tax return source. In fact, Bucovetsky and Wilson required by investors in the world (1991) also demonstrate that a small re- economy. In contrast, the region’s resi- gion should meet all of its revenue needs dents have savings and labor supply by taxing only labor income, although la- curves with finite elasticities. Thus, tax- bor-leisure decisions are distorted. Again, ing the income from savings and labor the intuition is that capital investment is minimizes the “deadweight loss,” or “ex- in infinitely elastic supply for a small open cess burden,” from taxation. economy, whereas the labor supply elas- The availability of a residence-based ticity is finite. If, instead, regions are large capital tax has important implications for enough to have influence over the equi- tax competition. Bucovetsky and Wilson librium after-tax return on capital, then a (1991) examine this issue within the con- region’s optimal tax system again includes text of a system of identical regions. They a source-based tax on capital income.

17 In contrast, Razin and Sadka (1991) and Frenkel, Razin, and Sadka (1991) suggest that tax competition is efficient even in the absence of residence-based taxation. In particular, Frenkel, Bazin, and Sadka conclude, “tax competition leads to a constrained optimum, relative to the set of tax instruments that is available” in cases where the countries “cannot effectively tax their residents on their income from capital invested in the rest of the world” (p. 206). The apparent conflict is resolved by noting that they consider only two small countries that face a fixed world interest rate determined in the rest of the world. There is no such rest of the world in Bucovetsky and Wilson (1991). 18 See Blumenthal and Slemrod (1995) for some estimate of these costs. 19 Gordon and MacKie-Mason (1995) and Gordon and Bovenberg (1996) provide additional explanations for why a government might desire to use source-based taxes or subsidies.

282 Forum on Mobility and Tax Analysis

For both cases, Bucovetsky and Wilson Huber (1999) investigates tax competi- (1991) demonstrate that regions under- tion in a model that essentially merges the provide the public good in equilibrium. But basic tax competition model with a finite- now the argument centers on a mechanism type version of the Mirrlees (1971) model for interregional externalities that differs of optimal income taxation. Each of many from the one in the basic tax competition identical regions contains two types of model. Suppose that a single region raises residents, distinguished by the type of la- its labor tax to finance additional public bor they provide, “high skilled” and “low good provision. As a result, the supply of skilled.” Each government uses a nonlin- labor declines, and firms respond by reduc- ear tax on wage income and a source- ing their demands for capital. More capi- based capital tax to finance public good tal is then available to other regions, rais- provision and redistribute income. As in ing the marginal products of labor in these the Mirrlees model, a government seeks regions. The resulting increase in the wage to maximize an objective function that rates in other regions encourages workers depends on the welfare levels of all resi- there to supply more labor, thereby par- dents. But unlike the Mirrlees model, the tially offsetting the distortionary effect of two types of labor enter production func- the labor tax on labor supplies. Thus, a rise tions as separate complementary factors. in one region’s labor tax creates positive As a result, they should be viewed as sepa- interregional externalities, but now these rate commodities, and thus an optimal externalities occur through interactions income tax system does not constitute an between the undistorted capital market optimal commodity tax system, because and the distorted labor markets. Their pres- the latter would require that the two types ence implies that tax rates and public good of labor income face different tax sched- levels are set too low. In particular, every- ules. The absence of optimal commodity one could be made better off if a central taxation creates a role for a source-based authority forced governments to tax wage capital tax. In particular, it should now be income at a higher rate to finance greater used to distort investment decisions in a public good provision. way that reduces the spread between the The availability of multiple tax instru- skilled and unskilled before-tax wage ments introduces a second source of inef- rates. Doing so partially compensates for ficiencies involving government behavior: the government’s inability to apply dif- the choice among tax instruments. For ex- ferent rate schedules to these two types ample, suppose that the variable supply of labor income. But the equilibrium capi- of savings in the Bucovetsky–Wilson tal tax may be positive or negative, de- model is replaced by the assumption of pending on the way in which a rise in the fixed capital endowments, as in the basic capital stock affects the relative marginal model. It remains true that small regions products of the two types of labor. will continue to tax only labor income, but In either case, Huber’s analysis suggests the efficient tax policy for the system of that the equilibrium capital tax will be regions now consists of taxing only capi- inefficiently low under reasonable condi- tal, given its fixed supply. Thus, the de- tions. To identify the interregional exter- centralized choice of tax instruments gets nality, suppose that the capital tax is in- things completely wrong from the view- creased above its equilibrium level in one point of efficiency. We see once again that region. To clear the capital market, we can extending the basic tax competition model expect this higher tax rate to be partially in various directions can produce new capitalized into the after-tax return on types of inefficiencies. capital, i.e., the return falls by some small

283 NATIONAL TAX JOURNAL amount.20 In the empirically reasonable changing the results, provided we con- case where high-skilled workers own tinue to assume that only capital is taxed. more capital than low-skilled workers, the To set the stage for the analysis of mul- fall in the after-tax return will create a tiple tax instruments, I begin by describ- more equal distribution of income in all ing two such models. regions. Thus, the capital tax continues to Brueckner (1999) retains the two-factor create a positive interregional externality, setup of the basic model but allows each but now this externality consists of ben- individual not only to choose where to eficial equity effects. It follows that the invest capital, but also to pick the com- equilibrium tax rate on capital is ineffi- munity in which to work and consume. ciently low, as in the basic tax competi- He assumes a large number of competi- tion model. The analysis nicely demon- tive “developers,” who choose public strates that whether the equilibrium tax good levels and tax rates on mobile capi- on capital is positive or negative does not tal to maximize the “profits” from com- tell us whether it is too low from the view- munity development (which equal zero in point of the entire system of regions. equilibrium). The public good has the To conclude, it appears that competi- properties of a private good, meaning that tion for capital can lead to inefficiently low the per capita cost of providing a given levels of taxation even if the tax instru- amount to each resident is independent ments available to the government extend of the number of residents, i.e., no scale significantly beyond a source-based capi- economies in public good consumption. tal tax. An important task for future re- The mobile individuals differ only in their search is to explicitly model the reasons preferences for the public good. In equi- for why particular sets of tax instruments librium, communities offer different tax are utilized by governments. The informa- rates and public good levels, and indi- tion and political economy approaches viduals sort themselves across communi- discussed below represent some initial ties according to their preferences.21 But efforts along this line. the capital tax continues to create a posi- tive externality, resulting in inefficiently Labor Mobility low tax rates and public good levels. In fact, the rules for equilibrium public good Since individuals are normally free to provision given by equations 2 and 3 re- choose where to reside within their coun- main unaltered. As found in the “Correc- try of citizenship, the models developed tive Policies” section, however, not every- above are perhaps more suitable for tax one must lose from tax competition. In the competition between countries (except for present case, individuals with relatively those groups of countries with free migra- “low” preferences for the public good may tion, such as the European Union). How- be better off under tax competition than ever, labor mobility can be added to the they would be in a fully efficient equilib- basic tax competition model without rium with head taxes used to finance pub-

20 Huber does not show that the after-tax return must fall in all cases. Assuming that it does fall, the assumption that the region is small implies that its residents’ capital income falls by a negligible amount. But when we sum the changes in capital income across the residents in the many small regions, we obtain a non-negligible change. 21 If the public good were pure (i.e., no congestion), then its per capita costs would fall with the number of residents, creating incentives for regions to grow in size. Employing this assumption, Perroni and Scharf (1997) analyze a model in which each region contains many types of residents and majority rule serves as the mechanism by which taxes and public good levels are chosen. They argue that increased capital mobility may make everyone better off in equilibrium, but that tax competition reduces efficiency “in less extreme sce- narios.” 284 Forum on Mobility and Tax Analysis lic good provision. Intuitively, these indi- Wilson’s (1995) surprising conclusion is viduals benefit from living where capital that this rule for efficient provision con- taxes are relatively low and, therefore, tinues to hold when there exist scale capital-labor ratios are relatively high. economies, although the head tax no The model developed by Wilson (1995) longer satisfies the marginal-cost-pricing yields similar conclusions in the absence rule. The presence of scale economies in- of scale economies. There exists a large but troduces a need for other taxes, since the fixed number of identical regions, each per capita cost of public good provision possessing a fixed amount of a third fac- then exceeds the marginal cost. Govern- tor, land. Each region competes for both ments respond by employing the property mobile capital and identical mobile work- tax and manipulating the head tax to par- ers, with the objective of maximizing the tially compensate for the distortionary value of the region’s land. Capital is now effects of the property tax. But they do not taxed through the use of a “property tax,” deviate from efficient public good provi- levied uniformly on the values of both sion. We see, then, that the taxation of capital and land. Assuming this is the only mobile capital need not imply that public tax, the rules given by equations 2 and 3 goods are underprovided in equilibrium. remain valid, but with the capital outflow This last conclusion becomes even more replaced by a reduction in the per capita pronounced if the head tax is replaced value of the region’s capital and land, since with a labor tax that distorts labor-leisure the latter now serves as the per capita tax choices. If scale economies are absent, then base. As before, a rise in the property tax the results are similar to those in the shifts some of this base to other regions, Bucovetsky–Wilson (1991) model of a sys- creating the usual positive externalities. tem of many regions: only labor is taxed If we were to add a head tax to this but public goods are underprovided. model while retaining the assumption of Adding scale economies creates a role for no scale economies, then the response property taxation, but the public good would be exactly the same as in the basic need no longer be underprovided. In fact, tax competition model: governments the Samuelson rule is now satisfied in the would abandon the property tax and use special case of a pure public good (an ex- only the head tax. But now the head tax treme case of scale economies).22 Once would be used to efficiently control mi- again, we cannot identify cases of gration. In particular, each government underprovision by examining whether would make each resident pay a head tax regions choose to tax capital. equal to the marginal cost of supplying Yet another consideration is the man- the public good to another individual. No ner in which labor is mobile. In the mod- other taxes would be needed to balance els discussed so far, individuals choose the the government budget because the ab- region in which to reside and work. In sence of scale economies implies that this contrast, Braid (1996) examines a model marginal cost equals the per capita cost with commuting. The world economy of public good provision. With efficient consists of many metropolitan areas, each head taxation, governments would choose containing a fixed number of identical lo- the efficient level of public good provision, cal jurisdictions with fixed residential which satisfies the Samuelson rule. populations. As in Wilson (1995), competi-

22 In contrast to the head-tax case, however, this equilibrium is inefficient. Wilson (1995) provides conditions under which a rise in every region’s public good supply beyond the equilibrium level reduces welfare if the financing comes from the labor tax, but increases welfare if the property tax is used instead (Prop. 4, p. 349). These results suggest that the primary inefficiency in this case is an inefficient tax mix, not an inefficient level of public good provision. 285 NATIONAL TAX JOURNAL tive firms in each jurisdiction use immo- The Common-Agency Approach bile land and mobile capital and labor to produce output, but now labor is mobile The research strategy reported in the because individuals can costlessly com- previous subsections involves limiting the mute to other jurisdictions. The public government’s power to tax in interesting good in this model is assumed to benefit ways and then investigating the resulting residents, not commuters. To finance it, implications for tax competition. An alter- jurisdictions can employ source-based native approach is to derive such limits taxes on wage and capital income, and as the outcome of specific aspects of the also a property tax, levied uniformly on economic environment. One such aspect both capital and land. Jurisdictions pre- is incompleteness in a government’s in- fer the property tax over a tax on mobile formation about the firms that it is at- capital alone, since the former includes the tempting to tax. For example, firms may nondistortionary land tax. To compete for differ in the degree to which they are commuters, they employ both the prop- interregionally mobile, but such differ- erty tax and the wage tax, rather than ences may be difficult for the government solely relying on the wage tax. Thus, to observe. In this case, the government Braid’s model also endogenizes the cannot tax a firm in a way that directly taxation of property. He ties the level of depends on its unobserved mobility char- taxation to the degree of competition for acteristics. Instead, the government must commuters, as indicated by the number base its tax on observable aspects of firm of jurisdictions. As this number rises, the behavior that may serve as signals of these wage tax declines, the property tax rises, characteristics, such as the firm’s invest- and the public good provision declines.23 ment decisions. This is a principal-agent Thus, Braid’s work shows that compe- problem, with the government serving as tition for commuters increases the degree the principal and the firm as the agent. See to which public goods are underprovided, Osmundsen, Hagen, and Schjelderup while causing greater reliance on the prop- (1998) for an example of such a model. erty tax. In contrast, recall that Wilson When two or more governments com- (1995) finds no such positive relation be- pete for a share of a mobile firm’s profits, tween public good underprovision and this problem becomes a common-agency reliance on the property tax. One impor- problem, with the governments now serv- tant difference between the two models ing as multiple principals. A Nash equi- is that scale economies in public good pro- librium is established when each govern- vision are central to Wilson’s results but ment chooses its optimal policy, given the irrelevant in Braid’s model, since he as- policy chosen by the other government. sumes that only residents consume the The firm’s private information about its public good, and their number is fixed in attributes allows it to earn an “informa- each jurisdiction. It would be interesting tion rent,” adding a new dimension to the to see whether the two models yield more welfare analysis of tax competition. In similar conclusions if commuters also ben- Mezzetti (1997), for example, tax compe- efit from a jurisdiction’s public expendi- tition (i.e., his “independent contracting”) tures. raises the information rent earned by the

23 This decline in public good provision also occurs when only the property tax is available. In another paper, Braid (1998) obtains a similar relation between commuting and public good provision, using a spatial model to capture the costs of commuting. In particular, public good provision falls as commuting costs fall. Braid (1997) adds absentee ownership of land to the model developed by Braid (1996) and finds that overprovision may result, but this is due to “tax exporting,” which I have discussed in the “Capital Tax Competition” section. 286 Forum on Mobility and Tax Analysis firm relative to its level when the two prin- region contains a fixed number of identi- cipals cooperate, and the result is lower cal residents, whose utility depends posi- aggregate welfare for the principals. Thus, tively on the consumption of a private tax competition again worsens welfare, good and a public good, and negatively but for new reasons. on the supply of labor. Public good expen- The common-agency approach seems ditures are financed by a tax on private especially useful for analyzing the taxa- good consumption, levied on an origin tion of multinationals. Many countries basis, and Mintz and Tulkens examine the attempt to tax the income of foreign sub- Nash equilibrium in these tax rates. The sidiaries, suggesting a common agency use of origin-based taxes means that each problem in which the home and host region’s government collects a uniform countries are the principals. Bond and tax on only the output of domestic firms, Gresik (1996) present an interesting model regardless of where this output is ulti- in which the home and host countries in- mately consumed. As a result, the region’s dependently confront the multinational residents can escape the tax by incurring with subsidies and trade taxes. Relative the transport costs necessary to cross the to setting these policy instruments coop- border and purchase the private good in eratively, tax competition is shown to the other region. In contrast, a destination- lower the countries’ aggregate welfare based commodity tax would enable a re- and leave the multinational worse off.24 gion to collect a tax on all of its residents’ In other words, the inefficiencies associ- private good consumption. This could be ated with tax competition turn out to be done through the use of border adjust- detrimental to all parties. ments, under which the tax is collected While there is thus some evidence that from domestic firms, but a tax rebate is the detrimental effects of tax competition given for exports of these goods and a tax in the basic model carry over to the more is collected on imports. (Imports and ex- complex common agency problem, much ports take place through cross-border remains to be done in this exciting area of shopping in the Mintz–Tulkens setup.) In research. As increasingly complex govern- this case, a region’s residents would not ment policies are considered, however, we escape their government’s tax by crossing are led to increasingly question the basic the border. But border adjustments are assumption that government officials seek administratively difficult to enforce, and to maximize welfare, rather than engage in some areas, they have been effectively in self-interested behavior. A model that eliminated, most notably in the European centers around such behavior is discussed Community. See Lockwood (1993) for a in the “Political Economy” section. detailed comparison of commodity tax competition under the destination and COMMODITY TAX COMPETITION origin principles. In contrast to the models of competition In addition to the literature on compe- for capital, Mintz and Tulkens (1986) and tition for scarce capital, a literature on de Crombrugghe and Tulkens (1990) “commodity tax competition” has also describe cases where the equilibrium is been developed. Mintz and Tulkens (1986) fully efficient. However, these cases occur introduce a model where this form of com- when transport costs are so high that no petition occurs between two regions that cross-border shopping occurs, either in are linked by cross-border shopping. Each equilibrium or in response to small tax

24 See also Bond and Gresik (1998), who examine the case where the two governments are asymmetrically in- formed about the firm’s production costs. 287 NATIONAL TAX JOURNAL changes.25 In such cases, there are none sible for the rise in the low-tax region’s of the interregional externalities described tax rate to harm residents of the high-tax previously. It seems difficult, however, to region, i.e., create a negative externality. describe these cases as “tax competition,” Hence, it is not possible to show that both because the governments are not really regions set their tax rates “too low” in competing over a tax base. equilibrium. Instead, tax competition in these mod- To conclude, the pattern of interregional els is associated with tax rate differences externalities is more complex than in the that are sufficiently large to overcome basic tax competition model, producing transport costs, so that the low-tax region less clear-cut results. It is interesting to “exports” the good to cross-border shop- note, however, that the literatures on capi- pers from the high-tax region. In this case, tal tax competition and commodity tax a rise in the high-tax region’s tax rate competition do seem to obtain similar re- raises the amount of shopping done by its sults concerning the advantage of small- residents in the low-tax region, thereby ness. See, in particular, Kanbur and Keen’s increasing the latter’s tax base. This tax (1993) analysis of a spatial model of cross- base change is called the “public con- border shopping.26 sumption effect.” As in the basic tax com- petition model, it represents a positive OTHER FORMS OF COMPETITION externality, implying that the high-tax region’s tax rate is inefficiently low. In There are many ways in which govern- particular, Mintz and Tulkens (1986) show ments can compete for mobile factors or that if a central authority were to force the shoppers other than through the use of two regions to change their tax rates in taxes or subsidies. One possibility is to directions that made both of them better compete through the use of public inputs off, then any such tax changes would in- that improve the productivity of capital. volve a rise in the high-tax region’s tax Keen and Marchand (1997) argue, for ex- rate. ample, that the equilibrium pattern of ex- But would this efficiency-improving tax penditures is inefficiently weighted to- change also involve a rise in the low-tax ward too much public input provision and region’s tax rate? The recent work by too little public good provision, since the Haufler (1998) concludes “not necessar- latter benefits residents but does not at- ily.” When the low-tax region increases its tract capital.27 There is also a literature tax rate, it not only creates the public con- suggesting that regions compete too ag- sumption effect mentioned above, but also gressively for mobile firms through the a “private consumption effect,” consisting use of inefficiently lax environmental poli- of the welfare loss that cross-border shop- cies, creating a “race to the bottom.” Wil- pers experience from the increased price son (1996a) reviews this literature and of the private consumption good. As a concludes that the possibility of a “race” result of these conflicting effects, it is pos- depends critically on assumptions about

25 See Mintz and Tulkens (Prop. 9b) and also de Crombrugghe and Tulkens (Prop. 2). 26 A limitation of this model is that governments care only about maximizing tax revenue. Trandel (1994) and Haufler (1996) analyze spatial models with less extreme objective functions, but they do not directly address the welfare comparisons in Kanbur and Keen (1993). See also Braid (1993). Unlike the other spatial models, he allows lump-sum taxation in addition to commodity taxes. 27 Noiset (1995) and Bayindir-Upmann (1998) obtain less clear-cut results, and Sinn (1997) claims that “public infrastructure,” modeled as reducing the “cost” of investment in a region, is efficiently provided. But neither paper also models the endogenous provision of public goods. If they did, then public expenditures would be inefficiently weighted away from public good provision and toward public inputs or infrastructure 288 Forum on Mobility and Tax Analysis the available tax and subsidy instru- These various examples illustrate the ments.28 As mentioned in the Introduc- multifaceted nature of competition among tion, policymakers have also express- governments, which is likely to create dif- ed concern about the possibility of a ficulties for the design of cooperative “race to the bottom” in welfare benefits, agreements to reduce the wasteful aspects brought on by the mobility of welfare of this competition. It may be possible to recipients. coordinate one policy variable, but doing Other government regulatory programs so may simply cause governments to com- can be subject to wasteful competition pete more vigorously by means of another between governments. See, for example, policy variable. Sinn’s (1997) demonstration that indepen- dent governments choose inefficiently THE POSSIBILITY OF EXCESSIVE low product quality standards, given that TAXATION their products are interregionally traded. In addition, he discusses the failure of in- The general thrust of much of the tax dependent governments to provide ad- competition literature is that tax compe- equate amounts of social insurance in the tition leads to inefficiently low taxes. In presence of factor mobility.29 contrast, two types of tax competition The degree to which existing tax laws are have been found to produce inefficiently enforced may also serve as a strategy vari- high taxes: vertical tax competition and able. Cremer and Gahvari (1996) consider tax competition with double taxation con- two countries that compete with each other ventions. By examining recent work, how- for cross-border shoppers, using both tax ever, I find that this conclusion may be too rates and tax audit probabilities as strate- hasty. gic variables. In this case, any coordination between the two countries in the setting of Vertical Tax Competition tax rates will be offset to some extent by competition for shoppers through reduc- The types of tax competition discussed tions in the audit probabilities. to this point can be characterized as “hori- In some cases, governments may design zontal tax competition,” in the sense that nontax policies that will commit them to the governments doing the competing are change the degree to which they compete all at the same level. There also exists a in taxes with other governments. See, literature on “vertical tax competition” for example, Jensen and Toma’s (1991) between different levels of government, analysis of the use of national debt as a such as the federal government and state strategic variable in a tax competition governments within the United States. model. By committing to debt now, a The basic problem is that each level of government can signal to other govern- government imposes a tax on the same tax ments that it will have to impose higher base. Whereas one state’s tax increases the taxes on capital in the future. If these tax base available to another state under other governments respond by raising horizontal competition, now the tax im- their own capital taxes in the future, then posed by one level of government dimin- the first government will have benefited ishes the size of the tax base available to from this reduction in competition for the other level of government. In the case capital. of capital taxation, for example, a rise in

28 For formal models of a race, see Kim and Wilson (1997), Markusen, Morey, and Olewiler (1996), Oates and Schwab (1988), and Rauscher (1995). 29 On the other hand, the existence of factor mobility may alter voters’ preferences and make them elect govern- ments with more liberal views about redistributive policies. See Persson and Tabellini (1992). 289 NATIONAL TAX JOURNAL the federal government’s tax rate reduces Jensen (1996). But this is not the only national savings, thereby lowering the reasonable assumption. The sequential amount of capital available to each state move models described in the “Corrective government. A rise in a single state’s tax Policies” section simplify matters by as- rate has a similar, but smaller, effect, re- suming that each government chooses its ducing the tax base available to the fed- policies only once. In a more complicated eral government. setting, state governments might both re- Since tax increases now create negative act to the past decisions of the federal gov- externalities, rather than positive exter- ernment and also make new decisions that nalities, we might expect conclusions from influence the federal government’s future the basic tax competition model to be re- behavior. Models with simultaneous versed: taxes are now set too high. This moves sweep all such complications away conclusion is too hasty, however, because by putting the state and federal govern- other aspects of the political-economic ments on an equal footing in terms of the environment must be specified. First, timing of their moves. there is the issue of what objective func- Clearly, the best case for efficiency will tion is employed by the federal govern- occur when the federal government is ment. A benevolent federal government benevolent and is able to move first, so desires to maximize some measure of the that it can influence the behavior of the aggregate welfare of all residents, whereas state governments. Boadway, Marchand, a benevolent state government is con- and Vigneault (1998) consider this case in cerned with maximizing the welfare of the a model where the federal government state’s own residents. Unlike horizontal and identical state governments utilize an tax competition, then, the objectives of the income tax that redistributes income federal government and state govern- among a diverse set of residents. They dis- ments overlap to some extent, reducing tinguish between cases where migration conflict. If the federal government pursues is possible or impossible, with the former objectives that are independent of resident case allowing horizontal tax competition welfare, then we might want to look not to exist. For both cases, the equilibrium is only at ways to facilitate more efficient efficient, given the available policy instru- interactions between the federal govern- ments. In other words, the federal govern- ment and state governments, but also at ment could do no better if it directly con- ways to improve the internal functioning trolled the states’ policy instruments. This of the federal government. outcome can be readily understood. An The other issue to consider is the individual state engages in vertical tax timing of the actions undertaken by the competition by not accounting for the federal and state governments. The “Cor- negative effects of its tax on the federal rective Policies” section discusses optimal budget. But the federal government has fiscal federalism under the standard as- sufficient tax tools and foresight to undo sumption that the federal government any inefficiencies in state government be- “moves” first, committing itself to a set of havior. In other words, vertical tax com- policies that state governments then treat petition seems to be happening at the state as fixed when choosing their own policies. level but not the federal level, because the In contrast, the models of horizontal tax federal government “sees through” the competition assume that all governments state budget constraints when it makes its choose their policies simultaneously. In own policy choices. The result is an effi- some cases, institutional features argue for cient equilibrium. the higher-level government having the If the federal and state governments first-mover advantage; see Hoyt and set their policies simultaneously, then the 290 Forum on Mobility and Tax Analysis federal government can no longer alter we have a situation where a firm’s income state policy choices by changing its own is being taxed by two different countries, policies. An efficient equilibrium is home and host. As a result, double taxa- unachievable in most cases, but a benevo- tion is viewed as a potential problem, and lent federal government can still manipu- three methods have been used to alleviate late its policy instruments to at least par- the problem. First, the home government tially offset inefficiencies at the state level. can provide a tax credit for taxes paid to See Hoyt (1996) for a recent analysis of this the host government. Second, it can allow case.30 foreign investors to deduct these taxes Keen and Kotsogiannis (1996) provide from their taxable income. Finally, it can a good example of the inefficiencies that exempt foreign-source income from taxa- can result when governments are no tion. Of these three methods, the deduc- longer benevolent.31 In their model, the tion method is the least commonly used. federal and state governments care only The formal analysis of how indepen- about maximizing tax revenue. Although dent governments choose their tax poli- there exists both horizontal tax competi- cies under these methods begins with tion between states and vertical tax com- Bond and Samuelson (1989). They use a petition between the federal government two-country model in which a capital-ex- and states, the vertical tax competition porting (home) country and capital-im- problem dominates in the case where all porting (host) country use their tax rates governments move simultaneously. The on foreign-source income as strategy vari- equilibrium tax rates are found to be too ables. The surprising result is that if the high, even relative to those that maximize home country provides tax credits, then tax revenue. the Nash equilibrium involves taxes so To conclude, vertical tax competition high that all international capital flows seems to be a more slippery concept than cease. In striking contrast to the basic tax horizontal tax competition. It seems most competition model, the problem here is likely to create inefficiencies in models not that taxes are too low, but rather that where the federal government is unable they are too high. The basic intuition is to optimally influence the choice of policy that the host country always has an incen- instruments by local governments, due to tive to raise its tax rate at least to the level commitment problems, information prob- levied by the home country, since the lems, or objectives other than welfare latter’s government treasury effectively maximization. Future work should seek pays the tax by providing tax credits to to develop more realistic political pro- foreign investors. But as long as capital cesses within and between the different exports occur, the home country should levels of government. keep its tax rate on foreign-source income above the host country’s rate, since doing Double Taxation Conventions so allows it to exercise its market power on the world capital market, i.e., to drive An important issue in the taxation of up the equilibrium after-tax return on multinationals is the use of double taxa- capital by reducing incentives to invest tion conventions. Given that a home coun- abroad. As a result, capital tax rates are so try attempts to tax foreign-source income, high in equilibrium that all capital exports

30 Hoyt (1996) also shows that a uniform matching grant can be used to ensure that the combined policies of the two levels of government are efficient, given the restrictions on the tax instruments. This result is limited to the case where the “localities” controlled by lower-level governments are identical. For other recent work on the use of intergovernmental grants to offset fiscal externalities, see Boadway and Keen (1996) and Dahlby (1996). 31 See also Keen’s (1998) review of the literature on “vertical tax externalities.” 291 NATIONAL TAX JOURNAL cease. In contrast, capital flows do occur are assumed, and there always exist posi- under tax deductions. Thus, we are left tive capital flows in equilibrium, regard- with the empirical puzzle of why credits less of which double taxation convention are so prevalent in the world economy. is employed. However, the choice among Subsequent research has attempted to them is no longer a matter of indifference. reconcile theory with practice. Davies and Davies endogenizes this choice by allow- Gresik (1998) allow foreign subsidiaries to ing each country to independently choose finance their investments by borrowing which method to employ, prior to the additional capital from the host country’s choice of tax rates. His results provide domestically owned firms, subject to col- some indication that deductions will be lateral requirements. This additional source the preferred method. In particular, they of investment funds dramatically changes are always used in the special case where the effects of tax credits. In particular, the the two countries have identical charac- countries no longer set their tax rates so teristics. While this conclusion once again high that all capital flows are eliminated. seems at odds with the popularity of tax Thus, the use of credits is no longer as bad credits, further results demonstrate the as suggested by Bond and Samuelson desirability of cooperative agreements (1989). This outcome does not fully explain that eliminate the use of deductions. Per- the use of credits, however, since Davies haps the most striking conclusion is that and Gresik still find that the home country in the case of identical countries, eliminat- weakly prefers deductions to credits. ing the use of deductions results in a fully Whereas Bond and Samuelson (1989) efficient equilibrium. Although inefficien- allow the home country to tax at different cies reappear in the asymmetric case, this rates the income its residents earn at home limited efficiency result offers an intrigu- and abroad, Janeba (1995) considers the ing contrast to the inefficiencies found in case where no such discrimination is per- the basic tax competition model. mitted. A surprising implication of this as- To conclude, while Bond and sumption is that the home country sets its Samuelson’s (1989) conclusion that tax tax rate equal to zero under both the tax credits eliminate all capital flows remains credit and exemption methods. The basic a striking result, subsequent research has idea is that a positive tax rate would in- generated less extreme results, ranging all crease capital exports, creating undesir- the way to the efficiency results obtained able terms-of-trade effects. But Janeba also by Davies (1998). In all cases reviewed here, finds that each country’s equilibrium level two-country models are employed, imply- of national income is independent of ing that terms-of-trade effects are central which method is used (credits, deduc- to the analysis. As seen in the “Large Re- tions, or exemptions). Thus, his model gions” section, extending the basic tax provides a solution to the empirical puzzle competition model to the case of asymmet- raised by Bond and Samuelson. Under all ric tax competition between large countries methods, however, the equilibrium tax is itself enough to significantly change the rates continue to be inefficiently set. The welfare implications. Thus, there is per- problem is not that tax rates are too low, haps not as much inconsistency between as in the basic tax competition model, but the two types of models as might appear. that the relative rates induce the home country to export too little capital. EFFICIENCY-ENHANCING In a recent paper, Davies (1998) consid- COMPETITION ers two-way capital flows, where each country exports capital to the other coun- Whereas the Tiebout model produces a try. Again, nondiscriminatory tax policies form of “efficient tax competition,” we 292 Forum on Mobility and Tax Analysis have seen that departures from the ideal- to the per capita cost of public good pro- ized setting of this model can produce vision, which is below the marginal cost. various forms of wasteful tax competition. Nevertheless, bidding for firms never re- Recently, researchers have turned their duces the social efficiency of firm location, attention to the possibility that such de- and in some cases, this bidding causes partures may also create an efficiency-en- firms to locate more efficiently. Black and hancing role for competition involving tax Hoyt conjecture, however, that the use of or nontax policy instruments. This section public services to compete for firms will describes some work in this direction. not produce an efficient outcome. They also demonstrate that firm location deci- Bidding for Firms sions are inefficient in cases where the firm possesses private information about how The discussion so far has assumed that its production costs differ between the two capital investments may be made in small regions, which it is unable to reveal to the increments. In many cases, however, in- two regional governments.32 vestments come only in large increments, The dynamic model developed by i.e., they are lumpy. Regions then find King, McAfee, and Welling (1993) also de- themselves competing for these lumps, parts from the Tiebout world by introduc- often taking the form of large firms. Com- ing uncertainties about firm productivity. petition for automobile plants is one ex- In particular, the social value of a firm is ample. It is often assumed that regions are given by the “surplus” that it generates able to target subsidies and other tax by producing in a region, but this surplus breaks to these mobile firms, rather than is uncertain to both the firm and regional altering the entire tax system. This com- governments prior to actual production. petition may be referred to as “bidding for Two regions compete for the firm over two firms.” In this case, the effect of tax com- periods. After choosing a location in the petition on the public goods provided to first period, the firm is free to relocate (at residents is no longer a central issue. But a cost) in the second period. The firm’s we can still ask whether the subsidies or location in each period is determined by tax breaks provided to firms are efficient an auction mechanism, and this location in the sense that they lead to efficient firm is found to be efficient. The second part location decisions while not creating any of the paper allows each region to invest unnecessary costs for the system of re- in “infrastructure.” Before the auction gions as a whole. Additional efficiency takes place, the two regions play a Nash issues arise in cases where firms receive game in investment levels, under which the benefits of public services and infra- each region sets its investment level opti- structure. mally, given the level chosen by the other Two well-known papers produce mod- region. The authors demonstrate that only els in which the bidding for a large firm an asymmetric Nash equilibrium exists, enhances efficiency. In Black and Hoyt where the equilibrium investment levels (1989), two regions compete for a large differ. In the first period, the firm locates firm because its presence attracts more where investment is highest. However, residents, which lowers the average cost the losing region may choose a positive of providing a public good to existing resi- (but lower) investment level, because this dents. The Tiebout assumption of efficient raises the probability that the firm will head taxes is absent from this model. In switch locations in the second period. This particular, each resident pays a tax equal possibility of relocation implies that the

32 Martin (1997) focuses on the problems associated with bidding for firms that possess private information. 293 NATIONAL TAX JOURNAL losing region’s investment is not socially through the provision of subsidies may wasteful. In fact, the authors show that the be better than no subsidies, but the equi- equilibrium is efficient. librium levels of these subsidies are not Both of these models follow the previ- likely to be optimal from the nation’s ous literature by assuming that each gov- viewpoint. ernment is concerned with the welfare of its own citizens. In contrast, Biglaiser and Imperfect Competition Mezzetti (1997) investigate a model in which attracting mobile firms provides a If firms are large, then the issue of im- state governor with the opportunity to perfect competition becomes potentially engage in activities that imperfectly sig- important. The international trade litera- nal his “ability” to voters. When two or ture on “strategic trade policy” has already more governors with re-election concerns exhaustively explored the effects of imper- compete for firms, the resulting location fect competition on a country’s optimal of the firm will not necessarily be efficient. trade policies. In particular, this literature However, such inefficiencies are not a has justified the use of output subsidies to problem associated with large projects or encourage the country’s firms to compete firms per se, but rather with the imper- more aggressively on world markets, fect political institutions. thereby increasing profits at the expense It is also important to recognize that in- of foreign firms. Janeba (1998a) combines terregional externalities may play an im- these strategic trade motives with a model portant efficiency-reducing role, even in of tax competition. He first follows the cases where subsidies can be targeted to strategic trade literature by specifying a individual firms. But now such externali- model with two countries, each contain- ties can easily work in favor of too little ing a single firm that sells output in a third competition for firms. For example, if one market. In this case, the two governments state in the United States attracts a foreign compete by offering subsidies to their firm, then all states may benefit in the firms. But then Janeba allows each firm to form of lower prices, due to reduced trans- be mobile between the two countries, port costs. This example has similarities meaning that it locates where its after-tax to the pecuniary externalities discussed in profits are highest. The governments now the context of large regions in the “Large recognize that their subsidies will affect Regions” section: by “importing” the firm, not only firm output decisions, but also a region creates desirable price effects for location decisions. In particular, each gov- other regions.33 But such price effects can ernment may seek to attract the other work the other way. If all states in the country’s firm and thereby capture some United States are competing for a foreign of its profits through taxation. Janeba as- firm that faces limited opportunities for sumes that the tax system must be non- locating its plant outside the country, discriminatory, meaning that a country then they will possess market power that imposes the same tax rate on the outputs can be exercised by competing less vigor- of all firms that operate within its borders, ously. Competition for foreign firms whether domestic or foreign.34

33 Rauscher (1995) identifies this type of externality as a possible reason for why regions might not seek to attract a polluting firm, even though it is desirable to do so (a phenomenon called “not in my backyard”). 34 Although countries do typically discriminate to some extent, they are restricted in this practice by the nondis- crimination rules established by international agreements such as GATT or the laws of the European Union. Janeba and Peters (1999) focus on the different implications of nondiscriminatory taxes and taxes that dis- criminate between domestic and foreign firms operating within a country, but they abstract from market structure issues by assuming that each country has access to an “immobile tax base” and competes for a “mobile tax base.” 294 Forum on Mobility and Tax Analysis

Janeba’s (1998a) surprising conclusion welfare-improving effects of tax competi- is that competition for mobile firms causes tion by failing to recognize possible the countries to compete their tax rates sources of inefficiency that might exist in down to zero. No country offers a tax rate the absence of capital mobility. In the “Po- below zero, because it would then attract litical Economy” section, inefficiencies both of the firms but be hurt by the trans- from the functioning of the political pro- fer of subsidy revenue to the foreign firm. cess are considered. Janeba is also able to generalize the zero- tax result to include cases where the firms’ Commitment Problems outputs are sold to the consumers in one of the two countries, rather than in a third The basic tax competition model as- country. In these cases, the country con- sumes that governments commit to a tax taining these consumers cares about con- system, and then capital owners choose sumers’ surplus, along with tax revenue where to invest their capital. In practice, and its firm’s profits. In neither case is the commitment issues arise because firms or equilibrium fully efficient, since the inef- capital become partially immobile once ficiencies associated with imperfect com- location decisions are made. One way to petition are still present. But tax competi- deal with this problem is to commit to ini- tion does improve welfare. tial subsidies or “tax holidays” for new Thus, imperfect competition seems to firms, thereby shortening the period of dramatically alter the welfare implications time in which commitment is required. But of tax competition.35 Recall, however, that such policies have the disadvantage of in- regions would choose zero taxes on mobile creasing firm turnover. In other words, capital in the basic tax competition model some fraction of firms may choose to leave if they had access to head taxes. Capital a region after the initial tax break has ex- taxes are avoided because each region faces pired, perhaps seeking tax breaks in other an infinitely elastic supply of capital. Simi- regions. Wilson (1996b) models the use of larly, each country’s production capacity is initial subsidies to attract new firms, find- infinitely elastic in Janeba’s model; it will ing that excessive firm turnover is indeed lose all of its capacity if its tax rate is in- the equilibrium outcome. Bond (1981) creased slightly above the other country’s finds empirical evidence of the problem. rate. But the two models differ in how gov- Janeba (1998b) demonstrates that tax ernments behave when the borders are competition may actually help solve com- closed. The equilibrium is fully efficient in mitment problems. He investigates a one- the basic model, whereas wasteful subsi- firm, two-region model in which decisions dies emerge when there is imperfect com- are made in three stages. First, the firm petition. Thus, tax competition is able to undertakes a single project consisting of play an efficiency-enhancing role under investment in “capacity” in each country. imperfect competition, but not in the com- Oil pipelines are a possible example. Sec- petitive environment of the basic model. ond, each government chooses the rate at This comparison suggests the basic tax which to tax the project output within its competition model could be accused of borders. Finally, the firm chooses its out- “stacking the deck” against the possible puts.36 A commitment problem arises here

35 Note also the importance of the nondiscrimination assumption. If discrimination in taxes were possible, then the two governments would continue to provide wasteful subsidies to their own firms. 36 In contrast, Kehoe (1989) considers a two-country model in which savings, but not the location of investment, is chosen prior to tax rate decisions. If the two countries collude in setting their tax policies, then they tax away savings, since it is fixed at the time taxes are chosen. As a result, nobody chooses to save. Tax compe- tition is preferred, because governments forgo capital taxation in an effort to attract investment. 295 NATIONAL TAX JOURNAL because the governments are able to amine this view formally in various “Le- choose tax rates after the firm has fixed viathan models,” where governments are its capacity levels. If there were a single concerned in part with maximizing the government, then it would have an incen- size of the public sector. Their conclusions tive to tax all profits away. The firm would about the welfare implications of tax com- recognize this incentive at the time of its petition are mixed, but all three papers initial investment decision and choose not assume that governments retain some to invest in capacity. When there are two degree of “benevolence,” perhaps caused governments, they will “compete” in tax by re-election concerns that are not for- rates if the firm has excess capacity and is mally modeled. therefore able to reallocate output be- It is difficult to ascertain empirically tween the two regions in response to dif- whether the welfare-improving or wel- ferences in the tax rates.37 Provided invest- fare-worsening view of tax competition is ment costs are sufficiently low, the firm more accurate, since both views seem to then chooses to undertake the project by predict that an increase in the number of investing in excess capacity as a means of competing governments should reduce creating tax competition. To conclude, the total size of government. Moreover, the commitment problems provide another empirical tests of the Leviathan model possible role for tax competition as an ef- conducted by Oates (1985, 1989) and oth- ficiency-enhancing activity. ers have encountered difficulties in even confirming that there is a relation between Political Economy aggregate government size and the decen- tralization of fiscal decisions among inde- A common assumption in the literature pendent governments, let alone identify- reviewed so far is that each regional gov- ing the welfare implications of such a re- ernment seeks to act in the best interests lation.38 It therefore appears useful to ex- of some set of residents or factor owners. plore alternative models of government Indeed, the basic tax competition model decision making that do not necessarily and many of its extensions assume away create this relation. differences between residents, so that Wilson and Gordon (1998) depart from there is no conflict about which residents’ the assumption in Leviathan models that preferences are given the most weight. In a single monolithic entity exercises con- the basic model, the only potential source trol over the entire range of tax and pub- of inefficiency is tax competition, making lic expenditure instruments. Their model it relatively easy for this competition to recognizes that some policy instruments turn out to be a bad thing. A rather differ- might be more accurately modeled as ef- ent perspective is taken by the public fectively under the control of electorates choice literature. Brennan and Buchanan or their representatives, whereas others (1980) argue that tax competition im- are largely delegated to self-interested proves welfare, because the size of gov- government officials, leaving the elector- ernment would be excessive in the ab- ate with only rudimentary methods of sence of this competition. Rauscher (1996, control. In particular, government officials 1998) and Edwards and Keen (1996) ex- are assumed to choose public expenditure

37 The concept of excess capacity is well-defined here, because the model assumes that the demand for the firm’s output is completely inelastic at prices below the consumers’ reservation level. 38 After reviewing his and other work, Oates (1989) concludes, “The empirical literature on fiscal centralization and government size thus contains a number of puzzles and inconsistent findings” (p. 582). More recently, Anderson and Van den Berg (1998) find no evidence of a relation between fiscal decentralization and govern- ment size. 296 Forum on Mobility and Tax Analysis policies but not tax rates, which are cho- bor, and thus there are no opportunities sen to maximize welfare. In another pa- for residents to confront public officials per, Gordon and Wilson (1998) argue at with a tax base that can be expanded length that this asymmetric treatment of through additional public good provision. tax and expenditure policies appears to be Rather, other incentive devices must be a good description of the situation of most relied upon. The model assumes that of- government bureaucrats. For example, ficials can be replaced with a probability while legislatures have substantial control that is related to their job performance, over the choice of tax rates, they cannot and that the effective penalty from being deal adequately with the innumerable replaced can be directly controlled specific expenditure and regulatory deci- through the choice of the officials’ sala- sions that affect the tax base, and they ries. must therefore delegate these decisions to By creating expenditure competition, others. Moreover, the electorate can moni- opening the economy increases the “effi- tor easily what happens to tax rates but ciency” with which government officials has a harder time monitoring the quality utilize tax revenues; there is less “waste” of the many different expenditure deci- in government. Consequently, expendi- sions. ture competition improves welfare in all Following Niskanen (1971) and the sub- regions. But regions also engage in a form sequent Leviathan models, Wilson and of tax competition similar to what occurs Gordon (1998) assume that government in the basic tax competition, causing tax officials benefit personally from the bud- rates to be too low. In the present case, get they control and, as a result, face in- each region fails to design its tax system centives to pursue activities that increase to fully exploit the potential incentive ef- the size of the budget. A positive connec- fects created by labor mobility. Higher tax tion between the tax base and public good rates would strengthen the connection levels is modeled by assuming a system between the level of tax revenue and pub- of many identical regions, each consisting lic good provision, thereby providing of a fixed amount of land, and allowing greater incentives for public officials to labor to be perfectly mobile across these increase public good supplies. But the re- regions. Each region employs a linear in- gions compete for mobile labor through come tax, consisting of a head tax or sub- reductions in tax rates. If they were to all sidy and a uniform tax rate on all labor raise their tax rates simultaneously, then and land income earned within the the incentive would be strengthened with- region’s borders. Given this tax system, out any outflows of labor. the government officials in each region Despite the efficiency losses from tax engage in “expenditure competition” by competition, welfare in the open economy increasing their provision of public goods exceeds welfare in the closed economy to attract more labor, thereby expanding because of the existence of expenditure the tax base. competition. The potential tax incentives To isolate the efficiency-enhancing ef- associated with mobile labor may not be fects of expenditure competition, Wilson fully exploited, but at least such incentives and Gordon (1998) compare the equilib- are present, in contrast to their complete rium in this “open economy” with the absence in the closed economy. equilibrium in a “closed economy,” in By reducing waste in government, which the lack of interregional factor mo- opening the economy may reduce the bility means that all factor supplies are effective “price” of the public good fixed. In particular, each resident pos- enough to induce residents to provide the sesses fixed endowments of land and la- public sector with more tax revenue than 297 NATIONAL TAX JOURNAL is available in the closed economy. Thus, ernmental competition for mobile factors opening the economy may raise or lower has been shown to play a beneficial role. the total “size” of government, although Briefly stated, this competition may in- welfare unambiguously rises. This conclu- duce government officials to reduce waste sion is consistent with the empirical am- in government. This possibility brings us biguities identified by Oates (1985, 1989) back to the contrast between Tiebout mod- and others. els and tax competition models made Although Wilson and Gordon (1998) do in my introductory remarks. Tiebout not model capital mobility, it is likely to models are motivated by the view that play a similar efficiency-enhancing role. competition among independent govern- In this case, government officials will en- ments is similar to competition in the gage in expenditure competition by in- private sector and therefore has desirable creasing those public inputs that enhance efficiency properties. In contrast, tax the productivity of capital. But it is also competition models often take the view possible that this competition for capital that intergovernmental competition de- might inefficiently distort the pattern of parts from the assumptions of the stan- public expenditures away from expendi- dard competitive model in ways that tures on public goods or inputs that do negate its efficiency properties. This pa- not enhance capital productivity, an out- per has stressed the role of interregional come similar to the Keen–Marchand externalities in this regard. As mentioned (1997) finding mentioned above. in the Introduction, Sinn (1997) stresses the nature of the goods and services pro- CONCLUDING REMARKS vided by governments: since they tend to be those goods and services for which The intellectual history of tax competi- competitive markets do not perform tion seems to have taken a normal route, well, reintroducing competition among going from simple models yielding governments in their provision is likely straightforward results to more compli- to reintroduce market failures. The politi- cated and less clear-cut conclusions. The cal approach to modeling intergovern- original insight that tax competition can mental competition takes a middle lead to inefficiently low taxes and public ground. On the one hand, it follows the good levels has been shown to hold in Tiebout approach by recognizing that this more general settings than originally in- competition introduces efficiency-enhanc- vestigated. However, the literature has ing incentives similar to the profit motives also identified circumstances under which facing competitive firms. On the other other inefficiencies occur, and it has also hand, it departs from Tiebout models by investigated competition involving a va- recognizing that such incentives operate riety of nontax instruments. Competition in an environment characterized by mar- among governments is now seen as a less ket failures that make a fully efficient equi- straightforward phenomenon than per- librium unattainable. As such, competi- haps originally envisioned. In fact, recent tion among governments has both good work has begun to examine models in and bad aspects, the importance of which which this competition has beneficial as- vary across the attributes of the goods and pects. services that the governments provide. The literature has also begun to inves- This assessment suggests a role for inter- tigate models in which political processes vention by a central authority, but both involving self-interested government of- political considerations and information ficials take center stage. Here, intergov- problems should be carefully addressed.

298 Forum on Mobility and Tax Analysis

Acknowledgments Boadway, Robin, Maurice Marchand, and Marianne Vigneault. This paper grew out of a series of lec- “The Consequences of Overlapping Tax tures on tax competition that I presented Bases for Redistribution and Public Spend- during a visit to the Center for Economic ing in a Federation.” Journal of Public Eco- Studies, University of Munich, in Decem- nomics 68 (1998): 453–78. ber 1996. I am grateful to Hans-Werner Bond, Eric W. Sinn and other members of CES for a “Tax Holidays and Industry Behavior.” Re- stimulating visit, and I thank Ralph Braid, view of Economics and Statistics 63 (1981): 88– Sam Bucovetsky, William Hoyt, Eckhard 95. Janeba, Patricia Wilson, and John Yinger Bond, Eric W., and Thomas A. Gresik. for helpful comments and suggestions on “Regulation of Multinational Firms with this paper. Two Active Governments: A Common Agency Approach.” Journal of Public Econom- ics 59 No. 1 (January, 1996): 33–53. References Bond, Eric W., and Thomas A Gresik. “Incentive Compatible Information Transfer Anderson, John E., and Hendrik Van den Berg. Between Asymetrically Informed Princi- “Fiscal Decentralization and Government pals.” Department of Economics, Pennsyl- Size: An International Test for Leviathan vania State University. Mimeo, 1998. Accounting for Unmeasured Economic Ac- Bond, Eric W., and Larry Samuelson. tivity.” International Tax and Public Finance 5 “Strategic Behavior and the Rules for Inter- (1998): 171–86. national Taxation of Capital.” Economic Jour- Bayindir-Upmann, Thorsten. nal 99 (1989): 1099–1111. “Two Games of Interjurisdictional Compe- Braid, Ralph M. tition When Local Governments Provide In- “Spatial Competition Between Jurisdictions dustrial Public Goods.” International Tax and Which Tax Perfectly Competitive Retail (or Public Finance 5 (1998): 471–87. Production) Centers.” Journal of Urban Eco- Beck, John H. nomics 34 (1993): 75–95. “Tax Competition, Uniform Assessment, Braid, Ralph M. and the Benefit Principle. Journal of Urban “Symmetric Tax Competition with Multiple Economics 13 (1983): 127–46. Jurisdictions in Each Metropolitan Area.” Biglaiser, Gary, and Claudio Mezzetti. American Economic Review 86 No. 5 (Decem- “Politicians’ Decision Making with Re-Elec- ber 1996): 1279–90. tion Concerns.” Journal of Public Economics Braid, Ralph. M. 66 (1997): 425–47. “Property Taxes, Wage Taxes, and Tax Ex- Black, Dan A., and William H. Hoyt. porting with Multiple Jurisdictions in Each “Bidding for Firms.” American Economic Re- Metropolitan Area.” Department of Econom- view 79 (1989): 1249–56. ics, Wayne State University. Mimeo, 1997. Blumenthal, Marsha, and Joel B. Slemrod. Braid, Ralph M. “The Compliance Cost of Taxing Foreign- “A Spatial Model of Tax Competition with Source Income: Its Magnitude, Determi- Multiple Tax Instruments.” Department of nants, and Policy Implications.” International Economics, Wayne State University. Mimeo, Tax and Public Finance 2 (1995): 37–54. 1998. Boadway, Robin, and Michael Keen. Brennan, Geoffrey, and James Buchanan. “Efficiency and the Optimal Direction of The Power to Tax: Analytical Foundations of a Federal-State Transfers.” International Tax Fiscal Constitution. New York: Cambridge and Public Finance 3 (1996): 137–55. University Press, 1980.

299 NATIONAL TAX JOURNAL

Brueckner, Jan K. Davies, Ronald B., and Thomas A. Gresik. “Tiebout/Tax-Competition Models.” Work- “Tax Competition and Foreign Debt.” Work- ing Paper No. 69. Champaign: The Institute ing Paper No. 4-98-1. University Park: De- of Government and Public Affairs, Univer- partment of Economics, Pennsylvania State sity of Illinois at Urbana–Champaign, 1999. University, 1998. Brueckner, Jan K., and Luz A. Saavedra. de Crombrugghe, Alain, and Henry Tulkens. “Do Local Governments Engage in Strate- “On Pareto Improving Commodity Tax gic Tax Competition.” Department of Eco- Changes Under Fiscal Competiton.” Journal nomics, University of Illinois at Urbana– of Public Economics 41 (1990): 335–50. Champaign. Mimeo, 1998. DePater, James A., and Gordon M. Myers. Bucovetsky, Sam. “Strategic Capital Tax Competition: A Pecu- “Asymmetric Tax Competition.” Journal of niary Externality and a Corrective Device.” Urban Economics 30 (1991): 67–181. Journal of Urban Economics 36 (1994): 66–78. Bucovetsky, Sam. Dhillon, Amitra, Carlo Perroni, and Kimberley “Rent Seeking and Tax Competition,” Jour- Scharf. nal of Public Economics 58 (1995): 337–63. “Implementing Tax Coordination.” Journal Bucovetsky, Sam, and John D. Wilson. of Public Economics (forthcoming). “Tax Competition with Two Tax Instru- Diamond, Peter A., and James Mirrlees. ments” Regional Science and Urban Econom- “Optimal Taxation and Public Production.” ics 21 (1991): 333–50. American Economic Review 61 (1971): 8–27 Bucovetsky, Sam, Maurice Marchand, and and 261–78. Pierre Pestieau. Dixit, Avinash, and Victor Norman. “Tax Competition and Revelation of Prefer- Theory of International Trade. London: Cam- ences for Public Expenditure.” Journal of bridge University Press, 1980. Urban Economics 44 (1998): 367–90. Edwards, Jeremy, and Michael Keen. Burbidge, John B., and Gordon M. Myers. “Tax Competition and Leviathan.” European “Population Mobility and Capital Compe- Economic Review 40 (1996): 113–34. tition.” Regional Science and Urban Econom- Fischel, William A. ics 24 (1994): 441–59. “Fiscal and Environmental Considerations in Burbidge, John B., James A. DePater, Gordon the Location of Firms in Suburban Commu- M. Myers, and Abhijit Sengupta. nities.” In Fiscal Zoning and Land Use Controls, “A Coalition-Formation Approach to Equi- edited by Edwin Mills and Wallace Oates, librium Federations and Trading Blocs.” 119–74. Lexington, MA: D.C. Heath, 1975. American Economic Review 87 (1997): 940– Frenkel, Jacob A., Assaf Razin, and Efraim 56. Sadka. Cremer, Helmuth, and Firouz Gahvari. International Taxation in an Integrated World. “Tax Evasion, Tax Policy and Economic Cambridge, MA: MIT Press, 1991. Integration.” Department of Economics, Gordon, Roger H. University of Illinois at Urbana– “Taxation of Investment and Savings in the Champaign. Mimeo, 1996. World Economy.” American Economic Review Dahlby, Bev. 76 (1986): 1086–1102. “Fiscal Externalities and the Design of Gov- Gordon, Roger H., and Jeffrey K. MacKie- ernment Grants.” International Tax and Pub- Mason. lic Finance 3 (1996): 397–412. “Why is There Corporate Taxation in a Small Davies, Ronald B. Open Economy? The Role of Transfer Pric- “The OECD Model Tax Treaty: Tax Compe- ing and Income Shifting.” In The Effects of tition and Two-Way Capital Flows.” Depart- Taxation on Multinational Corporations, edited ment of Economics, Pennsylvania State by M. Feldstein, J. Hines, Jr., and R. Hubbard. University. Mimeo, 1998. Chicago: University of Chicago Press, 1995. 300 Forum on Mobility and Tax Analysis

Gordon, Roger H., and A. Lans Bovenberg. Hoyt, William H. “Why is Capital so Immobile Internation- “Competitive Jurisdictions, Congestion, and ally? Possible Explanations and Implica- the Henry George Theorem: When Should tions for Capital Income Taxation.” Ameri- Property Be Taxed Instead of Land?” Re- can Economic Review 86 (1996): 1057–75. gional Science and Urban Economics 21 Gordon, Roger H., and John D. Wilson. (1991a): 351–70. “Tax Structure and Government Behavior: Hoyt, William H. A Principal-Agent Model of Government.” “Property Taxation, Nash Equilibrium, and Department of Economics, Michigan State Market Power.” Journal of Urban Economics University. Mimeo, 1998. 34 (1991b): 123–31. Gresik, Thomas. Hoyt, William H. “The Taxing Task of Taxing Transnationals.” “Tax Policy Coordination and Optimal Taxa- Department of Economics, Pennsylvania tion in a System of Hierarchical Govern- State University. Mimeo, 1998. ments.” Department of Economics, Univer- Haaparanta, Pertti. sity of Kentucky. Mimeo, 1996. “Competition for Foreign Direct Invest- Hoyt, William H., and Richard A. Jensen, ments.” Journal of Public Economics 63 (1996): “Precommitment in a System of Hierarchi- 141–53. cal Governments.” Regional Science and Ur- Hamilton, Bruce W. ban Economics 26 (1996): 481–504. “Zoning and Property Taxation in a System Huang, Yophy. of Local Governments.” Urban Studies 12 Tax Competition with Involuntary Unemploy- (1975): 205–11. ment. Ph.D. diss. Bloomington: Indiana Uni- Hamilton, Bruce W. versity, 1992. “Capitalization of Intrajurisdictional Differ- Huber, Bernd. ences in Local Tax Prices.” American Eco- “Tax Competition and Tax Coordination in nomic Review 66 (1976): 743–53. an Optimum Income Tax Model.” Journal of Haufler, Andreas. Public Economics 71 (1999): 441–58. “Tax Coordination with Different Prefer- Janeba, Eckhard. ences for Public Goods: Conflict or Har- “Corporate Income Tax Competition, mony of Interest?” International Tax and Double Taxation Treaties, and Foreign Di- Public Finance 3 (1996): 5–28. rect Investment.” Journal of Public Econom- Haufler, Andreas. ics 56 (1995): 311–26. “Asymmetric Commodity Tax Competition— Janeba, Eckhard. Comment on de Crombrugghe and Tulkens. “Tax Competition in Imperfectly Competi- Journal of Public Economics 67 (1998): 135–44. tive Markets.” Journal of International Eco- Haufler, Andreas, and Ian Wooton. nomics 44 (1998a): 135–53. “Country Size and Tax Competition for For- Janeba, Eckhard. eign Direct Investment.” Journal of Public “Tax Competition When Governments Lack Economics 71 (1999): 121–39. Commitment: Excess Capacity as a Hines, James R., Jr. Countervailing Threat.” Department of Eco- “Tax Policy and the Activities of Multina- nomics, Indiana University. Mimeo, 1998b. tional Corporations.” In Fiscal Policy: Les- Janeba, Eckhard, and Wolfgang Peters. sons from Economic Research, edited by “Tax Evasion, Tax Competition and the Alan J. Auerbach, 401–45. Cambridge, MA: Gains from Nondiscrimination: The Case MIT Press, 1997. of Interest Taxation in Europe.” Economic Hines, James R., Jr. Journal 109 (1999): 93–101. “Lessons from Behavioral Responses to In- Jensen, Richard, and Eugenia F. Toma. ternational Taxation.” National Tax Journal “Debt in a Model of Tax Competition.” Regional 52 No. 2 (June, 1999): 305–22. Science and Urban Economics 21 (1991): 371–92. 301 NATIONAL TAX JOURNAL

Kanbur, Ravi, and Michael Keen. Lockwood, Ben. “Jeux Sans Frontieres: Tax Competition and “Commodity Tax Competition Under Des- Tax Coordination When Countries Differ in tination and Origin Principles.” Journal of Size.” American Economic Review 83 (1993): Public Economics 52 (1993): 141–62. 877–92. Markusen, James R., Edward R. Morey, and Keen, Michael. Nancy Olewiler. “Vertical Tax Externalities in the Theory of “Competition in Regional Environmental Poli- Fiscal Federalism.” International Monetary cies When Plant Locations Are Endogenous.” Fund Staff Papers 45 (1998): 454–84. Journal of Public Economics 56 (1996): 55–77. Keen, Michael, and Christos Kotsogiannis. Martin, Laurent. “Federalism and Tax Competition,” Depart- “Bidding for Firms: An Asymmetric Auc- ment of Economics, University of Essex. tion Model of Interjurisdictional Competi- Mimeo, 1996. tion.” Department of Economics, University Keen, Michael, and Maurice Marchand. of Maryland. Mimeo, 1997. “Fiscal Competition and the Pattern of Pub- Mezzetti, Claudio. lic Spending.” Journal of Public Economics 66 “Common Agency with Horizontally Differ- (1997): 33–53. entiated Principals.” Rand Journal of Econom- Kehoe, Patrick J. ics 28 (1997): 323–45. “Policy Cooperation Among Benevolent Mieszkowski, Peter, and George Zodrow. Governments May Be Undesirable.” Review “Taxation in the Tiebout Model.” Journal of of Economic Studies 56 (1989): 289–96. Economic Literature 27 (1989): 1089–1146. Kennan, John, and Raymond Riezman. Mintz, Jack, and Henry Tulkens. “Do Big Countries Win Tariff Wars?” “Commodity Tax Competition Between International Economic Review 29 (1988): Member States of a Federation: Equilibrium 81–5. and Efficiency.” Journal of Public Economics Kim, Jinyoung, and John D. Wilson. 29 (1986): 133–72. “Racing to the Bottom with Multiple Tax In- Mirrlees, James A. struments.” Japan and the World Economy 9 “An Exploration in the Theory of Optimum (1997): 537–51. Income Taxation.” Review of Economic Stud- King, Ian, R. Preston McAfee, and Linda ies 38 (1971): 175–208. Welling. Niskanen, William A., Jr. “Industrial Blackmail: Dynamic Tax Bureaucracy and Representative Government. Competition and Public Investment.” Chicago: Aldine, 1971. Canadian Journal of Economics 26 (1993): 590– Noiset, Luc. 608. “Pigou, Tiebout, Property Taxation, and the Krelove, Russell. Underprovision of Local Public Goods: “The Persistence and Inefficiency of Prop- Comment.” Journal of Urban Economics 38 erty Tax Finance of Local Public Expendi- (1995): 312–16. tures.” Journal of Public Economics 51 (1993): Oates,Wallace E. 415–35. Fiscal Federalism. New York: Harcourt Brace Lee, Kangoh. Jovanovich, 1972. “Tax Competition with Imperfectly Mobile Oates, Wallace E. Capital.” Journal of Urban Economics 42 “Searching for Leviathan: An Empirical (1997): 222–42. Analysis.” American Economic Review 75 Lee, Kangoh. (1985): 748–57. “Factor Ownership and Governmental Stra- Oates, Wallace E. tegic Interaction: An Efficient Nash Equi- “Searching for Leviathan: A Reply and Some librium.” Department of Economics, Further Reflections.” American Economic Towson University. Mimeo, 1998. Review 79 (1989): 578–83. 302 Forum on Mobility and Tax Analysis

Oates, Wallace E., and Robert M. Schwab. Sinn, Hans-Werner. “Economic Competition Among Jurisdic- “The Selection Principle and Market Failure tions: Efficiency Enhancing or Distortion In- in Systems Competition.” Journal of Public ducing?” Journal of Public Economics 35 Economics 66 (1997): 247–74. (1988): 333–54. Tannenwald, Robert. . Osmundsen, Petter, Kare P. Hagen, and “Devolution: The New Federalism—An Guttorm Schjelderup. Overview.” New England Economic Review “Internationally Mobile Firms and Tax (1998): 1–12. Policy.” Journal of International Economics 45 Tiebout, Charles M. (1998): 97–113. “A Pure Theory of Local Expenditures.” Jour- Perroni, Carlo, and Kimberley A. Scharf. nal of Political Economy 64 (1956): 416–24. “Tiebout with Politics: Capital Tax Compe- Trandel, Gregory A. tition and Constitutional Choices.” Working “Interstate Commodity Tax Differentials and Paper No. 1997–05. Copenhagen: Economic the Distribution of Residents.” Journal of Policy Research Unit, Copenhagen Business Public Economics 53 (1994): 435–57. School, 1997. Wellisch, Dietmar. Persson, Torsten, and Guido Tabellini. The Theory of Public Finance in a Federal State. “The Politics of 1992: Fiscal Policy and Eu- New York: Cambridge University Press ropean Integration.” Review of Economic (forthcoming). Studies 59 (1992): 689–701. White, Michelle J. Rauscher, Michael. “Firm Location in a Zoned Metropolitan “Environmental Regulation and the Loca- Area.” In Fiscal Zoning and Land Use tion of Polluting Industries.” International Controls, edited by Edwin Mills and Wallace Tax and Public Finance 2 (1995): 229–44. E. Oates. Lexington, MA: D.C. Heath, Rauscher, Michael. 1975. “Interjurisdictional Competition and the Ef- Wildasin, David E. ficiency of the Pubic Sector: The Triumph “Nash Equilibria in Models of Fiscal Com- of the Market over the State?” Working Pa- petition.” Journal of Public Economics (1988). per 732. Kiel: Institute of World Econom- Wildasin, David E. ics, 1996. “Interjurisdictional Capital Mobility: Fiscal Rauscher, Michael. Externality and a Corrective Subsidy.” Jour- “Leviathan and Competition Among Juris- nal of Urban Economics 25 (1989): 193–212. dictions: The Case of Benefit Taxation.” Wildasin, David E. Journal of Urban Econonomics 44 (1998): 59– “Some Rudimentary Duopolity Theory.” Re- 67. gional Science and Urban Economics 21 (1991): Razin, Assaf, and Efraim Sadka. 393–421. “International Tax Competition and Gains Wilson, John D. from Tax Harmonization.” Economics Letters “A Theory of Interregional Tax Competi- 37 (1991): 69–76. tion.” Journal of Urban Economics 19 (1986): Richter, Wolfram F., and Dietmar Wellisch. 296–315. “The Provision of Local Public Goods and Wilson, John D. Factors in the Presence of Firm and House- “Trade, Capital Mobility, and Tax Competi- hold Mobility.” Journal of Public Economics tion.” Journal of Political Economy 95 (1987): 60 (1996): 73–93. 835–56. Ross, Stephen, and John Yinger. Wilson, John D. “Sorting and Voting: A Review of the Lit- “Tax Competition with Interregional Differ- erature on Urban Public Finance.” Center ences in Factor Endowments.” Regional for Policy Research, The Maxwell School, Science and Urban Economics 21 (1991): 423– Syracuse University. Mimeo, 1998. 52. 303 NATIONAL TAX JOURNAL

Wilson, John D. Bhagwati, edited by Robert C. Feenstra, Gene “Mobile Labor, Multiple Tax Instruments, M. Grossman, and Douglas A. Irwin, 225– and Tax Competition.” Journal of Urban Eco- 44. Cambridge, MA: MIT Press, 1996b. nomics 38 (1995): 333–56. Wilson, John D. Wilson, John D. “Property Taxation, Congestion, and Local “Capital Mobility and Environmental Stan- Public Goods.” Journal of Public Economics dards: Is There a Theoretical Basis for a Race 64 (1997): 202–17. to the Bottom?” In Harmonization and Fair Wilson, John D., and Roger H. Gordon. Trade, Vol. 1, edited by Jagdish Bhagwati and “Expenditure Competition.” Department of Robert Hudec, 395–427. Cambridge, MA: Economics, Michigan State University. MIT Press, 1996a. Mimeo, 1998. Wilson, John D. Zodrow, George R., and Peter Mieszkowski. “The Tax Treatment of Imperfectly Mobile “Pigou, Tiebout, Property Taxation, and the Firms: Rent-Seeking, Rent-Protection, and Underprovision of Local Public Goods.” Rent-Destruction.” In The Political Economy Journal of Urban Economics 19 (1986): 356– of Trade Policy: Papers in Honor of Jagdish 70.

304