Journal of Economic Literature Vol. XXXVII (September 1999) pp. 1120–1149 Oates: An Essay on Fiscal Federalism An Essay on Fiscal Federalism WALLACE E. O ATES1 1. Introduction planning that, in the view of many, has failed to bring these nations onto a path F ISCAL DECENTRALIZATION is in vogue. Both in the industrialized and in the developing world, nations are of self-sustaining growth. But the proper goal of restructuring the public sector cannot simply be de- turning to devolution to improve the per- centralization. The public sector in formance of their public sectors. In the nearly all countries consists of several United States, the central government different levels. The basic issue is one has turned back significant portions of of aligning responsibilities and fiscal in- federal authority to the states for a wide struments with the proper levels of gov- range of major programs, including wel- ernment. As Alexis de Toqueville ob- fare, Medicaid, legal services, housing, served more than a century ago, “The and job training. The hope is that state federal system was created with the in- and local governments, being closer to tention of combining the different ad- the people, will be more responsive to vantages which result from the magni- the particular preferences of their con- tude and the littleness of nations” (1980, stituencies and will be able to find new v. I, p. 163). But to realize these “dif- and better ways to provide these ser- ferent advantages,” we need to under- vices. In the United Kingdom, both Scot- stand which functions and instruments land and Wales have opted under the are best centralized and which are best Blair government for their own regional placed in the sphere of decentralized parliaments. And in Italy the movement levels of government. This is the sub- toward decentralization has gone so far ject matter of fiscal federalism. As a as to encompass a serious proposal for subfield of public finance, fiscal feder- the separation of the nation into two in- alism addresses the vertical structure of dependent countries. In the developing the public sector. It explores, both in world, we likewise see widespread inter- normative and positive terms, the roles est in fiscal decentralization with the ob- of the different levels of government jective of breaking the grip of central and the ways in which they relate to one 1 Professor of Economics, University of Mary- another through such instruments as land, and University Fellow, Resources for the Fu- intergovernmental grants. 2 ture. I am most grateful for a host of helpful com- ments on an earlier draft from Robert Inman, Ronald McKinnon, Daniel Rubinfeld, Robert 2 This economic use of the term “federalism” is Schwab, John Wallis, Barry Weingast, and three somewhat different from its standard use in politi- anonymous referees; for research assistance from cal science, where it refers to a political system Tugrul Gurgur; and for the splendid editorial with a constitution that guarantees some range guidance of John Pencavel and John McMillan. of autonomy and power to both central and 1120 Oates: An Essay on Fiscal Federalism 1121 My purpose in this essay is not to should have the basic responsibility for provide a comprehensive survey of fis- the macroeconomic stabilization func- cal federalism. I begin with a brief re- tion and for income redistribution in view and some reflections on the tradi- the form of assistance to the poor. In tional theory of fiscal federalism: the both cases, the basic argument stems assignment of functions to levels of gov- from some fundamental constraints on ernment, the welfare gains from fiscal lower level governments. In the ab- decentralization, and the use of fiscal sence of monetary and exchange-rate instruments. I then turn to some of the prerogatives and with highly open new directions in recent work in the economies that cannot contain much of field and explore a series of current top- the expansionary impact of fiscal stim- ics: laboratory federalism, interjurisdic- uli, provincial, state, and local govern- tional competition and environmental ments simply have very limited means federalism, the political economy of fis- for traditional macroeconomic control cal federalism, market-preserving feder- of their economies. Similarly, the mo- alism, and fiscal decentralization in the bility of economic units can seriously developing and transitional economies. constrain attempts to redistribute in- Some of this research is expanding the come. An aggressive local program for scope of the traditional analyses in im- the support of low-income households, portant and interesting ways. This will for example, is likely to induce an influx provide an opportunity both to com- of the poor and encourage an exodus of ment on this new work and to suggest those with higher income who must some potentially fruitful avenues for bear the tax burden. 3 In addition to further research. these functions, the central government must provide certain “national” public 2. The Basic Theory of Fiscal goods (like national defense) that pro- Federalism: Some Comments vide services to the entire population of the country. The traditional theory of fiscal feder- Decentralized levels of government alism lays out a general normative have their raison d’etre in the provision framework for the assignment of func- of goods and services whose consump- tions to different levels of government tion is limited to their own jurisdic- and the appropriate fiscal instruments tions. By tailoring outputs of such goods for carrying out these functions (e.g., and services to the particular pre- Richard Musgrave 1959; Oates 1972). ferences and circumstances of their At the most general level, this theory contends that the central government 3 It is straightforward to show that a system of decentralized poor relief is characterized by a gar- den-variety externality that results in suboptimal decentralized levels of government. For an econo- levels of support for the poor. More specifically, mist, nearly all public sectors are more or less fed- increases in support payments in one jurisdiction eral in the sense of having different levels of gov- confer external benefits in the form of a reduced ernment that provide public services and have number of poor households elsewhere. On this, some scope for de facto decision-making authority see Charles Brown and Oates (1985). There is, (irrespective of the formal constitution). In retro- moreover, evidence for the U.S. that state-level spect, it seems to me that the choice of the term decisions on levels of welfare support are interde- “fiscal federalism” was probably an unfortunate pendent; Luz Amparo Saavedra (1998), among one, since it suggests a narrow concern with budg- others, finds that states have responded to de- etary matters. The subject of fiscal federalism, as I creases (increases) in benefit levels in other states suggest above, encompasses much more, namely by reducing (raising) their own benefits to welfare the whole range of issues relating to the vertical recipients. For an excellent survey of this whole structure of the public sector. issue, see Jan Brueckner (1998). 1122 Journal of Economic Literature Vol. XXXVII (September 1999) constituencies, decentralized provision say, “local” in its incidence. The spe- increases economic welfare above that cific pattern of goods and services pro- which results from the more uniform vided by different levels of government levels of such services that are likely will thus differ to some extent in time under national provision. The basic and place. 5 This is to be expected. point here is simply that the efficient Nonetheless, there remains much to be level of output of a “local” public good said for the basic principle of fiscal de- (i.e., that for which the sum of resi- centralization: the presumption that the dents’ marginal benefits equals mar- provision of public services should be ginal cost) is likely to vary across located at the lowest level of govern- jurisdictions as a result of both differ- ment encompassing, in a spatial sense, ences in preferences and cost differen- the relevant benefits and costs.6 tials. To maximize overall social welfare Let me offer three observations on thus requires that local outputs vary the general theory. First, the founda- accordingly. tions of the Decentralization Theorem These precepts, however, should be need some elaboration. The theorem is regarded more as general “guidelines” itself a straightforward normative propo- than firm “principles.” As has been sition that states simply that “ . . . in pointed out in the literature, there is the absence of cost-savings from the certainly some limited scope for decen- centralized provision of a [local public] tralized macroeconomic efforts (Ed- good and of interjurisdictional exter- ward Gramlich 1987) and for assistance nalities, the level of welfare will always to the poor. In particular, there is a be at least as high (and typically higher) theoretical case for some poor relief if Pareto-efficient levels of consumption at local levels (Mark Pauly 1973), and are provided in each jurisdiction than if the fact is that state and local govern- any single, uniform level of consump- ments undertake a significant amount of tion is maintained across all jurisdic- redistributive activity. 4 tions” (Oates 1972, p. 54). The theorem Moreover, this prescription is a quite thus establishes, on grounds of eco- general one. It does not offer a precise nomic efficiency, a presumption in fa- delineation of the specific goods and vor of the decentralized provision of services to be provided at each level of public goods with localized effects. government. And indeed the spatial pat- While the proposition may seem almost tern of consumption of certain goods trivially obvious, it is of some interest and services like education and health is both in terms of setting forth the condi- open to some debate. As a result, we tions needed for its validity and, with find in cross-country comparisons some 5 For two useful treatments of the assignment of divergence in just what is considered, specific public services to the appropriate level of government, see Anwar Shah (1994, ch. 1) and 4 However, Martin Feldstein and Marian Vail- Ronald McKinnon and Thomas Nechyba (1997). lant Wrobel (1998) present some recent evidence 6 In Europe, proponents of fiscal decentraliza- suggesting that state government attempts to re- tion refer to the “principle of subsidiarity.” The distribute income are largely unsuccessful. They precept here is that public policy and its imple- find that progressive state income taxes in the U.S. mentation should be assigned to the lowest level have had little impact on the net-of-tax relative of government with the capacity to achieve the ob- wage rates of skilled versus nonskilled workers. jectives. This principle has been formally adopted Their claim is that the mobility of workers across as part of the Maastrict Treaty for European state borders undoes efforts at redistribution—and Union. Its intellectual roots, interestingly, are does so very quickly. The result is no redistribu- found in twentieth-century Catholic social philoso- tion, only deadweight losses from inefficient phy. On this see Robert Inman and Daniel Rubin- locational decisions. feld (forthcoming). Oates: An Essay on Fiscal Federalism 1123 some further analysis, for providing suggests that the magnitude of these some insights into the determinants of gains depends both on the extent of the the magnitude of the welfare gains from heterogeneity in demands across juris- fiscal decentralization (Oates 1998). dictions and any interjurisdictional dif- But there is more to the story. The ferences in costs. In particular, we find presumption in favor of decentralized that the potential gains from decentrali- finance is established by simply assum- zation stemming from interjurisdic- ing that centralized provision will entail tional differences in demand vary in- a uniform level of output across all ju- versely with the price elasticity of risdictions. In a setting of perfect infor- demand. If the costs of provision are mation, it would obviously be possible the same across jurisdictions, but de- for a benevolent central planner to pre- mands differ, then the extent of the scribe the set of differentiated local welfare loss from a centrally imposed, outputs that maximizes overall social uniform level of output increases, other welfare; there would be no need for things equal, with the price inelasticity fiscal decentralization (although one of demand. 7 There is a large body of might wish to describe such an outcome econometric evidence that finds that as decentralized in spirit!). The re- the demand for local public goods is sponse to this observation has been two- typically highly price inelastic. This sug- fold. First, one can realistically intro- gests that the potential welfare gains duce some basic imperfections (or from decentralized finance may well be asymmetries) in information. More spe- quite large. 8 cifically, individual local governments Pursuing this point into the realm of are presumably much closer to the peo- positive economics, we might expect ple and geography of their respective the magnitude of the potential gains jurisdictions; they possess knowledge of from fiscal decentralization to have both local preferences and cost condi- some explanatory power. Where these tions that a central agency is unlikely to gains are large, we would expect to find have. And, second, there are typically that the public sector is more decentral- political pressures (or perhaps even ized. In exploring this issue some years constitutional constraints) that limit the ago, I found some (perhaps vague) evi- capacity of central governments to pro- dence in its support: in a sample of vide higher levels of public services in countries, the fiscal share of the central some jurisdictions than others. These government varied inversely with an constraints tend to require a certain de- gree of uniformity in central directives. 7 In tax analysis, we are accustomed to a quite There are thus important informational different result: the deadweight loss varies directly with the price elasticity of demand. Here it is just and political constraints that are likely the reverse, since the distortion takes place on the to prevent central programs from quantity, rather than the price, axis. But interest- generating an optimal pattern of local ingly, if the source of the difference in efficient local outputs is cost differentials, then the gains outputs. from fiscal decentralization bear the opposite rela- My second observation concerns the tionship to the case where their source is differ- magnitude of the welfare gains from fis- ences in levels of demand: these gains then vary directly with the price elasticity of demand (Oates cal decentralization. We can, in princi- 1998). ple, measure the gains from the decen- 8 For surveys of this econometric literature, see tralized provision of public goods Rubinfeld (1987) and Oates (1996a). For an at- tempt actually to measure the welfare gains from relative to a more uniform, centrally de- decentralization, see David Bradford and Oates termined level of output. The theory (1974); they find large gains. 1124 Journal of Economic Literature Vol. XXXVII (September 1999) index of “sectionalism,” a measure of there were absolutely nothing mobile— the extent to which people in geo- households, factors, or whatever—there graphical subareas of a country identify would still exist, in general, gains from “self-consciously and distinctively with decentralization. The point here is sim- that area” (Oates 1972, pp. 207–208). ply that even in the absence of mobility, More recently, Koleman Strumpf and the efficient level of output of a “local” Felix Oberholzer-Gee (1998), in a more public good, as determined by the Sam- sharply focused study of states and uelson condition that the sum of the counties in the United States, find that marginal rates of substitution equals the decision to allow counties a local marginal cost, will typically vary from option to legalize the consumption of one jurisdiction to another. To take one alcoholic beverages depends signifi- example, the efficient level of air quality cantly on a measure of the heterogene- in Los Angeles is surely much different ity in preferences across counties within from that in, say, Chicago. each state. There is, I think, some inter- This point is of importance, because esting work to be done in exploring the the Tiebout model is often viewed as a extent to which the potential gains from peculiarly U.S. construction. The rela- decentralization can explain the ob- tively footloose households that it envi- served variation in actual governmental sions, responding to such things as local structure and policies. 9 schools and taxes, seem to characterize Third, I sense a widespread impres- the U.S. much better than, say, most sion, suggested in some of the litera- European countries. As a result, ob- ture, that the gains from decentraliza- servers outside the U.S. tend to believe tion have their source in the famous that this strand of the theory of local Tiebout model (Charles Tiebout 1956). finance is of limited relevance in their In this model, highly mobile households settings. While there may well be some “vote with their feet”: they choose as a truth to this, it most emphatically does jurisdiction of residence that locality not follow that there are no longer any that provides the fiscal package best significant welfare gains from the suited to their tastes. In the limiting decentralized provision of public goods. case, the Tiebout solution does indeed generate a first-best outcome that mim- 3. Fiscal Instruments in ics the outcome in a competitive mar- a Federal System ket. But the gains from decentraliza- tion, although typically enhanced by To carry out their functions, the vari- such mobility, are by no means wholly ous levels of government require spe- dependent upon them.10 In fact, if cific fiscal instruments. On the revenue side, governments will typically have ac- 9 Another interesting case is the setting of fed- cess to tax and debt instruments. But in eral standards for safe drinking water. After man- a federal system there is a further dating a set of standards for the quality of drinking water to be met in all jurisdictions in the Safe Drink- method for allocating funds among the ing Water Act of 1974, the federal government has different levels of the public sector: in- backed off and now allows a range of exceptions in tergovernmental grants. One level of recognition of the large interjurisdictional differ- ences in per-capita costs of meeting the standards government may generate tax revenues (U.S. Congressional Budget Office 1997). in excess of its expenditures and then 10 In certain settings, mobility can itself be a transfer the surplus to another level of source of distorted outcomes. See, for example, the seminal paper by Frank Flatters, Vernon government to finance part of the lat- Henderson, and Peter Mieszkowski (1974). ter’s budget. I want to review and Oates: An Essay on Fiscal Federalism 1125 comment briefly on the use of these highly mobile economic units (be they fiscal instruments in a federal fiscal households, capital, or final goods). But system. this in itself is not correct. The real im- plication is that decentralized levels of 3.1 Taxation in a Federal System government should avoid nonbenefit The determination of the vertical taxes on mobile units. Or, more accu- structure of taxes is known in the litera- rately, the analysis shows that on effi- ture as the “tax-assignment problem” ciency grounds decentralized govern- (Charles McLure 1983). And the basic ments should tax mobile economic units issue here is the normative question: with benefit levies (Oates and Robert Which taxes are best suited for use at Schwab 1991; Oates 1996b). Such eco- the different levels of government? The nomic units, in short, should pay for the question is typically posed in a setting benefits that they receive from the pub- in which there exists a nation state with lic services that local governments a central government, where there is lit- provide to them. tle or no mobility across national bor- The most well-known case of this is ders; at decentralized levels, in con- the earlier-discussed Tiebout model in trast, economic agents, goods, and which local jurisdictions use benefit resources have significant mobility taxes that effectively communicate to across jurisdictional boundaries with households the cost of consuming dif- the extent of this mobility increasing at ferent levels of local public goods; this successively lower levels of government. results in an efficient pattern of con- “Local” government, for analytical pur- sumption of these goods. But this is poses, may sometimes be characterized true not only for households. If local as operating in a setting in which eco- governments provide local inputs that nomic units can move costlessly among increase the productivity of capital em- jurisdictions. ployed in their jurisdictions, then they The difference in the mobility of should levy benefit taxes on capital in taxed units at the central and decentral- order to provide the set of signals ized levels has important implications needed for the efficient deployment of for the design of the vertical structure capital across localities (Oates and of taxation. Taxes, as we know, can be Schwab 1991). In sum, efficiency re- the source of distortions in resource al- quires not only that decentralized juris- location, as buyers shift their purchases dictions refrain from nonbenefit taxa- away from taxed goods. In a spatial set- tion of mobile economic units, but that ting, such distortions take the form of they actively engage in benefit taxation locational inefficiencies, as taxed units where the public sector provides (or owners of taxed items) seek out ju- services to these units. risdictions where they can obtain rela- The public sector must for various tively favorable tax treatment. High reasons rely to a substantial extent on excise taxes in one jurisdiction, for ex- nonbenefit taxes. Redistributive pro- ample, may lead purchasers to bear un- grams that provide assistance to the productive travel costs in order to pur- poor, for example, simply transfer in- chase the taxed items in jurisdictions come. But, as noted earlier, such pro- with lower tax rates. grams are not well suited to use at de- Such examples can suggest the con- centralized levels of government, where clusion that decentralized levels of gov- the mobility of economic units across ernment should avoid the taxation of local boundaries can undermine the 1126 Journal of Economic Literature Vol. XXXVII (September 1999) workings of such programs. It is for this analysis, moreover, establishes a pre- reason that the literature suggests that sumption for the taxation of relatively nonbenefit taxes, to the extent they are immobile economic units. A particularly needed, are best employed by higher attractive tax base is unimproved land, levels of government. since a tax on a factor or good in per- But provincial, state, and local gov- fectly inelastic supply will not be the ernments do, in fact, make use of some source of any locational inefficiencies. such levies. 11 In a seminal treatment of Such taxes (and any associated benefits this issue making use of an optimal taxa- from spending programs) will simply be tion framework, Roger Gordon (1983) capitalized into local land values. Thus, has explored the ramifications of the fiscally hard-pressed city governments decentralized use of a wide range of have at their disposal a tax base that nonbenefit taxes. And Gordon finds sev- cannot escape them through mobility. eral forms of potential distortion that There is some evidence in this regard result from an individual jurisdiction’s that the city of Pittsburgh, which has ignoring the effects of its fiscal deci- used a graded property tax under which sions elsewhere in the system; these in- land is taxed at five times the rate on clude inefficiencies involving, for exam- structures, has experienced an expan- ple, the “exporting” of tax burdens, sion in building activity that might not external congestion effects, and impacts have been forthcoming in the presence on levels of revenues in other jurisdic- of a higher tax on mobile capital (Oates tions, as well as certain equity issues and Schwab 1997). associated with a generally regressive pattern of tax incidence. 12 3.2 Intergovernmental Grants and The analysis suggests, moreover, Revenue Sharing some guidelines for the use of such taxes. A reliance on resident-based taxes Intergovernmental grants constitute a rather than source-based taxes, for ex- distinctive and important policy instru- ample, can lessen tax-induced distor- ment in fiscal federalism that can serve tions by reducing the scope for tax-ex- a number of different functions. The lit- porting (Inman and Rubinfeld 1996; erature emphasizes three potential roles McKinnon and Nechyba 1997). 13 The for such grants: the internalization of spillover benefits to other jurisdictions, 11 There is a lively and important debate in the fiscal equalization across jurisdictions, local finance literature over whether or not local and an improved overall tax system. property taxation, as employed in the U.S., consti- tutes benefit taxation. Bruce Hamilton (1975, Grants can take either of two general 1976) and William Fischel (1992) make the case forms. They can be “conditional grants” that local property taxes combined with local zon- that place any of various kinds of re- ing ordinances produce what is effectively a sys- tem of benefit taxation. Peter Mieszkowski and strictions on their use by the recipient. George Zodrow (1989) take the opposite view. Or they can be “unconditional,” that is, 12 See Inman and Rubinfeld (1996) for an excel- lent restatement and extension of the Gordon analysis. David Wildasin (1998a) provides a valu- source-based taxes (or “origin taxes”) involve tax- able survey of the various implications of factor ing factors where they are employed and goods mobility both for economic efficiency and for the and services where they are purchased. Under redistributive impact of public policy. resident-based taxation, governments have much 13 Resident-based taxes (also called “destination- less capacity to export the incidence of their taxes based taxes”) are levies on factors of production onto economic units elsewhere. Source-based (such as land, labor, and capital) based on the taxes, however, are often easier to administer and, owner’s residence and on goods and services based in certain forms, tend to be more commonly used on the residence of the consumer. In contrast, by state and local governments. Oates: An Essay on Fiscal Federalism 1127 lump-sum transfers to be used in any a necessary feature of fiscal federalism way the recipient wishes. The theory (Dan Usher 1995; Robin Boadway prescribes that conditional grants in the 1996). Economists normally think of re- form of matching grants (under which distributive measures from rich to poor the grantor finances a specified share of as those that transfer income from high- the recipient’s expenditure) be em- to low-income individuals. Intergovern- ployed where the provision of local ser- mental equalizing transfers require a vices generates benefits for residents of somewhat different justification based other jurisdictions. The rationale here on social values. 16 In practice, such is simply the usual Pigouvian one for equalizing grants play a major role in subsidies that induce individuals (in this many countries: in the fiscal systems of case policy-makers or the electorate) to Australia, Canada, and Germany, for ex- incorporate spillover benefits into their ample, there are substantial transfers of decision-making calculus. The magni- income from wealthy provinces or states tude of the matching shares, in such in- to poorer ones. In the United States, in stances, should reflect the extent of the contrast, equalizing grants from the fed- spillovers. 14 eral to state governments have never In contrast, unconditional grants are amounted to much. Intergovernmental typically the appropriate vehicle for grants in the U.S. typically address spe- purposes of fiscal equalization. The pur- cific functions or programs, but usually pose of these grants is to channel funds do not accomplish much in the way of from relatively wealthy jurisdictions to fiscal equalization. At the levels of the poorer ones. Such transfers are often states, however, there are many such based on an equalization formula that programs under which states provide measures the “fiscal need” and “fiscal equalizing grants to local jurisdictions— capacity” of each province, state, or lo- notably school districts. cality. These formulae result in a dis- Fiscal equalization is a contentious is- proportionate share of the transfers go- sue from an efficiency perspective. ing to those jurisdictions with the Some observers see such grants as play- greatest fiscal need and the least fiscal ing an important role in allowing poorer capacity. 15 jurisdictions to compete effectively with Although widely used, equalizing in- fiscally stronger ones. This view holds tergovernmental grants are by no means that, in the absence of such grants, fis- 14 Matching grants (possibly negative) can, in cally favored jurisdictions can exploit principle, also serve to correct some of the distor- their position to promote continued tions associated with the decentralized use of economic growth, some of which comes nonbenefit taxes (Gordon 1983). 15 Fiscal equalization can also make use of 16 The issue here is that from the perspective of matching grants. If the objective of the equaliza- redistributing income from rich to poor, equaliz- tion program is to equalize taxable capacity, the ing intergovernmental grants are bound to have granting government may choose to supplement some perverse effects. For such grants, although the revenue base of fiscally poorer jurisdictions by transfering income from wealthy to poor on aver- matching any revenues they collect by some speci- age, will inevitably result in some income transfers fied percentage. Such a measure has the potential from poor individuals who reside in wealthy juris- of allowing all jurisdictions to raise the same tax dictions to rich persons in generally poor areas. In revenues per capita for a given tax rate (irrespec- this sense, such equalizing measures are not as ef- tive of the actual size of their tax base). This form fective as programs that redistribute income from of fiscal equalization is sometimes called “power- rich to poor individuals. But a society may well equalization” and has gotten some attention in the wish, for other reasons, to provide additional sup- U.S. for state programs to achieve various equity port for the provision of local public services (such goals—most notably in the area of school finance as schools) in relatively low-income areas (e.g., (e.g., Feldstein 1976; and Nechyba 1996). Inman and Rubinfeld 1979). 1128 Journal of Economic Literature Vol. XXXVII (September 1999) at the expense of poorer ones. Fiscal centrally administered, nonbenefit taxes equalization, from this perspective, with a single rate applying to the na- helps to create a more level playing tional tax base will not generate the field for interjurisdictional competition.17 sorts of locational inefficiencies associ- But the case is not entirely persua- ated with varying rates across decentral- sive. Others have argued that fiscal ized jurisdictions. Moreover, central equalization can stand in the way of taxes can be more progressive, again needed regional adjustments that pro- without establishing fiscal incentives for mote development in poorer regions. relocation. There is, in fact, consider- McKinnon (1997a), for example, con- able evidence to indicate that state and tends that in the United States, the eco- local systems of taxes are typically more nomic resurgence of the South follow- regressive than central taxation (e.g., ing World War II resulted from Howard Chernick 1992). There is thus relatively low levels of wages and other some force in an argument for “revenue costs. It was this attraction of low wages sharing” under which the central gov- and costs that ultimately induced eco- ernment effectively serves as a tax-col- nomic movement to the South, bringing lecting agent for decentralized levels of with it a new prosperity. Fiscal equali- government. 18 The central government zation, from this perspective, may actu- then transfers funds, in a presumably ally hold back the development of unconditional form, to provinces, states, poorer areas by impeding the needed and/or localities. It is certainly possible, interregional flow of resources (both where the polity wishes, to build equal- emigration and immigration) in response izing elements into these transfers. to cost differentials. While there is here a real case for the But the primary justification for fiscal use of intergovernmental grants, a most equalization must be on equity grounds. important qualification is that such a And it is as a redistributive issue that it system of grants must not be too large continues to occupy a central place on in the sense of undermining fiscal disci- the political stage. In some cases, as in pline at lower levels of government Canada, it may provide the glue neces- (more on this later). sary to hold the federation together. In The prescriptive theory of intergov- other instances, like Italy, it may be- ernmental grants thus leads to a vision come a divisive force, where regions, of a system in which there exists a set of weary of large and longstanding trans- open-ended matching grants, where the fers of funds to poorer areas, actually matching rates reflect the extent of seek a dissolution of the union. Fiscal benefit spillovers across jurisdictional equalization is a complex economic and boundaries, and a set of unconditional political issue. grants for revenue sharing and, per- The third potential role for intergov- haps, equalization purposes. Such a ernmental grants is to sustain a more conception has, however, only modest equitable and efficient overall tax sys- tem. For reasons we have discussed, 18 This argument has even more force where, as in some developing countries and emerging democracies, provincial and local governments 17 As Boadway and Flatters (1982) have shown, simply lack the capacity for effective tax admini- equalizing grants may be required to offset distort- stration. In this setting, central transfers and/or ing locational incentives where some jurisdictions the piggybacking of supplementary rates on top offer pecuniary fiscal advantages to potential resi- of centrally administered taxes may be the dents resulting, for example, from large, taxable only realistic options. See, for example, Inman natural resource endowments. (forthcoming). Oates: An Essay on Fiscal Federalism 1129 explanatory power. We do, in fact, find ward to show that a lump-sum grant to federal matching programs that have a group of people is fully equivalent in supported a number of state and local all its effects, both allocative and distrib- activities with spillover effects, includ- utive, to a set of grants directly to the ing, for example, grants for interstate individuals in the group. Moreover, this highway construction. However, on result applies to an important class of closer examination, important anoma- collective-choice procedures, encom- lies appear. These grants are often passing several of the major models em- closed, rather than open, ended. They ployed in the public-finance literature. thus do not provide incentives for ex- These theorems, known as the “veil hy- pansion at the margin. Moreover, the pothesis,” thus imply that a grant to a federal matching shares are typically community is fully equivalent to a cen- much larger than justifiable by any tral tax rebate to the individuals in the plausible level of spillover benefits. community; intergovernmental grants, More generally, in a careful study of the according to this view, are simply a intergovernmental grant system, Inman “veil” for a federal tax cut. (1988) concludes that the economic the- The difficulty is that this hypothesis ory of intergovernmental grants does has not fared well in empirical testing. not provide a very satisfactory explana- It implies that the budgetary response tion of the structure of U.S. grant pro- to an intergovernmental transfer should grams; he finds that a political model be (roughly) the same as the response can do a much better job of explaining to an equal increase in private income U.S. grant programs. 19 in the community. But empirical studies Some years ago, David Bradford and of the response to grants have rejected I (1971a,b) tried to lay the foundations this equivalence time and again. Such for a positive theory of the response to studies invariably find that state and lo- intergovernmental grants by setting cal government spending is much more forth a framework in which the budget- responsive to increases in intergovern- ary decisions of the recipients of such mental receipts than it is to increases in grants are treated explicitly in a collec- the community’s private income. And tive-choice setting. In short, we treated this has come to be known as the “flypa- these grants, not as grants to an individ- per effect”—money sticks where it hits. ual decision-maker, but rather as grants While this finding may not be all that to polities that make budgetary deci- surprising, it is not so easy to reconcile sions by some collective algorithm (such with models of rational choice, for it as simple majority rule). This exercise suggests that the same budget con- produced some intriguing equivalence straint gives rise to different choices de- theorems. For example, it is straightfor- pending on what form the increment to the budget takes. There is now a large 19 As Inman and Rubinfeld (1996) point out, the literature that tries in a variety of ways prescriptive theory of grants presumes a central planner or political process that “will select so- (some quite ingenious) to explain the cially preferred policies” (p. 325). However, the flypaper effect. 20 James Hines and public-choice literature makes clear the potential Richard Thaler (1995) have suggested of central-government political mechanisms to make inefficient choices concerning policies that recently that this is just one of a more affect various groups differently. In addition, a general class of cases where having grant-distributing agency may have its own objec- tives; for an excellent study of how such objectives 20 For surveys and interpretations of this litera- can influence the pattern of grants, see Chernick ture, see Gramlich (1977), Ronald Fisher (1982), (1979). Oates (1994), and Hines and Thaler (1995). 1130 Journal of Economic Literature Vol. XXXVII (September 1999) money on hand (e.g., from grants) has a and troublesome issues of measurement much different effect on spending be- and interpretation in the Stine study. havior than where the money must be Subsequently, using national aggregate raised (e.g., by taxation). data on the state and local government Much of the early empirical work on sector, Shama Gamkhar and I (1996) the expenditure response to intergov- were unable to reject the hypothesis ernmental grants studied the period that the expenditure response to in- from the 1950’s through the 1970’s, creases and decreases in intergovern- when these grants exhibited a continu- mental grants has the same absolute ing path of expansion. As a result, much value per dollar of grants. Our findings of the interest focused on the budgetary are thus consistent with the proposition response to increases in grants. How- that the flypaper effect operates sym- ever, in more recent times, efforts at metrically in both directions. But much fiscal retrenchment and devolution have clearly remains to be done on this issue. led to large cuts in a wide range of fed- eral grant programs. And this has raised 4. A Note on Jurisdictional Boundaries the interesting and important question of whether the response to cuts in The treatment to this point has im- grants is similar in sign and magnitude plicitly taken as given a pattern of to the response to increases in these boundaries that divide the nation-state grants. Gramlich (1987), for example, into a set of jurisdictions for decentral- observed that during this period of re- ized governance. The existence and trenchment, state and local govern- magnitude of spillover effects from lo- ments responded to the cutbacks in calized public policies clearly depend grants by picking up much of the slack: on the geographical extent of the rele- they increased their own taxes and re- vant jurisdiction. One way to deal with placed in large part the lost grant funds such spillovers is to increase the size of so as to maintain levels of existing pro- the jurisdiction, thereby internalizing grams. If Gramlich is right, then we all the benefits and costs. The problem, should observe a basic asymmetry in re- of course, is that such an extension may sponse: the spending of recipients involve welfare losses from the reduced should be more responsive to increases capacity to differentiate local outputs. in grant monies than to decreases in There is clearly some kind of tradeoff these revenues. This issue is of some here between internalizing spillover importance if we are to understand the benefits (and costs) and allowing local budgetary implications of the ongoing differentiation. process of fiscal decentralization. In the In practice, much of the problem first study of this issue, William Stine stems from a set of existing boundaries (1994), examining the response of that are largely historically and cultur- county governments in Pennsylvania, ally determined and that may make lit- found just the opposite of Gramlich’s tle sense in terms of the economic and prediction: his estimates imply that geographical realities. Consider, for ex- these county governments not only ample, the United States. Suppose that failed to replace lost grant revenues, we were to begin with a tabula rasa, a but that they reduced their spending completely undefined set of boundaries from own-revenues on these programs for states and localities. And we set for as well, giving rise to a “super-flypaper ourselves the task of laying out both a effect.” There are, however, some tricky rational set of levels of government and Oates: An Essay on Fiscal Federalism 1131 borders for the jurisdictions at each pollutants that travel long distances level of government. One thing seems across the midwestern and northeastern clear: such a system of jurisdictions parts of the United States has led, un- would bear little resemblance to our ex- der congressional legislation in 1990, to isting map. The states, in particular, are the formation of an Ozone Transport quite poorly designed to deal with the Region (OTR) for the coordination of provision of certain important public efforts to manage air quality in eleven goods, notably environmental resources. eastern states and the District of Co- To take one example, rivers were used lumbia. Such regional organizations can historically (for understandable rea- be seen as the outcome of a kind of sons) to mark off one state from an- Coasian process in which interjurisdic- other. But from the perspective of ef- tional externalities are addressed through fective management of a public good, negotiation and coordinated decision- this is the worst sort of border. It means making. The history of such enterprises, that two independent and autonomous however, attests to their difficulty. The jurisdictions are making decisions that fascinating study by Bruce Ackerman et affect the public good whose output al. (1974), for example, of the attempt they jointly share. It seems clear that it to create a “model regional agency” in would make much more sense to place the form of the Delaware River Basin such resources within a single jurisdic- Commission reveals all the complexities tion. My own surmise is that a much and perverse incentives that can bedevil more rational map would probably en- such joint enterprises. Nevertheless, tail (1) some fairly sizeable regional such coordination does, in principle, of- governments that extend over water- fer an important avenue for addressing sheds, air sheds, and other environ- such interjurisdictional concerns. mental resources; (2) metropolitan gov- ernments that encompass center cities 5. Laboratory Federalism and and the suburbs that house many city Welfare Reform workers; and (3) smaller local govern- ments that allow groups of residents to It seems ironical in the light of the determine services of relevance mainly preceding treatment of principles (or to themselves. guidelines) for fiscal federalism to find But political realities being what they that welfare reform is in the vanguard are, we can expect to continue our col- of U.S. moves toward fiscal decentrali- lective life with much the same map in zation. The analysis suggests that the place. There does, however, remain threat of mobility of both low and high some flexibility in terms of creating use- income households will result in decen- ful compacts or associations of jurisdic- tralized policies that provide too little tions to deal with particular issues. The assistance to the poor (sometimes de- management of the Chesapeake Bay, scribed as a “race to the bottom”). Nev- for example, is in important organiza- ertheless, the decision has been made tional ways now the joint enterprise of to shift the primary responsibility for the relevant states (Delaware, Mary- poor relief back to the states. Under land, Pennsylvania, and Virginia), and measures signed into law in 1996, the Washington, D.C., with an important role federal government has replaced the also played by the federal government. longstanding federal entitlement pro- Likewise, the recognition that the man- grams, which came with both detailed agement of ground-level ozone involves rules and generous matching grants to 1132 Journal of Economic Literature Vol. XXXVII (September 1999) the states, by a system of block grants . . . It is one of the happy incidents of the with few strings attached. The states federal system that a single courageous State may, if its citizens choose, serve as a labora- now have broad scope to determine tory; and try novel social and economic ex- both the form and levels of assistance periments without risk to the rest of the under their programs to assist poor country. (Osborne 1988) households. 21 It is my sense that this is the primary How are we to understand this re- thrust behind the current welfare re- form? Does it represent an outright re- form. There exists much disappointment jection of the economic principles of and dissatisfaction with the operation and fiscal federalism? My answer is a quali- results under the traditional federal wel- fied no. There exists widespread recog- fare programs. But we really don’t have a nition of, and concern with, the likely clear sense of how to restructure them to shortcomings of a decentralized system achieve our societal goals of providing of poor relief. Policy makers are well needed relief and, at the same time, es- aware of the threat of strategic cuts in tablishing an effective set of incentives state levels of welfare support. But, as I to move people off welfare and into jobs. read it, we have decided to live with The recent legislation that transfers the this threat in order to seek out superior responsibility for these programs back to policy alternatives. And this brings us to the states represents, I believe, a recog- another dimension of fiscal federalism: nition of the failure of existing programs laboratory federalism. and an attempt to make use of the states In a setting of imperfect information as “laboratories” to try to find out what with learning-by-doing, there are poten- sorts of programs can work.22 tial gains from experimentation with a There are, in fact, a number of im- variety of policies for addressing social portant and intriguing examples of poli- and economic problems. And a federal cies whose advent was at the state or system may offer some real opportuni- local level and that later became fix- ties for encouraging such experimenta- tures of federal policy. Unemployment tion and thereby promoting “technical insurance, for example, was a state-level progress” in public policy. This point policy before the federal government was made long ago by James Bryce made it effectively mandatory on a na- (1888) who, in his insightful study of tional scale in the 1930s. More recently, the U.S. system of government, ob- in the area of environmental policy, the served that “Federalism enables a peo- experience in a number of states with ple to try experiments which could not their own forms of Emissions Trading safely be tried in a large centralized was an important prelude to the adop- country” (Vol. I, p. 353). Better known tion, in the 1990 Clean Air Act Amend- is a later statement by Justice Louis ments, of a national trading program in Brandeis, who wrote in 1932 that sulfur allowances to address the prob- There must be power in the States and the lem of acid rain. Without this experi- Nation to remould, through experimentation, ence in a number of states, I seriously our economic practices and institutions to meet changing social and economic needs doubt that policy-makers would have been willing to introduce such a new 21 For an excellent and recent review of this and unfamiliar policy measure as trade- whole debate in a historical context, see Therese able emissions rights on a national McGuire (1997). Rebecca Blank (1997) provides a concise and insightful treatment of the new 22 For a concurring view, see Craig Volden welfare legislation and its potential implications. (1997). Oates: An Essay on Fiscal Federalism 1133 scale. More generally, since the dawn of sarily need states as the “laboratories” the nation, programs successfully devel- for experiments. At the same time, one oped at the state level have often pro- might suspect that relatively indepen- vided models for subsequent federal dent efforts in a large number of states programs. will generate a wider variety of ap- States, of course, may learn from oth- proaches to public policy than a set of ers so that the diffusion of successful centrally designed experiments. policy innovations may be horizontal as A basic problem here is that there well as vertical. Both forms of diffusion has been little in the way of a real the- have been the subject of study by a ory of laboratory federalism to organize number of political scientists. Virginia our thought and to guide empirical Gray (1973) and Everett Rogers (1983), studies. However, the beginnings of for example, have found that the cumu- some theory are emerging, and they are lative distribution of states by date of quite illuminating. Susan Rose-Acker- adoption takes the S-curve shape, famil- man (1980) and, more recently, iar from the study of the spread of Strumpf (1997) have taken two quite other forms of innovation. Others, like different formal approaches to policy Jack Walker (1969), James Lutz (1987), innovation in a federal system. One in- David Huff et al. (1988), and David sight emerging from their analyses is an Nice (1994), have explored the geo- important, if familiar and unsurprising, graphical and other determinants of the one. There exists a basic “information pattern of adoptions by states. Empiri- externality” in that states that adopt cal studies of vertical diffusion are less new and experimental policies generate numerous. Thomas Anton (1989), Keith valuable information for others. And Boeckelman (1992), and Michael Sparer this creates a standard sort of incentive and Lawrence Brown (1996) have exam- for free-riding. From this perspective, ined the extent to which federal mea- we might expect too little experimenta- sures draw on the experience of the tion and policy innovation in a highly states. Some of this literature is rela- decentralized public sector. Indeed, as tively skeptical of the link. Sparer and Strumpf shows, it is unclear whether a Brown, for example, argue that (at least centralized or decentralized outcome for health care) “These laboratory adop- will result in more policy innovation. 23 tions and adaptations are probably more The underprovision of experimenta- the exception than the rule” (p. 196). tion at state and local levels can be ad- What are we to make of all this? A dressed through a system of subsidies to little reflection suggests first that there encourage these activities. And this is nothing in principle to prevent the raises another point regarding existing central government from undertaking welfare reform in the U.S. Under ear- limited experiments without commit- lier programs, federal aid took a match- ting the nation to an untested and risky ing form such that the federal govern- policy measure. Indeed, there have been ment effectively shared the costs and a number of such social experiments risks of new state-level programs. But with, for example, income-maintenance and housing-allowance programs that 23 The Rose-Ackerman and Strumpf analyses, have generated valuable information incidentally, also produce a number of subtle and about how programs work and the re- more surprising results. Strumpf finds, for exam- ple, that a state with a higher expected return sponse of participants to various values from experimentation can have a lower propensity of the key parameters. We don’t neces- to experiment. 1134 Journal of Economic Literature Vol. XXXVII (September 1999) under the new welfare reform mea- recognized. It manifests itself clearly on sures, matching aid has been replaced both sides of the Atlantic. We see it in by block grants. This in itself serves to Europe under the nomenclature of the reduce incentives for experimentation. “principle of subsidiarity,” where it is There are some conflicting incentives explicitly enshrined in the Maastrict here. On the one hand, the new legis- Treaty as a fundamental principle for lation gives the states broader scope for European union. In the U.S., it often experimentation, but it places the full appears more informally as an aversion cost of any new measures on the state to the “one size fits all” approach. with no sharing from the center. The Somewhat paradoxically, however, net outcome on the amount of this view is the subject of a widespread experimentation is thus a priori unclear. and fundamental challenge both at the More generally, we need a lot more theoretical and policy levels. The work on the implications of fiscal de- source of this challenge is the claim centralization for both the amount and that interjurisdictional competition kinds of policy experimentation and in- among decentralized levels of govern- novation. As I have suggested, there are ment introduces serious allocative dis- some clear and important cases where tortions. In their eagerness to promote innovation and experimentation at state economic development with the crea- and local levels have led to new policy tion of new jobs (so the argument goes), measures that have had broad national state and local officials tend to hold application. But it is much less clear how down tax rates and, consequently, out- we are to understand this experience in puts of public services so as to reduce terms of the overall effectiveness of a the costs for existing and prospective federal system in policy innovation. business enterprise. This results in a “race to the bottom” with suboptimal 6. Interjurisdictional Competition outputs of public services.24 and Environmental Federalism: This argument has a substantial his- A Challenge to the Basic View tory. Some thirty years ago, for exam- The preceding sections have set forth ple, George Break (1967) made the case an economic conception of a federal for the detrimental effects of interjuris- system. It is one in which the central dictional competition: government plays the major role in The trouble is that state and local govern- macroeconomic stabilization policies, ments have been engaged for some time in an takes the lead in redistributive mea- increasingly active competition among them- sures for support for the poor, and pro- selves for new business . . . In such an envi- vides a set of national public goods. De- ronment government officials do not lightly propose increases in their own tax rates that centralized levels of government focus go much beyond those prevailing in nearby their efforts on providing public goods states or in any area with similar natural at- whose consumption is limited primarily tractions for industry . . . Active tax competi- to their own constituencies. In this way, tion, in short, tends to produce either a gen- they can adapt outputs of such services erally low level of state-local tax effort or a state-local tax structure with strong regres- to the particular tastes, costs, and other sive elements. (Break 1967, pp. 23–24). circumstances that characterize their own jurisdictions. 24 Competition may also take place between dif- The general idea of decentralizing ferent levels of government. On such “vertical the provision of public services to the competition” (as well as horizontal competition), jurisdictions of concern has been widely see Albert Breton (1998). Oates: An Essay on Fiscal Federalism 1135 Fear of losing local business and jobs vided basic support for the centraliza- thus leads to suboptimal levels of state tion of environmental management in and local public goods. Such competition the United States. can involve regulatory as well as purely What I want to stress here is the fun- fiscal policies. John Cumberland (1979, damental character of this challenge to 1981) has extended the Break argument the basic model of fiscal federalism. to encompass the setting of standards for The claim is that the decentralized pro- local environmental quality. In the Break vision of public services is basically spirit, Cumberland contends that state flawed; in the words of one recent and local governments engage in “de- U.S. observer, we need centralization structive interregional competition.” In in order to “Save the States from order to attract new business and create Themselves” (Peter Enrich 1996). 25 jobs, public officials compete by reduc- But is this claim in fact true? This ing local environmental standards to turns out to be a very complicated ques- lower the costs of pollution control for tion both in theoretical and empirical firms that locate within their borders. In terms. There is now a substantial this instance, interjurisdictional competi- theoretical literature that addresses this tion leads to excessive environmental issue. In one set of papers, my col- degradation. The implication of the league Robert Schwab and I have devel- Cumberland view is that national stan- oped a series of models that explore the dards for environmental quality are conditions under which horizontal com- needed to prevent the excessive levels of petition among governments is effi- pollution forthcoming under state and ciency-enhancing (Oates and Schwab local standard setting. 1988, 1991, 1996). It turns out that it is More recently, Alice Rivlin (1992) straightforward to develop an analogue has echoed these views in her “rethink- to perfect competition in the private ing of U.S. federalism.” Although advo- sector. In such a setting, governments cating an extensive devolution of pub- compete with one another for a mobile lic-sector responsibilities to state and capital stock that both generates in- local government, Rivlin sees it as al- come for local residents and provides a most axiomatic that competition among tax base for them—and such competi- the states results in inadequate levels of tion leads local officials to adopt effi- public services. Her remedy is a system cient levels of outputs of public goods of shared taxes under which the reve- and tax rates. In these models, the in- nues from a new national value-added visible hand works in much the same tax would be shared among the states. way as in the private sector to channel This, she argues, would free the states policy decisions in individual jurisdic- so that they would not have “to worry so tions into an efficient outcome from a much about losing businesses to neigh- national perspective. boring states with lower tax rates” These models, moreover, are quite (p. 142). rich in terms of the variety of policy in- This line of argument has proved struments. Public officials provide not quite powerful in the policy arena. There are strong forces for the “har- 25 There is, incidentally, a very extensive, inter- monization” of fiscal and environmental esting, and lively debate on this matter among le- measures in Europe that draw heavily gal scholars. Recent issues of the law journals are full of papers on interjurisdictional competition on this proposition. Likewise, the case and its consequences. See, for example, Richard for the “race to the bottom” has pro- Revesz (1992) and Daniel Esty (1996). 1136 Journal of Economic Literature Vol. XXXVII (September 1999) only outputs for local residents, but kind of baseline from which one can in- public inputs that enhance the produc- troduce a range of quite plausible and tivity of locally employed capital, and realistic modifications that can be the environmental regulations that impose source of allocative distortions. A large costs on local business and improve lo- number of papers explore outcomes cal environmental quality. They finance either where jurisdictions are suffi- these public outputs with a set of taxes ciently large to have some influence on local residents and capital. And over the price of capital or where local there is no race to the bottom here. In- governments are restricted in their ac- stead, jurisdictions find it in their own cess to policy instruments and must, for interest to charge benefit taxes that example, tax business and household lead to efficient decisions in both the capital at the same rate. Many of these public and private sectors. 26 papers employ game-theoretic ap- The problem is that these models proaches in which there is strategic in- make some strong assumptions. Let me teraction among the jurisdictions note three of them here: jurisdictions (Wildasin 1988). In such settings, we behave as price-takers in national or in- find that outcomes can easily occur that ternational capital markets; public offi- involve suboptimal levels of public cials seek in their decisions to maximize outputs. 27 the welfare of their constituencies; and The theoretical literature thus gener- these officials have access to the ates some diverse findings on this issue. needed fiscal and regulatory policy in- There seem to be some basic efficiency- struments to carry out their programs enhancing aspects of interjurisdictional efficiently. It is not hard to show (or competition, but there are clearly a surprising to find) that violations of any range of “imperfections” that can be the of these conditions can lead to distorted source of allocative distortions. The real outcomes. Suppose, for example, that issue here is the magnitude of these dis- local policy makers are Niskanen-type tortions. Are we dealing with minor de- agents that seek to maximize, not the viations from efficient outcomes—or well-being of their constituencies, but does such competition produce major rather the size of the local public welfare losses? The pure theory can’t budget. It is then straightforward to help us much in answering this ques- show that they will set excessively lax tion. Moreover, some of the terminol- environmental standards in order to en- ogy is not very helpful. In particular, courage a larger inflow of capital so as the description of interjurisdictional to enlarge the local tax base (Oates and competition as involving a “race to the Schwab 1988). bottom” seems quite misleading. Such The Oates-Schwab models provide a a descriptive image may well be an effective rhetorical device: it conjures 26 I should emphasize here that all public out- up a vision of one jurisdiction cutting puts (including environmental quality) are entirely its tax rates and lowering its environ- local in these models; there are no spillover effects mental standards, only to be outdone into other jurisdictions. The analysis, incidentally, by a neighboring jurisdiction, in a pro- extends not only to fiscal instruments, but regula- tory ones as well (such as environmental stan- cess that leads to a downward spiral to dards). The analysis of “regulatory federalism” the “bottom” (suggesting a very bad is, in principle, analogous to that of fiscal federal- ism. The same general principles concerning decentralization apply to fiscal and regulatory 27 See John Wilson (1996) for an excellent instruments. survey of this literature. Oates: An Essay on Fiscal Federalism 1137 outcome indeed). However, the models 7. Fiscal Federalism: Expanding the that generate these results are nothing Scope of the Analysis of the sort. They are often game-theo- The normative framework for most of retic models that produce Nash equilib- the literature in fiscal federalism (and ria with suboptimal public outputs as for my treatment in this essay as well) the outcome. What matters here is the consists of the traditional principles of extent of the suboptimality. And the welfare economics. From this perspec- race-to-the-bottom terminology tends tive, institutions are evaluated in terms to obscure this issue. of their impact on efficiency in resource Unfortunately, we do not have many allocation and the distribution of in- empirical studies to bring to bear on come. However, the choice of a system this matter. There is a substantial de- of governance involves other values as scriptive literature addressing economic well: the extent of political participa- competition among state and local gov- tion, the protection of individual rights, ernments in the U.S., with some inter- and the development of various civic esting findings (Timothy Bartik 1991). virtues. Political theorists throughout But this body of work really does not the ages have explored the ways in shed much light on the normative ques- which different political systems ad- tion of whether such competition is ef- dress these various objectives of the ficiency-enhancing or not (Paul Courant polity. In addition, the vertical struc- 1994). In an interesting study that is of ture of government may have important relevance, Anne Case, James Hines, and implications for the way in which the Harvey Rosen (1993) find evidence of public sector functions and its impact strategic interaction in state-level fiscal on the operation of a system of markets. policies. Using a similar methodology, In this section, I want to explore some Jan Brueckner (1998) finds empirical of the new (and older) literature that support for policy interdependence in addresses some broader implications of the adoption of growth-control mea- fiscal federalism. sures by local governments in Califor- nia. But at this juncture, I think it is 7.1 Economic and Political Objectives fair to say that the jury is still out on in a Federal System this matter. The welfare implications of interjurisdictional competition remain The first issue involves extending the the subject of a lively ongoing debate conceptual horizon to encompass addi- with a real need for further empirical tional political objectives. What might work to supplement the large theo- this add to our more narrowly focused retical literature. In my own view, the economic view of fiscal federalism? In- existing work is not sufficient to make a man and Rubinfeld, in one strand of compelling case for the abandonment of their important new work on fiscal fed- (or basic amendment to) the principle eralism, have (and are) exploring this of fiscal decentralization. The case re- issue in an attempt to redefine and mains strong, it seems to me, for leav- extend the analytical framework to en- ing “local matters in local hands.” compass some of these additional politi- Moreover, as we shall see shortly, there cal and constitutional dimensions of is another literature that takes a very public-sector structure. different (and unambiguously positive) The approach of Inman and Rubinfeld view of the role of interjurisdictional (1997a,b,c) incorporates explicitly cer- competition. tain political goals into a more extended 1138 Journal of Economic Literature Vol. XXXVII (September 1999) objective function. In such a setting, we teresting that the Court has seen fit to find ourselves examining tradeoffs be- set aside, in certain instances, the pre- tween such goals as economic efficiency sumed economic consequences of cer- and political participation. In one such tain state regulations in favor of decen- illustration, they present a “federalism tralized political choices, so long as they frontier” in which (over the relevant “were decided by an open, participatory range) increased political participation political process, as evidenced by state comes at the expense of economic legislative involvement” (1997a, p. 1252). efficiency (1997a, p. 1230). It seems unlikely that we can ever The basic presumption here is that hope to quantify such tradeoffs in a for- more decentralized political systems are mally satisfying way. But the Inman-Ru- conducive to increased citizen impact binfeld work does suggest that careful on political outcomes and political par- analysis can certainly help to clarify the ticipation. The evidence on this issue, nature of the tradeoffs involved in the in truth, is somewhat mixed, but overall vertical design of the political system it suggests on balance “that both citizen and allow economics to play a broader influence and effort increase as the role in the debate. It is interesting, size of government declines” (1997a, moreover, that the political objectives p. 1215). The basic political objectives seem, on the whole, to strengthen the thus strengthen the case for increased case for fiscal decentralization. decentralization; they point to a system that is more decentralized than one 7.2 Public-Sector Institutions: chosen simply on the grounds of an Market-Preserving Federalism exercise in economic optimization. An alternative approach to federalism, While this is suggestive at a general related to the “new institutional econom- level, it raises the more difficult ques- ics,” sees political decentralization in tion of how one addresses these trade- terms of its capacity to sustain a pro- offs in the actual design of fiscal institu- ductive and growing market economy. tions. How, for example, can we define From this perspective, Barry Weingast and measure in a meaningful way the (1995), Ronald McKinnon (1997a), and marginal rate of substitution between their colleagues have explored the insti- economic efficiency and political par- tutional structure of a system that prom- ticipation and incorporate this into the ises to provide a stable framework for a design of a political system? To ap- market system (see also McKinnon and proach this question in a substantive Nechyba 1997 and Qian and Weingast way requires the study of more specific 1997). Weingast’s point of departure is issues. And here Inman and Rubinfeld a “fundamental political dilemma of an (1997a) provide a provocative beginning economic system,” namely that “a gov- with a careful study of “anti-trust state- ernment strong enough to protect prop- action doctrine.” This involves an intrigu- erty rights and enforce contracts is also ing series of Supreme Court decisions strong enough to confiscate the wealth in which state programs, that—had they of its citizens” (1995, p. 1). 28 been designed and introduced by pro- The attraction of federalism for ducers themselves, would have consti- Weingast is its potential for providing a tuted a violation of anti-trust laws—were 28 However, as Martin McGuire and Mancur Ol- upheld on the basis of state legislative son (1996) have shown, even a self-aggrandizing sovereignty. Although the history of this autocrat (if secure) has powerful incentives for doctrine is a complicated one, it is in- supporting an economically efficient system. Oates: An Essay on Fiscal Federalism 1139 political system that can support an effi- capital projects over their useful life. cient system of markets. In a provoca- But they have no recourse to public tive treatment, Weingast lays out a set sources for funding this debt; they op- of three conditions for a federal system erate in private credit markets just like that characterize what he calls “market- private borrowers. These markets them- preserving federalism.” These condi- selves, through the determination of tions require that (1) decentralized gov- credit ratings and other forms of moni- ernments have the primary regulatory toring fiscal performance, create an en- responsibility over the economy; (2) the vironment in which the fiscal authori- system constitutes a common market in ties must behave in responsible ways. 29 which there are no barriers to trade; These markets, by creating a hard and (3) decentralized governments face budget constraint in terms of debt fi- “hard budget constraints.” By this last nance, have imposed a very useful disci- condition, Weingast means that lower- pline on decentralized fiscal behavior. 30 level governments have neither the ca- More generally, a hard budget con- pacity to create money nor access to un- straint implies that decentralized gov- limited credit. And it implies further that ernments must place a basic reliance on the central government does not stand their own sources of revenues. They ready to bail them out in instances of must not be overly dependent on trans- fiscal distress. fers from above. I discussed in an ear- Weingast goes on to argue in histori- lier section the potential role for inter- cal terms that eighteenth century En- governmental grants, but Weingast and gland and the United States in the nine- McKinnon (as well as others) remind us teenth century were effectively such of the important discipline that stems systems of market-preserving federal- from self-financing. It is especially im- ism, and that this fostered in important portant that intergovernmental grants and fundamental ways the process of not be expansible in the sense that re- economic growth. It proved critical, ar- cipients can turn to the grant system to gues Weingast, to the industrial revolu- bail them out of fiscal difficulties tion in England and supported a system (Wildasin 1998b). In particular, public of “thriving markets” in the United States authorities need to fund their own throughout the nineteenth century. expenditures at the margin. 31 McKinnon (1997a) has explored in The institutional perspective reminds more detail the importance of Wein- us that there is more to the design of a gast’s last condition of a hard budget constraint. Crucial to this view is the 29 James Poterba and Kim Rueben (1997), for separation of monetary and fiscal pow- example, have found that those states with tighter anti-deficit rules, and more restrictive limitations ers. In a federal system, if the central on the authority of the state legislature to issue government controls the common cur- debt, pay lower rates of interest on their bonds. 30 McKinnon (1997b) has gone on to argue that rency, then lower-level governments much of the impetus for European Monetary will be limited to fiscal instruments and Union has as its source a collectively imposed will not have access to the “soft” option budgetary retrenchment. His interesting argument of monetized debt. As McKinnon points is that European decision makers, realizing that they cannot achieve fiscal stability with continued out, state and local governments in the access to monetary powers, are seeking through United States engage in extensive debt EMU to create the hard budget constraints that are finance for capital projects. This makes the prerequisite for responsible fiscal management. 31 This is subject to the qualification that good economic sense in terms of matching grants may be needed to internalize spreading the payments for long-lived interjurisdictional spillover benefits. 1140 Journal of Economic Literature Vol. XXXVII (September 1999) federal fiscal system than just the allo- The Brennan-Buchanan view suggests cation of functions to the appropriate the hypothesis that the overall size of levels of government. In addition, we the public sector “should be smaller, ce- need sets of formal and informal institu- teris paribus, the greater the extent to tions that embody the rights sorts of in- which taxes and expenditures are de- centives for public decision makers (Ol- centralized” (1980, p. 185). The evi- son 1990). These rules or procedures dence on this hypothesis is, however, at must make the costs of public programs best mixed. For example, I was unable as fully visible as their benefits in ways to find any systematic relationship be- that make public officials accountable tween public-sector size and the extent for their decisions (Shah 1998). of fiscal decentralization (Oates 1985). The treatment of fiscal structure in However, some later and more disag- this section is not unrelated to Geoffrey gregated studies have found some ten- Brennan and James Buchanan’s (1980) dencies of this kind (See Oates 1989 for view of fiscal decentralization as a a survey of this work.). mechanism for controlling the size of More generally, there is not much the public sector. Drawing by analogy evidence on the relationship between on the conventional theory of monopoly fiscal decentralization and economic in the private sector, they envision the performance. But there is some. Jeff government sector as a monolithic Huther and Anwar Shah (1996) at the agent, a “Leviathan,” that seeks its own World Bank have assembled a large and aggrandizement through maximizing diverse set of indices for eighty nations. the extraction of tax revenues from the These indices encompass a wide variety economy. From this perspective, the of measures of economic and political design of the constitution and associ- structure and performance: quality of ated institutions has as a major objec- governance, political freedom, political tive the placing of a set of constraints stability, debt-to-GNP ratios, measures that limits Leviathan’s access to tax and of income, the degree of equality in the other fiscal instruments. Fiscal decen- distribution of income, and many more. tralization can, in their view, play a In examining the statistical associations most important role in constraining among these various indices, they find public sector growth. Competition among in nearly every case a statistically sig- decentralized governments for mobile nificant and positive correlation be- economic units greatly limits the capac- tween increased decentralization and ity of Leviathan to channel resources improved performance (either in politi- into the public sector. As Brennan and cal or economic terms). There are obvi- Buchanan put it, competition among ous and important qualifications here. governments in the context of the “in- Such associations do not prove causa- terjurisdictional mobility of persons in tion. In particular, the degree of fiscal pursuit of ‘fiscal gains’ can offer partial decentralization is itself the outcome of or possibly complete substitutes for ex- a complex of political and economic plicit fiscal constraints on the taxing forces. Nonetheless, the initial results power” (1980, p. 184).32 are suggestive and invite further ex- ploration. Elsewhere, Sang-Loh Kim 32 In a more formal treatment of this matter, (1995) in an intriguing econometric Dennis Epple and Allan Zelenitz (1981) have study making use of an international shown that while competition among jurisdictions can constrain government rent-seeking behavior, panel data set, has estimated a Barro- it cannot altogether eliminate it. type growth model. In addition to the Oates: An Essay on Fiscal Federalism 1141 usual explanatory variables, he included that has no doubt occurred to the a measure of fiscal decentralization reader. In the earlier section on inter- that, in most of his estimated equations, jurisdictional competition, the central has a significant and positive partial as- concern was that such competition sociation with the rate of economic leads to too little in the way of public growth. Kim’s findings thus support outputs. There it was argued that com- Shah’s contention that fiscal decentrali- petition for new firms and jobs may lead zation enhances economic perfor- to public budgets that are too small, mance—in this case, more rapid eco- and to overly lax environmental stan- nomic growth. In contrast, Heng-fu Zou dards. In contrast, the thrust of this sec- and his colleagues have found a nega- tion has been on the beneficial effects tive relationship between economic of competition as a disciplining force growth and fiscal decentralization in that restrains the tendencies in the pub- two studies, one examining a sample of lic sector towards excessive spending forty-six countries over the period and other forms of fiscal misbehavior. 1970–89 (Davoodi and Zou 1998) and One’s view of the role of intergovern- the other a study of the growth of prov- mental competition clearly depends inces in China (Zhang and Zou 1998). on how one views the operation of the Much obviously remains to be done at public sector more generally! the empirical level in order to give us a better sense of the relationship of fis- 8. Fiscal Decentralization and cal decentralization to economic and Economic Development political performance. There is also much more to do at the When examining international cross- conceptual level. While Weingast’s in- sectional data on intergovernmental itial forays into market-preserving fed- structure, one is immediately struck by eralism are certainly provocative, they the sharp contrast in the extent of fiscal raise at least as many questions as they decentralization in the industrialized answer. It is fair, I think, to charac- and developing countries. In a study of terize the analysis as fairly “loose” at my own involving a group of forty-three this stage. For example, are Weingast’s countries (Oates 1985), the sample sta- conditions for market-preserving feder- tistics revealed an average share of cen- alism to be regarded as necessary or tral-government spending in total pub- sufficient (or both) for an effective po- lic expenditure of 65 percent in the litical foundation for a private market subsample of eighteen industrialized economy? Jonathan Rodden and Susan countries, as contrasted to 89 percent in Rose-Ackerman (1997) have raised a the subsample of twenty-five develop- number of probing questions concern- ing nations. In terms of total public ing the Weingast analysis. There is revenues, the central-government share clearly much to chew on here. The next for this same subsample of developing step, it seems to me, is to attempt to countries was over 90 percent! formalize these relationships more ex- Although there are real concerns with plicitly so as to get a better sense of the accuracy of some of these fiscal data how different political and budgetary (Richard Bird 1986), the general pre- institutions influence the functioning of sumption that the developing countries a market system. are characterized by relatively high de- Finally, it is impossible to leave this grees of fiscal centralization seems section without noting an obvious irony firmly grounded. And this, moreover, is 1142 Journal of Economic Literature Vol. XXXVII (September 1999) not something new. Writing over forty prove to be very helpful. If we look at years ago, Alison Martin and W. Arthur the United States, for example, we find Lewis (1956) noted that “the weakness that in the late nineteenth century the of local government in relation to cen- public sector was both very small and tral government is one of the most strik- highly decentralized. At the turn of the ing phenomena of under-developed century, the public sector accounted for countries” (p. 231). only about 8 percent of GNP in the What are we to make of this? Some U.S., while the central-government observers attribute the poor economic share of total public expenditure was performance of many of the developing around 30–35 percent. By 1955, the countries in large measure to the failure central-government share of public of central planning and make a strong spending had roughly doubled from case for the devolution of fiscal respon- one-third to two-thirds. 33 The fiscal sibilities. But the issue is clearly more records of other industrialized nations complicated than this. In particular, the like Great Britain reveal roughly similar question arises as to whether fiscal de- patterns. centralization is a cause or a result of The point is that the trend over this economic development. Roy Bahl and period of economic growth was not one Johannes Linn (1992), for example, ar- of increasing fiscal decentralization; it gue that as economies grow and mature, was just the reverse! It is worth noting, economic gains from fiscal decentraliza- however, that these centralizing tenden- tion emerge. As they put it, “Decen- cies seem to have played out around the tralization more likely comes with the middle of the century. For most of the achievement of a higher stage of eco- industrialized countries, fiscal centrali- nomic development” (p. 391); the zation ratios appear to have peaked in “threshold level of economic develop- the decade of the 1950’s, and since that ment” at which fiscal decentralization time, they have actually declined becomes attractive “appears to be quite slightly in most cases (Oates 1978; high” (p. 393). From this perspective, it Werner Pommerehne 1977). What typi- is economic development that comes cally seems to be taking place is a com- first; fiscal decentralization then fol- plicated process of intergovernmental lows. But not all would agree. More evolution. We see efforts at devolution generally, it seems to me, we must re- in a number of OECD countries accom- gard intergovernmental structure as panied, at the same time, by the emer- part of a larger political and economic gence of a new top layer of government system that both influences and is de- in the European Community. termined by the interplay of a variety of But all this may not have much rele- political and economic forces. It may vance for the developing nations. This well be that fiscal decentralization itself is because they have a very different has a real contribution to make to im- starting point for the growth process. As proved economic and political perfor- Diana Conyers (1990) stresses, “Most mance at different stages of development. less developed countries inherited rela- To gain further insight into this issue, tively centralized systems of govern- we might turn to the historical experi- ments from their colonial powers, and ence of the industrialized countries and in the first years of independence there examine the course of fiscal decentrali- 33 See John Wallis and Oates (1997) for a de- zation through extended periods of eco- scription and analysis of the evolution of American nomic growth. This, in fact, does not federalism in the twentieth century. Oates: An Essay on Fiscal Federalism 1143 was often a tendency to maintain—if improved growth performance depends.34 not strengthen—central control and The prescriptive literature on fiscal centralized systems of planning, in or- structure for the developing countries der to encourage a sense of national unity harks back directly to several of the and reinforce the new government and points made in the preceding sections. its policies” (p. 16). Thus, many of these In particular, there is a heavy emphasis countries entered upon nationhood with on reliance on own finance in order to highly centralized government sectors; create hard budget constraints. This can they have not undergone anything like have special relevance in the develop- the process of public-sector evolution ex- ing-country context, where decentral- perienced in the industrialized countries. ized governments often have very lim- The implication of all this is that the ited access to their own major sources potential of fiscal decentralization for of tax and other revenues and are heav- improving economic and political per- ily dependent on transfers from above. formance must be evaluated in terms of In some instances, provincial or state the specific circumstances that charac- governments may even have access to terize the current state of a developing the public banking system to absorb nation. There remains, in my view and their debt issues. This predictably leads that of some others (Shah 1994), a to large budgetary deficits and both strong case on traditional grounds for a fiscal and monetary instability. significant degree of decentralization in This literature makes reference to public-sector decision-making in the the problem of “vertical imbalance,” developing nations. This case, as we meaning a disparity between different have discussed, rests both on the poten- levels of government in their expendi- tial economic gains from adapting levels ture commitments and their access to of public outputs to specific regional or revenues. Although the concept suffers local conditions and on the political ap- from certain ambiguities, it does focus peal of increased participation in gover- attention on the important issue of the nance. The economic case has been widespread inadequacy of revenue made formally in purely static terms (as sources at decentralized levels of gov- noted earlier in the treatment of the ernment. The often heavy reliance of Decentralization Theorem), but it may provincial, state, and local governments well have some validity in a dynamic on transfers from above undercuts in- setting of economic growth. Develop- centives for responsible fiscal decision- ment policies that are sensitive to par- making; fiscal decisions become out- ticular regional or local needs for infra- comes of politically driven negotiations structure and even human capital are between central and “local” authorities, likely to be more effective in promoting not the result of weighing benefits and economic growth than are centrally de- costs of prospective public programs. termined policies that largely ignore The case for establishing adequate these geographical differences. There exists, incidentally, no formal theory of 34 Some observers, like Remy Prud’homme fiscal decentralization and economic (1997), argue that the case for fiscal decentraliza- tion has been much exaggerated. Prud’homme growth; it might be useful to set out claims that many of the premises of the fiscal fed- such a theory, for a framework that in- eralism vision are typically not satisfied in the corporates jurisdiction-specific invest- developing-country setting; decentralized govern- ment bodies, he argues, are frequently unrespon- ment programs might provide some in- sive to the needs of their constituencies and sights into the parameters on which manifest widespread corruption. 1144 Journal of Economic Literature Vol. XXXVII (September 1999) and effective tax systems at decentral- through all these dimensions of fiscal ized levels of government is one of the reform is the crucial attention to fiscal critical issues of fiscal federalism in the decision-making institutions and proce- developing world. And it is a truly chal- dures themselves to introduce mecha- lenging problem (Bahl and Linn 1992; nisms that provide incentives for public Bird 1992). The earlier section dealing officials to act in the public interest; this with the tax-assignment problem set means largely, as Shah (1998) stresses, forth some of the properties of “good” establishing channels for account- taxes at decentralized levels of govern- ability. 35 In the interim, provincial and ment. But provincial and local govern- local governments cannot be left to fend ments in developing countries often entirely for themselves; depending on the face serious obstacles to the use of specific circumstances, there will often these tax bases. The scope, for example, be a need for significant transfers from for using local property taxes is circum- the center, especially to impoverished scribed in many instances by the ab- jurisdictions. But the general direction sence of the requisite institutions for of needed reform seems clear. tax administration. As Bahl and Linn The ongoing efforts to decentralize (1992) point out, there is typically more the public sectors of former socialist potential for such taxes in urban than in states encounter much the same set of rural areas in most developing coun- issues. But the problems are in some tries. The obstacles are real, but there ways even more complicated, inasmuch are ongoing and extensive efforts to as the process of decentralization is go- build up the administrative capacity for ing on alongside a process of privatiza- more effective revenue systems. tion; the complicated and sometimes Fiscal reform efforts in the develop- chaotic transition from a command ing world thus must focus on (1) Re- economy to a market system does not structuring systems of intergovernmen- provide a stable environment within tal grants, in some instances to reduce which to restructure the public sector. the extent of financing that they pro- Nevertheless, a comprehensive process vide to decentralized levels of govern- of fiscal decentralization is underway in ment, and, more generally, to remove much of Central and Eastern Europe, the perverse incentives that they often and it involves the same issues of defin- embody for fiscal behavior on the part ing the fiscal responsibilities of the dif- of recipients; (2) Redesigning revenue ferent levels of government and intro- systems so as to provide decentralized ducing the fiscal instruments and levels of government a much expanded procedures needed both to support access to own-revenues to finance their emerging private markets and to deliver budgets and thereby reduce their de- needed public services (Bird, Ebel, and pendence on transfers from above; and Wallich 1995). (3) Reviewing the use and restrictions on debt finance to ensure that debt is- 9. Some Concluding Observations sues are not a ready way to finance defi- The evolution of the vertical struc- cits on the current account. All three of ture of the public sector continues in these avenues of reform contribute in important ways to the establishment of 35 See Govinda Rao (1998) for an illuminating a hard budget constraint, but one that treatment in the Indian context of the wide range of mechanisms (or “subterranean transfers” as he permits decentralized levels of govern- calls them) through which central government ment to do their job. Finally, running subsidizes the states. Oates: An Essay on Fiscal Federalism 1145 interesting and novel ways. As I noted sector. Recent decades have seen the earlier, the first half of the twentieth creation of special districts to provide century was characterized by a strong particular public services and the for- trend toward increased fiscal centraliza- mation of metropolitan area govern- tion. Indeed, some acute political ob- ments to bring center cities and their servers in the nineteenth century fore- suburbs into a single jurisdiction (again cast this trend. Tocqueville, writing in for purposes of addressing specific the first half of the nineteenth century, needs such as transportation and hous- predicted that “in the democratic ages ing). It is especially striking to witness which are opening upon us . . . cen- in the European Community the moves tralization will be the natural govern- toward devolution in many member ment” (1945, Vol. II, p. 313). And countries, while, at the same time, the nearer the end of the century, Lord Community develops a set of supra- Bryce reiterated this forecast (at least national institutions for governance and for the U.S.). After reviewing both the economic management. Other coun- “centrifugal” and “centripetal” forces at tries, like South Africa and the former work in American government, Bryce socialist states, are struggling with their concluded that while the centrifugal own sets of pressing issues in their at- forces were “likely, as far as we can see, tempts to find effective mechanisms for to prove transitory . . . the centripetal political and fiscal decentralization. forces are permanent and secular While the existing literature in fiscal forces, working from age to age” (1901, federalism can provide some general Vol. II, p. 844). Bryce then proceeded guidance on these issues, my sense is to forecast that “ . . . the importance of that most of us working in the field feel the States will decline as the majesty more than a little uneasy when proffer- and authority of the National govern- ing advice on many of the decisions that ment increase” (1901, Vol. II, p. 844). must be made on vertical fiscal and po- Later, Edward McWhinney (1965) went litical structure. We have much to on to generalize all this to what he learn! calls “Bryce’s Law,” the proposition that REFERENCES “ . . . federalism is simply a transitory Ackerman, Bruce et al. 1974. The Uncertain step on the way to governmental unity” Search for Environmental Quality. NY: Free (p. 105). Press. But such forecasts have not been Anton, Thomas J. 1989. American Federalism and Public Policy. Philadelphia: Temple U. Press. borne out. The second half of the twen- Bahl, Roy W. and Johannes F. Linn. 1992. Urban tieth century has seen the extent of cen- Public Finance in Developing Countries. Ox- tralization in most of the industrialized ford: Oxford U. Press. Bartik, Timothy J. 1991. Who Benefits from State countries reach some sort of peak with and Local Development Policies? Kalamazoo, a modest swing back in the direction of MI: W.E. Upjohn Institute. devolution of public sector activity. Bird, Richard M. 1986. “On Measuring Fiscal Centralization and Fiscal Balance in Federal There are, as Bryce suggests, important States,” Gov. Pol., 4, pp. 384–404. forces working in both directions, and ——— . 1992. Tax Policy and Economic Develop- one can expect the net effect to move in ment. Baltimore: Johns Hopkins Press. Bird, Richard M.; Robert D. Ebel, and Christine different directions as nations evolve I. Wallich, eds. 1995. Decentralization of over time. the Socialist State. Washington, DC: World What does seem to be taking place is Bank. Blank, Rebecca. 1997. “Policy Watch: The 1996 a growing complexity and specialization Welfare Reform,” J. Econ. Perspect., 11:1, pp. in the vertical structure of the public 169–78. 1146 Journal of Economic Literature Vol. XXXVII (September 1999) Boadway, Robin. 1996. “Review of ‘The Uneasy Good Economic Policy If You Tripped Over Case for Equalization Payments’ by Dan One?” Nat. Tax J., 47, pp. 863–81. Usher,” Nat. Tax J., 49:4, pp. 677–86. Cumberland, John H. 1979. “Interregional Pollu- Boadway, Robin and Frank R. Flatters. 1982. “Ef- tion Spillovers and Consistency of Environ- ficiency and Equalization Payments in a Fed- mental Policy,” in Regional Environmental Pol- eral System of Government: A Synthesis and icy: The Economic Issues. H. Siebert et al., eds. Extension of Recent Results,” Can. J. Econ., NY: NYU Press, pp. 255–81. 15:4, pp. 613–33. ———. 1981. “Efficiency and Equity in Interre- Boeckelman, Keith. 1992. “The Influence of States gional Environmental Management,” Rev. Re- on Federal Policy Adoptions,” Pol. Stud. J., gional Stud., 2, pp. 1–9. 20:3, pp. 365–75. Davoodi, Hamid and Heng-fu Zou. 1998. “Fiscal Bradford, David F. and Wallace E. Oates. 1971a. Decentralization and Economic Growth,” J. Ur- “Towards a Predictive Theory of Intergovern- ban Econ., 43, pp. 244–57. mental Grants,” Amer. Econ. Rev., 61:2, pp. Enrich, Peter D. 1996. “Saving the States from 440–48. Themselves: Commerce Clause Constraints on ———. 1971b. “The Analysis of Revenue Sharing State Tax Incentives for Business,” Harvard in a New Approach to Collective Fiscal Deci- Law Rev., 110, pp. 378–461. sions,” Quart. J. Econ., 85:3, pp. 416–39. Epple, Dennis and Allan Zelenitz. 1981. “The Im- ———. 1974. “Suburban Exploitation of Central plications of Competition Among Jurisdictions: Cities and Governmental Structure,” in Re- Does Tiebout Need Politics?” J. Polit. Econ., distribution Through Public Choice. Harold 89, pp. 1197–217. Hochman and George Peterson, eds. NY: Co- Esty, Daniel C. 1996. “Revitalizing Environmental lumbia U. Press, pp. 43–90. Federalism,” Mich. Law Rev., 95, pp. 570–653. Break, George F. 1967. Intergovernmental Fiscal Feldstein, Martin S. 1975. “Wealth Neutrality and Relations in the United States. Washington, DC: Local Choice in Public Education,” Amer. Brookings Institution. Econ. Rev., 65:1, pp. 75–89. Brennan, Geoffrey and James Buchanan. 1980. Feldstein, Martin and Marian Vaillant Wrobel. The Power to Tax: Analytical Foundations of a 1998. “Can State Taxes Redistribute Income?” Fiscal Constitution. Cambridge: Cambridge U. J. Pub. Econ., 68:3, pp. 369–96. Press. Fischel, William. 1992. “Property Taxation and the Breton, Albert. 1998. Competitive Governments: Tiebout Model: Evidence for the Benefit View An Economic Theory of Politics and Public Fi- from Zoning and Voting,” J. Econ. Lit., 30:1, nance. Cambridge: Cambridge U. Press. pp. 171–77. Brown, Charles C. and Wallace E. Oates. 1987. Fisher, Ronald C. 1982. “Income and Grant Ef- “Assistance to the Poor in a Federal System,” J. fects on Local Expenditure: The Flypaper Ef- Pub. Econ., 32, pp. 307–30. fect and Other Difficulties,” J. Urban Econ., 12, Brueckner, Jan E. 1998a. “Testing for Strategic In- pp. 324–45. teraction Among Local Governments: The Case Flatters, Frank; Vernon Henderson, and Peter Mi- of Growth Controls,” J. Urban Econ., 44, pp. eszkowski. 1974. “Public Goods, Efficiency, and 438–67. Regional Fiscal Equalization,” J. Pub. Econ., 3, ———. 1998b. “Welfare Reform and the Race to pp. 99–112. the Bottom: Theory and Evidence,” Institute of Gamkhar, Shama and Wallace Oates. 1966. “Asym- Gov. and Public Affairs, U. Illinois, WP 64. metries in the Response to Increases and De- Bryce, James. 1901. The American Common- creases in Intergovernmental Grants: Some wealth. London: Macmillan [first published Empirical Findings,” Nat. Tax J., 49, pp. 501–12. 1888]. Gordon, Roger. 1983. “An Optimal Tax Approach Case, Anne; James R. Hines, Jr., and Harvey S. to Fiscal Federalism,” Quart. J. Econ., 97, pp. Rosen. 1993. “Budget Spillovers and Fiscal Pol- 567–86. icy Interdependence: Evidence from the Gramlich, Edward M. 1977. “Intergovernmental States,” J. Pub. Econ., 52, pp. 285–307. Grants: A Review of the Empirical Literature,” Chernick, Howard. 1979. “An Economic Model of in The Political Economy of Fiscal Federalism. the Distribution of Project Grants,” in Fiscal Wallace Oates, ed. Lexington, MA: Heath-Lex- Federalism and Grants-in-Aid. P. Mieszkowski ington, pp. 219–39. and W. Oakland, eds. Washington DC: Urban ———. 1987. “Federalism and Federal Deficit Institute, pp. 81–103. Reduction,” Nat. Tax. J., 40, pp. 299–313. ———. 1992. “A Model of the Distributional Inci- Gray, Virginia. 1973. “Innovation in the States: A dence of State and Local Taxes,” Pub. Fin. Diffusion Study,” Amer. Polit. Sci. Rev., 67, pp. Quart., Oct., pp. 572–85. 1174–85. Conyers, Diana. 1990. “Centralization and Devel- Hamilton, Bruce W. 1975. “Zoning and Property opment Planning: A Comparative Perspective,” Taxation in a System of Local Governments,” in Decentralizing for Participatory Planning. Urban Stud., 12:2, pp. 205–11. P. de Valk and K. Wekwete, eds. Aldershot: ———. 1976. “Capitalization of Intrajurisdictional Avebury. Differences in Local Tax Prices,” Amer. Econ. Courant, Paul N. 1994. “How Would You Know a Rev., 66:5, pp. 743–53. Oates: An Essay on Fiscal Federalism 1147 Hines, James R. Jr., and Richard H. Thaler. 1995. Union,” in Macroeconomic Dimensions of Pub- “The Flypaper Effect,” J. Econ. Perspect., 9, pp. lic Finance: Essays in Honor of Vito Tanzi. 217–26. Mario Blejer and Teresa Ter-Minassian, eds. Huff, David L.; James M. Lutz, and Rajendra Sri- London: Routledge, pp. 73–93. vastava. 1988. “A Geographical Analysis of the ——— . 1997b. “EMU as a Device for Collective Innovativeness of States,” Econ. Geogr., 64, pp. Fiscal Retrenchment,” Amer. Econ. Rev., 87, 137–46. pp. 227–29. Huther, Jeff and Anwar Shah. 1996. “A Simple McKinnon, Ronald I. and Thomas Nechyba. 1997. Measure of Good Governance and Its Appli- “Competition in Federal Systems: The Role of cation to the Debate on the Appropriate Level Political and Financial Constraints,” in The New of Fiscal Decentralization.” Washington, DC: Federalism: Can the States be Trusted? John World Bank. Ferejohn and Barry Weingast, eds. Stanford: Inman, Robert P. 1988. “Federal Assistance and Hoover Institution Press, pp. 3–61. Local Services in the United States: The Evolu- McLure, Charles E., Jr. ed. 1983. Tax Assignment tion of a New Federalist Fiscal Order,” in Fiscal in Federal Countries. Canberra: Australian Na- Federalism. Harvey Rosen, ed. Chicago: U. Chi- tional U. cago Press, pp. 33–74. McWhinney, Edward. 1965. Comparative Federal- ———. Forthcoming. “On Designing Intergovern- ism. 2nd ed. Toronto: U. Toronto Press. mental Transfers With an Application in the Mieszkowski, Peter, and George R. Zodrow. 1989. New South Africa,” in Environmental and Pub- “Taxation and the Tiebout Model: The Differ- lic Economics. Arvind Panagariya, Paul Portney, ential Effects of Head Taxes, Taxes on Land, and Robert Schwab, eds. Cheltenham, UK: Ed- Rents, and Property Taxes,” J. Econ. Lit., 27:3, ward Elgar. pp. 1098–146. Inman, Robert P., and Daniel L. Rubinfeld. 1979. Musgrave, Richard M. 1959. The Theory of Public “The Judicial Pursuit of Local Fiscal Equity,” Finance. NY: McGraw-Hill. Harvard Law Rev., 92, pp. 1662–750. Nechyba, Thomas. 1996. “A Computable General ———. 1996. “Designing Tax Policy in Federalist Equilibrium Model of Intergovernmental Aid,” Economies: An Overview,” J. Pub. Econ., 60:3, J. Pub. Econ., 62, pp. 363–99. pp. 307–34. Nice, David C. 1994. Policy Innovation in State ———. 1997a. “Making Sense of the Antitrust Government. IA: Iowa U. Press. State-Action Doctrine: Balancing Political Par- Oates, Wallace E. 1972. Fiscal Federalism. NY: ticipation and Economic Efficiency in Regula- Harcourt Brace Jovanovich. tory Federalism,” Texas Law Rev., 75, pp. ——— . 1978. “The Changing Structure of Inter- 1203–99. governmental Fiscal Relations,” in Secular ———. 1997b. “The Political Economy of Feder- Trends of the Public Sector. H. Recktenwald, alism,” in Perspectives on Public Choice: A ed. Paris: Editions Cujas, pp. 151–60. Handbook. D. Mueller, ed. Cambridge: Cam- ——— . 1985. “Searching for Leviathan: An Em- bridge U. Press, pp. 73–105. pirical Study,” Amer. Econ. Rev., 75, pp. 748– ———. 1997c. “Rethinking Federalism,” J. Econ. 57. Perspect., 11:4, pp. 43–64. ——— . 1989. “Searching for Leviathan: A Reply ———. Forthcoming. “Subsidiarity and the Euro- and Some Further Reflections,” Amer. Econ. pean Union,” in The New Palgrave Dictionary Rev., 79, pp. 578–83. of Economics and Law. ——— . 1994. “Federalism and Government Fi- Kim, Sang Loh. 1995. Fiscal Decentralization, Fis- nance,” in Modern Public Finance. J. Quigley cal Structure, and Economic Performance: and E. Smolensky, eds. Cambridge, MA: Har- Three Empirical Studies. Unpublished Ph.D. vard U. Press, pp. 126–51. Dissertation, U. Maryland. ——— . 1996a. “Estimating the Demand for Pub- Lutz, James M. 1987. “Regional Leadership Pat- lic Goods: The Collective Choice and Contin- terns in the Diffusion of Public Policies,” Amer. gent Valuation Approaches,” in The Contingent Polit. Quart., 15, pp. 387–98. Valuation of Environmental Resources. D. Martin, Alison and W. Arthur Lewis. 1956. “Pat- Bjornstad and J. Kahn, eds. Aldershot: Edward terns of Public Revenue and Expenditure,” Elgar, pp. 211–30. Manchester Sch. Econ. Soc. Stud., 24, pp. 203– ——— . 1996b. “Taxation in a Federal System: The 44. Tax-Assignment Problem.” Pub. Econ. Rev., 1, McGuire, Martin C. and Mancur Olson, Jr. 1996. pp. 35–60. “The Economics of Autocracy and Majority ——— . 1998. “On the Welfare Gains from Fiscal Rule: The Invisible Hand and the Use of Decentralization,” U. Maryland Econ. Dept. Force,” J. Econ. Lit., 34, pp. 72–96. WP 98–05. McGuire, Therese J. 1997. “Intergovernmental Oates, Wallace E. and Robert M. Schwab. 1988. Fiscal Relations and Social Welfare Policy,” in “Economic Competition Among Jurisdictions: Intergovernmental Fiscal Relations. Ronald C. Efficiency-Enhancing or Distortion-Inducing?” Fisher, ed. Boston: Kluwer, pp. 173–98. J. Pub. Econ., 35, pp. 333–54. McKinnon, Ronald I. 1997a. “Market-Preserving ——— . 1991. “The Allocative and Distributive Fiscal Federalism in the American Monetary Implications of Local Fiscal Competition,” in 1148 Journal of Economic Literature Vol. XXXVII (September 1999) Competition Among States and Local Govern- tralization,” in Evaluation and Development: ments. D. Kenyon and J. Kincaid, eds. Washing- The Institutional Dimension. Robert Picciotto ton, DC: Urban Institute, pp. 127–45. and Eduardo Wiesner, eds. Washington, DC: ———. 1996. “The Theory of Regulatory Federal- World Bank, pp. 83–96. ism: The Case of Environmental Management,” Sparer, Michael S., and Lawrence D. Brown. in The Economics of Environmental Regulation. 1996. “States and the Health Care Crisis: The Wallace Oates. Aldershot: Edward Elgar, pp. Limits and Lessons of Laboratory Federalism,” 319–31. in Health Policy, Federalism, and the American ———. 1997. “The Impact of Urban Land Taxa- States. Robert F. Rich and William D. White, tion: The Pittsburgh Experience,” Nat. Tax J., eds. Washington, DC: Urban Institute Press, 50, pp. 1–21. pp. 181–202. Olson, Mancur Jr. 1990. “The IRIS Idea.” IRIS, Stine, William F. 1994. “Is Local Government U. Maryland. Revenue Response to Federal Aid Symmetri- Osborne, David. 1988. Laboratories of Democ- cal?” Nat. Tax J., 47, pp. 799–816. racy. Boston: Harvard Business School Press. Strumpf, Koleman S. 1997. “Does Fiscal Decen- Pauly, Mark V. 1973. “Income Redistribution as tralization Increase Policy Innovation?” unpub- a Local Public Good,” J. Pub. Econ., 2, pp. 35– lished paper. 58. Strumpf, Koleman S. and Felix Oberholzer-Gee. Pommerehne, Werner W. 1977. “Quantitative As- 1998. “Endogenous Institutions and Policy De- pects of Federalism: A Study of Six Countries,” centralization: Local Liquor Control from in The Political Economy of Fiscal Federalism. 1934–1970”: unpublished paper. Wallace Oates, ed. Lexington, MA: Heath-Lex- Tiebout, Charles. 1956. “A Pure Theory of Local ington, pp. 275–355. Expenditures,” J. Polit. Econ., 64, pp. 416–24. Poterba, James M. and Kim S. Rueben. 1997. Tocqueville, Alexis de. 1945. Democracy in Amer- “State Fiscal Institutions and the U.S. Munici- ica. NY: Vintage Books Random House: first pal Bond Market,” NBER WP 6237. published 1838. Prud’homme, Remy. 1995. “On the Dangers of US Congressional Budget Office. 1997. Federal- Decentralization,” World Bank Res. Observer, ism and Environmental Protection: Case Stud- pp. 201–10. ies for Drinking Water and Ground-Level Qian, Yingyi, and Barry R. Weingast. 1997. “Fed- Ozone. Washington, DC: US Govt. Printing Of- eralism as a Commitment to Preserving Market fice. Incentives,” J. Econ. Perspect., 11:4, pp. 83– Usher, Dan. 1995. The Uneasy Case for Equaliza- 92. tion Payments. Vancouver, BC: The Fraser In- Rao, M. Govinda. 1998. “Invisible Transfers in In- stitute. dian Federalism,” Australian National U. Econ. Volden, Craig. 1997. “Entrusting the States with Dept. WP 98/1. Welfare Reform,” in The New Federalism: Can Revesz, Richard L. 1992. “Rehabilitating Inter- the States Be Trusted? John Ferejohn and Barry state Competition: Rethinking the ‘Race-to-the- Weingast, eds. Stanford: Hoover Institution Bottom’ Rationale for Federal Environmental Press, pp. 65–96. Regulation,” NYU Law Rev., 67, pp. 1210–54. Walker, Jack L. 1969. “The Diffusion of Innova- Rivlin, Alice. 1992. Reviving the American Dream. tion Among the American States,” Amer. Polit. Washington, DC: The Brookings Institution. Sci. Rev., 63, pp. 880–99. Rodden, Jonathan, and Susan Rose-Ackerman. Wallis, John J. and Wallace E. Oates. 1998. “The 1997. “Does Federalism Preserve Markets?” U. Impact of the New Deal on American Federal- Virginia Law Rev., 83, pp. 1521–72. ism,” in The Defining Moment: The Great De- Rogers, Everett. 1983. Diffusion of Innovations. pression and the American Economy in the 20th 3rd ed. NY: Free Press. Century. M. Bordo, C. Goldin, and E. White, Rose-Ackerman, Susan. 1980. “Risk Taking and eds. Chicago: U. Chicago Press, pp. 155–80. Reelection: Does Federalism Promote Innova- Weingast, Barry R. 1995. “The Economic Role of tion?” J. Legal Stud., 9, pp. 593–616. Political Institutions: Market-Preserving Feder- Rubinfeld, Daniel L. 1987. “The Economics of the alism and Economic Development,” J. Law Local Public Sector,” in Handbook of Public Econ. Org., 11, pp. 1–31. Economics, Vol. II. Alan Auerbach and Martin Wildasin, David E. 1988. “Nash Equilibria in Feldstein, eds. Amsterdam: North-Holland, pp. Models of Fiscal Competition,” J. Pub. Econ., 571–645. 35, pp. 229–40. Saavedra, Luz Amparo. 1998. “A Model of Welfare ———. 1998a. “Factor Mobility and Redistribu- Competition with Evidence from AFDC,” Inst. tive Policy: Local and International Perspec- Govt. and Public Affairs, U. Illinois, WP 63. tives,” in Public Finance in a Changing World. Shah, Anwar. 1994. The Reform of Intergovern- Peter B. Sorensen, ed. London: MacMillan, pp. mental Fiscal Relations in Developing and 151–92. Emerging Market Economies. Washington, DC: ———. 1998b. “Externalities and Bailouts: Hard World Bank. and Soft Budget Constraints in Intergovern- ———. 1998. “Fostering Fiscally Responsive and mental Fiscal Relations,” World Bank Policy Accountable Governance: Lessons from Decen- Research WP 1843. Oates: An Essay on Fiscal Federalism 1149 Wilson, John Douglas. 1996. “Capital Mobility and Robert Hudec, eds. Cambridge, MA: MIT Environmental Standards: Is There a Theo- Press, pp. 393–427. retical Basis for a Race to the Bottom?” in Fair Zhang, Tao and Heng-fu Zou. 1998. “Fiscal De- Trade and Harmonization: Prerequisities for centralization, Public Spending, and Economic Free Trade? Vol. I. Jagdish Bhagwati and Growth in China,” J. Pub. Econ., 67, pp. 221–40.