تراکم و بخش دولتی

I present a two-region model in which the actions of local governments can cause the agglomeration of population. A local public good is a centripetal force. The centrifugal force is competition among interregionally mobile workers for housing. In the model the supply of the local public good rises in the number of workers in a region. This creates incentives for immigration. Depending on the relative strength of the centrifugal force, full agglomeration of all workers in one region can result even with perfectly identical regions. This creates an incentive for the local governments in each region to provide the public good strategically leading to overprovision and suboptimal agglomeration.

The question why agglomerations of economic activity form is an old one but recently regained a lot of attention in the economic literature. The recent literature on agglomeration basically treats two types of reasons for the clustering of firms and households (see Fujita and Thisse, 2000). The first type of explanation are the so-called Marshallian external economies (spillovers, labor market externalities, and shared inputs). The second and more prominent explanation are the pecuniary externalities which are modeled by the New Economic Geography (Krugman, 1991, Venables, 1996 and Fujita et al., 1999). In both theories, the agglomeration forces arise from the interaction of many private agents that behave nonstrategically. The public sector is almost never considered as an agent so that public actions and agglomeration are rarely linked1. This neglect of the public sector in the agglomeration literature is both concerning and surprising. It is concerning since the governments are powerful actors that try to influence the spatial allocation of factors for many reasons (see Funck, 1995). Theories of the spatial distribution of economic activity, which omit this force, are incomplete and not applicable to real world issues. It is also surprising because there already is a well-established literature which discusses the local governments’ influence on the interregional allocation of mobile factors in public finance: the Tiebout literature (see Tiebout, 1956, Stiglitz, 1977 and Wildasin, 1987) and the literature on tax competition (see Wilson, 1986, Wilson, 1999 and Wellisch, 2000). Especially the Tiebout literature is very close to some of the problems that are discussed in the agglomeration literature. It analyzes how households react to the provision of local public goods, i.e. how the public sector influences the spatial distribution of mobile households. Although this question is closely linked to the question how agglomerations form, the Tiebout literature has a different focus than the agglomeration literature. The main question analyzed is whether competition among local authorities leads to the efficient decentralized provision of public goods and thus can solve the freerider problem. In most models, the agglomeration question cannot be analyzed explicitly because they treat a large number of identical regions without strategic interactions and have very general assumptions concerning preferences and technologies. In such a setup, agglomeration never arises or is excluded by assumption (see, e.g. Richter, 1994) since it is considered as a border solution which does not allow the derivation of optimal tax policies. An exception is Stiglitz (1977), which presents a two-region model. In this model, it is explicitly shown that partial or total agglomeration of all mobile households in one of the two regions can arise because of the provision of a pure local public good financed by a tax on the residents of the respective region. The agglomeration force is the following circular causality: a larger population in one region means that the taxbase is broader so that more of the public good can be provided. This increases the utility level in that region so that further immigration from the other region is beneficial. In this model, I pick up the idea that local public goods might be a centripetal force and introduce this insight into a model which has a strong New Economic Geography flavor. One important contribution is to incorporate the public sector into the theories of agglomeration. Although the insight that the public sector might be a reason for agglomeration is not new, the recent works on agglomeration have completely ignored the whole public finance literature2. The model is a general equilibrium model with two regions, two private goods (consumption good and housing) and a pure local public good. There are two types of individuals: mobile workers and immobile houseowners. Both of them consume all three goods. The local governments of the two regions can only tax the houseowners rents. Using this tax revenue, they buy the consumption good and transform it into the local public good. The government's objective functions are the utility functions of the immobile houseowners. It is shown that the local governments can use the supply of the public good as a strategic variable in order to raise their share of the total mobile population. This benefits the houseowners. The supply of the public good increases in the number of a region's inhabitants which can result in a circular causation. The more of the public good is provided in a region, the more attractive is this region for immigrants. The more individuals migrate to a region, the higher are the government's revenues so that more of the public good can be produced. If the local governments have a very productive technology to produce the public goods at their disposal, full agglomeration of the mobile population in one of the two regions can arise. If the government's technology is not so productive, these border equilibria do not result. Then, with identical regions, there is only the symmetric equilibrium in which workers are evenly distributed on both regions. Yet with strategic behavior of governments, too much of the public good will be produced. If one region has a higher productivity of labor than the other, strategic behavior leads to too little agglomeration and overprovision of the public good. Even if my model is similar to the model in Stiglitz (1977), there are important differences. Stiglitz shows for two regions that the provision of a local public good might lead to total, partial, or no agglomeration of population. However, he does not explicitly discuss the conditions for each kind of equilibrium. From Stiglitz (1977) we learn that agglomeration might arise but not when it will. It is not possible in Stiglitz's model to analyze the agglomeration phenomenon in more detail since the model lacks the necessary structure to do so. For this reason, Stiglitz (1977) is not an agglomeration model in a narrow sense. It is rather a public finance model since it is built in order to study under which conditions the market equilibrium and the Pareto-efficient allocation of mobile households fall apart in a model with local public goods. Section 2 presents the model in detail. I determine the equilibrium value of the model's central variable, the interregional allocation of mobile workers, and state conditions for partial and total agglomeration. In Section 3, I discuss how the local governments behave under the assumption that the utility of immobile houseowners is maximized. Section 4 summarizes the results.

I have demonstrated that the provision of a public good, which is financed by a tax on an immobile local factor only, can induce households to immigrate. If this results in a higher income of the immobile factor, the supply of the public good can be larger triggering further immigration. A fiscal externality works as a centripetal force which can cause agglomeration. Local governments, which maximize the utility of residents receiving income from housing, have an incentive to provide the public good strategically. Regional competition for mobile population thus causes overprovision of the public good. With regional productivity differences, the regions strategic behavior distorts the interregional labor allocation so that too few workers live in the more productive region. These inefficiencies can serve as arguments against the decentralized provision of local public goods. If the regional supply of the public good were determined centrally, the negative externalities would be internalized. Two comments with respect to the public finance literature are worth making. It is a well established result of the tax competition literature that non-cooperative behavior of governments generally causes underprovision of public goods and suboptimally low tax rates (Zodrow and Mieszkowski, 1986, Fujita et al., 1999 and Wellisch, 200011. Contrary to this literature, there are no taxes on the mobile factor in my model. Therefore, my result is difficult to compare with the results of the tax competition literature. A second difference to results from public finance is that a tax on immobile property is not neutral with respect to migration in my model12. In the Tiebout literature, a tax on land rents usually does not affect the residential decision of mobile households because land is supplied inelastically in these models (see Wildasin, 1987, p. 1142, 1153, 1156). Since in my model all individuals—i.e. the houseowners, too—demand both the consumption good and housing and only houseowners are taxed, the supply of housing is not fixed. But even if it were inelastic, the tax would influence the migration decision because the mobile individuals benefit from the public good without having to bear a tax burden. This difference is important because immigration due to higher taxation of immobile property is the model's central mechanism. Although my model works like a New Economic Geography model, it does not rely on increasing returns to scale in production or technological externalities. There is, of course, an externality which is due to a pure public good. The agglomeration forces completely arises from a pecuniary or fiscal externality. This contrasts sharply with the existing theories explaining agglomeration. Looking for causes of agglomeration, one is not restricted to Marshallian externalities or market size effects as in Krugman's New Economic Geography. In addition, governments can influence the location decision not only of firms but of households as well. The agglomeration of population is thus not necessarily a consequence of the agglomeration of firms but can have its own causes. The aim of this paper is to broaden the narrow view of the agglomeration literature and to introduce the public sector into the analysis of agglomerations. This seems necessary for theoretical and empirical reasons. A possible extension of this simple model could be to endogenize the productivity differences. It would be interesting to analyze the interactions between the agglomeration forces caused by the public sector and the ones caused by the private agents. This might produce a new perspective on regional competition.