Fiscal federalism

Theories of multi-layer institutional organization builds upon the Oates De­centralization theorem (Oates, 1972), according to which the upper government level comes in only when the lower government level proves inefficient for some reason.

In the Oates model, spillover effects across jurisdictions are assumed to be non-existent. The implicit hypothesis is that every ‘package’ of taxes and public goods offered by a local government is perfectly oriented to its own community, because in each jurisdiction all individuals share the same ‘juris­dictional’ preference for taxes and public goods. Yet, the decentralization theorem is very heroic, due to the ‘correspondence principle’ which postu­lates a complete overlapping between the territory of the political jurisdiction and the area over which the economic effects of public policies are spread. On the other hand, the capacity of the packages to satisfy the particular preferences of each community does not rule out the possibility of externali­ties. In other words, the jurisdiction does not necessarily correspond to the ‘economic’ jurisdiction. The public goods produced in a jurisdiction often have positive spillovers, as they can be used by individuals belonging to other jurisdictions (for example, health care, transportation and so on).

As a solution to these questions, the Tiebout model (Tiebout, 1956) has suggested that competition among lower-level jurisdictions in offering pack­ages can be instrumental in fostering economic efficiency. Given similar jurisdictions with the same tax rate, the following conditions should hold for decentralized taxation and public goods provision to be efficient: (a) zero cost mobility of individuals (and financial capital); (b) perfect supply elastic­ity of competitive jurisdictions; (c) public goods are provided at the minimal average cost; (d) complete information (and common knowledge) about pref­erences and the offered packages of tax and public goods; and (e) inter-jurisdictional externalities are absent. The resulting efficiency condition is that the taxes levied to each individual equalize the cost of production of the public goods desired by residents. The production of public goods re­sponds to the well-known Samuelson condition whereby the marginal benefit equalizes the marginal cost. If goods and services are exposed to congestion, for the competition among the lower-level jurisdictions to be efficient, the additional qualification has to be added that the public sector production abides by the condition of the Buchanan’s ‘theory of club goods’: the average cost of one more resident must equal the marginal cost of one more resident (Buchanan, 1965).

Tiebout’s rationale for inter-jurisdictional competition is to mimic the func­tioning of the perfect competition market.

The Tiebout model is interesting because it allows the existence conditions for a competitive market for public goods to be neatly analysed. First, the supply of public goods is efficient when it creates benefits only for residents and taxes are levied only on them. The consideration of jurisdictions which are different for even a single factor produces interdependences among them, thus preventing the efficiency conditions from being met.

A typical differential factor is represented by technological conditions of production across jurisdictions. Inside large jurisdictional entities, such as the US and the EU, different technology levels are a main determinant of the divide between the centre and the periphery. Let us assume that economies of scale have fostered territorial concentration of research laboratories and firms operating in advanced sectors - such that vacancies are concentrated in a certain area of a federation. This raises a problem. On the one hand, for the number of jurisdictions compatible with perfect competition to exist, the communities should be uniformly spread over the territory. On the other hand, the high labour demand in the ‘agglomeration jurisdiction’, where most dynamic industrial sectors and services (and most job vacancies) are concen­trated, attracts many residents from other jurisdictions.

Note that this example abides by the Tiebout hypotheses. In particular, it does not conflict with condition (e). The presence in the agglomeration area of firms operating a demand for labour which attracts workers from outside does not represent an externality but an interdependence, as it can be solved by the system of prices absorbing the excess demand. Were the supply of public transportation not equalized with the commuters’ demand, the com­muting private costs could be too high and the mobile individuals would be forced to move to the agglomeration area. This possible outcome fails to meet condition (b) of efficient supply of jurisdictions, as for the optimal dimension for all public goods to be obtained the infinite supply of jurisdic­tions should be violated. The advantage of living in an advanced area may be counterbalanced by the price differentials (for instance, in housing) which opens vis-a-vis the other jurisdictions due to the industrial concentration, which encourages commuting. By defining B(1) the benefits and MC(1) the marginal costs of congestion in the agglomeration area and B(n - 1) the benefits and MC(n - 1) the marginal costs of congestion in the other (n - 1) jurisdictions, the inequality B(1) - MC(1) < B (n - 1) - MC(n - 1) follows. To cope with the augmented congestion costs, taxation is increased in the other jurisdictions (let us call them ‘residence jurisdictions’) and the supply of public goods will shrink. This public sector reaction to excess demand disturbs the conditions for perfect competition among jurisdictions. The fail­ure to meet the residence choice of the commuters reveals that the conditions of perfect competition for the supply of ‘residence jurisdictions’ are no longer fulfilled.

In the opposite case, the increased demand for public transportation is ‘accommodated’ and an inefficient outcome ensues. Since the residents’ choice conflicts with the economies of agglomeration, in the residence jurisdictions the excess demand for public transportation by the commuters working in the agglomeration area determines a much higher dimension of public transpor­tation with respect to the one that would equalize the demand for the other public goods (assuming that they equalize with the federal average).

Let us now waive the condition of a given quantity of capital, which is unrealistic once capital market liberalization is considered. On the one hand, every excess of public expenditure to be financed gives to each residence jurisdiction the incentive to attract foreign investments looking for the best profit conditions in liberalized financial markets. In order that the optimal dimension be financed and the public budget be in equilibrium, the tax rate is increased. On the other hand, under the hypotheses that individuals will move into those jurisdictions offering the packages that best suit their preferences, and the income level being the main determinant of preferences, each juris­diction involved in the strategic game has the incentive to attract the individuals with the largest tax base by reducing the tax rate and targeting the public goods preferred by them. The ensuing tax differentials may undermine the efficiency level of the transfers and services ranging from the risk insurance to the purely redistributive institutions of social protection (health care, edu­cation, poverty subsidies and so on).

The outcome is that the type composition of jurisdictions changes: they will no longer comprise heterogeneous individuals, but will drift towards ‘homogeneity’, as some jurisdictions will essentially be composed of rich individuals, while poor individuals will concentrate in other jurisdictions. A possible explanation is that the goods offered by the public sector are merit and public goods, for which competitive markets fail. In these service sectors characterized by market failures, the efficiency-enhancing incentives corre­sponding to the profit motivation of competitive firms cannot operate. Introducing competition among local governments in the supply of packages amounts to reproducing the same failures of private markets of perfect com­petition (Sinn, 2003). The suggestion is that a devolution process endowing lower-level jurisdictions with substantial fiscal autonomy should not be devised along the lines of the Tiebout model.

As far as social theory is concerned, at the beginning of the last century Max Weber perceived the trend towards homogeneous communities and the­orized that individuals as rational maximizers pursue the formation of ‘social closures’ because of their fall-out in terms of identity of interests (Weber, 1922 [1974]). Empirical research work on social segmentation in the US seems to confirm that the reduction in income redistribution depends on the increasing income homogeneity in local jurisdictions, as an effect of the community formation process according to the Tiebout model (Epple and Sieg, 1999). They even make recourse to a policy of ‘exclusionary zoning’, that is, regulations aimed at excluding from the area people with the same preferences as the community but a much lower income level (Cooter, 2000).