Federalism, Competition and Public Choice

A simple way to solve the public good problem is to privatize public goods. Charles E. Tiebout (1956) has shown that local public goods can be privatized if localities are separated from each other.

Tiebout competition generates important adjustment processes. First, local communi­ties will specialize in particular services and hence induce citizens with similar preferences to cluster in specialized local communities. Second, as far as the market is a discovery process (Hayek 1968 [2002]), local jurisdictions under competition will search for new products and processes, hence dynamically improve the quality of local public services. Third, competitive federalism is a laboratory for new ideas (see Oates’s concept of “labo­ratory federalism” in Oates 1999). Fourth, federalism may crowd out inefficient bureau­cratic and Leviathan solutions (Weingast 1995); and fifth, federalism may bring a local community on to a steeper growth path (Feld and Schnellenbach 2011).

The pure Tiebout model does not provide for public choice. Local governments work in the same way as markets work. They find the efficient solution themselves. Public choice is not only not necessary, it would even interfere with the market and destroy Tiebout’s efficiency. Public choice can, however, help to support federalism where the latter is not strong enough to bring about an efficient spatial specialization.

Suppose that the 100000 inhabitants of jurisdiction J, which consists of two local communities A and B with 55000 and 45000 inhabitants respectively, are debating the introduction of a progressive school system X or a conservative school system Y, see Table 2.

If jurisdictions A and B vote in their entirety between the two school systems, the pro­gressive school system X will be chosen. The preferences of 55 000 voters will be fulfilled, while 45 000 inhabitants are outvoted. If the inhabitants of A and B vote separately,

Table 2 Collective choice of school system X and Y

Sources: Wagner (1983); Blankart (2011).

however, the A inhabitants would choose the conservative school system Y, while the B inhabitants would opt for the progressive school system X. As can be seen, only 35000 inhabitants are outvoted: this is a clear advantage of decentralized voting (under the assumption that the inhabitants in A are equally concerned about school systems as those in B). Decentralized vote reveals that voters have different preferences.

Note that with the same overall, but another inter-local, distribution of votes, both communities might vote for the progressive school system so that decentralized decision making does not seem necessary. The equality of the two outcomes is, however, not an argument against decentralized voting, for how could one know that an aggregated vote is enough if not proved by decentralized voting? Therefore it is more reliable to start with decentralized government before centralized solutions are considered. This insight has been anchored in Oates’s decentralization theorem:

For a public good - the consumption of which is defined over geographical subsets of the total population, and for which the costs of providing each level of output of the good in each jurisdiction are the same for the central or the respective local government - it will always be more efficient (or at least as efficient) for local governments to provide the Pareto-efficient levels of output for their respective jurisdictions than for the central government to provide any specified and uniform level of output across all jurisdictions.

The conclusion is that a democracy can support federalism where its own forces are not strong enough to bring about an efficient outcome, but that democracy cannot replace federalism; for how could the vote districts be formed without the prior evidence pro­vided by federalism?

Oates’s decentralization theorem requires calculating the relative costs of alternative levels of government in each case. This, too, is often not possible. A more practical alter­native is provided by the subsidiarity principle. The principle says that if a problem can be tackled individually or within an individual’s family it should not be delegated to the local or a higher-level government. The burden of the proof of centralization is always with the next higher entity. If the subsidiarity test is negative the issue remains with the lower level. The subsidiarity principle has been accepted as a general principle in politics. It is also embedded in article 5 of the Treaty of Lisbon of the European Union: “The use of Union competences is governed by the principles of subsidiarity and proportionality.” The subsidiarity principle is related to the principle of institutional congruency. Only when the citizens of a jurisdiction bear the full costs of a public service can they calculate rationally whether it is worthwhile to increase taxes for its extension or vice versa to decrease taxes. However, if ambitious politicians believe that they can obtain power by redistributing costs they will not hesitate to do so. Therefore, the subsidiarity principle is weak in politics. It is often violated in order to attract more votes.