but we explore the implications of this assumption because it is implicit in the
reasoning of many advocates of tax competition.
Our study offers a synthesis of the traditional Public Finance view of tax
competition and the view of the Public Choice school by embedding a probabilistic
voting model of the type proposed by Persson and Tabellini (2000, ch. 3) in a tax
competition model similar to the one used by Edwards and Keen.4 In our model
the often vague concept of ‘political distortion’ has a very precise meaning. Our
indicator of the degree of political distortion depends on the size of the public
sector lobby and on the relative political influence of an individual lobby member,
measured by the derivatives of the voting function maximised by politicians. The
greater the sensitivity of voting behaviour to a change in economic benefits offered
to a lobby member, the greater is his political influence relative to the influence
of a voter outside the lobby.5
In our framework tax competition is a powerful institutional device which
may completely wipe out rents to public sector workers, but only at the cost of an
underprovision of public goods. Unlike the Leviathan literature, this paper offers
a theory of the ’political transmission mechanism’ through which tax competition
leads to reduced rent creation as well as reduced public goods provision. In our set-
up politicians may capture more votes from public sector workers by paying them
higher wages, and they may also attract votes by creating additional high-paying
public sector jobs. At the same time politicians may gain votes by offering higher
private consumption opportunities through lower taxes. In political equilibrium,
political candidates strike a balance between these competing ways of gaining
votes, accounting for the government budget constraint. When tax competition
is allowed, the amount of private consumption that must be sacrificed to raise
4 Our paper may also be seen as an extension of some ideas in Wilson (1989) who studies
optimal constraints on the tax base in a world where the tax rate is controlled by a policy maker
who diverts resources from spending on public goods towards a favoured group of consumers.
Such behaviour by the policy maker could be interpreted as an attempt to buy votes from an
influential interest group, but unlike us, Wilson (op.cit.) does not explicitly model the political
process, and he does not consider the effects of tax competition.
5 To limit the scope of the paper, we do not consider whether tax competition leads to
less corruption and whether it can be used to generate valuable information to voters. Much
empirical work on the efficiency effects of fiscal federalism has focused on the relationship between
fiscal decentralisation and corruption. In a cross-country panel study using the International
Country Risk Guide’s corruption index, Fissman and Gatti (2002) find a significant negative
relationship between corruption and decentralization. Huther and Shah (1998) obtain similar
results using similar indicators of corruption constructed by the World Bank.