Tax competition and federal equalization schemes with decentralized leadership

Wilson (1986) and Zodrow and Mieszkowsky (1986) pioneered the literature on tax competition, which is now quite voluminous. See, e.g., Brueckner (2000), Bucovetsky (1991), Bucovetsky and Smart (2006), Janeba and Wilson (2011), Wildasin (2003, 2011), Wilson (1991, 1999, 2005), and Wilson and Wildasin (2004). This paper contributes to this literature by showing that tax competition may not be an impediment for either efficiency or equity if the center implements fiscal and income equalization in a federation with decentralized leadership and symmetric regional governments. Bucovetsky (1991) examines asymmetric tax competition where regions differ in terms of population sizes. He shows that the Nash equilibrium is inefficient if production functions are quadratic. Wilson (1991) shows that the inefficiency result holds under quite general circumstances. Bucovetsky and Smart (2006) demonstrate that there are circumstances under which fiscal equalization schemes can be optimally designed in order to induce efficient choices by competing regional governments. However, in their model, preferences are quasi-linear. Hence, there is no social need to redistribute income across regions.

Silva and Caplan (1997) point out that the European Union is a good example of a federal system in which subcentral governments have the ability to commit to provision of public goods and services. They called the hierarchical structure of such a system, “decentralized leadership.” See also Akai and Sato (2008), Aronsson (2010) and Caplan et al. (2000) for models of decentralized leadership.

Breuillé et al. (2010) also consider the efficiency properties of fiscal equalization in federations characterized by decentralized leadership and in which the regional governments compete in the setting of taxes on capital. Like Köthenbürger (2004), Breuillé et al. (2010) neglect income redistribution because in their model consumers also have quasi-linear preferences.

In order to ensure that the solutions of all maximization problems considered in this paper are interior, one can assume that the utility function satisfies the standard Inada conditions. This assumption is unnecessary for the analysis that follows since interior solutions can be obtained for a broader class of utility functions.

A good example is \(f\left( {k_j } \right) =k_j^\beta z^{1-\beta }\), where \(1>\beta >0\).

Superscripts are used to denote functions henceforth.

It should be clear from Eq. (2) that \({\partial g^{i}\left( . \right) }/{\partial s_h }=0\) for all \(i,h=1,\ldots ,J, h\ne i\).

In a more general model, there may be other equilibria in the first stage besides the symmetric equilibrium. Since the symmetric equilibrium yields the socially optimal allocation, it Pareto dominates all other potential equilibria. If one applies a further refinement of Nash equilibrium (in addition to subgame perfection) such as strong perfect Nash, which requires the players to select the Pareto superior equilibrium, one will then be able to predict that the equilibrium in the first stage will be symmetric, preventing the potential coordination failure.

In future work, one can extend the basic model of Köthenbürger (2004) in order to allow free migration across regions and income transfers (as, e.g., in Myers 1990). In this setting, the regional governments should have incentives to internalize the fiscal externalities and hence to provide efficient levels of regional public goods if they anticipate that regional utilities are equalized in the migration equilibrium.

An interesting avenue for future research would be to combine capital tax competition with nonlinear income tax competition in the presence of mobile workers of different abilities. One can build on Wilson et al. (2014). This paper makes an interesting contribution with a model of nonlinear income tax competition.

Buettner and Wildasin (2006) find empirical evidence that may support strategic manipulation of revenue-raising abilities in large cities relative to small cities in light of ex post fiscal equalizing transfers, since the results suggest that large cities rely more on intergovernmental grants. Considering this together with the positive result in this paper, one wonders whether the empirical evidence is demonstrating that the model considered here is too simple and incomplete because federal income transfers do not actually equalize marginal utilities of income across jurisdictions, or regional authorities are not as forward-looking and sophisticated as it is assumed in this paper, or even that there are substantial informational asymmetries that prevent regional authorities to see through the center’s equalization schemes.

See Silva (2015) for more discussion of the soft budget literature.