The aim of the paper is to develop a procedure that allows policy makers to make an ex-ante assessment of a general compulsory amalgamation policy, providing quantitative indications about the possible financial effects. The amalgamation of small municipalities is a widespread practice all over the world. Policy makers usually justify amalgamation policies with the hypothesis that local public service provision is characterized by economies of scale and scope. However, population size is not the only determinant of economies of scale, which depend on many other factors. For these reasons, the expected effects of any amalgamation policy are uncertain, and ex-post empirical analyses are unable to offer unambiguous indications to policy makers since all programs differ. After a brief discussion of the relevant issues concerning amalgamation, we present the procedure used to simulate the economics and administrative effects of a general compulsory amalgamation policy. The procedure is tested with reference to the municipalities of Veneto, a region of Italy, for which we provide the results of a number of simulations under alternative amalgamation policies. The main result is that amalgamation policies based only on the a priori rule that small municipalities should merge may be very inefficient, because the expenditure reduction following an amalgamation policy may depend to a considerable extent on other territorial and socio-economic characteristics of the municipalities involved.
Keywords: municipality; amalgamation; local government expenditure; economies of scale.
Municipal amalgamation is a widespread practice all over the world. (1) Policy makers usually justify amalgamation policies with the hypothesis that local public service provision is characterized by economies of scale and scope, meaning that a reduction in per capita expenditure or a lower tax and debt burden can be realized by amalgamating small municipalities. However, this argument has many shortcomings.
A number of empirical studies, such as those by Breunig and Rocaboy (2008), Sole-Olle and Bosh (2005) and Bonisch et al. (2011), confirm the existence of a U-shaped municipal per capita expenditure curve. (2) Others, such as Sampaio de Sousa et al. (2005) and Gimenez and Prior (2007), conclude that municipal efficiency increases almost linearly with population size. (3) In contrast, the study by Loikkanen and Susiluoto (2005) suggests that small municipalities are more efficient. (4) These contrasting results are not entirely driven by the different socio-economic contexts examined, as they also reflect different techniques and methods used in the analyses. However, the key criticism to the standard approach to amalgamation is that population size is not the sole determinant of economies of scale, which depend on many other factors.
As demonstrated by Drew and Dollery (2014), in those countries where local government services provision consists mainly of services to property, both the number of households and the number of households plus the number of business entities represent better proxies for measuring local government output. In these cases, the use of per capita values, and population as a measure of output, leads to underestimate the unit cost of the services and overestimate the point at which diseconomies begin.
A deep insight into this question is provided by Drew et al. (2015), who showed that, as far as the Australian case in concerned, population is not a relevant determinant of local government's performance. Important factors are instead the population density, quantum of liabilities and the length of roads. However, this result largely depends on the specification of the DEA model, which is, according to the authors, the critical point for this kind of analysis. The relevance of population density in assessing the economies of scale is particularly highlighted by Drew et al. (2014).
Trying to verify the existence of a U-shaped cost function in New South Wales' local government (Australia), Fahey et al. (2016) showed that the shape of the average cost curve does not depends on council size, measured by the number of rate-paying units. Evidence of economies of scale arises, however, with reference to general purpose councils, for which economies of scale are important for functions like public order, transport and economic affairs. The authors stressed the fact that "different functional categories may prescribe conflicting optimal size jurisdictions.
For these reasons, a vast amalgamation program involving only small and preconceived municipalities may prove unsatisfactory, since it could result in an increase in public expenditure.
This paper develops a procedure that allows policy makers to make an ex-ante assessment of a general compulsory amalgamation program, giving them quantitative indications of the possible financial and administrative effects of such a program. The method is case-specific, since it is elaborated on the basis of the specific institutional, financial and structural characteristics of the local administrations to be amalgamated. Since population size cannot be considered as the only parameter that triggers the amalgamation procedure, a number of alternative amalgamation policies must be defined in order to implement an amalgamation program. These policies will represent a crucial part of the analysis. For each amalgamation policy, the proposed procedure will provide indications of the expected administrative impacts of the reform, in terms of the number and size of the local administrations involved in the amalgamation program.
However, the main purpose of the procedure is to assess the expected financial impact of the program, evaluated in terms of expenditure reduction. This task involves not only estimating a municipal expenditure function, but also defining spatial relationships between municipalities, the amalgamation criteria used by policy makers and, in particular, the method used to compute the expected financial gain.
The paper is organized as follows: in the second section, we discuss the key questions that justify the ex-ante assessment of municipal amalgamation and explain our reason for concentrating on compulsory amalgamations. The procedure used to simulate the economic and administrative effects of a compulsory amalgamation policy is presented in the third section. In Sections 4 and 5, we estimate an expenditure function and discuss the presence of economies of scale in the municipalities of the Veneto Region, located in northeast Italy, and report the results of simulations carried out using four alternative amalgamation policies.
THE EXPECTED IMPACTS OF MUNICIPAL AMALGAMATIONS: THE KEY ISSUES
2.1. FROM VOLUNTARY TO COMPULSORY AMALGAMATIONS
Oates's fundamental study on decentralization highlighted the importance of preference heterogeneity in determining the optimal local government dimension (Oates, 1972). Preference heterogeneity is also at the basis of Tie-bout's (1956) model, in which households choose between communities on the basis of their provision of public services and their tax rates. (5) The model predicts an improvement in allocative efficiency if the number of jurisdictions is sufficiently high.
Preference heterogeneity is the main obstacle to voluntary amalgamations, which require the exploitation of economies of scale, and also allocative efficiency (i.e. if the amalgamation is approved by local communities).
In a heterogeneous context, the higher the concern about democracy and compliance with local preferences, the lower the optimal government dimension will be. The relevance of this aspect appears to be clear, considering the huge fragmentation of local government that occurred in many Central and Eastern European countries after 1990. (6)
A policy that encourages voluntary amalgamations provides compliance with local preferences and the success of the initiative. Hanes and Wikstrom (2010) showed that voluntary amalgamations in Sweden proved to be more efficient than compulsory amalgamations. Even the Local Government Association of Queensland (Dollery et al., 2013, page 226) stressed the importance of the voluntary nature of any amalgamation proposal for its ultimate success.
However, if there are strong differences in local identities, voluntary amalgamations are not easy to realize, even in the presence of economies of scale. In Italy, for example, where the recent federalist wave has stressed the importance of local autonomy, only a few amalgamations occurred in the last twenty years, despite the fact that national legislation provides strong incentives for voluntary amalgamations. (7) Pirani (2012) showed that only nine municipal amalgamations have taken place in Italy since 1995, involving just 24 municipalities out of a total of more than 8,000.
The use of national or regional incentives to promote voluntary municipal amalgamation is another questionable point. In fact, while most of the benefits are enjoyed by the residents of amalgamated municipalities (or, in case of positive spillovers, by residents in a few neighboring municipalities), incentives paid by national or regional governments are mostly financed by taxpayers not involved in the amalgamation. Therefore, those taxpayers would only suffer from an increase in their tax burden.
Structural reforms of local government based on voluntary amalgamations are difficult to govern and the expected economic impact of such reforms is essentially uncertain because it is difficult to predict which municipalities will actually merge. A number of attempts to solve this problem can be found in Sorensen (2005), who analyzed the Norwegian case, Miyazaki (2013), who considered the situation in Japan, and Dur and Staal (2007), who based their study on a theoretical model. In any case, the authors stressed the importance of the expected efficiency gains and the fact that the...