Using spatial econometrics to analyse local growth in Sweden
This paper investigates factors that determine the average income growth and net migration rates in Swedish municipalities during the period 1981 to 1999. The main issue is to test the hypothesis that, conditional on a set of other possible determinants of regional growth, the growth rate in one municipality is affected by the growth rates in its neighboring municipalities. We also test the hypothesis of conditional convergence, that is, the hypothesis that initially 'poorer' regions tend to
... regions tend to grow faster than initially 'richer' regions conditional on the other explanatory variables in the model. We find a positive correlation between net migration rates in neighboring municipalities, which suggests that net migration tend to 'spill over' to neighboring municipalities. When it comes to average income growth, our results indicate spatial dependence in the error terms during the 1980's. Such dependence is important in the sense that it indicates that shocks into the system not only affect the municipality where the shock has its origin but spread across the country. In addition, and in contrast to previous empirical findings based on Swedish data, we do not find any clear evidence in favour of the hypothesis of conditional convergence. Instead, our results predict conditional divergence between municipalities located in the Stockholm region throughout the period and also for municipalities outside the Stockholm region during the 1990's. K eyw ords: Spatial effects, conditional convergence, regional growth, net migration. JEL classifi cation: C31, R11, R58. . † Special thanks to Sofia Lundberg and Krister Sandberg for their very valuable comments on previous versions of this paper. This research has been financed by the research project Urban Design and STINT. Using Spatial Econometrics to Analyze Local Growth in Sweden 2 base, and also the local governments ability to uphold national standards in the provision of local public services, depends largely on the average income level within the municipality and the extent to which the municipality is attractive to high-income individuals. Considering that the average private income levels and net migration rates vary quite much between municipalities and that the expansion of the local public sector mainly has been driven by political decisions made at the national level of government, the national government has felt a need to compensate regions with relatively small tax bases. This has been accomplished either by subsidizing the local government or the local private sector. A typical example of the former policies is the grant-in-aid system and location subsidies to the industrial sector are an example of the latter. The national government has also tried to affect local 'conditions' via investments in the infrastructure and the location of new universities, university colleges and government authorities. This paper concerns the dynamics of the local tax base in Sweden during the period 1981 to 1999. Following Aronsson et al. (2001) and Lundberg (2003) , the growth rate of the local tax base is decomposed into two components, the average income growth and net migration rates. The main issue is to test the hypothesis that, conditional on the other explanatory variables, the average income growth and net migration rates in one municipality 'spillover' and affect the growth rates in neighboring municipalities. Such spatial externalities may exist for different reasons. For instance, if one municipality (i) is highly attractive to migrants, this could also have a positive effect on the net migration rate in neighboring municipalities if this causes housing prices to increase in municipality i. That is, some individuals who would like to live in municipality i settle for one of municipality i's neighbors due to high housing prices in municipality i. On the other hand, high net migration rates for municipality i may also be at its neighbors expense if neighboring municipalities are regarded as relatively 'unattractive' to migrants making residents within these municipalities to migrate to municipality i. When it comes to average income growth, one reason for spatial externalities relate to private consumption. As the average income level increase within one region, some of this money may be spent in neighboring regions, which might have a positive effect on the economic activity and the average income levels within these regions. Another reason relate to 'spin-offs' from existing companies. This could either be as new businesses started up by former employees, subcontractors to existing companies, or other types of service providers for company employees or companies. Such businesses may locate in neighboring municipalities, which could have effects on the average income level where they locate.