اثرات تمرکز زدایی بر منابع درسی

Sweden has undertaken major national reforms of its school sector, which, consequently, has been classified as one of the most decentralized ones in the OECD. This paper investigates whether local tax base, grants, and preferences affected local school resources differently as decentralization took place. We find that municipal tax base affects per pupil spending in the same way regardless of whether the school sector is centralized or decentralized, but has a smaller effect on teacher–pupil ratio after the reforms. The less-targeted grants are the fewer teachers per pupil do the municipalities employ. The results for local preferences are less clear-cut.

While the trend in many US states has been to centralize school funding in order to avoid inequalities in school district spending, Sweden has undertaken national reforms to decentralize the responsibility and funding of the school sector. In 1991, the responsibility for compulsory and upper secondary school provision was transferred to the lower level government (municipalities) along with a less centralized system of targeted grants to schooling. In 1993, a major grant reform transformed the system of targeted grants into a general grant system. The latter reform implied a fundamental change of the organization of school funding. In 1996, teacher wages, which until then had been set through central negotiations implemented nationwide, started to be set at the local level. Cross-country comparisons undertaken after the implementation of these reforms rank Sweden as having one of the most decentralized schooling sectors in the OECD (1998). The purpose of this paper is to empirically investigate whether the reforms that decentralized the school sector affected local school resources and, if so, in what way.1 Two complementary measures of school resources are used in the analysis: per pupil spending and the teacher–pupil ratio. The following questions are analyzed: (i) To what extent does the impact of local tax capacity on school resources differ before and after decentralization? Studying the different effects of tax base is clearly a relevant issue since the drawback of decentralization is the risk of creating inequalities in school resources and thereby giving pupils different chances for the future depending on where they attend school. Furthermore, (ii) does it matter in what form (general or targeted) grants are delivered? Whether general and targeted grants have different effects is of great importance for all countries with more than one level of government. Finally, (iii) what is the impact of local preferences, and did it change in connection with the reforms? Adjusting supply to local conditions (economic as well as political) is indeed one of the main theoretical arguments for decentralizing public tasks to lower level governments (see e.g., Oates, 1972).

In this paper, we have analyzed how the decentralization of the Swedish school sector affected the allocation of school resources. We find that local income does not seem to matter more after the decentralization than before for per pupil spending. Hence, even after the cap on tax rates was lifted, tax base had the same effect on spending as before. For the teacher–pupil ratio, on the other hand, the effect of own tax base is smaller after the decentralization than before and tax base does not seem to matter at all after the 1993 grant reform. There is hence no evidence supporting that where you live has become more important for school resources. Turning to the effects of intergovernmental grants, we find that the form the grants arrive in (general or targeted) is important for the teacher–pupil ratio but not for per pupil spending: the less targeted grants are the less will be spent on the number of teachers. We also find that during the period with decreasing grants (1993–1995), grants have no statistically significant effects on either per pupil spending or the teacher–pupil ratio. One explanation for the lack of effects during the period with decreasing grants could be that municipalities react asymmetrically to increases and decreases in grants, something that is also found by Karlsson (2006). Finally, our results for local preferences are less clear cut. The political majority in the municipal council, if anything, matters less after the decentralization than before, which is strange given the low degree of local impact possible before the decentralization.18 One explanation for this might be that municipalities with a right-wing government were unable to increase spending above the minimum standards stipulated by laws, since they were not allowed to increase taxes during most of the period. Females affect per pupil spending more, but the teacher–pupil ratio less after the decentralization, whereas the share with high education does not seem to matter neither before, nor after the decentralization.