This study estimates municipal efficiency and economies of scale of municipalities in Bosnia and Herzegovina by employing data envelopment analysis (DEA) with variable (VRS) and constant (CRS) returns to scale. The results indicate low overall municipal efficiency, with economies of scale reached in very few municipalities. The average municipal efficiency score is 0.7115 under DEA VRS assumption, where only 16% of municipalities are found efficient. The average scale efficiency is 0.7458 with full scale efficiency reached by only 11% of municipalities in Bosnia and Herzegovina. Furthermore, the analysis shows strong positive impact of number of inhabitants on overall municipal efficiency. Politically motivated fragmentation of municipalities, aiming to bring peace and stability to the country, did not go hand in hand with improved economic efficiency.

Dayton Peace Agreement (DPA) has resulted in asymmetric decentralization in Bosnia and Herzegovina (BiH) in all aspects: political, administrative and fiscal. Decentralization driven by non-economic reasons is rather usual concept, and often the main reasons are political, ethnical or overall country stability. Decentralization model implemented in BiH became huge obstacle not only in reaching efficiency in provision of public services, but also to further economic development. While the purpose of DPA was to stop armed conflict, and while valuable back in time, there is no excuse to keep these solutions for more than 20 years. This paper aims to provide deep insight into experience of selected comparative countries where non-economic reasons initiated decentralization. In addition, it aims to identify patterns and features of administrative, fiscal and political arrangements that perform better in the environment similar to BiH. Analysis of the experience in developed countries identified different models in organizing ethnically divided societies and establishing different forms of cooperation between sub-national government units to increase efficiency. Transitional countries experience shows mixed results in terms of positive effects of decentralization on overall efficiency and citizens' well-being, but there is valuable experience and number of features, which may improve municipal efficiency in BiH as well. Having in mind very limited literature focused on specific BiH context as well as the need to improve efficiency at local community level, this paper takes an important first step in this direction by providing a systematic review of decentralization design in countries that had similar challenges as BiH. The focus of comparative analysis is on the administrative decentralization (territorial organization and responsibility designation), political decentralization (addressing democratic principles) and fiscal independency. Paper has identified certain mechanisms that do not require any or require minor changes in core legislation introduced by DPA. These primarily include activation of cooperation mechanisms already allowed by law as well as improving system of revenue and grant allocation. Democratization still did not reached proper level as mechanisms introduced by DPA do not address rights of minorities, and this has to be changed. Improving municipal efficiency in BiH by applying experience of developed and transitional countries therefore may range from better cooperation according to the existing laws, to substantive changes of legislation.

Abstract This paper uses a DEA-VRS methodology for the estimation of municipal efficiency to provide empirical evidence of the impact of decentralization in BiH under the Dayton Peace Agreement (DPA) on overall municipal efficiency. In particular, the paper focuses on the analysis of the overall efficiency of 33 new municipalities established under DPA. The findings suggest that the average municipal efficiency in BiH is rather low, and only 23 or 16% of municipalities in BiH are efficient. The average efficiency achieved is around 0.71. This means that with the same level of inputs (budget revenue) outputs may be increased by almost 30%, on average. The results of DEA-VRS efficiency estimation suggest that new municipalities have lower average efficiency (0.60) in comparison to “older” municipalities (0.74). In our sample, only six percent of newly created municipalities are efficient, 12% exhibit some level of efficiency, while the remaining 82% are inefficient, with significant share (39%) of very inefficient municipalities