Efficiency levels of sub-national governments: A comparison of SFA and DEA estimations

Purpose - By utilizing the two most commonly used approaches to generate best practice frontier to estimate efficiency of observed units, the purpose of this research paper is to estimate technical efficiency for total population of 200 Slovenian municipalities for the 2011 fiscal year. Design/methodology/approach - Stochastic frontier analysis (SFA) and data envelopment analysis (DEA) methods are used to estimate technical efficiency levels. Namely, the majority of studies have utilized these two traditional approaches. Since the advantages of one method often represent the disadvantages of the other method, the two methods have been selected to compare the results obtained on the technical efficiency levels. Findings - The results suggest that mean technical inefficiency should be approximately 22-25 percent (SFA method), whereas DEA method suggests the inefficiency in the range 12-18 percent. The DEA approach also suggests that the paper has many more technically efficient units compared to the SFA estimates. Nevertheless, the SFA assessment has revealed that, although on average the inefficiency should be larger compared to the DEA assessment, more than one-third of municipalities should exhibit relatively low levels of inefficiency (less than 5 percent). Originality/value - This study utilizes both parametric as well as non-parametric approaches to assess the technical efficiency, which is not very common in the empirical literature. Besides, it focusses on the local government efficiency in a post-socialist country. © Emerald Group Publishing Limited.