engleski

Assessment of regional efficiency in Croatia

Achieving balanced national development and reducing interregional socio-economic disparities is a comprehensive economic challenge. In this study, regional efficiency of Croatian counties is measured and compared over the three-year period (2005-2007) using Data Envelopment Analysis (DEA). The set of inputs and outputs consists of seven indicators suitable for measuring achieved level of regional socio-economic development. Analysis is carried out using models with assumption of variable returns-to-scale, under both input and output orientation. DEA identifies efficient counties as examples of good operating practices (benchmark members) and inefficient counties that are analysed in detail to determine not only the sources and but also the amounts of their inefficiency in each source. To enable proper longitudinal monitoring of development dynamics and make conclusions on behaviour of the county (whether its efficiency has improved, deteriorated or stagnated), window analysis is applied. Based on the results, guidelines for implementing necessary improvements to achieve efficiency are given. Analysis reveals great disparities among counties, regardless if each year is considered separately or compared with the other two years, and the majority of the results suggest that on average all the counties have some degree of inefficiency. In order to alleviate naturally, historically and politically conditioned unequal county positions that affect regional efficiency but over which economic policy makers do not have total control, categorical approach is introduced as an extension to the basic DEA models. This approach, combined with window analysis, changes relations among efficiency scores in favour of continental counties. By identifying regional disparity determinants, the results obtained using an improved DEA model should assist policy makers in formulating the most beneficial regional policy decisions.

data envelopment analysis ; regional efficiency ; county ; window analysis ; categorical variables

nije evidentirano