Multi criteria efficiency analysis of fuel stations

Abstract

As the oil prices increase, the oil products distribution sector becomes more attractive in Turkey. In recent years, new distributor companies entered the market and the number of gas stations increased significantly, while it is known that rapidly expanding industries experience “overshoot and collapse” behavior in the middle run. In this study, a multi-criteria efficiency analysis of gas stations of a distributor firm is conducted. Two different methods, the Data envelopment Analysis (DEA) and the Stochastic Production Frontiers (SPF) are used for this purpose. The CCR, the BCC models of the DEA, with and without Assurance Region (AR) extensions are applied. Results of the DEA studies suggest that, after a threshold level of physical resources, scale of station does not provide competitive advantage and better performance depends on management practices. Stations are grouped by cities and a considerable difference in average efficiency scores is observed among different cities. On the other hand, best performing cities coincide in the CCR and the BCC formulations to a great extent, whereas efficiency scores of inefficient units may change dramatically. The AR restrictions have different effects on different stations; some having higher efficiency scores and some having lower, while the statement about scale is still valid and the ranking of cities change considerably. Correlation of efficiency scores to input and output factors are also analyzed. In the SPF study, different outputs are aggregated into one single output. In this case, results differ considerably from results of the DEA study, especially when the BCC formulation of the DEA (which allows variable returns to scale), is considered. On the overall, results confirm usefulness of the DEA in multi output problems and provide useful information about performances of gas stations, which is especially valuable when investing in a new station.