عنوان مقاله [English]
Financial data and ratios have been commonly used in data envelopment analysis (DEA) models in order to produce a unified measure of performance. However, several researchers have indicated that the use of financial ratios in DEA models creates biased efficiency estimates in firms performance evaluation. Namely, the efficiency scores may be influenced by the choice of the ratios adopted, and units could be erroneously classified as efficient or inefficient. Thus, the research problem that has been posed is how to perform performance evaluation with the use of the DEA methodology producing unbiased results. This paper answers the research problem by bootstrap technique. The main proposed hypothesis is that bootstrapping of the DEA efficiency scores can remove the sensitivity of the efficiency scores relevant to the financial ratios used in performance evaluation and avoid errors in efficiency estimation. To test the research hypothesis and examine the sensitivity of the efficiency scores relevant to the financial ratios used, seven different DEA models have been created. The idea behind every model is to test whether the efficiency scores are sensitive to the financial ratios used in our analysis. In order to implement the proposed models, we have used data from audited financial statements of 20 insurance companies in Iran that are listed in the Tehran stock exchange (TSE). The sample period is one fiscal year from march 2010 to march 2011.According to the literature on the performance measurement of the insurance companies, three inputs have been used in our analysis, namely total assets, equity, and insurance expenses. Moreover, three financial ratios (profitability ratios) have been used as outputs in order to capture the performance of the companies. These are the ratios of underwriting profit to average assets, the investment profit to average investment, and the directly-written premiums to operating expenses. The results reveal that the bootstrap technique can resolve bias problem in efficiencies. So, efficiency scores will have less sensitivity relevant to the choice of the financial ratios or data adopted.