Evaluating the relative efficiency of banks in Malaysia: a hybrid DEA-AR/AHP approach / Nur Azlina Abd Aziz ... [et al.]

One of the most broadly used non-parametric methodsfor calculating the effectiveness boundary is Data Envelopment Analysis (DEA), a mathematical optimization approach. In comparison to other institutions in the sector, DEA simply proves which ones are operating extra proficiently. Plus, DEA can highlight the poorly performing area that needs to be improved (Ar & Kurtaran, 2013). As a result, most financial institutions used this method to determine the relative efficiency which allows individual bank performance to be measured and compared to other banks. However, there are several disadvantages when using DEA applications. Unrealistic weight and shortage discrimination power dispersion are two interrelated issues that have long been recognized (Bal et aL, 2010). When the amount of DMUs is small in comparison to variables, this common DEA problem becomes even more complicated (Despotis, 2002). When a DMU has a zero or extreme value, it is frequently evaluated as productive. Thus, the unrealistic weight dispersion problem occurs. The shortage of discriminating power arises when the amount of DMUs under review is insufficient compared to the total number of variables. In other words, DEA generally evaluates many inputs and outputs as 1 which indicates that the unit is efficient. The aim of this study is to measure the efficiency of the 20 commercial banks in Malaysia by referring to the financial annual report from each bank from 2016 to 2020. This study uses secondary data which was obtained from Bursa Malaysia.