Unveiling black-box of bank efficiency : An adaptive network data envelopment analysis approach

Purpose This paper aims to examine the efficiency of Islamic vs conventional banks in Malaysia by unveiling the traditional efficiency concept – black box – with a three-stage network structure of bank operations. Design/methodology/approach This paper applies data envelopment analysis (DEA) for examining bank efficiency. An adaptive three-step network DEA (NDEA) model is demonstrated for redefining the traditional black box of banking operations. Slack-based variable returns to scale approach is used. Data from all 43 commercial banks in Malaysia are examined over a six-year study period (2010-2015). Inputs and outputs of the model are selected based on CAMELS rating. Undesired output is also considered in time of examining bank efficiency in Malaysia. Findings The empirical results of this study signify that only a few banks in Malaysia have been performing well in converting deposits and equities into profit as well as minimizing loan loss provisions. Islamic banks in Malaysia have performed better both in production (converting deposits and equities into earning assets) and profitability (converting loans into net income). Conventional banks, however, have over scored in intermediation (converting earning assets into loans). Originality/value An adaptive NDEA approach proposed in this paper defines the core banking process instead of traditional approaches in examining bank efficiency based on individual functions (nodes in the network model). This approach has proven to provide better benchmark capacity.