Measuring efficiency of Vietnamese banks : Accounting for nonperforming loans in a single-step stochastic cost frontier analysis

Purpose This paper aims to examine alternative methods for treating nonperforming loans (NPLs) in bank cost-efficiency studies using stochastic frontier analysis (SFA). Design/methodology/approach The authors consider three methods of treating NPLs in SFA: as an additional control variable, as an environmental factor or as a deduction from total loans. Using data from the Vietnamese banking system (2003-2010), the authors then compare these results with those of the base model (where total loans is used regardless of the NPLs) to see which one is more appropriate for this study. Findings The authors observed that the first two methods are inappropriate for the analysis: one cannot find the significant relationship between NPLs and the banks’ total cost, and the other cannot account for any inefficiency at all. The authors suggested that the third method of separating NPLs from total loans can provide better insights. Using the proposed method, the authors showed that the cost-efficiency of Vietnamese banks over the period examined was moderate with a slight decreasing trend. When NPLs are separated, the cost-efficiency decreases in state-owned banks and big banks, whereas it increases in small and private banks. Research limitations/implications Research is limited to Vietnamese banks during a certain period, and it would be useful to apply the same technique to other data sets. Practical implications The paper suggests a new approach to account for NPLs in cost SFA studies in banking. Originality/value The paper provides a much more searching analysis of NPLs in banking than has generally been seen in previous research.