Since there is no agreement on the consistency of their estimates, the purpose of this paper is to investigate whether parametric stochastic frontier analysis (SFA) and nonparametric data envelopment analysis (DEA) generate consistent bank efficiency assessments.
The authors utilize four alternative efficiency computation models: two DEA technical efficiency models based on constant and variable returns to scale, and two SFA cost efficiency models employing Translog and Fourier functional specifications. An unbalanced panel of 59 Indian banks over 1990‐2007 is employed as a model, developing country, banking market.
The Translog and Fourier specifications in SFA and the constant and variable returns to scale assumptions in DEA are found to rank and identify “best‐practice” and “worst‐practice” approximately in the same order. The association between DEA efficiency estimates and non‐frontier standard performance measures, however, is mixed and inconclusive. Unlike DEA scores, SFA efficiency assessments were found to be consistent with cost and profit ratios and hence are “believable”.
For regulators and bankers alike, the authors' findings highlight the importance of investigating the consistency of efficiency scores across various research methods. They should ensure that frontier‐based efficiency assessments are not simply “artificial constructs” of models' assumptions/specifications.
This paper extends the existing literature by checking jointly the statistical consistency of both DEA technical efficiency scores and SFA cost efficiency scores. The prior studies focus either on technical efficiency or cost efficiency, but not both. Moreover, as far as the authors are aware, this is the first cross‐methodological validation study to focus on bank efficiency in the context of a developing country banking market.
