Bias in Malmquist Index and cost function productivity measurement (and how to correct it)

• Autores: A. Lozano Vivas, David B. Humphrey
• Localización: Documentos de trabajo de análisis económico, Nº 3, 2000, ISBN 84-95191-20-2, págs. 1-28
• Idioma: español
• Texto completo no disponible (Saber más ...)
• Resumen

  • Many current Malmquist Index studies of banking industry productivity are biased, typically overstating the rate of productivity ad-vance. A similar bias exists for banking productivity estimated with a stochastic frontier cost function or a growth accounting approach. The paper attempts to explain how this bias arises in banking studies and show that the majority of measured productivity change in many existing studies is due to this bias and thus does not re¿ect "true" productivity advance. The bias is eliminated when all outputs and inputs are included in the analysis, rather than focusing on only a subset of them. The bias is not due to the technique used but rather in how it is applied. A rst approximation to the "true" value of pro-ductivity using the Malmquist Index was obtained by subtracting an asset/liability index from a computed, but biased, Malmquist Index