When policies and plans are employed to influence the structure of manufacturing industries, as is the case in Bangladesh, knowledge of the technical efficiency of firms by size is critical for formulating policies. No other econometric methodology is more suited for discerning differences in technical efficiency at the micro level, in particular as between small and large enterprises, than the translog stochastic frontier production function. In the present paper, the indicated flexible production functions are applied to 17 four-digit manufacturing industries in Bangladesh, using the firm-level panel data of the CMI Merge File. By virtue of the availability of panel data (rare for enterprises), technical efficiency.

Empirical results are quite striking: Fifteen of the 30 industries (17 studied by the TL function and 13 by the CD function, the latter because of shortage of observations) experienced no significant productivity change over the 12-year period. Only 5 experienced significant acceleration in TFP change. Several suffered from deceleration. The overall picture that emerges is one of stagnation in productivity. Technical efficiency of small firms is lower than large ones. An exercise to explaining the inter-firm differences and temporal variations in technical efficiency in terms of scale of the enterprise, skill composition of the labor force, capital intensity, and material intensity failed because the explanatory power of the regressions is found very low and the explanatory variables have insignificant coefficients for most of the industries.