Identifying and Measuring Technical Inefficiency Factors:Evidence from Unbalanced Panel Data for Thai Listed Manufacturing Enterprises

This study employs stochastic frontier analysis (SFA) and two-stage DEA approaches to predict firm technical efficiency and analyse an inefficiency effects model. Aggregate translog stochastic frontier production functions are estimated under the SFA approach using an unbalanced panel data of 178 Thai manufacturing enterprises listed in the Stock Exchange of Thailand (SET), covering the period 2000 to 2008. The maximum-likelihood Tobit model is used to conduct the second-stage of the two-stage DEA model to investigate the relationship between technical inefficiency and environmental variables. Both parametric and nonparametric approaches are found to produce consistent results. The empirical evidence from both approaches highlight that Thai listed manufacturing firms had been operating under decreasing returns to scale over the period 2000 to 2008. The SFA approach reports that technical progress decreased over time, and relied on labour input. Both estimation approaches suggest that leverage (financial constraints), executive remuneration, managerial ownership, exports, some types of listed firms (i.e., family-owned firm and foreign-owned firm), and firm size have a negative (positive) and significant effect on technical inefficiency (technical efficiency). The empirical results obtained from both approaches also suggest that liquidity, external financing, and research & development (R&D) have a significantly positive (negative) effect on technical inefficiency (technical efficiency)