Probabilistic and Deterministic Models in Measuring Technical Efficiency

Abstract

The main aim of the present study is to derive the Technical Efficiency of the probabilistic Stochastic Frontier Production Models viz., Normal Half-Normal Stochastic Frontier Production Model (NHNSFPM), Normal Exponential Stochastic Frontier Production Model (NESFPM), Normal Truncated Skewed Laplace Stochastic Frontier Production Model (NTNSLSFPM) and the deterministic Linear Programming Data Envelopment Analysis model. newlineThe following distributional assumptions are considered in the derivation of Stochastic Frontier Production Models newline(i) and#119907;and#119894;~and#119894;.and#119894;.and#119889;.and#119873;(0,and#120590;and#119907;2) newline(ii) and#119906;and#119894;~and#119894;.and#119894;.and#119889;.and#119873;+(0,and#120590;and#119906;2), that is as non-negative Half- Normal and Exponential and Normal Truncated Skewed Laplace. newline(iii) and#119907;and#119894; and and#119906;and#119894; are distributed independently of each other and of the regressors. newlineThe parameters of all the technical efficiency estimates of the Stochastic Frontier Production Models are estimated using the method of Maximum Likelihood Estimation (MLE). newlineAs a case study, the present study incorporates Stochastic Frontier Analysis and Data Envelopment Analysis techniques to measure the technical efficiency of the three sectors of schools namely government, private and aided schools at their secondary and higher secondary levels on the basis of their performance and views on the Mathematics and Science subjects and school related factors viz., student-teacher ratio, socio-economic status, syllabus, learning disability, teaching related factors and school facilities. The primary data was collected from all the three sectors of schools in Coimbatore and Tiruppur districts, Tamil Nadu. The study also identifies the most technically efficient sector of school and the key variable affecting the efficiency of the schools among the seven input variables considered for the study. The thesis also presents the results of comparative studies on the basis of the technical efficiency estimates obtained from the probabilistic and deterministic techniques which can be categorized as follows: newline(i) Comparison among the models under study