New Approach in Stochastic Frontier Analysis Estimation for Addressing Joint Assumption Violation of Heteroscedasticity and Multicollinearity

Efficiency analysis in production units has long been a key area of interest in economics, particularly with the development of methodologies like Stochastic Frontier Analysis (SFA). Originating from seminal works by Aigner & Cain (1977) and Meeusen & Van Den Broeck (1977), SFA has been instrumental in assessing the efficiency of entities by separating technical inefficiency from random production fluctuations. Despite its widespread application, the SFA model faces challenges, especially when underlying assumptions such as multicollinearity and heteroscedasticity are violated. This study introduces a novel estimator called "Weighted Principal Component Analysis Estimation for Stochastic Frontier Analysis" (WPCA-SFA), which combines the methodologies of weighted least square estimation (WLS) and principal component analysis (PCA) to jointly address these assumption violations. Monte Carlo simulation experiments were conducted to evaluate the performance of the proposed estimator. The results demonstrate that the WPCA-SFA estimator significantly outperforms the standard SFA model by effectively mitigating the adverse effects of both heteroscedasticity and multicollinearity. Based on the findings, the study recommend that researchers and practitioners in the field of efficiency analysis consider employing the WPCA-SFA estimator, particularly in scenarios where multicollinearity and heteroscedasticity are likely to compromise the accuracy of parameter estimations. Neglecting these issues can lead to suboptimal results, whereas the WPCA-SFA has proven to provide more reliable and accurate predictions. This advanced correction methodology should be adopted to enhance the robustness of empirical analyses in stochastic frontier models.