Deng 和Huang (2008) 延伸了Fan et al. (1996)所提出的橫斷面資料下的半參數隨機生產邊界模型為縱橫資料下的時間變量(time-variant)半參數隨機邊界模型，其方法容許技術效率隨時間改變。本論文參考Deng 和Huang (2008)的方法，探討的是縱橫資料下時間不變量(time-invariant)半參數隨機邊界模型，文中假定的技術效率是不會隨著時間的改變而改變，並且利用無母數迴歸的技巧來建構模型組合誤差項未知參數的估計量。經由蒙地卡羅模擬研究結果發現，本文所提出的估計方法是具有一致性(consistent)，因此適合應用在隨機邊界生產函數的估計。最後採用了96間台灣電子產業公司的資料來進行實證分析。 Deng and Haung (2008) extend the semiparametric stochastic frontier model of Fan et al. (1996). A semiparametric estimation procedure suitable for the case when panel data are available is proposed and time varying technical efficiency allowed therein. This dissertation considers the extended model with panel data and time-invariant technical efficiency. The proposed semiparametric estimation procedure is similar to that of Deng and Huang (2008) and also provides consistent estimators of the model parameters as demonstrated by the Monte Carlo simulations. An empirical application of the proposed procedure, which employs the panel data from 96 Taiwanese electronic firms over the period 1996-2001, is carried out.
