本篇論文利用一隨機邊界模型調查五十五個美國產業類別 (分類標準依據 2002 North American Industry Classification System) 的技術效率, 同時估計一隨機邊界生產函數與無效率因素分析。研究期間為1997年到2011年, 總共15年。 與現有應用隨機邊界模型的研究文獻不同的是, 在現有研究與文獻中並沒有廣泛討論到未能觀察到的衝擊 (unobserved common shocks) 所產生之影響,若忽略這些影響將使得估計結果產生偏誤且不ㄧ致的現象。 本文的計量模型考慮並且控制這些影響進行估計並且得出估計果, 除估計出隨機生產逼界函數與無效率因子分析式之外, 也依據所估計出的效率值將五十五個美國產業類別的技術效率表現進行排序, 並且檢視跨橫斷面平均技術效率隨十五年間的變動情形。 本篇論文的研究結果指出一個重要訊息: 美國的主要產業跨橫斷面平均技術效率在1997到2011年間呈現出一波動且逐漸下降的趨勢。 除此之外, 本研究亦發現美國主要產業跨橫斷面平均技術效率的波動型態與美國實質經濟成長率變動似乎有著特定的關聯, 兩數列的波動型態非常一致但有著時間先後的差異。 對於此現象的一個預先推論是廠商對於未來前景的預期將會影響其對於調整技術效率的決策。 This paper uses an unconventional panel stochastic frontier model to investigate the technical efficiency of 55 industries in the United States during 1997-2011. Unlike the existing research adopting the stochastic frontier model, this paper takes account of and controls the common correlated effect caused by unobserved common shocks when estimating the panel stochastic frontier model. The parameters in the model and the indicators of technical efficiency are obtained through estimation the results estimated by conventional techniques are also provided as a comparison. Based on the results, the technical efficiency of the 55 industries and the overall technical efficiency are examined. The technical efficiency ranking of the 55 industries is also given, and their overall technical efficiency reveals important information: the overall technical efficiency of these industries has a decreasing trend and fluctuating movement. Additionally, it is found that the overall technical efficiency movement seems to have a specific relation to the real GDP growth rate in the United States. A preliminary explanation of this finding is based on future expectations and the decisions made by firms in these industries for adjusting their technical efficiency.
