本研究旨在運用DEA法衡量台灣地區國際觀光旅館之經營效率，並以Tobit迴歸分析探討影響經營效率之因素，並運用Malmquist生產力指數與其所分解之技術變動指數、效率變動指數探討旅館生產力之變化情形。另由於效率衡量方法可區分為參數法與無參數法，故本研究比較屬參數法的隨機邊界法(SFA)與屬非參數法的DEA法衡量所得旅館經營效率之差異。 以Tobit迴歸分析探討經營效率影響因素之結果顯示，餐飲收入與客房收入之比率、住用率與平均房價對效率值的影響為正向且顯著，餐飲部門坪效與餐飲部門樓地板總面積對效率值的影響則為負向且顯著。此外民國81年至90年之旅館經營效率在不同經營型態、規模別與地區別上存在顯著差異性。就生產力變動而言，在研究期間內全體旅館之生產力提昇8.91﹪，其提昇原因主要為管理技術之成長。連鎖經營型態雖因加入連鎖體系使技術提昇，但在整體資源配置上仍有待加強，致生產力提昇情形略低於非連鎖經營型態。 在DEA法與SFA法之效率衡量比較中， 不論是單一年度或於整個研究期間，兩種方法所得效率值之間皆呈低度相關，顯示不同衡量方法之採用會影響效率估計結果。由於DEA法之效率值於樣本期間內較穩定，因此本研究建議DEA法為估算旅館效率之較佳方法。 The principal objective of this study is to estimate their operation efficiencies of International Tourist Hotels in Taiwan by applying DEA and furthermore to investigate the determinants influencing efficiency by applying Tobit regression. Malmquist Productivity Index and its decomposed indexes including technical change and efficiency change index are used to analyze the change in productivity. Besides, the comparison of efficiencies estimated by DEA and Stochastic Frontier Approach (SFA) is conducted in order to evaluate the appropriate method for the International Tourist Hotel Sector The result from the Tobit regression analysis reveals that the ratio of restaurant income over room income, occupancy rate and average room rate have significantly positive influence on the efficiency while unit value of catering area and total space of catering area significantly negative influence. Over the period 1992-2001, the disparity of efficiencies among various hotels is found with respect to business type, hotel size and location. 8.91% increase in productivity mainly due to the improvement in managerial technique is estimated over the period under study, 8.91%. The increase in productivity for the chain business type is slightly lower than the independent business type. No matter in term of single year or the whole period, the efficiencies estimated from DEA and SFA show the insignificant correlation. It indicates that different methods lead to different results. Due to its stability of the efficiencies estimated from DEA, the study suggests DEA is the more appropriate approach for efficiency estimation of the International Tourist Hotel Sector.
