This paper uses stochastic frontier analysis (SFA) to study 12 international airports in the Asia-Pacific region based on data from the period 1998-2006. The SFA model uses a translog production function that is more flexible than a Cobb-Douglas one that is popularly used in the literature. As an extension of a previous study, this paper mainly discusses the implications of the parameter ... [Show full abstract] estimates of translog function. The use of a translog function is founded on testing the statistical hypothesis. According to the preliminary results, we suggest that airports should emphasis more on investment than on human resources. In addition, we find that the inefficiency in the SFA model largely results from technical inefficiency, and the inefficiency effects are time-varying and decrease over time. By using the model including environmental factors to explain the inefficiency, we suggest that the inefficiency would be less if the airport is more privatized.