Propensity Score Matching and Treatment Effects of Technology Purchase

• Main Authors: Yu-Han Chan, 詹雨涵
• Other Authors: 張嘉玲
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/15931541336009536078

碩士 === 國立中興大學 === 應用經濟學系所 === 102 === Innovation activities are critical drivers of technological advancement. In addition to Research and Development (R&D), technology purchase is another important way of investment in innovation activities. To explore the treatment effects of purchasing foreign technologies, this study used the propensity score matching method to match manufacturers in Taiwan. The propensity score (PS) was estimated based on the adoption decisions of firms, and the counterfactual effects were built to analyze the treatment effects of technology purchase on labor productivity and output. The objective of this study was to explore differences in treatment effects between firms that had purchased foreign technologies and firms that had not. Data were the sampling data of the manufacturing industry extracted from DGBAS' 2006 Industry, Commerce and Service Census. Differences in four-digit and two-digit industry data were analyzed. Empirical results show that propensity to purchase foreign technologies is negatively related to firm size but positively related to proportion of technical employees, foreign investments, industry purchasing intensity, and area purchasing intensity. As for the treatment effects, the analysis of four-digit industry data shows that for labor productivity, estimates of average treatment effect (ATE) range between 15.848 and 17.102, and estimates of average treatment effect on the treated (ATT) range between 3.987 and 4.556; for output, ATE estimates range between 1383.683 and 4810.913, and ATT stands at 1,436.978. These statistics indicate that purchasing foreign technologies can lead to improvements in labor productivity and output. Besides, the analysis of two-digit industry data shows that the treatment effects of purchasing foreign technologies on labor productivity and output were positive for most of the industries except the apparel and clothing accessories industry and the beverage industry.