Dynamic Analysis of Firm’s Survival、Exit and Switch：Application of Survival Analysis

• Main Authors: Ken-Chung Chiu, 邱耿中
• Other Authors: Hui-Lin Lin
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/31088374561828764217

碩士 === 國立臺灣大學法教分處 === 經濟學研究所 === 91 === The purposes of the thesis are to study the exit and switch rate of Taiwanese Electronics Firms. On the basis of data which comes from the Survey of Ministry of Economic Affairs during the period of 1992-1999, we applied the survival analysis to investigate the determinants of exit and switch behavior. The empirical results show that firm’s duration models follow approximately loglogistic distribution. That means the probability of exit and switch increases for certain time before declining. In addition, we consider firms may never shut down eventually. Thus we applied a split population duration model in which the probability of eventual failure is less than one. For analyzing the difference of behavior of exit and switch , we use three different samples:“exit”, ”switch” and “exit & switch” to empirically study firm- and industry- specific characteristics on the exit and switch rate. The empirical results are summarized as follows. (1) For both exit and switch rate, If R&D expenditure increases, then exit and switch rate will be decreased. Displacement effect holds by the variable, last year entry rate. Moreover, The exit and switch rate of new firms is much higher than that of incumbents. Other firm -and industry- factors, such as firm profit ratio, industry growth etc, show the similar evidence as past literature. (2) Two factors, firm size and export proportion affect exit and switch rate in different ways. We find that the larger the firm’s size, the lower the exit rate, but switch rate will be higher. The more the export proportion, the lower the exit rate, but switch rate will be higher. (3) We use a split-population duration model to examine the determinants of firm survival and firm exit & switch hazard rate. The empirical finding indicates that the firm failure to allow less than one represents an important misspecification with underestimate of the exit or switch rate and biased implication for inference on the factors of exit or switch rate.