Impact of Subjective Well-being on Unwinding of Taiwan Stock Index Futures

• Main Authors: Hsing-Kuo Lai, 賴興國
• Other Authors: Shean-Bii Chiu
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/83599034597216563337

碩士 === 臺灣大學 === 財務金融學研究所 === 98 === In addition to the disposition effect of behavioral finance, the subjective well-being (SWB) in positive psychology significantly influences unwinding decision. We conclude that the positive events of higher returns, price volatility, and the uptrend market bring investors the higher SWB and lead investors less inclined to unwind. Bullish investors are divided into two groups– the “double-good”, the investors earn profit in the upward market, and the “double-bad” as losing money during the downtrend market, and the special finding is with sensitivity analysis, the “double-good” investors imply higher SWB so as making less unwinding decision than the “double-bad” ones. Furthermore, hazard rate of the former is 18.3% less than the latter, which is completely contrary to the ratiocination of the disposition effect. The data during the most prosperous trading period in Taiwan futures market from 2007/1/1 to 2008/6/30 is adopted. There are 18,444,489 transactions in Taiwan stock index futures and small-scale ones by retail investors. We cope with the left, right and random censoring data as setting the research period in 40-trading-days just before the maturity date of each product; it also provides three advantages as (1) focusing on most exuberant products (2) more likely to examine the effect of the SWB comparing with the limitation of the disposition effect (3) less influenced by personalities as examining the variance of short-term SWB which is mainly affected by the externalities. There is no literature introducing SWB into investment decisions, but we contribute on proving nine hypotheses showing the influential of unwinding decision by SWB through testing the validity of explanatory variables by Cox proportional hazard model.