| The Return Predictability of Foreign Investors` Trading from Tax Havens in the Korean KOSPI Market |
|
Cheol Won Yang |
| 조세피난처 외국인 거래의 주가예측력 |
|
양철원 |
|
|
|
|
| Abstract |
|
Tax haven provides investors anonymity. It means that foreign trades from tax havens may contain substantial private information. This paper investigates whether the foreign trades from tax havens predict future stock returns in the Korean KOSPI market. The methodologies of portfolio approach and regression are employed for empirical analyses. When stocks are sorted into quintile by net purchases of tax havens investors scaled by daily dollar trading volume, the portfolio with higher ratio performs better than that with lower ratio. The hedge portfolio that shortsells the lowest quintile stocks and buy the highest has significant return of 0.28% per day (6.2% per month). The results are also significant after the portfolio returns are risk-adjusted by using the CAPM and Fama and French (1993) three factor model. When the portfolios are double-sorted by firm characteristics and net purchase from tax havens, the predictability is concentrated on the firms with small size, high book-to-market ratio, low foreign ownership, and high bid-ask spread. The result of regression also shows that future stock returns are positively related to the net purchase of tax haven investors. These results above consistently show that the foreign trades from tax havens predict future stock returns in Korea and foreign investors from tax havens are informed traders. One of plausible explanations for their return predictability is that the foreign investors from tax havens are insiders of Korean firms and use their insider information by trading Korean stocks. Therefore, the empirical results of the paper suggest that the Korean financial supervisors should oversee the trades from tax havens closely and design a related new policy. |
| Key Words:
외국인 거래,조세피난처,주가예측력,포트폴리오 접근법,회귀분석,Foreign Trades,Portfolio Approach,Regression,Return Predictability,Tax Haven |
|
PDF Links
Full text via DOI
Download Citation
