The Three Links policy and stock returns

Abstract

This study examines whether the Three Links (3L) policy affected stock returns before, during and after the 2008 Taiwan leadership election. The evidence shows that companies in the industries benefitting from (being damaged by) the 3L policy earned positive (negative) cumulative abnormal returns. Although the political party connection/affiliation seems to have had an impact as well, it was not persistent. When both factors – the policy effect and the political connection effect – are considered together, we find that the policy effect persisted during the election, while the political connection effect did not. It is obvious that the policy effect is a powerful factor explaining changes of stock returns during the 2008 leadership election in China’s Taiwan.

Keywords:

• The Three Links (3L) policy
• political connection
• cumulative abnormal returns

Notes

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Also see the studies of Di, Luft, and Zhong (2019) and Qiu, Zhan, and Wei (2019) for the discussion of the trade war between the US and China.

2 This study uses the 3L policy to examine the impact of policies, which is different from the setting of Lin et al. (2016) and Shen et al. (2017) that use different policy measures.

3 Also see the studies of Ashraf and Shen (2019), Datta, Doan, and Iskandar-Datta (2019), Xu (2020), Papadimitri et al. (2021) for the policy effect.

4 For example, air travel from Shanghai to Taipei could reduce over an hour and reduce 8 hours for the sea travel. In addition, China’s Hong Kong is estimated to lose one million travellers from China’s Taiwan annually because of the 3L policy.

5 For example, in the 12 months before the election, the average monthly growth ratio of all imports from the Chinese mainland is 1.88%. However, the average monthly growth ratio of imports from the Chinese mainland in the Communications and Internet industry is 6.26%, confirming that firms in this industry highly compete with firms of the Chinese mainland.

6 Also see the studies of Fisman (2001), Faccio (2006), Knight (2006), Claessens, Feijen, and Laeven (2008), Faccio and Parsley (2009) and Cooper, Gulen, and Ovtchinnikov (2010). We collect data on the political tendency from the TWSE, political party websites, the Market Observation Post System (MOPS) and the Central News Agency’s United Daily News. We also collect data on political donations from the Control Yuan.

7 In addition, we present the correlation coefficient matrix for the variables in Appendix. The correlation coefficients between CAR (−40, 5) and other variables are less than −0.533, indicating that multicollinearity is not a concern.