Beware of the crash risk: Tail beta and the cross-section of stock returns in China

ABSTRACT

We investigate the pricing of systematic tail risk measured by tail beta in the Chinese equity market. Using an array of tests, we examine the performance of more than 3,300 stocks for the years 1999 through 2018. Contrary to evidence from developed markets, we demonstrate a strong negative relationship between the tail beta and future returns. The effect is robust to many considerations and cannot be explained by established pricing factors or alternative risk or illiquidity measures. We link our findings to specific characteristics of the Chinese stock market.

KEYWORDS:

• systematic tail risk
• tail beta
• extreme value theory
• China
• equity market
• asset pricing
• return predictability
• low-risk anomaly

JEL CLASSIFICATION:

• G12
• G14

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

¹ For robustness, we also examine the data from CSMAR (China Securities Market and Accounting Research) and find no major difference in our results. We opt for Datastream due to better coverage and broad acknowledgment in international studies.

² See van Oordt and Zhou (2016) for details.

³ Dupuis (1999) and Matthys and Beirlant (2000), among others, offer some sophisticated methods for threshold selection. Nevertheless, these typically require estimation of additional parameters and may yield unstable results. In consequence, Gabaix, Gopikrishnan, and Plerou (2006) advocate simply fixing the cut-off point at the 5% level.

⁴ For robustness, we calculate also the factors manually based on our Datastream-based dataset; this operation leads to no qualitative change in results.

⁵ For robustness, we also apply the Fama-MacBeth regressions to the abnormal returns from the three-factor model of Fama and French (1993), as in Avramov, Kaplanski, and Subrahmanyam (2018), among others. This test yields consistent results.

Additional information

Funding

This paper is a part of the project [no. 2016/23/B/HS4/00731] of the National Science Centre of Poland. This project is also funded by the China Postdoctoral Science Foundation [no. 2018M640543].