Size Still Matters!

Wenyun Shi,Yexiao Xu

DOI: https://doi.org/10.2139/ssrn.2486002

2014-01-01

SSRN Electronic Journal

Abstract:Recent studies have challenged the cross-sectional explanatory power of the size variable for stock returns. In this paper, we reexamine the size effect by disentangling the delisting effect at the same time. We believe that firms to be delisted from their current stock exchanges usually experience some fundamental changes in financing or structures, which should introduce certain abnormal components into returns right before the delisting date. Therefore, the risk-return relation during this period would be very different from the ordinary patterns. In particular, we find that when a firm is close to be delisted from its current stock exchange, its returns are positively related to firm size, and this opposite effect offsets the original negative size effect. By excluding the delisting period, we are able to restore the negative size effect in a rather robust way. Additionally, we reconstruct the commonly used SMB, HML and UMD proxies without stocks that are in their delisting process, and show a significant improvement in the time-series fit over the existing proxies. Overall, we illustrate that size effect is still approximating for certain risk factors in empirical studies. However, there does exist a very different risk-return behavior during the delisting period, which should be reasonably distinguished from other observations.