monash_63207.pdf (1.28 MB)
The market reaction to stock splits and the ability to earn abnormal returns
thesisposted on 2017-02-02, 02:36 authored by Nguyen, Phuong Anh
A stock split is often regarded as a pure cosmetic accounting treatment and yet prior research shows that the market reacts positively upon the arrival of the split announcement. However, up to now, there has not been any convincing explanation for this favourable response while there is intense debate amongst researchers about whether these positive abnormal returns persist in the future. We revisit the issues related to the performance of splitting companies both around and following the announcement date. This allows us to study the information content of the event and assess whether the market has incorporated the implication of such information in a timely manner. In addition, we hope to draw meaningful inference about the profitability of trading following the announcement date. Our findings suggest that there is information in the split announcements, which is positively valued by the market. However, abnormal returns cannot be earned with certainty following the event. This is evident in both the option market and the stock market. Specifically, if informed investors use the option market to trade on their information, then our results indicate that informed investors do not believe in the success of a strategy that buys splitting companies subsequent to the announcement date. This is because the post-split announcement drift does not exist following every split; it is conditioned on whether the firms will split again in the future. While prior studies argue that the long-run abnormal returns are sensitive to the time period, we find that the aggregate long-run abnormal returns are higher in a time period where there is a large proportion of companies that split multiple times. Nevertheless, knowing whether the companies have split multiple times in the past will not lead to positive abnormal returns ex-ante; these returns can only be guaranteed if investors are able to forecast accurately which sample firms will implement another split in the future. Once the split again condition is controlled for, there is no role for the time period to influence the magnitude and significance of the abnormal returns. We also discover that firms that have not split before consistently outperform firms that have. This implies that instead of buying every company that splits, investors can achieve higher returns by focusing on those that have not split in the recent past. However, the profitability of this strategy depends on the state of the market (bull versus bear market). In summary, the thesis shows that while stock splits are perceived as good news by investors, abnormal returns cannot be guaranteed following the announcement date. The information contained in a stock split is incorporated into stock prices in a timely manner, however, what type of information this event is capturing remains an open question.