SSRN-Long Horizon Mean Reversion for the Brussels Stock Exchange: Evidence for the 19th Century by Jan Annaert, Wim Van Hyfte

If people look at the same data over and over again, it should not be surprising that eventually people find things. Moreover, if different people look at the same data set, they are likely to find the same things. For this reason, it is always nice when there is an out of sample data set that can be used to verify the initial findings. Annaert and Van Hyfte provide us this opportunity.

SSRN-Long Horizon Mean Reversion for the Brussels Stock Exchange: Evidence for the 19th Century by Jan Annaert, Wim Van Hyfte

They "present new evidence on the time-varying behavior of stock prices using a completely new and unique dataset of historical stock returns from the Brussels Stock Exchange that has never been studied before. To the best of our knowledge, this is probably the most comprehensive and accurately constructed historical index representing more 1500 different common stocks during the period 1832-1914. The excessive use of the CRSP return data in examining predictability and the data mining risks involved, render this independent return database a adequate out-of-sample test for different asset pricing anomalies identified in the literature."

With this cool data set, their key finding is that:

"Contrary to Fama and French (1988) and Poterba and Summers (1988), our results show that stock prices do not contain autoregressive stationary components but instead resemble a random walk. Capital appreciation returns exhibit stronger time-varying behavior than total returns. Belgian stock returns demonstrate strongly significant seasonality in January notwithstanding the absence of taxes. Moreover, long horizon mean reversion is present, however, completely concentrated in January."

While this is all interesting, what will likely force me to redo my notes is that the January effect does NOT (at least at first glance) appear to be tied closely to taxes or the small firm effect!

The authors:

"did not find any official sources or records making reference to the Belgian government levying taxes on capital gains or dividends during that period. Second, our results show that larger rather than smaller companies achieve abnormal returns throughout the month of January disputing the tax-motivated size premium. Last, abnormal returns earned during January appear to be related to more fundamental factors like dividends rather than taxes as the month of July, another high dividend-yield month, is subject to the same effect. Further research on dividends and how asset prices respond to dividend information is required to examine these effects in more detail.

Interesting.

Cite:

Annaert, Jan and Van Hyfte, Wim, "Long Horizon Mean Reversion for the Brussels Stock Exchange: Evidence for the 19th Century" (December 20, 2004). EFA 2005 Moscow Meetings http://ssrn.com/abstract=676006