In addition to such well-known calendar-based anomalies as the turn-of-the-month effect and the day-of-the-week effect, there is some support for the other January effect, which implies that the 11-month holding period returns following positive January results are higher than those following negative January results. The authors show that this result may be sample sensitive rather than a specific anomaly.
What’s Inside?
The other January effect suggests that the 11-month holding period returns following positive January returns are typically higher than those following negative January returns. Supported by data over the period of January 1940–December 2003, the other January effect suggests that returns in January are a predictor for the returns over the remaining financial year. Although this anomaly is supported by the data for 1940–2003, it does not hold when considering data covering an extended period from July 1926 to January 2012.
How Is This Research Useful to Practitioners?
Such anomalies in market pricing as the turn-of-the-month effect and the day-of-the-week effect are often used by investors to exploit potential opportunities to generate risk-free profits. Support for these anomalies is generally found with long sample periods over which to test the findings. But the authors find that one of the effects discovered recently, the other January effect, is not supported by a period longer than the original data sample. Extending the sample period from 1940–2003 to 1926–2012 reveals that the difference in returns for an 11-month holding period after positive returns in January is not significantly different from the returns following negative returns in January. In addition to this evidence of sample sensitivity, there is no support for the anomaly in non-US markets.
This anomaly is also not unique to the month of January; holding period returns following a positive February are typically lower than those following a negative February. And although September generally has fewer positive than negative returns, both the three-month and six-month holding periods following positive September returns are significantly higher than those following a negative September. The effects for other months are negligible, with the exception of positive other April and negative other August effects for a six-month holding period.
Generally, any positive effect during the first six-month holding period is reversed during the five months thereafter.
How Did the Authors Conduct This Research?
To test the validity of the other January effect, the authors retest earlier findings using an extended data period covering July 1926–January 2012 instead of January 1940–December 2003, which is used in previous research. Rather than solely focusing on the month of January and the subsequent 11-month holding period, the authors explore the relationship between the returns in every calendar month and the returns in the subsequent 1-, 3-, 6-, and 11-month holding periods. Their results indicate that the predictive power for the month of January is not as strong as previously thought. February’s returns, for example, are negatively correlated with the subsequent one-, three-, and six-month holding period results, whereas the September returns are good indicators for the following three- and six-month holding periods. April and August have some predictive power but only for six-month holding periods.
The authors’ findings support other research that concludes that the other January effect only holds in US markets and cannot be generalized to other markets.
Abstractor’s Viewpoint
The research presented in this article reveals that the predictive power of January on holding period returns of 11 months is potentially less significant than was previously thought. Although there are anomalies in pricing that could be exploited, investors would do well to tread carefully.