A new perspective on the anomalies in the monthly closings of the Dow Jones Industrial Average
Date
2003
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Southern New Hampshire University
Abstract
This study explores three types of month effects in the Dow Jones Industrial Average: (a) for a given period, if the mean of monthly percentage changes of each month was different from zero, (b) for a given period, if the mean of monthly percentage changes for a month was different from the means of all the other months, and (c) for a given period, if the variance of the monthly percentage changes for a month was different from the variances of all the other months. For our entire data set (May 1896 to December 2002) we find that the means of monthly percentage changes of only July, August, January and December were significantly greater than zero (months put in descending order). But the means of none of these three months were significantly higher compared to the means of all the other months. With a mean percentage change of -1.25%, only September appears with significant negative returns. And this mean is significantly lower compared to the means of all the other months. In other words, for the entire data set, we have a negative September effect.
Month effect with respect to variance (variance of monthly percentage changes for a month being significantly different from all the other months) was found for January, February and December (lower variances), and April (higher variance).
When we look at the first half of the twentieth century versus the second half, we see more pronounced month effects in the second half - considering all three types of effects we analyze. December exhibited all three types of effects in this period.
When we sub-divide the last century into four 25-year periods, we find more pronounced month effects in the last quarter than in the previous three quarters.
When we sub-divide the data into 10-year periods, we do not find any consistent and discernible pattern. The month effect varies with the time period we consider and the type of effect we analyze. Though one would expect the DJIA stocks to be free from seasonal patterns since each one of them are closely followed by a large number of analysts, the existence of any type of month effect is surprising. However, given that no discernible pattern is detectable is a reflection of efficiency of the DJIA stocks to a large degree.
Description
Version of Record
item.page.type
Working Paper