What Drives the January Effect?

Virginia Tech Working Paper

47 Pages Posted: 11 Dec 2000

See all articles by Honghui Chen

Honghui Chen

Department of Finance, University of Central Florida

Vijay Singal

Virginia Tech

Date Written: April 2001


The January anomaly has attracted much academic interest and has been explained in different ways. However, the multitude of explanations has created confusion about the validity and relative importance of those explanations. In some cases, the hypotheses are examined individually though the evidence may be consistent with more than one hypothesis. Furthermore, prior work has not adequately controlled for the bid-ask bounce. Therefore, the results leave the reader somewhat confused regarding the January effect: is it caused by tax-loss selling, window-dressing, information, bid-ask bounce, or a combination of these causes? In this paper, we try to disentangle different explanations of the January effect and identify its primary cause. We find that past losers are more likely to be sold in December than in January to realize the tax advantage of capital losses. Past winners are more likely to be sold in January than in December to postpone payment of taxes. The selling is accompanied by changes in volume around turn of the year consistent with the tax-related selling hypotheses. The results are not materially affected when we use the midpoint of quotes instead of actual prices: the bid-ask bounce accounts for about 20-25% of the observed returns. To verify the window-dressing hypothesis, we examine stock returns around June-July, the period of semi-annual reporting by institutional managers that is not contaminated by tax-related trading. We do not find an economically meaningful difference between the 5-day return at the end of June and the 5-day return at the beginning of July, which is not consistent with window dressing. If the January effect occurs due to release of new information in January that affects the information-poor firms more than the information-rich firms then the returns in January should be related to availability of information (for example, with the number of analysts as a proxy). We do not find a correlation consistent with the information hypothesis. There is no information-related effect in June-July. The evidence here supports the tax-related selling hypotheses as the drivers of January effect.

Keywords: January Effect, Seasonality, Window-Dressing, Tax Loss Selling, Tax Gain Selling

JEL Classification: G14, G10

Suggested Citation

Chen, Honghui and Singal, Vijay, What Drives the January Effect? (April 2001). Virginia Tech Working Paper, Available at SSRN: https://ssrn.com/abstract=252190 or http://dx.doi.org/10.2139/ssrn.252190

Honghui Chen

Department of Finance, University of Central Florida ( email )

PO Box 161400
Orlando, FL 32816
United States
407-823-0895 (Phone)

Vijay Singal (Contact Author)

Virginia Tech ( email )

250 Drillfield Drive
Blacksburg, VA 24061
United States
5402317750 (Phone)

Here is the Coronavirus
related research on SSRN

Paper statistics

Abstract Views
PlumX Metrics