The stock market crisis and momentum. Some evidence for the Spanish stock market during the 1990s

Abstract

In this article, we test the momentum effect in the Spanish stock market during the 1990s. Though there is evidence of momentum, it disappears after the 1997 crisis. While momentum profits are associated with both size and turnover effects, neither of these factors is a determinant in explaining the momentum effect. The turn of the year effect also lacks sufficient explanatory power to account either for the appearance or disappearance of this effect. An important role is played in this puzzle by the winner portfolio, particularly when it is constructed from small, high turnover stocks. Analysis of the characteristics and evolution of this portfolio may, therefore, help to explain the momentum effect.

Acknowledgement

The authors would like to thank the helpful comments made by anonymous referees. This study has received financial support from the ERDF and the Spanish Ministry of Science and Technology (SEC2003-07808-C03).

Notes

¹Lee and Swaminathan (2000) use turnover as proxy for volume.

² In their article, both the formation period, J, and the holding period, K, take values of: 3, 6, 9 and 12, making a possible total of 16 different momentum strategies.

³ The returns are measured in calendar time by the method proposed by Jegadeesh and Titman. Nevertheless, the results are very similar when the returns are measured in event time, using the Newey–West adjustment for the t-statistic. The results are available from authors upon request.

⁴ In the cited work, Jegadeesh and Titman extend the estimation period of their original study to include the interval between 1990 and 1998 and find evidence of significant positive momentum returns during the 1990s.

⁵ Since returns in the first month of the formation period are positive and, for the most part, significant, their omission causes a decline in observed momentum. This lack of short-term reversal may be due to microstructural differences between the two markets.

⁶ See, for example, Glaser and Weber (2001) and Lee and Swaminathan (2000).

⁷ The reason for considering the momentum portfolios in terms of terciles is based on the characteristics of the Spanish market, where there is a limited number of continuous market stocks, which could result in insufficient diversification of the portfolios formed on the basis of size, turnover and BTM.

⁸ The results for each momentum strategy across the different BTM subgroups, are available to any readers who may be interested.

⁹ Similar results are obtained using a joint size and turnover control technique, following the procedure used by Grinblatt and Moskowitz (2004).

¹⁰ Moreover, the adjusted for size and turnover rest of the year momentum returns are very similar to those in Panel C of Table 6 for nonrisk adjusted returns. The results are available from the authors upon request.

¹¹ The J = 9, K = 6 strategy was selected because it presents the highest return of the whole sample period and, as far as time stability is concerned, performs in a similar manner to the remaining strategies. These results are available from the authors upon request.

¹² In the January returns for 1997, as in the average January returns across the sample, there is a negative impact that generates an apparent break in the trend.

¹³ The results of the remaining strategies are not reported for the sake of simplicity, but are available from the authors upon request.

¹⁴ These unreported results are available from the authors upon request.

¹⁵ The reference in the article is to volume, but this is measured by average stock turnover.