Media Makes Momentum Alexander Hillert, Heiko Jacobs and Sebastian M¨ uller∗

February 2013 Appendix attached

Abstract Relying on 2.2 million articles from 45 national and local U.S. newspapers between 1989 and 2010, we find that firms particularly covered (neglected) by the media exhibit ceteris paribus significantly stronger (weaker) momentum. The effect is more pronounced in states with higher investor individualism and among stocks predominantly held by overconfident fund managers. Findings suggest that media coverage can exacerbate investor biases, leading return predictability to be strongest for firms in the spotlight of public attention. In line with prominent models such as Daniel et al. (1998), our results collectively support an overreaction-based explanation of the momentum effect. Keywords: Momentum, media, overreaction, attention effects, investor biases JEL Classification Codes: G12, G14

∗ Alexander

Hillert, Chair of International Finance and CDSB, University of Mannheim, L9, 1-2, 68131 Mannheim. E-

Mail: [email protected]. Heiko Jacobs, Chair of Banking and Finance, University of Mannheim, L 5, 2, 68131 Mannheim, Germany. E-Mail: [email protected]. Sebastian M¨ uller, Chair of Banking and Finance, University of Mannheim, L 5, 2, 68131 Mannheim, Germany. E-Mail: [email protected]. We thank seminar participants at the Second Helsinki Finance Summit on Investor Behavior, the University of Mannheim, the LMU Munich, the Campus for Finance conference 2013, as well as the University of California, Berkeley for valuable comments.

1

1

Introduction

The momentum effect is among the strongest and most pervasive return anomalies. While its existence has been convincingly documented in different time periods, countries, indices, and asset classes1 , a central issue is still far from being resolved: What are the underlying causes of momentum? In other words: Why exactly do winner stocks of the recent past tend to outperform loser stocks of the recent past? The goal of this paper is to employ an extensive media data set to shed new light on this long-standing debate. Recent research demonstrates that media coverage directly affects the way in which investors collect, process, and interpret information (e.g. Barber and Odean (2008), Engelberg and Parsons (2011), Tetlock (2011), Engelberg et al. (2012)). Collectively, findings suggest “a potentially important role for the media in shaping the behavior of the stock market” (Hong and Stein (2007), p. 118). The interesting link to the momentum literature lies in the fact that investors’ attention and information processing also play a crucial role in prominent behavioral theories of momentum.2 Whether momentum will ceteris paribus be weaker or stronger among high coverage stocks, is, however, not clear ex ante. On the one hand, media coverage may lead to faster information diffusion and thus weaken momentum. On the other hand, media coverage may contribute to investor biases, thereby strengthening momentum. As we outline in the literature review, theoretical considerations offer support for both views, making exploring the role of the media an important empirical task. To this end, we rely on a novel and carefully constructed data set of newspaper articles. It comprises approximately 2.2 million news stories in four leading national as well as 41 local U.S. newspapers from 1989 to 2010, which we match to 7,815 firms. The essence of our findings is captured in figure 1.

Insert figure 1 here

The cumulative profits to a winner minus loser long-short portfolio are displayed separately for stocks in the highest media coverage quintile (red solid line) and lowest quintile (blue dashed line). Media coverage 1 The

momentum effect was first described in Jegadeesh and Titman (1993). For U.S. stock momentum in earlier and

later timer periods see e.g. Chabot et al. (2009) and Jegadees