Printed News Is The Old News: The Role Of Local Media In Local Trading Activity And Local Stock Returns
University of Gothenburg – Centre for Finance – School of Business, Economics and Law
Goteborg University – School of Business, Economics and Law
When portfolio managers get started in the business, their investing style often changes over the years. However, when Will Nasgovitz bought his first stock when he was 12, he was already zeroing in on value investing, and he didn't even know it. Nasgovitz has been with mutual fund manager Heartland Advisors for almost 20 years, Read More
University of Gothenburg – Centre for Finance – School of Business, Economics and Law
July 26, 2016
We examine the role of local media in local trading activity and local stock returns. Local media appears to have an influence on individuals’ post buy-sell imbalance in local stocks and abnormal profitability on the local trades. However, local media has a larger impact on the profitability and the local trading activity of individuals who moved into the district (other locals) than on individuals who live in their birth districts (native locals). This result indicates that native locals are more prone to utilize their social networks and turn to the old-fashioned word of mouth method when it comes to obtaining and sharing local information, whereas other locals may lack this opportunity and instead obtain their local information from the Internet. Thus the effects on local trading activity and profitability may not be driven by local media itself, since local investors may have already obtained the information, perhaps form the same sources. Our results enable us to separate informed local trades from uninformed local trades. We also show that widely documented disposition effect stays remarkably consistent and robust in local trading activity.
Printed News Is The Old News: The Role Of Local Media In Local Trading Activity And Local Stock Returns – Introduction
We examine the role of local media in local trading activity and local stock returns. In efficient markets stock prices should incorporate all available value-relevant information instantaneously. However, we know from earlier studies that investors only follow and process a subset of information on a subset of firms, for instance, they exhibit local bias, which may have ambiguous effects on stock returns.
Although the evidence is scarce and mixed, some studies examining the sources of information channels exist. For instance studies examining the role of media in financial markets suggest that media, in particular local media, may serve as watchdogs (see Miller, 2006, Dyck, Volchkova, and Zingales, 2008, and Dyck, Morse, and Zingales, 2010). Local media is also more likely to discover information from employees and local suppliers and may report the news that is not yet disclosed by a firm. Thus local media may utilize its proximity to the local news sources and any potential asymmetric disclosure provided by firm insiders (Kothari, Shu, and Wysocki (2009).
This advantage may not only cause local media to publish the stories before distant newspapers but may also allow them to interpret and publish the same underlying news differently from national and foreign media. Local media may act as cheerleaders (referred to as “hype”) and may provide favorable news to attract more local businesses (advertising etc.) form the local firms. In other words, local media may exhibit slant (see Gentzkow and Shapiro, 2006; Gurun and Buttler, 2012). Thus, although the stories about local firms are based on the same underlying event, selective omissions by local analyst predictions and choice of words in the local media may convey a different impression of the event than what is presented by national media.
In view of this, examining the sources of local information and testing its direct effects on local investments would improve the earlier evidence on local bias and could enhance informational efficiency in financial markets. Some remote investors may not be aware that local investors already have obtained the information from local media and traded on the news. In this case the news might be stale, which may induce remote investors to behave differently, perhaps show some biases such as trading more aggressively, or simply reacting later, unaware that the news is delayed information. This may lead to temporary movements in stock prices (see Tetlock, 2007 and 2011).
The empirical evidence on such media slant on stock returns (see Groseclose and Milyo, 2005; Gurun and Butler 2012) is mixed. Cutler, Poterba, and Summers (1989) and Hirshleifer, Myers, Myers, and Teoh (2008) show no relationship between earnings announcements and stock returns. The post earnings announcement drift does not seem to be related to the trades of individual investors. In contrast, recent evidence shows a strong relationship between news coverage and stock returns. Engelberg and Parssons (2011) find a significant relationship between individual trading in local markets and local newspaper coverage of earnings announcements. Trading volume and return volatility seem to increase with communication activity measured by messages in internet chat rooms (see Antweiler and Frank, 2004). Although it is unrelated to the returns, the information content of the ambient noise level in the Treasury Bond futures trading pit seems to influence trading volume, volatility, and depth (Coval and Shumway 2001).
We add to the knowledge in these previous studies by providing new evidence regarding whether or not the local news about local businesses induces, amplifies, or reflects investors’ information advantage, local trading activity, and local stock performance. We measure local trading activity by the buy-sell imbalance (BSI), which indicates the net trade and thus changes in the individuals’ position (see Hong and Kumar 2002; Zhu, 2002). We measure the performance of the local trades by the Sharpe ratios. Unlike previous studies, which cover mostly local newspapers, our media data contains a richer information set as it covers both newspapers and webpage news provided by local, national, and foreign media sources. This distinction is important as webpage news have become significant information channels with the development of the Internet.
To our knowledge none of the above mentioned studies examine the media sources besides the printed national press. One exception is Antweiler and Frank (2004), however, the study does not particularly examine the relation between local media and local trading activity. We extend this literature by providing new evidence whether local media provided in sources other than the printed press has substantial influence on local trading activity and local trading profitability.
The main contribution of our study is the use of a more extensive database on media to test the local information hypothesis. Our media data allows us to separate news as i) printed press and webpage news, and as ii) local, national, and foreign news. We are also able to identify the region in which the media firm is headquartered, which allows us to identify whether or not a local media provides information about a local firm. It is possible that remote investors may read local webpage news about a firm in the region. However, our data shows that there are many small local webpage sources that remote investors may not be aware of at all. In particular, examining the extent to which remote investors pay attention to the local news will enrich our understanding of the relation between local information hypothesis and local investments. In this paper we also aim at separating the informed local trades from uninformed local trades. This separation would assist us in examining whether the observed results are based on superior information or familiarity to the firms.
Our investor data comes from Euroclear Sweden and cover quarterly holdings of all the investments in firms listed on Swedish stock markets during a period stretching from 2006 until 2014. This provides us with 36 waves of panel data covering about 1.8 million individual investors and their 30 million trades of which 16.29% consist of local trades and 83.71% consist of remote trades. We obtain the number of news articles covered in national, local, and foreign printed press, as well as web pages about these listed firms that appeared during the past month of each quarterly holding period, and match this information to our investor data. The media data is obtained from Retriever through University of Gothenburg’s library and the stock price information comes from Datastream.
See full PDF below.