Yahoo Finance Search And Earnings Announcements
University of California, Berkeley – Haas School of Business
James P Ryans
London Business School
Boston University – Questrom School of Business
Yahoo! Research Labs
July 1, 2016
We use Yahoo Finance search to examine the effects of investor attention at earnings announcements. We find that Yahoo Finance search at earnings announcements is a major factor explaining earnings responses and is predictive of subsequent returns. Moreover, we show that other measures of investor attention (e.g., Google search, EDGAR search) are less informative in explaining earnings responses and subsequent returns. The findings suggest the importance of investor attention, and in particular, retail attention in the pricing of financial information.
Yahoo Finance Search And Earnings Announcements – Introduction
The purpose of this study is to investigate the pricing effects of investor attention at earnings announcements. Despite the prevalence of research examining the market reaction to earnings (e.g., Ball and Brown 1968; Bernard and Thomas 1989), little is still known concerning how investor attention affects the pricing of earnings. Traditional asset pricing models generally conclude that new information is immediately available to traders and reflected in prices; accordingly, earnings responses should not depend on the extent of investor attention. However, theoretical models that assume not all investors observe the information suggest that the fraction of investors observing the information determines how fully the information will be priced (e.g., Grossman and Stiglitz 1980; Hellwig 1980).
Investigating the pricing effects of investor attention on information events has been inherently challenging as direct measures of attention have only recently become available. Prior research uses the number of shareholders (e.g., Bushee, Core, Guay, and Hamm 2010), index membership (e.g., Shleifer 1986; Chen, Noronha, and Singal 2004), stock listings (e.g., Kadlec and McConnel 1994, Foerster and Karolyi 1999), trading volume (e.g., Gervais, Kaniel, and Mingelgrin 2001; Barber and Odean 2008), news and headlines (e.g., Barber and Odean 2008; Yuan 2015), extreme returns (e.g., Barber and Odean 2008), institutional holdings (e.g., Lehavy and Sloan 2008), analyst following (e.g., Irvine 2003), price limits (e.g., Seasholes and Wu 2007), and day of the week (e.g., DellaVigna and Pollet 2009; deHaan, Shevlin, and Thornock 2015) as indirect measures of investor attention. Many of these measures are invariant over short time periods and can be driven by factors unrelated to investor attention. To better identify attention via the demand for firm-specific financial information, recent research uses Google Trends search volume (e.g., Da, Engelberg, and Gao 2011, 2015) and the Securities and Exchange Commission’s EDGAR search volume (Drake, Roulstone, and Thornock 2015; Dechow, Lawrence, and Ryans 2016).
We use proprietary Yahoo Finance (finance.yahoo.com) search from 2014 to 2015 to measure company-specific investor attention at earnings announcements for U.S. publicly-listed stocks. Yahoo Finance is the most popular web site for financial information in the U.S. with over 30 million unique daily users (Bordino, Kourtellis, Laptev, and Billawala 2014; Yahoo 2015), the vast majority of which are retail investors rather than professional investors. Due to the large number of daily users, the web-traffic patterns on Yahoo Finance are likely representative of the U.S. retail investor population. For comparison, there are approximately 60,000 unique daily EDGAR users, less than 0.3 percent of the number of daily Yahoo Finance users.1 We are unable to make comparisons to the frequencies of Google search volumes because Google only discloses relative search volumes (i.e., the search volume index, or SVI). Unlike Google search volumes, where it is not possible to identify whether investors ultimately view financial information, Yahoo Finance search reflects views of firm-specific financial- and news-related information.
Yahoo Finance presents a set of financial information pages for all publicly-listed companies. Each firm’s Summary page has 23 subpages, providing information on prices, news, ownership, analyst estimates, financial results and ratios, and SEC filings, where applicable data are available. Hence, it is possible both to identify that users are primarily consuming financial information on the searched firm, and to observe the types of financial information consumed when users search beyond the Summary page. In our analyses we consider both the effects of overall search and the specific categories of financial information search: i) Summary page search; ii) News search; iii) financial statement and SEC filing search (Financial); iv) Analyst Information search; and v) Other search.
We find that there is significant variation in Yahoo Finance earnings announcement search across U.S. publicly-listed firms. Specifically, firms in the highest quintile of abnormal search have increases of over 680 percent on the earnings announcement day, while firms in the lowest quintile of abnormal search have increases of only 34 percent, reflecting a spread of approximately 20 times. Size explains very little of the variation in abnormal search and as a result there are many large firms in the lowest quintiles of abnormal search. Not only is there significant variation in abnormal search at earnings announcements across firms but there is also significant variation within firms from quarter to quarter. For example, only 33 percent of firm-quarters stay within the same abnormal search quintile as in the past quarter, indicating that 67 percent of firm-quarters have significant changes in abnormal search from one earnings announcement to the next.
See full PDF below.