Ebook Consecutive Earnings Surprises: Small and Large Trader Reactions
Prior research demonstrates that investors respond differently to earnings surprises that occur as part of a string than to those that do not, but is inconclusive about why the market ascribes such importance to these earnings patterns. In particular, Barth, Elliott and Finn (1999) find that firms with patterns of increasing earnings have higher return reactions to the earnings surprises and higher price-earnings multiples as the pattern progresses relative to other firms. To shed light on the cause, I compare the trading responses of investors who make large trades (large traders), whom prior research has shown to be sophisticated, and investors who make small trades (small traders), whom prior research has shown to be relatively naïve, to earnings surprises that occur during strings of either positive or negative surprises.
Since the two classes of investors differ systematically in sophistication but not in risk preferences, differences in their responses to earnings strings shed light on whether mispricing or risk-based explanations are more likely. To provide direct evidence on whether small investors drive the market response to earnings strings, I also examine whether small trading patterns are associated with the stock returns surrounding the announcement of earnings surprises that occur as part of a string.
I test for differences in small and large trader responses to patterns in earnings growth and earnings surprises (both based on random-walk models and analyst forecasts). To measure investor reactions, I use New York Stock Exchange Trades and Quotations (TAQ) data. Following prior literature, I split trades based on the dollar value of the trade, and calculate measures of trade initiation using the Lee-Ready algorithm. This measure captures the direction (buy vs. sell) of the reaction. I examine each trade group’s level of reaction to “strings” of consecutive increases in seasonally-adjusted quarterly earnings. I find that small traders buy more intensely as a string of earnings increases lengthens while large traders’ buying and selling activity does not depend on the length of the string. These inferences hold when I focus on sequences of “strong positive” and “strong negative” surprises (defined to be in the top 30% and bottom 30% of earnings surprises) and when I define strings based on analyst-based earnings surprises. Small traders buy (sell) more strongly as a sequence of positive (negative) surprises lengthens, while large traders do not.
I relate small trade imbalances to concurrent returns around the earnings announcement, to verify that small trade imbalances are associated with returns, and find a strongly significant positive relation between small trades and the overall return response, even after controlling for large trades and the earnings surprise level. This positive association between the small trade response and concurrent returns remains significant throughout the earnings surprise series.
Together, these results suggest, first, that small traders have a preference for consistent performance, while large traders do not value consistency as highly. Second, the results suggest that “earnings momentum” trading is a cause of the return patterns documented in Barth, Elliott and Finn (1999), Kasznik and McNichols (2002), and Myers, Myers and Skinner (2007). These results contribute to our understanding of the market response to earnings patterns by providing evidence that small traders influence the market response to consistent patterns in earnings growth and earnings surprises.
This study contributes to the literature by providing evidence of whether it is less sophisticated small traders who react positively to earnings consistency or whether more sophisticated large traders value consistency as well. Bhattacharya (2001) and Battalio and Mendenhall (2005) increase our understanding of post-earnings-announcement drift by showing that it is small traders who respond most strongly to random-walk-based earnings surprises. This study increases our understanding of the market reactions to the time-series-pattern in earnings by building on the literature started by Bhattacharya (2001), extending the literature to examine small and large trade reactions to earnings time-series-patterns, and documenting significant trade response variations related to these patterns.
Prior work that examines market responses to earnings patterns has focused on returns-based tests (e.g. Barth, Elliott and Finn 1999; Kasznik and McNichols 2002; Myers, Myers and Skinner 2007). This study’s direct examination of trading patterns of small and large investors has the advantage of using each firm as its own control, assessing how different investors respond to the same announcement rather than comparing one set of firms with an unavoidably imperfect control sample. This research design increases our understanding of the market response, suggesting that “earnings-momentum” trading of unsophisticated investors is an important cause, and complementing the prior studies. .
The paper proceeds as follows. Section II describes the hypothesis development and related literature. Section III explains the empirical methods and data used in this paper, including the data and methods used to measure earnings surprises and trade reactions. Section IV presents results and Section V presents robustness checks. Section VI concludes.
Download
PDF Ebook Consecutive Earnings Surprises: Small and Large Trader Reactions
Posted in :
