UC Berkeley

This dissertation examines the factors that influence investors' attention to the stock market and the relationship that exists among attention and real output variables including stock returns, trading volume, and volatility. Traditional asset pricing models assume that information is effortlessly obtained and instantaneously incorporated into pricing. This assumption requires that investors devote sufficient attention to the asset, and ignores the existence of various channels through which public information is disseminated. In reality, attention is a scarce cognitive resource which is related to the effort that investors must expend to obtain information; the implications of this contingency of attention on these limitations have been remarkably under-researched in the past.

In the first chapter of this study, I familiarize readers with Google Trends data and explain why such data is a better source to proxy for attention than the measures previously used in the literature. Next, utilizing this data, I describe how to measure investors' attention with regard to M

amp;A announcements, and show that attention is not instantaneous with the release of information, but is, instead, spread over a period surrounding the announcement. Retail investors pay attention and demand information about a firm as the announcement date approaches, during the announcement, and for days afterward. Finally, I present three aggregate measures of attention in the stock market, which are also based on search volume from Google. After constructing these measures, I study how they correlate with, but differ from, existing proxies of attention.

In the second chapter, I consider whether limited attention explains the announcement effect bias found in the M

amp;A literature concerning merger and acquisition announcements. More specifically, I ask: How does variation in investors' attention affect the capital market response to Mamp;A announcements? To answer this question I rely on the measure for attention to Mamp;A announcements described in the previous chapter and find that high abnormal attention on the day of announcement predicts high adjusted abnormal returns the day after. This effect is strongest among firms with high standard deviations and betas, and it partially reverses over the following months.

The third chapter argues that negative stock market performance attracts more attention from retail investors than comparable positive performance. Specifically, I rely on the three aggregate measures of attention in the stock market to test and confirm the hypothesis that retail investors pay more attention to negative rather than positive extreme returns. Empirical results strongly support that with respect to stock returns investors display this negativity bias in attention allocation. Across all specifications, lagged negative extreme returns are stronger predictors than positive extreme returns of high attention at the stock and market level. I rule out that negative returns are stronger simply because they are more unusual or because negative and positive returns are not symmetrical events to stockholders.