Can Facebook Predict Stock Market Activity? - University of Miami

CanFacebookPredictStockMarketActivity? ∗ YigitcanKarabulut † Goethe University Frankfurt FirstDraft:August29,2011 ThisDraft:October17,2011 -PreliminaryDraft- Pleasedonotquotewithoutpermission Abstract Using a novel and direct measure of investor sentiment, I find that Facebook’s Gross National Happiness (GNH) has the ability to predict changes both in daily returns and trading volume in the US equity market. For instance, a one standard deviation increase in GNH predicts an increase in market returns equal to 11 basis points over the next day. Moreover, the impact of GNH appears to be stronger among small-cap stocks, and in the face of turmoil. Keywords: Investor sentiment, social media, behavioral finance, Facebook. JEL Codes: D81, G11, G12. ∗ I would like to thank Brad Barber, Paul Tetlock, Gabriele Lepori, Erik von Schedvin, and semi- nar

participants at the NBER Behavioral Finance Conference, and Sveriges Riksbank for very helpful comments and discussions. The initial draft of this paper was written in the Summer of 2011 while I was visiting as a dissertation intern the Research Division at the Sveriges Riksbank; kind hospitality is gratefully acknowledged. † Yigitcan Karabulut is with Goethe University Frankfurt, Retail Banking Competence Center, House of Finance, Gruneburgplatz 1, 60323, Frankfurt, Germany, E-mail: karabulut@finance.uni-frankfurt.de, Phone: +49 (0) 69 798 33859. The usual disclaimer applies. 1 Introduction The question of whether sentiment affects stock market has long attracted a great deal of attention from academics. Following the influential work of De Long et al. (1990), a number of studies address this issue, and show that noise trader sentiment can persist in financial markets and influence asset prices (e.g. Lee et al., 1991, Barberis et al., 1998; 1 Daniel et al., 2001; Tetlock, 2007). Thus the relevant question is not any longer whether sentiment has an effect on stock prices, but rather how to measure investor sentiment more accurately and ascertain its effects on the stock market (Baker and Wurgler, 2006). In this paper, I investigate the interactions between investor sentiment and stock market using a new and direct measure of sentiment. Specifically, I employ Gross National Happiness (hereafter referred to as, GNH) of Facebook, which captures daily sentiment using content from the individual status updates of almost 100 million Facebook users in the US. I attempt to quantify the effects of Facebook sentiment on stock market by examining three questions, which are derived from the behavioral models of securities market (e.g. De Long et al., 1990; Campbell et al., 1993). First, I ask whether the sentiment measure compiled from Facebook displays any ability to predict both daily returns and trading volume in the stock market. Second, I analyze the question of how sentiment impacts on the cross-section of stock returns rather than its effect on the aggregate market returns. Finally, I study the interactions between Facebook sentiment and market returns during unusually volatile episodes of stock market. To answer these questions, I adopt a vector autoregressive (VAR) framework, using daily stock market and sentiment data over the four year period from September 10, 2007 through September 9, 2011. Estimation results indicate that Facebook sentiment, as measured by GNH, displays ability to predict daily market returns. The results are statistically significant and eco- 1 For...

Website: bus.miami.edu | Filesize: 558kb
No of Page(s): 32
Download Can Facebook Predict Stock Market Activity? - University of Miami ....pdf