Research

Published Papers

Social Media and Financial News Manipulation
with Shimon Kogan and Tobias J. Moskowitz

Forthcoming at the Review of Finance

We study fake news in financial markets using a novel dataset from an undercover SEC investigation. Our setting measures both the direct and indirect effects of market manipulation. Fake articles directly induce abnormal trading activity and increase price volatility, but in addition, the awareness of fake news from the SEC investigation indirectly affects legitimate articles, causing market participants to discount all news from these platforms. These spillover consequences significantly reduce the social network’s impact on information dissemination, trading, and prices. The results are particularly acute among small firms with high retail ownership and for the most circulated articles. The equilibrium response of consumers and producers of news on these networks is consistent with models of trust, providing novel evidence on the importance of social capital for financial activity.

Initial Coin Offerings: Financing Growth with Cryptocurrency Token Sales
with Sabrina Howell and David Yermack

Lead article and Editor's Choice, The Review of Financial Studies, Issue ​33 (9) 2020, pp 3925–3974

Initial coin offerings (ICOs) have emerged as a new mechanism for entrepreneurial finance, with parallels to initial public offerings, venture capital, and pre-sale crowdfunding. In a sample of more than 1,500 ICOs that collectively raise $12.9 billion, we examine which issuer and ICO characteristics predict successful real outcomes (increasing issuer employment and avoiding enterprise failure). Success is associated with disclosure, credible commitment to the project, and quality signals. An instrumental variables analysis finds that ICO token exchange listing causes higher future employment, indicating that access to token liquidity has important real consequences for the enterprise.

Why Don't We Agree? Evidence from a Social Network of Investors

with Anthony Cookson

Journal of Finance Vol 75, No 1 (February 2020), pp. 173-228

We study sources of investor disagreement using sentiment of investors from a social media investing platform, combined with information on the users' investment approaches (e.g., technical, fundamental). We examine how much of overall disagreement is driven by different information sets versus differential interpretation of information by studying disagreement within and across investment approaches. Overall disagreement is evenly split between both sources of disagreement, but within-group disagreement is more tightly related to trading volume than cross-group disagreement. Although both sources of disagreement are important, our findings suggest that information differences are more important for trading than differences across market approaches.

Is Investor Attention for Sale? The Role of Advertising in Financial Markets
with Joshua Madsen

Journal of Accounting Research Vol 57, Issue 3 (June 2019), pp. 763-795

Prior research documents capital market benefits of increased investor attention to accounting disclosures and media coverage, however little is known about how investors and markets respond to attention-grabbing events that reveal little nonpublic information. We use daily firm advertising data to test how advertisements, which are designed to attract consumers' attention, influence investors' attention and financial markets (i.e., spillover effects). Exploiting the fact that firms often advertise at weekly intervals, we use an instrumental variables approach to provide evidence that print ads, especially in business publications, trigger temporary spikes in investor attention. We further find that trading volume and quoted dollar depths increase on days with ads in a business publication. We contribute to research on how management choices influence firms' information environments, determinants and consequences of investor attention, and consequences of advertising for financial markets.


 

Working Papers:

Can Social Media Inform Corporate Decisions? Evidence from Merger Withdrawals
with J. Anthony Cookson and Christoph Schiller

This paper examines how social media informs corporate decisions by studying the decision of firm management to withdraw an announced merger. A standard deviation decline in abnormal social media sentiment following a merger announcement predicts a 0.64 percentage point increase in the likelihood of merger withdrawal (16.6% of the baseline rate). The informativeness of social media for merger withdrawals is not ex- plained by abnormal price reactions or news sentiment. Consistent with learning from external information, we find that the social media signal is more informative after a firm adopts a corporate Twitter account, which offers a conduit for listening to investor feedback. In addition, most of the informativeness is driven by investors who reference fundamental information, not price trends, and is driven by longer tweets that likely contain investment analysis. The social media signal is also more informative for complex mergers in which analyst conference calls take a negative tone, driven by the Q&A portion of the call. Overall, these findings imply that social media is not a sideshow, but an important aspect of firm information environment.

Are Cryptos Different? Evidence from Retail Trading
with Shimon Kogan, Igor Makarov, and Antoinette Schoar

Trading in cryptocurrencies has grown rapidly over the last decade, primarily dominated by retail investors. Using a large dataset of more than 200,000 retail traders from eToro, we show that they have a different model of the underlying price dynamics in cryptocurrencies compared to other assets. Retail traders in our sample are contrarian in stocks and gold, yet the same traders follow a momentum strategy in cryptocurrencies. Individual charac- teristics do not explain the differences in how people trade cryptocurrencies versus stocks, suggesting that our results are orthogonal to differences in investor composition or clientele effects. Neither lack of cashflow information, inattention, or preference for lottery-like stocks explain our findings. We conjecture that retail investors hold a model of cryptocurrency prices, where positive returns increase the likelihood of future widespread adoption, which in turn will drive up asset prices.

 

The social Signal
with J. Anthony Cookson, Rinjing Lu, and William Mullins

Best Paper Award in Honor of Jack Brick, 11th Michigan State FCU Conference.


We examine social media attention and sentiment from three major platforms: Twitter, StockTwits, and Seeking Alpha.  We find that attention is highly correlated across platforms, but sentiment is not: its first principal component explains only 6 percentage points more variation than purely idiosyncratic sentiment.   We attribute differences across platforms to differences in users (e.g., professionals vs. novices) and differences in platform design (e.g., character limits in posts).  We also find that sentiment and attention are both positively related to retail trading imbalance, but contain different return-relevant information.  Sentiment-induced retail trading imbalance predicts positive next-day returns, in contrast to attention-induced retail trading imbalance, which predicts strongly negative next-day returns.  These results highlight the practical importance of distinguishing between social media sentiment and attention, and suggest caution when studying the social signal through the lens of a single platform.

 

Strategic Disclosure Timing and Insider Trading

Revision requested at Management Science

I provide evidence that managers strategically manipulate their company’s information environment to extract private benefits. Exploiting an SEC requirement that managers disclose certain material corporate events within five business days, I show that managers systematically disclose negative events when investors are more distracted, causing returns to under-react for approximately three weeks. Strategic disclosure tim- ing is concentrated among smaller firms with high retail-investor ownership and low analyst coverage. Furthermore, I use the fact that most insider sales are scheduled in advance to demonstrate that top managers are more than twice as likely to strategically time disclosures if the return under-reaction benefits their insider sales. Finally, I find that firms that systematically disclose negative news on Fridays have higher levels of earnings management.

Bad News Bearers: The Negative Tilt of Financial Press

with Eric C. So

We show the financial press is more likely to cover firms with deteriorating performance. Our main tests illustrate the nature of the media's story selection process (i.e., what events to cover) and the usefulness of this selection process for forecasting firms' future earnings news and returns. We first show the media is approximately 11-to-19 percent more likely to cover a firm's earnings announcements if they convey poor performance. Similarly, in forecasting tests, greater media coverage predicts subsequently announced declines in firms' profitability and negative analyst-based earnings surprises. A simple long-short strategy betting against firms with high media coverage yields an average return of roughly 40 basis points per month, suggesting media coverage helps forecast future returns because the story selection process is titled toward novel negative events. Together, our findings highlight the usefulness of the media's coverage decisions in estimating expected returns, as well as a potential inference problem when researchers use media coverage to measure the extent of information dissemination and/or whether an information event occurred.

Does Disagreement Facilitate Informed Trading? 
with J. Anthony Cookson and Vyacheslav Fos

Using high-frequency disagreement data from the investor social network StockTwits, we find that greater investor disagreement facilitates informed trading by activists and short sellers. These findings are unexplained by sentiment, news and retail order flow, and they remain when we measure disagreement overnight, which alleviates concern that disagreement and informed trading respond to a common shock. When short selling is costly, the facilitating effect of disagreement on trading is dampened for informed buyers but is amplified for sellers. These findings suggest that informed traders respond meaningfully to valuation changes induced by disagreement.

Finance TV
with Diego Garcia and Ryan Lewis
(draft coming soon)