Working Papers

Social media influencers in the stock market are commonly assumed to have a significant impact on investor decision-making. We have examined how YouTube influencers at different power levels exert varying degrees of impact on trading volume and stock returns. We present four main findings (1) Theory and common belief are nearly correct in that high-influencer groups exert stronger effects on stock returns and trading volume than lower-influencer groups; (2) Furthermore, the uploading behavior of high-influence groups is associated with significantly positive BHAR over the 10- to 60-day horizon; (3) However, on average, compared with the most-high and less-high influencer group, the less-high influencer group showed the more effect on short-term stock returns and trading volume; (4) The weaker effect of most-high influencer group tends to be associated with their coverage of large firms and their choice of a more cautious tone in their videos. We have also found evidence of spillover effects from influencers, implying that video uploads by high-influencer groups lead to more uploads from others. Furthermore, influencers' effects are more pronounced under short-selling constraints.

We examine institutional investors' presentations at charity-hosted investment conferences and show presenters act as information intermediaries at these events by contextually evaluating presented firms. Our study uses a unique setting where the information processing and analyses of sophisticated institutional investors are publicly disseminated, allowing us to examine how expert interpretation of public information aids the information processing of other market participants. Consistent with presentations improving other investors' information processing, prices reflect earnings news faster for presented firms in the quarters after presentations. The effect is stronger for presentations with in-depth analyses, multiple arguments, and longer presentation notes. Presenters are distinct information intermediaries as they use novel arguments compared to pre-conference analyst reports. After the conferences, the content of analysts' reports changes and they improve their earnings forecast accuracy. Our study sheds light on the previously unexplored role of institutional investors as information intermediaries and introduces a novel mechanism-termed the expertise effect-which complements the traditional awareness effect. Together, these effects underscore how institutional investors' presentations reduce information integration costs for other market participants.

We leverage rich social media data and large language models (LLMs) to examine the relationship between investor trading strategies, sentiment, and market outcomes. Extracting trading strategies  embedded in 96 million social media posts, we find that  strategy adoption is heterogeneous and dynamic, with substantial differences in performance outcomes. Our results show that news arrivals decrease users' reliance on technical signals and increase their utilization of fundamental signals. Technical sentiment negatively predicts stock returns, particularly among short-term or inexperienced users, whereas fundamental sentiment positively forecasts returns. Additionally, message sentiment correlates positively with aggregate retail buying, with technical sentiment strongly associated with aggressive buying by Robinhood investors. Our study demonstrates the promise of using AI to understand investor behaviors and their implications for market dynamics.

Boycotts triggered by public companies' practices perceived as ideologically polarizing can lead to negative investor reactions. In this study, I examine how the stock market responds to such boycotts and whether ideology-driven social media discourse shapes this response, given investors' increasing reliance on social media information for decision-making. On average, polarizing boycotts are associated with a 1% (2.3%) drop in equity value over the 7 (60) trading days after gaining online traction. Immediate price decline is more pronounced when social media discussions are dominated by users ideologically aligned with the boycotters, particularly when their posts attract online engagement, emphasize financial impact, or come from influential, prolific users. I also find modest evidence that return volatility following boycotts increases when the ideological beliefs of social media posters are more diverse. My findings suggest that polarizing boycotts against corporate actions have stock market ramifications, and that ideology-driven social media opinions seem to amplify both price decline and volatility.

We use data from HotCopper, Australia's largest stock message board, and exploit its unique announcement-specific thread structure to examine whether investors' accounting-related discussions help process less readable corporate announcements. We find that less readable disclosures, particularly unanticipated nonearnings announcements, generate more accounting-related discussions and are more pronounced for firms with low institutional ownership, limited coverage, and high operational complexity. Market reaction analyses reveal that these discussions are associated with stronger price and volume responses and improve market reactions around less readable announcements, suggesting that social media can function as a low-cost supplement for traditional information intermediaries. Further comparing New Zealand based stock message board, Sharetrader.co.nz, with the absence of announcement-linked threads shows weaker and less consistent effects, underscoring that social media's informativeness in capital markets hinges not just on investor participation but critically on platform design. Our findings contribute to the disclosure processing cost literature by showing that structured forums reduce cognitive frictions and improve market efficiency, especially in retail investor-dominated settings.

This paper shows that intense discussion on the Reddit social-media platform increases noise-trader risk for informed investors, resulting in lower price informativeness about earnings and delayed mispricing correction. Increased social discussion is associated with a decreased magnitude of pre-earnings-announcement drift, increased earnings-response coefficients, and a sizeable return reversal, demonstrating the decline in price informativeness around earnings announcement dates. In addition, stock prices of firms with high social discussion incorporate future earnings news more slowly. Social discussion leads to a decrease in trading activity by informed investors, such as hedge funds and short sellers. Consequently, social discussion results in delayed price correction of well-documented anomalies for up to two months; a corresponding trading strategy earns about 1.4% monthly. The main findings are corroborated using a matched sample. The findings suggest that intense social discussion reduces the production of value-relevant information.

Using debit and credit card data, we find that a one standard deviation increase in local firms' earnings surprises leads to a 2% increase in biweekly household consumption among households located near the firms' headquarters. This relation is stronger when earnings news is relevant to the local economy, widely disseminated, linked to increased information acquisition, or a credible signal of firm performance. Consumption responses (a) span various households, including small business owners, investors, employees, and other unaffiliated households, (b) are more pronounced among financially sophisticated households likely able to access earnings news, and (c) are concentrated in inexpensive, discretionary items, such as dining out. A two-round survey of 432 households before and after earnings announcements in late 2024 shows that household consumption responds to local earnings news, primarily because households update their beliefs about financial prospects, with media and word of mouth serving as key news dissemination channels. Our findings altogether indicate that earnings news informs local household consumption decisions.

This paper examines the existence, drivers, and implications of toxic content in financial social media. Using state-of-the-art machine learning algorithms to measure toxicity on Seeking Alpha, we find persistent toxicity primarily in the comments rather than articles, with over 50% of firms on the platform experiencing toxicity in recent years. Comment toxicity is greater for firms with more investor attention and disagreement, and for those led by female CEOs. We find three key results. First, toxicity displays a feedback loop in platform participation: past toxicity predicts more future toxic contributors for a given firm. Second, firms receiving more toxic comments have greater retail trading volume but less informative retail trades. Third, toxicity is associated with slower price discovery around earnings announcements, indicating potential broader market efficiency implications. Our findings suggest financial social media toxicity influences both user behavior and market outcomes, raising important considerations for platform governance in financial markets.

This study investigates how social media network structure influences short-term stock market reactions to buy recommendations issued by self-identified social media analysts (SMAs) within online retail investor communities. Consistent with social learning theories-and in contrast to social utility theories-we find that communities with lower network cohesion experience more positive short-term market reactions to buy recommendations. However, this negative relation diminishes at higher levels of network cohesion. Further supporting the social learning theories, the impact of network cohesion is more pronounced for SMAs with higher recommendation accuracy. In line with behavioral finance theories, we also find the effect of network cohesion to be stronger for stocks facing greater limits to arbitrage. Our exploratory analysis suggests that part of the initial price reaction reflects investor overreaction, as indicated by subsequent price reversals.

We study non-traditional investment research as a source of information for corporate strategic decisions, such as investments into innovation. Using a comprehensive sample of social media analyst reports from Seeking Alpha and exogenous variation in social media analysts' coverage overlaps, we show that firms are more likely to invest into technologies similar to firms covered by the same analyst. The effect varies with social media analysts' characteristics and differences in their contributed content that capture their unique information set. Overall, our results are consistent with non-traditional investment research enhancing firms' information environment as an additional source of information that guides corporate strategic decisions.

Does social media amplify bank fragility absent systemic crises (e.g., the SVB crisis)? Using a sample of U.S. commercial banks from 2009 to 2022, we show that heightened Twitter attention increases the sensitivity of non-core deposits-but not core deposits-to bank performance deterioration. This effect intensifies for banks with greater liquidity mismatch and when Twitter discussions are more influential. Neither enhanced bank transparency nor negative sentiment in social media discussions explains these results. Our findings indicate that social media is not merely an information transmitter; it heightens depositors' awareness of peer attention to banks, amplifying deposit outflow sensitivity to weak fundamentals even during calm periods.

Our research augments the expanding body of literature concerning the capability of prevalent LLM tools in supporting financial professionals. We introduce a framework to leverage ChatGPT to assess market sentiment through the analysis of social media data. We use the LLM models to construct market sentiment indicators based on Twitter tweets and use those indicators to explain Bitcoin return. Our analysis uncovers that sentiment indicators crafted with ChatGPT4o/ChatGPT3.5 significantly affect Bitcoin returns, even when accounting for a broad array of control variables and other pre-established sentiment indicators. These insights imply that ChatGPT4o/ChatGPT3.5 could empower financial professionals to discover sentiment information from Twitter tweets that were overlooked by previously introduced sentiment indicators concerning Bitcoin.