Research

Using hourly Robinhood holdings, we find retail investors do not adjust their portfolios in response to firms’ ESG disclosures, even though they trade on earnings and other non-ESG news. This revealed-preference evidence questions the case for ESG disclosure aimed at retail investors. Cited in connection with the SEC’s climate-disclosure rulemaking.

Financial fake-news authors time disinformation to high-attention earnings windows but avoid firms with stronger disclosure and the days just after earnings releases. Accounting information thus protects investors twice: by reducing incentives to produce fake news and by dampening its market impact.

Local earnings news released before a U.S. presidential election shifts voters toward the incumbent party, with an effect comparable in size to state GDP growth. A post-election falsification ties the result to the information in earnings rather than to underlying economic conditions.

In a pre-registered field experiment across 2,731 U.S. public companies, White investors receive meeting approvals with corporate insiders 34% more often than Black investors, and the gap persists even when investors signal professional sophistication.

Presenting earnings visually leads retail investors to trade on earnings more, but not better. Visualization lowers the cost of forming a signal yet does not improve returns, as investors trade against the information’s predictive content.

Using proprietary data on how long investors listen to earnings calls, we introduce sustained attention — the duration of engagement — as distinct from whether investors show up. Sustained attention predicts higher trading volume, larger absolute returns, and faster price discovery, even after controlling for the number of listeners.

Using complete data on retail investors’ interactions with an AI investment adviser, we find that greater human intervention degrades performance — adding risk through return-chasing and concentration without raising returns. Investors’ own input attenuates the benefits of AI-assisted investing.

Using spinoff Form 10 filings, we show that consolidated reporting leaves investors with incomplete visibility into material business-unit information. Investors respond most to business-unit content that is new relative to the parent’s prior 10-K, and that information spills over to industry peers.

This registered report elicits consumers’ willingness-to-accept for sharing real bank and brokerage data, delivering the first incentive-compatible estimate of what financial privacy is worth — and how it varies by data scope, institutional trust, income, and who is watching.

Combining an LLM-persona methodology with experimental evidence, we measure how the public perceives the SEC over time and link that perception to financial-market engagement. More favorable perception is associated with greater participation and stronger responses to earnings news.

Investors price ESG news mainly when its direction aligns with the firm’s most recent earnings, not as a standalone signal — cautioning against designs and strategies that treat ESG news as independent information.

Retail investors use stock-price movements as a signal that new public information exists, then seek out and trade on that information. Price changes thus act as an attention-triggering monitoring device.

Using the 2007 NBA betting scandal as a natural experiment, we show that a directly observed informed trader increases price volatility and lowers liquidity, with effects that vanish once the trader exits — including direct evidence of strategic price bluffing that Kyle’s (1985) model rules out under linear strategies.