Published Papers

We study the nature and effects of cultural biases in choice under risk and uncertainty by comparing peer-to-peer loans the same individuals (lenders) make alone and after observing robo-advised suggestions. When unassisted, lenders are more likely to choose co-ethnic borrowers, facing 8% higher defaults and 7.3pp lower returns. Robo-advising does not affect diversification but reduces lending to high-risk co-ethnic borrowers. Lenders in locations with high inter-ethnic animus drive the results, even when borrowers reside elsewhere. Biased beliefs explain these results better than a conscious taste for discrimination: lenders barely override robo-advised matches to ethnicities they discriminated against when unassisted.

We find evidence of selective exposure to confirmatory information among 400,000 users on the investor social network StockTwits. Self-described bulls are five times more likely to follow a user with a bullish view of the same stock than are self-described bears. Consequently, bulls see 62 more bullish messages and 24 fewer bearish messages than bears do over the same 50-day period. These "echo chambers" exist even among professional investors and are strongest for investors who trade on their beliefs. Finally, beliefs formed in echo chambers are associated with lower ex post returns, more siloing of information, and more trading volume.

In this article, we examine whether comovement in the stock prices of pairs of Chinese firms connected to the same political network are systematically shaped by the prevailing coordination versus competition incentives of that network's politicians. We find strong evidence from 2000 to 2012 (Jiang's and Hu's regimes) that stock price comovement is affected by the embeddedness of the firm-politician ties within the network. Among pairs of firms connected to a network through a common politician, we document an increase in stock price comovement. For those pairs of firms connected to a common network via separate politicians (rather than a common politician), we document a relative decrease in stock price comovement. This negative effect suggests that politicians' relationships within these political networks are generally adversarial rather than cooperative in nature. These results become significantly weaker during Xi's regime from 2013 to 2017, suggesting that Xi's anti-corruption campaign and state-owned enterprise reforms may have attenuated these political network effects on the firms. Our additional tests also show that stock price comovement becomes even more positive (negative) in settings which are expected to increase the coordination or decrease the competition (decrease the coordination or increase the competition) of the politicians.

Using hand-collected data spanning more than a decade on European banks' sovereign debt portfolios, we show that the trust of residents of a bank's countries of operation in the residents of a potential target country of investment has a positive, statistically significant, and economically important association with its cross-border exposures. In identifying cultural stereotypes at the bank level, we show that corporate culture at bank headquarters is influenced by foreign subsidiaries for several reasons, including banks' tendency to hire internally across borders for high-level managerial positions. We therefore leverage the geography of multinational bank branch networks to construct a bank-specific measure of culture that differs across banks headquartered in the same country, at the same point in time, with regard to the same target country. This allows us to compare how sovereign exposures are affected by cultural stereotypes while ruling out confounding factors at country and country-pair levels. The effect of stereotypes is persistent overtime, stronger for less diversified banks, and weaker for target countries whose bonds appear more frequently in bank portfolios. Cultural stereotypes are particularly salient when governments are hit by sovereign debt crises.

Using machine learning-based algorithms, we measure key impressions about sell-side analysts using their LinkedIn photos. We find that impressions of analysts' trustworthiness (TRUST) and dominance (DOM) are positively associated with forecast accuracy, especially after recent in-person meetings between analysts and firm managers. High TRUST also enhances stock return sensitivity to forecast revisions, especially for stocks with high institutional ownership. In contrast, the impression of analysts' attractiveness (ATTRACT) is only positively associated with accuracy for new analysts or when a firm has a new CEO or CFO. Furthermore, while high DOM helps male analysts' chances of attaining All-Star status, it reduces female analysts' accuracy and the likelihood of winning the All-Star award. In addition, the relation between TRUST and accuracy is modulated by the disclosure environment and is attenuated by Regulation Fair Disclosure. Our results suggest that face impressions influence analysts' access to information and the perceived credibility of their reports.

We analyze the partisanship of Commissioners at the SEC and Governors at the Federal Reserve Board. Using recent advances in machine learning, we identify partisan phrases in Congress, such as "red tape" and "climate change," and observe their usage among regulators. Although the Fed has remained relatively nonpartisan throughout our sample period (1930-2019), we find that partisanship among SEC Commissioners rose to an all-time high during the 2010-2019 period, driven by more-partisan Commissioners replacing less-partisan ones. Partisanship at the SEC appears in both the language of new SEC rules and the voting behavior of SEC Commissioners.

Maintaining economic output during the COVID-19 pandemic results in benefits for firm shareholders but comes at a potential cost to public health. Using store-level data, we examine how a CEO's political leaning impacts this trade-off. We document that firms with a Republican-leaning CEO experience a relative increase in store visits compared to firms with a Democratic-leaning CEO. The increase in store visits is associated with higher sales and positive abnormal stock returns. However, we also document higher COVID-19 transmission rates and more employee safety complaints in communities where establishments with higher store traffic are managed by a Republican-leaning CEO.

Using data from a Twitter-like investor social platform, we document evidence consistent with self-enhancing transmission bias. We find investors are more likely to post about their better-performing stocks. Their followers are more likely to buy the posted stocks than other non-posted stocks. The effect of the postings on follow-up purchases is consistent with an attention-based interpretation: the postings bring the discussed stocks into the followers' choice set and increase their purchases. We also find that postings' effect on follow-up purchases is related to postings' perceived credibility. The performance-postings relationship is stronger among more volatile stocks and the relationship between postings and follow-up purchases is stronger among stocks with higher recent returns, shedding light on the spread of high-variance and extrapolative strategies. We also document that the social network features influential nodes.

We educate investors about the benefits of dividend reinvestment and costs of misperceiving dividends as free income. The intervention increases planned dividend reinvestment in survey responses. Using trading records, we observe a causal increase in dividend reinvestment in the field of roughly 50 cents for every euro received. This holds relative to investors' prior behavior and various control samples. Investors who learned the most from the intervention update their trading the most. The results suggest the free dividends fallacy is a significant source of dividend demand. Our study demonstrates that simple, targeted, and focused educational interventions can affect investment behavior.

We examine the effect of personal, two-way communication on the payment behavior of delinquent borrowers. Borrowers who speak with a randomly assigned bank agent are significantly more likely to successfully resolve the delinquency relative to borrowers who do not speak with a bank agent. Call characteristics related to the human touch of the call, such as the likeability of the agent's voice, significantly affect payment behavior. Borrowers who speak with a bank agent are also significantly less likely to become delinquent again. Our findings highlight the value of a human element in interactions between financial institutions and their customers.

We study the influence from social interactions on equity trading. Using unique data on stock transactions, we exploit the quasi-random assignment of students to classrooms in a financial training program to identify how peer experience affects investor behavior. We find that individuals react more to peer gains than to peer losses. Students enrolled in courses where peers have positive outcomes: (i) are more likely to start trading, (ii) purchase similar stocks as their classmates, and (iii) are disproportionally attracted to stocks with extreme returns. These stocks have low subsequent returns, and new investors reacting to peer gains underperform other investors.

We study the role of social networks in hedge fund activism. Actively managed funds whose managers are socially connected to activists are more likely than unconnected managers to invest in target stocks; their investment decisions are profitable. Importantly, such effects are greater for funds facing more severe information asymmetry. Connected funds are 14.2 percentage points more likely to support activists in proxy contests and contribute to reducing proxy contest costs. Our evidence shows that social ties benefit both connected investors and activists, and suggests that social networks reduce information asymmetry around activist campaigns by facilitating information exchange and increasing trust.