Working Papers

We provide evidence for a new propagation mechanism of wealth inequality and mobility. Using unique administrative data and a quasi-field experiment of exogenous assignment, we find that educated entrants, faced with greater local wealth inequality and salient cases of wealth mobility, take financial, real, and self-employment risks and reach higher positions in the wealth distribution, while the less educated do not. This is driven by poorer communities with more salient cases of wealth mobility, consistent with peer exposure rather than supply-side effects. We find no evidence for other channels, such as obtaining higher-paying more secure jobs, relocating, or reducing debt.

This paper tests how people's moral values influence their views of debt contracts. We ask participants to make decisions about debt contracts in different hypothetical situations (vignettes). We separately measure their moral values using the Moral Foundations Questionnaire (Graham et al., 2009). We have three main sets of findings. First, differences in moral values strongly explain the cross-section of participants' debt decisions. Participants with more conservative values show more support for credit score-based loan pricing, stricter forms of collateral, and tougher bankruptcy resolution. Second, when we randomly change the economic costs and benefits of debt within our vignettes, we find that participants change their answers in the direction predicted by economic theory. Third, participants' beliefs of the functioning of the credit market strongly correlate with their moral values. Participants with conservative values are more likely to believe that strict enforcement and risk-based loan pricing provide incentives and are economically efficient. More liberal participants believe that insurance against unlucky shocks are important. Consistent with moral values being distinct from Bayesian beliefs, financial literacy does not attenuate moral values in shaping beliefs about what is economically efficient.

Using friendship data from Facebook, we find that among three aspects of social capital, Economic Connectedness - the fraction of one's social network with high income, has the strongest and most robust relationship with stock market and saving participation. One standard-deviation greater Economic Connectedness is associated with 10.6% greater stock market participation and 9.2% greater saving participation. Evidence from non-local friendships supports a causal link between household financial behavior and the income of one's friends. Our results indicate that the effect of Economic Connectedness on participation derives from opportunities to interact with high-SES individuals rather than from class-based friending propensities.

We document a causal effect of social interactions on investor behavior using the number of local soccer games as a measure of social interaction intensity. Social transmission is identifiable in buy but not sell trades. Social Interaction Intensity (SII) increases the sensitivity of buying to past buys, particularly in riskier stocks. This sensitivity is an increasing and convex function of past returns, with higher SII further amplifying the effect. Social interactions cause an extremity shift wherein existing shareholders increase their positions, especially within demographically homogeneous communities. Higher social interaction intensity increases the sensitivity of individual investors' trading volume and portfolio riskiness to past trades. At the market level, SII increases the sensitivity of stock trading volume and retail ownership percentage to past buys.

We study the informational value of trading networks in over-the-counter (OTC) markets. Using detailed transaction-level data from the corporate bond market, we show that investors with larger dealer networks make superior trading decisions before changes in credit fundamentals and yield better risk-adjusted performance. We trace these investors' superior trading decisions to trading connections where dealers are most likely to have access to novel credit-relevant information, supporting the interpretation that these investors obtain private information through their larger trading networks. Collectively, our evidence highlights the importance of trading relationships for investors' private information acquisition.

Using data from The Motley Fool's social prediction platform CAPS we document substantial differences in stock predictions across investment horizons. Short- and long-horizon investors respond differently to macroeconomic events and firm news announcements. At the onset of the Covid-19 pandemic the sentiment of short-horizon predictions became sharply more negative while long-horizon predictions remained optimistic. Short-horizon investors also react more than twice as strongly as long-horizon investors to earnings surprises and technical view events. Around acquisition rumors short- and long-horizon investors update in opposite directions about the target: short-term investors become more optimistic while long-term investors become more pessimistic. Motivated by these findings we develop a firm-day measure of horizon disagreement spanning from 2006 to 2022 and find it relates significantly to abnormal trading. Additionally the disagreement-trading relation strengthens on earnings announcement days providing new evidence on the role of model disagreement.

We study how political beliefs influence institutional investors’ portfolio choices and asset prices, using a novel dataset linking fund manager partisanship to holdings. Following Donald Trump’s surprise 2016 election, Republican-majority mutual fund teams sharply increased purchases of high beta stocks and raised equity exposure by approximately 2%, partly financed by elevated inflows. These reallocations affected asset prices: high-beta stocks more exposed to Republican mutual funds earned abnormal returns roughly 25 basis points higher per month. These returns are not explained by firm- or industry-specific news. Finally, Republican-majority fund Sharpe ratios declined, consistent with a deterioration in performance.

Ideology often shapes belief formation, which is central to asset pricing. However, the role of ideological narratives as a source of asset pricing risk remains largely unexplored. Using cryptocurrencies as a laboratory, I examine the role of two ideological narratives-anarchism and decentralization-in the cross-section of cryptocurrency returns. Leveraging social media data and large language models to measure ideology dynamics, I find that fluctuations in these two ideology dynamics are priced in the cross-section of cryptocurrency returns. A two-factor model based on ideological narratives explains the cross-section of cryptocurrency returns better than a three-factor model of crypto market, size, and momentum. Positive shocks to ideology salience are associated with a significant positive spread between more ideology-aligned and less aligned cryptocurrencies, indicating a relative increase in demand for more aligned cryptocurrencies when collective attention to ideological narratives heightens. Consistent with the view that factors proxy for state variables, ideology factors contain distinct information about future crypto market returns and user network growth. Neither investor sentiment nor attention explains the results of the ideology factors. Moreover, the role of ideological narratives extends beyond cryptocurrencies. Stocks with greater exposure to the anarchism narrative yield abnormally high returns that cannot be explained by common stock factor models. The results highlight how ideological narratives contribute to the emergence and adoption of new assets.

Using novel data of social interactions and individual trading records in the Bitcoin market, we document evidence of social learning which leads to sentiment contagion. Investors significantly update their beliefs about Bitcoin in the same direction of average peer sentiment although it is not informative about future price. Our findings indicate inefficiency in social learning, consistent with echo chamber effect and selective interpretation of signals. Moreover, social learning affects both individuals' trading decisions and aggregate market outcomes. We construct a novel measure for the intensity of sentiment contagion due to social learning. It significantly predicts Bitcoin volatility, volume and crash.

Using IRS tax filing data, I show that social network and word-of-mouth communications play an important role in stock market participation decisions. Using a novel dataset from Facebook, I construct a measure of social network friends' participation for US counties and find that a one-percentage point increase in friends' participation increases the focal county participation by 14 to 25 basis points in the following year. For identification strategy, I employ the revelation of financial misconducts as an exogenous negative shock to local participation rate and show that the instrumented change in friend participation significantly and positively predicts the change in focal county participation rates. The increase in participation rates among the low-income households induced by friends' participation decreases the Gini coefficients in metropolitan counties in the following two years. The evidence suggests that social influences and peer effects contribute to the cross-sectional differences in the stock market participation rates across US counties and may lead to lower income inequality.

We study how intercommunity social networks influence demand and supply dynamics in fintech lending platforms. Demand for online loans rises following increases in online borrowing activity in geographically distant but socially connected areas. On the supply side, borrower-area social proximity to deposit-rich regions enhances funding likelihood and is associated with better ex-post loan performance. We establish causality with instrumental variables obtained from natural disasters (demand-side) and financial adviser misconduct (supply-side). We further show that social connectedness is particularly effective in expanding both loan demand and supply in disadvantaged communities, without increasing delinquency rates. These results are consistent with intercommunity social networks raising awareness of alternative lending platforms and transmitting otherwise hard-to-obtain information that mitigates community-level information asymmetry. Although our findings suggest that social networks improve capital allocation and expand credit access for underserved communities, the welfare benefits are muted for low FICO score individuals.

Epidemiological models of belief formation put social interactions at their core; such models are widely used by scholars who are not economists to study the dynamics of beliefs in populations. We survey the literature in which economists attempting to model the consequences of beliefs about the future - expectations - have employed a full-fledged epidemiological approach to explore an economic question. We draw connections to related work on contagion, narrative economics, news/rumor spreading, and the spread of internet memes. A main theme of the paper is that a number of independent developments have recently converged to make epidemiological expectations (EE) modeling more feasible and appealing than in the past.