NATIONAL BUREAU OF ECONOMIC RESEARCH

## Behavioral Finance

Members of the NBER's Behavioral Finance Working Group met in Cambridge on November 2. Research Associate Nicholas C. Barberis of Yale University organized the meeting. These researchers' papers were presented and discussed:

Cary Frydman, University of Southern California, and Lawrence J. Jin, California Institute of Technology

Efficient Coding and Risky Choice

Frydman and Jin present a model of risky choice in which the perception of a lottery payoff is noisy and optimally depends on the payoff distribution to which the decision maker has adapted. The perceived value of a payoff is precisely defined according to a core idea in neuroscience called the efficient coding hypothesis, which indicates that more perceptual resources are allocated to those stimuli that occur more frequently. The researchers show that this principle implies that, for a given choice set of lotteries, risk taking varies systematically with the recently encountered distribution of payoffs. The model is tested in two laboratory experiments. In the first experiment, the researchers manipulate the distribution from which payoffs are drawn. Consistent with efficient coding of lottery payoffs, they find that risk taking is more sensitive to payoffs that are encountered more frequently in the choice set. Furthermore, sensitivity to extreme payoffs is initially small, but grows over time after repeated exposure. The second experiment consists of a purely perceptual task, in which subjects classify which of two symbolic numbers is larger. The researchers find that accuracy depends on the distribution of numbers to which the subject has adapted, which provides support for the key model assumption that perception of numerical payoffs is noisy and changes across environments.

Huseyin Gulen, Purdue University; Mihai Ion, University of Arizona; and Stefano Rossi, Bocconi University

Credit Cycles and Corporate Investment

Gulen, Ion, and Rossi study the real effects of credit market sentiment on corporate investment and financing for a comprehensive panel of U.S. public and private firms over 1963-2016. In the short term, they find that high credit market sentiment in year t correlates with high corporate investment and debt issuance in year t+1, particularly for financially constrained firms. In the longer term, high credit market sentiment in year t correlates with a decline in debt issuance in years t+3 and t+4, and with a decline in corporate investment in years t+4 and t+5. This pattern of increased investment in the short term and declined investment in the longer term is more pronounced for firms with larger analysts' earnings forecast revisions and comes with larger analysts' forecast errors, supporting theories of over-extrapolation of fundamentals into the future.

Alexander M. Chinco, University of Illinois at Urbana-Champaign

The Madness Of Crowds And The Likelihood Of Bubbles

Market participants are constantly swimming in a sea of psychological biases and trading constraints. And yet, in spite of all these biases and constraints, large pricing errors such as speculative bubbles are rare. Why is this? How often should we expect that some psychological bias will cause some trading constraint to bind? Chinco proposes a model to answer this question.
In the model, the number of speculators excited about an asset varies over time due to social interactions. As long as the asset's past returns remain below a critical threshold, r < r_**, these social interactions will disperse any crowd of speculators that happens to get excited about the asset. But, as soon as the asset's past returns rise above this critical threshold, r > r_**, the exact same social interactions will suddenly make the excited-speculator population boom. The resulting population explosion amplifies the effect of speculators' pre-existing psychological biases, causing arbitrageur constraints to bind and a speculative bubble to form.
The model predicts that speculative bubbles will be more common among assets whose speculators are more sensitive to fluctuations in past returns. More importantly, it also suggests how to estimate this key sensitivity parameter using data collected during normal times -- i.e., when there's no speculative bubble currently taking place. And, Chinco verifies this prediction empirically: after similar price run-ups, industries in the top sensitivity quintile are more than twice as likely to experience a speculative bubble than those in the bottom quintile.

Manuel Adelino, Duke University and NBER; Antoinette Schoar, MIT and NBER; and Felipe Severino, Dartmouth College

Perception of House Price Risk and Homeownership (NBER Working Paper No. 25090)

Adelino, Schoar, and Severino analyze the importance of households' perceptions of house price risk in explaining homeownership choice. While a majority of US households (71%) believes that housing is a "safe" investment, renters are much more likely to perceive housing as a risky investment (conditional on income, savings, location and all other observables). Risk perceptions vary across demographic groups, but significant differences persist after controlling for observables. Current housing decisions and future intentions to buy versus rent are strongly correlated with perceptions of house price risk. Households' exposure to housing risk due to financial constraints, expected mobility or labor income risk affect the decision to buy versus rent but do not mitigate the impact of risk perceptions on housing choices. Finally, the researchers show that households update their beliefs about the riskiness of housing in response to recent local house price changes. But renters are much slower to update than owners, which might explain the staggered entry into home ownership of different groups in response to house price increases.

Carolin Pflueger, University of British Columbia and NBER; Emil Siriwardane, Harvard University; and Adi Sunderam, Harvard University and NBER

A Measure of Risk Appetite for the Macroeconomy (NBER Working Paper No. 24529)

Pflueger, Siriwardane, and Sunderam document a strong and robust positive relationship between real rates and the contemporaneous valuation of volatile stocks, which they contend measures the economy's risk appetite. The researchers proxy for risk appetite explains 41% of the variation in the one-year real rate since 1970, while the valuation of the aggregate stock market explains just 1%. In addition, the real rate forecasts returns on volatile stocks, confirming the interpretation that changes in risk appetite drive the real rate. Increases in the measure of risk appetite are followed by a boom in investment and output.

Shimon Kogan, MIT; Tobias J. Moskowitz, Yale University and NBER; and Marina Niessner, AQR Capital Management

Fake News: Evidence from Financial Markets

Using a unique dataset of fake stock promotion articles prosecuted by the Securities and Exchange Commission, Kogan, Moskowitz, and Niessner examine the impact of fake news. In addition, they use a linguistic algorithm to detect deception in expression for a much larger set of news content using the fake articles as a training sample. The researchers find increased trading activity and temporary price impact from fake news about small firms, but no impact for large firms. Using the SEC investigation as a shock to investor awareness of fake news, the researchers find a marked decrease in reaction to news, particularly content deemed less authentic, but also legitimate news. These findings, including the indirect spillover effects on other news, are most pronounced for small firms with high retail ownership and for the most circulated articles. Understanding the motivation behind the fake articles, the researchers find that small firms engage in corporate actions and insider trading designed to profit from the fake articles, consistent with concerns of coordinated stock price manipulation. No such patterns are observed for large firms. The setting offers a unique opportunity to quantify the direct and indirect impact of fake news.

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email: info@nber.org