Behavioral Finance Meeting, Fall 2021

Nagel and Xu examine subjective risk premia implied by return expectations of individual investors and professionals for aggregate portfolios of stocks, bonds, currencies, and commodity futures. While in-sample predictive regressions with realized excess returns suggest that objective risk premia vary countercyclically with business cycle variables and aggregate asset valuation measures, the researchers find subjective risk premia extracted from survey data do not comove much with these variables. This lack of cyclicality of subjective risk premia is a pervasive property that holds in expectations of different groups of market participants and in different asset classes. A similar lack of cyclicality appears in out-ofsample forecasts of excess returns, which suggests that investors' learning of forecasting relationships in real time may explain much of the cyclicality gap. These findings cast doubt on models that explain time-varying objective risk premia inferred from in-sample regressions with countercyclical variation in perceived risk or risk aversion. Nagel and Xu further find a link between subjective perceptions of risk and subjective risk premia, which points toward a positive risk-return tradeoff in subjective beliefs.

Whether teams attenuate or exacerbate the behavioral biases which are pervasive at the individual level is an open question. To address this question, Barahona, Cassella, and Jansen use the mutual fund industry as a laboratory. Their focus is on how return extrapolation, a bias in investor behavior that has received considerable attention in recent work, is transmitted from individual fund managers to the team-managed funds they join. The researchers show that teams heavily attenuate the influence of extrapolation bias on funds' trading behavior. Additional analysis reveals that this attenuation is not due to differences in investment experience, compensation contracts, workload, and investment styles between solomanaged and team-managed funds. Rather, Barahona, Cassella, and Jansen's evidence suggests that the elicitation of team members' inner cognitive reflection can be responsible for teams' reduction in behavioral biases.Their results highlight the attenuation of the extrapolation bias as a potential benefit of team-based asset management.

Motivated by evidence of systematic forecast errors by market participants and professional forecasters, Valente, Vasudevan, and Wu construct a model of exchange rate determination where investors each (1) receive noisy private signals about the future path of interest rate differentials between the US and other countries and (2) overestimate the persistence of interest rate differentials. Their model is able to explain the forward premium puzzle, a well-known failure of the uncovered interest rate parity condition implied by traditional models (UIP), in a manner consistent with the survey evidence, in addition to a number of additional puzzles that existing models have struggled to simultaneously explain. These include the initial underreaction and delayed overreaction of currencies in response to monetary news; positive short-horizon and negative long-horizon autocorrelations of currency excess returns; and the lower return predictability of interest rate differentials for UIP trades implemented with longer maturity bonds. Valente, Vasudevan, and Wu's model is also useful for understanding the strong relationship between survey-based measures of macroeconomic news and exchange rates despite the weak relationship between macroeconomic fundamentals and exchange rates, the persistence of subjective beliefs, and the seeming reversal of the failure of UIP in recent years. Valente, Vasudevan, and Wu's results highlight the important role that investors' beliefs may play in exchange rate behavior.

With "2020 hindsight,'' the 2000s housing cycle is not a boom-bust but rather a boom-bust-rebound at both the national level and across cities. Chodorow-Reich, Guren, and McQuade argue this pattern reflects a larger role for fundamentally-rooted explanations than previously thought. The researchers construct a city-level long-run fundamental using a spatial equilibrium regression framework in which house prices are determined by local income, amenities, and supply. The fundamental predicts not only 1997-2019 price and rent growth but also the amplitude of the boom-bust-rebound and foreclosures. This evidence motivates their neo-Kindlebergerian model, in which an improvement in fundamentals triggers a boom-bust-rebound. Agents learn about the fundamentals by observing "dividends'' but become over-optimistic due to diagnostic expectations. A bust ensues when over-optimistic beliefs start to correct, exacerbated by a price-foreclosure spiral that drives prices below their long-run level. The rebound follows as prices converge to a path commensurate with higher fundamental growth. The estimated model explains the boom-bust-rebound with a single fundamental shock and accounts quantitatively for cross-city patterns in the dynamics of prices and foreclosures.

Individual forecasts of economic variables show widespread overreaction to news, but laboratory experiments on belief updating typically find underinference from signals. Fan, Liang, and Peng provide new experimental evidence to connect these two seemingly inconsistent phenomena. Building on a classic experimental paradigm, Fan, Liang, and Peng study how people make inferences and revise forecasts in the same information environment. Subjects underreact to signals when inferring about underlying states, but overreact to signals when revising forecasts about future outcomes. This gap in belief updating is largely driven by the use of different simplifying heuristics for the two tasks. Additional treatments link their results to the difficulty of recognizing the conceptual connection between making inferences and revising forecasts.

In the past decade, researchers in psychology and neuroscience studying human decision-making have increasingly adopted a framework that combines two systems, namely "model-free" and "model-based" learning. Barberis and Jin import this framework into a simple financial setting, study its properties, and link it to a wide range of applications. Barberis and Jin show that it provides a foundation for extrapolative demand and experience effects; resolves a puzzling disconnect between investor allocations and beliefs in both the frequency domain and the cross-section; helps explain the dispersion in stock market allocations across investors as well as the inertia in these allocations over time; and sheds light on the persistence of household investment mistakes. More broadly, the framework offers a way of thinking about individual behavior that is grounded in recent evidence on the computations that the brain undertakes when estimating the value of a course of action.