Research

Long abstract: Is excessive risk-taking in credit cycles driven by incentives or biased beliefs? I propose a framework suggesting that the two are actually related and, specifically, that procyclical overoptimism can arise rationally from risk-taking incentives. I show that when firms and banks have a limited liability payoff structure, they have lower incentives to pay attention to the aggregate conditions that generate risk. This leads to systematic underestimation of the accumulation of risk during economic booms and overoptimistic beliefs. As a result, agents lend and borrow excessively, further increasing downside risk. Credit cycles driven by this new "uninformed" risk-taking are consistent with existing evidence such as high credit and low risk premia predicting a higher probability of crises and negative returns for banks. My model suggests that regulating incentives can decrease overoptimistic beliefs and thus mitigate boom-and-bust cycles.

Long abstract: We find evidence suggesting that surveys of professional forecasters are biased by strategic incentives. First, we find that individual forecasts overreact to idiosyncratic information but underreact to common information. Second, we show that this bias is not present in forecast data that is not subject to strategic incentives. We show that our evidence is consistent with a model of strategic diversification incentives in forecast reporting. Our results caution against the use of surveys of forecasts as a direct measure of expectations, as this would overestimate the likely deviations from rational expectations, the information precision, and the degree of disagreement.

Long abstract: We study a small open economy in which the exchange rate aggregates private information about macro fundamentals, yet imperfectly because of non-fundamental financial flows. Foreign exchange interventions influence economic conditions via traditional portfolio balance and novel informational effects. Publicly announced interventions reveal the central bank's knowledge about macro fundamentals, while secret interventions alter the informational content of the exchange rate. When the private sector overreacts to public information, the optimal policy is to intervene secretly and tolerate some nonfundamental fluctuations in the exchange rate in order to limit its informativeness. Our results speak to practical questions about exchange rate policy design.

Long abstract:  We study how different sources of uncertainty shape household expectations about inflation. Using the Survey of Consumer Expectations, we show that belief stickiness declines when prior information is uncertain, but rises when new information is uncertain. While broadly consistent with Bayesian updating, households overreact relative to the Rational Expectations benchmark. We use this framework to decompose the recent rise in inflation expectation uncertainty: during the pandemic, macroeconomic volatility rendered priors more uncertain; during the high-inflation period, uncertainty stemmed from noisier signals. Belief stickiness offers a valuable statistic to distinguish sources of uncertainty in expectations data.

Long abstract: I show that information choice can explain the puzzling positive relation between bilateral investment and trade across countries. I present a model of endogenous information with both investment in assets and income from trade. While standard model of risk-hedging would require agents to invest in non-trading countries to diversify income risk, I show that limited information capacity and preferences for early resolution of uncertainty reverse this result. The intuition is that investors collect more information on trading partners to reduce income uncertainty, and therefore perceive their equity as less risky. I find that allowing for information choice reduces the role of risk hedging on portfolio decisions. I test my model’s implied relation between trade and attention in the data and find robust empirical support.