Zu Yao Hong

Abstract: This paper examines how beliefs of tail risk events influence macroeconomic expectations in a Bayesian learning model with noisy signals. We show theoretically and empirically that the misperception of tail risk results in overreaction to first and second moment shocks, in contrast to a Gaussian model. First moment shocks cause excessive optimism and pessimism in individuals as they provide valuable information about tail risk. Second moment shocks lead to more pessimistic forecasts as higher uncertainty is linked to an increased likelihood of disasters. As signals become noisier, the response to news regarding a first moment shock becomes more pronounced. Our findings shed light on factors driving overreaction in expectations and highlight the importance of uncertainty shocks in propagating macroeconomic stability.

Abstract: This paper examines how price-based learning with misperception of policy effects influences the conduct of macroprudential policy. We show that distorted beliefs, reflected in overoptimism, arise endogenously from policy interventions when agents mistakenly attribute policy-driven price movements to changes in fundamentals. As a result, beliefs about the aggregate state are updated incorrectly even when underlying fundamentals remain unchanged. Embedding this mechanism in a standard open-economy framework, we find that when regulators fail to internalize how their policies affect prices and beliefs, intervention can yield lower welfare than non-intervention. The optimal policy adjustments depend on whether price-based mislearning amplifies or offsets the intended effects of regulation.

*Previously circulated under the title "Learning from Prices, Credit Cycles, and Macroprudential Policies"

Abstract: We propose a theory of international uncertainty contagion in a two-country model with non-rivalrous news production. In each country, information acquisition occurs through news production and is allocated between a global aggregate process and a country-specific idiosyncratic process. Because aggregate news is non-rivalrous and can be reproduced across countries, one country’s information choices generate externalities for the other. When uncertainty about the foreign idiosyncratic process rises, these cross-border informational externalities induce strategic substitutability: the foreign country reallocates news production toward its country-specific process, while the domestic country shifts toward aggregate news. This raises posterior uncertainty about domestic conditions in both countries, resulting in international uncertainty contagion. The effect is stronger under high aggregate uncertainty, when reliance on reproduced foreign aggregate news is greater. Embedded in a two-country RBC model, our mechanism can reduce the output divergence typically generated by country-specific shocks and generate stronger cross-country output comovement. It also provides a complementary channel for risk-off-type dynamics, in which currency depreciations and capital outflows in riskier countries arise from country-specific rather than global shocks. 

Abstract: This paper introduces information quality in a real business cycle model. Information quality relates to the idea that information obtained can inaccurately reflect the actual state of the economy. Using the Survey of Professional Forecasters, I document that forecast errors are larger during downturns, even if agents acquire more information. I then augment a rational inattention model with information search frictions that generate variable information quality. Information depends on both data abundance and information search intensity. Unlike rational inattention models, which are demand driven, I allow for time-varying data abundance, or information supply, generating fluctuations in information quality. The model delivers pro-cyclical information quality, which rationalizes puzzling evidence that information acquisition and uncertainty increase in downturns. A Bayesian estimation of the model for the US economy shows that information quality accounts for sizable fluctuations in uncertainty and output. The model also generates: (i) systematic mistakes when agents do not internalize fluctuations in information quality, (ii) variation in information processing costs, which produce higher frequency and dispersion in price changes during downturns, and (iii) production externalities, as firms do not internalize that more activity generates data abundance, which reduces uncertainty.