Elastic attention, induced uncertainty and aggregate fluctuations

Abstract

This dissertation explores the macroeconomic impact of induced uncertainty that stems from decision maker behavior: (i) finite information-processing capacity, or state uncertainty due to rational inattention; (ii) concern about model misspecifications, or model uncertainty due to robustness. It is illustrated that by incorporating these two microeconomic frictions I can better account for the inconsistency between standard full-information rational expectation models and empirical stylized facts. The first part of this thesis examines the implications for international real business cycles of a small open economy (SOE); the second part analyzes the effect on households' saving and aggregate wealth.

In Chapter 1 (joint work with Yulei Luo and Jun Nie), we propose a novel explanation for one of the six major puzzles in international economics, the international consumption correlation puzzle. To the best of our knowledge, this is the first attempt to consider finite information-processing capacity in solving this puzzle. In a canonical open economy model under the complete market assumption, risk-averse consumers will insure country-specific risk using international financial markets, which leads to highly, even fully correlated consumption regardless of output correlations. However, the empirical evidence suggests that cross-country consumption is far from perfectly correlated; in fact they are lower than output correlation in most cases for G7 or OECD countries. To mitigate the discrepancy, we introduce optimal (or elastic) attention into an otherwise standard small open economy model. Agents in the model are assumed to face information-processing costs thus have limited capacity to process information when making major economic decisions. The optimal information-processing capacity is chosen in response to the amount of fundamental uncertainty, which differs across countries by nature. This can help endogenously generate heterogeneous and gradual consumption adjustments to income shocks across countries, thus improving model-fit with respect to empirical data in explaining international consumption and income correlations as well as some other key stochastic properties of the joint dynamics of consumption and income in individual countries.

In Chapter 2, I analyze the determinants of aggregate wealth accumulation in an infinite-period overlapping-generations model with uninsurable income risk. The work incorporates finite information-processing capacity and the preference for robustness into the model in which decision makers are aware of model misspecifications of the distribution of future shocks and optimize against the worst-case model specification. These two types of induced uncertainty interact with income risk and growth in determining the accumulation of wealth. The interactions help better explain the relative volatility of wealth to income documented in U.S. survey data. Furthermore, it shows that the role of the induced uncertainty is prominent especially in an economy experiencing high growth.