MODES · Modes of Expectation Formation in Macro and Finance

Expectations are central in macroeconomics and finance. I propose to study modes of expectation formation in macro and finance: the ways in which expectations are constructed. These modes can differ in the inputs used and in how these inputs are processed. For instance, agents could consider relevant macroeconomic state variables and process this information through a developed mental model of the economy. A different approach could be based on spontaneous recall of easily accessible personal experiences and use of a simple heuristic.Empirical studies suggest that different modes are relevant, but we do not understand their relative importance, drivers, decision-relevance, and aggregate consequences. This poses a barrier to a unified understanding of belief formation, preventing the field from moving forward. My objective is to answer these questions for the most central domains in macro and finance. I use a novel portable, survey-based approach building on an AI chat, laying the groundwork for a unified account of macroeconomic expectation formation.In Part 1, I study modes of forming inflation expectations among households, managers and experts. I use experiments to understand how rational inattention and memory shape modes of belief formation. I use a New-Keynesian model to study macro consequences.In Part 2, I study modes of expectation formation about aggregate stock returns using surveys with investors. I use linked administrative bank account data to study the consequences of the mode of belief formation for the pass-through of beliefs to decisions. I explore aggregate implications in an asset pricing model.In Part 3, I study modes of expectation formation about aggregate unemployment. I conduct household surveys linked with administrative register data to understand drivers such as cognitive skills, exposure to aggregate fluctuations, and first- and second-hand past job loss experiences. I use a search and matching model to study macro implications.