FAS Theses and Dissertations

Macroeconomics and financial economics have long taken expectations seriously in theory and ignored them in measurement. Both fields rest on canonical assumptions — rational expectations in macroeconomics, efficient markets in financial economics — that have become increasingly hard to defend. This dissertation pushes past these benchmarks. Three studies develop disciplined behavioral alternatives grounded in cognition, formalize their implications, and test them in the data. The first essay extends the canonical decomposition of unexpected stock returns to allow for biased cash-flow beliefs, and shows that a meaningful share of what the literature reads as risk premium news is belief-driven rather than risk-driven. The second essay develops and causally tests a memory-based theory of macroeconomic expectations, establishing that selective retrieval of past experiences shifts beliefs about the macroeconomy independently of any change in information. The third essay provides a cognitive micro-foundation for animal spirits and confidence multipliers, embedding memory-based retrieval in aggregate demand so that current outcomes cue past experiences that shape expectations about the future. Collectively, the three chapters of this thesis argue, across very different empirical environments, that the foundational assumptions of rational expectations and efficient markets are systematically violated in tractable and measurable ways.