Thesis - Open Access
Master of Science (MS)
Department / School
Adaptive Learning Gains, Asset Pricing, Equity Risk Premium, Information Structure, Kalman Filter, Uncertainty
This paper delves into the complexities of asset pricing, emphasizing the need to go beyond prevailing paradigms and constant learning gain assumptions. We examine the influence of personal experiences, adaptive learning processes, and subjective return expectations on asset pricing. By incorporating the concept of time-varying learning gain, we provide a more realistic portrayal of asset pricing. Empirical analysis reveals a consistent negative correlation between experienced real payout growth and subsequent returns, indicating counter-cyclical behavior. Our findings also support the mean-reversion hypothesis in stock returns, although caution is needed due to some scenarios lacking statistical significance. Theoretical exploration uncovers that higher uncertainty or variability compels investors to seek additional compensation, thus elevating the equity risk premium. Moreover, the information structure does not form a filtration, leading to no convergence to a specific value in the long run. Agents perceive future increments as negatively serially correlated but lack the memory to effectively exploit this correlation for forecasting. Consequently, the Law of Iterated Expectations does not hold. We propose the "resale" valuation method as ideal for agents with adaptive learning gains. These findings contribute to an innovative asset pricing model with adaptive learning gains, enhancing our understanding of market dynamics. While this study does not provide calibration or validation, we outline the model’s theoretical foundations and implications for future research. Our work adds to the evolving landscape of asset pricing theory, highlighting the significance of adaptive learning in capturing complex dynamics.
South Dakota State University
Lokossou, Sedealy Juste, "Adaptive Learning Gain in Asset Pricing" (2023). Electronic Theses and Dissertations. 716.