Asset Pricing Impacts of AI in a Model with Heterogeneous Risk Aversion

Abstract

We develop a heterogeneous-agent model with two channels through which AI affects the economy: an increase in expected growth rate and volatility of output, and higher agents' risk tolerance. We use this framework to study the impacts of AI on portfolio choices and asset prices at equilibrium. Specifically, we show that AI changes the optimal portfolios of both more and less risk-averse agents, increases the price-dividend ratio and stock volatility, and reduces the interest rate, price of risk, and aggregate leverage compared to an economy without these two AI channels. Our framework can be extended and used to study the financial market implications of AI.

Keywords: risk aversion, asset pricing, artificial intelligence, ChatGPT, preference heterogeneity

JEL Classification: G12, O30, O33, G11