When making economic decisions, individuals may struggle with uncertainty from two sources: doubt about their own forecasting ability, and doubt about the accuracy of the data guiding their decision.
Martin Ellison, Pascal Paul and Thomas Sargent examine whether or not asset prices can be better explained by incorporating both these forms of ambiguity about consumption dynamics into an economic agent’s problem. Specifically, they ask whether fear of misspecification in the model used for forecasting consumption and fear of misspecifying the data used for those forecasts can explain asset pricing.
Ellison outlined this work in a poster session at the Ambiguity and Robustness in Macroeconomics and Finance conference. In their model, individuals are unsure about the state of the economy. Even if they know the current state, they aren’t sure about how it influences tomorrow’s state. The authors use previous theoretical work by Sargent and Hansen to separate the two sources of uncertainty; they apply that work to distill what happens to asset pricing due to the contribution of uncertainty about the state of the world alone. This allows them to layer on uncertainty about forecasting rule as well, to see what additional changes in behavior occur when both types of uncertainty are present.
When individuals are uncertain about their forecasting rule, they act as if the distribution of future consumption growth rates is shifted downward; they are pessimistic about growth possibilities. Adding uncertainty about the state of the world to the model, the authors show that individuals also act as if the persistently low-growth state of the world is more likely.
Ellison and his coauthors show that ambiguity concerning the forecasting rule primarily impacts the level of the market price of risk, without strongly impacting the volatility. Adding ambiguity about the state of the world adds volatility to the risk premium. In this way, ambiguity can be used to control both the level of the risk premium and its volatility.
This work advances the use of ambiguity as a tool to understand the equity premium and the volatility of that premium by studying fear of model misspecification both about forecasting rules and of the data used for those rules. This allows them to introduce enough flexibility to explain both the level of the market price of risk and the variance of risk.