A Decomposition of Conditional Risk Premia and Implications for Representative Agent Models

161 Pages Posted: 1 Dec 2020 Last revised: 11 Feb 2021

See all articles by Fousseni Chabi-Yo

Fousseni Chabi-Yo

University of Massachusetts Amherst - Isenberg School of Management

Johnathan Loudis

University of Notre Dame - Mendoza College of Business

Date Written: January 29, 2021

Abstract

We develop a methodology to decompose the conditional market risk premium and risk premia on higher moments of excess market returns into components related to contingent claims on down, up, and normal market returns. We call these components the downside, upside, and central risk premia. The decompositions do not depend on assumptions about investor preferences nor do they depend on assumptions about the market return distribution. They can be computed in real time using a cross-section of option prices. The components' contributions to total risk premia vary over time and across investment horizon, as do the total risk premia themselves. Our risk premium decompositions offer powerful tools for evaluating representative agent models in a conditional setting. We develop a related methodology to estimate analogous conditional decompositions implied by prominent representative agent models and compare these to the data-implied decompositions. Although many representative agent models are able to match the unconditional market risk premium thus “explaining” the risk premium puzzle, they generally do a poor job matching conditional risk premia and their components. Our results provide a host of new empirical facts regarding sources of conditional risk premia and identify a set of new challenges for representative agent models.

Keywords: Market risk premium; Variance risk premium; Crash risk; Conditioning information; Risk-neutral moments; Preferences; Stochastic Discount Factor

JEL Classification: E44; G1; G12; G13

Suggested Citation

Chabi-Yo, Fousseni and Loudis, Johnathan, A Decomposition of Conditional Risk Premia and Implications for Representative Agent Models (January 29, 2021). Available at SSRN: https://ssrn.com/abstract=3734689 or http://dx.doi.org/10.2139/ssrn.3734689

Fousseni Chabi-Yo

University of Massachusetts Amherst - Isenberg School of Management ( email )

Amherst, MA 01003-4910
United States

Johnathan Loudis (Contact Author)

University of Notre Dame - Mendoza College of Business ( email )

Notre Dame, IN 46556-5646
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
128
Abstract Views
523
rank
265,492
PlumX Metrics