A Predictive System with Heteroscedastic Expected Returns and Economic Constraints

Maxime Bonelli, Daniel Mantilla-Garcia: We propose a variation of a predictive system that incorporates two (additional) economically motivated assumptions about the dynamics of expected returns, namely 1) their positivity, and 2) a time-varying volatility correlated with economic regimes.

Author(s):

Maxime Bonelli

Research Engineer, Koris International

Daniel Mantilla-Garcia

Research Associate, EDHEC-Risk InstituteHead of Research & Development, Koris International

The implications of the modified system are consistent with well established empirical facts of stock returns, in particular, the simpler version of the modified system without predictors can explain the well documented countercyclicality of the dividend-price ratio’s predictive power. In empirical tests we find that the system can produce significantly lower forecast errors than the historical mean out-ofsample as well as some improvement compared to the original system.

Type: Working paper
Date: le 07/07/2014
Research Cluster : Finance

See Also

EDHEC-Risk Instute paper on value in sovereign bond markets accepted by the Journal of Fixed Income
News
- 12-09-2019
We are pleased to enclose an EDHEC-Risk Institute research article published in the...
EDHEC Faculty welcomes Oxford professor Renée B. Adams for a reseach seminar
News
- 11-09-2019
On September 12, 2019, EDHEC faculty will be delighted to welcome Oxford professor...
Riccardo Rebonato will unveil the results of the 12th EDHEC-Risk European ETF & Smart Beta Survey on Sept 23 in London
News
- 03-09-2019
Riccardo Rebonato, Professor of Finance, EDHEC Business School, EDHEC-Risk Institute,...
Launch of the
News
- 03-09-2019
EDHEC Business School and Scientific Beta have announced the launch of the “Advanced...