Research

Publications

We develop a novel empirical asset pricing framework to estimate time-varying risk premia, building on the recently introduced score-driven conditional betas model. First, we extend the conditional betas theory by establishing the asymptotic distribution of standard tests statistics for parameter constancy in the conditional regression. These tests allow to assess the significance of a given factor in the regression. In addition, we introduce a residual bootstrap procedure for the Wald statistic and establish its validity. Second, we propose a two-step estimation procedure to recover time-varying factor risk premia from individual stock returns. We illustrate the performance of our tests and risk premia estimation procedure through simulations. Third, we present an application in which we assess the existence of a time-varying risk premium associated with a carbon risk factor in the cross-section of U.S. industry portfolios.


We consider an extension of ARCH(∞) models to account for conditional asymmetry in the presence of high persistence. After stating existence and stationarity conditions, this paper develops the statistical inference of such models and proves the consistency and asymptotic distribution of a Quasi Maximum Likelihood estimator. Some particular specifications are studied and we introduce a Portmanteau goodness-of-fit test. Additionally, test procedures for asymmetry and GARCH validity are derived. Finally, we present an application on a set of equity indices to reexamine the preeminence of GARCH(1,1) specifications. We find strong evidence that the short memory feature of such models is not suitable for peripheral assets.

Supplementary appendix

Working papers

Factor models are highly common in the financial literature. Recent advances allow to relax the constancy of slope coefficients (the so-called betas) by considering conditional regressions. The theory on the estimation of these dynamic conditional betas however usually relies on short memory volatility models, which can be restrictive in empirical applications. Moreover, exogenous variables have proven useful in recent studies on volatility modeling. In this paper, we introduce a multivariate framework allowing for time-varying betas in which covolatilities can exhibit higher persistence than the standard exponential decay. Covariates are included in the dynamics of both conditional variances and betas. We establish stationarity conditions for the proposed model and prove the consistency and asymptotic normality of the QML estimator. Monte Carlo experiments are conducted to assess the performance of the estimation procedure in finite sample. Finally, we discuss the choice of potential relevant exogenous variables and illustrate the pertinence of the model on real data applications.


Tracking macroeconomic data at a high frequency is difficult as most time series are only available at a low frequency. Recently, the development of macroeconomic nowcasters to infer the current position of the economic cycle has attracted the attention of both academics and practitioners. The specifications usually rely on a Markov-switching dynamic factor model with mixed-frequency data whose states allow for the identification of recession and expansion periods. However, such models are notoriously not robust to the occurrence of extreme shocks such as Covid-19. In this paper, we show how the addition of time-varying volatilities in the dynamics of the model alleviates the effect of extreme observations and renders the dating of recessions more robust. Both stochastic and conditional volatility models are considered and we adapt recent Bayesian estimation techniques to infer the competing models parameters. We illustrate the good behavior of our estimation procedure as well as the robustness of our proposed model to various misspecifications through simulations. Additionally, in a real data exercise, it is shown how, both insample and in an out-of-sample exercise, the inclusion of a dynamic volatility component is beneficial for the identification of phases of the US economy.

PhD Thesis

Infinite ARCH processes, dynamic betas, and financial applications

PhD advisors: Christian Francq & Jean-Michel Zakoïan

PhD jury: Giraitis, L. (Queen Mary University of London), Luati, A. (Imperial College London & Università di Bologna), Roueff, F. (Télécom & Institut Polytechnique de Paris), & Scaillet, O. (Université de Genève)

My PhD manuscript is available here.

Presentations in scientific conferences and seminars

* presentation by a coauthor