- Score-Driven Asset Pricing: Predicting Time-Varying Risk Premia Based on Cross-Secional Model Performance
This paper proposes a new parametric approach for estimating linear factor pricing models with dynamic risk premia. Time-varying risk prices and exposures follow an observation-driven updating scheme that reduces the one-step-ahead prediction error from a cross-sectional factor model at the current observation. This agnostic approach is particularly useful in situations where predictors are unknown or of uncertain quality. Updating schemes for elliptically distributed returns are derived and propose cross-sectional regression errors as driving sequence for the parameter dynamics. Estimation and inference are performed by likelihood maximization. A simulation study confirms that the novel method is capable of filtering and predicting substantial risk price movements. The empirical performance of the method is illustrated by an application to a panel of size-sorted equity portfolios.
Presentations: DGF 2019 Doctoral Tutorial (Essen), VfS 2020 (Online), IRMC 2020 (Online), ES European Winter Meeting 2020 (Online), SWFA 2021 (Online), AFFI 2021 (Online), ES North American Summer Meeting 2021 (Online), SoFiE (Pre-)Conference 2021 (Online), FMA European Conference 2021 (Online), CEF 2021 (Online), IAAE 2021 (Online), ES Asian Meeting 2021 (Online), 7th International Young Finance Scholar's Conference (Online), Statistical Week 2021 (Kiel), DGF 2021 (Innsbruck), FEBS 2022 (Portsmouth), Frontiers of Factor Investing Conference 2022 (Lancaster) .
Best Doctoral Paper Award DGF 2019
- Currency Returns and FX Dealer Balance Sheets (with Stefan Reitz)
- Dynamic Mixture Vector Autoregressions with Score-Driven Weights (with Alexander Georges Gretener and Matthias Neuenkirch)
We propose a novel dynamic mixture vector autoregressive (VAR) model in which time-varying mixture weights are driven by the predictive likelihood score. Intuitively, the state weight of the k-th component VAR model in the subsequent period is increased if the current observation is more likely to be drawn from this particular state. The model is not limited to a specific distributional assumption and allows for straightforward likelihood-based estimation and inference. We conduct a Monte Carlo study and find that the score-driven mixture VAR model is able to adequately filter the mixture dynamics from a variety of different data generating processes which most other observation-driven dynamic mixture VAR models cannot appropriately cope with. Finally, we illustrate our approach by an application where we model the conditional joint distribution of economic and financial conditions and derive generalized impulse responses.
Presentations: CFE 2021 (Online), SNDE Symposium 2022 (Online), CEF 2022 (Dallas), IAAE 2022 (London), VfS 2022 (Basel), Statistical Week 2022 (Muenster).
- (Almost) Recursive Identification of Monetary Policy Shocks with Economic Parameter Restrictions (with Jan Pablo Burgard and Matthias Neuenkirch)
Recursively identified vector autoregressive (VAR) models often lead to a counterintuitive response of prices (and output) shortly after a monetary policy shock. To overcome this problem, we propose to estimate the VAR parameters under the restriction that economic theory is not violated, while the shocks are still recursively identified. We solve this optimization problem under non-linear constraints using an augmented Lagrange solution approach, which adjusts the VAR coefficients to meet the theoretical requirements. In a generalization, we allow for a (minimal) rotation of the Cholesky matrix in addition to the parameter restrictions. Based on a Monte Carlo study and an empirical application, we show that particularly the “almost recursively identified approach with parameter restrictions” leads to a solution that avoids an estimation bias, generates theory-consistent impulse responses, and is as close as possible to the recursive scheme.
Presentations: VfS 2022 (Basel).