Work in progress

Dynamic clustering of multivariate panel data, joint with Igor Custodio Joao, Andre Lucas, and Julia Schaumburg.
(ECB WP) (most recent version) (most recent Web Appendix)
R&R, Journal of Econometrics.
A reduced-form model for studying time-varying group structures in high- and multi-dimensional data.
Transitions across bank business model groups are related to group dissimilarity and differences in bank profitability.

Modeling extreme events: time-varying extreme tail shape, joint with Andre Lucas and Xin Zhang.
(ECB WP) (TI DP) (Riksbank WP).
Reject & resubmit at Journal of Business and Economic Statistics.
An extreme value theory-inspired statistical model to study time variation in tail parameters for long univariate time series.
Central bank asset purchases lean against the tail risk of extremely adverse market outcomes while active.

Euro area sovereign bond risk premia during the Covid-19 pandemic, joint with Stefano Corradin and Niklas Grimm.
(ECB WP) (A recent version), submitted.
A statistical framework for decomposing sovereign bond yields into their underlying risk premium components.
ECB monetary policy and EU fiscal policy announcements affected bond yields in different ways.

The risk management approach to macro-prudential policy, joint with Sulkhan Chavleishvili, Robert F. Engle, Stephan Fahr, Manfred Kremer, and Simone Manganelli.
(ECB WP) (A recent version) (Web Appendix), submitted.
Macro-prudential policy makers need to trade off an economy's short-term upside potential against its medium-term downside risk.
We use a quantile VAR model to relate this intertemporal trade-off to measures of financial stress and medium-term vulnerabilities.

Nonlinear dynamic factor models with interacting level and volatility, joint with Geert Mesters and Siem Jan Koopman.
(in progress)
We propose a nonlinear dynamic factor model with interacting conditional mean and volatility factors.
The model allows us to relate euro area sovereign bond yields to interacting risk factors.