Academic Background and Research

Academic Background

My academic background is in mathematics, statistics, and economics. I obtained degrees from the Vienna University of Technology (Dipl.-Ing. in 2009, with distinction, thesis supervised by Manfred Deistler) and “ingénieur généraliste” at Ecole Centrale Paris (Promotion 2010, thesis supervised by Frédéric Abergel).

After an internship in the Global Arbitrage Trading department at Credit Suisse in London, I started my PhD in mathematics at the Vienna University of Technology and changed after one year to the economics department at the University of Vienna, where I passed my PhD under supervision of Manfred Deistler and Benedikt M. Pötscher (May 2015, with distinction). During my PhD, I could further my international network through extended research visits at the Australian National University in Canberra (November 2012 and November 2013, invited by Brian D. O. Anderson) and University of Pennsylvania (January until May 2014, invited by Frank Schorfheide).

Research Interests

I focus on data-driven modeling to characterize and identify structural economic shocks without relying on (possibly implausible) restrictions derived from economic theory, as is the case in Dynamic Stochastic General Equilibrium (DSGE) models.

Early projects (without innovation) were concerned with cointegration analysis (project at the end of BSc. equivalent at TU Wien), factor GARCH models with applications to financial time series (project at Ecole Centrale Paris), and dynamic principal component analysis, i.e. the frequency domain equaivalent of (static) principal component analysis, (project at the end of MSc. equivalent at TU Wien).

My PhD thesis comprises contributions on singular vector autoregressive (VAR) models, identifiability from mixed frequency data, as well as solution sets and identifiability of rational expectations (RE) models. Solutions of singular VAR models are stationary multivariate processes with innovation covariance matrix of reduced rank. They are the essential building block of generalized dynamic factor models. Given mixed frequency data, it is interesting to ask the question whether it is possible to identify an underlying model which generates data at the highest frequency. In linear RE models, I worked on characterizing the full solution set and on identifiability by constructing an analytical mapping from internal and external characteristics.

During my postdoc at the University of Helsinki, I have worked on possibly non-invertible VARMA models and generalized dynamic factor models. Possibly non-invertible VARMA models take informational asymmetry between economic agents and outside observers (e.g. econometricians) into account. They are a data-driven alternative to DSGE models which are based on economic theories and often rely on strong assumptions.

The outcomes are several publications in top-tier journals, including the Journal of Econometrics, Econometric Theory, and the Journal of Time Series Analysis, see Publications.

If you are working with one of my R packages or build on my work in any other way, I would be happy to hear from you. Especially, I am happy to provide feedback or help out with feature requests regarding non-invertible VARMA models.