Gianluca De Nard
University of Zurich
Department of Economics
NYU Stern School of Business
44 West 4th Street
For the academic year 2019/2020 I have been working at the NYU Stern Volatility and Risk Institute as visiting PhD scholar supervised by Nobel prize winner Prof. Robert F. Engle. For my stay in New York the Swiss National Science Foundation (SNF) has awarded a fellowship. Due to my work and projects of joint interest, I became a member and research fellow of the NYU Stern Volatility and Risk Lab.
Volatility and Risk Institute:
In V-Lab we report in real-time various risk measures for financial markets. The analysis is based on volatilities, correlations, systemic risk, tail risk, geopolitical risk and climate risk.
There are two promising ongoing research projects together with Robert F. Engle and Bryan Kelly. The projects are related to climate finance and financial econometrics. The main idea is to derive a climate change index based on textual analysis of newspapers and compute a mimicking portfolio to hedge climate (change) risk.
Finished SNF Funded Projects:
We present the research plan for the Multivariate Volatility Modeling project at the NYU Volatility Institute with direct collaboration and supervision of Prof. Robert F. Engle. We develop a new generalized linear shrinkage target, which can be used for large-dimensional covariance matrix estimation. The proposed covariance matrix estimator can be applied for various fields of research (e.g. Statistics, Economics, Biology, Engineering, etc.). However, we focus on the accurate large-dimensional multivariate volatility estimation of multi-asset class returns and its application to portfolio selection, which has been neglected in the recent literature. Additionally, we plan to extend and combine the generalized shrinkage estimator with factor and dynamic conditional correlation models.
We expect that the proposed generalized constant-variance-covariance shrinkage estimator will outperform even latest shrinkage techniques, as it is based only on a statistical framework and therefore has less restrictive assumptions. Especially it should deliver more efficient portfolio selection and detection of anomalies in the cross-section of asset returns and thus will contribute significantly towards the covariance matrix estimation literature. We plan to run an extensive empirical simulation to evaluate the out-of-sample performance of the proposed estimator.
In summary, the goal of the project is to contribute to the development of the next generation of multivariate volatility models that combine dynamic conditional correlation and factor models with a new generalized and well-conditioned shrinkage estimator.