New Results on Correlation Matrices

Seminar abstract

In this presentation the authors present new results for a new parametrization of correlation matrices, that can be viewed as a generalization of the Fisher transformation of a correlation coefficient.

The parametrization has desirable properties and can be used to generate random correlation matrices, as we illustrate.

They also present results for block correlation matrices, for which a canonical representation greatly simplify estimation in static models. They also use the canonical representation to introduce a new type of multivariate GARCH models. Several results will be visualized using the Julia Programming language.

How to join this seminar

This seminar is free to attend. The seminar will take place online in a password protected Zoom room.

Please contact the organiser for the password.

Speaker bio

Peter Hansen is the Henry A. Latané Distinguished Professor in Economics at the University of North Carolina, Chapel Hill.

He holds a MSc in Mathematics and Economics from University of Copenhagen and a PhD. in Economics from University of California, San Diego.

He previously held academic positions the European University Institute in Florence, Italy, Stanford University, and Brown University.

Professor Hansen is a leading researcher on forecasting and volatility modeling, and he was included on the Thomson Reuters/Clarivate list of the World’s Most Influential Scientific Minds in 2014, 2015, 2016, 2017.

His research is in econometrics, including forecasting, volatility measurement and modeling, cointegration, structural changes, and multiple testing, with some of his main contributions being associated with the Test for Superior Predictability, the Model Confidence Set, the Realized Kernel Estimator, and the Realized GARCH framework, which won the Richard Stone Prize in Applied Econometrics in 2014.