Engle programme

Econometric Institute and Princeton University Press organize the intensive PhD-course:

"Multivariate Volalility Modelling with Dynamic Correlations"

 Prof. R.F. Engle (Stern School of Business, New York University)

May 21-23, 2003
Erasmus University Rotterdam

Course outline

This intensive three-day PhD-course provides an in-depth overview of modelling volatility of financial assets, with a strong emphasis on recent developments concerning multivariate models with time-varying (dynamic) correlations. Obtaining accurate measures and forecasts of time-varying correlations is of utmost importance for funds management, risk management, and pricing derivatives. Popular (multivariate) volatility models include time-series models, factor models, and correlations implied from market instruments. The most popular time-series model is the Autoregressive Conditionally Heteroskedasticity (ARCH) model, introduced by Professor Robert Engle in 1982. Since then Generalized ARCH (GARCH) models have become popular in practice. Obviously, multivariate GARCH models are required to model the dependence between stocks, stock markets or the factor returns in a factor model for asset prices. Recently Professor Engle introduced a new multivariate GARCH model, the Dynamic Conditional Correlation model, which solves some of the difficulties surrounding the estimation of multivariate GARCH models whilst maintaining the desired property of time-varying correlations. We are very pleased to announce that Professor Engle will personally provide 3 lectures on his new model, from an introduction into multivariate GARCH modelling to illustrating applications for portfolio construction and dynamic hedging. The complete course consists of 5 lectures of 2 hours, given on 21 May (full day), 22 May (full day) and 23 May (morning). Dr. Dick van Dijk will provide an introductory lecture on volatility modelling (in particular univariate GARCH-type models) in the morning of 21 May. This is then followed by 2 lectures by Professor Engle on 21 May (afternoon) and 22 May (morning). Dr. Martin Martens will then discuss some alternative (multivariate) time-series models and factor models on 22 May (afternoon). This will provide a solid foundation for Professor Engle’s final lecture on 23 May (morning) in which he will apply his Dynamic Conditional Correlation model to factor returns. The resulting predicted correlations for the factor returns can then be used to predict the correlation matrix for the individual assets. A more extensive outline of the lectures on the Dynamic Conditional Correlation model is given below.

May 21


Room A-1


Room C-2





Lecture by Dick van Dijk

Introduction by prof. dr. E.G.J. Vosselmans, Dean of the Economic Faculty

Lecture by Robert Engle

May 22

Room G2-41

Room G2-21



Lecture by Martin Martens

Lecture by Robert Engle  


There will be breaks of 15 minutes the lectures.   

May 23

Room M3-15


Lecture by Robert Engle  

The lecture on May 23 is part of a symposium. You are all welcome at the organized lunch and also to attend the lectures in the afternoon by Kay Giesecke on default correlations and Antoon Pelsser on interest rate correlations and Antoon Pelsser on interest rate correlations.

Except for coffee on Wednesday morning and Friday (coffee, lunch and drinks) you have to take care of drinks and lunches yourself.


Prof. Robert F. Engle (Stern School of Business, New York University)
Dr. Martin Martens (Econometric Institute, Erasmus University Rotterdam)
Dr. Dick van Dijk (Econometric Institute, Erasmus University Rotterdam).