Tuesday 7 October 2 p.m. - 3:40 p.m. (Europe/Amsterdam)
Abstract
Key to risk metrics in corporate finance is the dynamic covariance between assets. It may be necessary to estimate up to 10,000 covariances and to do so quickly. The Gaussian copula approach to covariance estimation is computationally efficient but is it robust? We will explore estimation of covariances using time series approaches, and adapt these approaches to quickly changing market conditions. Our methodology will provide short-term covariance forecasting and long-term covariance scenario estimation. Computational speed will be central to our considerations.
