Curso Avançado: EXTREMES AND DEPENDENT EXTREMES IN FINANCIAL RISK MODELLING

EXTREMES AND DEPENDENT EXTREMES IN FINANCIAL RISK MODELLING

CURSO AVANÇADO Extremes and Dependent Extremes in Financial Risk Modelling Curso de curta duração Centro de Estatística e Aplicações da Universidade Lisboa Data: 7 a 10 Outubro de 2002 Prof. Responsável: Alexander McNeil Professor of Mathematics, ETH Zurich Mathematik – ETH Zentrum; Web: www.math.ethz.ch/~mcneil CH-8092 Zurich Programa do Curso Abstract: We concentrate on the Value-at-Risk problem in financial risk modelling, where we are required to estimate a high quantile of the loss distribution for a given portfolio of investments over a fixed future time period. This problem presents an extremely interesting area of application for statistics with many new challenges. Whereas much of traditional statistics concerns the average, the normal and the expected, in the VaR problem we are more concerned with the extreme, the abnormal and the unexpected. It is no surprise to learn that extreme value theory (EVT) has an important role to play in this area, but to apply EVT properly we have to take the strong serial dependencies in financial time series data into account and learn something about the econometric models that are applied to financial data, such as the class of GARCH models, and their extremal behaviour. Financial risk modelling is also an inherently multidimensional problem since the basic object of interest is a portfolio of many underlying instruments whose values change continually. We must also consider multivariate models for extremes and address the issue of possible joint extreme movements in many underlying investments. In this area the ideas of copulas and tail dependence become important. In this short course we will give an introduction to these ideas. Alongside theory we will present several applied examples involving the analysis of real financial data. Provisional Contents: 1. The VaR Problem: Introduction 2. Standard Statistical Techniques of Extreme Value Theory 3. Financial Time Series Models 4. Applying EVT to Financial Time Series 5. Multivariate Models, Copulas and Tail Dependence 6. Implementing these Ideas in Practice HORÁRIO DO CURSO 2ª feira (07/10/2002) a 5ª feira (10/10/2002) Horário do curso : 9 horas no total, em 4 dias. 2ªfeira: 14:30-18:30, com coffee-break das 16:00 às 16:30; 3ª, 4ªfeira: 14:00-16:00 ; 5ªfeira:14:00-15:30 . INSCRIÇÃO O número de inscrições está limitado a 40 participantes. A inscrição deverá ser realizada até 20 de Setembro, e a participação será confirmada até 27 de Setembro. O pagamento da inscrição – 100 Euros – deve ser efectuado até dia 4 de Outubro de 2002, por cheque à ordem da Fundação da Faculdade de Ciências da Universidade de Lisboa e enviado para o Centro de Estatística e Aplicações. CONTACTOS: Centro de Estatística e Aplicações Faculdade de Ciências (DEIO) Bloco C/2 – Piso 2 Campo Grande 1749-016 Lisboa Telf: 21 – 7500120 E-mail: ceaul@fc.ul.pt BOLETIM DE INSCRIÇÃO: Nome: Profissão: Local de Trabalho: Direcção para correspondência: Telefone de Contacto: Fax: E-mail: