Berlin 2008 – scientific programme

Parts | Days | Selection | Search | Downloads | Help

AKSOE: Arbeitskreis Physik sozio-ökonomischer Systeme

AKSOE 13: Poster Session (posters on display 10:00-19:00)

AKSOE 13.12: Poster

Wednesday, February 27, 2008, 17:30–19:00, Poster G

Realized Volatility and Realized Covariance in Heavy-Tailed Financial Data — •Oliver Grothe and Christoph Müller — University of Cologne, Research Training Group Risk Management

Realized volatility and realized covariance have recently been used intensively for measuring and forecasting volatility and dependency of intraday financial data. For these estimators, nice convergence properties may be derived under standard assumptions. However, they face two important problems when actually working with high frequency financial data: market microstructure effects and heavy tails in return data. The former introduces a bias to the estimators, the latter may lead to infinite variances of the estimators. While recent research suggested several solutions to overcome the bias, the influence of heavy tails on the estimators remains mainly unexplored.

We analyze this influence and show that the standard estimators tend to get useless if the tail indices of return distributions approach values as commonly observed in financial data. However, we proof that other estimators such as the bipower variation remain accurate.