UCL logo

UCL Discovery

UCL home » Library Services » Electronic resources » UCL Discovery

Heavy tailed time series: testing and numerical issues for dependent observations

Charemza, W; Francq, C; Makarova, S; Zakoian, J-M; (2011) Heavy tailed time series: testing and numerical issues for dependent observations. In: (Proceedings) 5th CSDA International Conference on Computational and Financial Econometrics (CFE'11).

Full text not available from this repository.

Abstract

The paper tackles some practical problems related to computations and applications of the quasi marginal maximum likelihood estimator of alpha stable stationary process. Different estimators of the long run covariance matrix are compared for their finite-sample accuracy. In particular, a size distortions of the statistics based on these matrices (e.g. Student t-ratios) for various alpha-stable distributions under the changing strength of the dependences is investigated. The theoretical and simulation results are accompanied by the empirical study of the world stock market indices. For the stock market returns the market inefficiency has been evaluated tested through the maximum likelihood estimation of the characteristic parameter (alpha) and long-run covariance matrices.

Type: Proceedings paper
Title: Heavy tailed time series: testing and numerical issues for dependent observations
Event: 5th CSDA International Conference on Computational and Financial Econometrics (CFE'11)
Location: London, UK
Dates: 17 December 2011 - 19 December 2011
UCL classification: UCL > Provost and Vice Provost Offices > UCL SLASH
UCL > Provost and Vice Provost Offices > UCL SLASH > SSEES
URI: http://discovery.ucl.ac.uk/id/eprint/1356655
Downloads since deposit
0Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item