Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/29566 
Authors: 
Year of Publication: 
1998
Series/Report no.: 
Diskussionspapier No. 21/1998
Publisher: 
Friedrich-Alexander-Universität Erlangen-Nürnburg, Lehrstuhl für Statistik und Ökonometrie, Nürnberg
Abstract: 
In almost all studies concerned with the distribution of financial data skewness and leptokurtosis will be measured by the third and the fourth standardized moments. Additionally, there is the problem of some severe outliers in the data. Therefore, skewness and leptokurtosis will be overestimated because the standardized moments are very sensitive with respect to these outliers as the investigation of the influence function of higher order standardized moments shows. In the literature concerned with adaptive and robust statistical methods there are alternative proposals for measuring skewness and kurtosis. These measures depend on means only defined on a part of the support of the considered distribution or on quantiles. Also for these measures the influence functions will be derived to discuss the influence of isolated outliers
Document Type: 
Working Paper

Files in This Item:
File
Size
127.65 kB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.