Heavy Tails in High-Frequency Financial Data

23 Pages Posted: 22 Jan 1997

Date Written: Undated

Abstract

We perform a tail index estimation of financial asset returns in two markets: the foreign exchange market and the interbank market of cash interest rates. Thanks to the high-frequency of the data, we obtain good estimates of the tail indices and we are able to analyze their stability with time aggregation. Our analysis confirms that the variance of the return is finite but points to the non-convergence of the kurtosis. Both financial markets present similar tail behavior of the returns. A study of the extreme risks reveals the need to depart from the Gaussian assumption by taking the fat tails fully into account. A study of tails under temporal aggregation, also investigating data from theoretical price formation processes, shows that ARCH-type processes represent the true behavior better than unconditional distribution models.

JEL Classification: G10

Suggested Citation

Dacorogna, Michel M. and Pictet, Olivier V., Heavy Tails in High-Frequency Financial Data (Undated). Available at SSRN: https://ssrn.com/abstract=939 or http://dx.doi.org/10.2139/ssrn.939

Michel M. Dacorogna

DEAR-Consulting ( email )

Scheuchzerstrasse 160
Zurich, 8057
Switzerland
+41795447327 (Phone)

Olivier V. Pictet

Pictet Asset Management ( email )

Geneva
Switzerland

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
1,157
Abstract Views
3,352
rank
23,939
PlumX Metrics