HEC Research Papers Series,
HEC Paris

Abstract: For Central Banks, institutional, and individual investors it is crucial to understand the frequency and importance of drops or sudden rises in financial markets. Extreme value theory (evt) is an interesting tool providing answers to questions such as:

-with what frequency do we find variations of returns beyond a given threshold ?

-over a given period, what type of extreme variation can be expected?

- with what type of unconditional distribution of returns are the tails of returns compatible?

-in a cross country setting of emerging and mature financial markets do extreme variations behave in a similar manner?

- can we learn about the evolution of returns of presently developing economies from the early returns of presently mature markets?

- do countries behave similarly in terms of up or down crashes for a given level of development?

In the following paper we start with a review of theoretical elements of evt. In the empirical section of this study we consider five mature markets, nine Asian, six Eastern European, and seven Latin American emerging markets. The tail-behavior of returns is found to be compatible with the existence of up to the third moment but not beyond. The estimation of the tail distribution as a Generalized Pareto Distribution shows that great care has to be taken for emerging markets where little data is available and returns' distribution is subjet to violate the iid assumption. Using a subsample of countries we demonstrate the limitations of evt. We also show that little can be learned from 19th century US data about presently emerging markets' tail behavior.

Keywords: extreme value theory; generalized Pareto distribution; stock market returns

JEL-codes: C13; C22; G15; O16

57 pages, April 1, 1999

Full text files

3b27b9c744ba836ee62c8078d6dd0e05.pdf  

Download statistics

• HEC Paris
• Ordering Working Papers
• Home page for this series

Questions (including download problems) about the papers in this series should be directed to Antoine Haldemann ()
Report other problems with accessing this service to Sune Karlsson ().

This page generated on 2024-02-05 15:47:24.