Extreme observations in developed and emerging equity markets

It is widely accepted that the distribution of financial returns has heavy tails. In this context it is important to understand the frequency and importance of extreme events in financial markets. Extreme Value Theory is the appropriate framework for studying the tail behaviour of a distribution. Tail index estimators, such Hill index can be used to measure the shape of the tail, but it depends for its accuracy on a correct choice of the threshold where the tail begins, and the tail shape estimates are inefficient unless the threshold is accurately determined. We will apply bootstrap methodology to select the correct sample fraction for tail estimation. Also Generalized Pareto Distribution (GPD) model is a convenient framework for the approximation of the threshold exceedances distribution and correct inference depends for its accuracy on a correct choice of the threshold. In this paper we will use corrected threshold selection to explore similarities and differences between tail distributions in equity markets using stock index time series with different characteristics, such as size and maturity. We will try to find out if the probability of extreme events is similar in big or small markets and in mature or emerging markets.