Detecting contagion between US and Latin-Americans stock markets

Presenting author
Jose Marcos Vera Leyton
Universidad Aut onoma de Colombia
Bogota Colombia
marcosveraleyton@gmail.com

This paper investigates the contagion e ect between United States and some Latin-American stock markets and between them. Using daily data over the sample period 2001-2010, and dividing the series by utilizing the iterated cumulative sums of squares (ICSS) algorithm of Inclan and Tiao (1994) we could detect sudden changes in variance of returns and identify the stable and non-stable periods. By analyzing the mean of returns, we could also identify some turmoil in the non-stable periods, and then we could detect if there where contagion e ect between the stock markets by using Dynamic Conditional Correlation, Engle (2002). We could detect evidence of contagion e ect in the stock markets of Chile, Brazil, Mexico, Colombia and EEUU, in the 2001, 2006 and 2008 crisis.

Key words:

Contagion e ect, ICSS algorithm, DCC GARCH methodology, Turmoil periods.
References Inclan, C.,Tiao., 1994. Use of cumulative sums of squares for retrospective detection of changes of variance. Journal of the American Statistical Association 89, 913-923.
Engle, Robert F., 2002. Dynamic conditional correlation: A simple class of multivariate generalized autoregresive conditional heteroskedasticity models, Journal of business and Economic Statics 20, 339350.
Forbes, Kristin and Roberto, Rigobon, 2002. No contagion , only interdependence: Measuring Stock market comovements, Journal of nance 57, 2223-2261

Fecha: 7 de Septiembre
Hora: 7pm
salon: F307