This study presents an analysis of stock market volatility in eleven countries. The study used the autoregressive conditional heteroscedasticity models, assessing which among these models was more efficient in modelling the market volatility in these countries. Furthermore, using correlation, covariance and Granger causality indices, the returns and volatility of stock markets were compared among the BRICS countries, countries in Latin America, developed countries and Brazil in specific. The results suggest that stock market volatility is best modelled using asymmetric GARCH methods (EGARCH and TGARCH), which displayed leverage effects in the series studied.