Value-at-risk predictive performance: a comparison between the CaViaR and GARCH models for the MILA and ASEAN-5 stock markets

Purpose This paper tests the accuracies of models that predict Value-at-Risk (VaR) for Market Integrated Latin America (MILA) and Association Southeast Asian Nations (ASEAN) emerging stock markets during crisis periods. Design/methodology/approach Many VaR estimation have been presented in literature. In this paper, is estimated using Generalized Autoregressive Conditional Heteroskedasticity, EGARCH GJR-GARCH under normal, skewed-normal, Student-t skewed-Student-t distributional assumptions compared with predictive performance (CaViaR) considering four alternative specifications proposed by Engle Manganelli (2004). Findings The results support robustness CaViaR model out-sample forecasting MILA ASEAN-5 evidence based on backtesting approach analyzed according to their accuracy. Originality/value An important issue market risk inaccurate since different lead measures, which means there not yet an accepted method all situations markets. particular, quantifying crucial evaluating global biggest exchange ASEAN region accounted 11% total foreign direct investment inflows 2014. Furthermore, Development Bank, projected average 7% annual growth 2025.