BACKTESTING THE VALUE AT RISK FOR LATIN AMERICAN STOCK AND CURRENCY MARKETS

Contenido principal del artículo

Werner Kristjanpoller Rodríguez
Andrés Barahona Ossa

Resumen

In this article three methodologies are analyzed for calculating the value at risk (VaR): parametric, semi-parametric and non-parametric models. In order to evaluate their validity, a representative model was chosen for each: EGARCH for the parametrics, CAViaR for the semi-parametrics and the historic simulation for the non-parametrics. To validate these methodologies, the model proposed by Candelon et al. (2011) was used, a backtest based on the general method of moments. Variables to be forecast were the exchange rates and main stock-market indexes of the principal Latin American markets (Argentina, Brazil, Chile, Colombia, Peru, and Mexico). Results show that the CAViaR model is the best at forecasting the var for the markets and currencies during the periods that were analyzed.

Detalles del artículo

Cómo citar
Kristjanpoller Rodríguez, W., & Barahona Ossa, A. (2015). BACKTESTING THE VALUE AT RISK FOR LATIN AMERICAN STOCK AND CURRENCY MARKETS. Investigación Económica, 73(287). Recuperado a partir de https://www.revistas.unam.mx/index.php/rie/article/view/50507