Abstract
The purpose of this article is to evaluate the usefulness of performance trends for forecasting the future performance of the IGBC (Colombian exchange market index). To this end, 18 different specifications of the GARCH-M model and high frequency data were used. The models in review considered the leverage, day-of-the-week, hour-of-the-day, and day-hour effects. 115 different forecasts for the next 10 minutes were assessed for each of the 18 models, using descriptive statistics and the Granger's and Newbold (1977) and Diebold's and Mariano (1995) tests. The best model was found to be the one that does not consider the day-hour effect on the mean or the variance.
Highlights
O objetivo do artigo é avaliar a utilidade de padrões de comportamento para prever o comportamento futuro do Índice Geral da Bolsa da Colômbia (IGBC)
Se empregaram 18 diferentes especificações do modelo GARCH en media (GARCH-M) e dados de alta frequência
Número óptimo de rezagos para el proceso ARMA en la ecuación de la media
Summary
Ph.D en Economía, Iowa State University, Estados Unidos. Profesor tiempo completo y Director CIENFI (Centro de Investigaciones en Economía y Finanzas), Universidad Icesi, Colombia. Dirigir correspondencia a: Universidad Icesi, Calle 18 No 122-135, Cali, Colombia
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
