Modelos garch con innovaciones con colas pesadas: aplicación empírica a la volatilidad de los mercados de acciones y de divisas en países con ingresos altos y latinoamericanos
No Thumbnail Available
Date
2021-09-28
Journal Title
Journal ISSN
Volume Title
Publisher
Pontificia Universidad Católica del Perú
Abstract
En este trabajo se utilizan datos diarios de los mercados de acciones y divisas para
estimar los modelos GARCH y GJR comparando países emergentes con países de
altos ingresos. En ambos modelos, se toma la distribución Normal como base y se
consideran distribuciones con colas pesadas: la distribución t-Student, la distribución
GED, la distribución NIG y la distribución NRIG, tanto en su versiones simétricas
como asimétricas para capturar las características de los retornos. Los principales
resultados son los siguientes: (i) en todos los mercados y países se seleccionan modelos
con distribuciones de colas pesadas: t-Student (S) y error generalizada (GED);
(ii) es importante incluir efecto apalancamiento para el mercado de acciones, pero
esto no es concluyente para el mercado de divisas; (iii) incorporar una distribución
asimétrica en los retornos resulta necesario para todos los mercados de acciones,
aunque esto no es necesario para algunos mercados de divisas.
This paper uses daily data from stock and Forex markets in order to estimate GARCH and GJR models comparing emerging countries with high-income countries. In both models, the Normal distribution is taken as the basis and some heavy-tailed distributions are considered: the Student t-distribution, the GED distribution, the NIG distribution and the NRIG distribution, all in their symmetric and skewed versions to capture the characteristics of the returns. The main results are as follows: (i) in all markets and countries, models with heavy-tailed distributions are selected: Student’s t (S) and generalized error (GED); (ii) it is important to include the leverage effect for stock markets, but this is not conclusive for the Forex markets; (iii) incorporating an asymmetric distribution in returns is necessary for all equity markets, although this is not necessary for some currency markets.
This paper uses daily data from stock and Forex markets in order to estimate GARCH and GJR models comparing emerging countries with high-income countries. In both models, the Normal distribution is taken as the basis and some heavy-tailed distributions are considered: the Student t-distribution, the GED distribution, the NIG distribution and the NRIG distribution, all in their symmetric and skewed versions to capture the characteristics of the returns. The main results are as follows: (i) in all markets and countries, models with heavy-tailed distributions are selected: Student’s t (S) and generalized error (GED); (ii) it is important to include the leverage effect for stock markets, but this is not conclusive for the Forex markets; (iii) incorporating an asymmetric distribution in returns is necessary for all equity markets, although this is not necessary for some currency markets.
Description
Keywords
Acciones (Bolsa)--Modelos econométricos--Países en desarrollo, Acciones (Bolsa)--Modelos econométricos--América Latina, Acciones (Bolsa)--Modelos econométricos--Países desarrollados
Citation
Collections
Endorsement
Review
Supplemented By
Referenced By
Creative Commons license
Except where otherwised noted, this item's license is described as info:eu-repo/semantics/openAccess