Valor en riesgo anual de los mercados accionarios de México y Estados Unidos: VaR tradicional vs VaR cópulas elípticas

Abstract

Economical and financial links between neighboring countries are undeniable, in this sense, it is of paramount importance to analyze the implied risks to improve the decision making process. In this research risk is measured through the potential loss of Value-at-Risk (VaR) between Mexican and USA stock markets. Different traditional VaR methodologies are used, such as Delta-Normal VaR, Historical Simulation VaR, and Monte Carlo Simulation VaR, these in turn are verified against the Elliptic Copulas VaR methodology, which estimates more accurately inherent extreme events of the stock market prices returns. The analysis considers daily time series data during 40 years, from 1975 to 2014; these are divided into segments to carry out a yearly analysis. The stock markets sample includes the Mexican (IPC) and USA (S&P 500) stock indexes. The empirical evidence identifies the variation over time of potential losses patterns of the stock markets being analyzed. Additionally, it verifies the precision and robustness of VaR measurement through the copula methodology which aims to capture the non-linearity and therefore avoids the underestimation of the losses