A utilização das metodologias de VaR : value at risk, simulação histórica e paramétrica usando a distribuição normal, para explicar as variações do índice ibovespa no período de jan/2016 a dez/2017

Abstract

The purpose of the present monographic work is to analyze the applicability of backtesting to evaluate which Value at Risck (VaR) methodology best explains the variations of Ibovespa in the period from January 2016 to December 2017. For this, two VaR calculation methods were used: The first method is Variance-Covariance (parametric): uses the Normal distribution and the exponential smoothing moving average (EWMA) as a mechanism for estimating volatility. The application of this method and estimation mechanism was chosen because it is one of the most used methods in the literature and in practical applications. The second method is the Historical (non-parametric) Simulation: based on the empirical distribution of the data. Being a non-parametric method, there is no need to know the probability distribution of the data. After the analysis, it is concluded that the Historical Simulation method is close to the limit of 5% (five percent), as presented in the Backtesting result, characterizing it as the most suitable for the VaR calculation in the series of Ibovespa. It should be pointed out that the present study is of practical importance for professionals working in the variable income market, since it helps them to make decisions regarding the risk of this segment.