Análise empírica do value-at-risk por simulação histórica com atualização de volatilidade para fundos de ações no Brasil

• Autor(a) principal: Dias, Leonardo Ethur
• Data de Publicação: 2010
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional do FGV (FGV Repositório Digital)
• Texto Completo: http://hdl.handle.net/10438/6874
• topic: Economia; Finanças; Risco (Economia); Mercados financeiros futuros; Mercado financeiro; Especulação; Ações (Finanças)

For more then a decade, Value-at-Risk (VaR) has bee n choosen by financial and non-financial institutions to measure and control market risk for their portfolios. Because parametric methods usually assume that daily returns of market variables are normally distributed, some risk managers choose to use historical simulation to calculate their portfolios VaR. The main criticism to the traditional historical simulation is that it gives the same weight in the distribution for all the returns of the period used. This work tests the model of historical simulation with volatility updating, proposed by Hull and White (1998), for the Brazilian stock market and compares its performance to the traditional model. The model showed superior performance in forecasting losses for the portfolios and in the speed it adapted itself in times of rupture in volatility levels.