This paper uses a Value at Risk （VaR） approach to evaluate a country financial vulnerability, by analyzing the risk exposure of its Central Bank, as if their assets are subject to market risk. The Brazilian currency...This paper uses a Value at Risk （VaR） approach to evaluate a country financial vulnerability, by analyzing the risk exposure of its Central Bank, as if their assets are subject to market risk. The Brazilian currency exchange swaps contracts （USS/Brazilian Reais） are submitted to a delta-normal VaR method, in order to evaluate the market risk of each swaps series, by modeling the variance of the daily returns, from August 1999 to January 2003. All daily returns series exhibited heteroscedasticity in the conditional variance and sudden changes in the unconditional variance. The points of changes of the unconditional variance were determined through the Iterative Cumulative Sum of Squares （ICSS） algorithm, and the conditional variance was modeled with Markov-Switching-Generalized Autoregressive Conditional Heteroscedasticity （SWGARCH） in order to capture heteroscedasticity and regime change. The results lead to two main conclusions： First, a VaR model must incorporate heteroscedasticity and regime switching in order to describe the variance of the tested series, submitted to brisk changes of economic and political scenarios. Second, a volatility-based VaR do not necessarily generate forward-looking indicators, but rather coincident indicators of possible financial vulnerabilities. The future research will evolve towards evaluating the effects of the Basel III recommendations as if they could be applied to this crisis period.展开更多
文摘This paper uses a Value at Risk （VaR） approach to evaluate a country financial vulnerability, by analyzing the risk exposure of its Central Bank, as if their assets are subject to market risk. The Brazilian currency exchange swaps contracts （USS/Brazilian Reais） are submitted to a delta-normal VaR method, in order to evaluate the market risk of each swaps series, by modeling the variance of the daily returns, from August 1999 to January 2003. All daily returns series exhibited heteroscedasticity in the conditional variance and sudden changes in the unconditional variance. The points of changes of the unconditional variance were determined through the Iterative Cumulative Sum of Squares （ICSS） algorithm, and the conditional variance was modeled with Markov-Switching-Generalized Autoregressive Conditional Heteroscedasticity （SWGARCH） in order to capture heteroscedasticity and regime change. The results lead to two main conclusions： First, a VaR model must incorporate heteroscedasticity and regime switching in order to describe the variance of the tested series, submitted to brisk changes of economic and political scenarios. Second, a volatility-based VaR do not necessarily generate forward-looking indicators, but rather coincident indicators of possible financial vulnerabilities. The future research will evolve towards evaluating the effects of the Basel III recommendations as if they could be applied to this crisis period.
