ОЦЕНКА НА СТОЙНОСТ ПОД РИСК НА ПОРТФЕЙЛ ОТ СВЕТОВНИ ИНДЕКСИ ПОСРЕДСТВОМ ВЕКТОРНИ АВТОРЕГРЕСИОННИ МОДЕЛИ
ESTIMATION OF THE VALUE AT RISK OF WORLD INDICES PORTFOLIO USING VECTOR AUTOREGRESSION MODELS

Summary/Abstract: The main objective of this paper is to test whether approaches that model cross correlation structure of the multivariate data can lead to more accurate estimates of the market risk of a portfolio in comparison with classical methods like empirical Cumulative Distribution Function (CDF) or Risk Metrics TM IGARCH(1,1) process without drift. The data contains daily closing values for S&P 500, DAX and Nikkei 225 Indexes and the period is 2006 – 2022. Forecasts about portfolio with equal weights Value at Risk one month in the future are sequentially calculated starting from2010 on the basis of all available data up to that moment. The actual quantile (5% and 1%) of the portfolio return is then used as a benchmark for the forecast’s accuracy. Obtained results show that Vector Autoregression Models outperform the other considered methods although they do not directly grasp heavy tails and volatility clustering.

• Details
• Contents