Value at Risk prediction: the failure of RiskMetrics in preventing financial crisis. Evidence from Romanian capital market

摘要

Financial markets are not perfect and the risk cannot be totally eliminated, that's why risk reduction became more and moreimportant for the financial markets, since the 2008 financial crisis.The most commonly used tool for risk measure is Value at Risk, being considered a crucial milestone, because it shows themaximum loss in the value of a portfolio asset.The first comprehensive market risk management methodology was developed by JP Morgan in 1994, and was calledRiskMetrics, which become extremely popular due to its easy implementation.This paper analyzes the capacity of RiskMetrics in forecasting the high volatility during the financial crisis for the financialRomanian market and to see if there is some differences regarding the value of decay factor estimated based on squared errorloss, the RiskMetrics approach, and the values obtain from implementing the check error loss function in estimating the decayfactors. We found that in the case of BET and BET-FI, the RiskMetrics estimations underestimate the decay factor, by attaching alower weight to the most recent variance. Moreover, we proved that RiskMetrics model was good enough to forecast thevolatility on Romanian financial market during the financial crisis period, only if we estimate the decay factor based on checkerror loss function.