Estimation of the market risk exposure of Vietnamese banks’ portfolios using VaR approach

Abstract: This paper analyses the effectiveness of different methods to estimate Value-at-Risk (VaR) of VN-index, proxy of a Vietnamese bank’s portfolio. Both parametric and non-parametric approaches are employed to estimate daily VaRs for two sets of data, one of those sets is 8 months behind the other. We find that non-parametric methods are more reliable than parametric methods when employed to estimate VaR for a bank’s portfolio in Vietnamese market. Volatility weighted methods perform better for the first set of data, where there is a sudden jumps in the returns. Meanwhile, Basic historical simulation work best for the data that fluctuates dramatically but have no impulsive jumps over time. Our conclusion is that to correctly estimate the maximum loss of a bank’s portfolio, especially in the unstable market like Vietnam, it is important to choose the suitable window size to estimate VaR.