Improvement Of VaR Algorithm Based On Historical Simulation And M - C Method

The common understanding of risk include:risk is the possibility of loss; risk of loss probability; risk of actual results with the expected results of the deviation; risk is the degree of variation of future results and so on. But from the economic behavior and activity is essentially risk refers to the uncertainty of future revenue. That is the situation we may become better future is good, it can be our future situation worse.Standard historical simulation principle:the individual risk factors in the past on an event or change in distribution accurately portrayed, as these risk factors change in the future distribution or changing circumstances, on this basis, through the establishment of risk factors and self-check the mapping between the portfolio value expressions to simulate the self-examination combined profit and loss in the future root condition, then calculate a given confidence level of VAR. Obviously, the standard historical simulation method does not assume market risk factor obey some probability distribution, but directly with the distribution of risk factors, past changes indicative of future changes in the distribution, so the standard historical simulation method does not require parameter estimation, which is a non-parametric estimation method full valueln essence, the historical simulation method for VaR is based on the history of the future is simple to reproduce this hypothesis, but the time-weighted historical simulation method for standard historical simulation method unrealistic assumption of equal probability, risk factors proposed for different periods given different weights.Time-weighted historical simulation method from the more recent history of the current data given greater weight, time to reflect such a universal fact:risk factors have occurred from now, the closer again in the future behavior of the likelihood of recurrence larger, or that risk factors from now, the more recent change scenarios for predicting future changes to provide more information on the distribution...