Which of the following statements are true in relation to Historical Simulation VaR?

CORRECT TEXT

Which of the following statements are true in relation to Historical Simulation VaR?

I. Historical Simulation VaR assumes returns are normally distributed but have fat tails

II. It uses full revaluation, as opposed to delta or delta-gamma approximations

III. Acorrelation matrix is constructed using historical scenarios

IV. It particularly suits new products that may not have a long time series of historical data available
A . II
B . II and III
C . I and IV
D . All of the above

View Answer

Answer: A

Explanation:

Historical Simulation VaR is conceptually very straightforward: actual prices as seen during the observation period (1 year, 2 years, or other) become the ‘scenarios’ forming the basis of the valuation of the portfolio. For each scenario, full revaluationis performed, and a P&L data set becomes available from which the desired loss quantile can be extracted.

Historical simulation is based upon actually seen prices over a selected historical period, therefore no distributional assumptions are required. Thedata is what the data is, and is the distribution. Statement I is therefore not correct.

It uses full revaluation for each historical scenario, therefore statement II is correct. Since the prices are taken from actual historical observations, a correlationmatrix is not required at all. Statement III is therefore incorrect (it would be true for Monte Carlo and parametric Var).

Historical simulation VaR suffers from the limitation that if enough representative data points are no available during the historical observation period from which the scenarios are drawn, the results would be inaccurate. This is likely to be the case for new products. Therefore Statement IV is incorrect.