The approach to this method is to run Siemen’s foreign exchange rate exposure against a set of
historical exchange rates, and in that way yield a distribution of losses. The VaR can then be found
by computing a percentile corresponding to the confidence level you want, e.g. 95% or 99%. The
steps in calculating Siemens historical VaR can be expressed as follows:
1. Obtaining historical exchange rate data (the data must be long term sufficiently75)
2. Computing currency return, weighted in accordance with Siemens exposure in the USD and
RMB respectively
3. Derive VaR at different percentile’s corresponding to different confidence levels.
Thus, the assumption behind the historical simulation method is that history repeats itself in
regards to the exchange rate distributions, and an advantage of the historical simulation method is
that returns do not have to follow the normal distribution76
.
Historical Value at Risk has been calculated for the EUR/USD and EUR/RMB monthly currency
returns, weighting the currencies in accordance with the estimated cash flow exposures, as it was
the case with the variance-co variance method. So also here USD currency return weights
approximately 86% and the RMB 14% of the joint currency return.
The historical VaR has been calculated using data from the three periods; January 1981- April
2013, January 1999 – April 2013 and last July 2009 – April 2013. The monthly returns are shown as
a scatter plot to the left side in figure 13, where the horizontal axis represents observations
ordered by date starting from January 1981 and ending April 2013. A six month’s moving variance