wilmott.com

Hi, can someone check if the following equation is correct to find the VaR (99% confidence) of the electricity futures contracts:F = Futures priceSigma = Implied VolA = Contract SizeVaR = [ F * exp(-0.5 * Sigma^2 * 1/252 + Sigma * (1/252)^(.5) * 2.33) - F) * (A)I was given the above equation, but have no clude why it was set up like that. Thanks!

Hi, does anyone have a clude where the equation is from? Thank you!!! Any hints are welcome!!!I thought I can just calculate VaR as follow:VaR = 2.33 * Sigma * F * (252)^(-.5) * A

Depends on if the sigma is calculated using log prices or not. If so you should invert back using exponent. Otherwise your second setup is correct./Rutger

Just out of curiosity, is GBM any good as a model for electricity futures? Is there any spikey behaviour as observed in the within day electricity prices?Thanks,Sam

QuoteOriginally posted by: AthleteScholarVaR = [ F * exp(-0.5 * Sigma^2 * 1/252 + Sigma * (1/252)^(.5) * 2.33) - F) * (A)VaR = 2.33 * Sigma * F * (252)^(-.5) * A For small Sigma, the Sigma^2 term in the first equation can be neglected and exp(Sigma*2.33*252^-0.5) is approximately 1 + Sigma*2.33*252^-0.5. Using that substitution, the two formulae are identical.The second formula is the VaR if you assume P&L has a normal distribution. This is the usual practice. The first formula is the VaR assuming P&L follows a lognormal distribution and Sigma is the standard deviation of the underlying normal, not the lognormal. This is not a common practice. It only matters for large Sigma, it takes an 82% annual standard deviation to make a 5% difference in the VaR calculations. There are much more important errors inthe VaR calculation than this one, and I see no reason that it would give a more accurate VaR.

There is a very different behaviour of futures prices before and during the contract month. Prior to the month, you don't really need to worry about spikes - so GBM is fairly reasonable. During the month, you would probably want to take spikes into account.