0
$\begingroup$

It is well know that one uses the Black 76 model to price commodity derivatives. I would however like to perform a Monte Carlo simulation that ties back to this number.

How would one go about this process? Is there a way to make use of the known future prices to simulate suitable spot prices that will result in the Monte Carlo tying back to the formula approach?

Improve this question
$\endgroup$
5
  • $\begingroup$ I am a little confused as to your goal. A statistic is any function of data. The pricing models are Frequentist point estimators. So, to me, your question is a bit like asking is there a way to simulate data whose sample mean is 5. The answer is yes, infinitely many. What is the goal you are mentally working to. $\endgroup$
    – Dave Harris
    Commented Dec 31, 2016 at 2:11
  • $\begingroup$ My thinking is as follows: I know that under Black's model the forward prices are log normally distributed. Therefore I can simulate forward prices using this distribution. $\endgroup$
    – User16473932
    Commented Jan 2, 2017 at 6:58
  • $\begingroup$ My thinking is as follows: I know that under Black's model the forward prices are log normally distributed. Therefore I can simulate forward prices using this distribution. I want to re price an option though. So my question is, can I make use of the log-normally simulated forward prices to price the option (and if so, how would I do this?) Or would I need to simulate spot prices by perhaps using gbm, where the drift is no longer the risk free but rather the growth implied by the forward prices I would like a simulation that is calibrated to the forward price as well as option price. $\endgroup$
    – User16473932
    Commented Jan 2, 2017 at 7:50
  • $\begingroup$ If all you want to do is simulate log-normal prices choose a package, such as R, and generate random numbers. The code in R would just be: y<-exp(rnorm(1000,mu,sigma)) $\endgroup$
    – Dave Harris
    Commented Jan 2, 2017 at 22:51
  • $\begingroup$ I am well aware on how to simulate a log normal distribution. I am however not sure how to use this to reprice an option as what I have simulated if future prices. Does this mean under each simulation I again use Blacks model (though this seems like I would double count volatility) or do I need to convert the future prices to spot prices and then get a terminal value for the option and discount? $\endgroup$
    – User16473932
    Commented Jan 3, 2017 at 7:35

1 Answer 1

Reset to default
1
$\begingroup$

No need to convert futures prices to spot prices. Your simulation should look like:

$S_{T}=S_{0}*exp(\mu T-0.5\sigma^{2}T+\sigma \sqrt{T}z)$

where $\mu$ is the drift of spot prices. If you use the spot-forward relationship $F=S_{0}*exp(\mu T)$, you can rewrite the equation in your simulation to be:

$S_{T}=F*exp(-0.5\sigma^{2}T+\sigma \sqrt{T}z)$

Put simply, Black76 is just standard BS rewritten to use forward/futures prices instead of spot prices

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.