3
$\begingroup$

My goal is to develop a model to simulate long term FX movements. (I am not sure if long term makes any difference, but if it does I am more interested in long term fx movements)

These Monte Carlo simulations will not be used for pricing but from a risk-management perspective, to calculate how much the portfolio is exposed to FX risk.

I was wondering if there is a suggested model (or paper or anything) that I could use as a starting base.

Some further notes, after Matt's comment:

I am not interested in some sort of a trading strategy (i think i misused the words: portfolio and exposure). And i don't want to hedge or price anything.

Imagine that you have lend some money (through some instrument) that you will receive in 20-30 years in a foreign currency. Now you believe that forex rates have some correlation with your instrument and some other factors and you want to see how much you stand to lose in a worst case scenario.

My idea was that I would run an MC where i will have some sort of forex evolution correlated with the evolution of my instrument and i set up some sort of a stress test.

Is my line of thinking correct, or you believe that i should be trying something else?

If we agree that my approach is decent, my question is how would you evolve a forex rate.

$\endgroup$
14
  • $\begingroup$ You do not need a forecasting model to calculate your portfolio risk exposure. Aside that this site is not a repository for trading strategies, one of the hardest earned asset. If you need to calculate fx exposure for portfolios simple map out each asset in the portfolio and group its notional exposure (or delta for options) by currency to get what you want. For trading strategies I recommend to look elsewhere because I am pretty sure most users are more than hesitant to hand you the fruits of their hard labor free of charge. $\endgroup$
    – Matt Wolf
    Jun 20, 2013 at 10:42
  • $\begingroup$ Hi Matt, thank you for your answer. I've been reading this site for some time and i've learn a ton from your answers. I think i didn't phrase my question correctly. Please bear with me, I am still a student, I will try to reformat my question. $\endgroup$ Jun 20, 2013 at 12:45
  • 1
    $\begingroup$ @following your edit, could you please pose a specific example. What do you mean with "instrument"? Your worst case is a perfect correlation of your "instrument" with the "foreign" currency you receive back in x years and if I understand correctly you look to model the future paths of that currency? Worst case then is a massive devaluation of such currency vs your base currency. You can use an fx model to simulate such paths but I tell you upfront that you will not gain much with that. Long-term fx rates are heavily a function of fundamentals such as monetary policy, nations' debt servicing... $\endgroup$
    – Matt Wolf
    Jun 20, 2013 at 13:40
  • $\begingroup$ I see that I abused terminology one more time. by worst case i meant: an average (expected) loss for some parameters. In terms of instrument, what i have in mind is pretty simple. Along the lines of having euros or dollars and lending really long term money in some "weird" currency. This practically means that you can't really hedge the forex risk. I didn't want to convolute the example that's why i avoided being super specific. $\endgroup$ Jun 20, 2013 at 15:58
  • 1
    $\begingroup$ Ok, I tell you how it is, then you can still dig out the annual Bloomberg article on the top fx forecasters. You put 200 apes into a room. You ask each of them where they see USDJPY a year from now. A year later you find out that 10/200 apes were relatively close in their estimate to the actual rate. Same game repeated...a year later 1/10 again hits the jackpot. That guy is pronounced the king of fx analysts. Chance? Talent? I chuckle each time I come across the same spiel. If I saw the same guy in the top 3 over 5-10 years I would resort it to talen but not 1-2 data observations. $\endgroup$
    – Matt Wolf
    Jun 20, 2013 at 16:16

0

Your Answer

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

Browse other questions tagged or ask your own question.