In order to simulate an interest rate yield curve, can I just estimate a covariance matrix of historical key rate data, simulate with a normal copula, spline my simulated key rates, then price my assets as a function of the simulated yield curves?

Alternatively, could I just spline historical key rate values and estimate risk as a function of those values?

Why is this insufficient? / Where does the need for stochastic models arise?



One reason is that your system needs to correctly price today's yield curve and interest rate options market , otherwise it is not arbitrage free versus existing instruments. Historical data will not in general achieve that.

  • $\begingroup$ "need" is very strong. In some cases, yes your model should price the market correctly, but it's not a necessary condition for all possible uses. The approach described is very useful in risk management for example. $\endgroup$
    – crunch
    Aug 17 '17 at 14:08
  • $\begingroup$ ok, since the OP referred to pricing of assets I assumed that consistency with the market would be important. $\endgroup$
    – dm63
    Aug 18 '17 at 2:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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