How would you analyse the correlation between LIBOR and MXWD?

My initial intention was to get the log return of MXWD and take the opposite of the log return of the LIBOR yield (because when yield goes up, the price of the LIBOR instrument goes down).

I was considering looking at LIBOR 1m, 3m, 6m, 12m. All values are annualised.

How would you do it and why?

  • $\begingroup$ I've recently come across an interesting suggestion by López de Prado who said to partially difference your time series in order to find an optimal trade off between stationarity and long term memory. Then give it a try. Use ADF test to find the best fractional exponent: start with $d=0.1$ and increase until ADF p-value is small enough to reject the trend hypothesis, then give a look at your scatterplot for the model to fit. $\endgroup$ – Lisa Ann Apr 30 '19 at 18:29
  • 1
    $\begingroup$ I would take the :"price of the Libor instrument" to be $\frac{1}{1+T r}$ where $r$ is Libor and T is the maturity (for ex: $T$=0.25 for 3 month libor). The compute the log return on this price as on the price of any security. $\endgroup$ – Alex C Apr 30 '19 at 20:56
  • $\begingroup$ Thanks Alex C, quick question why 1/1+Tr and not 100/1+Tr? I am reasoning in terms of 100 = par but I may be wrong. $\endgroup$ – tweedi May 1 '19 at 8:16
  • 1
    $\begingroup$ 100 works fine. Whichever you prefer. $\endgroup$ – Alex C May 1 '19 at 13:41

Your Answer

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

Browse other questions tagged or ask your own question.