It is a known fact that default rates seem to exhibit cyclic behavior. Most probability of default models use one-year averages of default rates to calibrate the models. The one-year averages should be observed long enough to take into account the effects of a full business cycle. But what if historic data does not do back far enough to explain the effects of a full business cycle?

I guess one could link macro economic factors to default rates and extrapolate. Then the value should be somewhat cycle neutral. But from what data could there be extrapolated? Also S&P and Moody's default rates go back to the 1980's but are they representative in all case to extrapolate from? And I do wonder, how do financial institutions make 'artificial' data to represent full business cycles? Thanks for any suggestion or help.

  • 1
    $\begingroup$ If the underlying credits have credit ratings you can use the excellent historical data on the expected default frequencies of corporate credits conditional on credit ratings. $\endgroup$ – BKay Dec 15 '15 at 16:23
  • $\begingroup$ Sometimes portfolio's consist of more than just one credit rating or sometimes this rating is not known for grouped data. Could there still be a way to examine the cyclic behavior in a different manner? Still thanks for the help I will definitely look if this can make life easily. $\endgroup$ – user6717 Dec 15 '15 at 21:23

Your Answer

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