Credit scoring models and other questions

[MSharky]

Hi,

I am having a bit of a hard time trying to grasp the idea of credit scoring models and how they relate to Structural models of default. For example, aren't both the KMV model as well as Altman's Z-score ways to calculate default probabilities? Isn't Mooody's KMV Credit Monitor a scoring model in its own right?

Also, what is the difference between Altman's Z-score (Parametric discrimination) and the CreditPortfolioView model? They both seem to involve probit/logit models to transform the regressed variable into default probabilities.

Thanks,

 

[ShaktiRathore]

Hi,
please see these links if of any help:
https://forum.bionicturtle.com/threads/altmans-z-score.5345/#post-14842
https://forum.bionicturtle.com/threads/credit-portfolioview-model.1302/#post-4749
https://forum.bionicturtle.com/threads/credit-portfolio-models-bis-asrf-model.679/#post-2531
Thanks

 

[David Harper CFA FRM]

Not a response to your (expansive) question but I did want to mention that Edward Altman himself was recently interviewed by CFA Institute (3 parts): https://blogs.cfainstitute.org/inve...s-a-bubble-building-in-global-credit-markets/

 

[NNath]

Not sure if this has been discussed in the forum before, but the performance of credit scoring models is also confusing

The performance measurement approach, cumulative accuracy profile (CAP). With the introduction of Crouhy this year (2016), its covered in 2 places, deServigny and Crouhy. The definition is complicated and so far I haven't understood its difference from ROC and AP calculation from CAP.