SAS Macro Code for Delinquency Roll Rate Analysis

Reply
Contributor
Posts: 59

SAS Macro Code for Delinquency Roll Rate Analysis

SAS Macro Code for Delinquency Roll Rate Analysis

 

I am looking for a SAS Macro or a generic SAS Program that would enable me to construct a Roll Rate Matrix / Markov Chain Matrix

based on Delinquency Data which can then be used for Loss Forecasting in Consumer Credit Risk

 

I would appreciate any help that is available

 

Thanks

Grand Advisor
Posts: 17,464

Re: SAS Macro Code for Delinquency Roll Rate Analysis

Getting specific subject matter expertise is unlikely but if you post and example of your data and calculation required you can probably get some help. At least it will be faster. 

Contributor
Posts: 59

Re: SAS Macro Code for Delinquency Roll Rate Analysis

Hi,

I appreciate your advice.

Thanks for responding to my request.

Basically the idea is to use empirical data to estimate the probability of
a customer changing their delinquency state within a unit of time (such as
a month) and to embed this information within a Markov Transition Matrix.

I will work on your suggestion, but in the meanwhile we can only hope...

Regards

##- Please type your reply above this line. Simple formatting, no
attachments. -##
Grand Advisor
Posts: 10,251

Re: SAS Macro Code for Delinquency Roll Rate Analysis

SAS also has a Risk Managment module that you may want to investigate.

Grand Advisor
Posts: 9,596

Re: SAS Macro Code for Delinquency Roll Rate Analysis

I suggest you post it at IML forum, since it is about matrix operator. @Rick might have something for you .

 

http://blogs.sas.com/content/iml/2016/07/07/markov-transition-matrices-sasiml.html

 

http://blogs.sas.com/content/iml/2016/07/13/absorbing-markov-chains-in-sas.html

Contributor
Posts: 59

Re: SAS Macro Code for Delinquency Roll Rate Analysis

Thank you for your response to my request.

 

I have accessed and read the two DO LOOP articles published by Rick Wicklin.

 

I have also directed my initial request to Rick for his advice.

 

Regards

Ask a Question
Discussion stats
  • 5 replies
  • 902 views
  • 0 likes
  • 4 in conversation