It has been a while since I posted here in AF. I am currently doing credit monitoring for a multinational bank with borrowers across Asia and the Pacific. One of my upcoming project is to come up with a systematic way of monitoring early warning triggers in credit monitoring using publicly available data.
Just in case, may I ask for your suggestions if what data (e.g. publicly available, macro economic indicators) to look out for that can serve as early warning triggers
Ratings - get a report each day which lists those entities which have ratings changes by any of the 3 major rating entities
Price Changes - get a report each day which lists those entities which have price changes of >10%.
New alerts - set up news alerts for each of the companies in the portfolio that have a list of key words in them i.e. default, warning or error… or something like that.
Get your crappy credit analysts to actually do some work and monitor their counterparties.
Pick an industry which is heavily correlated to the underlying businesses… i.e. maybe you have a bunch of producers or consumers that heavily correlated to aluminum price. Set up some alert for aluminum price…
Tough question. The only approach is to use an AI with non parameterized principal component analysis to analyze the relationship between credit and factors like global temperature, maine lobster population, volcanic activity, and holiday sales at Walmart. You’ll need some heavy computational ability to process all this data, so make a budget of about $75 million for supercomputers. Following that, adopt a cat and have the cat pick securities to simulate randomized behavior. The cuddles are a secondary benefit. Please update us on your progress. Tanks.