When will credit card company's begin using FICO 9?

[Anonymous]

---

@CreditMagic7 wrote:

---

Hi CGID.

Just to touch momentarily on trended data aspect.

In your opinion IF and/or more likely WHEN, a cardholder(s) "revolves" a balance or balances AT ALL, is this activity whether it is once. twice etc. compared to PIF routinely on all other accounts, trended to some comparison in order for distinctions to be made to someone who revolves vs transacts.

In others words, are we close to right to assume that this data to the revolving activity will be collected and scored by percentages to weigh if the Credit User is considered one or the other?

This will be a great topic for further discussion in another thread to maybe better hammer out who will fall into which category and if that might be distributed on a percentage basis (how many times a balance is rolled) compared to PIF routine.

Thanks for any answer in advance.

Can you tell us what you know?

---

Great questions.  I think very little is know about how existing or future R-T analysis programs will work.  But in my opinion I think your "big picture" intuition is right.

That is to say, the developers will likely assign the person's R or T behavior along a spectrum.  It won't be simply on or off -- either you have been paying in full on every card for the last 30 months (T) or else you are an R.

At the pure T end of the spectrum will be people who:

    * Do use their cards

    * Always pay in full

    * Have done so for every month in the last 24 months

At the other end will be the most risky kind of revolving behavior:

    * All cards carrying a balance

    * Only making minimum payments

    * Balances tend to either only go down a tiny bit each month or are going up

    * This is the typical behavior each month for the last 24 months

In the middle will be people who do not full neatly into either category.  You can imagine people getting a score from 1 to 5 (say).  1 would be almost pure T and 5 would be high risk R.

If a person in 2017 or 2018 decides he wants to be an R to take advantage of a 0% card, here are things that the R-T analytic engline would likely look for as markers of low risk:

*  Is he a full T on most of his cards?  (I.e. he's paying in full every month on the others)

*  Is he regularly paying much more than the minimum payment on the one account he is carrying a balance on?

*  Has he been a full T on several months of the 24-month history (i.e. his R-behavior is not the norm)

*  Has he been a full T in the last six months?  (I.e. the R-behavior has not happened recently)

Bear in mind that I am describing how I imagine an R-T analysis would work from the perspective of a lender considering you for an installment loan or a charge card.  For them, their sole concern is you becoming delinquent.  A credit card issuer would use trended data in much more complex ways.  For them a customer being a revolver would not necessarily be a bad thing, since he pays interest.  They'd use trended data to (as much as possible) identify customers who are relatively low risk but who are also immensely profitable (either via swipe fees or paying interest).

[Anonymous]

---

@HeavenOhio wrote:

---

 With the other cards, it's pretty easy to tell if I've been revolving or transacting. And in the couple of cases where I was revolving (0% promos), it's easy to tell that I was making payments substantially above the minimum.

---

Great to hear that.  See my post immediately above, where I talk about things people can do who want to carry a balance on a 0% card but who also want to be classified as largely a T (or a very low risk R) by an R-T analysis.  Again, that's all speculation on my part, but it certainly seems reasonable.

[CreditMagic7]

---

@Anonymous wrote:

---

@CreditMagic7 wrote:

---

Hi CGID.

Just to touch momentarily on trended data aspect.

In your opinion IF and/or more likely WHEN, a cardholder(s) "revolves" a balance or balances AT ALL, is this activity whether it is once. twice etc. compared to PIF routinely on all other accounts, trended to some comparison in order for distinctions to be made to someone who revolves vs transacts.

In others words, are we close to right to assume that this data to the revolving activity will be collected and scored by percentages to weigh if the Credit User is considered one or the other?

This will be a great topic for further discussion in another thread to maybe better hammer out who will fall into which category and if that might be distributed on a percentage basis (how many times a balance is rolled) compared to PIF routine.

Thanks for any answer in advance.

Can you tell us what you know?

---

Great questions.  I think very little is know about how existing or future R-T analysis programs will work.  But in my opinion I think your "big picture" intuition is right.

That is to say, the developers will likely assign the person's R or T behavior along a spectrum.  It won't be simply on or off -- either you have been paying in full on every card for the last 30 months (T) or else you are an R.

At the pure T end of the spectrum will be people who:

    * Do use their cards

    * Always pay in full

    * Have done so for every month in the last 24 months

At the other end will be the most risky kind of revolving behavior:

    * All cards carrying a balance

    * Only making minimum payments

    * Balances tend to either only go down a tiny bit each month or are going up

    * This is the typical behavior each month for the last 24 months

In the middle will be people who do not full neatly into either category.  You can imagine people getting a score from 1 to 5 (say).  1 would be almost pure T and 5 would be high risk R.

If a person in 2017 or 2018 decides he wants to be an R to take advantage of a 0% card, here are things that the R-T analytic engline would likely look for as markers of low risk:

*  Is he a full T on most of his cards?  (I.e. he's paying in full every month on the others)

*  Is he regularly paying much more than the minimum payment on the one account he is carrying a balance on?

*  Has he been a full T on several months of the 24-month history (i.e. his R-behavior is not the norm)

*  Has he been a full T in the last six months?  (I.e. the R-behavior has not happened recently)

Bear in mind that I am describing how I imagine an R-T analysis would work from the perspective of a lender considering you for an installment loan or a charge card.  For them, their sole concern is you becoming delinquent.  A credit card issuer would use trended data in much more complex ways.  For them a customer being a revolver would not necessarily be a bad thing, since he pays interest.  They'd use trended data to (as much as possible) identify customers who are relatively low risk but who are also immensely profitable (either via swipe fees or paying interest).

---

I'm sure many of us would agree with this Theory as a rough baseline of what to expect of it.

As things move more into this style and formulation of credit data history and measurings certain aspects of it should begin to become more easily narrowed down, probably as well as many other elements of FICO scorings that are much simpler these days to draw a more accuarate data read on our own profiles.

Thanks CGID.

[Anonymous]

Well then, it looks like trended data won't be used for some time.

If FICO 9 has been in the works for 3-5 years at this point and still isn't widely used yet, we can infer that FICO 10 (and trended data) are likely another 3-5 years away beyond FICO 9 becoming more common.

[Anonymous]

Hi BBS.  Trended data won't be used inside a major FICO model for a while.  But it doesn't mean it won't be used by lenders.  It's already a part of the underwriting software used for most mortgage loans (though I believe FHA loans are exempted -- can't remember for sure).  And that started in Sept of last year (top down decision by Fannie Mae).

Another thing to remember is that sometimes lenders skip FICO versions.  Big companies do this with computer operating systems too.  Many big companies are migrating from Windows 7 to Windows 10 -- my workplace is doing that this year I believe.  So it's quite possible that a company could jump from FICO 8 to FICO 10.  (FICO NextGen was a resounding failure -- many or most companies skipped over it.)

I think CC issuers will be really interested in using trended data to identify lucrative customers.  Specialty products (some of them likely being developed by FICO) may be used as supplements to the FICO 8 or 9 score.

We'll know more as time goes on certainly.

[Anonymous]

I agree that trended data will be extremely useful to creditors in the upcoming years.  For those of us out there that are Transactors, it will give us an advantage over others that currently can't be realized.  Not that the "game" of credit is a competition, but in terms of being the best "fit" customer for a product it's good to know that sound Transactor behavior will allow such a person an advantage over someone that is a Revolver.