topic Re: Question about AAoA and AUs in Understanding FICO® Scoring

Question about AAoA and AUs

Anonymous — Fri, 29 Mar 2019 16:17:39 GMT

Hello. Quick question—is the age of account considered on AU accounts when calculating my AAoA, and in calculating my FICO and Vantage scores? It appears so, but I’m wondering if it somehow is treated differently than accounts on which I am the primary cardholder. And when I apply for a new card, does the computer system they use to approve or deny see that my oldest account is an AU? Likewise, when determining my overall utilization, is the AU card considered? Thank you so much for any insight you can give me!

Re: Question about AAoA and AUs

Anonymous — Fri, 29 Mar 2019 16:37:43 GMT

Certain AU accounts "count" and certain ones don't. If the account is being counted, it will impact your utilization, age of accounts factors, etc. As far as whether or not a "counted" AU account is factored into a lending decision, that would really matter on a lender-specific level. Some computers may value it well and it could help, where upon a MR a human being could discount the account entirely knowing that it's sort of artificially impacting your own profile.

Re: Question about AAoA and AUs

Anonymous — Sat, 30 Mar 2019 06:00:53 GMT

Only way to know for sure is to do a "efficacy" test, as I call it. Otherwise, you don't know whether the anti-abuse algorithm has flagged it. I'm doing one now, as I've been taught by more experienced members. Pay all accounts to 0, except the AU account. Then watch for the no revolving balance penalty. If you get it, the AU account isn't counting. https://ficoforums.myfico.com/t5/Understanding-FICO-Scoring/AU-efficacy-test/td-p/5557347/

Re: Question about AAoA and AUs

Revelate — Sat, 30 Mar 2019 16:46:01 GMT

To be excruciatingly clear though, AU's count full monty on everything before FICO 8; namely, 99.99% of the mortgage underwriting in the US they will count.

FICO 8 has an anti-abuse algorithm and I believe FICO 9 has the same though there were some early reports of it's counting differently nothing came of that.

Re: Question about AAoA and AUs

Anonymous — Sat, 30 Mar 2019 17:31:10 GMT

@Revelate True enough they would count on old mortgage algorithms. Which presents a question: would it do one any good if one already had an older account? If the AU was simply another old account that buttressed AAoA? Would it even be worthy of inclusion considering the low point gains for AAoA? IF there’s already a older established AooA?

Re: Question about AAoA and AUs

Anonymous — Sun, 31 Mar 2019 04:44:57 GMT

It's also worth noting that mortgages always involve a MR by a human being, so they'll able to "see through" the artificial gain from an AU account. It will still positively impact all mortgage scores, though, so improving that middle score is always a great thing.

Re: Question about AAoA and AUs

Revelate — Sun, 31 Mar 2019 09:39:16 GMT

AAOA appears to be a scorecard segmentation bit: that is always a good thing.

As for manual review BBS, they won’t discount it because the neither GSE’s nor the secondary market will... it’s just about the FICO score and the overwhelming majority of lenders if not all these days (really any without stupid overlays) when talking anything but jumbos or other portfolio loans, will UW a file that only has one AU tradeline on it to generate a FICO score and nothing else.

Money trumps any reason to kill that app cause of an AU affecting the score.

Re: Question about AAoA and AUs

Anonymous — Sun, 31 Mar 2019 19:20:25 GMT

Very interesting, I have not yet heard that AAoA was a scorecard segmentation factor. Correct me if I’m wrong, I thought it was 1. clean/derogatory, 2. Young or old (AooA), 3. thick/thin, 4. new accounts/no new accounts.

I thought those factors were scorecard determinates and that then based on other factors within that scorecard, your score was determined. Please correct me and instruct me where I am wrong or misunderstand.

Re: Question about AAoA and AUs

Revelate — Sun, 31 Mar 2019 19:39:44 GMT

@Anonymous wrote:
Very interesting, I have not yet heard that AAoA was a scorecard segmentation factor. Correct me if I’m wrong, I thought it was 1. clean/derogatory, 2. Young or old (AooA), 3. thick/thin, 4. new accounts/no new accounts.

I thought those factors were scorecard determinates and that then based on other factors within that scorecard, your score was determined. Please correct me and instruct me where I am wrong or misunderstand.

Hrm well I may have somewhat misspoke there but going through it: i'm not sure if thick / thin factors at all though we've seen some strangeness; however, given how so much of the algorithm is percentage based it shows up in other ways and we've seen max scores with 2 credit cards only IIRC.

Derogatory scorecards: type of derogatory (public record / collection or deliquency) and some factor of age presumably of the derogatory though we don't know what that breakpoint is and might be different based on type of derogatory.

It doesn't break down into 4 nice neat bins though I don't think on the clean scorecards: 30D lates play there too. New accounts matter on all models and I think oldest does on some, but thick/thin I'm kinda dubious on for several reasons, and AAOA dominates the reason codes for me certainly and it might be some combination of oldest / AAOA or even just AAOA on the newer models.

TT has some good pictures but I don't think there's a pat answer on the age factors and we certainly don't have a great idea on the AAOA breakpoints... though it does appear that length of revolver / installment tradelines which might well be AOOA do factor in the Industry Option scorecards for FICO 8, and that might be precisely oldest account but I don't recall ever seeing those reason codes in the classic score.

I'm sure TT will disabuse me if I'm wrong here .

Re: Question about AAoA and AUs

Anonymous — Sat, 06 Apr 2019 00:51:13 GMT

TT’s illustrations are where I learned what I said from. Although I must make one correction. According to his illustration once you determine which group of scorecards via thin or not thin, Then, if it’s thin, the next determine is whether or not there are new accounts. If the profile is not thin, then the next determine is whether or not there are any new accounts.

These are simply my interpretations of TT’s illustrations. And if I am incorrect please teach me and correct me. But that’s what led me to believe AAOA is not a scorecard determinate.

If you can teach me how to insert a picture from mobile I’ll insert the illustration.

Re: Question about AAoA and AUs

Anonymous — Mon, 01 Apr 2019 01:29:51 GMT

My understanding has always been that only AoYA and AoOA are scorecard assignment factors, not AAoA. It's possible though that AAoA could be a factor that may influence signal strength of other factors... I don't think that's out of the question.

Re: Question about AAoA and AUs

Anonymous — Mon, 01 Apr 2019 04:31:22 GMT

This is my understanding of scorecard segmentation.: Credit: Our own TT. @ThomasThumb Care to comment?

Re: Question about AAoA and AUs

Anonymous — Tue, 02 Apr 2019 02:31:53 GMT

Right and the "aged" sections of that diagram are referring to AoOA and AoYA, not AAoA.

Re: Question about AAoA and AUs

Anonymous — Fri, 05 Apr 2019 22:14:59 GMT

Well I think aged is referring to AooA and new accounts refers to AoYA, I think ... except when it’s referring to minors of course.

Re: Question about AAoA and AUs

Thomas_Thumb — Sat, 06 Apr 2019 00:54:01 GMT

The two age factors for score cards are:

1) New credit as defined by AoYA (age of youngest account)

2) File age as defined by AoOA (age of oldest account - open or closed)

Average age of accounts is a scoring attribute but not a scorecard assignment factor. People that see AAoA drop due to new accounts being added may change scorecards but, if so, the reassignment almost certainly relates to new credit.

File thickness is a scorecard assignment factor as well. However, it is a bit hard to nail this one down as it appears to comprise both type/mix of accounts and QTY of accounts. There is a gray region that I would classify as neither thin or thick (say over 4 accounts but less than 11 accounts). I'm in this category since 4 closed accounts dropped off my file.

SJ's file seems to react more strongly to certain scoring attributes than others with thick files. I suspect this is due to maintaining multiple new accounts under 12 months age. I would place him on the new credit-thick file-long credit history scorecard.

As we know Fico 8 has 8 "clean" scorecards and they have stated segmentation factors include credit history (age of file), new credit and depth of credit. Beyond that the exact segmentation is not published. My hypothesis on segmentation as mentioned in other threads is:

I go back and forth on newness segmentation as 1 year or 2 years and depth of credit dividing line as thin/non thin or thick/non thick. Also, mild delinquency was not originally listed in presentations on Fico 8 segmentation. It showed up later. So I think it is a tag along and score limiting attribute within "clean" scorecards as opposed to a true segmentation factor.

Re: Question about AAoA and AUs

SouthJamaica — Sat, 06 Apr 2019 02:27:22 GMT

@Thomas_Thumb wrote:
The two age factors for score cards are:
1) New credit as defined by AoYA (age of youngest account)
2) File age as defined by AoOA (age of oldest account - open or closed)

Average age of accounts is a scoring attribute but not a scorecard assignment factor. People that see AAoA drop due to new accounts being added may change scorecards but, if so, the reassignment almost certainly relates to new credit.

File thickness is a scorecard assignment factor as well. However, it is a bit hard to nail this one down as it appears to comprise both type/mix of accounts and QTY of accounts. There is a gray region that I would classify as neither thin or thick (say over 4 accounts but less than 11 accounts). I'm in this category since 4 closed accounts dropped off my file.

SJ's file seems to react more strongly to certain scoring attributes than others with thick files. I suspect this is due to maintaining multiple new accounts under 12 months age. I would place him on the new credit-thick file-long credit history scorecard.

As we know Fico 8 has 8 "clean" scorecards and they have stated segmentation factors include credit history (age of file), new credit and depth of credit. Beyond that the exact segmentation is not published. My hypothesis on segmentation as mentioned in other threads is:

I go back and forth on newness segmentation as 1 year or 2 years and depth of credit dividing line as thin/non thin or thick/non thick. Also, mild delinquency was not originally listed in presentations on Fico 8 segmentation. It showed up later. So I think it is a tag along and score limiting attribute within "clean" scorecards as opposed to a true segmentation factor.

Fascinating. I guess you've got me pegged. I guess I would be in C-2.

No derogatory items

Age of oldest account > 30 years

Definitely non-thin: total of open & closed accounts in the neighborhood of 60

Lots of new accounts, average age of accounts < 4 years