topic Re: Was I re-bucketed or something? in Understanding FICO® Scoring

Was I re-bucketed or something?

wasCB14 — Tue, 29 Nov 2016 01:39:35 GMT

My TU according to both Discover and Barclaycard-provided scores (both less than a week old) fell from about 780 to 733 in the last month. Puzzled, I got the full report from annualcreditreport.com.

Nothing amiss in the report. It's just a sea of green "OK"s.

No new accounts in the last 6 months. One TU inquiry 18 months old. No baddies. 2% utilization overall, with balances on 7 of 13 open cards (which is normal for me, and no single-card utilization was high). No installment history, just 13 open cards (and one closed). AAoA just under 2 years. Everything has "Current" status.

Is it likely I was re-bucketed? 47 points is a big drop.

The only other explanations I can think of would involve TU thinking a card was inactive. One card hasn't had a balance report for 8 months. Another hasn't had a balance report for 6 months. A third had a balance report for the first time in the last 14 months. I was using them during that time, but not much, and generally paying before the statements cut.

No apps planned for a few months, I'm mostly just curious what might explain this.

Re: Was I re-bucketed or something?

Sbrooks1 — Tue, 29 Nov 2016 02:24:30 GMT

I'm thinking it is the card that had not reported a balance in 14 months. Just my opinion though.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 02:57:51 GMT

For clean profiles, scorecard assignment in FICO 8 is based on three factors:

(1) Age of oldest account

(2) Age of youngest account

(3) Total number of accounts (i.e. whether or not you are a thin profile)

#3 couldn't be at issue since you have so many accounts.

What is #1 and #2 for you? (Include closed accounts if appropriate.)

Re: Was I re-bucketed or something?

wasCB14 — Tue, 29 Nov 2016 03:21:48 GMT

@Anonymous wrote:
For clean profiles, scorecard assignment in FICO 8 is based on three factors:

(1) Age of oldest account
(2) Age of youngest account
(3) Total number of accounts (i.e. whether or not you are a thin profile)

#3 couldn't be at issue since you have so many accounts.

What is #1 and #2 for you? (Include closed accounts if appropriate.)

Oldest account just over 3 years. Open.

Youngest account just over 6 months. Open.

13 open, 1 closed...consisting of:

11 open major network credit cards. 1 open store-only card. 1 open Amex charge card. 1 closed (on good terms) Amex charge card.

No Amex backdating, for that matter.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 03:50:21 GMT

Your rebucketing hypothesis has some real possibility. Right around the time that your score changed, your age of oldest account crossed over from 2.9 to 3.0. Age of oldest account is one of the three factors used in scorecard assignment.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 13:11:55 GMT

I don't know how I feel about rebucketing. I think maybe the ranges on the buckets are a bit bogus. Using the data from the OP, I don't think that one should ever lose 47 points from being rebucketed to a "better" bucket based on "better" profile data... such has crossing from 2 year AAoA to 3 year AAoA for example. Losing a few points sure, but 47 is significant and could make a difference in obtaining the best interest rates in many cases. Maybe not for the OP considering his scores, but for others it surely could. Ironically, assuming AAoA was the deciding factor here, if the OP were to open a new account today, next month when it reported he'd likely get back those 47 points, maybe 40 net after the inquiry/new account.

Re: Was I re-bucketed or something?

newhis — Tue, 29 Nov 2016 13:40:44 GMT

Going over the 3 year mark on oldest card and 6 months newest card, doesn't make sense. Unless there is a higher risk someone having oldest card over 3 years (than 2), with AAoA of less than 1 year, I mean, crossing the 3 year mark on oldest account.

My DW is close to 3 year mark on oldest card. Her AAoA just crossed the 2 year mark. I don't expect a score drop.

OP, if you can, pay the 3 cards you mention and report $0 again.

Any of your cards is above 30%? or higher and was not that high last month?

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 14:36:12 GMT

Hi BBS, hi newHis. You guys are both using the abbreviation AAoA connected with the possibility of re-bucketing (scorecard reassignment). Just for clarity, the factor that might be involved is AOA not AAoA. AOA is age of oldest account -- AAoA is average age of accounts.

I am definitely not saying that re-bucketing is for certain (or even probably) what happened. But given that his AOA went from 2.99 to 3.0 at almost exactly the same possible, the answer to the question in the subject line is: Yes, given that one of your three scorecard factors may have crossed a threshold value, it is possible.

By way of contrast, if none of those three factors had crossed a possible threshold, I woild have said "it's gotta be something else."

To give a metaphor that may make rebucketing seem less baffling, here is one from high school that should be familiar. Suppose you have two football teams, the one from junior high (grades 6-8) and the one from high school (grades 9-12). Johnny is a huge success in his last year on the junior high team -- he's bigger and stronger and has been playing longer than most of the other kids. Then he crosses over into high school. He's now on a different team. Suddenly the guys are much bigger and he's the one with the LEAST exerience. He's getting run over! The good news is that if he hangs in there for just a little while, he will by comparison start being much better than some of the other guys.

Rebucketing is like that. I agree that the score drop is bigger than I'd expect so reB may not be the explanation or only past of it. But we should expect (by way of the football story above) the possibility of a score drop. Johnny can't stay on the junior high team forever. And when he is switched over, it will seem hard and even unfair at first.

Re: Was I re-bucketed or something?

newhis — Tue, 29 Nov 2016 15:02:39 GMT

Thank you CreditGuyInDixie.

I tried to explain that a 2 year AOA with less than 1 year AAoA could be better than 3 year AOA and less than 1 year AAoA (different bucket).

DW will be crossing 2->3 year AOA but last month she crossed the 2 year AAoA. My guess is that her score will not drop as hard as OP, because she will be 3 year AOA and 2 year AAoA, and because she is paying down her BT card and car loan, maybe the score will not drop at all.

Do you have any information about cards without balance for 6 or 14 months, sudenly reporting a balance, if that changes the score?

I don't use a card every month, and last month I let it report and I lost about 12 points (I think because of this), from 801 to 789. Maybe the last time I used it was 6 months ago.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 15:38:59 GMT

CGID,

I misread what you wrote regarding AoOA earlier; I read it as AAoA which is why I responded the way I did. My apologies!

I like the football analogy above and think it makes sense with respect to rebucketing. However, one thing is missing from the equation. With the football analogy, we have the luxury of human rationalization. We can easily rationalize why Johnny was better than average on the 6-8 team but below average on the 9-12 team. To assign some FICO scores to this for fun, perhaps Johnny was a 780 on the younger team and he's only a 660 on the new team for now until he grows, gains experience, etc. Someone scouting the team however has the ability to consider previous data (his previous 780 prowess) and can likely deduce that the current 660 is in fact not necessarily representative of Johnny's upside. I don't think in this situation Johnny would be looked adversely upon.

With respect to credit rebucketing, to my knowledge the algorithms don't take into account past data. There's no human rationalization in the score produced and it's more or less taken at face value. There's no way on the surface of looking at the new, lower score and determining if it's a "Johnny" and will be a 780 in a year or two or if it will stay at 660. Hopefully that makes sense. A creditor pulling someone's credit report gets a report and score and can't see what "bucket" someone is in; they can't determine what team Johnny is on. All they can go off of is what they see right now and the score that goes along with it.

I guess my argument is that FICO scores are supposed to be a product of the credit profile and accurately (as loosely as that word is used) predict risk. If factors are changing that are lowering the perceived risk, such as aging accounts for example, it seems counter intuative that a positive change results in a score that's less representative of that updated profile.

Re: Was I re-bucketed or something?

SouthJamaica — Tue, 29 Nov 2016 17:17:22 GMT

@wasCB14 wrote:
My TU according to both Discover and Barclaycard-provided scores (both less than a week old) fell from about 780 to 733 in the last month. Puzzled, I got the full report from annualcreditreport.com.

Nothing amiss in the report. It's just a sea of green "OK"s.

No new accounts in the last 6 months. One TU inquiry 18 months old. No baddies. 2% utilization overall, with balances on 7 of 13 open cards (which is normal for me, and no single-card utilization was high). No installment history, just 13 open cards (and one closed). AAoA just under 2 years. Everything has "Current" status.

Is it likely I was re-bucketed? 47 points is a big drop.

The only other explanations I can think of would involve TU thinking a card was inactive. One card hasn't had a balance report for 8 months. Another hasn't had a balance report for 6 months. A third had a balance report for the first time in the last 14 months. I was using them during that time, but not much, and generally paying before the statements cut.

No apps planned for a few months, I'm mostly just curious what might explain this.

I think it's just a glitch. TU does that some times. There's absolutely nothing in your profile that should have caused a 47 point drop.

I'm betting it pops right back.

The problem is that when it does pop back, you won't know for awhile, because you're getting your score the super-slow way, from the credit card site.

Re: Was I re-bucketed or something?

wasCB14 — Tue, 29 Nov 2016 17:33:38 GMT

@newhis wrote:
+1

Going over the 3 year mark on oldest card and 6 months newest card, doesn't make sense. Unless there is a higher risk someone having oldest card over 3 years (than 2), with AAoA of less than 1 year, I mean, crossing the 3 year mark on oldest account.

My DW is close to 3 year mark on oldest card. Her AAoA just crossed the 2 year mark. I don't expect a score drop.

OP, if you can, pay the 3 cards you mention and report $0 again.
Any of your cards is above 30%? or higher and was not that high last month?

I'm paying off the one that had the first balance report for 14 months. The other two have no current balance.

My highest single-card utilization is around 8%. That's very low compared to my normal spending and payment habits. My charge card balance is presently under $100.

Also, if it is a bucket thing, 6 full months is the longest I've been in the garden. So that could be two of the three factors changing: Length of history, and time since last account.

I also had an inquiry fall off in the last month, going from 2 to 1.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 17:38:32 GMT

Hi BBS. The way that (for me) the football analogy works is this.

In the footbal example, Johnny was a big fish in a small pond. It was natural for him to shine on the Junior High team, because all the other kids on his team were small, weaker, and had played football for less time than he had.

Then he moves to the high school and the guys on his team are all bigger than he is! Bigger, stronger, faster, and know more about football. He was the star player six months ago -- but now in football practice every afternoon he's being run over by the guys on his team. He's no longer the star player. He's the weakest player! He may be on the bench a lot. And so on. (But as I also said, over time that will change. If he stays at it, eats his Wheaties, reads his playbook, he will eventually be one of the big stars on the high school team. If he then goes on to play college ball, he'll have another rude awakening his freshman year there -- he was rebucketed again, darn it!)

FICO scorecards are like that. FICO groups people it regards as being "similar" into 12 groups. (I think that's the number of scorecards in FICO 8.) The exact details behind how a person gets classified is a trade secret, but some things are known. For example, there are score cards for dirty profiles and those that are clean. (A person can have a couple 30-day lates and be clean in this sense.) Clean profiles are then placed into 8 scorecards, depending on the three factors I gave earlier. (That the method involves these and only these three factors is also known.)

A person is then scored by FICO against the people in his group (scorecard). Another high school analogy is that it is like being graded on a curve for an exam. Your exam score is based on well you do relative to the other people in that class -- in that "scorecard." The same exam might be given to two different classes with person X in one class and person Y in the other -- with X and Y getting exactly the same questions right. Nonetheless, if the students in X's class knew the material better than the students in Y's, then X will get a worse grade.

Hope this makes a bit more sense.

What you may be wondering is why FICO does scorecards at all. Why not just let everyone be on the same team and rank everyone against each other? Why split everyone into 12 subteams and rank them against the people only on their team?

I am not a statistician but the answer must be that FICO has to determined that this in general creates a more reliable score for lenders. And it's not just FICO that does this. Every major scoring model does this. E.g. Vantage 3 does scorecards too.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 17:45:42 GMT

@wasCB14 wrote:
@newhis wrote:
+1

Going over the 3 year mark on oldest card and 6 months newest card, doesn't make sense. Unless there is a higher risk someone having oldest card over 3 years (than 2), with AAoA of less than 1 year, I mean, crossing the 3 year mark on oldest account.

My DW is close to 3 year mark on oldest card. Her AAoA just crossed the 2 year mark. I don't expect a score drop.

OP, if you can, pay the 3 cards you mention and report $0 again.
Any of your cards is above 30%? or higher and was not that high last month?
I'm paying off the one that had the first balance report for 14 months. The other two have no current balance.

My highest single-card utilization is around 8%. That's very low compared to my normal spending and payment habits. My charge card balance is presently under $100.

Also, if it is a bucket thing, 6 full months is the longest I've been in the garden. So that could be two of the three factors changing: Length of history, and time since last account.

I also had an inquiry fall off in the last month, going from 2 to 1.

I am pretty sure that scorecard assignment works on integer year values. But nobody knows for sure so it is certainly conceivable that a "youngest account" breakpoint is set by FICO 8 for exactly six months. More likely in my opinion is that it is at one year (or possibly even two years).

Because of all that, if you have been rebucketed (that's an IF), then I'd guess it was age of oldest turning 3.

Re: Was I re-bucketed or something?

Thomas_Thumb — Tue, 29 Nov 2016 17:54:43 GMT

@wasCB14 wrote:
My TU according to both Discover and Barclaycard-provided scores (both less than a week old) fell from about 780 to 733 in the last month. Puzzled, I got the full report from annualcreditreport.com.

Nothing amiss in the report. It's just a sea of green "OK"s.

No new accounts in the last 6 months. One TU inquiry 18 months old. No baddies. 2% utilization overall, with balances on 7 of 13 open cards (which is normal for me, and no single-card utilization was high). No installment history, just 13 open cards (and one closed). AAoA just under 2 years. Everything has "Current" status.

Is it likely I was re-bucketed? 47 points is a big drop.

The only other explanations I can think of would involve TU thinking a card was inactive. One card hasn't had a balance report for 8 months. Another hasn't had a balance report for 6 months. A third had a balance report for the first time in the last 14 months. I was using them during that time, but not much, and generally paying before the statements cut.

No apps planned for a few months, I'm mostly just curious what might explain this.

I think you answered your own question - cause of score drop due to cards that have been inactive 6 or more months reporting activity.

Others have reported score drops due to this situation. My understanding is scores should rebound the next reporting cycle. Your score drop is a bit severe but, you have re-activated 3 cards and one has been inactive over 12 months.

There could be something else going on but, the inactive cards reporting activity is my bet.

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 18:00:21 GMT

I wish that there was a definite way to figure out WHY you experienced the score drop. Either way (re-bucketing or card inactivity) it would tell us something really interesting.

With re-bucketing it would give us some evidence that a scorecard breakpoint may be at 3 years (age of oldest). That's REALLY cool to know if true.

With card inactivity, it would mean that using the card while it continues to report $0 each month is not sufficent to make FICO think the card is actually active. Some people think using the card and then paying to zero (PTZ) is enough to prevent FICO from categorizing it as inactive after a while. But your case would show that allowing cards to report a balance and then PIF is better at avoiding the inactivity issue.

But I agree with SouthJ that we may never know what caused it.

If you think that inacativity might be the culprit (even a chance) you could consider switching to PIF rather than PTZ (your current strategy for some of your cards). Allowing balances to report and then using autopay may be easier too. There's been some recent threads that are increasingly bringing to everyone's attention that FICO 8 may not care with you have a lot of $0 balances or not. That used to be the received wisdom (and it is certainly true about earlier FICO models) but there may be little advantage in PTZ even in the month right before a big credit pull (unless it is a mortgage). FICO 8 may simply not care whether 5 of your 6 cards are at $0 or whether only 1 card is at $0.

Re: Was I re-bucketed or something?

SouthJamaica — Tue, 29 Nov 2016 18:13:27 GMT

@Anonymous wrote:
I wish that there was a definite way to figure out WHY you experienced the score drop. Either way (re-bucketing or card inactivity) it would tell us something really interesting.

With re-bucketing it would give us some evidence that a scorecard breakpoint may be at 3 years (age of oldest). That's REALLY cool to know if true.

With card inactivity, it would mean that using the card while it continues to report $0 each month is not sufficent to make FICO think the card is actually active. Some people think using the card and then paying to zero (PTZ) is enough to prevent FICO from categorizing it as inactive after a while. But your case would show that allowing cards to report a balance and then PIF is better at avoiding the inactivity issue.

But I agree with SouthJ that we may never know what caused it.

If you think that inacativity might be the culprit (even a chance) you could consider switching to PIF rather than PTZ (your current strategy for some of your cards). Allowing balances to report and then using autopay may be easier too. There's been some recent threads that are increasingly bringing to everyone's attention that FICO 8 may not care with you have a lot of $0 balances or not. That used to be the received wisdom (and it is certainly true about earlier FICO models) but there may be little advantage in PTZ even in the month right before a big credit pull (unless it is a mortgage). FICO 8 may simply not care whether 5 of your 6 cards are at $0 or whether only 1 card is at $0.

Although I use each of my cards every month, the vast majority have no reported balance, and it's possible that many of them have never reported a balance, because I'm the type of person who likes to pay them off even before the statement cuts. I don't think the bureaus, or the FICO algorithms, think the accounts are inactive.

Re: Was I re-bucketed or something?

Thomas_Thumb — Tue, 29 Nov 2016 18:21:46 GMT

SJ -

Accounts that show payments - which your's do if they were used and paid off show activity. Thus, they would not be tagged as inactive.

Re: Was I re-bucketed or something?

SouthJamaica — Tue, 29 Nov 2016 18:28:52 GMT

@Thomas_Thumb wrote:
SJ -

Accounts that show payments - which your's do if they were used and paid off show activity. Thus, they would not be tagged as inactive.

Thanks.

OP had stated that he was using the cards and making payments: " I was using them during that time, but not much, and generally paying before the statements cut."

Re: Was I re-bucketed or something?

Anonymous — Tue, 29 Nov 2016 18:35:01 GMT

Thanks SJ!

Thanks for drawing TT's attention to the fact that our OP has been using the cards and paying to zero most or every month

This is an area that Thom Thumb knows much more about than myself, but I would love to see some clear answers on it. I think it would have practical implications for a lot of us with cards we shoebox or PTZ. I would love to see some concerted attempts by several people to answer the following questions:

Do these "inactivity" scoring penalties exist for certain? On what grounds do we know this?

If yes.....

Is it in both the mortgage models and FICO 8?

Does using a card but paying to zero prevent FICO from considering it inactive?

Does the scoring penalty kick in only after the "inactive" card is used again?

Does "used" mean a transaction or allowed to report a positive balance to the CRA?

There should be easy ways to test all of these claims empirically and by a variety of people and profiles.