@Anonymous wrote:
@SouthJamaica wrote:If I see 2 guys who pay their bills promptly and who manage their credit deftly, and one has a little bit of his debt left to pay off while the other guy has paid it all off, I'm not going to say the 2nd guy is less creditworthy because he doesn't owe any money any more.
Paying your bills down to zero is what you're supposed to do. People who do that are the most creditworthy.
I agree with what you're saying above, but that doesn't really get to the meat and potatoes of my question to you which is at what point does past payment history (on loans or revolvers) matter less than current history to you? I do think the algorithm is flawed in that it only looks at a single moment in time and if there isn't a reported balance it "looks" like you aren't using your revolvers. I have no problem with the penalty for no revolving credit use, although it certainly shouldn't work the way it does with respect to reported balances. I think that's a fair gripe, but it is what it is. If the FICO Gods can come up with a way to "see" revolving credit use outside of reported balances the problem will be solved.
But back to my question to you, how about the element of time that I was asking about in paying back debts? If two otherwise equal profiles have exactly one closed [paid as agreed] auto loan on them, one being closed 1 year ago and the other closed 6 years ago, to me the profile with the closed 1 year ago loan is "better" because that installment loan account is more recent. That person has more recently shown the ability to pay back an installment loan as agreed. The longer the element of time elapses, the greater the chance for life factors to change which could make payback of an installment loan today more difficult. Same thing with revolvers. If someone cuts up all of their revolvers and leaves the accounts open all paid down to $0, 3 years from now we don't know if they'd still be able to pay those cards down to $0 if they had balances. I'm just saying that it makes sense for a penalty to be imposed for no revolving credit use. The argument really is how it's imposed, I suppose.
I'm reasonably certain that the FICO gods can in fact see and do in fact know whether an account has been active or not, without regard to whether it reports a balance or not.
There is even a specific data field "last activity".
@Anonymous wrote:Right, so hopefully a future algorithm will use data other than a reported balance to see revolving credit use.
I'd still like to know along a timeline at what point you feel past account data becomes less meaningful than current account data.
I don't know the answer to that. As a human -- as opposed to a computer -- I would want to know both before making a lending decision.
@Anonymous wrote:Hi SouthJ. I think that DOLA is not always submitted by every creditor on every account for every month. I once asked about that here on the forums and couldn't get anyone to confirm that it was. Even if that field is more complete in recent years (which I do not know to be true) it may well have been spottier in the past, which is when FICO 8 and earlier models were developed.
(Aside: when you pull your full reports at annualCreditReport.com, do you see a field called DOLA on the report?)
Examples: DOLA is submitted when accounts are delinquent, but perhaps much more unevenly if an account is clean (depending on the issuer).
If so, that would explain why they used the presence of at least one CC reporting a balance as a proxy to answer the question "Is this person using his cards?"
1. I don't know, I look at the reports from MyFICO, and haven't pored over them to see how consistent the reporting is.
2. In the internet age, I just don't feel that reported balance is a very sound proxy for account activity. If it's a vestige from an earlier time, that would be a benign and innocent explanation all right, but not a good excuse for failing to update.
@Anonymous wrote:Yep, there's a number of things that I have been hoping will make their way into the big FICO models (i.e. FICO 10). But they all depend on creditors beginning to supply reasonably complete data -- the fields are already there in the database, as you point out. But in order for FICO to use them we have to start seeing the creditors submitting data for them.
(1) Dropping the All Zero Cards penalty.
(2) Dropping the emphasis on ultralow utilization, and beginning to replace it with an emphasis on whether the person is primarily a transactor (pays his statement balance in full, which as you know is different from paying his cards to zero). TU did some studies a while back and showed that the Transactor-Revolver distinction is far more predictive for risk than utilization.
#1 could be accomplished with either the DOLA field (as you observe) or with the newer Trended Data.
For #2 we'd need creditors to submit the full trended data on their cards, which some do.
I'd also like to see...
(3) A stronger emphasis on a person's history of successfully managing and paying off loans (FICO 8 has no emphasis on this at all). TD would be helpful here but is not really necessary.
You'll get no argument from me on any of those items from your wishlist 