topic utilization - percentage vs. absolute dollar amount in Understanding FICO® Scoring

utilization - percentage vs. absolute dollar amount

expatCanuck — Sat, 20 Oct 2018 20:35:29 GMT

Greetings -
I was wondering when/whether, for revolving credit, absolute dollar amount ever factors into the FICO calculation.

For example, a year ago, I had an aggregate credit limit of 40K -- 10K posting to the bureaus would mean 25% utilization. A year later, I have about four times that credit limit; that same 10K works out to 6.25% utilization.

My score a year later is decidedly improved, but I'm wondering -- does a credit issuer turn the same 'blind eye' to that 10K? I'd think that, since one is asked for one's income on an application (and sometimes required to substantiate it), DTI factors into the calculation.

And, now that I've 'talked through' this, I think may be beginning to get it -- in part. The FICO score is income ignorant -- but presumably, at some point, it takes absolute dollar amount into account.

Re: utilization - percentage vs. absolute dollar amount

Anonymous — Sat, 20 Oct 2018 17:15:28 GMT

Some scoring theorists on this forum (e.g. Thomas Thumb) are very open to the possibility that FICO takes absolute dollar amounts into consideration. There are a lot of reasons to consider that some FICO models might do this.

Others are adamant that no FICO model does this ever.

Personally I am agnostic on the issue. I'd be delighted to see some well-designed tests run this year or next that focus very specifically just on this issue (tests that could be run by more than one person). Such tests certainly could be done, if someone really wanted to -- and as you say it's an interesting question.

From a practical perspective it often doesn't matter:

* When you are not applying for credit, your score doesn't matter. (Unless your scores are so low that you are at constant risk of lenders taking adverse action: in which case just implement AZEO and work on bringing your scores up.)

* When you are applying for credit (and if you need some extra points) just implement AZEO with a small dollar value on one card.

Re: utilization - percentage vs. absolute dollar amount

Anonymous — Sun, 21 Oct 2018 00:06:18 GMT

I firmly below that total amounts owed is factored into at least some fico models. I think Experian for sure. The worst bracket being 20k+ owed between non mortgage loans and revolving.

It is a known fact that "average balance" is absolutely a FICO factor on some models (0 to 49$, 50$ to 100, $101 to 200 etc) which is similar to "total amounts owed" just a different way of looking at it.

Re: utilization - percentage vs. absolute dollar amount

Anonymous — Tue, 23 Oct 2018 19:33:52 GMT

ExpatCanuk said: “I'd think that, since one is asked for one's income on an application (and sometimes required to substantiate it), DTI factors into the [FICO] calculation.” I disagree with your conclusion. Yes, it is possible that DTI factors into the calculation on whether to approve a loan or not; but it does not factor into your FICO score calculation. Firstly, there are many other factors that could have helped your score. Mainly your CL increased 400% thus lowering your total utilization accordingly. That is a big factor. Also your accounts aged a year and maybe some inquiries dropped off. There are a ton of other factors that could have helped you but I don’t think that it had anything to do with your income.

I really wanted to check out some of the other information stated on this thread. So what I did was pull up my last 3-B FICO report (I get the monthly pulls). NotTomCruise, I’ve heard the phrase “average balance” before, but for the life of me I cannot find it anywhere on any CR for any type of loan. The reference to “average balance” that I am familiar is with respect to checking, savings, investment, and money market accounts; but they have nothing to do with CRs and FICO scoring.

Every one of my loan types provide the exact same categories. Some items are left blank (like credit limit on mortgage, auto and personal loans) but the categories remain. My revolving, personal, auto, and mortgage loans all say: “Last Updated, Payment Status, Worst Delinquency, Balance [and] Credit Limit.” Under the More Details section it states: “Open Date, Last Activity, Terms, Scheduled Payment, High Balance, Loan Type, Responsibility, Company Name, Account Number, Contact Info [and] Comments.” Nowhere does it state “average balance.”

I do not understand what you mean when you say “…‘average balance’ is…similar to ‘total amounts owed’ just a different way of looking at it.” The only way I can see average balance as any variable would be in revolving accounts (other than the banking instruments I mentioned). Installment (auto, mortgages, personal, and student) loans have a decreasing balance (assuming you always pay on time and don’t get late charges, fees etc.) They do list “High Balance” which in these cases is the original loan amount, but there is no “average” that has any reporting value.

Revolving (CC or HELOC) accounts could have an average, but I’ve never seen one on my reports nor would it make any sense. The CC companies report your balance on your statement date (or reporting date to the CRAs if different). They report your balance at that time and your CL. I know some show a minimum payment for that month under “Scheduled Payment” but those are rare companies and do not represent any type of “average.” Once again this assumes you pay on time and the companies report once per month/cycle.

CGID why would TT and other theorists think that “…absolute dollar amounts [are a] consideration?” What would be the “…reasons to consider that some FICO models might do this?”

Rhetorically I ask; is an annual average of a million dollars in debt a large or small value? To me it would be large but to Elon Musk it is pocket change. Yet, the FICO models have no input location for our specific data; thus it is valueless output. That is why UWs require DTI for certain loans (mortgages for sure but personal, auto and CCs loans as well in some cases).

But back to revolvers. How would an average balance be of value even in this case. I understand that if an individual always carries a balance, then the FICO model could (over some defined time period) calculate an average; but where is that in the FICO stated score determinators? If a borrower always paid the minimum due for 11 months they would see the monthly debt relatively unchanged; however, if on the 12th month they paid off the balance in full, then what would their average monthly balance be? FICO does not back track, so there would have to be a month-over-month calculation on top of their monthly utilization calculation. FICO does not state that month-over-month calculation is performed and thus I must accept it is not a factor.

Also, how would AZEO work in that case if the period were a monthly cycle. If a borrower had a $10,000 CL and used that $10K in credit that month, but paid it in full before the reporting date; what would be their average balance? $5K for the month, maybe $9K if they purchased at the beginning of their cycle? How much? We all know that this is not the case for their reported balance is $0 and their utilization is also 0% for that cycle.

So I revert to my original statement. Average balance is a factor for banking accounts, but not for FICO from what I understand. You can put me in the “adamant [category] that no FICO model does this ever.”

Re: utilization - percentage vs. absolute dollar amount

Thomas_Thumb — Wed, 24 Oct 2018 02:40:49 GMT

@Anonymous wrote:
     ExpatCanuk said: “I'd think that, since one is asked for one's income on an application (and sometimes required to substantiate it), DTI factors into the [FICO] calculation.” I disagree with your conclusion.

     I do not understand what you mean when you say “…‘average balance’ is…similar to ‘total amounts owed’ just a different way of looking at it.” The only way I can see average balance as any variable would be in revolving accounts (other than the banking instruments I mentioned). Installment (auto, mortgages, personal, and student) loans have a decreasing balance (assuming you always pay on time and don’t get late charges, fees etc.) They do list “High Balance” which in these cases is the original loan amount, but there is no “average” that has any reporting value.

     CGID why would TT and other theorists think that “…absolute dollar amounts [are a] consideration?” What would be the “…reasons to consider that some FICO models might do this?”

     So I revert to my original statement. Average balance is a factor for banking accounts, but not for FICO from what I understand. You can put me in the “adamant [category] that no FICO model does this ever.”

Y

George Bush said: "Read my lips". Fico says - read my reason codes.

If you want to know what Fico looks at, read their reason statements. Nothing theoretical about it. Theory only comes in play when considering how important is the factor in scoring, does the factor have multiple thresholds and is it applied differently depending on scorecard. The below is pretty clear - Fico looks at a revolving utilization (ratio) and at both aggregate revolving balance and total balance [X mortgage(s)] in absolute $$ terms. Looking at an average balance is NOT mentioned. Average has no significance to risk by itself.

As we know someone making $50k can have a $500k aggregate revolving CL and someone makiing $500k may only have a $50k aggregate revolving CL. Looking at absolute $$ in addition to utilization may be a sensitivity adjustment. Models include attributes that an analysis of data show are statistically significant for improved correlation. It really has little to do with "reasoning" that colors everyones individual perspective.

Re: utilization - percentage vs. absolute dollar amount

Anonymous — Wed, 24 Oct 2018 03:32:35 GMT

Good post from TT above. Reason statements don't lie.