Utilization has no memory... But should it?

[Anonymous]

I was thinking about this today and am sort of conflicted about it.  I think very valid arguments could be made on both sides here and I've wrestled with both sides of the arguement myself and am not sure I came away with any strong feeling toward either side in the end.  I'd very much like to hear the opinions of all of you.

As we know, utilization has no memory as we say quite often on this forum.  One can have 100% maxed-out utilization one month and pay it off completely the next month to 0% and upon 0% reporting scores will reflect that exact snapshot in time with zero influence on the maxed-out utilization a month prior.

Obviously the most current information (today) is what is most impactful with respect to credit scores and the older that information gets the less of an influence it has on scores.  With utilization this is very much the case, but I find it odd that previous [recent] history isn't considered at least in part.  Payment history certainly has memory, as if you were late 12 months ago on a payment even with 11 months of on time payment history since, that late payment will continue to hurt your scores.  The reasoning for this of course is that if someone was late once, the probability of them being late again is greater.  Greater risk, lower score.  Simple enough.

What about utilization though using this mindset.  If someone maxes out their cards one month and gets them paid back down, do we think that they are more likely to max out their cards again in the future?  Does this make them more of a risk?  If so, should their scores reflect that?  On the flipside if someone has near zero balances (1-5% utilization) month after month or even for years and then 1 month their utilization significantly spikes, wouldn't most consider them less risky still than someone that has carried 50-70% utilization for the same period of years even if they just paid down their balances to 1-5% overall utilization this month for that single snapshot?

I look forward to the different opinions on this.

[JagerBombs89]

No.

[Anonymous]

Hi BBS!  Always interesting to hear from you.

The sole reason that FICO has historically never considered in its models the history of Amounts Owed (i.e. the amount you owed each month and also the amount you paid each month, stretching back for say the last 30 months) is because the CRAs never collected that data.  As soon as the "amount owed" was updated, that became THE amount owed, and all previous amounts were wiped.

This is why FICO developed this monomanical obsession with customers having preposterously low utilizations (e.g. 1-9%).  The much more valuable metric was whether you were a person who was always paying his CC bills in full each month.  .  But the CRAs gave them no ability to tell whether a person was gradually running up a lot of debt and only paying the minimum amount (highly associated with risk) or whether you were running up 2/3 of the limit each month but always paying it off (associated with far less risk).  Since FICO couldn't tell the difference it had to rely on the less helpful metric of what is your current utilization -- people who had maxxed out their cards would at least be identified that way, even if other far less risky people were caught in the same net.

Now you might be thinking.... wait a minute, I have seen my credit report, and it does show the amount owed on my credit cards as well the amount I paid, not just for August 2016, but also July, June, May, April.... for many many months going back.  Yup.  But that is a comparatively RECENT event.  The CRAs started doing that a few years ago.  But for 40+ years they were not.  The only time they tagged a historical event and kept track of it is if it was a late payment.

Since it is such a recent switch for the CRAs to begin collecting these data, FICOs big models have not begun using it.  They are still stuck in the stone age of knowing nothing about your spending and payment patterns over time (as long as you make at least the minimum payment on time).  Even FICO 9 is not using it, as far as I know, though I expect FICO 10 and Vantage 4 will make a much heavier use of it.

Incidentally, a person who always pays his credit cards in full is called a Transactor.  A person who often does not is called a Revolver.  There's lots of stuff out there about the T-R distinction and how it is becoming very important.  Here is just one article I grabbed at random by googling those two words. 

https://www.creditsesame.com/blog/mortgage/revolver-transactor-why-you-should-worry-if-youre-revolve...

You'll note that just because the T-R distinction may not be part of your current FICO score, that doesn't mean creditors might not be looking carefully at it, using other tools. 

[sarge12]

I do not care which method they use....I always PIF every single card, every single month. I am aware though that some people run a balance, not because they can not pay, especially near the last 1/4 of the year. Many have retirement accounts and such that they may not wish to withdraw until the new year to prevent the tax being due this year. As long as the CRA's let everyone know that historical utilization data was going to affect the score, everyone would know in time to adjust their habits accordingly. The way I do it, it would not matter either way. My utilization is actually under 5%...100% of the time. With about 150,000 dollars of credit limits, 5% would be 7500 dollars. I would never allow this much debt on credit cards. The other factor to consider is this. If I only got credit limits because I thought I may need the credit would owing 5000 dollars if I only had 7000 dollars credit limit, make the 5000 dollars any easier to pay off...no!!!! The amount of debt you can repay largely depends on income to pay that debt, and your low utilization does not affect your ability to pay that debt at all. We get CLI's normally not to run up the debt we owe, but solely to make it easier to maintain a low utilization...it's their game, we just play it by the rules they set forth. I could look at someones revolving debt load plotted on a graph for a 1 year period and tell you exactly who was likely to get in credit trouble. It would be the person who's revolving debt load was going up every month. Why?...because they are spending more than they are making every month. If you really want to stay immune to the effects of AA, CLD's and never really have to worry about utilazation, do as I do, treat your credit cards as if they were charge cards.

[Thomas_Thumb]

---

@sarge12 wrote:

I do not care which method they use....I always PIF every single card, every single month. I am aware though that some people run a balance, not because they can not pay, especially near the last 1/4 of the year. Many have retirement accounts and such that they may not wish to withdraw until the new year to prevent the tax being due this year. As long as the CRA's let everyone know that historical utilization data was going to affect the score, everyone would know in time to adjust their habits accordingly. The way I do it, it would not matter either way. My utilization is actually under 5%...100% of the time. With about 150,000 dollars of credit limits, 5% would be 7500 dollars. I would never allow this much debt on credit cards. The other factor to consider is this. If I only got credit limits because I thought I may need the credit would owing 5000 dollars if I only had 7000 dollars credit limit, make the 5000 dollars any easier to pay off...no!!!! The amount of debt you can repay largely depends on income to pay that debt, and your low utilization does not affect your ability to pay that debt at all. We get CLI's normally not to run up the debt we owe, but solely to make it easier to maintain a low utilization...it's their game, we just play it by the rules they set forth. I could look at someones revolving debt load plotted on a graph for a 1 year period and tell you exactly who was likely to get in credit trouble. It would be the person who's revolving debt load was going up every month. Why?...because they are spending more than they are making every month. If you really want to stay immune to the effects of AA, CLD's and never really have to worry about utilazation, do as I do, treat your credit cards as if they were charge cards.

---

Yes and No.

A key element in addition to PIF is aggregate credit limit relative to natural monthly spend.  Assuming a typical $5000 monthly spend, a person relatively new to credit with $100k yearly income but only $20k aggregate CL would be at 25% utilization. That is well above optimum. To avoid this many play the pre-pay before statement closes game. 

I advocate letting all charges post on statements and then paying statement balances 3 to 5 days before due date. That is the only strategy I have ever used. It's simple and maximizes float time on money (other than carrying a balance on a card during a 0% promo period) However, this PIF approach does require having a sufficient total credit limit - total CL of at least 20x typical monthly spend for good results - meaning keeping Ag UT under 9%. [for reference my Ag CL is about 40x average monthly spend and staying below my benchmark of 6% Ag UT is not difficult].

A reasonable average monthly spend on credit cards relative to income - depends on a lot of factors. Nonetheless, I'd suggest not exceeding 30% for average monthly CC spend relative to average monthly income as a loose guideline.

[sarge12]

---

@Anonymous wrote:

I was thinking about this today and am sort of conflicted about it.  I think very valid arguments could be made on both sides here and I've wrestled with both sides of the arguement myself and am not sure I came away with any strong feeling toward either side in the end.  I'd very much like to hear the opinions of all of you.

As we know, utilization has no memory as we say quite often on this forum.  One can have 100% maxed-out utilization one month and pay it off completely the next month to 0% and upon 0% reporting scores will reflect that exact snapshot in time with zero influence on the maxed-out utilization a month prior.

Obviously the most current information (today) is what is most impactful with respect to credit scores and the older that information gets the less of an influence it has on scores.  With utilization this is very much the case, but I find it odd that previous [recent] history isn't considered at least in part.  Payment history certainly has memory, as if you were late 12 months ago on a payment even with 11 months of on time payment history since, that late payment will continue to hurt your scores.  The reasoning for this of course is that if someone was late once, the probability of them being late again is greater.  Greater risk, lower score.  Simple enough.

What about utilization though using this mindset.  If someone maxes out their cards one month and gets them paid back down, do we think that they are more likely to max out their cards again in the future?  Does this make them more of a risk?  If so, should their scores reflect that?  On the flipside if someone has near zero balances (1-5% utilization) month after month or even for years and then 1 month their utilization significantly spikes, wouldn't most consider them less risky still than someone that has carried 50-70% utilization for the same period of years even if they just paid down their balances to 1-5% overall utilization this month for that single snapshot?

I look forward to the different opinions on this.

---

There is another very good reason that the CRA's will not do this. Due to BK those that get way overextended do not make the banks money. Likewise, due to rewards those that PIF every month are not very profitable either. It is those in the middle, that carry a balance, but not more than their income can handle that are the most profitable. These are the very people who would have a lot lower score, if utilization history were used, or it would encourage them to be in the PIF group that is the less profitable. Utilization counting for only the present will do just as well as history to identify those who are a huge risk of defaulting without seriously discouraging people from carrying a balance.

[Anonymous]

Thanks for the replies everyone.

Outside of credit score hardcore people like us, I doubt many average individuals understand much about credit scores and the criteria that go into them at all.  Most people probably can't even define utilization much less how differences in it can impact scores.  I think we often think from our point of view on this forum (educated with respect to credit), understandably so, but in the big picture we are largely in the minority here.

Credit scores are supposed to accurately predict risk.  I guess my argument here is that looking at utilization through two paper towel tubes (an exact moment) rather than a slightly larger image may not be the best possible representation of risk [with respect to the utilization sector of scoring].  I'm not suggesting that utilization should impact scoring for 7 years like payment history can, 1 year or even 6 months.  It would seem to me, though, that a period of the last several months could be beneficial.  I don't think it would have to be a significant difference; it could be that 2/3 of your utilization calculation is based on the moment and the final 1/3 based on your recent history, say 3 months.  Overall this wouldn't significantly impact ones scores either positively or negatively IMO.

I just struggle with believing that someone that carries zero balances or minimal utilization for months or years that spikes their utilization once should be viewed as being equally as risky [with respect to utilization scoring] than someone that's had extremely high or near maxed out utilization over the course of the same period of time.  Conversely, someone that's had extremely high utilization for years that suddenly pays it all off being able to achieve the same score immediately [again with respect to only utilization here] as someone that's proven for years that they can keep their utilization low or at 0 just doesn't seem to make perfect sense when scores are supposed to accurately predict risk.

[EW800]

> Outside of credit score hardcore people like us, I doubt many average individuals understand much about credit scores and the criteria that go into them at all.

I absolutely agree!  Given my somewhat of an obsession with credit scoring, like most others here, at times I have discussed credit reporting and scoring with family and friends if somehow the topic comes up.  Many were certain, and still believe, that there is only ONE credit score and such.  They look at me like I am crazy when I tell them there are dozens of scores, that can vary significantly.  I've been looked at like I am even crazier when I have mentioned auto-enhanced, bank card enhanced and so on.   

I believe I see your point in regard to utilization not having any memory.  I do see it both ways.  In my case, I have not had a balance on any revolving account, other than about $5 on one account, for about the last 36 months.  I have often wondered if there should be just a tad bit of credit given for going so long with such low utilization, as I would think this could be an indicator that it is unlikely that I am going to jump into a ton of debt.  But, I see the point of just looking at the current Utilization as well. 

Thanks for bringing up the great topic!

[Anonymous]

Hi BBS.  I think you are asking a great question.  What I was trying to say in response is that future scoring models will almost certainly analyze the "amounts owed" on credit cards in a far richer, sophisticated and reliable fashion than the current comparatively clumsy technique of looking only at your profile as a single snapshot in time (the current instant) with whatever your balances happen to be.

You are right to suspect that analyzing amounts owed over a long time is more reliable -- particularly if it enables the scoring algorithm to detect a tendency to run up balances and then pay only small portions of them.  This is critical -- what FICO wants is to be able to compare not only "utilizations" over time but also in conjunction with that payment patterns over time.

And the reason they have had to rely on the relatively clumsy method of analyzing only at any instant is because for decades the CRA data did not permit you to do anything more.  The data of what you owed each month (and how much you paid each month) was simply not available.  (These data are called trended data.)

That's changed in the last few years.  So we'll be seeing a lot more of scoring systems using trended data analysis of credit cards.  Already Fannie Mae has built this in to their Destop Underwriter software, which if I read the news stories correctly has already been switched on -- so anyone who thinks they might be getting a mortgage (even as soon as late this year) should be looking at their profile and trying as much as they can to adopt Transactor-like behavior as opposed to that of a Revolver.

[iv]

---

@Anonymous wrote:

I just struggle with believing that someone that carries zero balances or minimal utilization for months or years that spikes their utilization once should be viewed as being equally as risky [with respect to utilization scoring] than someone that's had extremely high or near maxed out utilization over the course of the same period of time.  Conversely, someone that's had extremely high utilization for years that suddenly pays it all off being able to achieve the same score immediately [again with respect to only utilization here] as someone that's proven for years that they can keep their utilization low or at 0 just doesn't seem to make perfect sense when scores are supposed to accurately predict risk.

---

Can't say I'd agree on that... sudden PIF->Carrying is higher-risk than existing long-term carrying!

Yeah, someone who has had high utilization for years isn't the best credit risk... but years of high util (with no missed payments) is stable. (And making lenders plenty of interest income.)  And when they pay it off or down? That's just a straight improvement.

But someone who has never carried a balance before who suddenly has a large increase in util? Yikes.  Yes, could be a large one-time purchase that will be PIF, or could be taking advantage of a 0% offer... (and with most folks around here, that's exactly what it would be).  But - it could also be a bustout, or just a sign of sudden financial distress.  Either way, it's not stable.  And unstable == risk.

Really, someone who was able to PIF for years suddenly carrying balances is a big red flag for the average consumer account...

You know what I'd personally consider a good scoring metric? Amount of interest paid in the past two years, as a percentage of average balances.  It's not reported by lenders, and it won't be, nor would FICO ever include it in scoring (given that non-delinquent accounts that pay interest are very profitable).  But I think that would be a much better measure of financial responsibility... but not a better risk/profit measurement.