TIME: Understanding its relevance to FICO scoring.

[Anonymous]

I'm not sure what got me thinking about this today.  I suppose all the threads/discussions on Trended Data and how looking at data over a period of time is far more meaningful.  The "big picture" to me carries far more weight than what is seen through two paper towel tubes over your eyes.

We often hear about "TIME" being the one factor with respect to FICO scoring that can't be altered.  You can't slow it down.  You can't speed it up.  People often ask questions like what they can do to fix their credit and while sound credit behavior is paramount, usually the answer is coupled with the fact that it needs to be given time.

While time is definitely a major factor in profile recovery when considering a "bad" profile, it's definitely not a significant factor when considering a "good" profile.  I mean it is with respect to AAoA, but that's a relatively insignificant portion of FICO scoring.  This is where I feel Trended Data in future algorithm versions like FICO 10 will play a big role.  And, I'm not even sure such a version would even take it as far as what I'm thinking.  Perhaps I'm considering more FICO 12 type stuff.  Here are a couple of examples of what I'm thinking.

Take a conventional 30 year mortgage.  If someone pays back a 30 year mortgage according to the terms of the loan (paid as agreed) and carries that loan for the full 30 years, that's quite significant IMO.  30 years for many is probably half of their credit lifespan when it's all said and done.  A lot can happen in 30 years.  Career changes, job loss, medical issues, divorce, loss of loved ones, etc.  This list is endless of things that could have prevented one from missing a payment at some point during 30 years.  If someone does go those entire 30 years though without ever being late on a payment, to me, that's an extremely strong indicator of creditworthiness.  I guess my point here is that longevity while not considered now in FICO scoring, probably should be in some future model.  A mortgage paid back over 30 years to me is far more significant of an event than paying back a 3-7 year auto loan, yet both "installment loans" are considered equally under FICO scoring.  I get it that when considering underwriting and the "human" factor when talking a manual review that these things can be considered, but it would make sense to me to have FICO scoring account for something like this in some sort of way.

Another thing, let's talk longevity the other way around instead of positive longevity, the negative side of it.  Say you have two otherwise equal profiles.  On one, someone compiles 10 late payments across 4 different accounts in a span of 3 months with the final month being December 2014.  On either side of these 3 months this person possesses flawless payment history.  On the other hand, you have a different person that also compiles 10 late payments across 4 accounts with the final month being December 2014, but those 10 late payments came scattered throughout a 2 year period.  If you take some later point in time, say mid 2016, these two otherwise equal profiles will probably have about the same credit score.  Same amount of negative information, same amount of time since their last infraction.  When considering this data however, IMO the person that exhibited poor credit behavior for the longer period of time (2 years) should be viewed as a greater risk and scored accordingly.  The individual that had the 3 month collapse likely encountered one of those life-changing events that caused them to abruptly miss lots of payments.  To their credit though they recovered quickly where the other person exhibited poor credit behavior for 2 years.  Longevity being considered in these examples, I'd consider the person with poor payment history for 2 years to be somewhat more of a risk as it represents a longer pattern of behavior.

I'm sure everyone on here can think of other "time" type examples.  I know that Trended Data is supposed to be considered for periods of time of say 12, 18, 24 or even 30 months with respect to things like utilization and whether or not people pay in full or "carry" balances.  I'd challenge these numbers and suggest that time should be considered even more.  Someone that has adopted a PIF lifestyle for 24 months is great, but if someone else with an otherwise equal profile has proven to exhibit this behavior for 72 months, should they not be considered less of a risk (and thus scored accordingly)?  I'm sure I'm jumping the gun a bit here since obviously we haven't even gotten to the point where Trended Data is really even considered yet under the current models, but I think these things are interesting to think about regardless.

I'd like to hear the opinions of everyone else regarding TIME, its relevance to FICO scoring and whether you feel longevity (positive or negative) should ultimately be considered in future FICO models.  Thank you all for your insight and contributions, as always.

[NRB525]

Thoughtful post, well written.

I will propose that many of the nuances you are looking for may already be in the algorithms.

Regarding the long string of lates, that is going to have an effect of aging the earliest, and those would begin to fall off at some point. If someone goes through a bad patch, with two months of a number of lates, well, that's indicative of something causing them financial stress. In the measures of FICO, the first of anything is the worst, then the impact of any others of that kind have less of an impact. So the two-month bad patch is somewhat limited in its impact. The longer running set of lates is also, to a certain extent, mitigated, but if there has been time since the last late, is it better in a risk model to think the lates are something in the past, less of a risk? I think that is an appropriate approach.

Regarding the 30-year mortgage, if someone is in one house for 30 years and pays that down to zero, kudos. But then, what sort of new mortgage or borrowing are they looking for now? The purpose of the FICO scores is to have some insight into future borrowing. Someone who stuck with one loan that long, most likely not at an age they will be borrowing much of anything any more. So yes, there is a positive attibute to it, but no real interest by the lenders in this person as a future customer.

Length of activity is critically important also, and is in the models already. This is why new files cannot obtain 850, most likely cannot obtain 800 for a few years. It takes time of payments to buildup that feature of the file, to build up that trust background.

[Revelate]

Unfortunately I can't give a more thoughtful response right now, but I don't know how we escape it.  The fact is the more data that is there, the improved resolution of the analysis... that's true for every credit file, not intrinsictly tied to clean or dirty bucket.

Also I can't see a world where someone with 20 years of positive history is viewed the same as someone with 2 years even if their scores are within spitting distance for whatever reason from a lender perspective.  FICO will never be the end all and be all of an underwriting decision.

[Anonymous]

There is something that is lost here...

Human brains can compute data not only faster than a computer can, but more efficiently. When you have a team of data scientists and software engineers together... well, that's much more computational power than you can imagine.

When we write software, it is imperfect. There are things we can do really well now, even with algorithms, but still not as well as the brain. Information in a computer currently only exists in bytes, in which that byte can be one of two values - 1 or 0. The more substantial the data, the greater quantity of 1's and 0's there are to process. With increased complexity comes increased time, and believe it or not, there are some processes that require such a substantial amount of data churning that it can exceed what our computers are even capable of processing. 

We know that newer algorithms will be even more robust than what we've got now, but they still won't be perfect. 

A set of numbers that are relatively similar, versus a complex web of data points  and structures... it's just an entirely different ballgame. 

[sarge12]

BBS...we are digging deep here, but IMO the most important factor that contributes to good credit is something laws will not allow fico to consider...AGE!! When I was 25, I was living payday to payday with no savings. As I aged, saving became a higher priority than dating and partying, and eventually things were paid off, and financial stability ensued. Now at 58, I have half a million in retirement savings and less than 40000 in debt. Being able to draw any amount of this 500k out I need to...what are my chances of default on debt? Almost 0..nada...none! As we age, most naturally accumulate more wealth. How the credit score can not consider savings or age is just crazy. Since my income, even without drawing any money from savings is at least 400 dollars more than my regular debts and expenses, it is super easy to pay my bills, I do not even have to try hard. People who have not had time to build this wealth are naturally going to have a lot harder time with a budget...even if they are responsible.

[Anonymous]

---

@NRB525 wrote:

If someone goes through a bad patch, with two months of a number of lates, well, that's indicative of something causing them financial stress.

---

Interesting statement and while it slightly veers off topic here, I have a point to make.  While in most cases it's indicative of something causing financial stress, it isn't always.  Sometimes it can just be simple life "stress" that has nothing to do with anything financial.  I had a friend that lost his wife unexpectedly.  For several months he was a shut in and handled the weight of the situation internally.  I can assure you that during these few months that paying his bills on time wasn't something he was capable of thinking about.  He was late on a lot of accounts because he was an emotional wreck.  Within 3-4 months all of his accounts were current once he overcame the initial shock of the situation and he's been fine credit wise since.  Financially he's very well off and at no time was he ever late on his accounts due to financial reasons.  I know in the eyes of FICO the "reasons" don't matter, but this situation is a perfect example of why someone may have a few months of negative information on their credit report that IMO shouldn't be scored as harshly as someone else with an equal amount of negative information that spans 2 years or a long enough period of time that's indicative of behavior rather than circumstance.

[Anonymous]

Really nice post, BBS.

One reason that all scoring models are limited in terms of how they can analyze time is because of the 7-year rule imposed by federal legislation on negative items in one's report.  I'll recap your biggest example (the 30-year mortgage) below:

Take a conventional 30 year mortgage.  If someone pays back a 30 year mortgage according to the terms of the loan (paid as agreed) and carries that loan for the full 30 years, that's quite significant IMO.  30 years for many is probably half of their credit lifespan when it's all said and done.  A lot can happen in 30 years.  Career changes, job loss, medical issues, divorce, loss of loved ones, etc.  This list is endless of things that could have prevented one from missing a payment at some point during 30 years.  If someone does go those entire 30 years though without ever being late on a payment, to me, that's an extremely strong indicator of creditworthiness. 

But the way that the 7-year rule works, a person could have had many 60-day and even 90-day lates scattered throughout the first 20 years of that mortgage.  If he then got his act together and had seven years of perfect payment history (years 21-27) it would then look (at the beginning of year 28) as though he had never missed a payment.  Of course, the fact that he's not missed a payment for the last 7 years is itself important, and no doubt predictive.  It's just that this is the furthest back in time any credit scoring program (or manual reviewer) can look.  The FCRA prevents the 30 year backward glance you suggest.

I do feel like you are htting on something important though, which is that the FICO 8 algorithm seems to be pretty weak on time.  It allows you to develop a really high score with very little time invested.  I think people have shown here that it is possible to break 800 in maybe a year, certainly two.  Very common is for a person with a six month total history to start with scores around 740 (e.g. our friend SubExistence quite recently).  

And that's not counting the preposterous tricks that can be played with AU accounts.  (In FICO's defense, it did attempt to completely abolish any scoring benefit from AU accounts when FICO 8 was gearing up for release, but FICO was defeated by a Housewives' Rebellion.  Seriously.  FICO was forced to back off because so many spouses lacked any credit cards of their own.  Thus we are stuck today with the ludicrous capacity of a dad to give his 18-year old daughter an instant 30 year credit history.)

My feeling is that a few years is simply not enough time to have enabled you to screw up -- for precisely the reasons you give.  Certainly six months is not!  And hence having a great score in a tiny length of time seems silly.  My view is that the Payment History category ought to be more tightly linked to Age of Oldest Account and Total Number of Accounts.  FICO 8 does that to so some extent with scorecard assignment (AoOA and Number of Accounts are two of the three scorecard assignment factors for clean accounts) but clearly IMO not enough.  There should be a world of difference between a person who's never missed a payment in six years vs. a person who's never missed a payment in six months.

[Anonymous]

---

@sarge12 wrote:

BBS...we are digging deep here, but IMO the most important factor that contributes to good credit is something laws will not allow fico to consider...AGE!! When I was 25, I was living payday to payday with no savings. As I aged, saving became a higher priority than dating and partying, and eventually things were paid off, and financial stability ensued. Now at 58, I have half a million in retirement savings and less than 40000 in debt. Being able to draw any amount of this 500k out I need to...what are my chances of default on debt? Almost 0..nada...none! As we age, most naturally accumulate more wealth. How the credit score can not consider savings or age is just crazy. Since my income, even without drawing any money from savings is at least 400 dollars more than my regular debts and expenses, it is super easy to pay my bills, I do not even have to try hard. People who have not had time to build this wealth are naturally going to have a lot harder time with a budget...even if they are responsible.

---

Very good post and great point about savings.  If savings could somehow be factored into FICO scoring I think that would be excellent and certainly a great indicator of whether or not someone would be able to repay a debt and thus a factor to evaluate risk.

I think most people have savings many many times less than their total credit limits.  However, if you have someone with savings many many times more than their total credit limits, I think between these two otherwise equal people it's a fair judgement call to state which one is in a better financial position to repay their debts.

[sarge12]

BBS...You may be right about the amount of savings...I mean, I did not exactly consciously save all that. I worked at a tire plant for 35 years, and 300k+ of that money was a lump sum payout of my company paid retirement which I could have chosen as a lifetime annuity. When I became disabled, the lump sum + 200k of 401k money became immediately available. I faced 650 dollars a month house payment and 300 a month for my goldwing, and I started recieving 2200 a month in SSDI benefits. At no time in life have I had so much disposible income, but due to poor health, never has money been so little needed. I could easily live on the SSDI income alone.

[Anonymous]

---

@sarge12 wrote:

BBS...we are digging deep here, but IMO the most important factor that contributes to good credit is something laws will not allow fico to consider...AGE!! When I was 25, I was living payday to payday with no savings. As I aged, saving became a higher priority than dating and partying, and eventually things were paid off, and financial stability ensued. Now at 58, I have half a million in retirement savings and less than 40000 in debt. Being able to draw any amount of this 500k out I need to...what are my chances of default on debt? Almost 0..nada...none! As we age, most naturally accumulate more wealth. How the credit score can not consider savings or age is just crazy. Since my income, even without drawing any money from savings is at least 400 dollars more than my regular debts and expenses, it is super easy to pay my bills, I do not even have to try hard. People who have not had time to build this wealth are naturally going to have a lot harder time with a budget...even if they are responsible.

---

I think all the laws govern is what data can be listed on one's credit report.  A consumer's age can't be listed.  But as far as I can tell, that doesn't prevent FICO or Vantage from using other report data as a proxy for age.  The particular field that comes to mind is Age of Oldest Account.  It's not a perfect method, since a person can be old while having all young accounts.  But if he does have an old account it means he can;t be young.  So, like any proxy, it's imperfect, but it does to some extent work.

And we know that FICO gives you a scoring benefit for having an Age of Oldest far beyind what it realistically can show in terms of credit history.  Realistically all FICO cam really see is your last seven years -- beyond that is a gray zone where you could have had millions of derogs and FICO wouldn;t be able to tell.  Nonetheless, FICO still gives people a scoring advantage for having far older accounts than that, which I suspect is their way of using Age of Oldest Account as a proxy for a consumer's physical age.