FICO scores have become so unreliable

[Peter1142]

There are literally dozens of different scores. I have an 800 auto score, 750 area credit card score according to some models, I go apply for a mortgage and my score is 695 and I can't get a good rate. Meanwhile that very same day the score from Experian is 754.

Why are mortgage scores so low? I have had a mortgage 5 1/2 years and haven't missed a single payment. I remember this issue when I first applied, too.

It really makes no sense I can have an 800 auto score and a 695 mortgage score... I don't have any missed payments on either. To me it really proves these scores are not as science based as they claim.

[Anonymous]

You answered your own question in identifying some of the many different FICO scoring models and thus different scores generated.

Those scores aren't "unreliable" if you're always looking at the same score and therefore comparing apples to apples.  If you're going for a mortgage and only focus on your mortgage scores, ignoring all of the rest (as they're irrelevant to your upcoming app) you'll find the mortgage scores to be perfectly reliable and consistent.  If you start comparing apples to oranges and look at a mortgage score verses an auto score, sure you can see a 100 point or higher variance between the two.  They're two different algorithms that handle the data in different ways, but more/less weight on different factors, etc.

The only FICO score(s) that ever matter at any given point in time is/are the ones that a potential lender with which you may be seeking credit uses.

[Trudy]

2 factors that can impact MTG scores more than other scores is # of accts reporting a balance and new accounts.  There's more to it but I see my MTG score fluctuate quite a bit when these factors change.  Not everyone but for me they have always been my lowest scores.  

As BBS states, each model and model version has different algorithms/formulas to maximize the score.

[Anonymous]

Peter1142 understand your questioning the FICO Scores you are finding. Your answers are already included from the other posters. Will add something for you, my scores as of this post:

850 (300-850) FICO 08
810 (300-850) FICO MORTGAGE 2
882 (250-900) FICO AUTO 08
885 (250-900) FICO BANKCARD 8
807 (334-818) BEACON 5 SCORE (Equifax)

Now some of my monthly credit cards report FICO or VANTAGE Scores. The scores provided by the credit card lenders match depending on which score they are utilizing and reporting to me.

As to the mortgage score which is low, makes me wonder? I deliberately owe a few dollars on one credit card so as to NOT lower my regular FICO Score. But with the Mortgage Score the lenders are seeing a mortgage and auto loan that have a utlization of greater than 50% (why would I take out a loan to get 30% utilization to keep a FICO Score happy?). Anyways, I do understand your questioning and can understand your frustrations.

To really peak your interest, I went into my local brick and motar bank of thirteen years to obtain a Personal Loan for some Home Improvements (Roof, Gutters and some doors and windows). Looking for $10,000 and I had the cash but, did not want to pay it all at once (the bank had really good rates) and liked my Beacon 5 Score. Well anyway, the bank said if most of their customers had my scores they would make more loans and then, turned me down (too much other available credit). Went to another local brick and motar bank I have used for seven years and the money was in my checking account within two minutes. Paid the loan off in 45 days.

Moral of the story ... "go figure" ...?

[Anonymous]

---

@Peter1142 wrote:

Why are mortgage scores so low? I have had a mortgage 5 1/2 years and haven't missed a single payment. I remember this issue when I first applied, too.

---

Hi Peter.  It's not that mortgage scores (for all people) are somehow always low.  The mortgage scores of some people are higher than their other scores.

Rather, the better question is: "why are my mortgages scores so much lower?"  The folks here can help you understand that and better still help you take steps to fix that.  They'll need to know a lot more about your profile.

Another thing to bear in mind is that what are usually called, as a kind of shorthand, the FICO "mortgage scores", were never created by FICO as models specifically for the mortgage industry.  This is in contrast to (say) FICO 8 Auto which was indeed tailor-made specifically for the auto industry.  The "mortgage scores" are simply very old models that entities called Fannie Mae and Freddie Mac settled on a long time ago and which all mortgage lenders are basically forced to use by FM.  Thus the old "mortgage models" do not weight specifically your history at paying off mortgages any more than they do your history at paying off auto or personal loans (say).

Are you looking to refinance your existing mortgage?  Planning to buy another house?  As a practical matter, we can help you best if we know ,more details about why you'd be concerned about your mortgage scores.  We also may be able to save you money if you are paying $30 or $40 per month for a credit monitoring service you may not need.

[Peter1142]

I gave up on the refinance, it is very expensive anyway versus the savings and I was already on the fence. I just want to be rid of the FHA mortgage insurance I am paying a huge rate for and can't get rid of any other way. But in the end it doesn't make sense to spend $7000 to get rid of it and draw out the mortgage with a higher rate too (I got 3.75% on my original loan.)

Like I said, I know there are dozens of different scores with different models. My point was that there is no justifiable reason why someone would be an 800 with regards to an auto loan and a 695 with regards to a mortgage, I believe there is nothing in a credit report that can account for such variability in creditworthiness on two installment loans with the same credit file. And to me, the fact that these numbers can be so all over the place depending on the model proves that these scores contain a lot of hocus pocus. That is all I was trying to say.

I do have a lot of accounts reporting with a tiny balance and this is perhaps why the model used for mortgages doesn't like me.

[Anonymous]

---

@Peter1142 wrote:

Like I said, I know there are dozens of different scores with different models. My point was that there is no justifiable reason why someone would be an 800 with regards to an auto loan and a 695 with regards to a mortgage, I believe there is nothing in a credit report that can account for such variability in creditworthiness on two installment loans with the same credit file. And to me, the fact that these numbers can be so all over the place depending on the model proves that these scores contain a lot of hocus pocus. That is all I was trying to say.

I do have a lot of accounts reporting with a tiny balance and this is perhaps why the model used for mortgages doesn't like me.

---

Sure there is.  They are completely different scoring models.  That's like saying that you know a Corvette ZR1 can do 0-60 in 2.9 seconds and that Prius does 0-60 in 9.7 seconds, but since they're both "cars" you think the Prius should do it in at least 7 seconds.  It would be an apples to oranges comparison, the same way comparing a mortgage model to an auto model would be.

You said that you do have a lot of accounts reporting [tiny] balances.  If they're tiny, you can bring them to $0 without an issue, correct?  Why don't you try bringing all but one down to $0 balances and see what happens with your mortgage scores.  Depending on your profile, you may see your 690 scores go to 720-730, basically putting you right there where you'd receive top-tier rates.

[Peter1142]

---

@Anonymous wrote:

---

---

Sure there is.  They are completely different scoring models.  That's like saying that you know a Corvette ZR1 can do 0-60 in 2.9 seconds and that Prius does 0-60 in 9.7 seconds, but since they're both "cars" you think the Prius should do it in at least 7 seconds.  It would be an apples to oranges comparison, the same way comparing a mortgage model to an auto model would be.

This is a completely false analogy. You are using two different cars, but there is only one borrower with one creditworthiness. I would be a single car in your analogy, getting two different readings from two different instruments. Obviously in my example one is measuring auto creditworthiness and one is not, but there is no reason that makes sense for them to be wildly different. No one is a top tier auto risk and a lukewarm mortgage risk.

Thanks for the advice, regardless.

[Anonymous]

---

@Peter1142 wrote:

I do have a lot of accounts reporting with a tiny balance and this is perhaps why the model used for mortgages doesn't like me.

---

That certainly is part of why the mortgage scores are lower.  Those old models are much more sensitive to the number of open accounts with a balance than is FICO 8 Auto.

Another part of the explanation is that FICO 8 Auto has a maximum ceiling of 900, whereas the mortgage scores have a much lower ceiling (e.g. the old EQ model used by the mortgage industry has a maximum score of 818).  Your Auto score is closer to the top than is your mortgage score, but not that much closer.

Another easy explanation would be if you had a completely clean report except for a collection that was under $99.  The old models are very punitive about all derogs but the much later FICO 8 model was redesigned to ignore small collections, which you might get for a parking ticket or a library book you forgot to turn in.

Keeping in mind the age of the different models is important.  If you were used to a 2018 car driving a certain way (and with features on the inside as well) you might be frustrated and baffled when you got in your neighbor's car -- until you found out that it was a vintage card from the 1960s.

It sucks perhaps that FM forces the mortgage industry to use very old models (the EX model was released in 1998!) when newer ones are available, but that's outside of FICO's control.  In fact, a few years after FICO 8 was released, FICO released a new version of FICO 8 designed especially for the mortgage industry.  But Fannie Mae wouldn't let the industry use it. 

[Anonymous]

---

@Peter1142 wrote:

---

I would be a single car in your analogy, getting two different readings from two different instruments.

---

That's fine to think of it as a single car, but the instruments are weighing data differently and measuring two different things.  Think of it like measuring torque vs horsepower.