cancel
Showing results for 
Search instead for 
Did you mean: 

Wild difference between TU and EX on same data

tag
Anonymous
Not applicable

Wild difference between TU and EX on same data

I just purchased my TU and EX scores from myFICO. TU is 742, EX is 808. That is a huge difference, and I don't understand it at all.

 

I have gone through both reports and I don't see any difference in the data. How in the world is it possible for two scores from the same data, using (we are told) the same formula to be totally different?

 

The TU score says my major negatives are short credit history (9 years 5 months oldest, 6 years average), and a high non-mortgage balance (~$3,600). The EX score says length of credit history and non-mortgage balance are positives, but both are exactly the same as TU!

 

FWIW, 808 seems about right, 742 seems far too low.

 

Since we are considering refinancing our mortgage, this is very worrying to us.

 

It's really hard to believe that FICO is a competent or reputable company when we see weird stuff like this...

Message 1 of 7
6 REPLIES 6
Macroman
Regular Contributor

Re: Wild difference between TU and EX on same data

The lever that I can see is paying down those balances. It sounds as if TU considers them revolving and EQ installment.

 

EQ and TU do not use the same formula. There are others here more knowledgeable than I but I do know that each CRA has their own scoring algorithms.

Message 2 of 7
Anonymous
Not applicable

Re: Wild difference between TU and EX on same data

There are a number of Fico scoring models, and there are slight variances for versions that are created to utilize the data from each different credit reporting agency. In other words, an EQ score based on a 2004 formula might vary slightly from a TU score based on a 2004 formula. This is because each CRA organizes their database differently and the scoring models try to take that into consideration.

 

In this case, however, the difference is most likely due to the fact that the EQ score sold by MyFico and Equifax uses a 2004-2005 era scoring model, while the TU score sold by MyFico uses a 1998 scoring model. If your lender pulls a TU'98 score, he will probably see similar scores to what you saw. If he pulls a TU'04 score (which is more likely), it's quite possible that it might be closer to 800.

 

Keep in mind that the mortgage lender will base approval and rates on your middle score, so even if the TU score is significantly lower than the other two, it won't affect your approval or rate.

 

When you do compare your own scores, or when you compare Fico scores you purchase with those your lender pulls, be very careful to compare data line-by-line. The same account can be reported to one CRA as Open, and to another as Closed. Or it can be reported to one CRA as Revolving and to another as Installment. HELOC accounts can be scored differently by TU than by the other two CRAs. Reports pulled on different days can return different scores. Etc., etc., etc.

Message 3 of 7
Anonymous
Not applicable

Re: Wild difference between TU and EX on same data

I know the different agencies have their own unique scoring systems, but I thought the FICO scores we can buy on myFICO are meant to be calculated the exact same way for both TU and EX?

 

The annoying thing about those revolving balances the TU report is grumbling about is that we always pay all balances in full every month. Households on higher incomes can easily run several thousand $$ per month on credit cards; as a percentage of household income, that's not meaningful. It seems really stupid for FICO to treat those as strong negatives just because they would be a red flag for a penniless student. I know the solution is to pay the balances off earlier in the month (on the statement date, maybe?), but that way we are wasting the grace period (= interest I could be earning elsewhere) just because FICO's algorithm is weak in this area.

 

Anyhow, even if I simulate paying the balance down to 0 using myFICO's simulator tool, I still don't get anywhere near my EX score, presumably because of the age of accounts negative.

 

And I'm still wondering how the exact same oldest account and average account ages can be considered good in EX and bad in TU.

 

I contacted myFICO support; so far all I have had is two rounds of irrelevant standard responses (probably machine-generated from keywords in my questions); they haven't come back to my latest request, as yet, so maybe my issue has finally reached a human being.

 

It's a little scary to know that people's whole lives are controlled by something that appears to be so arbitrary and unaudited. But that's what we get for having a legislature that is bought and paid for by big corporations...

Message 4 of 7
Anonymous
Not applicable

Re: Wild difference between TU and EX on same data

In case you are wondering why I appear to have ignored Revike's very helpful response, it's because that hadn't arrived when I started writing my follow-up post above.

 

To me, Revike's response just emphasises how random and arbitrary this stuff can be.

Message 5 of 7
Anonymous
Not applicable

Re: Wild difference between TU and EX on same data

One other point I like to make is that the "positive factors" and "negative factors" listed on credit reports very often confuse more than explain. This is especially true when scores start approaching the high 700s. The report format will try to extract X number of positive and X number of negative factors, even when they don't exist. So it's very common to see "factors" that contradict each other.

 

Most positive and negative factors are either blindingly obvious ("your 1% utilization is a positive factor" Duh!) or appear to be total nonsense. There are thousands of posts here and at CB from people who are freaking out because of nonsensical factors listed on their reports. If a factor doesn't seem to make sense, just ignore it.

 

In the vast majority of cases, unusual score differences are caused by errors in reporting by the creditor. In your case, like I said, the difference is probably due to the 6-year difference in scoring formulas.  Most of us have complained for years about the TU'04 score not being available to the public, and you will find dozens of posts by people who bought their TU'98 score from MyFico and were unpleasantly surprised when their lender happened to pull a TU'04 score and it was lower. Part of the blame for the continuing sale of the 1998 TU score lies with TransUnion.

 

In coming years, mortgage scores based on 2008-2009 scoring models will start to be used by lenders, so there will likely be even more discrepancies ...

Message 6 of 7
RobertEG
Legendary Contributor

Re: Wild difference between TU and EX on same data

FICO does not actually calculate any scores.  They cant.  They dont hold the immediate daily data necessary to do that.

FICO licenses a scoring algorithm to a CRA,  The CRA only gets their own licensed version from Fair Isaac.

When you request a FICO score, it is generated by the CRA by running their data through their licensed FICO algorithm.  They are licensed to call it a FICO score based on their contract with Fair Isaac, and the flavor of scoring algorithm they choose to use.  There is no plain vanilla.

However, with all that being said, Fair Isaac stated in one of its Webinars a few years ago that variances in scoring that exceed approx 30-pts between the CRA algorithms is unusual, and should raise some concern.  Usually, it is due to some data difference in your consumer files with the different CRAs.

The commercial credit reports you get here, and elsewhere, are not a  full report of your credit file.  I suspect there is a difference in your credit files with each of the CRAs.

My sugestion is to obtain the best credit report you can, which, in my opinion, is the annual, free credit report you are entited to at annualcreditreport.com. 

If you have a full and current CR from each of them, and they show identical data, I would be surprised.

 

 

Message 7 of 7
Advertiser Disclosure: The offers that appear on this site are from third party advertisers from whom FICO receives compensation.