FICO 2008: Better, But Not Good Enough

[Anonymous]

The problem with making FICO scoring take into account more variables is that, the more variables any index accounts for, the less accurately it reflects each one, and the more likelihood that a variable is included that doesn't actually figure into what the index is measuring.

And to have FICO factor income would introduce a whole slew of complexities...not the least of which is how income would be known and verified by FI. You've got employment income, you';ve got 1099 income, gifts. Major privacy issues involved in FI knowing the income of everyone in the country.

[FretlessMayhem]

Regarding income, people in areas with a higher cost of living would have better credit scores.

So, if I really wanted to boost my score up 50 points, I could just take a job in New York City or Washington DC.

[KingAdrock]

---

700Man wrote: (1) NO CONSIDERATION OF FINANCIAL ASSETS: I understand FICO is a credit-scoring system and not supposed to be an asset-scoring system. Still, if someone has sufficient financial assets -- bank CD's, a brokerage account, mutual funds, etc -- chances are that not only can those assets be used during financially difficult times to make payments, but it also indicates someone more financially responsible. Ditto those with 401(k)'s, IRA's, etc. If 2 people both earn $75,000 and have identical credit histories, but one has $50,000 in bank CD's and mutual funds worth $125,000, which one should be ranked higher on the FICO system ??

How do you suggest FICO do this? Should FICO have 24/7 access to your bank accounts, brokerage accounts and IRAs so they can know the current balances and change your score accordingly? I don't know about you but I would NEVER want FICO to have that. Lenders take your income into consideration. Not FICO. IMHO that's the way it should be.

(4) BORROWING CAPACITY: The focus on percentage of card limits being used is misplaced. You can shift balances around (as I have done) and if anything, it should be TOTAL CARD LIMITS relative to TOTAL CURRENT DEBT. If I have total card limits of $50,000 over 5 cards, but 3 are totally empty and 2 are maxed out to $10,000 each, what's the difference if I had 5 cards each with $4,000 on them ???

This IS considered already. FICO considers the utilization % both on each card individually AND your total available credit.

(5) JOB HISTORY: The credit bureau's don't really seem to have good information on job histories, especially those with lots of job changes (something I am familiar with). I'm currently with a very large company which should count for something more than if I was with 'Joe's Auto Repair' or something like that. If you have multiple jobs but not big unemployment gaps, that should also be a plus as should long periods of continuous employment with a single employer.

1. Stable employment IS a factor to credit risk, that's why lenders ask for job information when you apply for credit. Why should FICO?

2. Working at a "big" company is no proof you'll always have the job. You can always be downsized. Or even be working at the next Enron.

(6) FIXED PAYMENT HISTORY: Previous payments of fixed obligations like an auto loan, mortgage, or other installment credit don't seem to be ranked that highly. If I made dozens of payments in a row on a house or auto loan for thousands of dollars each month total, shouldn't that count more than if I missed a payment or was late on a $750 balance with a minimum payment of $35 on some dumb credit card ??

If you make "dozens" of payments (meaning years of monthly payments) on time, your score WILL go up dramatically. Likely more than one 30 day late would drop you.

(7) INCOME VERIFICATION: If annual history can't be retrieved by FICO, then a range should be established. You should have the chance to submit Federal 1040's establishing income to boost your score if FICO is unable to ascertain income levels or ranges. Would you submit your most recent 5 1040's with proof of income if it raised your score 50-75 points off the bat ??? The FICO scores were easily rigged and manipulated in the last few years by people who wanted to qualify for huge mortgages. If a few of my suggestions above were in place, fraudulent buyers would have been scored alot lower and never gotten the mortgages in the first place.

Why do you WANT FICO to know your income? The lenders ask for it, mortgage companies make you prove it, why should FICO know it at all?

Blaming the market crash on "rigged FICOs" is absurd. If that were true, wouldn't we hear in the news about 800+ FICO scorers defaulting all over the place?

---

[Anonymous]

STEVIE 2-- yes i know that's his point.  duh.  My point is to 700man.   Pretty snide of you. 

[marty56]

FICO is the salary independant portion of the descion making process of whether a person should be extended the credit or not.  It is the the risk portion in terms of repayment not the ability to make the payments.  Income and DTI would be the final piece of the puzzle to help the lender make a reasobale guess that this would be a good loan.

Income is a factor on how much to loan a DTI can serve as both a risk factor and amount factor.

Mortage companies are interested in DTI and it is quite possible to have a high FICO score and a high DTI which would prevent you from getting a mortgage becuase your DTI is to high.

Message Edited by marty56 on 04-29-2008 06:03 AM

[Anonymous]

I guarantee that FI has better evidence regarding the predictiveness of any particular variable than you or I do.

Well, the FICO scores have been shown to have been as reliable as the ratings from Moody's, S&P, etc, so I think I won this round. 

The FICO scores suffer from the same 'grade inflation' as Harvard:  too many people with 700+ scores and the probability of default at that level (5%) was clearly too LOW entering this financial tsunami. 

And yes, as someone who looks over the finances of people I can say that for equal FICO scores, if someone has financial assets they are MORE likely to keep paying even if their debt/income goes beyong 28/36.

[haulingthescoreup]

---

700Man wrote:

I guarantee that FI has better evidence regarding the predictiveness of any particular variable than you or I do.

Well, the FICO scores have been shown to have been as reliable as the ratings from Moody's, S&P, etc, so I think I won this round. 

The FICO scores suffer from the same 'grade inflation' as Harvard:  too many people with 700+ scores and the probability of default at that level (5%) was clearly too LOW entering this financial tsunami. 

And yes, as someone who looks over the finances of people I can say that for equal FICO scores, if someone has financial assets they are MORE likely to keep paying even if their debt/income goes beyong 28/36.

---

I know of very few institutions that lend solely on the basis of one's FICO score. If you read many posts here, you'd see repeated bafflement at being denied for this, that, and the other when supposedly their score was good enough.


eta: wow, just noticed that this was a 10-month-old thread that was resurrected!

Message Edited by haulingthescoreup on 02-25-2009 08:28 PM

[Anonymous]

Wow!  What an educational response.  Thanks for so much information.

[Anonymous]

Hey 700,

You're right that the new Fico isn't perfect but I'm not sure I agree with some of your proposed fixes. Most of what you fault Fico for not considering in calculating your score is more properly the role of the Lender in assessing the creditworthiness of borrowers at the point of purchase.

Verifying your income, assets, length of time in your job,  borrowing capacity, etc. etc.

all of these seem to speak more to the likelihood of a Lender being able to recover on the loan or minimize his losses in the event of default.  But since Fico never lends you any money I just don't see how any of that is relevant for Fico scoring purposes.

And as I understand it, Fico is willfully blind to these extraneous issues in the interest of fairness and justice. It's just not in the mission statement. Fico merely reports on your credit history and let's the Lenders decide the rest.

Personally, I wouldn't want to have a class based credit system where fat cats hog it all and the rest of us have to pay cash.That would be as bad, in my view, or worse, than considering age and race or sex in the calculations.

However, what frosts me is that Fico allowed people a piggyback loophole that lets newbies tag a free ride on the credit of well established borrowers. That's a slap in the face to people who actually earned their Fico credit scores.

They once talked about eliminating the piggyback loophole but last I heard they were backing away from the notion. That's a bad play on their part, for sure. Piggybackers haven't demonstrated anything as far as responsibly using credit. All they have proven is they have a cosigner's confidence.

Yeah, and a lot of people had confidence in Bernie Madoff--for what that was worth.

Regards,

Cato  

[Anonymous]

---

@Anonymous wrote:

I guarantee that FI has better evidence regarding the predictiveness of any particular variable than you or I do.

Well, the FICO scores have been shown to have been as reliable as the ratings from Moody's, S&P, etc, so I think I won this round. 

If your point was, at heart, how difficult it is to accurately predict the future behavior of people based simply on past behavior without any regard to "unnatural stressors," then yes, I guess you did. If that WASN'T your point, then, nope, you pretty much didn't.

MY point is that human behavior is, far more than it isn't, UNpredictable. As Lewis Black once said, "We're all just special little snowflakes." FI does it's best (and I think, despite my horrid scores) that it's doing a damned good job, considering the material it has to work with.

The FICO scores suffer from the same 'grade inflation' as Harvard:  too many people with 700+ scores and the probability of default at that level (5%) was clearly too LOW entering this financial tsunami. 

And yes, as someone who looks over the finances of people I can say that for equal FICO scores, if someone has financial assets they are MORE likely to keep paying even if their debt/income goes beyong 28/36.

---

You're applying too many statistical numbers and variables to something that can't be quantified so simply and resolutely. Granted, FICO's doing that, too, but it's got MILLIONS of data that it uses to come up with it's assumptions. All you have is your (limited) concept and your spattering of data you've gleaned from your job. Unless you've dealt with every single person in the USA who's ever gotten credit (as FICO quite obviously has), your sampling is too limited.

Not enough.