Opinions differ on effectiveness of FICO score

[Tuscani]

---

Boswd wrote:

I agree 100%. What ever happened to teaching econ/finance in schools.  Too many people live by the seat of their pants in search of the all elusive American Dream. No one has a saving's account anymore and no one seems to put money away for a rainy day.  They live paycheck to paycheck, credit card to credit card and then people wonder why the economy is such a mess. There's no get rich quick out there.  It's just hard work, diligence, patience and common sense.

I could not agree more.

---

What about PowerBall!!

[Boswd]

I play MegaMillions,  the solution to all my problems

[smallfry]

I agree totally. Although the sheeple have been sold a bill of goods by TPTB you have to have enough sense to buy only what you can afford. If you can't afford to buy a house using a convential 30 year fixed then chances are you will lose that house. RE doesn't always go up and interest rates can pop higher at any time. The buyer is ultimately at fault here not the sharks at the banks or mortgage companies. You cook it you eat it.

[smallfry]

FICO is blameless here. I don't blame the lenders either. I would be willing to bet that the lenders escape relatively unscathed. The loans are probably bundled and sold to pension plans. So the borrowers get bent over twice here.

[smallfry]

Not the school's fault either. Whatever happened to the individual spirit this country was founded on? It's up to each individual to educate himself not depend on Uncle to do it for him. It's this mindset that has us in the grave situation we are currently in. No culpability anywhere everyone is graded on a curve no failures. Yeah right.

[hdporter]

This discussion has moved on to debate consumer responsibility in their defaults (well, the debate is pretty one sided). However, the parties in the article weren't assigning "blame" for the defaults.

Instead, the issue was the variance in the number of actual defaults from those expected by the lenders. Lenders say FICO misled them; FICO asserts that lenders changed their lending practices from those that were assumed under the FICO estimates. This is the issue addressed in the article.

- Harry

Message Edited by hdporter on 05-03-2007 09:07 AM

[hdporter]

---

@Anonymous wrote:

How is the FICO score at fault? It hasn't changed the basic parameters. It's not like suddenly the score indicates that someone with a 550 score has only a 2% risk of defaulting.

---

No, but let's say that FICO predicts a 10% risk of default (or whatever value seems realistic to you).

Lenders would price the product so that an allowance for that default cost, assuming the 10% risk assumption, is covered by the profit they realize from borrowers who don't default.

Thus far, everyone is happy, provided that the default risk comes in reasonably close. However, if we find that the actual default rate is 50% higher than predicted (15%) and, as a consequence, lenders suffer a loss in lending to this sub-prime group, then a little finger pointing wouldn't be surprising.

If the lenders have conducted their underwriting in a manner that is comparable to the way they have historically done so (and was assumed in FICO's predictions of the default rate for this group), then there's cause to hold FICO responsible for having developed an inaccurate model. Bear in mind that FICO offers reasonable assurances of that accuracy in selling the product.

However, if lenders have changed their lending practices so that they're more generously extending credit to this group, than it would be expected that default experience would be higher than predicted in a FICO model that assumes more conservative lending practices.

The lenders and FICO are at odds over which of these two explanations are at the heart of the higher default rate. We can't make an accurate call -- again, I believe lender behavior is more likely at the root.

- Harry

[Anonymous]

Harry - I think you and I are agreeing to agree.

All the FICO data that goes into the formula is historical, based on millions of people's behavior over the past 20 years or so.  But the types of loans and the lending criteria have recently changed.  It may be too soon for today's foreclosures to be taken into account in the FICO risk prediction model. 

Ten years ago, subprime customers would not have been offered ARMS, interest only mortgages, etc.  Maybe they wouldn't have been offered mortgages at all based on their credit, income and debt load.  But the lenders recently changed their practices, the real estate poop is just starting to hit the fan, and eventually the default rates will be reflecting in FICO scoring.  Or something like that.

[Anonymous]

I think the Fico scoring system is extremely flawed. First of all, it doesn't have an exact points system so one cannot tell exactly how many points will be lost or gain if certain actions are taken on your credit report. One has to keep guessing and god forbids if you are missing those 2 little points that will move you from "higher risk" to "lower risk". Secondly, one cannot dispute the score directly, whatever score the systems gives, one has to accept. There are a myriad of other reasons that I don't have time to list and I think the system needs an overhaul. Lawyers need to take the people that develop this system to court so they can change this travesty.

[Anonymous]

One point FICO critics have to consider is that a scoring system is only as good as the data it receives. With all the chicanery people play with the system (like CAs re-reporting the same debt over and over to failing to update accounts to consumers adding other peoples' tradelines) the quality of this data is rather questionable.

I mainly blame the CRAs themselves. They could significantly improve the integrity of the system by refusing to allow the listing of any debt more than once, and by requiring ALL pertinent information associated with a debt (account number, original creditor, current creditor, data of delinquency, balance, and full contact information) be reported in order to list a debt. Requiring full data would eliminate a lot of the fraud and "noise" from the system. Just like ISPs use spam scoring systems, the CRAs could keep tabs on the quality of the data submitted by a CA, and ban that CA from its system if that CA was the source of too many incorrect or bogus reports.