A request to MyFICO Barry....

[just_curious]

I'd appreciate it if you could review the thread entitled "FICO doesn't understand the credit card game" and add your comments, either there, here, or somewhere...

The last posting on the thread summarizes the issue I have been raising - - - -

Another poster wrote -

---

masdeocho wrote:

The FICO score measures the PROBABILITY of paying back debt, not the ABILITY.

---

And I responded -

No - actually, it does not.

Think about what has been said on this thread - by myself and others who use CCs to their full advantage, including rewards programs.  I have never missed making any payment on time, and in full, in decades of using a wide range of credit instruments.  Yet FICO is so hung up on the % of my credit limit that is outstanding at the moment the monthly statement issues that it cannot possibly properly judge my "probability of paying back debt".  To do that Fair Issac needs to modify their algorithms to reflect the PIF behavior of many CC holders, which would only require a very minor adjustment in the data they consider.

This has been my point since my original posting - that a relatively minor improvement to the analysis would make a measurable improvement in the FICO scoring system, and I haven't seen anyone rebut this.  The comments that "it is what it is" and "just play the game" and so forth aren't really responsive.  I know how to "play the game" but the real question is...........

WHY SHOULDN'T FICO IMPROVE ITS SCORING SYSTEM ??

[MidnightVoice]

---

just_curious wrote:

I'd appreciate it if you could review the thread entitled "FICO doesn't understand the credit card game" and add your comments, either there, here, or somewhere...

The last posting on the thread summarizes the issue I have been raising - - - -

Another poster wrote -

---

masdeocho wrote:

The FICO score measures the PROBABILITY of paying back debt, not the ABILITY.

---

And I responded -

No - actually, it does not.

Think about what has been said on this thread - by myself and others who use CCs to their full advantage, including rewards programs.  I have never missed making any payment on time, and in full, in decades of using a wide range of credit instruments.  Yet FICO is so hung up on the % of my credit limit that is outstanding at the moment the monthly statement issues that it cannot possibly properly judge my "probability of paying back debt".  To do that Fair Issac needs to modify their algorithms to reflect the PIF behavior of many CC holders, which would only require a very minor adjustment in the data they consider.

This has been my point since my original posting - that a relatively minor improvement to the analysis would make a measurable improvement in the FICO scoring system, and I haven't seen anyone rebut this.  The comments that "it is what it is" and "just play the game" and so forth aren't really responsive.  I know how to "play the game" but the real question is...........

WHY SHOULDN'T FICO IMPROVE ITS SCORING SYSTEM ??

---

I am not a FICO/FI employee, but.....

First - FICO deals in statistical probabilty, not individual cases.  If you want individual situations to be considered, manual review is the answer.

Second - improvements.  From a business perspective, FI does what any business should do in a Capitalist society - it seeks to make profit whilst satisfying its customers.  We are not its customers, it has no need to satisfy us.  If its customers demand improvements, then it should be working on them.  If we demand improvements, we don't count until we get the legislature behind us.

[Barry]

---

just_curious wrote:

I'd appreciate it if you could review the thread entitled "FICO doesn't understand the credit card game" and add your comments, either there, here, or somewhere...

The last posting on the thread summarizes the issue I have been raising - - - -

Another poster wrote -

---

masdeocho wrote:

The FICO score measures the PROBABILITY of paying back debt, not the ABILITY.

---

And I responded -

No - actually, it does not.

Think about what has been said on this thread - by myself and others who use CCs to their full advantage, including rewards programs.  I have never missed making any payment on time, and in full, in decades of using a wide range of credit instruments.  Yet FICO is so hung up on the % of my credit limit that is outstanding at the moment the monthly statement issues that it cannot possibly properly judge my "probability of paying back debt".  To do that Fair Issac needs to modify their algorithms to reflect the PIF behavior of many CC holders, which would only require a very minor adjustment in the data they consider.

This has been my point since my original posting - that a relatively minor improvement to the analysis would make a measurable improvement in the FICO scoring system, and I haven't seen anyone rebut this.  The comments that "it is what it is" and "just play the game" and so forth aren't really responsive.  I know how to "play the game" but the real question is...........

WHY SHOULDN'T FICO IMPROVE ITS SCORING SYSTEM ??

---

Hi just_curious,

You make a good point in asking "why shouldn't FICO improve its scoring system?"  The scoring analysts at Fair Isaac ask that all the time, and are constantly working to improve the FICO score's ability to predict future risk.  In fact, since we first developed this scoring system almost 20 years ago there have been many many improvements to the formula's abilty to do what it's supposed to do --measure risk.

All changes to the FICO scoring formula are are made for the purpose of increasing its ability to predict risk and are based on research conducted by analyzing actual consumer behavior as reflected in millions of consumer credit files over many years.  So when we place a relatively high amount of weight to factors such as payment history or the ratio of balances to limits on credit cards, it's because the research has shown these factors to be highly predictive.  Other factors, such as inquiries, also tend to be predictive but to a lesser degree and thus don't carry as much weight.

So, you may be right in stating that factoring in PIF behavior would improve the scoring system.  Or not.

It all depends on what the empirical evidence -- consumer credit bureau data -- shows in terms of how well the addition of this factor improves the ability to predict future risk.  And while I don't know to what extent PIF behavior has been studied by the Fair Isaac analysts, I can tell you it has been considered, along with just about every other aspect of credit behavior you can imagine (and many you can't), and I would simply have to guess that PIF behavior hasn't been found to provide enough predictive value to include it in the formula.

So, long story short....you may be right, but what consitutes "improvement" lies with what the data tells us in terms of predicting future risk.

Hope that makes at least some sense....

Barry

[just_curious]

Barry wrote:

Hi just_curious,

[deleted text]

So, long story short....you may be right, but what consitutes "improvement" lies with what the data tells us in terms of predicting future risk.

Barry

Barry -  I guess I just hope the gurus will continue to look at this topic as the CC rewards business is much more important that it was 20 or even 5 years ago, and has clearly impacted CC usage and payment behavior, resulting in some scores that don't make logical sense (not just mine...).

And, by the way, understanding "credit behavior" was sort of critical to my career, so I wouldn't assume too much lack of knowledge there....

Anyway - Thanks for the response - it helps make this forum useful to have you around it.....

[RobertEG]

Barry hit it right on the head.  FICO scoring is a model created for the benefit, not of consumers, but for creditors, to evaluate benefits and risks in their lending.  It is not a model created to boost consumer ability to gain more credit, or to evaluate the full credit worthiness of consumers.  A primary component of the risk portion of the FICO scoring is the potential for serious late payments (60+) over the next one to two years..  That is what creditors want, and what Fair Isaac is in the business of delivering.   They have analyzed millions of credit reports to develop the risk correlations.  I think they are delivering what lenders want to hear.

A 30-day late is not considered a serious negative risk under Fico, but 60+ lates are… in fact, I might cynically venture to postulate the reason for this....occasional 30-day lates are really a financial windfall to the credit card companies.  Once you are late for just 30 days on one payment, anywhere in the world, that is reported to a CRA, that gives the credit card company, under their fine print in the consumer "agreement," the ability to dissolve ALL their good prior rates, and set them at any usorious rate of their choosing.  Tony Soprano, late vig.  They love this.  But once lates get repeated, it makes their Arrid less than dry, and the risk begins to become more of a concern than the prior usury benefit inflicted after the relativley low risk 30 day late used to justify their future, and forever, rate hike might mean that future payments may not be received.   

And you can also bet that buried deep within the FICO algorithms is also the benefit side to the lendors, their clients.  When a creditor lends money, they have their own interest burden over the life of their loan, and expect to secure more interest than they incur.  One who PIF’s is not a starling preferred customer.  They have derogatory words for these types of customers.  So don’t think for a minute that if you are a PIF kinda guy or gal, and the credit grantor can predict this, that they want a model score that steers them towards the grant of credit to one who will produce little or no profit for them.  I would suppose that, if they had it their way, an identified PIF kinda guy or gal would lead them to want a lower, and not higher, reported FICO score.

We do not know what is buried in the vaults of the FICO algorithm, but rest assured that the buried nuggets are for the benefits of the lendors, and not the consumers. 

Do you suppose that might be why most lendors do not see those with FICOs approaching 800 to be more profitable in their lending decisions than the 740 guy, who has some past, but small, dings?  Hmmmm?????

Message Edited by RobertEG on 11-29-2007 09:48 PM

Message Edited by RobertEG on 11-29-2007 09:56 PM

[Anonymous]

Oh dear Lord! I feel like I've just come out of a 3 hour lab lecture, in a college level micro economics course!

How very well stated Barry & RobertG! Thank you 2.

[RobertEG]

LOL!  I am am laughing with you, and not at you, Sylviatob!  And welcome to the dark world of FICO!   LOL!  Understanding, at what ever level one has the time to understand, is the key.  This site has done more for my credit understanding than anyone in the industry would like me to know.  The people at the credit card companies who evaluate FICO scores, as well as the gurus at Fair Isaac who generate them, all have advanced degrees.  So sometimes some economic jibber-jabber is necessary in the discussion. 

But the gurus dont openly lecture on the topics that we consumers all discuss openly in here, for they are closely guarded industry trade secrets for the purpose of making money.  So it is left to us to speculate and pontificate.  And the money they ultimately make from the FICO game comes from, you guessed it, we, the consumer.  If the secrets that they choose to keep were for the benefit of the consumer, they would openly dislcose them, and they would no longer remain trade secrets.  They dont.  It is for them to make money, and not to benefit the consumer. The ONLY reason that we know as much as we do about FICO scoring is because Congress required, under the Fair Credit Act, that the public be entitled to see the basics of what they are doing in the scoring.  We can only speculate on their secret algorithms and motivations, which the Act does not require them to disclose... and nor do I advocate that they should.  I belive in captitalism, and that is da name of da game..  But we can try to decipher enough to let us see the basics, and try to make the captitalism work for both sides.

[Anonymous]

Waidaminit while I drag out my eco book, and my Webster's! hahahahahaha.

Well put Robert. That's what makes this fun(if you can call it that). Gather unrelated bits of information and try to extrapolate it into some reasonable semblance of the theory of how things work. Kinda like a twisted jigsaw puzzle, with mazes filled with frustration. LOL.

[MidnightVoice]

Another thing to remember is that it is a model based on large numbers of people.  Hence there will always be people whose behavior does not fit the model - outliers, and extremes on the bell curve (I was thinking of saying "Gaussian Distribution", but thought, as this is not a stats lecture, I would say "bell curve"  ).

So there will always be people who are a much better risk than their score suggests, and people who are a greater risk than their score suggests.

[Anonymous]

Yes! M.V., I'm convinced that a LOT of the modeling, if based permutation, probability, and a LARGE dose of statistics. Statistics, combined with economics. They've concocted a powerful voodoo potion to ward off even the baddest of evil spirits. lol. It's magic I say! Magic!