topic Re: FICO scoring on installment loans is anal in Understanding FICO® Scoring

FICO scoring on installment loans is anal

DaveInAZ — Fri, 12 May 2017 19:02:13 GMT

If you don't have an installment loan reporting you're penalized for that in your credit mix. Get an installment loan, and you are penalized for the loan balance being "too high" until you get the balance down to less than 10% (or is it 9%?).

Say you take out a 60 month car loan. Assume your monthly payments all equally lower your principal. Yeah, they don't because more of your payment goes toward interest in the beginning and more toward principal at the end. But for simple math's sake let's say they all equally lower your principal.

So, for the first 54 of responsible monthly payments FICO gives you a "You naughty boy! Loan balance is too high!"

Then you get a whole 6 months of "Atta boy!, loan balance is acceptable to us."

Then you make the last payment paying it off, and you start getting a "Tsk, tsk! No recent installment loans!"

Right now, I'm getting both #1 & #2 of above. In Feb, 2015 I took out a car loan at a small local credit union at a great rate. It's almost half paid off, $5500 balance on a $10k loan. But they only report to EX & EQ. So on EX & EQ FICOs I get "Loan balance is too high!!!!", while on TU I get "Tsk, tsk! No recent installment loans!"

Yeah, nothing can be done about TU until I get an installment loan that reports to it. I have app'd for a Home Equity Line of Credit with Penfed, but since it's a LOC and not a fixed loan amount I don't think it count as installment, although it should report as a real estate loan which should help the account mix. I just do what's best for me financially, and while my FICO scores are important, I don't let that dictate my financial decisions.

But in my opinion FICO considering an installment loan balance being "too high" is ridiculous. I think a 50% loan balance would be reasonable, you've made around half the payments in a responsible, timely manner. Yeah, I/we aren't going to get FICO to change its scoring algorithms, just a rant & I'll be curious about the comments.

Re: FICO scoring on installment loans is anal

Glen_M — Fri, 12 May 2017 19:06:41 GMT

It's a math formula created by a bunch of accountants and statisticians, not an amusement park ride.

It doesn't need to "make sense", just make money.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 19:13:09 GMT

That's why the Alliant SSL "hack" is so popular even with people who have high FICO scores.

I'm basically screwed on that note without the hack. I own my home mortgage-free and I use private car services so I never will own a car again. Don't need a loan for anything so I just will keep the SSL going and renew a new one once that one expires.

As for your interpretation that 50% utilization is still "too high", don't forget that FICO spends millions (hundreds of millions over decades) analyzing real people and risk factors and their math holds up very well for investors in debt. 50% left on a loan still puts someone in a category where they may be a credit risk later -- if you lose your job, get sick, etc, you can't just PIF that loan most of the time like one at 9%. 50% may not even cover the current market value of the asset if you bought in a bubble.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 19:12:07 GMT

Hi Dave! A few quick thoughts:

(1) Part of your concern is purely practical. That is, you want to make sure that you are getting the full scoring benefit for your TU score (since you have no open installment loan on the TU report).

Happily that is easy to solve. Just take out a Share Secure loan with Alliant, pay almost all of it off, and keep it open for five years. The first three posts in this thread will tell you all you need to know about that:

http://ficoforums.myfico.com/t5/Understanding-FICO-Scoring/Adding-an-installment-loan-the-Share-Secure-technique/m-p/4506756

(2) You may be under the impression that you get no scoring benefit unless your installment utilization is < 9%. This is not so. You almost certainly get some benefit far earlier than that.

Furthermore, contributor Thomas Thumb speculates that the thresholds for benefit may be quite a bit higher if a person has been making payments on a loan in good standing for a while. What constitutes a "while" is uncertain, but I think TT doubts it is longer than 24 months.

(3) Remember that there isn't a single way that FICO handles installment utilization (how much of one's open debt has been paid off). FICO 8 cares about that. The mortgage model for TU and EQ (FICO 04) does not care at all. The mortgage model for EX (FICO 98) cares but perhaps not as much as FICO 8. And I don't think we have much hard evidence for exactly how much FICO 9 cares, though I have been trying to get people on the SS loan thread to provide that.

Future FICO models may very well track and reward the successful management and gradual payoff of loans that are 24+ months in length. Certainly someone with a history of five such loans (say) on his report (but no open debt) deserves more scoring benefit than someone with no such history but who is exploiting the SS loan trick.

But until the models arrive that you'd like, the good news is that you can certainly take practicial steps to prevent yourself from being penalized by the successful payoff of your open debt. Just make sure you have implemented the SS loan technique and you will be set.

(4) You write: "I just do what's bestfor me financially, and while my FICO scores are important, I don't let that dictate my financial decisions." Good for you! Great way to think.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 20:14:47 GMT

I had only 3 CCs and too out a personal loan and lost between 45-58 points on all FICO 28 scores. Most of them dropped me from the 700s. So I'm guessing as he's guessing. On 60 month loan I'm also going to be punished for a long time too

Re: FICO scoring on installment loans is anal

Gunnar419 — Fri, 12 May 2017 20:26:26 GMT

The Alliant Share Secure loan is definitely a good credit hack, but why should it be necessary?

FICO's scoring regarding installment loans may be the most irrational part of credit scoring. I can see the point of wanting an installment loan to be present in the credit mix, but not the point of slaughtering otherwise good scores the instant the installment loan gets paid off. If closed and paid credit cards count on our credit for up to 10 years, why should an installment loan ding us, and badly, the month after it's paid off?

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 20:38:08 GMT

@Gunnar419 wrote:
If closed and paid credit cards count on our credit for up to 10 years, why should an installment loan ding us, and badly, the month after it's paid off?

When has FICO been for the consumer? It isn't. It's for banks. FICO is a great way for banks to commoditize debt by risk factor with the explicit desire to make a profit.

Someone who doesn't carry a loan right now has no CURRENT report of how well they can manage it. How you managed credit in the past well has no reflection on your profitability for banks today. Even if your history up to last month was perfect payment, it still doesn't reflect how profitable you will be for a lender today.

Carrying a loan at the moment and showing the ability to manage a broad mix of credit tells banks you're a low risk, but you MIGHT be a profitable consumer.

FICO is not for consumers, it's for lenders.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 21:14:05 GMT

@Gunnar419 wrote:
The Alliant Share Secure loan is definitely a good credit hack, but why should it be necessary?

FICO's scoring regarding installment loans may be the most irrational part of credit scoring. I can see the point of wanting an installment loan to be present in the credit mix, but not the point of slaughtering otherwise good scores the instant the installment loan gets paid off. If closed and paid credit cards count on our credit for up to 10 years, why should an installment loan ding us, and badly, the month after it's paid off?

I'm not sure that's the best parallel to draw. Like closed and paid credit cards, closed and paid loans remain on one's credit reports for (typically) ten years.

And you would find that if all of your revolving accounts became closed, you would also experience a substantial scoring penalty. Even if only one revolving account among two or three becomes closed, that can cause you to take a score drop, in a manner analgous to the (possible) score loss that can occur with one loan amongst two or three open loans becoming closed. (The closure can affect "utilization.")

So far from open installment accounts being treated much differently than open revolving accounts, FICO 8 treats them in a similar way. With both, there is a "utilization" factor that applies only to open accounts. With both, there's a big penalty when all accounts of a single type becomes closed.

A better line of thinking, to my mind, would be one that argues that closed installment loans should be scored very differently from closed revolving accounts. I.e. that FICO should look at a closed loan that lasted for 24+ months with the borrower never missing a payment and should give him some scoring points for that, especially if he has more than one such closed loan. FICO does not do that especially for closed credit cards but (it could be argued) FICO should do that for closed installment loans. And it may well be that FICO is planning that for future models.

So as I mentioned in my first response, I sympathize with the OP and agree that he may be right that better models would consider closed loans as a scoring asset. But the way to get there is not to argue that closed revolvers and installment loans should be treated in a very similar way. They are -- which is likely the problem you are sensing.

Re: FICO scoring on installment loans is anal

DaveInAZ — Fri, 12 May 2017 21:21:56 GMT

@Anonymous wrote:
Hi Dave! A few quick thoughts:

(1) Part of your concern is purely practical. That is, you want to make sure that you are getting the full scoring benefit for your TU score (since you have no open installment loan on the TU report).

Happily that is easy to solve. Just take out a Share Secure loan with Alliant, pay almost all of it off, and keep it open for five years. The first three posts in this thread will tell you all you need to know about that:

http://ficoforums.myfico.com/t5/Understanding-FICO-Scoring/Adding-an-installment-loan-the-Share-Secure-technique/m-p/4506756

(2) You may be under the impression that you get no scoring benefit unless your installment utilization is < 9%. This is not so. You almost certainly get some benefit far earlier than that.

Furthermore, contributor Thomas Thumb speculates that the thresholds for benefit may be quite a bit higher if a person has been making payments on a loan in good standing for a while. What constitutes a "while" is uncertain, but I think TT doubts it is longer than 24 months.

(3) Remember that there isn't a single way that FICO handles installment utilization (how much of one's open debt has been paid off). FICO 8 cares about that. The mortgage model for TU and EQ (FICO 04) does not care at all. The mortgage model for EX (FICO 98) cares but perhaps not as much as FICO 8. And I don't think we have much hard evidence for exactly how much FICO 9 cares, though I have been trying to get people on the SS loan thread to provide that.

Future FICO models may very well track and reward the successful management and gradual payoff of loans that are 24+ months in length. Certainly someone with a history of five such loans (say) on his report (but no open debt) deserves more scoring benefit than someone with no such history but who is exploiting the SS loan trick.

But until the models arrive that you'd like, the good news is that you can certainly take practicial steps to prevent yourself from being penalized by the successful payoff of your open debt. Just make sure you have implemented the SS loan technique and you will be set.

(4) You write: "I just do what's bestfor me financially, and while my FICO scores are important, I don't let that dictate my financial decisions." Good for you! Great way to think.

Hi CreditGuyInDixie - Your thread on the Share Secure loan with Alliant is excellent,kudos!, and I've read much of it. Right now my FICO scores across the 3 CRAs are in flux, as the accounts included in my 2010 BK7 are falliing off. I got TU to early exclude all of them, and it jumped from 677 to 720 FICO. I got EX & EQ to exclude all but my not reaffirmed mortgage, but they're still mired at 678/685 EX/EQ, and they're the ones reporting my car loan (and at 1.99% APR I'm not paying it down to under 10% just to make FICO happy!).

So I'm not sold on the necessity of a secured loan for TU, but once things are all equal with BK accounts gone if I think it's worthwhile my local CU where I have accounts offers secured loans so I'd go with them. But your instructions on paying it down to under 10% and making minimal payments for as long as possible is an excellent guideline.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 21:38:14 GMT

@Anonymous wrote:
@Gunnar419 wrote:
If closed and paid credit cards count on our credit for up to 10 years, why should an installment loan ding us, and badly, the month after it's paid off?
When has FICO been for the consumer? It isn't. It's for banks. FICO is a great way for banks to commoditize debt by risk factor with the explicit desire to make a profit.

Someone who doesn't carry a loan right now has no CURRENT report of how well they can manage it. How you managed credit in the past well has no reflection on your profitability for banks today. Even if your history up to last month was perfect payment, it still doesn't reflect how profitable you will be for a lender today.

Carrying a loan at the moment and showing the ability to manage a broad mix of credit tells banks you're a low risk, but you MIGHT be a profitable consumer.

FICO is not for consumers, it's for lenders.

Hi ABCD! That sound plausible, but I am pretty sure it is not so. (Though again, I can totally see why you'd think it might be.)

FICO has been repeatedly clear that the big scoring models we are all familar with (Classic, Auto Enhanced, and even Bankcard Enhanced) produce a number that measures one thing: the likelihood that a consumer will become severly delinquent on one or more accounts in the next X months. The exact value for X might vary for the model: maybe it is 18 months for one model or 24 months for another. And "severely delinquent" might have a slightly different definition. (Though I wouldn't be surprised if it had a consistent meaning of "90 days.")

But FICO's published white papers are clear that this and only this is what the score measures.

And aside from what the white papers say, we can see this is empirically by looking at people who are extremely unprofitable for CC issuers (indeed they make the CC companies negative profit) but who score extremely high (800s) on FICO 8 Classic and FICO 8 Bankcard Enhanced. Indeed I am pretty sure that, for every point a person's score exceeds 780, that is correlated with an increasing probability that he will make the CC issuer less money (or even make the issuer negative money). It's possible for a person to have a score in the 840s and yet make the issuer a good deal of money, but that is very rare.

If the score was partly measuring profitability, this could not happen.

But where I think you are likely on target is that FICO may well be making other kinds of highly specialized data mining tools that assist a CCC (for example) in identifying lucrative customers. It's just that the main scoring model does not do that, save through the method of enabling the CCC to assess risk of default (something that certainly eats into profitability).

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 22:12:57 GMT

@Anonymous wrote:
Hi ABCD! That sound plausible, but I am pretty sure it is not so. (Though again, I can totally see why you'd think it might be.)

FICO has been repeatedly clear that the big scoring models we are all familar with (Classic, Auto Enhanced, and even Bankcard Enhanced) produce a number that measures one thing: the likelihood that a consumer will become severly delinquent on one or more accounts in the next X months. The exact value for X might vary for the model: maybe it is 18 months for one model or 24 months for another. And "severely delinquent" might have a slightly different definition. (Though I wouldn't be surprised if it had a consistent meaning of "90 days.")

But FICO's published white papers are clear that this and only this is what the score measures.

And aside from what the white papers say, we can see this is empirically by looking at people who are extremely unprofitable for CC issuers (indeed they make the CC companies negative profit) but who score extremely high (800s) on FICO 8 Classic and FICO 8 Bankcard Enhanced. Indeed I am pretty sure that, for every point a person's score exceeds 780, that is correlated with an increasing probability that he will make the CC issuer less money (or even make the issuer negative money). It's possible for a person to have a score in the 840s and yet make the issuer a good deal of money, but that is very rare.

If the score was partly measuring profitability, this could not happen.

But where I think you are likely on target is that FICO may well be making other kinds of highly specialized data mining tools that assist a CCC (for example) in identifying lucrative customers. It's just that the main scoring model does not do that, save through the method of enabling the CCC to assess risk of default (something that certainly eats into profitability).

FICO absolutely is a profit indicator. It isn't sold as such, but it absolutely is one for those who commoditize debt to sell to investors.

Remember: the bank does not generally care only for what YOU spend and generate in either interest rate profits or merchant processing fees but the bank also cares where your particular snapshot (as per FICO) will sit in an aggregate group of borrowers that the bank is bundling to let investors take part in. In fact, you can go back more than 20 years to see the beginning of this trend (see Katherine Samolyk's report from the early 00s).

So while a consumer layman may think a bank loves a consumer with a mid-high FICO score who pays interest, that's not how the bank sees it. A bank may make more money through bundling and commoditization of debt borrowers by adding more high FICO low return customers to a package, thereby skewing the default profile just enough to squeak out a tiny bit more return from investors (and investment AI software) who might have passed on the package had it been a fraction higher in default risk at the same price. (Jens Beckert & Patrik Aspers book from 2011 titled "The Worth of Goods").

Banks don't just make money on interest or merchant fees, they make money on bundling creditors to resell to bond investors (and other commoditized bundled packages) -- see Peter Eavis, 2012, NY Times.

FICO isn't sold as just a risk profile anymore. It probably was 20 years ago, but now FICO isn't just promoting it as a risk measure solely -- when you incorporate securitization into the game, it's also about overall profit incentive for investors buying up those debts as a bundle (Vinod Kothari, "Securitization: The Financial Instrument of the Future", 2006 ).

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 22:55:33 GMT

I admit to not following what you are saying fully. I guess it's because I do not know much about what you are talking about with respect to the commoditization of debt. I think this is touched on in the documentary "Inside Job" (about the financial meltdown of 2008 and the history of changes in the financial world leading up to it, e.g. derivatives). Is that what you are talking about: the bundling of loans and them being sold to investors?

I guess I was talking about something that feels a lot simpler.

The OP was wondering why the FICO 8 model would reward him as an individual consumer for having an open installment loan (that is mostly paid off) vs. having only closed loans. Why would he get a higher score? In your first reply, it sound like you were saying that the model does this because the score is measuring an individual's likelihood of being profitable (in some sense apart from his probability of becoming delinquent). I suggested that struck me as improbable for a couple reasons, one of which was that a person is increasingly less likely to be profitable to banks (at least those who issue CCs) as his FICO 8 Classic increases.

In your reply to me just now it sounds like (maybe?) you agree that the score (as it increases from 780 to 850) does not track with the increasing probability that an individual will be profitable (again outside the obvious benefit to profit that consumers who have very low rates of default will be more profitable, especially to lenders of installment loans). I am not sure but maybe you are saying that. In which case that's all I was saying too: the score does not predict a person's profitability, only default risk.

I don't have any difficulty imagining that banks could bundle high scoring (i.e. extremely low default risk) clients with other riskier borrowers to make a bundle of loans more appealing to investors. But we don't have to suggest that the model itself is designed to predict profitability of individuals to explain that. That would be true even if the model does exactly what it claims to do: predict risk of severe delinquency.

FICO goes back and forth in its tendency to make installment utilization an significant score factor. Some models (FICO 98) it matters. Then FICO removes it altogether (FICO 04). Then it puts it back in again (FICO 8). And I am not sure anybody knows for sure exactly how it is weighted in FICO 9. I think it is probably just FICO doing its best to assess risk of default, based on the data it has at the time.

If trended data end up taking off, we may see all kind of profitability assesment tools being created, by FICO and others.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 23:06:34 GMT

@Anonymous wrote:
Is that what you are talking about: the bundling of loans and them being sold to investors?

Indeed! It's about a way banks can profit from subprime borrowers with much less risk, while also profiting from superprime borrowers who don't pay interest!

I guess I was talking about something that feels a lot simpler.

Definitely but the sad truth is FICO is still about lenders, not borrowers. In fact, FICO is more about commercial banks today (investment banks) than it is about consumer banks -- unlike how it was 30 years ago.

In your first reply, it sound like you were saying that the model does this because the score is measuring an individual's likelihood of being profitable (in some sense apart from his probability of becoming delinquent). I suggested that struck me as improbable for a couple reasons, one of which was that a person is increasingly less likely to be profitable to banks (at least those who issue CCs) as his FICO 8 Classic increases.

My apologies, I worded it in a confusing fashion. A consumer's FICO score, say FICO 8, is an early indicator of what group of risk that consumer falls into, but on the flip side it can also be an indicator of potential for a consumer bank ("credit card company") to profit from using that consumer in a role to further the consumer bank's goal of selling that debt to a commercial bank's customers (investors).

In your reply to me just now it sounds like (maybe?) you agree that the score (as it increases from 780 to 850) does not track with the increasing probability that an individual will be profitable (again outside the obvious benefit to profit that consumers who have very low rates of default will be more profitable, especially to lenders of installment loans). I am not sure but maybe you are saying that. In which case that's all I was saying too: the score does not predict a person's profitability, only default risk.

I agree with this -- an individual consumer (borrower) may not be more profitable as a spender/consumer to that consumer bank as an individual, but as a unit in a larger group of consumers, that individual's FICO score can help the bank market a less risky aggregate portfolio of borrowers together as a more profitable bundle to commercial banks.

I don't have any difficulty imagining that banks could bundle high scoring (i.e. extremely low default risk) clients with other riskier borrowers to make a bundle of loans more appealing to investors. But we don't have to suggest that the model itself is designed to predict profitability of individuals to explain that. That would be true even if the model does exactly what it claims to do: predict risk of severe delinquency.

Even FICO themselves admits this is the case, though: "Scores play a critical role in the securitization of asset-backed securities such as auto loans, telco contracts and GSE-backed mortgages. FICO Scores are understood and trusted by investors as a key indicator of the default risk of a given pool of loans to be securitized, which informs how much the investor is willing to pay for that pool." To wit, the FICO score is an indicator of default risk for the explicit purpose of setting a price based on profit potential. http://www.fico.com/en/blogs/risk-compliance/truth-squad-is-a-fico-score-700-the-same-as-a-vantagescore-700/

Also: “As an extension of this position, we are excited to offer more comprehensive solutions to support the needs of the secondary market and securitization process for US mortgages. We believe this combination of Experian Capital Markets and FICO solutions will increase investor confidence and make for a more efficiently functioning market.” http://www.fico.com/en/newsroom/advanced-fico-analytics-incorporated-into-experians-mortgage-backed-securities-solution-10-04-2010

FICO goes back and forth in its tendency to make installment utilization an significant score factor. Some models (FICO 98) it matters. Then FICO removes it altogether (FICO 04). Then it puts it back in again (FICO 8). And I am not sure anybody knows for sure exactly how it is weighted in FICO 9. I think it is probably just FICO doing its best to assess risk of default, based on the data it has at the time.

Yep and this is based somewhat on changing metrics for profit positions by both consumer banks AND commercial banks. For example, before the prior housing bubble, it was common to see consumer banks with crazy low interest rates in order to try to profit themselves from the borrowers' positions. Nowadays, even SUPERPRIME borrowers are given way higher interest rates on their credit cards because the consumer banks themselves want to sell those positions to commercial banks to bundle to investors.

If trended data end up taking off, we may see all kind of profitability assesment tools being created, by FICO and others.

My comments in blue, of course!

BTW, I told probably 20 people I knew with 750-790 FICO08 scores to try the Alliant SSL and sent them your link -- will report back when I hear from them to see if any of them get the super pop to closer to 850. All of them are free of mortgage, auto and student loans and can't get above 790.

Re: FICO scoring on installment loans is anal

DaveInAZ — Fri, 12 May 2017 23:08:45 GMT

@Anonymous wrote:
@Anonymous wrote:
Hi ABCD! That sound plausible, but I am pretty sure it is not so. (Though again, I can totally see why you'd think it might be.)

FICO has been repeatedly clear that the big scoring models we are all familar with (Classic, Auto Enhanced, and even Bankcard Enhanced) produce a number that measures one thing: the likelihood that a consumer will become severly delinquent on one or more accounts in the next X months. The exact value for X might vary for the model: maybe it is 18 months for one model or 24 months for another. And "severely delinquent" might have a slightly different definition. (Though I wouldn't be surprised if it had a consistent meaning of "90 days.")

But FICO's published white papers are clear that this and only this is what the score measures.

And aside from what the white papers say, we can see this is empirically by looking at people who are extremely unprofitable for CC issuers (indeed they make the CC companies negative profit) but who score extremely high (800s) on FICO 8 Classic and FICO 8 Bankcard Enhanced. Indeed I am pretty sure that, for every point a person's score exceeds 780, that is correlated with an increasing probability that he will make the CC issuer less money (or even make the issuer negative money). It's possible for a person to have a score in the 840s and yet make the issuer a good deal of money, but that is very rare.

If the score was partly measuring profitability, this could not happen.

But where I think you are likely on target is that FICO may well be making other kinds of highly specialized data mining tools that assist a CCC (for example) in identifying lucrative customers. It's just that the main scoring model does not do that, save through the method of enabling the CCC to assess risk of default (something that certainly eats into profitability).
FICO absolutely is a profit indicator. It isn't sold as such, but it absolutely is one for those who commoditize debt to sell to investors.

Remember: the bank does not generally care only for what YOU spend and generate in either interest rate profits or merchant processing fees but the bank also cares where your particular snapshot (as per FICO) will sit in an aggregate group of borrowers that the bank is bundling to let investors take part in. In fact, you can go back more than 20 years to see the beginning of this trend (see Katherine Samolyk's report from the early 00s).

So while a consumer layman may think a bank loves a consumer with a mid-high FICO score who pays interest, that's not how the bank sees it. A bank may make more money through bundling and commoditization of debt borrowers by adding more high FICO low return customers to a package, thereby skewing the default profile just enough to squeak out a tiny bit more return from investors (and investment AI software) who might have passed on the package had it been a fraction higher in default risk at the same price. (Jens Beckert & Patrik Aspers book from 2011 titled "The Worth of Goods").

Banks don't just make money on interest or merchant fees, they make money on bundling creditors to resell to bond investors (and other commoditized bundled packages) -- see Peter Eavis, 2012, NY Times.

FICO isn't sold as just a risk profile anymore. It probably was 20 years ago, but now FICO isn't just promoting it as a risk measure solely -- when you incorporate securitization into the game, it's also about overall profit incentive for investors buying up those debts as a bundle (Vinod Kothari, "Securitization: The Financial Instrument of the Future", 2006 ).

I thought the comment would be interesting.

I was going to reply to your previous comment, ABCD: "When has FICO been for the consumer? It isn't. It's for banks.". because that hit me with a "Duh! Of course!" Many of us here get conditioned to thinking in terms of what the FICO score can do for us, the consumer - higher score, bigger CLs, lower APRs. But of course FICO was developed for the lenders, a quick & easy scorecard indicating how risky a borrower might be, but also how profitable they may be. There's so much in a credit report: Low current utilization? Great, nice FICO, but have they carried a balance in the past and paid interest? Ah yes, there it is, back in 2015. So this borrower is one who is willing to splurge for a large purchase and pay interest on it.

I agree with you, FICO absolutely is a profit indicator, and I think it is even privately sold as such. For the commodization of debt as you went into such excellent detail on, but probably a large number of things. Like is the borrower a "churner"? With so many, increasingly large new account bonuses, lenders want to know if you're a long term card holder, or someone who takes the bonus & runs. Large banks like Chase have come up with their own model, the 5/24, but for many smaller banks it's cheaper to use FICO.

Re: FICO scoring on installment loans is anal

Anonymous — Fri, 12 May 2017 23:30:00 GMT

@DaveInAZ wrote:
I agree with you, FICO absolutely is a profit indicator, and I think it is even privately sold as such. For the commodization of debt as you went into such excellent detail on, but probably a large number of things. Like is the borrower a "churner"? With so many, increasingly large new account bonuses, lenders want to know if you're a long term card holder, or someone who takes the bonus & runs. Large banks like Chase have come up with their own model, the 5/24, but for many smaller banks it's cheaper to use FICO.

Yep, if you spend enough time reading through FICO's own website, you'll find a lot of their own verbiage on this point. FICO sells FICO scores to consumer banks as a profit indicator, not just a risk one. Here's an example:

"In effect, the project moved The Co-operative Bank from a risk-based pricing approach to more a profit-optimized-based pricing approach," says Neill Crossley, principal consultant at FICO.

http://www.fico.com/en/wp-content/secure_upload/Coop_Bnk_Success_2465CS.pdf

"In another top bankcard issuer’s study, using the FICO® 8 Score for account management across the issuer’s entire card portfolio would produce an estimated $12 million in net new profit. The new score will assist the lender in segmenting their customer base for cross-sell promotions and credit line increases."

http://www.fico.com/en/newsroom/fico-8-score-is-new-standard-for-credit-risk-assessment-07-27-2010

"Typically, these decisions are single-shot decisions and designed to optimize objectives defined over a fixed time horizon — for example, decide on credit limits and interest rates for new accounts to maximize 2-year profit."

http://www.fico.com/en/blogs/analytics-optimization/unlocking-the-science-of-sequential-decisions/

FICO isn't meant to make sense for consumers. We are not the target audience, we're just the product that's sold to lenders ("retail bank") to sell to investors ("commercial banks"). If a consumer can start to understand FICO scoring, we stop being the product and start becoming the competitor to investors.

Once consumers start to understand a particularly popular version of a FICO score, we should expect a new score version to become popular... And then we start to analyze the new score to better compete with banks.

Re: FICO scoring on installment loans is anal

DaveInAZ — Sat, 13 May 2017 00:11:07 GMT

@Anonymous wrote:

The OP was wondering why the FICO 8 model would reward him as an individual consumer for having an open installment loan (that is mostly paid off) vs. having only closed loans. Why would he get a higher score?

No, actually I totally get that. Revolving CC debt vs. installment debt are two very different things. Revolving debt indicate how well you do on managing everyday expenses. Is your income such that you can easily handle everyday expenses? Low utilization indicates that, and also indicates you are responsible enough to not splurge in things you can't afford despite on having adequate credit limits to do that. Installment debt is for major purchases such as a car or home, and indicates you plan well enough to budget for that monthly expense. So having a mix of revolving & installment loans should have a positive effect on your FICO.

My bone of contention is that for the vast majority of the period of an installment loan you are penalized for what FICO considers a balance too high compared to the original loan amount, ignoring - or at least not giving adequate weight - to the fact that you have been responsibly making timely payments on the loan for years. That discourages me, and likely many others, from taking out an installment loans.

Re: FICO scoring on installment loans is anal

Anonymous — Sat, 13 May 2017 00:11:27 GMT

@DaveInAZ wrote:
... have they carried a balance in the past and paid interest? Ah yes, there it is, back in 2015....

I agree with you, FICO absolutely is a profit indicator, and I think it is even privately sold as such. For the commodization of debt as you went into such excellent detail on, but probably a large number of things. Like is the borrower a "churner"? With so many, increasingly large new account bonuses, lenders want to know if you're a long term card holder, or someone who takes the bonus & runs. Large banks like Chase have come up with their own model, the 5/24, but for many smaller banks it's cheaper to use FICO.

Hi Dave. It's ok for the conversation to move into a different (possibly more speculative) realm, but it's useful to track back to your original question. You were asking about a particular scoring model (FICO 8) and why does it score installment debt the way it does.

All the things you talk about in the text I quote above are things that FICO 8 cannot do. It cannot tell whether you are carrying a balance today, much less whether you did back in 2015. It cannot tell whether you have ever received a CC bonus. It does not penalize credit cards that are closed after a few months (except in the sense that such an account won't one day grow into a very old 11+ year account -- but even then that lack of benefit will only manifest itself in your score many years later, not now.

As I mentioned earlier, FICO might well be making highly specialized products now that attempt to leverage trended data or other data mining techniques, products it may indeed be selling to banks, and I think it is very likely we will see more of those kind of products (which assess an individual's profitability apart from default risk) in the future.

But the model you were asking about initially (the one that people talk about here with the 300-850 range) does not do those things.

Re: FICO scoring on installment loans is anal

Anonymous — Sat, 13 May 2017 00:14:45 GMT

@DaveInAZ wrote:
. That discourages me, and likely many others, from taking out an installment loans.

Correct, but remember banks want sensible customers and their best INITIAL metric is a snapshot of credit today.

Who makes more sense to give a loan to? Someone who has paid down a long term installment loan to under 9%, or someone who took on a loan that is still at 91%?

Who is more likely to want to get a new loan? The 9% guy or the 91% guy?

Who is less likely to want a loan? A guy who has no current loan, or a 9% one?

Re: FICO scoring on installment loans is anal

Anonymous — Sat, 13 May 2017 00:20:45 GMT

@Anonymous wrote:
BTW, I told probably 20 people I knew with 750-790 FICO08 scores to try the Alliant SSL and sent them your link -- will report back when I hear from them to see if any of them get the super pop to closer to 850. All of them are free of mortgage, auto and student loans and can't get above 790.

Delighted to hear that the SS loan thread was something you recommended to others. I was simply a mailman for that. The technique was discovered/invented by other folks far more knowledgeable than myself.

Re: FICO scoring on installment loans is anal

Anonymous — Sat, 13 May 2017 00:27:03 GMT

The scores can be used to segment client base 1) subprime give them the First Premier & Credit One deals 2) average 650-720 or there abouts they carry balances and will pay interest 3) prime transactors score above 720 no balance money made on swipe fees only. The segmenting is why so many Myficoers once there scores go above 705 ,710 quit getting mail pieces. They are not a profit center except for swipe fee (Amex type client) Capital 1 was one of the first heavy data miners and segmenters.