Should utilization matter more than debt?

[Anonymous]

---

@Thomas_Thumb wrote:

---

---

Unfortunately, there is no way to make an unbiased determination on that without knowing capacity to pay. Given income can not be considered, the ways to get a feel are looking at balance(s) relative to credit limit, payment history and possibly inferring something from installment loan amounts. Lenders and CC companies can/do look at income so some extrapolation can be done to determine a minimum likely income based on open. approved loans/loan payments.

I like the transparency of VantageScore and the white papers they publish validating scores against known results. Their analysis of data (chart posted up thread) continues to show utilization as a stronger predictive component than balances.

---

Sure there is.  It's simple statistics.  More people in the population have the capacity to pay a $200 debt than a $20k debt.  I don't really think that this can be argued.  100x more debt is an incredibly greater amount of exposure and represents a larger risk.  Again, if either person were to lose their income source it's extremely clear which person would have the greater capacity to pay back their debt.  Sure there are other factors not included here such as savings (assets) that are unknowns, but I'm speaking under the assumption that all other things are equal.

[Thomas_Thumb]

You have to have data sets to perform statistics. Answers are opinion without data and the supporting data analysis.

Of course, we all have opinions and that is what you are asking for in your subject line. In this case I choose to go with the analysis presented by VantageScore.

[Anonymous]

Why would I (or you) even need a data set to make such an obvious assumption?  I don't see how this is opinion.  I'm sure there's a data set out there that shows that a $20k debt is more risky than a $200 debt and that far more people would struggle paying back $20k over $200.  Anyone that has the capacity to pay back $20k obviously can pay back $200.  The converse certainly cannot be said, so by definition more people are able to pay back a $200 debt than a $20k debt.

I'm sure there's a data set out there that says that 3 years after someone achieves a doctorate level of education their average salary is greater than someone else 3 years after achieving just a high school GED.  I don't have access to that data set, but honestly I don't need to see it to know that the result of it is likely pretty cut and dry.  Are there outliers/exceptions to all data sets and statistics?  Of course.  That goes without saying.   

[Revelate]

---

@Anonymous wrote:

Why would I (or you) even need a data set to make such an obvious assumption?  I don't see how this is opinion.  I'm sure there's a data set out there that shows that a $20k debt is more risky than a $200 debt and that far more people would struggle paying back $20k over $200.  

I'm sure there's a data set out there that says that 3 years after someone achieves a doctorate level of education their average salary is greater than someone else 3 years after achieving just a high school GED.  I don't have access to that data set, but honestly I don't need to see it to know that the result of it is likely pretty cut and dry.  Are there outliers/exceptions to all data sets and statistics?  Of course.  That goes without saying.   

---

Taking a personal and absurd example:

Say I'm the one with 20K debt - I could write a check for it within a week.  Is that a significant risk?  I have near liquid assets well in excess of that, and a mortgage if I truly flaked that they could go after.  Many home owners fall into this category too.

Someone else, sans assets, living paycheck to paycheck, and a $200 debt they haven't cleared.

Which one is the higher risk when we're just talking possible default rates (and pointedly not lender exposure)?  I agree  somewhat based on statistical averages that 20K might be a more risky bet across the consumer base, but it's by no means a given.  As a result, since neither assets nor income are factored into the scoring, some average debt statistics don't make much sense.

That all said, I have the sneaking impression that within the next decade FICO will change their algorithms to look more at total debt than utilization: utilization metrics made more sense before outsized limits.  Also FICO's exclusion of it is why it's only one part of the underwriting process... where they absolutely they do factor in income and debt.  

[sarge12]

---

@Revelate wrote:

---

@Anonymous wrote:

Why would I (or you) even need a data set to make such an obvious assumption?  I don't see how this is opinion.  I'm sure there's a data set out there that shows that a $20k debt is more risky than a $200 debt and that far more people would struggle paying back $20k over $200.  

I'm sure there's a data set out there that says that 3 years after someone achieves a doctorate level of education their average salary is greater than someone else 3 years after achieving just a high school GED.  I don't have access to that data set, but honestly I don't need to see it to know that the result of it is likely pretty cut and dry.  Are there outliers/exceptions to all data sets and statistics?  Of course.  That goes without saying.   

---

Taking a personal and absurd example:

Say I'm the one with 20K debt - I could write a check for it within a week.  Is that a significant risk?  I have near liquid assets well in excess of that, and a mortgage if I truly flaked that they could go after.  Many home owners fall into this category too.

Someone else, sans assets, living paycheck to paycheck, and a $200 debt they haven't cleared.

Which one is the higher risk when we're just talking possible default rates (and pointedly not lender exposure)?  I agree  somewhat based on statistical averages that 20K might be a more risky bet across the consumer base, but it's by no means a given.  As a result, since neither assets nor income are factored into the scoring, some average debt statistics don't make much sense.

That all said, I have the sneaking impression that within the next decade FICO will change their algorithms to look more at total debt than utilization: utilization metrics made more sense before outsized limits.  Also FICO's exclusion of it is why it's only one part of the underwriting process... where they absolutely they do factor in income and debt.  

---

The problem is that creditors always wanted DTI as part of a credit score. Some of the members of congress objected that to do so would discriminate against the poor, and deny them access to the lowest interest rate, even when their payment history was perfect. Thus, as important as a credit score is to the access to credit and lower interest rates, the Government decided income and assets could not be used as part of the score, but can be considered by an individual creditor.

[Anonymous]

---

@Revelate wrote:

---

---

Taking a personal and absurd example:

Say I'm the one with 20K debt - I could write a check for it within a week.  Is that a significant risk?  I have near liquid assets well in excess of that, and a mortgage if I truly flaked that they could go after.  Many home owners fall into this category too.

Someone else, sans assets, living paycheck to paycheck, and a $200 debt they haven't cleared.

Which one is the higher risk when we're just talking possible default rates (and pointedly not lender exposure)?  I agree  somewhat based on statistical averages that 20K might be a more risky bet across the consumer base, but it's by no means a given.  As a result, since neither assets nor income are factored into the scoring, some average debt statistics don't make much sense.

That all said, I have the sneaking impression that within the next decade FICO will change their algorithms to look more at total debt than utilization: utilization metrics made more sense before outsized limits.  Also FICO's exclusion of it is why it's only one part of the underwriting process... where they absolutely they do factor in income and debt.  

---

I never said there weren't exceptions.  In fact, I said outliers exist with all statistics.  The fact remains that everyone that can pay a $20k debt can also pay a $200 debt, but not everyone that can pay a $200 debt can pay a $20k debt.  In my eyes, it's really that simple.  Risk of not being able to pay a $20k debt > risk of not being able to pay a $200 debt based on the numbers alone.  Add a zero to the numbers and they are even more glaringly obvious as your example even proves.  More people can pay a $2000 debt than a $200k debt; even a mortgage with some decent equity potentially couldn't get someone out of $200k.  But $2000 for them?  Probably no issue.

[Thomas_Thumb]

---

@Anonymous wrote:

Why would I (or you) even need a data set to make such an obvious assumption?  I don't see how this is opinion.  I'm sure there's a data set out there that shows that a $20k debt is more risky than a $200 debt and that far more people would struggle paying back $20k over $200.  Anyone that has the capacity to pay back $20k obviously can pay back $200.  The converse certainly cannot be said, so by definition more people are able to pay back a $200 debt than a $20k debt.

I'm sure there's a data set out there that says that 3 years after someone achieves a doctorate level of education their average salary is greater than someone else 3 years after achieving just a high school GED.  I don't have access to that data set, but honestly I don't need to see it to know that the result of it is likely pretty cut and dry.  Are there outliers/exceptions to all data sets and statistics?  Of course.  That goes without saying.   

---

 Big data shows utilization is a valid predictor of creditworthiness. Factors and weighting are based on a population as a whole and also segmented populations. Generally speaking many statistical models don't factor in true outliers.

Relative to the subject line, it's a bit broader than an example designed to push a certain point of view. As I said, utilization has been found to matter quite a bit and is weighed more heavily (per the above chart) than debt. Now should utilization look at monthly spend relative and monthly payments in addition to balance on a statement? I'd say yes, but that's just my opinion.

[sarge12]

---

@Thomas_Thumb wrote:

---

@Anonymous wrote:

Why would I (or you) even need a data set to make such an obvious assumption?  I don't see how this is opinion.  I'm sure there's a data set out there that shows that a $20k debt is more risky than a $200 debt and that far more people would struggle paying back $20k over $200.  Anyone that has the capacity to pay back $20k obviously can pay back $200.  The converse certainly cannot be said, so by definition more people are able to pay back a $200 debt than a $20k debt.

I'm sure there's a data set out there that says that 3 years after someone achieves a doctorate level of education their average salary is greater than someone else 3 years after achieving just a high school GED.  I don't have access to that data set, but honestly I don't need to see it to know that the result of it is likely pretty cut and dry.  Are there outliers/exceptions to all data sets and statistics?  Of course.  That goes without saying.   

---

 Big data shows utilization is a valid predictor of creditworthiness. Factors and weighting are based on a population as a whole and also segmented populations. Generally speaking many statistical models don't factor in true outliers.

Relative to the subject line, it's a bit broader than an example designed to push a certain point of view. As I said, utilization has been found to matter quite a bit and is weighed more heavily (per the above chart) than debt. Now should utilization be calculated according to monthly spend as opposed to a point in time balance on a statement? I'd say yes, but that's just my opinion

---

This is probably true, but only because the general population has not learned to game the system as members on these forums have by constantly increasing credit limits. Many of the educated on these forums have 300k and 500k credit limits, which makes low utilization very easy. If the public as a whole ever starts doing the same, utilization as a risk predictor will not be effective anymore.

[Anonymous]

---

@sarge12 wrote:

---

---

This is probably true, but only because the general population has not learned to game the system as members on these forums have by constantly increasing credit limits. Many of the educated on these forums have 300k and 500k credit limits, which makes low utilization very easy. If the public as a whole ever starts doing the same, utilization as a risk predictor will not be effective anymore.

---

Spot on - I agree completely.  What do the members of this forum and/or people well versed in this knowledge represent, a fraction of 1% of the population?

Perhaps my perspective since being part of the fraction is skewed a bit, so for me utilization doesn't seem like a very important risk factor.  I think calculating utilization based on monthly spend or of any number of the suggestions throughout this thread were incorporated we'd get a much better risk representation score that would not just appeal to the majority, but those of thus in the very small sector of the population that understand the ins and outs of utilization and scoring.

[sarge12]

---

@Anonymous wrote:

---

@sarge12 wrote:

---

---

This is probably true, but only because the general population has not learned to game the system as members on these forums have by constantly increasing credit limits. Many of the educated on these forums have 300k and 500k credit limits, which makes low utilization very easy. If the public as a whole ever starts doing the same, utilization as a risk predictor will not be effective anymore.

---

Spot on - I agree completely.  What do the members of this forum and/or people well versed in this knowledge represent, a fraction of 1% of the population?

Perhaps my perspective since being part of the fraction is skewed a bit, so for me utilization doesn't seem like a very important risk factor.  I think calculating utilization based on monthly spend or of any number of the suggestions throughout this thread were incorporated we'd get a much better risk representation score that would not just appeal to the majority, but those of thus in the very small sector of the population that understand the ins and outs of utilization and scoring.

---

Yes, truth is, most of the general public waits until their running too close to their credit limit to even request a cli. Most of us on this forum, however get cli for the sole purpose of being able to easily maintain a very low utilization...but we are the vast minority. I request credit limits on cards, and as a general rule I never exceed 10% utilization on any card. If I have a 10k limit, I only charge 1k on that card, with very few occasional exceptions for a large purchase. My low utilization is largely being controled by asking for CLI's until they are high enough to make low utilization easy. I also will make multiple payments to make sure utilization is always low if needed. The general public does not even understand the way utilization works, and may run utilization up to 80%, and PIF before due date, and then wonder why their score is so low when they apply for credit, never realizing they are allowing 80% utilization to report every month...even if they PIF. It is a game....I play it well, but within the rules!!!!