Utilization - per card or overall?

[Anonymous]

---

@Thomas_Thumb wrote:

I wouldn't be surprised if Fico 9 and VS 3 both have subroutines that incorporate the T/R feature - but the subroutine was "turned off" when rolled out for general scoring. Alternatively, - these models may have been set-up to allow future inclusion as a bolt on addition. Could be a 3rd party application.

Here is an interesting paper I just came across on T-R modeling which appears well thought out and robust.

http://support.sas.com/resources/papers/proceedings15/3217-2015.pdf

---

Awesome paper, Thom Thumb!  I work with a lot of SAS users so this is of great interest to me.  I have not read it yet but certainly intend to.

There is no date on the paper, but after doing a little bit of detective work I believe it looks like the paper was presented at the SAS Global Forum 2015 (April 26-29, 2015).  THere would have been a vetting and review period before hand, so a fair guess is that the paper was written during the summer of 2014 (or summer/fall).  So that gives folks here a sense of how recent it is. 

Interesting idea about FICO 9 having T-R scoring subroutines that have been turned off temporarily.  Very possible.  It's worth remembering that Vantage 3.0 was released almost three years ago (I saw a fact sheet for the model dated March 2013).  So it's unclear whether Vantage would have had time to incorporate this new T-R technology in the model by then.  Though by your idea, maybe they have since been building the capacity for VS 3 to enable a T-R module to be plugged in, either their own or someone else's.

[Thomas_Thumb]

Here is an Experian article that you may also find interesting. I don't personally subscribe to trending - time based weighted averages do the job better

http://www.experian.com/assets/consumer-information/white-papers/6985-cis-ts-whitepaper-exp2014.pdf

[Aahz]

---

@Anonymous wrote:

That's a great question, Aahz.  Chaouki's claim does not imply that the T/R distinction is part of current FICO models, though I can see why you might think it does.

It only would if FICO 8 was the only tool on the market for evaluating consumer credit..  But it isn't.  TransUnion, for example, makes and sells its own tools, and according to the NYT piece, Fannie Mae has partnered with TU and is going to be incorporating TU's trended data analysis technology into FM's flagship underwriting product ("Desktop Underwriter").

There's also a legitimate question to raise about whether Chaouki (who works for TU) might be exagerating the extent to which his company's technology has penetrated into widespread use.  Remember he's a VP of a unit that sells of financial services.

---

You're absolutely right! I fell into the dual trap of focusing too much on FICO and not critically examining Chaouki's use of the PR-pretty, but highly vague, term "popular".

While I've only skimmed over the SAS article Thomas posted (lack of finance/economics/statistics education makes it hard for me to digest) something I found there has brought me back to your earlier assertions on page 1 of this thread.  To wit-

---

@Anonymous wrote:

Your credit reports now record what your balance was every month for many months in a row.  (Until recently, they only recorded your most recent balance.)  More than that, for each month, your reports record how much of that balance you later paid.  That means that future lenders will be able to see whether you tend to always PIF.  People who adopt that style are known as transactors.  At the other end of the spectrum are people who rarely PIF.  These people only pay a part of the total amount owed.  They carry the remainder over to the next month.  (This is called carrying a balance.)  People who often carry a balance are known as revolvers.

---

My question is regarding the part I highlighted in red.

Full reports (as opposed to those visible on MF) do indeed show how much you paid on an account in any given month in the "Actual Payment Amount" field, but there does not seem to be any differentiation between the amount being paid before or after the statement cuts.  Making your distinction betwen the two uses of PIF meaningless for scoring purposes (FICO or otherwise).  In fact, the SAS article opens the section titled "General Model Setup" with-

---

At the high level credit card holder can be non-active, active, delinquent and defaulted. Active and nondelinquent credit cards holders are split up into two groups: revolvers and transactors. Revolver is user who carries a positive credit card balance and not pay off the balance in full each month – roll over. Transactor is user who pays in full on or before the due date of the interest-free credit period. Competent user does not incur any interest payments or finance charges.

---

While they do not further define the term inactive, can't we assume that the having any figure in the "Actual Payment Amount" field of a credit report would negate inactive status regardless of the value in the "Balance" field? Doesn't this imply that there would be no difference between your two PIF definitions when it comes to determining T/R status as well?

If, as I believe, general consensus holds that letting most cards report $0 balance is better for FICO scores and when the payment is made is not relevant in determining T/R status, then wouldn't one still be better off paying balances before they report?  Thus achieveing both lowest possible utility and Transactor status.

[Aahz]

---

@Thomas_Thumb wrote:

Here is an Experian article that you may also find interesting. I don't personally subscribe to trending - time based weighted averages do the job better

http://www.experian.com/assets/consumer-information/white-papers/6985-cis-ts-whitepaper-exp2014.pdf

---

Great find, Thomas!  And much easier to follow then the SAS paper, IMO.  

From page four of this Experian paper-

---

Experian conducted an analysis that grouped consumers based on how they paid their credit card balances - minimum payers ("MIN Payers") paid up to five percent of their balance, "STEADY payers" paid between five percent and 99 percent of their balance and "TRANSACTOR payers" paid their balances off in full.

As indicated in Figure 4, MIN payers have limited capacity to assume additional payment obligations on credit cards.  On the other hand, TRANSACTOR and STEADY payers are less likely to default as they take on more card debt.  This key piece of information is not captured in today's generic risk models. 

---

 This would indicate that there is more to the Revolver/Transactor breakdown with "Revolver" being split into both Min and Steady payers, with Steady being far more attractive to lenders.  This is excellent news for those of us who carry balances (whether as a result of need or 0% opportunities).

Traditionally I've focused on paying at least double the minimum payment when carrying balances.  But with this insight I now plan to ensure I'm paying at least 5% when choosing not to PIF.  While full Transactor status is clearly the gold standard (with ~10% default), being a Steady payer only increases the default rate to around 14% as opposed to the 22%-23% rate for Min payers.

[iv]

---

@Thomas_Thumb wrote:

Here is an Experian article that you may also find interesting. I don't personally subscribe to trending - time based weighted averages do the job better

http://www.experian.com/assets/consumer-information/white-papers/6985-cis-ts-whitepaper-exp2014.pdf

---

That is interesting... but I'd like to highlight a slightly different part of the whitepaper.

While T/R history data is somewhat useful for risk management, the expanded payment history data is FAR more useful for profit mangement.  Take a look at the graphs on page 10, and the quote at the top of the page:

Thirty-two percent of the U.S. population generates very little income for lenders, yet this population continues to receive offers for expensive rewards programs along with high lines of credit. A unique opportunity exists to allocate marketing dollars toward acquiring the most profitable segments. Lenders will be able to invest heavily into rewards and lines for high-yield high- spending segments, while limiting offers to other populations.

Figure 12 and the associated notes show the use of payment data history to target high credit lines and good rewards to high-spending transacting consumers (to get them to switch primary cards), and lower APRs to low-spending revolvers (to get revolving balances, as they said, "captured" onto their cards).  Notice that this high/low spend analysis is across all revolving lines, not limited to those of the lender doing the analysis... which might explain why some lenders still aren't reporting full payment data (Amex/Discover), or stopped reporting full data (Chase).

[Thomas_Thumb]

Additional food for thought - As some say, it's all about the money.

New frontiers in credit card segmentation - McKinsey & Company

[Anonymous]

---

@Aahz wrote:

---

@Thomas_Thumb wrote:

Here is an Experian article that you may also find interesting. I don't personally subscribe to trending - time based weighted averages do the job better

http://www.experian.com/assets/consumer-information/white-papers/6985-cis-ts-whitepaper-exp2014.pdf

---

Great find, Thomas!  And much easier to follow then the SAS paper, IMO.  

From page four of this Experian paper-

---

Experian conducted an analysis that grouped consumers based on how they paid their credit card balances - minimum payers ("MIN Payers") paid up to five percent of their balance, "STEADY payers" paid between five percent and 99 percent of their balance and "TRANSACTOR payers" paid their balances off in full.

As indicated in Figure 4, MIN payers have limited capacity to assume additional payment obligations on credit cards.  On the other hand, TRANSACTOR and STEADY payers are less likely to default as they take on more card debt.  This key piece of information is not captured in today's generic risk models. 

---

 This would indicate that there is more to the Revolver/Transactor breakdown with "Revolver" being split into both Min and Steady payers, with Steady being far more attractive to lenders.  This is excellent news for those of us who carry balances (whether as a result of need or 0% opportunities).

Traditionally I've focused on paying at least double the minimum payment when carrying balances.  But with this insight I now plan to ensure I'm paying at least 5% when choosing not to PIF.  While full Transactor status is clearly the gold standard (with ~10% default), being a Steady payer only increases the default rate to around 14% as opposed to the 22%-23% rate for Min payers.

---

Great points, Aahz.

The NYT article alludes to this too and earlier in this thread I gave exactly this advice to Jlynn0819.  I.e. that if she wants to continue to revolve on her 0% account, then she'll likely be scored much more leniently by the R-T module if she is always paying substantially more than the minimum payment and if additionally her balance is steadily going down over time.

[Anonymous]

---

@Aahz wrote:

---

@Anonymous wrote:

That's a great question, Aahz.  Chaouki's claim does not imply that the T/R distinction is part of current FICO models, though I can see why you might think it does.

It only would if FICO 8 was the only tool on the market for evaluating consumer credit..  But it isn't.  TransUnion, for example, makes and sells its own tools, and according to the NYT piece, Fannie Mae has partnered with TU and is going to be incorporating TU's trended data analysis technology into FM's flagship underwriting product ("Desktop Underwriter").

There's also a legitimate question to raise about whether Chaouki (who works for TU) might be exagerating the extent to which his company's technology has penetrated into widespread use.  Remember he's a VP of a unit that sells of financial services.

---

You're absolutely right! I fell into the dual trap of focusing too much on FICO and not critically examining Chaouki's use of the PR-pretty, but highly vague, term "popular".

While I've only skimmed over the SAS article Thomas posted (lack of finance/economics/statistics education makes it hard for me to digest) something I found there has brought me back to your earlier assertions on page 1 of this thread.  To wit-

---

@Anonymous wrote:

Your credit reports now record what your balance was every month for many months in a row.  (Until recently, they only recorded your most recent balance.)  More than that, for each month, your reports record how much of that balance you later paid.  That means that future lenders will be able to see whether you tend to always PIF.  People who adopt that style are known as transactors.  At the other end of the spectrum are people who rarely PIF.  These people only pay a part of the total amount owed.  They carry the remainder over to the next month.  (This is called carrying a balance.)  People who often carry a balance are known as revolvers.

---

My question is regarding the part I highlighted in red.

Full reports (as opposed to those visible on MF) do indeed show how much you paid on an account in any given month in the "Actual Payment Amount" field, but there does not seem to be any differentiation between the amount being paid before or after the statement cuts.  Making your distinction betwen the two uses of PIF meaningless for scoring purposes (FICO or otherwise).  In fact, the SAS article opens the section titled "General Model Setup" with-

---

At the high level credit card holder can be non-active, active, delinquent and defaulted. Active and nondelinquent credit cards holders are split up into two groups: revolvers and transactors. Revolver is user who carries a positive credit card balance and not pay off the balance in full each month – roll over. Transactor is user who pays in full on or before the due date of the interest-free credit period. Competent user does not incur any interest payments or finance charges.

---

While they do not further define the term inactive, can't we assume that the having any figure in the "Actual Payment Amount" field of a credit report would negate inactive status regardless of the value in the "Balance" field? Doesn't this imply that there would be no difference between your two PIF definitions when it comes to determining T/R status as well?

If, as I believe, general consensus holds that letting most cards report $0 balance is better for FICO scores and when the payment is made is not relevant in determining T/R status, then wouldn't one still be better off paying balances before they report?  Thus achieveing both lowest possible utility and Transactor status.

---

You write:

"Full reports (as opposed to those visible on MF) do indeed show how much you paid on an account in any given month in the "Actual Payment Amount" field, but there does not seem to be any differentiation between the amount being paid before or after the statement cuts."

I'd be very surprised if the Actual Payment Amount data are based on calendar months (e.g. July 1-31, Nov 1-30, etc.).  Rather, just in the same way that the balances are associated with a reporting period (typically the end of a billing cycle), I'd assume the Actual Payment Data is associated also with reporting cycles.  Thus, suppose my Citi Double Cash reports on Nov 10 with a balance of $400 based on a cyycle that ended on Nov 9.  Then the Actual Payment Data will look at how much payments I made Nov 10-Dec 9 -- i.e. in the following billing cycle.  If you do it that way, then Actual Payment Data always refers to what happens after the reported balance.

It's also possible that every single payment is recorded along with the date that payment is made.

You ask later:

"If, as I believe, general consensus holds that letting most cards report $0 balance is better for FICO scores...."

It's important to bear in mind that nobody claims that you get any long term benefit out of keeping most cards reporting a balance of $0.  That's a pure ephemeral snapshot aspect of scoring.  So there's really no need to keep all or most of your cards at $0.  Every tiny bit of advantage you get from that you can obtain from zeroing them out in the month before an important credit pull.

Here (as I see it) are the advantages to paying in full AFTER the statement cut:

(1)  It is really easy.  You can just set your cards up on auto pay.  Boom.  You're done.  PTZ (pay to zero) takes more work, and you have to make sure you don't use your card between the time you pay and the beginning of the next cycle.

(2)  We know that PIF (after statement) is in full keeping with what the T-R distinction models are looking for.  We know it will give you the full benefit of being a Transactor. This is more than we can say about PTZ.  Maybe PTZ will give you the full benefit of being a transactor -- maybe not.  Until we do know it just seems simpler to continue establishing your T history via the PIF route.

[NRB525]

At the end of that Experian Consumer Information White Paper, there's an interesting EX product offering: ERIC for Revolving, to estimate the interest paid by and APR of a consumer.

Interesting set of links folks. Thanks.

[Anonymous]

---

@Anonymous wrote:

---

@Aahz wrote:

---

@Anonymous wrote:

That's a great question, Aahz.  Chaouki's claim does not imply that the T/R distinction is part of current FICO models, though I can see why you might think it does.

It only would if FICO 8 was the only tool on the market for evaluating consumer credit..  But it isn't.  TransUnion, for example, makes and sells its own tools, and according to the NYT piece, Fannie Mae has partnered with TU and is going to be incorporating TU's trended data analysis technology into FM's flagship underwriting product ("Desktop Underwriter").

There's also a legitimate question to raise about whether Chaouki (who works for TU) might be exagerating the extent to which his company's technology has penetrated into widespread use.  Remember he's a VP of a unit that sells of financial services.

---

You're absolutely right! I fell into the dual trap of focusing too much on FICO and not critically examining Chaouki's use of the PR-pretty, but highly vague, term "popular".

While I've only skimmed over the SAS article Thomas posted (lack of finance/economics/statistics education makes it hard for me to digest) something I found there has brought me back to your earlier assertions on page 1 of this thread.  To wit-

---

@Anonymous wrote:

Your credit reports now record what your balance was every month for many months in a row.  (Until recently, they only recorded your most recent balance.)  More than that, for each month, your reports record how much of that balance you later paid.  That means that future lenders will be able to see whether you tend to always PIF.  People who adopt that style are known as transactors.  At the other end of the spectrum are people who rarely PIF.  These people only pay a part of the total amount owed.  They carry the remainder over to the next month.  (This is called carrying a balance.)  People who often carry a balance are known as revolvers.

---

My question is regarding the part I highlighted in red.

Full reports (as opposed to those visible on MF) do indeed show how much you paid on an account in any given month in the "Actual Payment Amount" field, but there does not seem to be any differentiation between the amount being paid before or after the statement cuts.  Making your distinction betwen the two uses of PIF meaningless for scoring purposes (FICO or otherwise).  In fact, the SAS article opens the section titled "General Model Setup" with-

---

At the high level credit card holder can be non-active, active, delinquent and defaulted. Active and nondelinquent credit cards holders are split up into two groups: revolvers and transactors. Revolver is user who carries a positive credit card balance and not pay off the balance in full each month – roll over. Transactor is user who pays in full on or before the due date of the interest-free credit period. Competent user does not incur any interest payments or finance charges.

---

While they do not further define the term inactive, can't we assume that the having any figure in the "Actual Payment Amount" field of a credit report would negate inactive status regardless of the value in the "Balance" field? Doesn't this imply that there would be no difference between your two PIF definitions when it comes to determining T/R status as well?

If, as I believe, general consensus holds that letting most cards report $0 balance is better for FICO scores and when the payment is made is not relevant in determining T/R status, then wouldn't one still be better off paying balances before they report?  Thus achieveing both lowest possible utility and Transactor status.

---

You write:

"Full reports (as opposed to those visible on MF) do indeed show how much you paid on an account in any given month in the "Actual Payment Amount" field, but there does not seem to be any differentiation between the amount being paid before or after the statement cuts."

I'd be very surprised if the Actual Payment Amount data are based on calendar months (e.g. July 1-31, Nov 1-30, etc.).  Rather, just in the same way that the balances are associated with a reporting period (typically the end of a billing cycle), I'd assume the Actual Payment Data is associated also with reporting cycles.  Thus, suppose my Citi Double Cash reports on Nov 10 with a balance of $400 based on a cyycle that ended on Nov 9.  Then the Actual Payment Data will look at how much payments I made Nov 10-Dec 9 -- i.e. in the following billing cycle.  If you do it that way, then Actual Payment Data always refers to what happens after the reported balance.

It's also possible that every single payment is recorded along with the date that payment is made.

You ask later:

"If, as I believe, general consensus holds that letting most cards report $0 balance is better for FICO scores...."

It's important to bear in mind that nobody claims that you get any long term benefit out of keeping most cards reporting a balance of $0.  That's a pure ephemeral snapshot aspect of scoring.  So there's really no need to keep all or most of your cards at $0.  Every tiny bit of advantage you get from that you can obtain from zeroing them out in the month before an important credit pull.

Here (as I see it) are the advantages to paying in full AFTER the statement cut:

(1)  It is really easy.  You can just set your cards up on auto pay.  Boom.  You're done.  PTZ (pay to zero) takes more work, and you have to make sure you don't use your card between the time you pay and the beginning of the next cycle.

(2)  We know that PIF (after statement) is in full keeping with what the T-R distinction models are looking for.  We know it will give you the full benefit of being a Transactor. This is more than we can say about PTZ.  Maybe PTZ will give you the full benefit of being a transactor -- maybe not.  Until we do know it just seems simpler to continue establishing your T history via the PIF route.

---

Everyone is making the assumption that its somehow "difficult" to asertain Transactor behavior when one uses PTZ payment method. Nothing could be further from the truth. The method of detecting Transactor behavior is to compare the difference between payments reported and previous months statement balance (Payments - Balance). If the value of payments minus previous balance is zero or any positive number then you are a "full transactor", but if the number is a negative value, then depending on the (payment/payment - balance) ratio, then one is a Revolver or Revolver/Transactor combination.