Reported balances vs starting limit

[BrutalBodyShots]

---

@Thomas_Thumb wrote:

---

---

A model that discriminates based on a transactor vs revolver behavior must be rigorous and without inference. If not, the institution is subject to litigation. Given statement balance is the only field reported for a specific account, certainty of T/R behavior appears elusive short of $0 balance reporting. Other data, such as payments, must be provided.

If payments and credits equal or exceed previous balance => PIF transactor.

---

@Thomas_Thumb - I still think you aren't getting what I'm talking about at all. This has nothing to do with discrimination, nor can it lead to litigation. If payment history is reported, the issuer would know flat out if someone is a Transactor or Revolver. No inferring. They can then use that information as they see fit when it comes to lending decisions. If (say) 90% of that information can be inferred from looking at reports (by a human being or by an algorithm designed by a human being), how would that lead to anything of the sort that you are implying?

Are you legitimately trying to tell me that if you look at the reported balances (not payments) on any given credit report you can't infer Transactor or Revolver behavior with those numbers?

Cornelius has reported balances over 12 months that are as follows:

$325, $790, $620, $111, $990, $1210, $30, $544, $330, $891, $72, $400

Rupert has reported balances over 12 months that are as follows:

$325, $390, $355, $411, $370, $320, $358, $399, $310, $344, $379, $322

Looking at those reported balances with no payment history details at all, what would you infer?

[Thomas_Thumb]

Rupert might be making mid cycle payments so reported monthly balance stay below a self imposed ceiling - and/or the account is primarily designated for regular reoccurring monthly charges that are allowed to report naturally. Alternatively Rupert could be making minimum payments but still using the card sparingly.

Cornelius may let balances report naturally with no micromanaging and/or may be use the account more sporadically for purchasing higher cost non essentials.

In neither case can I ascertain if one or both are transactors or revolvers.

I show both types of balance reporting. My PIF Fidelity account is primarily used for monthly medical insurance premiums and reports a rather consistent balance month over month. My PIF Discover card is used for discretionary spend and living expenses. I let balances report naturally but sometimes pay current balance  a few days before due date vs just statement balance. As a result reported balances fluctuate a lot with Discover.

Payment data is essential to prove my transactor behavior month over month and year over year - or possibly the minimum due as a whole number "tell".

Trended data can certainly enhance predictive risk analysis independent of T/R behavior. Such data analysis often applies some type of time weighted averaging for smoothing. The most recent data point(s) have the greatest impact on the average. The weighted average (say from the most recent 6 months) is then plotted and trended. For example:

B(avg) = 0.4B1 + 0.25B2 + 0.15B3 + 0.1B4 + 0.06B5 + 0.04B6

Where B is reported balance, (1) means current month and (2) is last month.

Is the plotted average trending up, remaining steady within boundry limits or trending down? 

Applying the above weighted averaging to C and R balances yields trended averages now => 6 months ago of:

C: 539, 660, 568, 521, 812, 662

R: 356, 380, 369, 377, 353, 344

C: 12 month avg bal = 526

R: 12 month avg bal = 357

Neither is showing a significant upward trend in balance reporting. Current weighted average is near respective 12 month average for both.

[FicoMike0]

Interesting debate. It seems to assume that revolver behavior represents higher risk. I agree that would seem likley, but I don't see any evidence that any of the score models use it as a factor. Fico and vs have had the opportunity to data mine millions of historic reports to extract such a corrolation. They seem to use utilization as a proxy. Since they make money developing risk models, a lot of it, wouldn't they include it if it were highly predictive?

[BrutalBodyShots]

---

@FicoMike0 wrote:

Interesting debate. It seems to assume that revolver behavior represents higher risk. I agree that would seem likley, but I don't see any evidence that any of the score models use it as a factor. Fico and vs have had the opportunity to data mine millions of historic reports to extract such a corrolation. They seem to use utilization as a proxy. Since they make money developing risk models, a lot of it, wouldn't they include it if it were highly predictive?

---

I don't think anyone debates the fact that those that carry balances are an elevated risk to default relative to those that pay in full monthly. As both TT and I agreed up thread, there's no way currently without payment history information reported by all issuers to be absolutely certain, so there's no way to currently include it in the algorithm. I'm sure if that data were available universally that it would definitely be included since it is in fact highly predictive. Almost everyone that defaults on their debts was carrying balances first, where it's extremely rare for a strict Transactor to do so.

@Thomas_Thumb I think a better way to frame this debate is to take pure "Transactor vs Revolver" from the discussion and take it back to what you can best infer from the data. You can infer without question that Cornelius using the numbers I provided up thread is making large payments monthly. It's no secret and obvious looking at the numbers. When looking at Rupert's numbers, you don't know. Sure you can come up with an example like you gave where perhaps Rupert is still paying in full (or nearly) but you know far less looking at his data set relative to that of Cornelius.

I don't see it as a black and white discussion. Transactor vs Revolver. There are tons of gray areas. For example, someone that pays "just the minimum" monthly on a (say) $1000 carried balance is a Revolver, but the highest risk Revolver there is. Now if they pay $400/mo? They are still a Revolver, but exhibiting less risky use of their revolving credit relative to the person only paying the minimum. If they pay $900 on the carried $1000 balance? Even less risk. 

So again, I agree that the pure distinction between Transactor and Revolver can't be determined at all times from just reported balance data without payment history data... but I do think there's plenty of data there in many cases to deduce far more than you're giving credit for.

[SRT4kid93]

@BrutalBodyShots 

Can you give your opinion on the "whole number" point I made earlier?

if someone were carrying a balance their min payments would be random numbers. Not only that, but as the interest continues to be tacked on, their minimum payments would continue to either 

1) continue in an upward trend if they are making min payments (since the interest may be more than they are paying)

2) if their payments are big enough to offset the interest, the numbers may not continue in an upward path, but they would certainly continue to be weird, not whole numbers 

[Vinjints]

Minimum payments are designed to cover the interest, no? Making minimum payments, without adding additional charges should show a slight decrease each month in the amount owed. 

[SRT4kid93]

No minimum payments attack only the principle. It doesn't even start to cover interest. And if it does, it's a very small percentage 

it depends on how high the balance is and therefore  how much interest but I've seen hundreds if not thousands of examples of people making only the minimum and every month they owe more than the month before. 

it's what people mean when they say "drowning in credit card debt" or you "can never get ahead"

[BrutalBodyShots]

@SRT4kid93, I agree with what you said regarding minimum payments, whole numbers, etc. That's a very good point and gauge that I hadn't considered prior so I appreciate you sharing it.

[Vinjints]

Your hundreds and thousands of examples are completely backwards. It takes years to payoff credit card debt (drowning in debt) because the bulk of your money is going to interest and fees rather than the principal. 

[SRT4kid93]

@Vinjints 

**palm to my face**

you are absolutely right, i totally mixed that up. apologies. 

yes, when making minimum payments you arent making a dent in the actual debt. 

Either way. It is very possible for your debt to go up even while making min payments. Your payments need to be large enough to cover the interest accrued.