[Karen37a]

And that is the true reason behind all of this.

You are not getting the sales away fromt the smart brokers who bring the client to you.

People always make up hairbrained schemes to create business when they are failing.


(And brokers don't because of the renewals)

This regulation argument is the low of the low though.

[CreditGuy]

Credit Guy,

I am not sure if you do credit but you do know that credit reports can vary 30-50 points depending on what version of software you are using within the same Credit Reporting Agency?

One lender can have one score, another a lower or higher score, even though its the same Credit Bureau on the same day? That could also affect the calculation.

I know one big lender using old software, I do not know if they know they are using it (blindly) or purposely using it to generate the lower score .

Everyone would have to use the same exact version of credit scoring software to get a TRUE apples to apple comparison. I guess the big credit reporting agencies will have to do something to stop the discrepancy in the scores . Lets Regulate them too !!!

Yes, I "do" credit and am aware there are various versions of a credit score provided by one bureau. Let me give you a brief lesson in credit bureaus, credit scoring, and model development.

To start, the software you are using has nothing to do with the score you get back. Because credit bureaus deal in PII, they have to know who is requesting the data and for what purpose they will be using the data. This classification is the first segmentation of users. What credit data from the bureau is exposed to a user is a function of who they are and for what they will be using it. An insurance company, landlord, and bank all get exposed to different levels of data in terms of details and records. Within a bank, the bank might get different scores based on the product for which you are applying. Once the bureau has segmented you out, they can determine which score version you get as a function of the data to which you are exposed. A landlord might not get a score at all, but get any public records or collections items that are open against their prospective tenant. A bank on the other hand gets everything and the latest and greatest score (FICO 8, usually). (If they want it. More on this in a second.)

Even within this space, a "lender" running a two man money shop on Long Island isn't going to get the same score version that a big lender like Kabbage is getting since the bureaus are going to treat the small shop more like a consumer or landlord and give them less and/or less granular data, a less granular score, etc.. The saying "you get what you pay for" holds for credit bureaus. If you only want to sign up online and pay $2 a pull, you are going to get the $2 quality credit bureau. If you go through a more stringent vetting process and pay more, you are going to have access to more robust data and be able to pick the score version most suitable to your line of business. Also, I would note here that within each bureau there are specialized score versions for large consumer segments like cards (Bankcard FICO), mortgage (Beacon 5.0) and auto (Auto FICO). Depending on which score version and level of data you want, you will pay different costs per inquiry. Mortgage uses the oldest model, for reasons we will talk about now. FICO 9 was just released, but nobody really uses it yet because it is too new.

Credit scoring. In lending, the purpose of scoring credit is to build a logit that expresses probability of default/likelihood to repay and then bound your credit policy/risk appetite to a subset within that range to create a favorable ROI across your portfolio of loans. Since we are are talking about FICO/consumer credit, let's stick with that. FICO ranges from 300-850 with lower values indicating a higher likelihood of default. For most of the lenders in our industry, 500 is the absolute floor picked arbitrarily and/or because everyone else is doing it. This is because by and large, borrowers with sub-500 FICOs more often than not default (pD>.50). So, how did the first person figure this out? Effective model development.

When modeling credit risk you want to prove that 500 FICO has a certain default rate and/or is an inflection point. In order to do that, you have to generate observations/loans and see how they perform over multiple vintages. The life cycle is: 1) pick a variable (or variables) you think are going to predict default, 2) make loans with that/those assumption(s), 3) compare actual performance to your score predictions/defaults in the real world. Rinse and repeat ad infinitum. When you are building models this way, you want to try to use the same data sources as inputs for purposes of model integrity. I'm guessing this is why the one big lender you mention is using an old version of the score. Their model was either copied from another industry/model that used the old score version, or they chose X score version when they started building their model from scratch and rather than rebuild or recalibrate the model every time a new version of the score they are using comes out they just use the legacy score. Chances are though that anyone that is being classified as a "big lender" has folks that know exactly why and what they are doing using the older score. This is also why the mortgage industry uses an old version (Beacon 5.0) of FICO. Their models are more complex, require more observations to build samples sets, and are harder to roll forward to a new version of FICO score, which provides many very highly weighted variables in their scoring models.

If you really want to make the argumentum ad absurdum that any variation in inputs used by lenders negates any attempts at standardization of outputs to consumers, that is you prerogative, but I would warn all those reading this that she hasn't the foggiest clue about credit scoring, setting baselines, regulating an industry at scale, and what appears to many other fundamentals about how this business of credit risk management really works.

[Karen37a]

LOL...very long post I will read it later.

I am assuming you are agreeing with me that Algorithms are a moving target and hard to quantify on a message board.

Everyone has their own personal agenda no matter how they dress it up. And mine is to go make money and not debate things that people will never change or change them to suit their personal needs.

[CreditGuy]

I am assuming you are agreeing with me that Algorithms are a moving target and hard to quantify on a message board.

Quite the opposite. I'm not even sure you know what the word algorithm means.

[Karen37a]

I am going to underwriting school soon, I am sure ill find out. ( sarcasm )


The FICO scores used for mortgages

Equifax Beacon 5.0
Experian/Fair Isaac Risk Model V2SM
TransUnion FICO, Risk Score, Classic 04
These are sometimes listed on the tri-merge mortgage report as follows:

Equifax/FICO classic V5 FACTA
Experian/Fair Isaac (Ver. 2)
Transunion/FICO Classic (04)


EQ 04 aka FICO 5 (on the myFICO report)
EX 98 aka FICO 2 (on the myFICO report)
TU 04 aka FICO 4 (on the myFICO report)


*** I am not doing Mortgages anymore so check that they didnt change

Banks/Lenders can use Fico or Vantage scores and their proprietary system

[HDF]

Equally relevent:

relevent.png

[staten]

lol.... go Karen. Hey credit guy-smart people usually only have to use small words.

[CreditGuy]

Your user name perfectly explains why you think I use big words.

[J.Celifarco]

Unbelievable that some people are still begging for this government to impose regulations on their business.

Its not begging for it, its realizing that it is definitely going to come and better to be involved and help to steer it so it does the least amount of damage. Sticking our heads in the sand and pretending nothing is going on isn't going to help anyone. The only way to survive long term is to be part of the solution

[Karen37a]

Eleanor Roosevelt said to stop it... from the grave


"Great minds discuss ideas; average minds discuss events; small minds discuss people."-Eleanor Roosevelt

[channin19]

You actually believe there is a large group of people in the industry that would self impose regulation and think big picture instead of how much money can I make this second this minute???? I wish you were correct but there is no way the majority of the industry would agree to anything unless it was imposed by the government

Far from it. I'm not as active as I once was on the board but I view most of the characters in this industry in the same light as you. I do think that a move like this was needed but I'm not sure cost was the right place to start. It gives a very definitive ceiling for regulators to start from. When I speak to merchants cost is one part of the discussion What they don't like is dealing with Tony the retired pizza delivery boy hitting them up every five minutes with a 1.45 over 3 months telling them they are a piece of ****.

[staten]

Arrogant and condescending.....lol I don't happen to live in Staten Island, quite far from it, that said, some of the brightest people I've met in this industry are from there. Wanna play the experience and education game? You lost. I don't have the same insecurity issues as you. mwuh!

[CreditGuy]

Meh.

[J.Celifarco]

Far from it. I'm not as active as I once was on the board but I view most of the characters in this industry in the same light as you. I do think that a move like this was needed but I'm not sure cost was the right place to start. It gives a very definitive ceiling for regulators to start from. When I speak to merchants cost is one part of the discussion What they don't like is dealing with Tony the retired pizza delivery boy hitting them up every five minutes with a 1.45 over 3 months telling them they are a piece of ****.

I agree also that is not the place to start. i personally think Barrier to entry is the biggest problem we have.. If it took a little work and knowledge to enter the industry you wouldn't have half the issues we do.. Right now anyone with 100 bucks to buy some UCC and a phone can call themselves an ISO and thats a problem.

[WhoisKingsley]

ILPA created the SMART Box. CAN is a member, but so are Kabbage and On Deck. Between the three of them, they make up a very large majority of dollars and units funded in this industry. That's a significant network effect and signal to borrower. If a broker presents two deals: one from CAN and another deal from ABC Capital, and the CAN one has the SMART box clearly showing cost of funds and total payback and the other is just a contract and they have to hunt for the terms, which are they going to feel more comfortable with? How long until a borrower says, "Your contract doesn't have a SMART Box so I can easily compare terms." ILPA is hoping that is a day that comes soon.

Breakout Capital is not a member of ILPA and doesn't use SMART Box, FYI.

When it comes to the Borrower/customer/client - the feature SMART Box includes (or any other movement)can make the difference. Take your head away from Wall Street for a second and think Main Street- Think like a person who has been subjected to mainstream ways of research and obtaining something. Whether it is a product (material) or a financial product, there are visual and simple features that give a customer the feeling of security in their decisions- something many direct and in-direct companies CANNOT give.

Regulation is not only disclosing information- it is providing a unified process. A unified process is something that is not likely to happen, thus the multiple groups trying to align this industry whether it's for self regulation or not. It goes without saying the the low barrier of entry along with the high level of enabling is one of the biggest reasons of why nothing will be done without a "boom". It will likely cater to the deeper pockets and banks because that's how 'Murica works.

[Karen37a]

As Stated REgulation is a bad bad bad thing>>>And that is the true reason behind all of this.

You are not getting the sales away fromt the smart brokers who bring the client to you.

People always make up hairbrained schemes to create business when they are failing.


(And brokers don't because of the renewals)

This regulation argument is the low of the low though.

[soscap]

As Stated REgulation is a bad bad bad thing>>>And that is the true reason behind all of this.

You are not getting the sales away fromt the smart brokers who bring the client to you.

People always make up hairbrained schemes to create business when they are failing.


(And brokers don't because of the renewals)

This regulation argument is the low of the low though.

In my opinion, regulation would be the best thing... 1. More institutional money. 2. Less fraud 3. Regulation brings checks and balances which = less stupidity and more funders staying in business 4. No more of those shady brokers.

[Carminooch]

In my opinion, regulation would be the best thing... 1. More institutional money. 2. Less fraud 3. Regulation brings checks and balances which = less stupidity and more funders staying in business 4. No more of those shady brokers.

Agree...this is such a saturated market. I came from an investment advisory world where there is probably too much regulation but no doubt about it, this industry is desperate for it. I run into too many incompetents on a daily basis that do not belong in any industry like this and would have been kicked out a long time ago if they were a broker/financial adviser, reason why they are in this industry probably to begin with.

[soscap]

Yes, I "do" credit and am aware there are various versions of a credit score provided by one bureau. Let me give you a brief lesson in credit bureaus, credit scoring, and model development.

To start, the software you are using has nothing to do with the score you get back. Because credit bureaus deal in PII, they have to know who is requesting the data and for what purpose they will be using the data. This classification is the first segmentation of users. What credit data from the bureau is exposed to a user is a function of who they are and for what they will be using it. An insurance company, landlord, and bank all get exposed to different levels of data in terms of details and records. Within a bank, the bank might get different scores based on the product for which you are applying. Once the bureau has segmented you out, they can determine which score version you get as a function of the data to which you are exposed. A landlord might not get a score at all, but get any public records or collections items that are open against their prospective tenant. A bank on the other hand gets everything and the latest and greatest score (FICO 8, usually). (If they want it. More on this in a second.)

Even within this space, a "lender" running a two man money shop on Long Island isn't going to get the same score version that a big lender like Kabbage is getting since the bureaus are going to treat the small shop more like a consumer or landlord and give them less and/or less granular data, a less granular score, etc.. The saying "you get what you pay for" holds for credit bureaus. If you only want to sign up online and pay $2 a pull, you are going to get the $2 quality credit bureau. If you go through a more stringent vetting process and pay more, you are going to have access to more robust data and be able to pick the score version most suitable to your line of business. Also, I would note here that within each bureau there are specialized score versions for large consumer segments like cards (Bankcard FICO), mortgage (Beacon 5.0) and auto (Auto FICO). Depending on which score version and level of data you want, you will pay different costs per inquiry. Mortgage uses the oldest model, for reasons we will talk about now. FICO 9 was just released, but nobody really uses it yet because it is too new.

Credit scoring. In lending, the purpose of scoring credit is to build a logit that expresses probability of default/likelihood to repay and then bound your credit policy/risk appetite to a subset within that range to create a favorable ROI across your portfolio of loans. Since we are are talking about FICO/consumer credit, let's stick with that. FICO ranges from 300-850 with lower values indicating a higher likelihood of default. For most of the lenders in our industry, 500 is the absolute floor picked arbitrarily and/or because everyone else is doing it. This is because by and large, borrowers with sub-500 FICOs more often than not default (pD>.50). So, how did the first person figure this out? Effective model development.

When modeling credit risk you want to prove that 500 FICO has a certain default rate and/or is an inflection point. In order to do that, you have to generate observations/loans and see how they perform over multiple vintages. The life cycle is: 1) pick a variable (or variables) you think are going to predict default, 2) make loans with that/those assumption(s), 3) compare actual performance to your score predictions/defaults in the real world. Rinse and repeat ad infinitum. When you are building models this way, you want to try to use the same data sources as inputs for purposes of model integrity. I'm guessing this is why the one big lender you mention is using an old version of the score. Their model was either copied from another industry/model that used the old score version, or they chose X score version when they started building their model from scratch and rather than rebuild or recalibrate the model every time a new version of the score they are using comes out they just use the legacy score. Chances are though that anyone that is being classified as a "big lender" has folks that know exactly why and what they are doing using the older score. This is also why the mortgage industry uses an old version (Beacon 5.0) of FICO. Their models are more complex, require more observations to build samples sets, and are harder to roll forward to a new version of FICO score, which provides many very highly weighted variables in their scoring models.

If you really want to make the argumentum ad absurdum that any variation in inputs used by lenders negates any attempts at standardization of outputs to consumers, that is you prerogative, but I would warn all those reading this that she hasn't the foggiest clue about credit scoring, setting baselines, regulating an industry at scale, and what appears to many other fundamentals about how this business of credit risk management really works.

Color me impressed! Not to many on this forum of your caliber.