Credit Score is Predicative. But should it be used?

[whatyouwantfinally]

but it would be if you were using inaccurate driving/accident history for a specific driver. The insurance scoring systems are based upon perfect & timely data. But the insurance scoring is pulled from mostly from credit score information from the big three & those are full of errors, inaccuracy & fat fingered data from closed/transferred accounts.

Sure, but if it was that inaccurate it wouldn't be predictive. Yet is is predictive.

Go pull address data from the data sellers. With our hot real estate markets of people moving, selling, etc, the majority of address checks for licensed drivers are not even close to accurate as to who is currently living in the house as a potential driver. So, imagine all the data inaccuracy of the credit scoring systems

That's not true. The majority are accurate.

I dont have a problem with accurate verifiable use of data, but have seen a ton of widowed women sitting on millions in the bank with no debt who get horrible insurance scores because they either never used debt or the deceased husband had the debt/credit in his name.

I'm not going to disagree that it isn't perfect. But the people you are talking abut are the exception, not the norm. I just don't see how the existence of statistical outliers (which exist in any probability model) translates to unfairness.

 

[Allen Trent]

I just don't see how the existence of statistical outliers (which exist in any probability model) translates to unfairness.

so, it would be ok if some outliers get charged auto premiums with a DUI listed on their driving record as long as most dont? the reason this never happens is the carrier is required by law to list the exact dates/convictions of driving violations & CLUE accidents. the insured than can see those & dispute when they are inaccurate. This isnt the case for credit based insurance scores. the individual that has their insurance score drop between renewals is not told what exact items impacted their score & corresponding discount level. they are merely provided with a vague list of a couple possible generic categories "reason codes" that could have caused the "adverse action"(too much/not enough). It really should be required that a specific account # caused the change

again, I benefit substantially from it, so not crying about it, just think there is room for better disclosure & improved accuracy. I have nearly max insurance score discounts, but I have found dozens of mistakes on my very, very limited account information & it takes months & numerous attempts at follow up to get it corrected with all 3 reporting agencies..................only to then have those same mistakes pop up again 2 years later when I run them again. most consumers would give up.

PS-- in regard to accuracy, this article show 34% of Americans have at least 1 mistake & insurance scores are based off these credit reports. So, sure, not a "majority" by definition, but massively higher than some fractional statistical anomaly. Keep in mind, the vast majority of Americans never check their credit reports. https://www.cnbc.com/2021/06/11/how-to-fix-those-mistakes-on-your-credit-report.html

 

[shawnmwalker]

@adjusterjack Insurtech is a place where we talk about the religion for three days a year: ITC Vegas 2021 (insuretechconnect.com)

 

[shawnmwalker]

See the schedule of topics discussed for three days. Pretty robust topics on the religion of Insurtech! "Get the Net!"- Waynes World

ITC Vegas 2021 (insuretechconnect.com)

armovieguys - Wayne's World - Are you mental? - YouTube

 

[whatyouwantfinally]

so, it would be ok if some outliers get charged auto premiums with a DUI listed on their driving record as long as most dont? the reason this never happens is the carrier is required by law to list the exact dates/convictions of driving violations & CLUE accidents. the insured than can see those & dispute when they are inaccurate. This isnt the case for credit based insurance scores. the individual that has their insurance score drop between renewals is not told what exact items impacted their score & corresponding discount level. they are merely provided with a vague list of a couple possible generic categories "reason codes" that could have caused the "adverse action"(too much/not enough). It really should be required that a specific account # caused the changel

A DUI is going to affect an insurance premium FAR more than, say, a wrongly reported late payment 2 years ago, so it's not really a fair comparison. I see sense in carriers having to be more specific, but again, I don't see it as unfair. If they told you the specific items, you'd still be on your own to deal with it.

again, I benefit substantially from it, so not crying about it, just think there is room for better disclosure & improved accuracy. I have nearly max insurance score discounts, but I have found dozens of mistakes on my very, very limited account information & it takes months & numerous attempts at follow up to get it corrected with all 3 reporting agencies..................only to then have those same mistakes pop up again 2 years later when I run them again. most consumers would give up.

That's kind of what I mean. Despite inaccuracies, you have a maxed out score. That wouldn't be the case with tickets/accidents/DUIs.

PS-- in regard to accuracy, this article show 34% of Americans have at least 1 mistake & insurance scores are based off these credit reports. So, sure, not a "majority" by definition, but massively higher than some fractional statistical anomaly. Keep in mind, the vast majority of Americans never check their credit reports. https://www.cnbc.com/2021/06/11/how-to-fix-those-mistakes-on-your-credit-report.html

But, like in your case, a large number of those inaccuracies don't have a significant effect on scores. The % of people who have bad credit due to inaccuracies is really low.

So, yes, I think better transparency in credit scoring would be good. I don't think the problem is near the level of considering its use in insurance scoring unfair.

 

[Allen Trent]

A DUI is going to affect an insurance premium FAR more than, say, a wrongly reported late payment 2 years ago, so it's not really a fair comparison. I see sense in carriers having to be more specific, but again, I don't see it as unfair. If they told you the specific items, you'd still be on your own to deal with it.



That's kind of what I mean. Despite inaccuracies, you have a maxed out score. That wouldn't be the case with tickets/accidents/DUIs.



But, like in your case, a large number of those inaccuracies don't have a significant effect on scores. The % of people who have bad credit due to inaccuracies is really low.

So, yes, I think better transparency in credit scoring would be good. I don't think the problem is near the level of considering its use in insurance scoring unfair.

fair enough

in my case, I believe there are about 4 levels of insurance score better than mine out of about 20 scoring levels. those 4 better levels amount to only about 10% greater discount at the top, but at $4000 in annual premium, that $400 difference seems like it could be explained.

the roof credit, matured insured credit, how far from fire dept or fire hydrant credit is very easy to confirm. how far you drive, your driving record, multi policy credit are straight forward & easy to confirm. Credit based insurance scoring can be from 0-80% credit amongst some carriers, so you would think that would owe some sort of verifiable explanation to the carrier/agent/consumer

 

[shawnmwalker]

Washington Insurance Commissioner Pursues January Date for Rule to Ban Use of Credit Scores (insurancejournal.com)

More articles at the bottom of the link explaining more.

 

[shawnmwalker]

@adjusterjack The two in the middle used to be DOI commissioners, If I remember correctly. The two on the outside, I believe sell data to carriers. If Insurance scores go away....I assume they'll have much to gain.

 

[shawnmwalker]

Some guys comments off of Reddit where I posted the same topic:

"The problem with predictive analytics and all data analysis is confirmation bias. What makes a credit score predictive? If individuals with lower credit scores already live in high claims areas, then their zip code is likely the problem and not their credit score. You can find a correlation between almost any two data sets, but that doesn't mean there's any factual causation. The credit scoring systems are terribly archaic and need a complete overhaul. That's not to say just hand people money and lines of credit etc. but I think by nature we always want to have a solution to a problem, and it leads us to potentially create problems that weren't actually there."

 

[scagnt83]

This isnt the case for credit based insurance scores. the individual that has their insurance score drop between renewals is not told what exact items impacted their score & corresponding discount level. they are merely provided with a vague list of a couple possible generic categories "reason codes" that could have caused the "adverse action"(too much/not enough). It really should be required that a specific account # caused the change

This is not necessarily true. MetLife Auto sent me a premium increase notice and listed the reason as my credit history. Specifically, "excessive amount of credit cards for credit profile". I had just opened up a new card through a store I shop at frequently, making it my 4th credit card, 2 normal cards & 2 store cards.

So in that instance, MetLife did not say exactly which account caused it, but it was obvious to me which one it was from the reason listed.

Ironic thing about that, my actual credit score increased because of that new card. But my insurance score dropped because of it. (at least my insurance score with metlife)

Now, perhaps that was because of MetLife's own claims experience with people in those circumstances. But they definitely lost a client over it since Allstate was less expensive after that increase. (and their online interface was stone-age quality so I had already considered switching)

----

I agree that there is a lot of ambiguity in the insurance scoring system. But imo, its up the the insurers how to utilize the info. It would benefit them from a customer service level and trust building level to be more transparent about it. But its on them to do so or not.

Im not sure the areas that generalize and create unfairness are really that huge of an issue. But I am a life/annuity agent, so I certainly could be wrong. I thought it was messed up that my premiums were raised, but Im also much happier with my current carrier in general, so it seems like in my case the free market worked properly.

Had MetLife taken the effort to ask about the specifics of the cards, they would have realized I really just had 2 traditional credit cards that were being used on any regular basis. So maybe their analytics are not advanced enough yet to discern the situation down to that level. Or maybe they dont care. But it seems they could have been more thorough in their review. Ultimately it was their loss. But I do see how this could possibly adversely effect people in more complicated insurance situations.