Does Zip Code Really Matter?

[TrueGeminiNC]

For those who think this is discriminatory - I would disagree.  

From my experience in the Insurance industry - premiums were driven by multiple components including your zip code.  This data element was supposedly utilized to help assess risk.  Also, credit itself is a controversial data element that is being used by Insurers.  

Conversely, why would a credit card company not use this same logic for risk assessment.  Additionally, I know for a fact that AMEX sells its cardholder information (e.g. zip code, income, etc.) to companies who consume this data for multiple reasons.  So - I wouldn't be surprised if credit card companies utilize zip codes in their algorithims to help predictt risks.

[UncleB]

---

@TrueGeminiNC wrote:

For those who think this is discriminatory - I would disagree.  

From my experience in the Insurance industry - premiums were driven by multiple components including your zip code.  This data element was supposedly utilized to help assess risk.  Also, credit itself is a controversial data element that is being used by Insurers.  

Conversely, why would a credit card company not use this same logic for risk assessment.  Additionally, I know for a fact that AMEX sells its cardholder information (e.g. zip code, income, etc.) to companies who consume this data for multiple reasons.  So - I wouldn't be surprised if credit card companies utilize zip codes in their algorithims to help predictt risks.

---

It's not quite the same.

With auto insurance, where you live (i.e. where you keep and drive your car) is rightfully used to help determine insurance rates because the actions of others in your area can impact claims.  Simply put, if you live in an area that is statistically known to have drivers with more accidents, your rate is (and should be) higher, because you are more likely to have a claim - even if by no fault of your own.

On the other hand, the people in my neighborhood/street/ZIP/etc. don't have any impact on how I pay my bills.  It can be argued that an area with more job losses could be an indicator, but that's not the question here.  Many people live in one ZIP, and work in another... sometimes several counties away.

I've already included a link to an older story about when Amex got caught "redlining" customers based on the merchants they frequented, and the outcry was loud enough that they supposedly ended the practice.  As to if they use an applicant's ZIP in consideration of a new account approval or a CLI, who knows (but Amex) but the practice is shady at best and provides questionable results at worst.

[TrueGeminiNC]

---

@UncleB wrote:

---

@TrueGeminiNC wrote:

For those who think this is discriminatory - I would disagree.  

From my experience in the Insurance industry - premiums were driven by multiple components including your zip code.  This data element was supposedly utilized to help assess risk.  Also, credit itself is a controversial data element that is being used by Insurers.  

Conversely, why would a credit card company not use this same logic for risk assessment.  Additionally, I know for a fact that AMEX sells its cardholder information (e.g. zip code, income, etc.) to companies who consume this data for multiple reasons.  So - I wouldn't be surprised if credit card companies utilize zip codes in their algorithims to help predictt risks.

---

It's not quite the same.

With auto insurance, where you live (i.e. where you keep and drive your car) is rightfully used to help determine insurance rates because the actions of others in your area can impact claims.  Simply put, if you live in an area that is statistically known to have drivers with more accidents, your rate is (and should be) higher, because you are more likely to have a claim - even if by no fault of your own.

On the other hand, the people in my neighborhood/street/ZIP/etc. don't have any impact on how I pay my bills.  It can be argued that an area with more job losses could be an indicator, but that's not the question here.  Many people live in one ZIP, and work in another... sometimes several counties away.

I've already included a link to an older story about when Amex got caught "redlining" customers based on the merchants they frequented, and the outcry was loud enough that they supposedly ended the practice.  As to if they use an applicant's ZIP in consideration of a new account approval or a CLI, who knows (but Amex) but the practice is shady at best and provides questionable results at worst.

---

From a risk assesment perspective  - you can apply this same logic to the credit card industry e.g. credit card companies analyzing default rates on a regional if not more granular level as a risk mitigateion strategy.  I think to some degree - all industries take advantage of similar concepts.  

[UncleB]

---

@TrueGeminiNC wrote:

---

@UncleB wrote:

---

@TrueGeminiNC wrote:

For those who think this is discriminatory - I would disagree.  

From my experience in the Insurance industry - premiums were driven by multiple components including your zip code.  This data element was supposedly utilized to help assess risk.  Also, credit itself is a controversial data element that is being used by Insurers.  

Conversely, why would a credit card company not use this same logic for risk assessment.  Additionally, I know for a fact that AMEX sells its cardholder information (e.g. zip code, income, etc.) to companies who consume this data for multiple reasons.  So - I wouldn't be surprised if credit card companies utilize zip codes in their algorithims to help predictt risks.

---

It's not quite the same.

With auto insurance, where you live (i.e. where you keep and drive your car) is rightfully used to help determine insurance rates because the actions of others in your area can impact claims.  Simply put, if you live in an area that is statistically known to have drivers with more accidents, your rate is (and should be) higher, because you are more likely to have a claim - even if by no fault of your own.

On the other hand, the people in my neighborhood/street/ZIP/etc. don't have any impact on how I pay my bills.  It can be argued that an area with more job losses could be an indicator, but that's not the question here.  Many people live in one ZIP, and work in another... sometimes several counties away.

I've already included a link to an older story about when Amex got caught "redlining" customers based on the merchants they frequented, and the outcry was loud enough that they supposedly ended the practice.  As to if they use an applicant's ZIP in consideration of a new account approval or a CLI, who knows (but Amex) but the practice is shady at best and provides questionable results at worst.

---

From a risk assesment perspective  - you can apply this same logic to the credit card industry e.g. credit card companies analyzing default rates on a regional if not more granular level as a risk mitigateion strategy.  I think to some degree - all industries take advantage of similar concepts.  

---

Actually, I don't see how the same logic can be applied.

With car insurance, other people impact your rates because they have the ability to (literally) crash into you, or even vandalize your vehicle.  Nobody in my ZIP code has the ability to impact if I pay my credit card bill or not.

We'll just have to 'agree to disagree'.

[TrueGeminiNC]

That is the point - it is not meant to isolate or drill down to the individual level. This would be a discramanotory.   Regardless if you pay your credit card bill on time or not - I am certain all credit card companies look at areas where they have high default rates.  

You may pay your credit card bill on time - however, others who live within your area may not.  Unfortunately, similar to insurace - you become a casualty regardless if you are a good driver or not.  

And too - lets be clear, I am not saying this is a singular data element - but, I do beleive this data element coupled with others does drive credit card company decisions.

[UncleB]

---

@TrueGeminiNC wrote:

That is the point - it is not meant to isolate or drill down to the individual level. This would be a discramanotory.   Regardless if you pay your credit card bill on time or not - I am certain all credit card companies look at areas where they have high default rates.  

You may pay your credit card bill on time - however, others who live within your area may not.  Unfortunately, similar to insurace - you become a casualty regardless if you are a good driver or not.  

And too - lets be clear, I am not saying this is a singular data element - but, I do beleive this data element coupled with others does drive credit card company decisions.

---

I understand that it's not a singular data element, but where your analogy breaks down is in the comparison to auto insurance.

With auto insurance you can become a casualty due to someone else's choices, negligence, etc.  Your rate is partially assigned based on a 'speculation' (albeit very educated) as to what your chances are of encountering one of your 'neighbors' on the road.

There is no similar 'encounter' with my neighbors that can impact how I pay my bills. 

I understand you're saying that creditors have the ability to do this, and I'm not arguing with that - they have the 'ability' to do many things.  I'm just saying that logically it makes no real 'sense', and any company that has business practices that allow for this had better be ready for the 'blow back' if word gets around, just as Amex found out the hard way a few years ago.

[Anonymous]

---

@UncleB wrote:

---

@TrueGeminiNC wrote:

That is the point - it is not meant to isolate or drill down to the individual level. This would be a discramanotory.   Regardless if you pay your credit card bill on time or not - I am certain all credit card companies look at areas where they have high default rates.  

You may pay your credit card bill on time - however, others who live within your area may not.  Unfortunately, similar to insurace - you become a casualty regardless if you are a good driver or not.  

And too - lets be clear, I am not saying this is a singular data element - but, I do beleive this data element coupled with others does drive credit card company decisions.

---

I understand that it's not a singular data element, but where your analogy breaks down is in the comparison to auto insurance.

With auto insurance you can become a casualty due to someone else's choices, negligence, etc.  Your rate is partially assigned based on a 'speculation' (albeit very educated) as to what your chances are of encountering one of your 'neighbors' on the road.

There is no similar 'encounter' with my neighbors that can impact how I pay my bills. 

I understand you're saying that creditors have the ability to do this, and I'm not arguing with that - they have the 'ability' to do many things.  I'm just saying that logically it makes no real 'sense', and any company that has business practices that allow for this had better be ready for the 'blow back' if word gets around, just as Amex found out the hard way a few years ago.

---

But remember that auto insurance credit score is used in rate setting in many states (those where it is not illegal) and equally there is no direct connection between that and your liklihood to be an accident.    With big data, companies can claim that things correlate, and certainly zip/region and default might.

[Anonymous]

---

@heychrisbaker wrote:

---

@Anonymous wrote:

ZIP code only matters as a method of identification.

---

What do you mean by "identification"?  

---

I mean it in the "Are you who you say you are" sense. One method of verification is whether your supplied information matches what's on your report. Addresses and employers are listed on your CR. 

[UncleB]

---

@Anonymous wrote:

---

@UncleB wrote:

---

@TrueGeminiNC wrote:

That is the point - it is not meant to isolate or drill down to the individual level. This would be a discramanotory.   Regardless if you pay your credit card bill on time or not - I am certain all credit card companies look at areas where they have high default rates.  

You may pay your credit card bill on time - however, others who live within your area may not.  Unfortunately, similar to insurace - you become a casualty regardless if you are a good driver or not.  

And too - lets be clear, I am not saying this is a singular data element - but, I do beleive this data element coupled with others does drive credit card company decisions.

---

I understand that it's not a singular data element, but where your analogy breaks down is in the comparison to auto insurance.

With auto insurance you can become a casualty due to someone else's choices, negligence, etc.  Your rate is partially assigned based on a 'speculation' (albeit very educated) as to what your chances are of encountering one of your 'neighbors' on the road.

There is no similar 'encounter' with my neighbors that can impact how I pay my bills. 

I understand you're saying that creditors have the ability to do this, and I'm not arguing with that - they have the 'ability' to do many things.  I'm just saying that logically it makes no real 'sense', and any company that has business practices that allow for this had better be ready for the 'blow back' if word gets around, just as Amex found out the hard way a few years ago.

---

But remember that auto insurance credit score is used in rate setting in many states (those where it is not illegal) and equally there is no direct connection between that and your liklihood to be an accident.    With big data, companies can claim that things correlate, and certainly zip/region and default might.

---

I do remember... and that's exactly my point. 

Correlation doesn't necessarily equal causation.  If anything, when my credit was bad I was more careful since I knew any accident would cause a significant financial burden, yet in some states (mine included) companies are allowed to use credit data to determine rates.

I'm not arguing that it can't be done, I'm simply saying it shouldn't be done, since there's no true causation (i.e. cause and effect).  We have laws against charging folks different rates based on race, since it makes no difference in reality regardless of what 'big data' says.  However, if you were to crunch enough numbers you'll find that one race likely has a "statistically" higher rate of safer drivers... should they get a discount?  Of course not.  This is why you can't always just rely on 'statistics'.

If somebody were to be really motivated, they could crunch enough numbers to prove how safe left handed drivers are compared to right handed drivers, and adjust their rates accordingly... "the data would support it".  Would this be fair - or even factually accurate?  Of course not.

That's my point with using where somebody lives to determine their interest rate, AA, CLD, etc.  If you 'torture' the data enough you can basically create justification to take any action you want... if it makes actual business sense do so is another matter.

[Anonymous]

---

@UncleB wrote:

---

@Anonymous wrote:

---

@UncleB wrote:

---

@TrueGeminiNC wrote:

That is the point - it is not meant to isolate or drill down to the individual level. This would be a discramanotory.   Regardless if you pay your credit card bill on time or not - I am certain all credit card companies look at areas where they have high default rates.  

You may pay your credit card bill on time - however, others who live within your area may not.  Unfortunately, similar to insurace - you become a casualty regardless if you are a good driver or not.  

And too - lets be clear, I am not saying this is a singular data element - but, I do beleive this data element coupled with others does drive credit card company decisions.

---

I understand that it's not a singular data element, but where your analogy breaks down is in the comparison to auto insurance.

With auto insurance you can become a casualty due to someone else's choices, negligence, etc.  Your rate is partially assigned based on a 'speculation' (albeit very educated) as to what your chances are of encountering one of your 'neighbors' on the road.

There is no similar 'encounter' with my neighbors that can impact how I pay my bills. 

I understand you're saying that creditors have the ability to do this, and I'm not arguing with that - they have the 'ability' to do many things.  I'm just saying that logically it makes no real 'sense', and any company that has business practices that allow for this had better be ready for the 'blow back' if word gets around, just as Amex found out the hard way a few years ago.

---

But remember that auto insurance credit score is used in rate setting in many states (those where it is not illegal) and equally there is no direct connection between that and your liklihood to be an accident.    With big data, companies can claim that things correlate, and certainly zip/region and default might.

---

I do remember... and that's exactly my point. 

Correlation doesn't necessarily equal causation.  If anything, when my credit was bad I was more careful since I knew any accident would cause a significant financial burden, yet in some states (mine included) companies are allowed to use credit data to determine rates.

I'm not arguing that it can't be done, I'm simply saying it shouldn't be done, since there's no true causation (i.e. cause and effect).  We have laws against charging folks different rates based on race, since it makes no difference in reality regardless of what 'big data' says.  However, if you were to crunch enough numbers you'll find that one race likely has a "statistically" higher rate of safer drivers... should they get a discount?  Of course not.  This is why you can't always just rely on 'statistics'.

If somebody were to be really motivated, they could crunch enough numbers to prove how safe left handed drivers are compared to right handed drivers, and adjust their rates accordingly... "the data would support it".  Would this be fair - or even factually accurate?  Of course not.

That's my point with using where somebody lives to determine their interest rate, AA, CLD, etc.  If you 'torture' the data enough you can basically create justification to take any action you want... if it makes actual business sense do so is another matter.

---

I guess I would disagree then!   If it were proved that left-handed drivers had a poorer safety record (living in a right-hand dominated world) in what sense would it not be fair to charge them more (why should everyone else have to pay the costs).    Like race, it's not their fault or choice, but neither is more poor eye/hand coordination that makes me more likely to hit things.  Now "protected class" rules can prevent discrimination, but that is social policy and doesn't mean the correlation or analysis is wrong.