> Right, which drives up transaction volume which in turn means the payment network and the seller get more volume and more money.
It only drives up transaction volume for high risk transactions, which is bad because those are the ones that increase the amount of fraud.
> The risk is accounted for in interchange.
The risk is magnified from fraud taking advantage of the insurance. You have to pay more because the insurance is a deep pocket to steal from with weak ability to protect itself because the actual parties are no longer vested in preventing it. Or the cost of the "insurance payout" is paid by the merchant who knows that they're innocent but the payment processor has no way to know that.
> The bigger your risk pool the lower your cost. That's how insurance works. The more people you can socialize big losses over, the less you have to charge each person.
That's not how insurance works at all. You buy insurance because you and some other people each have a 0.1% risk of losing $100,000 and you would each rather pay $100 with 100% probability than $100,000 with 0.1% probability. For this to work it makes no difference whether there are a million other people or a billion. It only matters if you get down to something like 1000 people and even then your expected value is still the same, but there is a greater chance that your premium ends up as either $200 or $0 instead of $100. But we're not talking about population numbers that low here anyway.
And the socialized losses are the issue, because that's a moral hazard. Once you're insured you take bad risks and rely on the insurance to eat it, which raises costs for everybody.
When you have to insure a $100K+ loss you can't afford to suffer, you're willing to eat that overhead. But to insure a $100 loss? It makes no sense. You're going to make enough $100 purchases yourself that you can "self insure" cheaper than an insurance company with bureaucratic overhead and moral hazard can do it for you. It costs less to lose your modest purchase price a small percentage of the time than to pay a higher than that percentage for the insurance. And the amount you lose buying the insurance is inherently more than the amount you lose of your own on average unless the insurance company is losing money.
> Your margin on top is a function of your business goals. It's not strictly true that more people equals more money, you can always pass on the costs. Depends on where you make your money. A 501(c)(3) that offers insurance for instance would just charge less.
If your net margins per transaction as an insurer are a given percentage, you make more money by processing more transactions unless that percentage is zero or negative. A non-profit may purposely have zero margins, but then in what sense is it "in their interest" to have greater volume? (And how many non-profit major banks are you aware of?)
> 3% for card present is actually really high, I was using a blended average of the ~2.5ish% charged for card-present, 3ish% charged for online and 3.5ish% charged for card-not-present transactions at the point of sale to small and medium sized businesses. You can expect this to be 1% lower for merchants of substantial scale.
You're implying that "merchants of substantial scale" are paying an average of around 2%, or around 1.5% for card present transactions, for cards paying 2% cash back. That seems a bit fishy, doesn't it?
Square is using "let us mush all this together and average it out for you" pricing, but that only really works if they reject merchants with high chargeback rates etc., and still requires them to raise their rates if cards that give higher rewards by charging higher fees become more popular.
Meanwhile most other payment processors are going to impose chargeback fees on top of that, which is what we're really trying to avoid here -- plus the major cost that isn't in any of these numbers which is the cost of lost merchandise and labor when there is a fraudulent chargeback for goods and services already rendered.
> Again, see [2].
Right, so the merchant is paying 2.9% to Square, but the poor customer has a 0% cash back card because with their credit history it's all they can do to get a credit card offer at all. Then too many of the merchant's other low income customers commit fraud and they get kicked off of Square and have to suffer the high end of "between 2.87 percent and 4.35 percent" from whichever of Square's competitors will take them.
This customer might be quite pleased to precommit to not doing a chargeback with a trustworthy merchant in exchange for a ~4% discount.
I’m not quite following you here. Are you suggesting that there is an alternative to the current system that allows buyers and sellers to opt out of the ~2% fee built into prices to cover the cost of fraud?
And that this system is currently precluded by law?
As I understand it, merchants are now allowed to offer a lower price for cash purchases. The fact that so many merchants use credit card systems seems to speak to their marginal value.
My off-the-cuff analysis is that there is a certain amount of fraud that is inevitable with any non-cash transaction, especially those that allow chargebacks.
The credit card companies have tried to enforce the lack of a cash discount because it enables them to spread the costs across virtually the entire system of users. Without the ability to do this, I don’t see how a system as widely adopted as Visa and MasterCard could have got off the ground. This is what makes it work.
If we assume the cost of fraud is independent of the structures that determine who bears the cost (a big assumption) then the argument is going to be about how that cost is apportioned among users.
Since the system has been adopted, overwhelmingly, by both buyers and sellers, it seems reasonable to conclude that it has been a (huge) value add to the whole economic system.
If you allowed participants to use the system but opt out of paying what is essentially the insurance policy against fraud, the total cost of fraud in the system can’t go down without the gross level of transactions going down.
You seem to be saying that there should be a tiered, opt out system of fees. Buyers wanting to use their card in high risk (of fraud or chargebacks) industries, like porn, should bear the cost of providing the fraud insurance buy paying higher fees?
It begs the question why such a service is not being provided by the market. Essentially, credit card companies have come to the conclusion of providing such a service is not profitable.
In a sense, they are offering an “opt out” for the extreme cases like porn, by making it hard for the porn providers to offer credit card services at all.
Forcing a universal fee structure on everyone provides two huge advantages:
It facilitates the build-out of the network itself, which provides tremendous value to all users.
And it lowers the total cost of determining what the actual cost of fraud insurance should be. This cost would fall across all parties.
The card providers would have to build the industrial infrastructure to understand the costs at a more granular level. It would require both capital investment and ongoing costs to manage such a system. It would also incentivize non-productive activity around trying to “game” the system. Essentially it would be a new attack surface.
It would also move some of the costs of evaluating the risk of a given purchase onto the consumer. This would result in a large net increase in total costs as individuals who lack the skill and scale to properly do this would be forced into it. This would also result in a massive “duplication of effort.”
Intuitively, I think this would result in a large decline in overall economic activity, as it simply would not be worth the investment to make certain kinds of purchases.
The current system puts the costs of fraud mitigation into the entity who can provide it at the lowest cost.
There is also of course the simple argument that the credit card system is opt-in. It is possible to conduct business with cash or check. I use a tech who repairs speakers that I collect. He adds 3% to use Paypal. I opt to send him a check.
You can also send cash through the mail. In this instance, a transaction where the seller won’t release the product until they receive the funds provides the kind of transaction you want, where the entire risk of fraud (or loss) is borne by the buyer.
I think most small businesses at least would accept cash in person or the mail.
It’s not ideal, but my guess is that how the net cost of fraud protection with electronic transactions is distributed is pretty close to ideal.
This is just one of the many problems that render crypto-currency systems to be hopelessly expensive. The lack of ability to build in “socialized fraud protection cost” will prevent mass adoption.
They also suffer from the problem that the cost of securing the network should rise in a system lacking inflation as the total value of transactions increase.
The game theoretical costs of providing the minimum of security a proof-of-work system offers are hard to parse. But the current cost of bitcoin transactions is outrageously high, when accounting for mining cost.
IMO, the bitcoin system is being supported by a combination of outright fraud, speculative investment, and money laundering.
The real costs land in a very disproportionate manner, for example, when exchanges go belly up, or when an individual is scammed out a large sum.
It only drives up transaction volume for high risk transactions, which is bad because those are the ones that increase the amount of fraud.
> The risk is accounted for in interchange.
The risk is magnified from fraud taking advantage of the insurance. You have to pay more because the insurance is a deep pocket to steal from with weak ability to protect itself because the actual parties are no longer vested in preventing it. Or the cost of the "insurance payout" is paid by the merchant who knows that they're innocent but the payment processor has no way to know that.
> The bigger your risk pool the lower your cost. That's how insurance works. The more people you can socialize big losses over, the less you have to charge each person.
That's not how insurance works at all. You buy insurance because you and some other people each have a 0.1% risk of losing $100,000 and you would each rather pay $100 with 100% probability than $100,000 with 0.1% probability. For this to work it makes no difference whether there are a million other people or a billion. It only matters if you get down to something like 1000 people and even then your expected value is still the same, but there is a greater chance that your premium ends up as either $200 or $0 instead of $100. But we're not talking about population numbers that low here anyway.
And the socialized losses are the issue, because that's a moral hazard. Once you're insured you take bad risks and rely on the insurance to eat it, which raises costs for everybody.
When you have to insure a $100K+ loss you can't afford to suffer, you're willing to eat that overhead. But to insure a $100 loss? It makes no sense. You're going to make enough $100 purchases yourself that you can "self insure" cheaper than an insurance company with bureaucratic overhead and moral hazard can do it for you. It costs less to lose your modest purchase price a small percentage of the time than to pay a higher than that percentage for the insurance. And the amount you lose buying the insurance is inherently more than the amount you lose of your own on average unless the insurance company is losing money.
> Your margin on top is a function of your business goals. It's not strictly true that more people equals more money, you can always pass on the costs. Depends on where you make your money. A 501(c)(3) that offers insurance for instance would just charge less.
If your net margins per transaction as an insurer are a given percentage, you make more money by processing more transactions unless that percentage is zero or negative. A non-profit may purposely have zero margins, but then in what sense is it "in their interest" to have greater volume? (And how many non-profit major banks are you aware of?)
> 3% for card present is actually really high, I was using a blended average of the ~2.5ish% charged for card-present, 3ish% charged for online and 3.5ish% charged for card-not-present transactions at the point of sale to small and medium sized businesses. You can expect this to be 1% lower for merchants of substantial scale.
You're implying that "merchants of substantial scale" are paying an average of around 2%, or around 1.5% for card present transactions, for cards paying 2% cash back. That seems a bit fishy, doesn't it?
Square is using "let us mush all this together and average it out for you" pricing, but that only really works if they reject merchants with high chargeback rates etc., and still requires them to raise their rates if cards that give higher rewards by charging higher fees become more popular.
Meanwhile most other payment processors are going to impose chargeback fees on top of that, which is what we're really trying to avoid here -- plus the major cost that isn't in any of these numbers which is the cost of lost merchandise and labor when there is a fraudulent chargeback for goods and services already rendered.
> Again, see [2].
Right, so the merchant is paying 2.9% to Square, but the poor customer has a 0% cash back card because with their credit history it's all they can do to get a credit card offer at all. Then too many of the merchant's other low income customers commit fraud and they get kicked off of Square and have to suffer the high end of "between 2.87 percent and 4.35 percent" from whichever of Square's competitors will take them.
This customer might be quite pleased to precommit to not doing a chargeback with a trustworthy merchant in exchange for a ~4% discount.