I my part of the EU I'd certainly feel like I'm missing out if I just kept on paying the ~€5 a month fee for a bank account and didn't look for a card that gave me some kind of return on my spend every month...
(It's Finland btw, and you need a Finnish, not "anywhere-in-SEPA", bank account in order to practically function in society here with the strong online authentication service)
No, it's taking the merchants' money and giving you back part of it. How much do merchants eat it vs passing on the cost? I don't think that's an easy question to answer and probably varies a lot by merchant type and product, but I think we can assume the answer is not always passes on 100% of the cost, so owners of the high end rewards cards are winning to some degree. You could argue that non-rewards cards are absorbing costs of rewards cards in the case where merchants do pass through costs, though.
Except that 100% of the cost isn't being transferred to the cardholder, either. You might have a 1% cash back card while the merchant is paying 3%. If only half of the cost is being passed on, you're still losing money. And price isn't the only variable. Even if the merchant ate the entire 3%, that might require them to cut costs in some way so you receive a lower quality product, or drive some competitors out of business and thereby allow the remaining companies to reduce quality without lowering prices because the company providing a better product for the same price was eliminated by the fees.
Right, but the interesting part of TFA is about how the rates paid by merchants are higher for the top 10% of cards. Your example assumes the same rate paid by everyone. Because only a small portion of transactions incur the high rate/high reward, it seems far less likely that the split between pass-through vs eat-it still won't benefit high end cardholders.
That still depends on what the split is. If the average fee is 3% and the high-end customer is getting 2% back, the merchant could be passing on e.g. 2.1% and causing you to come out behind while still passing on only 70% of the cost.
And for anyone getting less cash back the math is even worse, which from the same premise will be the majority of people or else the merchant's costs (and so the amount they pass on) would be even higher.
> If the average fee is 3% and the high-end customer is getting 2% back ...
You're not picking realistic numbers. The fees range from around 1.5% to ~3%, and the 2.5%-3%+ fees are limited to ~10% of customers. An average fee of 3% doesn't make sense. Very basic napkin math would be 0.9((1.5+2.5)/2)+0.1((2.5+3)/2) = ~2.1% average. So only in the 100% passthrough case does the high end cardholder actually lose. That's not likely.
And if you look at the chart, if 740 is 10%, there are probably far fewer than 10% of transactions averaging 2.75%, so this is likely still way overestimating.
Stripe is hella expensive! But they are easy to get started with.
Competitive rates depend on the type of business, i.e. their volume and their fraud risk. The most transparent form of pricing is called “interchange plus” where it’s a flat markup on the interchange schedule. High volume merchants should be able to find a markup in the fractions of a percent.
It is my understanding that the big Stripe customers negotiate lower rates with them as they scale.
And that's assuming you're depositing all of your revenue. If your business allows you to pay some of your suppliers in cash, you could have >50% of your revenue in cash and never pay a bank for cash handling because you're immediately spending it on business expenses rather than depositing it.
> If small businesses give discounts for cash it's because they're committing tax fraud, I presume.
I have seen governments charge a convenience fee for credit card processing. Is the government committing tax fraud?
> Cash handling is typically in the neighborhood of 0.2%-0.3%, so you were apparently overpaying […]
This is not universal, and banks in different countries charge different cash collections fees. The cash collection fees are also structured, e.g. whether the daily collection is required, or every other day, or once a week.
It does not end there.
Many banks still require the business to sort collected coins into separate money bags according to the coin denomination, e.g $1 coins go into one bag, $0.50 coins go into their own bag. Coin bags have a weight limit, 2 or 3 kg, which means that the business has to weigh the money bag up before handing it over, or it will not be accepted.
Now that we are done with material things, we also have to consider all things immaterial that the cash handling entails.
Before the money bag is handed over, the collected cash has to be counted and reconciled against the cash register records on premises, otherwise it will create annoying and time consuming to fix discrepancies in the accounting system. If a staff has accidentally mislaid a note or a few coins, amounts won't reconcile and incur a cash collection delay as the armoured truck can't wait for the reconciliation to complete. Which may consequently increase the risk of leaving cash in the shop overnight with all expected consequences of a potential burglary and losing the cash.
That is just some of the peculiarities of how cash is handled, and I am not sure whether cash handling turns out to be cheaper for an average business with a substantial number of cash payments a day.
Electronic payments, on the other hand do not have any of those shortcomings, vastly reduce the margin for human errors, automate the reconciliation and accounting and reduce the risk (i.e. no money is kept in the shop overnight).
And it's so much easier to pay by phone! I can be through the payment process in 20 seconds at a supermarket if I pay with my phone, and need nothing on me than my phone. Meanwhile cash takes fumbling, counting either by a machine or by a person, calculation in my head what I need to provide, making sure I always have not too little and not too much on me, making sure the total amount of coins doesn't get out of hand, etc.
> This is not universal, and banks in different countries charge different cash collections fees.
This would not seem to be relevant unless you're in one of those countries.
> Many banks still require the business to sort collected coins into separate money bags according to the coin denomination, e.g $1 coins go into one bag, $0.50 coins go into their own bag. Coin bags have a weight limit, 2 or 3 kg, which means that the business has to weigh the money bag up before handing it over, or it will not be accepted.
There are machines that sort and count coins. Coins have uniform weights, so the number of coins in a 3kg bag will always be the same and the machine's count tells you when you're at the weight limit.
Also, what kind of business are you in that you're accumulating small denomination coins? Typical behavior is the customer shows up with three $20 bills to buy something for $52.37 and then you need to disperse a $5, two $1s and some loose change. You don't bring coins to the bank, you bring them a stack of $20s and $50s and then withdraw more small denominations to make change with.
> Before the money bag is handed over, the collected cash has to be counted and reconciled against the cash register records on premises, otherwise it will create annoying and time consuming to fix discrepancies in the accounting system.
So you dump the cash into the counting machine at the end of shift. It's really not that complicated.
> If a staff has accidentally mislaid a note or a few coins, amounts won't reconcile and incur a cash collection delay as the armoured truck can't wait for the reconciliation to complete.
Why would you wait until the truck arrives to do the count? You count the money and put it in a safe. Also, why would you pay for armored car service? If you have $300,000 in annual revenue then your daily take is less than $1000.
> Which may consequently increase the risk of leaving cash in the shop overnight with all expected consequences of a potential burglary and losing the cash.
Shops typically have more value in inventory than they have in cash. Breaking into a safe to steal $1000 in cash isn't even worth the trouble when they can break the glass on the display case and walk away with $10,000 in electronics or tools or small appliances.
> Electronic payments, on the other hand do not have any of those shortcomings, vastly reduce the margin for human errors, automate the reconciliation and accounting and reduce the risk (i.e. no money is kept in the shop overnight).
And they'll be great as soon as we have a system to use them that has low transaction costs and preserves the buyer's privacy by not creating an electronic record of everything they buy tied to their government ID. Until then they can GTFO.
With all due respect, you are wildly projecting based on the lack of knowledge and experience in the problem domain citing the information you have gleaned into (but not gained from) using the first random link that a search engine has turned up in. The reality is vastly different from what you might have conjured up.
Cash handling is a fairly obscure process that is ridden with absurdities and edge cases that most people are oblivious of. The claim that accepting cash is cheaper for the business than taking card payments becomes false once the cash turnover exceeds a certain threshold.
> This would not seem to be relevant unless you're in one of those countries.
«Not happens in my backyard» is not helpful. Cash is handled in nearly all countries around the world, and the cash collection and handling process is more or less similar everywhere. The basics of the cash collection and accounting principles have not changed since the times of Sumerians.
> There are machines that sort and count coins. Coins have uniform weights, so the number of coins in a 3kg bag will always be the same and the machine's count tells you when you're at the weight limit.
Such machines exist in bank branches only, and the bank branches usually close before the business shuts the doors for the night. Most importantly, they do not scale. If it is a family run grocery store or an arts and craft shop, you do your banking yourself during business hours or send your eldest offpsring who is old enough to whack collected coins into a machine and bank the money.
It does not work with larger and big retailers, government agencies and alike with a substantial regular cash intake. Cash has to be collected, usually daily, and it comes with «processing» constraints, e.g. manual counting, receipting and bagging the cash.
A money bag has a weight limit that can't be exceeded, but it can be underfilled because there has been simply a smaller number of coins collected in a specific denomination. A money bag has to have a transaction receipt affixed to it stipulating the amount deposited into the bag – the bank requires it, and it also goes into an accounting system for reconcilliation. If amounts counted by the bank and by the business end up differing, there will be a reconcilliation problem in the transaction feed from the bank into the accounting system even if it is 1 cent off. This is why cash gets counted twice. If you make a mistake counting coins, you have to start over again.
> Also, what kind of business are you in that you're accumulating small denomination coins?
Any medium to large grocery chain that accepts cash for payments where by COB coins are overfilling cash registers. Schools on donation and charity days, e.g. the «gold coin donation» days, where each child brings a $1 or $2 gold coin with them. Service stations and small grocery stores in suburbs and in the country.
> Typical behavior is the customer shows up with three $20 bills to buy something for $52.37 and then you need to disperse a $5, two $1s and some loose change.
Yes, the 37 cents is the problem here. Please allow me to introduce you to the Australian (and previously in New Zealand, too) 50¢ coin that has a dodecagonal shape, has a mass of 15.55 g and a diameter of 31.5 mm, followed by the 20¢ coin with a diameter of 28.65 mm and a weight of 11.3 g. Five such 50¢ coins will burst most wallets and will warrant a purse to carry more.
Back to your example. The change for the $52.37 amount paid by $60 will entail:
1. A $5 note (almost weighless).
2. A $2 coin (6.60 g) or 2x $1 coins (2x 9.00 g == 18.0 g).
3. A 50¢ coin (15.55 g)
4. A 10¢ coin (5.65 g)
5. A 5¢ coin (2.83 g) – due to rounding rules because of 1¢ and 2¢ out of circulation.
The coin change will amount to a net 48.63 g (or 37.23 g for 1x $2 coin scenario) weight, but this weight for now will make a transfer out of the cash register into someone's pocket / wallet. So far so good.
And then you have a customer (in fact, many customers) who is fed up with lugging the metal scrap with them who will pay $7.58 in its entirety in coins (just to offload the burden), and that is how the business ends up with having to fill up money bags for coins at COB. This is ubiquitous, however, the bulk of the change still comes from a constant stream of small purchases in cash.
> Why would you wait until the truck arrives to do the count? You count the money and put it in a safe. Also, why would you pay for armored car service?
Because people still break in and moreso in low-socio areas where people with substance abuse problems will absolutely break in and take any money even if it is $5 in the safe – their conscience does not operate on the ROI level, and they take whatever they can get a hold of. And you would pay for the armoured service because most banks do not offer non-armoured services.
> Shops typically have more value in inventory than they have in cash. Breaking into a safe to steal $1000 […]
Stolen inventory has to be sold first, it can be traced (moreso so for high value inventory items), so there is no instant gain. Stealing even $10 is a net gain, the stolen note is untraceable and it is an instant net gain. Most breakins are small value ones, and the high value breakins are an entirely differen cattle of fish.
> And they'll be great as soon as we have a system to use them that has low transaction costs and preserves the buyer's privacy […]
The reason why we do not have them is because Visa and Mastercard colluded quite a while ago in their quest to eliminate cheap and low transaction fee national payments networks around the world (hello, Cirrus/Maestro, EFTPOS, Switch and similar) that were detrimental to their business. Technically, the national payment networks still do exist but are almost not seen anymore.
Initially, Visa and MC only had one product: credit cards. First, it was an exclusive product for affluent customers with a high net worth that later became a mainstream product for the masses, and the internet was a major driver for the credit cards to become mainstream.
Then Visa and MC figured that they could also tap into the cash payments and came up with the debit card product, approached national payment networks (granted, via connections in governments) to convince them first to co-brand the national payment networks with Visa/MC debit cards (a dual purpose card and the first fix is free) to later proceed to eliminate the national payment cards altogether, which has now successfully happened almost everywhere. Granted, Visa/MC debit card processesing and interchange fees are much higher than the corresponding national networks' ones and for a reason. Governments in many countries have had to intervene and cap the transaction and interchange fees in recent years.
So Visa/MC did not stop there, of course, and are now tapping into the transaction information to gain further from results of their collusion with each other, and they absolutely loathe any privacy related changes that will thwart their PII and behavioural information collection attempts.
The current trend seems to be national payment networks making a slow comeback by way of instant bank payments, payments to mobile numbers and similar ways, so Visa/MC are at odds and do not seem to have a strategy. At least not for now.
The next wave of instant/mobile payments is going to be exciting.
> That's just the interchange fee, not what merchants are actually paying.
What you pay Stripe is for the combination of interchange fees + Stripe's own service fees. The Stripe service component of the fee would be charged regardless. Whatever markup a merchant makes for Stripe's fees are not recoverable and irrelevant to the comparison.
By comparison, Stripe charges 0.8% + $0.00 for ACH, which includes any costs they might be paying to the bank, so the upper bound on the cost of their services is 0.8%. 4.4% - 0.8% is 3.6%.
Identify a payment processor that a small business making e.g. $60,000/year in $20 credit card transactions can use that would charge lower fees, if you can.
You're making a huge assumption that their markup is the same for credit card transactions! You're also making broad oversimplified assumptions about the what merchants do in response to transaction fees. Multiple implausible things need to be true for high end card users to be losing out on this scheme. Most likely the high end users come out at least slightly ahead and most of the cost is borne by lower end users and merchants.
> You're making a huge assumption that their markup is the same for credit card transactions!
We can take this from the other end though. A merchant who accepts ACH via Stripe pays 0.8%, one who accepts $20 on a credit card through Stripe pays 4.4%, so the merchant pays 3.6% more with a credit card than ACH, and 4.4% more than accepting physical cash. It doesn't matter how much of each is Stripe's profit unless there is some competitor a small business could use to process credit cards for less. Is there?
> You're also making broad oversimplified assumptions about the what merchants do in response to transaction fees.
In a competitive market, increasing every competitor's costs would cause them to pass on most/all of the costs, because the competition is keeping margins thin and their alternative is to go out of business. This is not an oversimplification, it's what actually happens in real commodity markets.
> Multiple implausible things need to be true for high end card users to be losing out on this scheme.
All that's required to happen is that the merchants are passing on more of the credit card processing fees than the amount of the rewards. Since the fees are higher than the rewards, this is not that implausible.
1. You're missing the very important detail that there are multiple combined fees, and you are trying to compare a rebate for one of those fees to the combined fees.
2. Commodity markets literally are an oversimplification.
> You're missing the very important detail that there are multiple combined fees, and you are trying to compare a rebate for one of those fees to the combined fees.
The combined fees are the cost of accepting credit card payments. The entire collection of them is avoided by accepting cash.
> Commodity markets literally are an oversimplification.
Only in the sense that everything is an oversimplification.
If OPEC cuts oil production, the price of gas goes up all over, because it's a commodity and the gas stations can't just eat the price increase. If the DRAM companies were colluding to constrain production and they get caught and have to stop, the price of DRAM goes down all over, because it's a commodity and buyers will take the lowest price. But if the price of DRAM goes down, that doesn't mean the price of iPhones go down, because iPhones are not a commodity -- only Apple makes them -- and then they don't necessarily have to lower their prices just because their costs went down.
Real competitive markets can actually behave like idealized commodity markets, or as close as makes no difference under reasonable sets of assumptions.
> You are conflating unassociated fees
If they're unassociated then how does a small business pay only the interchange fee and not the rest of them? If the answer is that you can't, they're not unassociated.
> Real competitive markets can actually behave like idealized commodity markets,
Sure sometimes, but transaction processing markets are not that simple. They are 2 sided markets, i.e. both the merchant and the merchant's customer are customers of the credit card company, who charges the merchant and the merchant customer varying fees based on how valuable the customer is to them. Then you have a merchant services layer on top of that. Then you have the variety of markets and goods that use all these services, all of whom may have different terms and fee structures with the credit card company and/or merchant services company. Merchants will pass on or absorb fees based on many factors, and possibly even varied within its goods and services. It's really complicated, and commodity markets are a gross oversimplification of how it works. Trying to model it would be a nightmare.
> If they're unassociated then how does a small business pay only the interchange fee and not the rest of them?
If interchange fees were 0 then you would still pay the other fees, which do not go to the credit card company. The fees are related, but not associated.
> It's really complicated, and commodity markets are a gross oversimplification of how it works. Trying to model it would be a nightmare.
But now your argument is "it's complicated and there's no way to know" which isn't a strong claim that the status quo is to the advantage of the customer.
Meanwhile the portion of the fee that doesn't go directly to the cardholder is a deadweight economic loss, which, in general, is only to the advantage of the parasite extracting it and to the disadvantage of everyone else.
Notice also that the most likely alternative to "it actually harms even the 2% cash back customer" is "it's barely better than breakeven to even the 2% cash back customer and harms everybody else." Which is hardly a reason to keep it.
> If interchange fees were 0 then you would still pay the other fees, which do not go to the credit card company.
Ah, but that's the issue. You wouldn't. Because the rewards programs are a monopolistic practice.
Suppose I want to start a competing payments network and my sales pitch is I charge low fees. I'm only charging 0.05%, and provide free code that does the basic thing Stripe does, and keep the costs down by using anti-fraud tech the existing networks don't care to invest in because they're shifting the cost of fraud to the merchants and payment processors. The merchants are immediately on board if I can get cardholders. But the cardholders won't use it because no rewards programs.
Take away the rewards programs and now the network with the lowest processing costs will be the most widely accepted, and then customers want those cards because they're more widely accepted and otherwise indistinguishable.
> What you pay Stripe is for the combination of interchange fees + Stripe's own service fees.
Yeah, Stripe is a bad example. Because Stripe is a payment service provider. They take all the various fees involved with managing payments and package it up, then put a pretty bow on top with some useful services, APIs, nice marketing. But this package deal is significantly marked up. Stripe absorbs the variable interchange fees and different rewards card markups because they are charging you 3% flat (more or less) and have a healthy margin in for themselves in the middle. Stripe makes a little less when you charge a Platinum AMEX, but they make a relative ton when you charge a secured mastercard. They know that less than 10% of the transactions are these higher cost cards, so they just absorb the lower profit on those transactions.
This is a wholly different game than lower level payment processors. For example at a company I worked for about a decade ago, we stuck a deal with WorldPay which is the largest payment provider in the world. We were paying interchange fees, a small fraud fee of a few cents, and then a worldpay fee of 20-40 basis points depending on the card. We were directly charged more on a premium rewards card, but the margin on WorldPay was a few basis points above cost. But they provided nothing really in terms of services. We had to find our own payment software, terminals, and everything else. They were just the raw service.
So imagine worldpay on one side, which is just brokering with the banks and requiring us to do everything else. On the other extreme, you have Stripe which is "turnkey" and you can sign up with zero sales volume on a pretty website. One is interchange + 30 basis points, the other extreme is a flat 3%+20¢.
Stripe is a great business. But not a good example here. They essentially abstract away all the complexity in this article by charging you more money (their raw negotiated cost with banks is probably 0.8-1.2% on 90% of their transactions), they are marking up 1-2% for themselves as a service provider. That is not to vilify Stripe. They serve a valuable role and the abstraction layers (SaaS subscriptions, free trials, etc) are well worth it for a lot of companies. But keep in mind, this is a service company on top of the credit card system. So its not a great example in this discussion.
I thougt that it works like when you make a purchase using a credit card that offers rewards, the issuer earns money from the transaction through interchange fees, which are fees charged to merchants for processing credit card payments.
That's the problem right there. Bank accounts should be free by force of law. A bank charging depositors for literally anything is just absurd. They should be paying you, not the other way around.
For many average people I don't see how they could be free. Maybe savings account with limitations should pay something and be free, but with checking account. You get what couple thousand passing through each month?
You have all the costs like KYC requirements, actual overhead of tracking and managing the balances and payments. And many of the payments options are free or low cost...
If it was such money maker, surely there would be lot of competition offering zero fee bank accounts.
> If it was such money maker, surely there would be lot of competition offering zero fee bank accounts.
In Czechia banks literally are competing like that, offering free bank accounts, sometimes even with bonuses on top if you pass enough money through it. So it seems like your model of banks' financials does not work.
> For many average people I don't see how they could be free.
I do. In fact it is free in my country. Anyone can go to literally any bank and get a free account with support for basic operations, including a checking account. I set up my GitHub Sponsors thing with a free bank account. I'm not paying the bank to get paid by GitHub.
Just turn it into a basic human right and don't look back. They should be glad we're depositing our money in there. It's just completely absurd to be charged even a fraction of a cent for the privilege of having some bankers profiting off of your money. It'd make sense to pay them if they were safekeeping it for you in their vaults. They aren't. They're lending it all out to third parties, gambling in stock markets. You name it, they're doing it. These bankers actually crash the economy now and then with their irresponsibility and they face approximately zero consequences for it.
So why should anyone pay them a cent for this "privilege"? That's just clown world levels of insane. They should be competing to see who can pay depositors the most, not charging them. Managing balances? It's the computer's job. Overhead? Literally not our problem. Just make them eat all those costs. Maybe that'll even make them start lobbying for less government bureaucracy so as to reduce their "overhead". AML/KYC is just the financial arm of global mass surveillance anyway and should not even be legal to begin with.
(It's Finland btw, and you need a Finnish, not "anywhere-in-SEPA", bank account in order to practically function in society here with the strong online authentication service)