This article describes the history of debit cards and interchange fees as they apply to debit cards. I'm interested in determining who would benefit and who would lose if this portion of financial reform goes through.
First, a primer.
When you go swipe a debit card, there are fundamentally 5 parties involved. There's the customer, the merchant, the customer's bank, the merchant's bank, and the network.
The customer (let's call the customer Sue, for simplicity of prose) deposits her paycheck into a checking account at the bank, where it can be accessed with a debit card. The debit card is on the Visa network.
Sue then goes and buys things from a merchant - say, Home Depot. She swipes her debit card to pay for the transaction, which is conveniently $100.
However, Home Depot gets less than that, say, $98, which is deposited into the merchant's bank (HDBank) in Home Depot's account. $1.74 goes to the customer's bank (SueBank). Visa gets a piece of it, about 19cents, for facilitating the transaction, and HDBank gets a tiny piece (maybe 7 cents). These aren't precise numbers, but they're order of magnitude illustrative.
If you look at that transaction relative to what would happen if Sue used cash:
Home Depot: +$100
Home Depot: +$98
Note that Home Depot loses 2 dollars, which are paid to the financial institutions. This is almost identical to what happens with a credit card, by the way, except instead of coming out of Sue's checking account, the money Sue spends is being lent to her by SueBank at an exorbitant interest rate.
Of course, Home Depot has control over its own pricing, and it has a mix of cash customers and debit/credit card customers. If it doesn't want to charge separate prices for cards and cash (most retailers don't), it will then adjust its price given the variable profit margins of the two types of consumers and the proportions of the customer base which they make up. This inevitably leads to higher prices. In effect, cash payers are subsidizing card payers for the convenience of using a card.
The article above actually outlines a similar situation from almost 100 years ago: checks. When banks cashed checks from other institutions, they would take a cut of the check for processing. The Fed made this illegal. While I don't have data, I'd be willing to bet you pretty hard that a profit-maximizing bank would have raised its interest rates slightly (it's less important to have lots of customers and more important to have profitable ones if you can't take a cut), but checks, which represent no risk to the bank and are merely a logistical transaction, are free.
Whether this is fair depends on how you view a bank's purpose. If the purpose of a bank is merely as a depository and lending institution, then the provision of checks represent a service for convenience and security so customers don't have to go to the bank and withdraw cash every time they want to pay somebody. From this perspective, if it costs a bank more to cash checks from other banks than it does to cash checks from its own accounts, then it's logical to allow banks to charge that fee.
Of course, this presents a substantial problem: it discourages the use of banks for transactions. Use of banks has tremendous positive externalities, as has entered the news again recently in a WSJ article examining how to promote investment and development in Africa. Savings accounts provide security, so you don't feel the need to carry cash constantly and thus spend it so it's not stolen (up to 20% of incomes in underbanked countries get spent on socially wasteful things like alcohol, prostitutes and drugs, largely because incomes can't be saved easily for fear of theft). They also provide tremendous convenience - imagine being paid your paycheck in cash? Finally, and most importantly, banks provide credit to the rest of the economy, so bank deposits allow other parts of the economy to thrive.
Thus, the Fed argued, almost certainly correctly, that it is socially optimal to force banks to recognize checks "at par" - at the face value they were written at.
The advent of new transactions has meant that this concept has arisen again. I get charged every time I make an online transfer from my savings account to my brokerage, which almost certainly falls under this category, given the security, time and convenience benefits of online transfers (virtually costless and riskless to the bank).
Debit cards may also fall under this category. On the one hand, stores benefit from being able to accept them, and consumers benefit from being able to use them, so asking them to pay isn't wholly unreasonable. However, they're riskless to the bank, they're an important development for commerce, they make transactions way more efficient for every single party, and yet we're making it more difficult for retailers to benefit and thus less likely for them to take debit cards (especially given that they're at potential risk from a customer overdraw). The US may be a more efficient economy, overall, for reducing or eliminating interchange fees, and it really would protect the consumer from having to carry cash into stores that don't take debit cards.
Credit cards have many of the same efficiency benefits, except in this case, the default risk is borne by the customer's bank instead of the retailer.
Of course, nothing comes without second order consequences. Without the ability to charge retailers, banks would probably start charging fees to many accountholders who have debit cards, especially the accountholders for which they don't need to compete as hard under fee-constrained conditions (i.e., low-asset accountholders, from whom revenues are mostly derived from fees, rather than high-asset accountholders, from whom revenues are mostly derived from re-lending deposits. The difference is due to the marginal cost of handling one customer, regardless of assets). You thus risk underbanking more low-income people in the US, and some debit accountholders will be cancelled.
Interest rates on credit cards would probably rise, or, if that's illegal, credit cards would again not be provided to some of the lowest-limit credit cardholders in the system. It's also inevitable that annual fees would have to come back, especially for customers who pay their balance, meaning some of the most fiscally responsible people in the country lose access to credit cards.
In short, killing interchange fees probably reduces access to cards, and perhaps even credit, for both low-income consumers and high-creditworthy consumers. Given how much everyone benefits from consumers having cards, it's probably the case that interchange fees should stay. We pay for our convenience in a) higher prices and b) more consolidated retailers (who can bargain down their fees with the banks/network), but that's probably worth it.
What isn't unreasonable is government intervention with the networks. Again, I almost always hate government intervention, but to argue that Mastercard and Visa don't have substantial monopoly/oligopoly power is blind, and they harm all of us more than they benefit themselves. The credit issuers/banks are somewhat more competitive, so there's less to be gained by targeting them.