Several credit cards are so rewarding that (1) banks are spending the entire amount of interchange on their customers, and (2) banks are losing money on the products.
Premium travel rewards is so competitive that banks may never earn back the cost to acquire a customer (acquisition bonus, plus related expenses). I’ve long argued that when the bank’s model shows payback in 7-10 years, what that means is that the person who built the model will no longer be the one managing the project when it comes time to see whether that prediction actually came true. Customers who keep cards for 5, 10 and 15 years as a whole turn profitable, but the acquisition cost to find those customers may mean that the product as a whole was a money-loser.
Saturday Night Live captured the phenomenon in their commercial for the ‘First CityWide Bank Of Change’. They lose money on every transaction, but make it up on volume!
For these cards the bet is that the bank will lose money on interchange, but make money on APR. They’re looking for customers who will revolve (but with premium cards they’re often attracting the customers least likely to revolve balances). Banks lose money on some customers (who pay off their bills each month) while making money on those who pay monthly interest charges.
There are other versions of this: card deals that pay so much to the co-brand partner, in terms of marketing fees and low credit card processing costs (think: the Citibank-Visa deal for Costco), that the bank is really buying an opportunity to lend to card customers and whether the deal works out depends entirely on how much consumers borrow.
A paper from Federal Reserve, IMF, and National University of Singapore authors is one of the first to look beyond interchange to find winners and losers in credit card rewards. And far from the naive story about cash customers subsidizing rewards credit card customers, they find that cross subsidies are actually from rewards customers who revolve to those who do not.
[W]e find that sophisticated consumers profit from reward credit cards at the expense of naive consumers who lose money both in absolute terms and relative to classic cards. We estimate an aggregate annual cross-subsidy of $15.5 billion. Notably, our results are not driven by income—while sophisticated high-income consumers benefit the most, naive high-income consumers pay the most.
A couple of interesting data points:
Reward cards account for over 80 percent of total credit card spending and for over 60 percent of all new credit card originations (CFPB, 2019). In 2019, the largest U.S. banks paid out $35 billion in credit card rewards.
Interchange aside, those $35 billion in card rewards compare to approximately $100 billion in card interest and fees.
Since banks are looking for customers who will borrow, and to find those customers they also need to find customers that don’t (and they can’t really fire those customers), those who don’t borrow get a great deal – one that in their specific case banks may even lose money on. Banks are looking to make money on a portfolio basis, rather than on each individual customer. But that doesn’t really mean there’s a subsidy, per se. And it doesn’t mean borrowers are losers since they may have a need to borrow. The loss, for them, comes in when they’re taking a less good deal for borrowing than they could otherwise achieve when doing so through their rewards products.
The lesson here is the one I outlined years ago: this game is only for you if you pay off your cards in full each month. It’s easy to moralize and say you shouldn’t carry a balance, but there are customers who need to repair their car to get to work or otherwise smooth their income against necessities and credit cards are a superior form of borrowing compared to the next-best alternative some customers face like payday loans. If you’re someone who uses cards for borrowing, though, you’re best off focusing on borrowing costs (like interest rate, and no fee 0% balance transfer offers) rather than rewards and certainly calculating the net expense.
(HT: Tyler Cowen)
