The core intuition is that a given ATM often has monopoly
power ex post, once you are there and need the money. Lower fees mean
fewer machines but that still might be better than facing the mark-up.
Here is the paper, whack it
down if you can.
Here are a few of my own thoughts.
First, in city centers ATMs don’t seem to be particularly
scarce or so far apart from each other to justify accusing many of them as
having locational monopolies. A few
months ago I was in Aveiro, Portugal, and in
quick need of cash. Even though Aveiro
is no great metropolis, I noticed three ATM machines each within (I’m estimating
here) less than a five-minute walk from the little shop that refused to accept
my credit card as payment (and, thus, caused me to seek out cash). (The machine I used charged a fee; I presume
that also the others would have charged me a fee.)
So the locational-monopoly story might work for the lone ATM
in some remote town nestled high in the Canadian Rockies, but surely it is a
suspect account of ATM fees charged for use of machines in cities.
Second, and relatedly, as the acceptability of credit cards
(at least in the U.S.) for making even very small payments increases, the
plausibility of the locational-monopoly account of ATM fees shrinks. Greater acceptability of credit cards as
payment for the likes of cups of coffee and ice-cream cones means that even the
lone ATM within a convenient walk of a Starbucks or a Baskin-Robbins has less
locational-monopoly power than it might have possessed just a few years ago. Are ATM fees becoming rarer?
Third, the locational-monopoly account fits very awkwardly
with the fact that banks typically do not charge ATM fees to their own
customers. Presumably it’s just as
inconvenient for a Bank of America customer to seek out another ATM (other than
the Bank of America ATM that he’s about to use) as it is for a Citibank bank
customer, standing behind him in line, to do so.
Fourth, if we grant (contrary to what I argue above) that
ATMs indeed are little locational monopolists, this putative fact would mean
that ATMs are at least far enough apart from each other that it’s not worth the
typical cash-seeking person’s time to walk or drive from one ATM to another in
order to save a dollar or two. By the
NBER paper’s authors’ admission, though, outlawing ATM fees reduces the number
of ATM machines. Thus, with fewer ATM
machines the average distance between each one still in operation increases, making
any need to trek from one to another even more burdensome.
So imagine a busy Saturday night in downtown Des Moines, where ATM
fees are prohibited. A long line exists
at a few, and one or two others are out of service. The inconvenience facing customers needing
cash to find the nearest working ATM is now greater than it would have been had
the government of Iowa not outlawed ATM fees. Are these
customers made better off by the prohibition on ATM fees?
Fifth, many automobile dealerships today offer free loaner cars when you bring a car that you bought from that dealership into that dealership for service or repair. But I know of no dealership that offers free loaners to persons driving other makes of automobiles. If governments mandate that any dealership that offers free loaner cars to service-customers who bought their cars from that dealership must also offer free loaners to service-customers who bought their vehicles from other dealerships, what would happen to the prevalence of "free" loaners?