ATMs have been around for almost a quarter of a century, but fees, especially double fees, for using them are a more recent phenomenon.

When ATMs were introduced in the 1970s, they were set up only inside or immediately outside their banks' branch offices. They were seen by banks largely as a way of saving money, by reducing the need for tellers. Even with the relatively expensive computer technology of the late '70s and early 80s, the cost of processing deposits and withdrawals via ATMs proved to be less than the cost of training and employing tellers to do the same work.

To encourage customers to embrace the technology and overcome their trepidations about putting their checks into a machine's slot rather than a teller's hands, banks originally didn't charge customers any fees for using ATMs. (Indeed, in time, some banks started charging customers for not using ATMs, through so-called "human teller fees" - a charge for each time a customer uses a teller for a service that could be performed by an ATM.)

Banks that embraced the ATM profited handsomely, often growing far faster than old-fashioned banks in the effort to get business from ordinary Americans.

At first, a bank's ATMs could only be used by consumers who already had checking or savings accounts with that bank, through the bank’s "proprietary ATM network."

However, by the early 80s, banks began to take advantage of improvements in telecommunications technology and formed "shared ATM networks" with other banks, allowing customers of one bank to withdraw money by using ATMs of other banks. Banks paid other ATM owners "interchange" fees, to cover the marginal cost of the "off-us" transactions by its customers on the owner's machines. Banks paid the network a "switch" fee per transaction, plus an annual "membership" fee, to cover the costs of the network. Originally, these fees were not directly passed onto consumers.

After all, from the perspective of a bank, banks that joined the network could advertise that their customers could get access to their money from far more locations than those banks who didn't belong. Yet, the bank would not only not have to pay for tellers; it wouldn't have even have to pay for the cost of building and maintaining most of the extra ATMs from which customers could access their funds. Big banks, seeking to maximize the value of the new shared networks, urged small banks to join, arguing that joining the network would be much less expensive for a small bank than building a competing network. But the big banks weren’t totally altruistic, even then. They needed to increase usage of their own machines to justify their own expense, and could do so most easily by adding volume from non-customer transactions.

At this point, some banks realized that many people were essentially hooked on ATMs and would be willing to pay some small amount of money to use them, especially when they were travelling. The banks were fortunate that this period coincided with an era of high anxiety about crime and a fear of carrying around large amounts of money.

Consequently, a number of banks slowly began to charge fees. However, originally, the networks themselves prohibited double fees, or surcharges. Some networks even feared that double surcharges would "kill the goose that laid the golden egg."

In the mid-1980s, then, some banks began imposing a fee on their customers for using another owner's ATM. These so-called "foreign" or "off-us" fees became more common in the 1990s.

By the early '90s, using ATMs had become an everyday part of life for a large percentage of Americans. Young people barely even knew what it was like to hand a deposit slip to a teller and ask for their $100 withdrawals in a mix of $5s, $10s, and $20s.

ATMs have been around for almost a quarter of a century, but fees, especially double fees, for using them are a more recent phenomenon.

When ATMs were introduced in the 1970s, they were set up only inside or immediately outside their banks' branch offices. They were seen by banks largely as a way of saving money, by reducing the need for tellers. Even with the relatively expensive computer technology of the late '70s and early 80s, the cost of processing deposits and withdrawals via ATMs proved to be less than the cost of training and employing tellers to do the same work.

To encourage customers to embrace the technology and overcome their trepidations about putting their checks into a machine's slot rather than a teller's hands, banks originally didn't charge customers any fees for using ATMs. (Indeed, even before surcharging became popular, banks started charging customers for not using ATMs, through so-called "human teller fees" - a charge for each time a customer uses a teller for a service that could be performed by an ATM.)

Banks that embraced the ATM profited handsomely, often growing far faster than old-fashioned banks in the effort to get business from ordinary Americans. But once their customers got hooked, banks began imposing fees. First, they began charging their own customer’s off-us fees when they used another bank’s machine.

To many observers, this seemed illogical. After all, it would seem to be in a bank's financial interest to encourage its customers to use other banks' machines, i.e. to provide customers with the same services without having to pay for the construction and maintenance of the ATM equipment.

However, there was more to the move than meets the eye.

For one, banks were - and always are - worried about market share. They're worried about losing customers to other

banks. So, they'd rather that their customers not hang around other banks' ATM machines for fear that their customers will take the next step and bring their checking and savings accounts - and their home and car loan business - inside the other banks' doors. So, even though the banks claimed that the fees were necessary to offset the costs of the "interchange" and "switch" fees, they wanted to keep their own customers closer to home.

For another, large banks in particular were eager to take advantage of their size - which was partly the product of new laws and regulations allowing for extensive interstate banking - when competing against smaller banks. The new fees gave the large banks the golden opportunity to advertise that they had far more ATMs than smaller banks, which meant that their customers would have to pay fewer fees for using "foreign" banks' ATMs.

Finally, the fee system was rigged to benefit all concerned - except the customer. According to the U.S. Office of Thrift Supervision, an average ATM transaction costs the ATM owner about $0.27, including the amortized cost of the equipment, the telecommunication costs, and the personnel to oversee the operation. However, as of 1998, the average "foreign" fee was $1.20. Of that, 10c (the "switch" fee) typically went to the network operators (e.g. Cirrus, Star), 60c (the "interchange" fee) went to the ATM owner, and the remaining 50c was kept by the customer’s bank. Of course, it had virtually no out-of-pocket expenses when its customers used another owner’s machine, except its pro-rated share of network membership, which it also offset by the "off-us" fees it received from other banks. After all, in addition to paying interchange fees when its customers use another’s machine, banks receive interchange fees when other banks’ customers use their machines.

As is often the case, this "nickel-and-dime 'em" strategy irritated consumers, but didn't enrage them - especially politically influential upper-income consumers, for whom a dollar was a small price to pay for the convenience of withdrawing $300 from an ATM at the top of Vail Mountain and other relatively obscure locations.

The banks also had the political wisdom to spend millions of dollars on campaign contributions and top lobbyists . . . and to not impose fees on ATMs located in or near government offices. A 1998 study found that the only Wells Fargo ATM that did not charge fees in the entire state of California was the one that sat in the basement of the state capitol in Sacramento.

The new fees proved to be a boon for the banks, contributing mightily to their profit turn-around, which started at about the same time.

By the mid-90s, a number of large banks decided to push their luck a little further . . . and, this time, perhaps too far.

The banks argued that they needed to charge a second fee for the exact same operation, but with a different justification. The rationale was the other side of the above equation; that banks should be allowed to charge non-customers for using their machines. In other words, the banks should be compensated for building the ATMs and for their wear and tear.

Superficially, the rationale for this so-called "ATM surcharge" sounds perfectly plausible. As bank spokespeople are wont to say, there should be no such thing as a free lunch, that banks shouldn't be forced to give away their services (e.g. the ATMs) to non-customers for free.

What those spokespeople neglect to mention is that not only are the banks saving money through ATMs by not having to hire tellers, but they are already being more than adequately compensated for the cost of the machines through the interchange fee that the non-customers' banks are already paying to them. (They're getting compensated even if the non-customer's bank doesn't charge its customer an "off-us fee" .That's because the non-customer's bank still is paying a so-called "interchange fee" to the network and the ATM owner, but, rather than passing that cost onto the consumer, they're absorbing it.)

At first, even many elements of the banking industry balked at the ATM surcharge concept. Banks fought to enact state laws allowing surcharging, and put pressure on the nation’s largest ATM networks, Plus and Cirrus, to allow surcharging across their network switches. But for several years, those national networks resisted.