What Would Happen if ATM Surcharges Were Banned?

In early September, Bank of America, the nation's largest owner of ATMs with about 17,000, increased the fee it charges to non-BoA customers for using BoA ATMs to $3 from $2, the highest such fee in the nation. (A timely move for a bank with about $5.5 billion in mortgage securities-related write-downs this year.)

And while ATMs have become ubiquitous in most parts of the country over the last two decades, growth rates and usage have declined in recent years. There were 33 ATM transactions per-capita in the U.S. in 2006 versus 45 per-capita in 2001. Between 1996 and 2001, the number of ATMs grew by 133 percent to 324,000 while between 2001 and 2006 the growth rate fell to 22 percent for a total of 395,000 ATMs countrywide.

The growth spurt in the mid-to-late 1990's coincided with widespread use of surcharges for the first time. But over the same period, the number of transactions per ATM fell by about 45 percent. In a recent paper, Gautam Gowrisankaran of Washington University and John Krainer of the Federal Reserve Bank of San Francisco looked at whether the ATM surcharges over-encouraged the proliferation of ATMs and whether we'd be better off if there were no surcharges and less ATMs.

To investigate, they turned to the border region of Iowa and Minnesota. While that might sound arbitrary or unusual, it turns out that for at time Iowa banned the used of surcharges, so the location actually produced a nice quasi-natural experiment. Using data from 2002, the year before surcharges were introduced in Iowa, Gowrisankaran and Krainer constructed different scenarios to see how the economic surplus would be affected by different pricing regimes.

They found that the overall surplus was very similar between a policy that allowed surcharges and one which prohibited it.

We find that a surcharge ban would moderately decrease the number of ATMs, increase consumer surplus, decrease firm profits, and result in roughly the same total surplus.

This means that the type of ATM surcharge policy that's chosen largely only impacts the distribution of the economic benefits between consumers and banks.

Not surprisingly, allowing for surcharges results in more ATMs (an average of 1.27 instead of 1.12 per 1,000 people) but fewer total transactions due to the higher prices. The implication is that consumers gain more from the lower prices without surcharges than they do from the lower travel time [because more ATMs mean that you're likely to be closer to one] when surcharges are allowed.

Surprisingly the optimal solution for society, Gowrisankaran and Krainer suggest, would come about if there was a central planner. In their experiment, the researchers found that this entity would choose no surcharge and about 50 percent more ATMs than in the surcharge scenario.

The fact that there is more entry implies that firms are adding to consumer surplus with their entry more than they are reducing profits to other firms by stealing their business. Because of the additional entry and the zero prices, the volume of transactions is higher than under the other regimes.

The researchers also found that despite the surge in ATMs after surcharging was introduced in 1996, a ban on this type of fee only reduced new ATM locations by 12 percent.

This relatively small effect stands in stark contrast to the observation that ATM deployment tripled between 1996 (when the networks lifted their surcharge ban) and 2001 but is completely consistent with the difference between Iowa and Minnesota border data. It is worth noting that ATM deployment was growing rapidly throughout the 1980s and early 1990s as well, apparently for reasons other than the price of ATM services.

Overall, the research suggests that -- if easier access to ATMs is seen as a societal benefit -- banks need to be encouraged to invest more in their ATM networks.

(Here is another recent paper on the topic which looks at the distribution of ATMs in Belgium and comes to similar conclusions.)