RDP 2012-03: ATM Fees, Pricing and Consumer Behaviour: An Analysis of ATM Network Reform in Australia 1. Introduction

For many countries, automated teller machines (ATMs) have become the main channel through which consumers withdraw cash from banks. In 2007, two out of every three cash withdrawals in Australia were made via ATMs (Emery, West and Massey 2008). Almost half of the ATM transactions made that year occurred at machines that were not owned by the cardholder's bank, that is, at so-called foreign ATMs.^[1] A number of fees can be associated with a single transaction at a foreign ATM: an interchange fee (that the cardholder's bank pays to the ATM owner); a foreign fee (that the cardholder pays to his own bank); and a direct fee (that the cardholder pays directly to the ATM owner).^[2] In Australia, interchange fees were prohibited in March 2009 by the Reserve Bank of Australia (RBA). This was part of an industry-led reform that also included a move to a direct-fee model for ATM pricing and other measures designed to increase competition and efficiency in the ATM system (RBA 2009a).

A number of theoretical models have been developed to examine banks' profit-maximising choices about how to structure ATM fees. A common feature of these models is that the elimination of interchange fees will cause a one-for-one increase in direct fees and a one-for-one fall in foreign fees, leaving the price faced by consumers for foreign ATM transactions unchanged in the short run. However, while the total price of foreign ATM transactions was unchanged in Australia following the 2009 reform, the adjustment of foreign and direct fees was almost twice as large as the reduction in the interchange fee.

This paper addresses this discrepancy by developing a model of ATM fees that can explain this feature of the Australian experience. Specifically, it extends an existing model of ATM fees (Croft and Spencer 2004) by relaxing a number of assumptions regarding the division and identification of costs and the number of banks in the market. Equilibrium fee strategies are solved for explicitly, and it is shown that when there are three banks in the market, the model still predicts that the total price of a foreign ATM transaction will be unchanged, but that the change in foreign and direct fees need not equal the eliminated interchange fee. The extensions to the model regarding costs also allow a direct assessment of whether certain costs of providing ATM services map directly into certain fees. However, the approach taken in this paper to modelling ATM fees, and the approach taken in the existing literature, cannot explain a striking feature of the Australian experience – the shift in consumer behaviour away from foreign ATM use. Two potential explanations for this are proposed.

The paper proceeds as follows. Section 2 describes the Australian ATM market pre- and post-reform. Section 3 surveys the literature on ATM fee determination and discusses theoretical and empirical findings, while Section 4 develops a model of ATM fees based on Croft and Spencer's (2004) approach and derives predictions for ATM fees following the removal of interchange fees. These results are then shown to be consistent with the Australian experience. The response of the consumer is examined in Section 5 and potential explanations for observed behaviour are put forward. Section 6 concludes.

Footnotes

In 2010, ATMs were still the primary channel through which individuals accessed cash, though their use relative to other cash withdrawal methods had diminished (see Bagnall, Chong and Smith 2011). This shift may reflect the March 2009 ATM reforms discussed in Sections 2 and 5. Data on ATM cash withdrawals can be found in Reserve Bank of Australia Statistical Table ‘C4 ATM Cash Withdrawals’. [1]

The direct fee is also known as the direct usage fee, the direct charge or the ATM surcharge in the literature. [2]