Branch density, staff productivity, a nation's stage of economic development, and the share of its financial assets controlled by banks have all been held responsible for variations in the cost of national banking systems. Of these factors, the first two have attracted the attention of individual banks searching for cost cuts, because they are matters that lie within their control.
But there is another factor that banks control, but to which they have paid scant attention: the pronounced link between the type of payment system used in a given country and the cost structure of its banking. Specifically, the more extensive the use of checks for non-cash payments, the higher the cost of the banking system. So while painful branch and staff cuts may, in some cases, have helped reduce costs, a more effective means to that end might have been to wean customers away from their checkbooks.
To take the aggregate evidence first: those OECD countries where checks are widely used for non-cash payments (the United States, Canada, Australia, and the United Kingdom) tend to have more costly banking systems than the European countries that rely to a greater extent on electronic transaction methods or non-bank giro systems (Switzerland, the Netherlands, and Germany) [ILLUSTRATION FOR EXHIBIT 1 OMITTED].
This link is stronger than that between the cost of national banking systems and any of the other factors traditionally used to explain differences in that cost. Take branch density, for example. Some high-cost countries, such the United States and the United Kingdom, have lower branch density than low-cost countries such as Switzerland and Belgium [ILLUSTRATION FOR EXHIBIT OMITTED].
What is true for an entire banking system also holds for individual banks: check use is important in determining overall cost structure. Banks with a low level of check use and a high level of electronic transactions operate with substantially lower costs than banks with a high level of check use.
The reason lies in a bank's cost structure. If a bank's customer base makes heavy use of checks, as much as 55 percent of its total costs can relate to retail payments [ILLUSTRATION FOR EXHIBIT 3 OMITTED]. In many cases, most of this cost can be attributed to the issuance, processing, and reconciliation of checks and other paper-based transaction instruments such as bank drafts.
The biggest block of retail payment costs sits in distribution, where a large number of staff tend to be occupied in handling paper-based transactions. Closing branches has little impact on such costs because it merely relocates check-related activities and associated costs. Centralization may permit processing scale efficiencies of 10 to 20 percent of overall retail payment costs, but savings as high as 90 percent could be achieved if checks were replaced by electronic alternatives. A typical over-the-counter check transaction, fully costed, costs a bank around A$3; an equivalent ATM transaction costs $0.50, an EFTPOS (electronic funds transfer at point of sale) transaction $0.30, and a direct debit or credit less than $0.10.
There are four main strategic implications for banks, bank regulators, and governments that run counter to conventional wisdom:
Subsidize. To reduce costs, banks need to change customers' behavior. To do that, they may have to subsidize emerging electronic channels as an incentive to switch from checks. Such a move is likely to be unpopular with banks already uncomfortable that their cost base has increased as channels intended to reduce it have proliferated. But the problem lies not in the cost of new channels, but in banks' failure to scale back the more expensive channels. Reducing transaction fees for electronic channels would be one solution: the less banks charge their customers to pay a bill through an ATM, for example, the better their chances of wooing customers away from branches and checks.
Don't target retail customers. The primary target of banks' campaigns to alter payment behavior has been the retail customer. In the United States, for example, retail customers write more than 80 percent of all checks [ILLUSTRATION FOR EXHIBIT OMITTED]. Yet the number of customers involved means progress is often slow and expensive.
A more effective way to reduce check use would be to focus on the main recipients - that is, businesses. More than 90 percent of checks are drawn on companies, many of which accept no other form of payment. Banks should therefore make sure that a consumer "push" for electronic payment instruments is complemented by a "pull" from those that accept these funds.
Don't compete, cooperate. Competitive pressures will limit the extent to which banks succeed in persuading customers to use alternative payment methods. If customers can switch to other banks that continue to promote the use of traditional methods at subsidized rates, little progress will be made.
This constraint will be accentuated by the network characteristics of electronic channels. In most countries, electronic channels are linked into wider networks to provide maximum benefit to customers. Networks encourage uniform price-setting in other words, all banks end up charging each other the same interchange fee for allowing one another's customers to use their network. This arrangement ultimately favors the bank with the highest costs, because these costs are passed on to customers.
This lack of market discipline may already be causing over-investment in electronic payment channels. Between 1991 and 1996, growth in ATM terminals in Australia outstripped that in ATM transactions [ILLUSTRATION FOR EXHIBIT 5 OMITTED], leading to higher rather than lower transaction costs on average.
Until it is cheaper for customers to use electronic payment systems, banks will have little chance of persuading them to switch - particularly in the case of older customers who hesitate to try other channels. Of customers aged 55 and over in Australia, for example, 68 percent have never used an ATM, 76 percent have never used an EFTPOS terminal, 92 percent have never banked over the telephone, and none has tried PC banking [ILLUSTRATION FOR EXHIBIT 6 OMITTED].
Banks therefore need to cooperate, at least temporarily, to discourage the setting of high interchange fees and encourage the movement of large groups of customers to new channels. This may raise concern about anti-competitiveness. But if a low-cost banking system is considered to be in the national interest, regulators may have to allow banks to collaborate rather than compete to achieve customer migration.
Governments should intervene. Governments can play an important role in influencing payment systems. Emerging economies will benefit if their governments encourage a rapid move to electronic systems before customers become reliant on checks. Governments of developed countries can drive change by using electronic systems for their own transactions - by making social security payments through rechargeable smart cards, for example, or providing telephone banking facilities for community services.
Encouraging these types of change will pay handsome rewards. We estimate that US banks could save up to a third of their operating costs - and reduce check volumes by a factor of four or five - if customers could be persuaded to behave like the Dutch or the Germans in the way they pay their bills. But change will come about only if individual banks, the industry as a whole, regulators, and governments start working together to the benefit of investors and consumers alike.
Many of the ideas expressed here build on themes from a review of financial system regulation undertaken by the Australian government in 1996-97. We would like to thank Lynne Curran from the Australian Federal Treasury for her comments.
Matt Bekier and Sam Nickless are consultants in McKinsey's Sydney office.…