Centralized versus decentralized supply of savings and payment services. With 81% of all non-cash transactions made by check (Table 2), the U.S. is the heaviest user of this payment instrument among developed countries. (Check use in India, at 99%, is even higher - Table 3. Check use is associated with a disaggregated banking system, a large geographic area, and a historical reliance on a privately-owned banking system for safekeeping of savings for the general public.

Most countries in Europe, in contrast, rely upon GIRO payments rather than checks. Although European banks are aggregated, while U.S. banks are not, this difference is not the main reason for Europe's relatively low use of checks. The primary reason is that, until recently, banks in Europe did not focus on providing savings or transaction services to the general public: this task has historically been the purview of the centralized and publicly-owned post office. This historical reliance on a centralized non-bank provider for savings services, and the fact that most countries in Europe are relatively small geographically, made the shift from cash payments to non-cash GIRO payments easy in Europe but impossible in the U.S. Instead, the historically cost-effective way to make payments in a disaggregated banking system within a large country has been by check.

Paper versus electronic payments. The same reasons that have determined the use of different payment instruments in the U.S. and Europe have also influenced the shift from paper to electronic payments. European countries have been much more successful in implementing electronic payments than has the U.S. for at least four reasons. Europe has: (1) a centralized postal GIRO and an aggregated banking system; (2) a system of nationwide postal and banking offices; (3) a greater emphasis on explicit pricing of payment services to users; and (4) a government/banking system policy of directing payment instrument use toward the most cost-efficient methods. The first two reasons lead to a high proportion of on-us payments so that electronic payments are a natural outgrowth of shifting from paper to electronic accounting within each bank or the post office. The last two reasons - explicit pricing and direct intervention in the market for payment services - has been more evident in Europe and helped to shift users from more to less costly payment alternatives.

Importantly, once a country is in the position of relying heavily on a paper payment instrument, it is very difficult to change that reliance even when new electronic instruments are developed which have demonstrably lower average costs. This is because the appropriate cost comparison is essentially between the average variable cost of the paper instrument and the average total (fixed plus variable) cost of the new electronic instrument. Only variable cost is relevant for the paper instrument because the fixed expenses have already been made - they are sunk costs - and, typically, the only new fixed expenditures are for replacement of fully depreciated capital and training of new workers. Since none of the fixed expenses of the new electronic instrument have yet been made, average total cost needs to be covered if the electronic instrument is adopted. Thus a new electronic payment instrument will only be adopted if its average total cost is less than the variable expenses of the existing paper instrument, which is a more difficult hurdle to surmount.

Non-bank competition in supplying payment services. In Europe today, both banks and the post office provide payment services. If one supplier offers a more desired instrument, their market share and payment revenues rise. Competition between bank and non-bank providers of payment services has been a factor in promoting the shift from paper to electronics.

A similar incentive does not exist in the U.S. as banks are (effectively) the only suppliers of both paper and electronic payment services. Shifting a customer from check use to electronic payments merely succeeds in shifting transaction business from one part of the bank to another. Market share and revenues are little affected as electronic payments cannibalize check volume. Even so, a redistribution of payment business within a bank would be in a bank's interest if strong scale economies existed for electronic payments. Although there is some evidence of such scale economies, banks have not priced their check and electronic payment products in a manner to make these cost differences evident to their customers.

Payment pricing and payment float. The price of different payment instruments should reflect their underlying cost so users, given their needs, can choose the lowest cost alternative. In the U.S., the per transaction cost of using cash, a check, or the equivalent of a GIRO payment is estimated to be, respectively, $.04, $.79, and $.29. In terms of real resources, cash is cheapest while checks are more expensive than GIRO (or even debit card) payments. However, once float costs/benefits are included, the user costs - real resource cost plus float - becomes $.09, $-.04, and $.29. When float is not charged for, checks are usually cheaper for users than is a GIRO or a debit card payment. But these cost differences, especially for non-cash payments, will not be clearly "seen" by users when banks charge for payment services by requiring a minimum balance to be held and/or pay a below-market interest rate on transaction deposits. Here the marginal cost of payment instrument use seen by the user is effectively zero, not positive and differentiated as it would be if a direct fee were assessed for different non-cash payment instruments. Without an explicit fee on payment instrument use, there is no real incentive for users to demand the lowest cost instrument. Although relative bank costs may favor electronic payments, as in the U.S., users have to see this cost difference reflected in the relative prices they pay in order to demand the low cost service. Because of distortions associated with check float and bank pricing policies, resources are wasted and payment costs to users are higher than what they would otherwise be. This illustrates the result that an efficient payment system can only arise in a market economy when payment services are properly priced so that users can correctly choose the lowest cost service.

Commercial bank and central bank payment functions. In principle, all of the functions of a payment system- user interface, clearing, settlement, and regulation - could be provided either entirely by commercial banks or by the central bank. In practice, these functions are divided. Commercial banks perform the user interface function and typically provide for all of the transfer of payment information needed to process, clear, and account for the payments being made. Central banks provide for the final settlement of interbank payments and control the vast majority of the regulatory function. This is the typical division of payment functions for all but large-value payments in most countries. Because of their importance for the smooth operation of commerce and financial markets, central banks often own and/or operate large-value payment networks themselves - and thus directly provide these specific payment processing and clearing services - rather than leave this function solely to commercial banks.

When there is a major difference in payment functions provided by commercial and central banks, it concerns the payment processing and clearing function for low-value consumer and enterprise payments. In the U.S., the central bank processes around one-fourth of all check payments and competes with commercial banks. This is the direct result of having a highly disaggregated banking system (with over 11,000 banks), which was due to a long history of politically motivated geographical restrictions preventing the development of a small set of very large, truly national banks that exists in virtually all other developed countries. When the banking system is concentrated and truly national in scope, commercial banks typically provide all of the processing and clearing of low value payments, postal GIRO payments excepted. The greater the concentration, the larger is the percentage of payments which can be cleared internally. With a national branch network, payments drawn on other institutions can often be presented for payment at the local office of other large banks rather than needing transportation to the (possibly) non-local office where the payor's account is actually located. In this event, payments drawn on other institutions can be cost-effectively collected using a bank's own resources rather than those of competing correspondent banks (or the central bank). This is the arrangement that currently exists in Canada.