Skip to main content

Interbank Settlement and the Emergence of Central Banks

2000 Annual Report

April 1, 2001


Edward J. Green Senior Policy Advisor, Federal Reserve Bank of Chicago
Richard M. Todd Former Vice President and Advisor to the CICD
Interbank Settlement and the Emergence of Central Banks

Here we examine in more detail the development of corresponding banking arrangements that paved the way for the emergence of central banks as hubs in national payment networks. (See Goodhart 1988 for a detailed analysis.) We cast our discussion in terms of check transactions, which were the principal form of transactions (both for large- and small-value payments) from the mid-19th century until the Federal Reserve introduced the precursor to Fedwire, its wire transfer service for large-value payments, in 1918. The points that we make here are as valid for electronic payments as for checks, however.

To begin, consider how transfers of bank balances are used to make payments in an economy with only one bank. A person (household or firm) holds wealth in a demand account, with zero or very low return, primarily in order to make payments. Payment by transfer of a bank balance is acceptable to a payee because it is secure against both theft and loss of market value and is verifiable. Payment by transfer of bank balances is mutually advantageous to the payor and payee because it is fairly inexpensive, so that the cost of making a payment does not eat up the gain to trade.

Now consider what happens when there are several banks. It would probably be infeasible, and would certainly be inefficient, for each person to have an account at every bank. Unless two traders happen to have accounts at the same bank, no individual banker can make payments for them in the way that has just been envisioned. Payment requires a way to get funds from one bank to another. Now, if there are relatively few banks (as in Canada, until recently), a solution to this problem is for every bank to have an account with every other bank. Suppose that, with this arrangement in effect, A writes a check for $1,000 to B, who has a different bank from A. B takes the check to his or her bank, which in turn presents it to A's bank. A's bank debits $1,000 from A's account and credits $1,000 to the account of B's bank at A's bank. B's bank then credits $1,000 to B's account. Over time, there will be payment flows from account holders at A's bank to account holders at B's bank and vice versa. Then—say, when the balance in each bank's account at the other is above $1 million—the banks can agree to reduce those balances by offsetting amounts of up to $1 million without any funds actually having to be transferred. Banks' ability to make such reductions of offsetting payments, known as bilateral netting, can keep the cost of making payments by interbank transfer almost as low as by transfer of balances within a single bank. Only where there is persistent asymmetry in the payment flows between the two banks does it become necessary to make an actual money transfer, which typically does involve significant cost. *

During the period 1837-1913, the United States did not have a central bank. The regime of interbank payments just described was, in principle, how the U.S. payments system operated. However, since there were too many banks for it to be advantageous for every bank to have an account at every other one, a system of correspondent banking arose. Actually, there was a hierarchy of correspondent banks. Each small city had one or more correspondents that served the local banks; each major city had several correspondents that served the correspondent banks of the smaller cities of that region; and New York City had a number of banks that were correspondents for the regional correspondent banks across the country. If B's bank did not have an account at A's bank, then it presented A's check to a third bank—the correspondent bank—at which both it and A's bank had accounts, and the correspondent bank transferred the amount of the check from the account of A's bank to the account of B's bank. Moreover, if there is a cycle of offsetting payments—$1,000 from A to B, $1,000 from B to C and $1,000 from C to A—then the payments that are induced between these payors' banks cancel. Correspondent banking thus provides the possibility of economizing in the payments process by multilateral netting, which reduces the need to make actual money transfers even below the level that would be required under bilateral netting.

Offsetting interbank payments such as we have just discussed typically are not simultaneous. If a correspondent bank waits until receipt of an offsetting payment in order to do netting, rather than debiting the bank on which the first check is drawn, then either the bank that presents the first check, or the correspondent bank, is extending credit to the paying bank of that first check. For example, if A's bank deposits a check to A from B in the morning and the correspondent bank promptly credits the amount of the check to the account of A's bank, while B's bank does not deposit a check for an equal amount to B from A (or payable to and from any two customers of the respective banks of B and A) until the afternoon, then the correspondent bank is making a loan to B's bank over the midday period. On the other hand, if the correspondent bank waits until an offsetting check is deposited with it to credit the account of A's bank, while not debiting the account of B's bank (which would constitute gross payment rather than net payment), then A's bank is extending credit to B's bank over midday, in effect. Because the correspondent bank has an ongoing relationship with each of its respondents, its credit is typically more acceptable to the presenting bank than the credit of a payor bank that the presenting bank may not know well. When the correspondent bank provides credit in this way, it has the option, in effect, to insure the value of the payment to the presenting bank. ** That is, the correspondent irrevocably credits the account of the paying bank at the time of presentment. Such an arrangement is said to provide immediate finality. Particularly in the case of large-value payments, interbank payments are made more efficient by the provision of legal and practical immediate finality in this way.

The roles that large correspondent banks played in netting interbank obligations and extending credit to facilitate interbank settlement were, in our view, the core payments system roles assumed by the Reserve Banks and other central banks.


* Before the Reserve Banks provided a streamlined interbank settlement service, there was a large, direct cost in the form of expensive shipment of currency or gold. Today there remains a cost, albeit a much smaller one, associated with the opportunity cost of holding wealth as balances to effect settlement rather than investing it in productive projects.

** That is, the correspondent bank has the option to offer its respondents a contract to this effect. In some cases, the correspondent may be required by law to do so.

Return to: Thoughts on the Fed's Role in the Payments System