The Future of Payment Systems

Stephen Millard, Andrew Haldane, Victoria Saporta

• Publisher: Routledge
• Year: 2007
• Print ISBN:
  9780415438605
• eBook ISBN:
  9781134071296
• Topic: Business
• Subtopic: Business General
• Index: Business

Drawing on wide-ranging contributions from prominent international experts and discussing some of the most pressing issues facing policy makers and practitioners in the field of payment systems today, this volume provides cutting-edge perspectives on the current issues surrounding payment systems and their future.It covers a range of continually im

Part I
Payment systems and public policy

1 Central banks and payment systems

Past, present and future

Stephen Millard and Victoria Saporta¹

Introduction

Central banking and payment systems – mechanisms that enable the transfer of monetary value – are inextricably linked. In the past, institutions that developed into modern central banks stood at the top of the inter-bank payments hierarchy, providing the ultimate settlement asset exchanged by commercial banks when settling payments with each other. At present, modern central banks devote a considerable proportion of their resources to operating, overseeing and influencing developments in payment systems. In the future, innovations in payment system technology might permanently change the role of central banks, possibly even leading to their demise.

And yet, the economics literature in the field is surprisingly scarce. With some honourable exceptions (including papers by the contributors to this volume), mainstream monetary economics has largely ignored the mechanics of how payments are actually made and banking theory has largely ignored the management of liquidity intraday. Even within central banks, payment systems are often treated as simply ‘the plumbing’ and left to technocrats.

The aim of this chapter is to paint a broad-brush picture of the economic links between central banks and payments in the past, the present and the future. The purpose is ambitious and impossible to cover comprehensively in a single chapter – hence ‘broad-brush’. In particular, we start by arguing that the modern roles of central banks can be seen as natural outgrowths of their historical role in the inter-bank payments hierarchy. We then proceed to ask what are the characteristics of payment systems modern monetary authorities should be interested in? And how should this interest be made operational? Should central banks own, operate and/or oversee payment systems? We conclude with some tentative thoughts on how the payments landscape may evolve in the future and what that may mean for the future role of central banks.

The chapter is organised as follows. We first provide background on the development of payment systems and central banking, arguing that historically they have been closely linked – the past. We then go on to analyse the role of modern-day central banks in the payment system, in particular in which systems should they be interested and how should they exercise this interest – the present. Finally, we offer ideas about the future direction of payment systems and of central bank involvement therein – the future.

Payment systems and central banking – the past

Natural pyramiding

Historically, the evolution of central banking can be traced back to the market’s natural demand for an efficient way to make payments. This natural demand can lead to the development of a hierarchy or pyramid in payments with the liabilities of a proto central bank at its apex, as the ‘settlement asset’ of choice. In other words, where institutions could provide a safe settlement asset that other banks use to settle obligations ultimately between themselves, they often developed the characteristics that in the twentieth century we came to associate with modern central banks.

Payment systems form the means by which monetary value is transferred. Agents have a natural demand for a safe and verifiable asset – money – that they can use to transfer value in exchange for goods. This demand is derived from the low probability of the ‘double coincidence of wants’ necessary for trade in a barter economy (Jevons (1875) and, in a modern context, Kiyotaki and Wright (1989, 1993)). Given this asset, agents will eventually wish to find a way of being able to make payments – transfers of this asset – without having to carry it. There are at least two reasons for this.

First, as suggested inter alia by He et al. (2005), money is susceptible to theft. Banks developed as places where people could deposit their gold for safekeeping. The banks would then issue their customers with receipts. These receipts represented a form of debt and, eventually, this debt became ‘transferable’ in the sense that it became possible for a merchant who wished to make a purchase to transfer the debt to the seller as payment for his goods. Final settlement occurred when the sellers went back to the bank to call in the debt.

Second, as suggested inter alia by Kohn (1999), it was hard to verify the true value of different coins (the predominant form of money at this stage). Banks developed as places where agents could have their money counted and valued by money changers. As it was efficient for this process to only happen once, agents would leave their money – once counted and valued – with the money changers who would issue them with receipts. Payments were made with both payer and payee present at the bank. Where the payee did not hold an account at the payer’s bank, he either opened one or could ask the bank to transfer the money to his own bank. Since banks were close to each other, this was done by the payer’s banker walking over to the payee’s banker with the money.

But, in an economy with many banks, it is inefficient for every agent to have an account with each and every bank and the banks themselves might be a long distance from each other. One solution is for each bank in the economy to have an account with all other banks and net obligations bilaterally with them. In a world with many banks this will tend to result in an inefficiently large number of inter-bank accounts. A more efficient solution is for a hierarchy – or pyramid – of banks to develop, with banks at the bottom of the pyramid having accounts with correspondent banks in its upper tier which in turn have accounts with banks at the apex of the pyramid. Indeed, there is plenty of historical evidence that such pyramiding evolved naturally in a free-banking environment without the need for the state to superimpose and/or guarantee a ‘settlement institution’ at the apex of the pyramid.²

One example is the case of England (and, later, the United Kingdom). The Bank of England was founded in 1694 and was granted a number of privileges by the British Government, in return for its services in raising finance and managing the Government’s accounts.³ Due to these privileges, the Bank has been the largest and best capitalised bank in the United Kingdom for most of its history. Its large capital base and creditworthiness meant that it became the ‘custodian’ of choice – other banks naturally felt that it was the safest institution in which to hold their gold reserves, which they exchanged against Bank of England notes. Consequently, Bank of England notes (and later deposits) became the ultimate settlement asset for making payments, placing the Bank at the top of the payments pyramid in the United Kingdom. But, for most of its history, despite being the Government’s banker, the Bank did not enjoy an explicit government guarantee, nor was there an explicit or implicit acceptance that if the Bank chose to put the capital of its shareholders at risk the Government would step in to cover any resulting loss. For example, in 1890 the Chancellor of the Exchequer refused a request by the Governor of the Bank to guarantee its shareholders against loss if it were to support Barings Bank.

A second example of natural pyramiding is the development of the Suffolk Bank system in Boston in the early nineteenth century. The development of this system is discussed in Goodhart (1988), Trivoli (1979) and Calomiris and Kahn (1996). At the time, ceteris paribus, Boston banks could issue fewer notes than their New England country competitors because the probability of a note being presented for payment varied negatively with the difficulty of travelling to the bank that issued it. This put the Boston banks at a competitive disadvantage to country banks and encouraged them to develop secure and systematic ways to redeem the various note issues that were circulating freely around the city. The Suffolk Bank ran the most successful system – it undertook to redeem at par the notes of country banks as long as they maintained sufficiently large deposits, topped up as necessary so as to make redemption at par possible. Moreover, the Suffolk Bank refused entry to its clearing system to banks it deemed not to have the requisite degree of integrity. In effect, it undertook an early form of supervision of banks.

A third example relates to the arrangements for inter-bank payments in the United States during the period 1837–1913 (when there was no central bank in the country). Green and Todd (2001) explain that a hierarchy of correspondent bank relationships developed. Each small city had one or more correspondent banks and New York City had a number of banks that facilitated interregional payments; that is, there was essentially a ‘mutualised cooperative’ at the top of the pyramid. Put in the words of Smith (1936):

The conspicuous position held by the banks of New York City in this respect – in 1912 six or seven of them held about three-quarters of all banks’ balances – seemed to point to the existence of spontaneous tendencies to the pyramiding and centralisation of reserves and the natural development of a quasi-central banking agency, even if one is not superimposed.

(our italics)⁴

There is, of course, the issue of whether natural pyramiding is socially optimal or whether the government may wish to intervene by creating an institution that sits at the summit. On one view – referred to as the ‘jaundiced view’ in Calomiris and Kahn (1996) – private systems such as the Suffolk Bank system are driven by large banks seeking to limit the supply of money and engage in monopoly pricing. Any gains are at the expense of the smaller banks and the public as a whole. An alternative view – the ‘sanguine view’ – is that such arrangements increase efficiency and reduce risk in the banking system. Calomiris and Kahn (1996) suggest that empirical evidence backs the sanguine view in the case of the Suffolk Bank system (see also Selgin and White (1994) for a similar view).⁵

But regardless of whether such natural pyramiding is socially optimal, the fact that it seems to occur raises the question of how many banks would naturally take this role at the top of the hierarchy? Does the market, in each currency, tend to one proto central bank or more? The relative standing of different banks and the structure of capital market flows in a country are important factors – as in the case of the Bank of England. Another important factor is the structure of the banking market.⁶ In an oligopolistic ‘free banking’ market with few banks, it may still be efficient for banks to hold bilateral correspondent accounts with each other, settling in each others’ monies, rather than in an outside settlement asset. According to Green and Todd (2001), in Canada banks did just this, until recently. In consequence, markets with a few large banks dominating the system may tend to develop flatter upper-tier structures. In contrast, in a unitbank system – that is, a system consisting of a large number of small independent units – efficiency considerations will lead the smaller units to seek an arrangement that would decrease the number of inter-bank relationships. In such systems ‘proto central banking agencies’ may develop naturally.⁷

Features of the settlement institution

What are the financial features that such proto central banks need to display to enjoy a comparative advantage in performing the functions of the settlement institution at the apex of the pyramid?

First, if a central bank is commercially-oriented – in practice, if it is privately-owned – it needs to find ways of overcoming various conflicts of interest. Banks whose payments are made via the settlement asset of the proto central bank would need assurances that it is not using the information in their accounts to compete unfairly with them. Further, there may be a tension between a bank’s pursuit of profit and its role as a central bank. There is plenty of historical evidence that suggests that when the provider of the ultimate settlement asset is also a commercial bank, conflicts of interest ensue, especially during periods of financial crisis. For example, in 1793, the Bank of England was asked to aid some large country banks – it refused to do so and some important failures occurred that spilled over to the London market. Henry Thornton (1802) writes that ‘a sense of unfairness of the burden cast on the Bank by the large and sudden demands of the banking establishments in the country, probably contributed to an unwillingness to grant them relief’. Goodhart (1988), inter alia, describes a number of other examples, involving commercial rivalries among the Suffolk Bank and the New England country banks, the Bank of England and the London bill brokers and the Banque de France and potential commercial competitors.

One way of overcoming the conflict of interest problem is for the central bank to stop competing with the other banks for non-bank business. In effect, this was how the Bank of England overcame the problem, withdrawing from all (new) commercial activity with non-bank entities between around 1880 and 1910 (Goodhart, 2004).

Second, the banks whose payments are made using the central bank’s liabilities as settlement asset need to be confident in the credit quality and liquidity of this asset – where liquidity should be taken to mean the acceptability of this asset as a means of payment by others. At least one of three features recur in the development of institutions as central banks and help explain other banks’ willingness to use these assets: (a) the provider’s bank notes and deposits are backed by a commodity with intrinsic value (such as gold); (b) the provider has a very large capital base such that the probability of failing to realise its obligations is very small; and (c) the provider holds an explicit or implicit government guarantee.

During different times in its history, the Bank of England had each of the features (a), (b) and (c). The fact that the Bank had the largest capital base of any bank in the United Kingdom well into the nineteenth century – feature (b) – was the key factor in explaining why the Bank’s liabilities became the settlement asset of choice. During this time, the Bank’s relative standing as the banker to the Government might have created the impression that it had an implicit government guarantee – feature (c) – but any such impression was certainly less firmly held than today. The Barings episode in 1890 is a concrete example of the Government refusing to underwrite the capital of the Bank.

It was not until 1844 that the Banking Act placed restrictions on the Bank’s ability to print notes that were not backed by gold. It stipulated that the Bank had to hold gold reserves against all the notes it issued in excess of a fiduciary issue of £14 million – essentially forcing the Bank to display feature (a). However, the regulation was suspended during subsequent periods; in particular, during the liquidity crises that occurred in 1847, 1857 and 1866, the government allowed the Bank to issue additional notes not backed by gold. Again, however, there was no indication that the government would underwrite any losses the Bank made as a result of intervention. In 1946, the Bank was nationalised – effectively giving it feature (c).⁸

But the Bank of England’s history may be unusual. In many cases, more than one bank exhibited features (a) and (b). Either you would expect to see these banks ‘jockeying for position’ as the central bank or you would expect to see flatter upper tier structures as in Canada or the New York Clearing House system with ultimate settlement (where necessary) carried out in gold or government bonds.⁹ In such cases, eventually central banks tended to be superimposed by the state to provide the ultimate settlement asset – as, for example, in the cases of the Bank of Canada, the Federal Reserve, the Reichsbank and the Swiss National Bank, among others. Often, but not always, this happened in response to banking crises that were perceived to result from the lack of a central bank being able to provide lender-of-last-resort assistance. For example, the Federal Reserve System was set up in 1914 in response to the banking panic of 1907 as a direct result of a perceived need for a government-backed lender of last resort in such circumstances. This is discussed in more detail below.

Proto central banks, financial and monetary stability

At present, central banks all around the world generally share two closely related core purposes – monetary and financial stability. In this sub-section we ask why central banks evolved as the natural candidates for taking on these two responsibilities and argue that the answer lies in the key role proto central banks played in the payment system. That is, these core purposes can be seen as natural outgrowths of a central bank’s role in payments. In Section 3, we reverse the question and ask, given that modern central banks are currently the public institutions commonly charged with the preservation of monetary and financial stability, what do these core functions imply for their modern interest, and active involvement, in payment systems?

Historically, privately-owned settlement institutions that supplied the settlement asset at the top of the payments pyramid had a natural interest in ensuring the ability of their client base – the banking sector as a whole – to meet the public’s demand for liquidity. The reason for this is that if it allowed a solvent commercial bank to fail as a result of a run, it would only aggravate the situation and this could ultimately result in a run on itself. Also, assuming the commercial bank stayed in business, the central bank would make a high return on this lending (given that lender-of-last-resort assistance would typically be given at high rates of interest). Put differently, profit maximisation is consistent with Bagehot’s (1873) rule that a central bank should always lend to liquid but solvent institutions against collateral – that is, be a ‘lender of last resort’. Historical evidence backs this assertion. For example, the Bank of England provided lender-of-last-resort assistance during the financial crises of 1857 and 1866.

Equally, the status of a proto central bank at the top of the payments pyramid derived from the fact that it was perceived to be ‘safe’ – that is, an institution with a large capital base, holding high quality assets. So a commerciallyoriented central bank would also need to be concerned about its own soundness. This would give it incentives to be careful about to whom it should provide settlement accounts and to monitor these banks; one can think of this as an early form of banking supervision. In addition, it also had to weigh carefully the advantages of providing lender-of-last-resort assistance to the banking system to avoid a drop in its revenue stream against the risk of lending to an insolvent institution and making a loss that could decrease its capital base and threaten its reputation as the supplier of the ultimate settlement asset. As a result, proto central banks were more likely to let healthy banks go down than risk lending to unhealthy banks by mistake. Hence, in a fractional reserve system, cent...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Figures
• Tables
• Contributors
• Foreword
• Acknowledgements
• General introduction
• Part I Payment systems and public policy
• Part II New approaches to modelling payments
• Part III Current payment policy issues
• Part IV Policy perspectives on the future of payments