Analysis of Systemic Risk in Multilateral Net Settlement Systems
By: Sujit Chakravorti
Source: Journal of International Financial Markets, Institutions and Money
Technological advances in telecommunications and computers along with the rapid growth of the volume and value of transactions in financial markets over the last two decades have contributed to the explosive growth of daily interbank funds transfers. Today, the gross value of payments cleared and settled by a major industrialized country’s large-value funds transfer systems in a few days is equivalent to the country’s annual gross domestic product. The sheer volume of payments over the domestic large-value funds transfer systems has led policymakers to consider various measures to reduce the probability of a stoppage of one of these systems.
The apparatus used to transfer monetary value is the payment system. Large-value payment systems are at the foundation of any nation’s financial system. Horii and Summers (1994) (p. 74) explain the importance of safe and efficient large-value payments systems: ‘‘The safe and efficient operation of large-value transfer systems has a bearing not only on the markets they directly serve but on a nation’s whole financial system.’’
A safe payment system minimizes the risks involved in the transfer of monetary value. These risks are often collectively referred to as payment system risk These risks include liquidity risk, settlement risk, and systemic risk. Liquidity risk is the risk that a participant does not have good funds at the time of settlement but can provide them at a later time. Settlement risk is the risk that one party to a transaction does not deliver the underlying asset in its entirety at the specified settlement time. Systemic risk, as defined by the Bank for International Settlements (1992) (pA2-7), is "the risk that the inability of one institution to meet its obligations when due will cause other institutions to be unable to meet their obligations when due."
Although financial analysts agree that large-value payment systems should be safe and efficient, debate continues regarding their optimal design and operation. One major point of departure in payment system design is whether to settle payments individually known as gross settlement or to settle the net position at the end of some specified time known as net settlement.
Recently, major industrialized countries have been moving towards real time gross settlement (RTGS) systems (Folkerts-Landau et al., 1996; Bank for International Settlements, 1997; Chakravorti, 1997b). In true RTGS systems, each payment is settled in real time with central bank reserves. All of the European Union member countries are in the process of establishing RTGS systems (Working Group on EU Payment Systems, 1993). Specifically, countries that have adopted the Euro as their currency have each established a national RTGS systems that are interlinked via the Trans-European Automated Real-time Gross settlement Express Transfer system, TARGET (European Central Bank, 1998). In addition, the Bank of Japan has decided to convert its designated-time settlement component of BOJ-NET to RTGS by the end 2000.
The primary motivation to move to RTGS systems is to reduce payment system risk. Although systemic risk is substantially reduced in RTGS systems, the potential for payments gridlock exists. Payments gridlock occurs when the flow of payments is disrupted because participants are waiting to receive payments before sending them. To avert such gridlocks, central banks may extend intraday credit or establish a queuing mechanisms. However, by extending intraday credit, the central bank may be exposed to greater credit risk. The introduction of a queuing mechanism may delay settlement. Folkerts-Landau et al. (1996) describe time delays that occur on the Swiss queuing system: ‘‘. . . in 1993, on average only 51% of the payment volume was settled by 2:00 P.M., while 95% was initiated by that time’’ on the Swiss RTGS system.
Payment systems that net transactions multilaterally allow banks to settle one net position at the end of the day. During the day, banks receive and send payment messages to other banks. Banks that owe funds are called due-to banks, and banks that receive funds are called due-from banks. A clearinghouse acts as an intermediary and collects good funds from due-to banks and releases good funds to due-from banks. Final settlement occurs when the clearinghouse has successfully completed this process.
The primary reason that net settlement systems exist is to reduce the cost to settle a given value of payments. If banks had to settle payments individually, they would on average need to hold more reserves. According to Padoa-Schioppa (1994) (p36), a ‘pure’ RTGS would require, for a given volume of gross transactions, a greater intraday supply of central bank money than a ‘pure’ netting system. Banks face higher costs by holding a greater level of reserves, because reserves held at the central bank are non-interest bearing assets.
In such systems, banks often interpret payment messages as good funds and release the funds to their customers. Federal Reserve Bank of New York (1991) (p17) explains the behavior of the Clearing House for Interbank Payments System (CHIPS) participants: ‘competitive pressures have, however, resulted in a common practice of permitting receivers access to funds immediately’. The time difference between when funds are released and when they are settled increases settlement risk. The risk that participants face in a net settlement system where payment messages are interpreted as good funds is the inability of other participants to deliver funds at settlement time. The potential inability of a participant to settle due to the inability of others to settle is the primary risk that policymakers attempt to minimize.
Although the defaulting participant may not have dealt with every other participant directly, its default could affect the whole settlement process. For example, if participant A sends a payment message to participant B and participant B uses it to make a payment to participant C, the default of participant A could affect participant C. Although many countries are implementing RTGS systems, net settlement systems continue to exist. In Japan and the United States, net settlement systems operate along side of gross settlement systems. The existence of these and other net settlement systems implies that net settlement systems offer benefits not found in RTGS systems. The dual existence of RTGS and net settlement systems increases liquidity and reduces cost to payment system participants. Some analysts have argued that policymakers need not choose between one or the other because there are advantages of having net settlement systems operate along side of RTGS systems (New York Clearing House Association, 1995). Furthermore, Folkerts-Landau et al. (1996) argue that the ongoing reforms to convert to RTGS systems may lead to private sub-netting systems that would serve as low-cost alternatives.
Policymakers and payment system operators have adopted various risk-reducing measures for net settlement systems. Such risk-reducing measures include collateral requirements, net debit caps, restricting participation, and loss-sharing arrangements. A net settlement system could mimic an RTGS system by requiring full collateralization of all payments. New York Clearing House Association (1995) estimates that in such a system, participants would face an opportunity cost of $180 million annually as opposed to $5.9 million for covering the largest net debit.
This article studies the effect of different size liquidity shocks resulting from bank defaults on the surviving banks based on differences in the level of risk-reducing measures. In the next section, the large-value net settlement system in the United States — CHIPS — is discussed. In Section 3, liquidity and the role of the central bank during crises is analyzed. In Section 4, a model is constructed to determine the amount of reserves held by due-to banks given an expectation of the potential number of bank defaults that may affect their end of period settlement. The model predicts the threshold point where the payment system operator can no longer collect good funds from the remaining banks to complete the settlement process. In Section 5, the central bank’s ability to avert a systemic collapse is considered. In Section 6, payment system policy options to reduce systemic risk are investigated. Section 7 concludes the article.