V. Controlling Settlement Risk in the Global Foreign Exchange Market
- International Monetary Fund
- Published Date:
- September 1996
The global foreign exchange market is at the center of the international financial system and its importance has grown even more with the ongoing globalization of finance and financial institutions. It is now the largest, the most liquid, the most innovative, and the only 24-hour global financial market in the world. In April 1995. when the BIS’s fifth triennial survey of foreign exchange markets was completed, the average daily net turnover in this market reached nearly $1.2 trillion, up from $800 billion in 1992, with about 150,000 wholesale currency transactions daily. Most of these transactions involved the dollar, the yen, and the deutsche mark, with a large number exceeding $500 million (see Table 5).27 Although the growth in cross-currency financial transactions has clearly contributed to the growth in foreign exchange markets, some have argued that the causation also runs the other way, namely, that a highly efficient and low-cost foreign exchange market has contributed to the growth in cross-currency transactions finance, particularly the growth of short-term cross-border money market and derivative transactions.
Because the foreign exchange market is at the core of the international financial system and the global network of interbank payments, a disruption in this market could have serious consequences for global trade and finance and for the international banking system. The risk of such a disruption arises from the possibility that sudden doubts about the solvency of one of the major market participants could cause its counterparties to withhold or delay payments for foreign exchange transactions that are not yet settled, thereby aggravating the difficulties, and possibly causing a chain reaction of other settlement failures. According to some of the major foreign exchange dealers, it is not uncommon for a major international bank to have unsecured overnight foreign exchange settlement exposures that exceed its capital. The risk of such a chain reaction of settlement failures is probably not large, but it may well have grown in line with the growth in trading volume and the size of transactions. The serious consequences of a disruption in the global foreign exchange market, should it occur, has become an important concern to the Group of Ten central banks and to market participants.
Risks in the Global Foreign Exchange Market
A key worry is that risk in the foreign exchange market is concentrated in a relatively small number of systemically important international banks. Indeed, the ten most active international foreign exchange dealers in London—all of them banks—account for nearly half of all trading in London; and 75 percent of all foreign exchange transactions in London are between banks. Thus any disruption in the foreign exchange market could adversely affect the stability of the global banking system. Furthermore, central banks are keenly aware that they may inadvertently be taking on more risk as lender of last resort to this market than they would prefer. It is unclear whether the risk of a disruption in the foreign exchange market and its impact on the global banking system, have been adequately priced into foreign exchange transactions. Indeed, market participants are quite candid in their view of the matter: most prefer to deal with money center banks because these banks have access to liquidity facilities at the central bank, suggesting that foreign exchange transactions are so cheap precisely because the risks are not currently borne by the market participants, but, instead, have been shifted to some extent to the central banks.
The concerns of central banks regarding foreign exchange settlement risk is related to their concerns about the intraday payments-related credit generated in many of the major wholesale payments systems in the Group of Ten countries. There too, the question arises whether the large volume of trading in money and derivative markets is facilitated by the intraday credit provided by the central banks. Significant reform efforts are being undertaken to reduce such intraday payments-related credit. For example, the U.S. Federal Reserve System has reduced the payments-related daylight overdrafts (frequently in excess of $70 billion) it extends to the banking sector during a normal business day by imposing fees on overdrafts, and European Union countries have agreed to introduce RTGS in their wholesale payment systems. However, public sector involvement in the payments and settlement aspects of the foreign exchange market has so far not been very extensive, despite the fact that the intraday, or even overnight, credit created through foreign exchange transactions is of a similar order of magnitude and is more concentrated than the wholesale payments-related credit in national payments systems related to domestic flows.28
An important part of the difficulties in settling payments in the foreign exchange markets arises from the cross-currency nature of the settlement procedures. It is not possible currently to make final payments simultaneously on the two legs of a currency transaction (as would occur in domestic single-currency financial transaction among banks that settle on the books of their central bank) because settlement must occur through the payments systems of the two countries, which are not always open at the same time. The inability to finalize payments simultaneously gives rise to the possibility that after one counterparty has already delivered funds on its side of the transaction, the other party might encounter financial difficulties and fail to deliver funds, as happened when Bankhaus Herstatt failed in 1974. The difference now is that the size and number of foreign exchange transactions have increased substantially, and settlement failures are more likely to involve the large players with the consequent potential for domino effects.
A recent survey of foreign exchange dealers shows that failures to make timely final payments on foreign exchange transactions occur routinely because of operational mishaps, back office procedures, and a host of other operational problems that have nothing to do with the financial condition of the counterparties.29 Most of these problems are resolved within the business day or overnight by drawing on interbank lines of credit. This procedure works for small amounts, but the magnitude of foreign exchange exposures is now so large, that it would not be uncommon for a major bank to have a $2 billion unsecured zero-interest foreign exchange settlement exposure with just one other bank as the result of one day’s transactions. A bank’s total foreign exchange settlement exposures can, and at times does, exceed its capital, and could well be in excess of its capacity to borrow short-term funds in the open market.
In the past, both officials and private participants have presumed that the probability of a payments failure by a major player in the foreign exchange market was negligible. There is now evidence that the potential for Herstatt-type events might be higher than previously thought. Four highly publicized and potentially threatening events involving foreign exchange settlement failures, or the possibility of one, have occurred since 1990: the collapse of Drexel Burnham Lambert in 1990. the collapse of BCCI in 1991. the attempted Soviet coup d’etat in 1991. and, most recently, the Barings crisis in 1995.30 Although these financial problems were resolved without important systemic consequences, each of them had the potential to develop into a global systemic problem if the major central banks had not been determined to resolve them promptly.
Foreign exchange settlement risk could be eliminated or substantially reduced by making it possible to finalize the two payments of a foreign exchange transaction at the same time. Two important changes in international payments arrangements could help facilitate this: eliminating existing gaps in operating hours of the major wholesale payment systems, and linking payments systems in real time to achieve intraday payments finality. This may sound simple enough, but the challenges in implementing these steps are formidable. They would entail significant reforms in the major domestic payment systems and an increased need for continuous coordination among the major central banks.
From a purely technical standpoint, it is clear that settlement risk can be substantially reduced. In fact, important progress already has been made in reducing the gaps in operating hours of the U.S. and Japanese and the U.S. and German payment systems.31 In addition, the major industrial countries are moving toward RTGS wholesale payment systems. However, linking the domestic payment systems of the major countries involves larger issues of public policy. Who should provide foreign exchange settlement services, the public sector through the official wholesale payment systems or the private sector through profit-oriented privately run payment systems? Will official or private sector solutions yield the greater reduction in global systemic risk? How will the status of reserve currencies be affected?
A major challenge in resolving these issues is to quantify the intertemporal costs and benefits (public and private) of substantially reducing settlement risk. There also are difficult practical and policy choices that are required before the necessary reforms and the linkage of payment systems would become politically acceptable to all of the major players. For example, achieving intraday payments finality for foreign exchange transactions on domestic systems would require central banks to supply credit, and incur credit risk, in order to smoothly provide intraday payments finality.32 The U.S. Federal Reserve System currently has an average of $50 billion a day in daylight overdrafts on Fedwire, and is therefore exposed to the possibility of incurring large losses in the event of defaults. If it were to decide to settle the large volume of gross dollar foreign exchange transactions—which are now first netted in New York on the Clearing House Interbank Payments System (CHIPS)—its credit exposure could increase by more than 40 percent. In settling domestic transactions, central banks have managed their credit exposures by shifting some of the implicit costs back to market participants by charging fees or requiring collateral. This has raised private settlement costs and encouraged the development of private sector arrangements for reducing the number of transactions that need to be finalized on the official systems. Even in some of the major countries where intraday finality is routine for domestic transactions—the United States, for example—the private sector chooses not to settle the bulk of their gross foreign exchange transactions on these systems.33 One reason why such private arrangements tend to be cheaper than the official settlement systems, is that they benefit from the fact that the major bank participants are covered by a too-large-to-fail safety net.
Establishing formal payment systems linkages—the second necessary step to substantially reduce settlement risk—also poses difficult policy choices for the major central banks. Direct linkages would increase exposure of payment systems and their participants to cross-bor-der spillovers from disturbances in other countries, and it ultimately entails providing multicurrency settlement services on domestic payment systems. There also would be the need for closer international coordination of macroprudential, supervisory, and lender-of-last-resort policies. Alternatively, the central banks could let private clearing agents act as buffers between market participants and the official payment systems, but this too would involve challenges for the central banks including the establishment of effective operations for multicurrency settlement; the regulation and supervision of the agents to ensure integrity and proper risk management; and the enforcement of legally binding rules for cross-border operations, internal loss-sharing, and access to lender-of-last-resort facilities.
The Group of Ten countries have undertaken considerable efforts to advance the understanding of the nature and implications of cross-currency settlement risk,34 culminating in the most recent paper by the CPSS, which concluded in March 1996 that the greatest reduction in foreign exchange settlement risk will come from improvements in banks’ own internal risk management practices, including greater use of existing netting arrangements. The report proposed a series of private sector initiatives to achieve this over the next two years, including actions by banks to improve control over foreign exchange settlement procedures, by industry groups to provide risk-reducing multicurrency services, and by central banks to induce rapid private sector progress. At the end of two years, the CPSS will decide if further official action is needed. The adoption of this private sector strategy may lead to reductions in, and better management of, the Herstatt type of settlement risk, and it does not prevent the central banks from quietly forging ahead with the necessary reforms. But there is as yet no firm timetable or modality for achieving major reduction in foreign exchange settlement risk. Market participants generally expressed a preference for a more comprehensive and coordinated approach to payments system reforms as a means of reducing foreign exchange settlement risk, rather than the experimental uncoordinated private sector initiatives currently under way.
As noted earlier, many of the major players in the foreign exchange markets find the official payments systems too expensive for settling the large and growing number of gross foreign exchange transactions. The foreign exchange business is highly competitive and profit margins are only a few basis points. These costs have encouraged dealing banks to establish private bilateral and multilateral netting arrangements.35
Netting arrangements allow member counterparties to reduce the number of final payments and amounts involved. This reduces settlement exposures, improves liquidity management, and reduces capital requirements to the extent netting agreements are recognized by the Basle Committee. Bilateral netting can reduce exposure by as much as 50 percent, depending on the number of participants and the nature of their payments to one another.
In March 1996. the Group of Twenty banking group proposed a multilateral multicurrency clearing bank as a possible mechanism for reducing costs and for substantially reducing the Herstalt-type foreign exchange settlement risk. The development of this and other private proposals is being encouraged by the CPSS. While a multilateral multicurrency clearing bank could be set up in a number of different ways, it is reportedly envisioned by the Group of Twenty as a clearing bank that would link gross currency payments, matching and settling the payments one at a time through the crediting or debiting of accounts held at the clearing bank by users.36 In order to effect payment finality, the clearing bank ideally would have access to the national RTGS systems of the currencies settled within the bank.37 but it could also achieve finality, albeit more expensively, by holding sufficiently large balances of its clearing members. If the national RTGS payments systems had overlapping hours, Herstatt-type risk could also be reduced substantially by allowing the clearing bank to settle both legs of a currency payment simultaneously during the window of overlapping business hours. Although risk would be much lower in this setup, the clearing bank may need to require users to place large sums in their accounts so that, during the window of overlapping hours, there are sufficient funds within the clearing bank to execute payment orders. Thus, a strain on money-market liquidity in the settling currencies during this window may result, making it somewhat harder for the central bank to maintain stable money market interest rates. The tradeoff between settlement risk and potential liquidity effects is at the heart of the discussion of this approach, which is still in its initial development stage.
Although the establishment of a clearing bank, such as the one proposed by the Group of Twenty, theoreticaily could substantially reduce Herstatt-type risk, it has drawbacks. As described above, the major central banks would have to operate their payments systems during overlapping hours. There also is some concern that the operations of such a clearing bank could potentially affect liquidity in the major money markets. Even if these problems are solved, there is the issue of costs and benefits. The obvious benefit of fully implementing RTGS systems would be the substantial reduction in the large overnight foreign exchange exposures and the associated Herstatt-type risk. This benefit has to be balanced against the costs of setting up and operating such a clearing bank.
Other private initiatives—some well under way—may go a long way. Through bilateral netting (FXNET) and multilateral netting (MULTINET and ECHO) systems, major banks have the opportunity to reduce the risk exposure in the foreign exchange market considerably, because large transactions are concentrated among the top I0-to-20 banks in the world. Although netting does not completely eliminate Herstatt-type risk, it does offer an economical way to reduce the exposure. At a minimum, netting arrangements are, and will continue to be, important until the major central banks lengthen their operating hours, and the RTGS clearing bank is established, but their importance may extend well beyond this time. The netting systems may even coexist with the clearing bank after it is established because of its cost advantage. Current netting arrangements can be further enhanced if the legality of enforcing netting can be firmly established in some countries (including some of the Group of Ten countries).
Although these private sector solutions are beneficial in some important ways, they do not remove all of the risks and may even add to them. From a systemic perspective, netting arrangements perpetuate delays in settlement, and netted transactions are still subject to Herstatt-type risk. Furthermore, there is still considerable uncertainty about the legal enforceability of net-ling arrangements in some Group of Ten jurisdictions. If netting arrangements are complemented or superseded by the introduction of global clearing houses and new multicurrency settlement procedures, different operational, legal, and financial risks will be introduced that would raise many of the same issues that central banks are concerned about in establishing direct linkages between payments systems. What all of this means is that the ability of global multicurrency clearing houses to reduce global systemic risk, rather than redistribute it or even increase it, is by no means assured.
For further details see Annex III.
The foreign exchange market has not yet adopted the formal regulatory or sell-regulatory structure that can be found in other major markets, such as the stock exchanges, fulures exchanges, equity markets, or the systems of primary dealers in some of the major government securities markets
Foreign Exchange Committee (1994).
See Bank for International Settlements (1996b).
The Federal Reserve has agreed to open Fedwire at 12:30 a.m. in 1997. overlapping with the final business hours in Japan and much of Europe’s business day.
Central banks have chosen various alternatives for managing these exposures: in the United States, the Federal Reserve impose credit limits and charges fees for daylight overdraft; payments systems in the European Union countries are being restructured into RTGS with collateralized lending by the central banks. In the Bank of Japan’s settlement system, most transactions are settled at fixed times during the day, although the network has the capability of RTGS.
To the extent that there are cost differences among the major wholesale systems, there are incentives to redenominate transactions into the low-settlement-cost payments medium—for now, the dollar. This could raise broader macroeconomic and macroprudential considerations, including exchange market pressures and their implications for domestic monetary policy, the major international banks, and financial systems.
Netting arrangements periodically aggregate the amounts owed between or among participants for a specific currency for delivery on a specific date, thereby lowering the payment flows requiring settlement. FXNET and the SWIFT Accord systems currently provide bilateral netting to 70 and 27 banks, respectively. ECHO and MULTINET are multilateral, multicurrency netting systems. ECHO is based in the United Kingdom and has 13 bank members; MULTINET is not yet operational, planning to begin in late 1996 with, at a minimum, its 8 founding members. See Box 13 in Annex III for more details on netting systems.
This outline of the Group of Twenty proposal is based on information received prior to May 1996. No formal documentation has yet been released.
The clearing bank would also need central bank coordination to assure “good funds” were provided so as to guarantee payment finality of both legs of a currency transaction.