Paper on the Feasibility of a Foreign Exchange Transactions Tax - March 27, 1999

A Feasible Foreign Exchange Transactions Tax

Rodney Schmidt*
Research Associate
North-South Institute
55 Murray Street, Suite 200
Ottawa, Ontario, Canada
K1N 5M3
Fax: (613) 241-7435

Program Advisor
Vietnam Economic and Environment Management
International Development Research Centre
40 Nguyen Van Ngoc
Hanoi, Vietnam
Fax: (84-4) 766-0469
E-mail: veem@hn.vnn.vn


March 1999

Abstract

There is virtually no formal infrastructure for trading foreign exchange. Traders in major banks around the world communicate directly with each other or through a broker. By contrast, the infrastructure for settling foreign exchange trades is becoming increasingly formal, centralised and regulated. This is due to new technology subject to increasing returns to scale and to cooperation between trading and central banks to reduce and eliminate settlement risk.

Settling a foreign exchange trade requires at least two payments, one for each of the currencies traded. Settlement risk is eliminated when these two payments are matched, traced to the original trade and made simultaneously. The technology and institutions now in place to support this make it possible identify and tax gross foreign exchange payments, whichever financial instrument is used to define the trade, wherever the parties to the trade are located and wherever the ensuing payments are made.

JEL Classification : F3, G2

Keywords : Tobin tax; foreign exchange; settlement; payment systems

* This paper was written for the North-South Institute. I would like to thank Roy Culpeper, President, the participants in a North-South Institute seminar and, especially, Patrick Georges for helpful comments on earlier drafts, without implicating them in the views expressed in this paper or in any remaining flaws.


A Feasible Foreign Exchange Transactions Tax
A tax on foreign exchange transactions was proposed 20 years ago by Nobel laureate James Tobin (Tobin, 1978). Tobin intended it to increase monetary policy independence and reduce nominal exchange rate volatility when capital flows freely across international borders. Recently Tobin’s tax has also been promoted to defend exchange rates from speculative attacks, manage transitions between exchange rate regimes and finance international public projects (Eichengreen et al., 1995; Felix, 1995; ul Haq et al., 1996).

Tobin’s tax is thought to be impracticable, however, by both those who object to it as an interference in an efficient market and those who support it in principle (Frankel, 1996; Garber and Taylor, 1995; ul Haq et al., 1996). Some think it impracticable because it would have to be applied globally:

...enforcement is a big problem. Certainly if some countries adopted the Tobin tax but others did not, the foreign-exchange trading would simply move to where it was not taxed. For this reason, everyone agrees that it would have to be imposed in virtually all countries, large and small. This would require more widespread support than seems possible politically (Frankel, 1996, p. 156).

Even if Tobin’s tax were internationally legislated, for example by making it a condition of membership in the International Monetary Fund (Eichengreen, Tobin and Wyplosz, 1995), others think it impracticable because foreign exchange trades are hard to monitor (Garber and Taylor, 1995). The international foreign exchange market is decentralised and without a systematic and comprehensive system for recording individual trades. Instead, to regulate private banks and financial institutions and measure capital flows, central banks and supervisory bodies require them to register overnight balance sheet positions. However, in today’s 24-hour global marketplace foreign exchange traders can hide positions by shifting them between branches in different time zones to stay continuously within working hours. Or traders can use derivative financial instruments that do not show on balance sheets (Garber, 1998). Finally, traders can avoid the foreign exchange market altogether by buying and exchanging securities, such as bonds or treasury bills, denominated in different currencies (Garber and Taylor, 1995).

Every foreign exchange trade has to be settled, however, with an exchange of assets, usually bank balances, denominated in different currencies. In contrast to the market for trading foreign exchange, international infrastructure in place to settle foreign exchange trades is increasingly formal, centralised and regulated. Can Tobin’s tax be applied, then, to foreign exchange trades as they are settled? The answer depends on whether the settlement infrastructure allows a one-to-one correspondence to be established between foreign exchange payments and their originating trades and on whether the tax can be applied to payments made offshore.

We contend that a transactions tax applied to foreign exchange payments is feasible. We base our contention on an analysis of recently established settlement technology and institutions designed to reduce and eliminate settlement risk. ‘Settlement risk’ exists when one party to a transaction makes an irrevocable payment of a currency or delivery of a security denominated in that currency before its counterpart makes the opposing payment to complete the exchange of assets and settle the transaction. The risk is that the counterpart will then default on the payment. Settlement risk is eliminated when the two payments are matched, traced to the original foreign exchange trade and made simultaneously. This is known, in the case of exchanges of bank balances, as ‘payment-versus-payment’ (PVP) settlement and, in the case of exchanges of securities, as ‘delivery-versus-payment’ (DVP) settlement. The technology and enforcement mechanisms underlying this capability at the various settlement institutions for foreign exchange are sufficient for a feasible Tobin tax.

Section one following outlines the argument that Tobin’s tax is feasible if it is applied to foreign exchange transactions as they are settled. Section two specifies the target of Tobin’s tax, namely individual trades made in the interbank or wholesale foreign exchange market. Section three describes foreign exchange settlement infrastructure. Section four discusses the minimal infrastructure requirements for tax feasibility. Section five concludes by suggesting how to implement Tobin’s tax.

1. A Tax on Interbank Foreign Exchange Payments is Feasible
As a result of rising foreign exchange trading volume, new technology and efforts to reduce settlement risk, the infrastructure in place to make interbank foreign exchange payments is increasingly formal, centralised and regulated. Major trading banks and other financial institutions (‘banks’) can now make foreign exchange payments among each other, by netting offsetting payments periodically in offshore netting systems or securities clearing houses, or through the domestic payment system of the country that issues the relevant currency. Usually both methods are used: offsetting payments are eliminated and then net payments are made in domestic payment systems. Payments cannot be made otherwise, unless a bank drops out of the interbank market and into the retail market for foreign exchange, where settlement is less formal and decentralised and transactions costs are correspondingly higher.

For a tax on interbank foreign exchange payments to be feasible, it must be possible to distinguish foreign exchange payments from domestic financial payments, tax gross foreign exchange payments and enforce the tax in offshore netting systems. Three features of the current settlement infrastructure in the interbank foreign exchange market accomplish this.

First, domestic payment systems eliminate settlement risk for domestic financial transactions. A domestic financial transaction is one where both payments made to settle the trade are denominated in the domestic currency. Most domestic payment systems in the G-10 countries, as well as in many other countries, are Real Time Gross Settlement (RTGS) systems, meaning that they process gross payments rather than batches of payments with netting. The significance of RTGS systems is that they support PVP settlement for domestic transactions. That is, domestic payments are matched and traced to the original financial trade and then processed simultaneously. If a payment is not matched it can be treated as a foreign exchange payment and taxed.

Domestic payment systems with RTGS capability do not currently support PVP settlement of foreign exchange transactions. The reason is that settling a foreign exchange trade requires two payments to be made in different domestic payment systems, which are often located in different time zones. If the working hours of the two payment systems do not continuously overlap, the payments cannot be matched and processed simultaneously. However, by mid-2000 a central global settlement bank, the Continuous Linked Settlement (CLS) Bank, will open. The CLS Bank will run around the clock throughout the week and, with direct links to numerous domestic payment systems, will support PVP settlement for foreign exchange transactions. Then foreign exchange payments at the CLS Bank can be directly identified and taxed.

Second, offshore netting systems operate PVP foreign exchange trade settlement by netting. That is, foreign exchange payments submitted for netting are matched and traced to the original foreign exchange trade before netting proceeds for both payments simultaneously. This is necessary to maintain the integrity of the netting system since, when payments are accepted for netting, the original foreign exchange payment obligations between the two parties to the trade are replaced by payment obligations between each party and the netting system. PVP netting ensures that the netting system does not take a net creditor, debtor or foreign exchange position. Thus, gross foreign exchange payments can be identified and taxed as part of the netting process.

Third, central banks or their supervisory bodies regulate offshore netting systems and can enforce those regulations. The same mechanisms can be used to enforce a foreign exchange payments tax.

The right of central banks to individually and collectively regulate offshore netting systems was codified by the BIS Committee on Interbank Netting Schemes of the Central Banks of the Group of Ten Countries in the Lamfalussy Minimum Standards in 1990, and re-affirmed by the same Committee in 1998. The Standards provide for measures to reduce settlement risk and otherwise promote domestic and international financial stability. Central banks enforce the regulations by refusing access of non-cooperating netting systems to the domestic payment system and by sanctioning banks that are members of both the domestic payment system and the offshore netting system. This is effective because offshore netting systems cannot process a payments of a currency unless they have access to the domestic payment system. Further, banks are not accepted as members of the netting system unless they trade in the interbank market for foreign exchange, which is to say they participate in a domestic payment system. Netting systems and netting activity are also easy to identify because the technology is subject to increasing returns to scale and therefore requires many participants to be viable. Also, most netting services, whether in formal systems or informally between pairs of banks, are delivered by a single third party, the Society for World-side Interbank Financial Telecommunications (SWIFT). SWIFT also provides a standardised and automated communications network for banks and netting and domestic payment systems.

The rest of the paper elaborates this argument.

2. The Tax is on Interbank Foreign Exchange Transactions
Foreign exchange transactions occur in the retail market for foreign exchange, between end-users or between major trading banks and their customers, and in the interbank market for foreign exchange among major trading banks. This paper considers a tax on interbank transactions. We select interbank foreign exchange transactions exclusively because they are the natural target for a tax that aims to modify exchange rate behaviour and because, with a well-defined and regulated settlement infrastructure that supports a transactions tax, the potential for avoiding the tax by shifting trading outside the interbank market is small.

Participants in the interbank market for foreign exchange are called ‘tier one’ banks, as opposed to those in the retail market which may be ‘tier two’ or ‘tier three’ banks (the appendix extends this description and interprets figures 1-4 which illustrate it and other statements made in the paper). Banks achieve tier one status because of their financial and trading importance and reputation for creditworthiness. Tier one status entails direct access to the domestic payment system, which often means having an account with the central bank, and the selection of tier one banks usually therefore involves the explicit or implicit agreement of the central bank as the ultimate guarantor of domestic financial stability. Like market-makers on securities exchanges, tier one banks make and accept offers to trade with other such banks as a matter of routine. These banks normally locate in major financial centres for the liquidity available to facilitate domestic currency payments. This is also why banks trading in a foreign currency maintain trading balances in correspondent banks in a financial centre in the country that issues the currency.

Tier two and lower banks operating in the retail market for foreign exchange incur substantially higher transactions costs than tier one banks. There are higher processing costs due to lost scale economies, more exposure to settlement, credit, liquidity and operating risks and worsened trading terms due to lost creditworthiness. To minimise these additional costs retail banks also keep trading balances at correspondent banks with access to the domestic payment system.

We do not address a tax on open currency positions on overnight bank balance sheets (Kenen, 1996), from which international capital flows are measured. Overnight balance sheets reflect neither within-the-day and after-hours deals nor off-balance sheet trades, such as those using derivatives. These omissions are important because banks tend to avoid taking open overnight balance sheet positions, regarding them as too risky because of the high volume of trading that continues abroad (Frankel et al., 1996). They do take open currency positions during the day and off balance sheets (Lyons, 1991, 1995; Lyons and Rose, 1995; Garber, 1998). Overnight positions are taken primarily by non-bank financial institutions, such as hedge funds (private closed-end investment funds) and institutional investors (mutual and pension funds and insurance companies) (IMF, 1993). These retail transactions are processed in the interbank market via correspondent banks. The information gleaned from them and other transactions made by banks in the interbank market determines the daily exchange rate. For these reasons an effective Tobin tax must be assessed on individual interbank foreign exchange transactions

3. How Interbank Foreign Exchange Transactions are Settled
Settlement infrastructure in both foreign exchange and securities markets is evolving in response to rising trading volume, new technology and heightened awareness of settlement risk. The tendency of the evolution is to formalise and centralise to exploit economies of scale and reduce or eliminate settlement risk. As a result, settlement systems in the interbank foreign exchange market look more like those on organized securities exchanges, which themselves are becoming more centralised through sharing of clearing houses.

This section describes the various ways to settle foreign exchange transactions in the interbank market. We show how and why settlement methods are becoming more formal and central and how the links between various settlement institutions are strengthening. We describe the end result of these changes, the form of which is expected in mid-2000 with achievement of full PVP foreign exchange settlement (CLS, 1998). We also discuss some of the implications of these settlement methods for Tobin’s tax. Section 4 identifies features of the settlement infrastructure needed for tax feasibility.

There are three ways to settle a foreign exchange transaction in the interbank market. First, one can make a payment in the domestic payment system of the country that issues the currency being sold. Second, one can net offsetting payments deriving from a numerous transactions in a netting system. Third, in the future one may be able to use a ‘contract for differences’ (CFD) which, like derivative financial instruments available for domestic financial transactions, only requires payment of profits or losses realised as a result of movements of the exchange rate relative to the notional principal amounts traded. We discuss these settlement methods by turns. In practice, they are used in combination and we also discuss the resulting connections between them.

3.1 Domestic Payment Systems
The most direct way to settle an interbank foreign exchange transaction is for each party to the trade to make a gross payment, consisting of a transfer of bank balances for the full sale of the currency, in the domestic payment system of the country that issues the currency (see figure 1 for an illustration). This means that the two or more payments needed to settle the transaction are made in different systems possibly located in different time zones. Unless the payments are coordinated and the operating hours of the payment systems overlap, the payments occur at different times and there is settlement risk. By contrast, in the case of domestic financial transactions, where all payments are denominated in the domestic currency, there is no settlement risk if the linked payments are made simultaneously in the domestic payment system.

Suppose, as is currently the case, foreign exchange transactions settled in domestic payments systems were exposed to settlement risk. Then it would not be possible to identify payments associated with a foreign exchange transaction, because only the domestic currency payment passes through the domestic payment system and it is not matched with its counterpart in another payment system. One could, however, do the opposite and identify payments associated with a domestic financial transaction, since linked domestic payments can be matched in the same payment system and processed together. If direct payment were the only way to settle foreign exchange transactions then Tobin’s tax could be assessed on all payments that were not identified as domestic financial payments.

When PVP foreign exchange settlement is available in domestic payment systems it will be possible for gross payments deriving from a common foreign exchange transaction to be matched and processed simultaneously. Then it will be easy to identify and tax foreign exchange payments directly.

There are two ways to achieve PVP foreign exchange settlement in domestic payment systems. One is to have all domestic payment systems operate around the clock so that opening hours will overlap continuously. The other is to create a global settlement institution to process payments continuously 24-hours a day with direct links to domestic payment systems. The latter is the objective of the new CLS Bank which, with headquarters in New York and operations in New York and London, is expected to open in mid-2000 (figure 4 illustrates) (CLS, 1998).

The CLS Bank concept was developed at the urging of the G-10 central banks by a group of major foreign exchange trading banks which organized in 1995 and created CLS Services Ltd (CLSS) in 1997. CLS Bank is designed to settle world-wide trading in all the major currencies (Financial Times, 27 June 1997). It will eliminate settlement risk by simultaneously making the two or more payments of a foreign exchange transaction on its own accounts. Payments will be final by legal dispensation of the countries whose currencies are being settled and, in most cases, will be operationalised by permitting CLS Bank to hold settlement accounts directly with the relevant central banks. Effectively CLS Bank will be an extension of participating domestic payment systems.

3.2 Netting Systems

An alternative way to settle interbank foreign exchange transactions is to periodically eliminate offsetting payments among banks in netting systems (figure 2). It is because of prior netting that Tobin’s tax cannot be imposed only in domestic payment systems.

Netting was traditionally done informally and between pairs of banks. It would be very hard to monitor such activity. Since the late 1980s, however, netting has increasingly been done in formal systems exploiting new technology and economies of scale and using standard processes (IMF, 1996; Summers, 1991). They are easy to identify and access.

The new netting technology permits continuous multilateral netting of gross payments in multiple currencies of members located around the world. Netting is more effective in terms of the volume and value of payments that can be offset against each other the greater the number of participants and currencies. For this reason the leading multilateral and multicurrency netting systems, Exchange Clearing House Organization (ECHO) and Multinet International Bank (MIB), are merging, with each other and with CLS Bank, to create a worldwide one-stop foreign exchange settlement institution. Netting systems are now able to settle up to 90% of foreign exchange payments, compared to about 25% for bilateral netting (Perold, 1995).

Netting software and interfaces are also becoming standardised and automated, in both formal systems such as ECHO and in informal bilateral activity. Netting for both is now done by a common third party, SWIFT, a non-profit cooperative owned by banks, through its Accord service (BIS, 1997; BIS, 1998a). Accord is, in effect, a virtual central netting system encompassing both physically central netting systems (such as the new ECHO in London) and physically dispersed netting between pairs of banks.

3.3 Contracts for Differences
Finally, a potential new way to settle foreign exchange transactions is via CFDs. Like some domestic financial derivative instruments, CFDs would allow parties to a trade to settle foreign exchange transactions by making a net payment equal to the difference between the purchase and sale transactions prices of the underlying foreign exchange contract, without having to transfer principal amounts. In contrast with netting systems, which net accumulated multilateral payments periodically, CFDs would permit continuous bilateral netting.

CFDs or similar derivative instruments are not yet available in the foreign exchange market (Garber, 1998). They are, however, being considered for hedging or speculation trades. Their successful introduction depends in part on whether necessary adjustments to market trading and settlement conventions can be made to support them (BIS, 1998b). In particular, just as in the case of over-the-counter (OTC) domestic financial derivative instruments, standard CFD contracts must specify the notional principal currency amounts traded, on which profits or losses due to movements in exchange rates are calculated. Electronic templates for this purpose would be devised with the status of legal contracts. ‘Plain vanilla’ contracts would incorporate notional principal amounts directly, while the market value of more complex contracts would be determined by a financial model, just as is already done routinely for domestic derivative instruments. Tobin’s tax could then be automatically assessed on notional principal amounts or the market value, as coded into the CFD template, when net profit or loss payments on CFDs are processed by netting and domestic payment systems. For example, Accord provides automated netting services for existing OTC derivatives, both in ECHO and bilaterally (BIS, 1998a).

3.4 Sequential Settlement
These three ways of settling foreign exchange trades are typically accessed sequentially. That is, in future a CFD may be used to define a trade. Then payments arising from numerous trades involving all kinds of financial instruments are netted, typically in an offshore netting system or securities clearing system. Finally, net amounts due are paid in domestic payment systems (figure 2illustrates the connection between netting and domestic payment systems for foreign exchange cash transactions and figure 3 does the same for foreign exchange securities transactions). Typically participants in netting systems are also members of the domestic payment system of the relevant currency, or keep foreign currency trading balances with a correspondent bank that is a member of the domestic payment system.

Because foreign exchange settlement institutions are used in combination, both Tobin’s tax and PVP foreign exchange settlement must apply in and be coordinated across the institutions. The rest of the paper shows how this can be done in netting and domestic payment systems. Since CFDs will be processed in the same institutions the discussion is complete.

4. Settlement Features Needed for a Feasible Tobin Tax
The minimal features required for tax feasibility are determined by the need to identify foreign exchange payments, tax gross payments and prevent evasion regardless of the financial instrument used for the trade, the location of the parties to the trade or where the payment is made. In light of this and the preceding discussion, necessary and sufficient conditions for a feasible Tobin tax on interbank foreign exchange transactions are three-fold.

First, domestic payment systems must continuously process gross payments (as submitted to the payment system, although they may be net payments from netting systems) rather than following the tradition of periodically processing payments in batches. Payment systems with this capability are known as Real Time Gross Settlement (RTGS) systems. They support PVP settlement of domestic financial transactions since linked payments in the common domestic currency are all made in the domestic payment system. Second, netting systems must operate PVP netting (DVP in securities clearing houses), whereby linked payments are matched and traced to the original foreign exchange trade before being netted simultaneously. Third, central banks or their delegates must enforce Tobin’s tax in offshore netting systems and coordinate the tax across netting and domestic payment settlement systems.

These three conditions are also necessary, but not sufficient, for full PVP settlement of foreign exchange trades. For sufficiency one must also have PVP foreign exchange settlement in domestic payment systems. Because Tobin’s tax is feasible without full PVP foreign exchange settlement there is no disincentive to trading banks to help achieve it. Indeed, once the CLS Bank is operational, Tobin’s tax can be designed to encourage banks to use it by setting a higher rate on unmatched foreign exchange payments than on ones that are matched and made together.

4.1 Real Time Gross Settlement in Domestic Payment Systems
For tax feasibility domestic settlement systems process gross payments for two reasons. The first is that domestic payment systems must be able to match linked domestic financial payments, equivalent to PVP financial settlement, in order to be able to distinguish them from foreign exchange payments. The second is that the system must be able to tax gross foreign exchange payments submitted directly to the payment system. The desire to achieve domestic financial PVP settlement is the reason nearly all G-10 countries and many other countries, including Thailand, Hong Kong, Korea, the Czech Republic and, soon, China, have RTGS domestic payment systems.

In RTGS systems small processing or user fees are already routinely imposed on payments (Summers, 1991). These can be continuously and automatically collected since payments are processed individually. User fees are equivalent in application and effect to Tobin’s tax except that they are applied to all payments rather than just foreign exchange payments.

When PVP foreign exchange settlement is available in domestic payment systems via the CLS Bank it will be possible to identify and tax gross foreign exchange payments directly.

4.2 Payment-versus-Payment in Netting Systems
Most foreign exchange payments are netted before submission to domestic payment systems. Hence, Tobin’s tax must also be applied to gross payments submitted for netting at the point or during the process of netting. For this netting systems must process gross payments individually, just as RTGS domestic payment systems do, and moreover must be able to identify payments deriving from foreign exchange transactions.

Netting systems can do this because, as a result of settlement risk controls imposed by central banks and their regulatory bodies, they operate PVP settlement in the netting process. This means that netting systems match multiple payments originating in a common transaction and then net them simultaneously, or not at all.

When a netting system accepts payments for netting, the original foreign exchange payment obligations between the two parties to a foreign exchange trade are legally replaced by payment obligations between each party and the netting system. That is, the netting system itself incurs payment obligations to ensure that payments made to and from the netting system are final, just as when payments are made in domestic payment systems. To eliminate the settlement risk exposure of the netting system and ensure that the netting system does not make net loans to or receive net loans from its member trading banks (beyond certain limits allowed to enhance settlement liquidity in the netting system), the netting system processes payments associated with a common foreign exchange trade simultaneously, according to the PVP principle. Then the netting system does not take open foreign exchange positions by netting foreign exchange payments. The transfer of the payment obligation from the foreign exchange trade counterpart to the netting system occurs when the gross payments have been successfully matched (BIS, 1998b; ECHO, 1998). In the process it would be easy to automatically and systematically identify foreign exchange payments and apply Tobin’s tax.

Securities exchanges world-wide also operate DVP netting and settlement in their clearing houses, for the same reasons as foreign exchange netting systems (BIS, 1992; Borio and Van den Bergh, 1993; The Economist,9 May 1998). Thus it is also easy to apply Tobin’s tax to foreign exchange transactions intermediated by an exchange of securities (figure 3).

4.3 Enforcing the Tax in Offshore Netting Systems
Although it is technologically possible to apply Tobin’s tax in both domestic payment and netting systems, it remains to show that the tax can be enforced in offshore netting systems. There are two dimensions to this problem. One concerns the ability of central banks or their delegates to regulate recognised offshore netting systems. The other concerns their ability to identify netting systems or less formal netting activity. We address these in turn.

4.3.1 Regulating Offshore Netting Systems
The interest and ability of central banks and supervisory bodies to regulate offshore netting systems, and the interest of netting systems to comply with such regulations, stems from the institutional links between the two systems. Net amounts owed by netting systems to their member banks, and vice versa, are paid in the domestic payment system of the relevant currency. This requires formal and legally defined links between netting and domestic payment systems. Netting systems are often offshore and process multiple currencies, so they maintain such links with numerous domestic payment systems.

Netting systems and their members maintain close ties to domestic payment systems for several reasons. First, net payments made in domestic payment systems have legal status, and the regulator of the domestic payment system can also confer legality on the gross payments that were settled by previous netting. This is important in the event of a default and need to distribute losses among members of the netting system. Legal status also renders payments made in the netting and domestic system irrevocable. Second, domestic money markets are integrated with the payment system and the central bank is ready to ‘lend as a last resort’ to safeguard the integrity of the system, so credit and liquidity to support payments are to hand. Third, the domestic financial and payment system is supervised and regulated to ensure the creditworthiness of member banks and control systemic risk. Only members of domestic payment systems have access to formal netting systems (Lamfalussy Report, 1990).

Central banks or their delegated regulatory bodies are also interested in maintaining ties with netting systems. If a netting system were to fail, the viability of participants who are members of the domestic payment system would be at risk, as would the domestic financial system.

The right of central banks to individually and collectively supervise netting procedures and risk control measures in offshore netting systems, for the purposes of controlling settlement risk specifically and financial stability generally, is codified in the Lamfalussy Minimum Standards and was recently re-affirmed (BIS, 1998b; Lamfalussy Report, 1990). The Standards establish three principles governing offshore supervision, which apply irrespective of the type of financial instrument or the contractual obligation netted:

Firstly, their application should ensure that cross-border systems are subject to review "as systems" by a single authority with responsibility to consider the system’s impact in different countries. Secondly, they should provide a cooperative approach to ensure that the interests of different central banks and supervisory authorities are reflected in the oversight of any one system. Thirdly, cooperation between central banks should, in particular, help to preserve the discretion of individual central banks with respect to interbank settlements in their domestic currency (Lamfalussy Report, 1990:7).

These principles support both effective coordinated supervision of netting systems and the prerogative of individual central banks to maintain their own interpretation of measures necessary to preserve domestic financial stability. For example, ECHO, based in London, is regulated by the Bank of England in consultation with other interested central banks under the terms of the Lamfalussy Report (CLS, 1998). To process a new currency ECHO requires the permission of the central bank that issues the currency, which is granted depending on ECHO’s settlement risk management safeguards and the legal enforceability of the netting process in the domestic payment system of the currency:

We [ECHO] therefore need to work very closely with the local regulators on understanding how the local payment system works and to know the law on netting...The central banks have agreed that trades in their respective currencies may be settled within the system and that the rules, operational structures and systems of ECHO are appropriate to their national markets (ECHO, 1998).

Similarly, the CLS Bank will be formed under US Federal law and supervised by the Federal Reserve: "The [CLS] consortium’s plan is in response to a demand by central banks that the private sector find a solution to the problems of settlement risk in foreign exchange markets..." (Financial Times, 27/06/97).

Individual central banks reserve the right to unilaterally regulate any offshore netting system that processes its currency. In accordance with the Lamfalussy Minimum Standards and to control settlement risk the US Federal Reserve now requires systems that net obligations denominated in the US dollar to monitor and limit net amounts owed by each participant. The Federal Reserve also requires such systems to have procedures in place to prevent contagion effects in the event that the participant with the largest net amount owing is unable to make the payment. These procedures may include reversing the netting operation yielding the payment due and using collateral or the system’s capital to cover the original gross amounts due (Federal Reserve, 1994).

To enforce these regulations, the Federal Reserve reserves the right to prohibit the use of Federal Reserve payment services to support funds transfers that are used to settle, directly or indirectly, obligations on large-dollar multilateral netting systems that do not meet the Lamfalussy Minimum Standards. ...Moreover, in order for Federal Reserve Banks to monitor the use of intraday credit, no future or existing privately operated large-dollar multilateral netting system will be permitted to settle on the books of a Federal Reserve Bank unless its participants authorise the system to provide position data to the Reserve Bank on order (Federal Reserve, 1994).

Hence, central banks can unilaterally enforce a netting tax in offshore systems netting payments of the domestic currency by refusing to process or settle net payments from non-co-operating systems. They can also impose less onerous sanctions on members of the domestic payments system that participate in non-cooperating offshore netting systems.

4.3.2 Identifying Netting Activity
It is easy to identify and regulate organized offshore netting systems which process a particular currency. For example, the preceding Federal Reserve requirements apply to all systems with three or more participants that net payments or foreign exchange contracts involving the US dollar, and have on any given day net payments in any currency or currencies combined larger than US$ 500 million, or routinely process individual payments or foreign currency contracts with a daily average dollar value larger than US$ 100,000. Further, netting systems that meet these threshold criteria are subject to the requirements if they or any of their participants are members of the Federal Reserve System, or if participants’ net payments are settled through a Federal Reserve settlement account (one that is on the books of the Federal Reserve), which account belongs either to the netting system or to the participants in the netting system or their agents individually.

It is now also relatively easy to identify less formal netting activity between pairs of banks. This is because much such activity is done by SWIFT, which also provides the netting services of many formal netting systems such as ECHO and the increasingly standard communications and messaging network among banks and between banks and netting and domestic payment systems (BIS, 1993). The ubiquitousness of SWIFT is a result of the cost effectiveness of having such sophisticated services provided by a central third party, a virtual netting and messaging system, and of standardising the services to provide an automated and seamless interface between all participants in the foreign exchange settlement system. The dual function of SWIFT as a netting and communications system enables electronic and automatic recording of the transit, matching and netting history of foreign exchange payments.