A foreign exchange transaction reporting system (FXTRSsm) for central banks

A foreign exchange transaction reporting system (FXTRSsm) for central banks

Futures 31 (1999) 759–777 www.elsevier.com/locate/futures A foreign exchange transaction reporting system (FXTRS sm) for central banks Alan F. Kay, H...

206KB Sizes 1 Downloads 19 Views

Futures 31 (1999) 759–777 www.elsevier.com/locate/futures

A foreign exchange transaction reporting system (FXTRS sm) for central banks Alan F. Kay, Hazel Henderson


P.O. Box 5190, St. Augustine FL 32085, USA

Abstract This paper builds on our 1996 work, “Introducing Competition to Global Currency Markets” [1], with respect to recent developments of the international capital and foreign exchange markets. Recent proposals are reviewed relevant to the desirability of a new architecture or structure to improve the operation of these markets. We propose in greater specificity than in our earlier paper a Foreign Exchange Transaction Reporting System, (FXTRS sm), whose specific technological features will supplement the beneficial affects of non-technological methods proposed by ourselves and others to stabilize and improve these markets, while reducing their social and environmental costs. Patents are pending on the FXTRS sm and royalties earned will be donated to the United Nations by the authors.  1999 Elsevier Science Ltd. All rights reserved.

1. Background and global context Since the publishing of our 1996 paper, global currency markets have reached some $1.5 trillion daily and have become more turbulent, as we predicted. After the 1997 Asian meltdowns of the Thai baht; the Malaysian ringgit: the Indonesian rupiah, the Korean won and the turbulence experienced in other Asian and emerging currency markets, debates over the need for currency boards, various kinds of capital controls, and even taxation of currency trading increased markedly [2]. These debates continued after the Russian default of August 1998 and the nearbankruptcy and bailout of The Long Term Capital Management (LTCM) hedge fund * Corresponding author. Tel.: +1-904-829-3140; fax: +1-904-826-0325. Hazel Henderson is a member of the Editorial Board of Futures. Alan F. Kay is President of the American Talk Issues Foundation. Both served on the Global Commission to Fund the United Nations. E-mail address: www.hazelhenderson.com (H. Henderson) 0016-3287/99/$ - see front matter.  1999 Elsevier Science Ltd. All rights reserved. PII: S 0 0 1 6 - 3 2 8 7 ( 9 9 ) 0 0 0 3 1 - 2


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

a few weeks later. A consensus was announced on October 30th, 1998 among G-7 finance ministers and central bank governors by Britain’s Chancellor of the Exchequer, Gordon Brown, on charting proposals to change the world’s “financial architecture” [3]. U.S. Federal Reserve Board Chairman Alan Greenspan’s speeches to Congress and others explained the Fed’s assistance in the LTCM bailout used similar terminology. Both Greenspan and former U.S. Treasury Secretary Robert Rubin referred to the destabilizing role of technology and interlinked currency markets as requiring this “new international financial architecture” [4,5]. Calls for such concerted policy development also came from France, Germany and Britain, whose Prime Minister, Tony Blair proposed a new Bretton Woods conference, and from Canada, whose Finance Minister, Paul Martin urged a new global supervisor of financial supervisory bodies — an idea now widely endorsed. Since then, the introduction of the euro on January 2, 1999 has absorbed eleven formerly traded European currencies — taking them off traders’ screens. Many in financial circles believe this has added to the pressure on other important currencies remaining, including those of Hong Kong, Canada, and Australia, the Czech koruna and those in Latin America, particularly the Brazilian real, which fell 40% after its flotation on January 15, 1999, but has since recovered substantially. Argentina reacted by raising the possibility that it would abandon its peso (already under a currency board) and adopt the U.S. dollar [6]. Clearly, the existing international financial architecture: the Bretton Woods Institutions, and the IMF’s financial assistance to the Asian economies and Russia — and even its $41 billion standby line of credit to Brazil — proved inadequate to deal with recent volatility and financial contagion. In spite of the recovery in Korea and other Asian stockmarkets, in Brazil and Wall Street’s lofty valuations — the global financial system is still fragile — with banking still a weak link. So far, few new policy initiatives have been offered beyond the standard set of prescriptions within the paradigm known as “the Washington Consensus”, i.e. exhorting countries to put their domestic economic houses in order and get their fundamentals right. In this still “pre-contagion” view, the remedies to prevent currency devaluations and speculative attacks are domestic matters for national policy makers. By now, it must be clear that the new global contagion and volatility can no longer be addressed by nations acting alone. They lie beyond the reach of domestic policy makers’ unilateral efforts and long-term solutions are to be found at the global systemic level — through harmonizing standards for disclosure, accounting, risk-assessment and via international agreements. Thus, IMF traditional prescriptions and conditionalities imposed on Asia’s domestic economies often produced counter-intuitive results — leading to its new contingency credit lines (CCL) for emerging economies. Orderly bankruptcy proceedings for distressed countries became a subject for debate. The most equitable option, based on Chapter 9 of US banking law is described in Henderson (1996) [7,8]. UK Chancellor Gordon Brown led the growing call for debt-cancellation and for the IMF to sell some of its gold for Special Drawing Rights (SDRs). Malaysia’s risky route of closing its capital markets — essentially removing the ringgit from the interest of most currency traders, seems to have paid off in stock rebounds and its success in January 1999 in borrowing $1.35 billion from a consortium of 12 international

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


banks at only three (3) percentage points over LIBOR. Other familiar proposals have been advanced or tried: trading bands, crawling bands, currency pegs, crawling pegs, fixed parity, raising interest rates, traditional currency intervention, early warning disclosures, currency boards, as well as versions of Chile’s successful use of partial controls on short-term inflows. China has more options, with its huge internal market and the limited convertability of its currency. Yet, most of these individual countrybased policies will continue to fail, since they cannot address the now global, systemic nature of technologically interdependent financial markets. Yet major central banks could individually, provide powerful leadership for improving and stabilizing such markets, as we show in this paper. These tightly-linked, real time globalized financial markets and the so-called contagion they create should have come as no surprise. These markets were deliberately deregulated during the 1980s and 1990s. Free markets, trade and privatization became the policy imperatives of many, if not most national governments, as the best path to accelerated GNP-growth. Few heeded the admonition encoded in the 1962 Fleming–Mundell model, still valid today: countries wishing to interlink their economies in world trade cannot simultaneously achieve: (1) stable exchange rates (2) autonomy of domestic economic policy and (3) free global capital flows.1 Ever since 1972, when the Bretton Woods system collapsed, national policy makers have been confronted with this axiom — they can achieve two of these goals — but not all three at once. By mid-1999, some clear-headed reappraisals of the headlong deregulation and liberalization of financial markets pinpointed the results: increasing unstability and risks of continued financial crises. Other factors have also entered the global currency markets since the launch of the euro which despite its early weakness, still is expected to denominate up to 30% of world trade. A chorus of financial “gurus” urged the linkage of the dollar, the yen and the euro into a new de facto global currency regime in bid “to end currency volatility” [9]. Such proposals are unrealistic, in our view, due to the rigidity they would impose over vastly different countries — unless national currencies and regional currency unions were strengthened and local currencies were allowed to coexist to clear purely local markets [10]. Meanwhile, the Australian Prime Minister, John Howard’s Task Force on International Financial Reform, chaired by Federal Treasurer Peter Costello MP with members including his Secretaries of Finance, Treasury, Foreign Affairs, the Governor of Australia’s central bank and the CEOs of Australia’s four largest banks, released its report on December 31, 1998. The Task Force acknowledged that a global response was required to address international markets’ volatility and recent inappropriate risk assessments and investment decisions in the private sector. The Task Force’s overall goal was “soundly-based and more stable capital flows” with “reforms needing to be progressed through international grouping, such as the G1 The Robert Mundell and J. Marcus Fleming model (IMF Staff Papers, 1962) essentially showed that governments and central banks overseeing open economies cannot simultaneously maintain (1) the independence of their domestic monetary policies, (2) stable exchange rates, and (3) uncontrolled global capital flows.


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

22—with active involvement of the private sector”. The report also stressed enhanced reporting and transparency [11]. Since then, the G-22, convened by the USA in April 1998, was expanded into the G-26, to include small countries. Other approaches at the global, international and systemic level have received much attention. They include revived interest in currency exchange taxes first proposed by John Maynard Keynes and later by U.S. economists James Tobin (1978) and Lawrence Summers (1989), now U.S. Secretary of the Treasury. We reviewed the debate concerning such a tax in our earlier paper [1, pp. 315–17] and since then these proposals have proliferated and received more intense scrutiny in many studies in Europe, the Americas and by the World Bank [12,13]. Economist Jeffrey Sachs, director of the Harvard Institute for International Development, proposed that the G-7 (now 8) should be expanded to G-16 to include eight democratically-governed developing countries including Brazil, India, South Korea, South Africa, Chile and Costa Rica. Sachs notes “for a decade, we have had a phony Washington Consensus — and almost no real discussions between rich and poor countries on the challenges facing a world of greater income inequity than ever before in history. The G-16 would establish parameters for renewed and honest dialogue”. Commenting on the Asian crisis, Sachs adds, “The IMF worked mightily and wrong-headedly to make the world safe for short-term money-managers. The IMF encouraged central banks from Jakarta to Moscow to Brasilia to raise interest rates to stratospheric levels…investors do not (thus) gain confidence. The more these countries tried to defend their currencies the more they incited panic” [14]. We pointed out this feedback loop regarding interest rate hikes in our 1996 paper [1, pp. 314]. Many high-level financial officials are now calling for the IMF’s role to be expanded to that of a global lender of last resort and also reversing their earlier views on the need for some small and emerging countries to use capital controls [15,16]. Few refer to the G-15, a powerful group of developing countries which has met continuously since the late 1980s, and produced an influential report [17]. Other global proposals include investor/philanthropist George Soros’ concept of an International Credit Insurance Corporation to undergird global markets. We agree with Soros’ analysis, and support many of his initiatives toward the goal of a “global open society” (where enhanced democratic political processes provide a social balance to market processes). However, such an International Credit Insurance Corporation, without additional restructuring of existing financial architecture — might well exacerbate current moral hazard problems. Indeed, Soros offers many other useful proposals to prevent what he sees as a “disintegration of the global capitalist system and the evident inability of the international monetary authorities to hold it together” [18, p. 171]. Soros also favors a special loan guarantee fund of at least $150 billion to enable developing countries with sound economic policies to regain access to international capital markets. This idea was floated by then U.S. Treasury Secretary Robert Rubin at the IMF annual meeting in October 1998 — but received little support. Soros believes such a loan guarantee fund should be financed with a new issue of SDRs, which he points out over European bankers’ objections, would not create additional money — but if they were ever issued, would merely fill a hole created by a default

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


[18, p. 177]. Soros also believes that IMF conditionalities should have included debtto-equity conversions of non-performing loans and tighter policing of banks’ capital adequacy ratios by the Bank for International Settlements (BIS). These BIS rules exempted banks from such adequate reserves in the case of loans to Korea because it had joined OECD, whose “rich country” membership enjoyed such “low-risk” exemptions. Currently the BIS is debating replacing its 8% reserve requirement with 6 graduated “risk-buckets” (the highest for loans to highly-leveraged hedge funds). Yet today’s risk-assessment models have now proved inadequate. Soros targets several kinds of derivatives as needing regulation because they engender “trend-following” or herd-behavior, including delta-hedging, “knock-out” options and suggests that all derivatives should be licensed and registered with the U.S. Securities and Exchange Commission (SEC) as “new issues” of securities. Clearly the failure of many of these hedging strategies lies in their models’ assumptions of efficient markets and perfect information — points we made in our earlier paper. Other remedies widely discussed include imposing margin requirements and “haircuts” on derivatives and other off-balance sheet transactions, as well as regulating on an equal basis hedge funds, proprietary trading operations of banks and investment banks’ in-house hedge funds. Interestingly, Soros does not address the issue of speculative currency-trading directly — although he does not deny that he engaged in it and that it is widespread and a factor in current volatility and contagion. Alan Greenspan also reminds us that this increase in volatility is good news for traders, who thrive thereby. Yet few, least of all the authors, blame traders. They are playing by the rules of the current game and are not empowered to change it. In addition, in the battle of the central banks vs the speculators they sometimes play a beneficial role. Thus, in systems’ terms, the global economy, by virtue of its real time technological interlinkages has become the newest de facto “global commons”, i.e. a common resource of all its users. Such commons (e.g. the world’s oceans, atmosphere, satellite orbits, the internet, etc.) require win-win agreements, rules and standards applicable to all users. If normal competitive behavior (win-lose) continues, the result is loselose as competition between players leads to sub-optimization and the system itself absorbs risks and eventually can break down. Since the Asian crisis, more attention is now devoted to mitigating this systemic risk. There has been widespread support for Robert Rubin’s view that creditors will no longer be bailed out — indeed, that they must be bailed in. The neoclassical assumptions that markets are self-correcting is now severely strained. Many now admit that free markets, trade and capital flows can indeed swamp many small emerging market economies with devastating effects rapidly felt worldwide. Even Londonbased The Economist, a bastion of free market orthodoxy, has changed its views. The magazine’s editorials often call for interventions in markets, while respectfully covering such formerly taboo subjects as capital controls, tighter bank regulations, taxing currency exchange and even advocated Japanese authorities to “start printing money” [19] (i.e. the “consumer vouchers” subsequently distributed to increase domestic spending). All the new systemic proposals, including those to deal with electronic com-


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

merce — which is now further loosening central banks’ control over money aggregates — should be rigorously studied and debated. The technological infrastructure of today’s global economy will not be dismantled — indeed it is becoming more complex, interlinked and faster-operating every day. Economist Walter Russell Mead of the U.S. based Council on Foreign Relations and others address these new issues by proposing an international central bank (first proposed in 1930 by John Maynard Keynes) which might stabilize foreign exchange markets by maintaining its own currency as an international unit of account. Many others have made similar proposals as well as for reference baskets of currencies (former U.S. Treasury Secretary James Baker); a “board of overseers” of international financial markets (Henry Kauffman) and variations on all these proposals we have discussed, from G. Fred Bergsten, (to revive the Japanese proposal for an Asian Monetary Fund), Willem Buiter and Anne Sibert (a universal debt-rollover option with penalty DROP), Barry Eichengreen [20], Marshall N. Carter, CEO of State Street Corporation, Jane D’Arista and others too numerous to mention. 2. The FXTRSsm The specific function of the FXTRS sm we propose is targeted precisely to the more efficient, transparent function of foreign exchange trading. It is not a magic bullet. Its main advantage is that it can be developed and begin operation in the short term, requires no additional bureaucracy or cumbersome international agreements, and helps stabilize FX markets. It can be licensed unilaterally to central banks. As it is installed as a superior operating system — the FXTRS sm can rapidly become a de facto technological global standard. Already a few central banks are beginning to trade on technologically advanced systems such as that offered by Bridge Information Systems. Private trading on the Electronic Broker System (EBS) is conducted by over 800 private banks, 20 of which own this screen-based intranet system. FXTRS must handle many FX market functions [1], but the most important aim of our research has been to present a technological structure that would reduce the likelihood, scope and force of a massive bear raid attack on a weak currency. Such attacks have sometimes played a role in crippling the economy of the target, disrupting societal arrangements, pushing the middle class below the poverty line, clogging trade pipelines, damaging international relations, causing factional armed conflict, producing masses of refugees, and toppling governments. Globally, such attacks might play a role in a worldwide depression or set back for decades hopes for a satisfactory world order. Nevertheless, at times and to some degree they are inevitable. We are not trying to eliminate them. We are trying in a fair, balanced, and logical manner to greatly reduce their likelihood and their severity. We have focussed on the mechanism of transactions not because FX traders are, per se, the cause of the problem. On the contrary, traders providing liquidity with generally razor thin bid-offer spreads and very low transaction costs, are essential to the satisfactory operation of the huge FX market. This desirable state is only possible because trader activities, including so-called speculation, produce a market

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


of such enormous size that it is economically possible for both high liquidity and thin margins to co-exist. Traders are reasonably compensated for supplying the atrisk capital which makes this possible. Bear raids on weak currencies can be viewed as battles. On one side are the central banks [1, pp. 307–8], which are the only market players at times ready, if necessary, to sell low and buy high to protect their national economies. On the other side are all others, individuals and institutions, not just speculators or hedge funds, but indeed anyone who, with no need for collusion, is ready to jump into the fray at some point in hopes of buying low and selling high. When a currency is weak, there is no doubt that to some extent its price should fall. Whether a bear raid succeeds does not primarily depend on how weak the currency is, but more on how much capital can be brought into the attack and how much capital is fleeing the country. Even strong currencies would succumb to a large enough raid. When it seems that a bear raid will succeed because of the size of the traders atrisk war-chests, then, without capital controls, individuals and institutions with large portfolios in (or denominated in) the attacked currency (i.e. nationals), when facing up to increasing and potentially disastrous losses, will seek to pledge or sell the attacked currency. In order to protect themselves from further losses, they are ready to do this even when the attacked currency has already been greatly reduced in value. This process further accelerates the fall of the currency. The bear raid forces a measure of discipline onto central banks which otherwise might not be too diligent about maintaining their economies or over-value their currencies. However, the amount of corrective discipline that the “invisible hand” of the market, i.e. the attackers and the flight capital of nationals exert, often has little relation to the amount the currency falls. Together they can make the currency, even at irrationally depressed prices, actually worth no more than the lowest — previously unimaginable — price it finally reaches. Bear raids were prevalent prior to the 1929 crash. The collapse of the U.S. market and ensuing depression helped elect Franklin Roosevelt president in 1932. Investment banker, Joseph P. Kennedy, was appointed by Roosevelt to head the newlycreated SEC. New regulations cleaned up the roles of bankers and brokers and made the stock market safer for investors. Based on his intimate knowledge of how the US securities markets worked, Kennedy introduced a number of changes in the transaction process itself. One was the “uptick” rule which prevented a broker from selling short if the last sale price of a listed stock was lower than the previous transaction price. This slowed the momentum of bear raids and they largely disappeared. Note that this rule utilized “ticker tape” action. The “ticker” was based on transaction reporting, which we place at the heart of the FXTRSsm. With technology undreamt of in the 1930s, a much smoother process can be implemented in the FXTRS sm to handle the geographically spread-out and much larger and more active FX markets. Technological provisions in the FXTRS sm will enable the relevant standards body to curb bear raids without impairing the functioning of the market in normal times, nor depriving (or in any way slowing) the execution of any transaction desired by willing buyer and seller at a mutually-agreeable price.


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

3. Assumptions and design for an FXTRSsm We still assume, as in our 1996 paper, that central banks will at some point begin to consider the possibility of initiating new screen-based intranet-type foreign exchange transaction reporting systems (FXTRSsm). Such a system in addition to existing trading systems is an advanced version of our earlier proposal of FXE. Ultimately a sizable group of central banks may agree to link their FXTRS sm systems and set-up a standards body with authority to oversee such a broader, interlinked system. The FXTRS sm would fulfill some of the needs cited by central bankers and finance ministers for “a new global financial architecture”. Thus we assume that these key financial policy makers will encourage broad participation of financial institutions in the FXTRS sm. In the rest of this paper, we lay out criteria and in key ways the structure of such a transaction reporting system to show how benign, useful, and profitable a system could be to all central banks and to the many diverse parties who would utilize it. The main purpose of this paper is thus to encourage central banks to begin to consider seriously and invite bids from information technology providers to build such foreign exchange transaction reporting systems. If properly designed and used, the FXTRS sm could lock-in features shown in Table 1, all now widely accepted as desirable components of a new global financial architecture. Table 1 System capability design goals 쐌 The system will introduce stability into the international currency markets via technologically inherent standards, without interfering with any of the execution methods and modalities used widely today and without unfairly penalizing any of the types of trading or reasons for trading that exist today, including “speculation”. The means described are so flexible that, in principle, various levels of stability can be achieved: i.e. slightly more that exists today, or a great deal more, or even so much more that the very rigidity achieved would be damaging to the functions that a proper market should perform. The standards body designated by each central bank working with the system developers can decide how much of the control capability discussed in this article would actually be built into the system. 쐌 The system will charge (a) fees for transaction reporting that all users of the system with minor exceptions will find unobjectionable and (b) fees described in our paper[1, p. 313] for valuable data services and access to the system database by financial organizations eager for such information. The flow of funds into the system from fees alone will be many times greater than the operating costs of the system. Arrangements can be made so that anticipated large surpluses can be transferred to currency reserve funds or other worthy purposes. Transaction taxes, or other taxes, will be unnecessary to achieve this outcome, but could be included in addition, if desired. 쐌 The operating capability, standards, principles, and practices of the system can be set up to be acceptable both politically and financially 쐌 to central banks, 쐌 to the diverse group of financial organizations which are the major users of currency exchange markets, 쐌 to vendors (Bloomberg, Reuters, Bridge, etc.), 쐌 to foreign exchange brokers and dealers, and 쐌 to the national political leaders and the general public of the participating countries. The support of these groups is essential to the success of FXTRS sm as described in our earlier paper [1, pp. 317–18]

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


To achieve the goals in Table 1, the central banks can assure that certain executions will be promptly reported to the system as follows: all, or most, of the transactions involving (on both sides of the transaction) (a) currencies of participating countries or (b) financial instruments, such as derivatives, the values of which are denominated in currencies of participating countries. Certain kinds of transactions may be exempted. The transaction reporting process can be handled as described in our paper [1, pp. 311–12]. The system standards body might also wish to consider developing system specific terminal(s). Such terminals would be attractive because of the primacy given by them to facilitating system functions. New participants will ultimately prefer a new terminal. Offering a system-specific terminal, whether through installation and maintenance by a vendor or otherwise, would yield an additional source of system revenue, and could be helpful in establishing price/service bargaining leverage by making the system less dependent on collective vendor performance. Whether through existing or new terminals, this business will be welcomed by all vendors as a source of ongoing revenues. As described in our paper [1] transaction reporting to the system — or retrieval by participants of transaction data and of system reports and statistics — can be accomplished with minimal key strokes (or mouse clicks), using easily learned codes that institutional participants, through their vendors, can tie into their own transaction processing software. The latter would represent a substantial amount of new business welcomed by the vendors.

4. Marketing comparison between government bonds and government issued currencies We have compared [1, pp. 313–14] the highly structured and stabilizing operation of the government bond market by the Federal Reserve System in the U.S., (the Fed) with the general absence of control by the Fed or any other government agency over the large private market for the purchase of U.S. currency, when payment is made in other currencies. The features proposed for the FXTRS sm parallel or are alternatives to Fed activities and procedures that keep the U.S. government bond market stable, so stable that this market has become the “safe haven” of choice for many individuals and institutions around the world. This “safe haven effect” has been a major factor in allowing the U.S. to refinance its $5 trillion debt at lower interest rates. It is proposed that central banks participating in the FXTRS sm consider encouraging “marketmakers” who, at the discretion of each central bank, may operate privately under a franchise. Such a franchise could be similar to the arrangement the Fed has with primary dealers in the U.S. government bond market. Alternatively, in many countries, marketmakers may be employees of an agency of the central bank working for salary and bonus. In either case, the marketmaker(s) for a given currency pair should be encouraged, even required, to quote generally the best bid and offer that is available world-wide. This means that marketmakers are financially incented and held accountable by some means like frequent rating on the three performance


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

measures of good marketmaking that are underlined here: most frequently having the narrowest spread and supplying the largest size. Whether an execution comes from marketmakers or any other system participant, an important aspect of how the system should function is illustrated in the currency pair, UK£/US$, market. The transaction fee is charged in dollars to the party buying dollars (in analogy with parties buying U.S. government bonds), while the other side of the trade, the party buying pounds is charged the transaction fee in pounds. As explained later the two fees may be considerably different.

5. Trade reporting for greater stability We explain in the balance of this paper the principle design and operating features of the system which can produce the desired range of control over destabilization, mentioned previously. We start by noting that trade reporting itself in existing markets generally helps stabilize the market. Quotations given by FXE dealers (and brokers) presumably tell customers what the dealer is willing to do. Their usefulness depends on the customer’s trust in the dealer. Since there are thousands of dealers, the degree of trust necessarily either comes somewhat second-handed (and trust does not travel well), or the customer limits the number of dealers it will use. If the latter, the customer may also be giving up executions at better prices. A quotation is truly only made firm and real when there is a resulting transaction. Observing the information flow of executed transactions to all participants in the proposed system, will give every customer a wealth of benchmarks on whether a particular quotation was competitive or not and every dealer an understanding of its own competitiveness. When a market lacks information, participants can sometimes be spooked or can too easily vacillate between over-caution and recklessness, characteristics exhibited by the global currency markets and their recent volatility, over- and under-shooting. Nevertheless, even stock exchanges with last-sale tickers are not immune to destabilizing forces. Trade reporting will help FXE, but further stabilizing mechanisms are still needed.

6. Fee structure concept A basic consideration is that, except for emergency fees, all fee rates will be so small as a percent of the transaction value that no transaction contemplated by a willing buyer and a willing seller at an agreed price will be derailed because of fee resistance. Transaction fees constitute the main, perhaps the only, source of revenue from the system’s foreign exchange transaction activity. There are potentially four considerations in determining the net total transaction fee: (1) the base fee, adjusted for size, (2) adjustment for trade purpose, (3) fortuitous timing component, and (4) emergency. The base fee, adjusted for size is the same for both buyer and seller. The adjustment for trade purpose is usually different for the two sides of a trade

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


because their reason for trading is often different. If there is a fee component for fortuitous market timing in a given trade, it is charged to one side and not the other. The same is true for the emergency fee. In view of the benefits perceived by users, particularly since the system will strengthen stability and inhibit or ameliorate bear market raids upon currencies of vulnerable countries, fees will be readily accepted. This attitude will be enhanced if fees are established by a fair, fully justified, open, and reasonable process. For all these reasons, except for emergencies, the existence of fees will be off the trader’s radar screen and out-of-mind. The system will have a wide range of algorithms to handle a variety of fee structure components built into software. Some may never be expected to be used and owners may choose to eliminate them. When needed fee options are activated or authorized by the relevant standards committee, these algorithms become available and are adjusted by a system operator selecting software parameter values. Once the process is decided upon, parameters may be virtually instantly changed. The system owners (central banks) are provided, as an extra comfort factor, a much greater range of flexibility of fee options than they will need. Each parameter may be adjusted all the way from zero (no transaction inhibition) to some large value, enough to significantly inhibit trading. The proper level can be readily found for adequate control of each fee component and the relative levels can be adjusted to fit the owners’ own informed, operating concepts. Starting from very low rate levels where there is no inhibition on trading due to fees, as the rates are increased, system revenue rises proportionately for a time. If the increase continues, rates hypothetically would reach a level where trading is greatly inhibited. At some point along the way, fee revenue is maximized. At this point the system is losing as much volume from dissatisfied departing customers as it is gaining from new customers unhappy with competing marketplaces. This is an unstable operating point and not recommended. As long as there are competitive marketplaces, better practice will be to operate at still lower fee rate levels. For public policy reasons, lowering rates even further will make it universally clear that the owners are operating the best among all competing marketplaces. If this thinking informs their decision-making, owners will capture all of the FX transaction business they seek. 6.1. Base fee and size dependence A base transaction fee of a small percentage, like 0.001% of the value of a baseline trade of $1 million (or equivalent in other currencies) would yield $10 per trade. Without any other charges if all trades were of this size, the total system revenue, when all major currency countries were participating, would then be $10 million per day or about $3 billion per year. Foreign exchange transactions vary in size from a few dollars at an exchange kiosk or retail bank counter to several billions of dollars in the international capital markets. The base fee would have to have some dependence on size. For example, with very small trades (under $10 000) the fee should probably remain fixed de minimis at $10. The FXTRS SM could be operated in one of two modes, either (1) seeking to accommodate and handle properly trades over the full range of size (about five or six orders of magnitude) or (2) restricting activity


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

to large trades, with a minimum of say, $1 million, as recommended in our 1996 paper [1, pp. 318–19]. With the former, several levels of marketmaker, capital, size and activity requirements, will be needed. 6.2. Adjustments for a taxonomy of trade purposes The purpose of a particular trade or some category of trades in FXE trading may be judged by worthiness from the point of view of transparency, legality, fairness and public policy. Fees, in principle, may be made to depend on the purpose of the trade by category. In Table 2, a list of a dozen categories are adequate to make useful distinctions for public policy and trade facilitating purposes. Categories may be modified, dropped, or replaced, but unlike changing other parameters, after the system is in operation, it may require a major effort to redefine purpose codes, since all users must be retrained to use new codes in reporting trades. The adjustment of fee for “purpose” might be accomplished by a schedule of multiplicative factors. A simple schedule illustrates the concept. Total fees could be cut for favored purposes such as transparency, doubled for dubious purposes, and unchanged for those in-between. Of course, much more complex schedules can be readily handled. Table 2 Categories of FXE transactions by purpose Transactions are executed in order to facilitate, make payments toward, or arrange for the following: 1. international delivery of all legal goods, services, and software (including cyber services). 2. direct investments in or acquisition of foreign, non-financial assets, operating companies and properties. 3. international travel. 4. (with respect to the portfolio of a resident/citizen of a country initially largely holding assets denominated in the currency of the country—including the currency itself) seeking to obtain more diversified holdings and to achieve a reasonable portfolio balance. 5. gift or contribution to individual/organization/charity. 6. personal or business loan, carrying no investment-features. 7. portfolio investments in foreign financial assets or financial companies. 8. (wholesale and retail) broker/dealer distribution function, efficiently placing portions of a relatively large amount of a currency originating from a transaction into the hands of many different parties who desire relatively small amounts. 9. hedging: balancing a portfolio as in (4) except that in this case the portfolio is not one being managed for, or on behalf of, a resident/citizen. 10. arbitrage (true): one or more trades that when executed simultaneously can produce a profit for the arbitrageur, the profit arising from the variation of bid/offer prices available at the same time on different exchanges/regional markets. 11. arbitrage (speculative): same as (6) except that one or more of the bid/offer prices is known only speculatively while the remaining quote(s) is/are known to be real. 12. (pure speculation) an execution by a party to a trade based on the party’s belief that coupled with other transaction(s), the transaction will produce a profit or reduce a loss, with the key proviso that this category covers only cases not covered under previous categories. Trades outside these 12 purposes or undisclosed can be assumed to be for tax-avoidance, money laundering, or other illegal purposes.

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


6.3. Market timing characteristics: price versus time As would be reported to the system, consider the time sequence of executed transactions in a single currency pair, A and B (for example UK£ and US$) where some rapid rate like ten to a hundred transactions a minute may well occur among busier currencies at busier times. Fig. 1 is an illustrative scatter diagram of the price ratio of two currencies over time which extends from time zero to a later time T. The price is normalized with respect to a reference such as the first trade of the segment from 0 to T. The vertical axis or ordinate is the percent change of B/A, normalized by dividing this ratio by its value at the time of the first trade at t=0 written as (B/A)/(B/A)0, which at t=0 is 1. The abscissa of Fig. 1 is zero percent change, and this first transaction point has coordinates (0,0). Fig. 2 considers the same case as Fig. 1, expanded after time T up to some other time, say, the end of the day. We also imagine that just after time T there is a clear change in the market and the value of A drops with respect to B. We show in Fig. 2 the trend line for trades before time T (the same as Fig. 1), and a new one for trades after time T. The scatter is no greater about the second line than the first. The trend line is the best fitting straight line where price points are weighted by some scale adjusting for trade size, according to mathematical formulas available from the authors on request.

Fig. 1. Trades from beginning of day to time T. Percent change in normalized ratio of amounts of currency B to currency A executed from time t=0 to t=T.


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

Fig. 2. Trades of Fig. 1 continued to end of day. Percent change in normalized ratio of amounts of currency B to currency A executed from time t=0 to end of day.

In Fig. 2 there are a few trades more than 0.5% above and a few more than 0.5% below the full day trend line. Those who bought pounds in this group close to the top and those who bought dollars in the group at the bottom, for their smart or lucky market timing, can be viewed as having received a “bonus,” specifically each did better than the counterparty by at least 1.0% for all such trades. There can be no substantive objection to the fee going up to 0.002% from 0.001%, representing an increase of 0.001%, that is one thousandth of the size of the “bonus” that arises from fortuitous market timing, whether this has occurred by luck, experience, market power, and/or negotiating skill. Fig. 3 shows a similar and typical situation for a day when there are three abrupt changes in the A/B price trend. If Fig. 2 were extended further, say for a week, an entirely different picture could emerge. Suppose, after the first day, the trend line climbs (or falls) radically, for example, rises 10% in a week. The buyers of A at the top in the first day would then, if they had not sold or hedged A, would be in a significant loss position by the end of the week. The buyers of B at the bottom in the first day, if they had not hedged or sold, would have an even greater bonanza from the week’s outcome and could have no complaint. What is a fair fee for fortuitous timing in this case? Those A buyers who bought at around the first day’s high still had a bonus compared to their counterparty, that was a thousand times greater than their timing fee. They could have no complaint

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


Fig. 3. From beginning to end of day. Percent change in normalized ratio of amounts of currency B to currency A executed from time T=0 to end of day.

that the market then turned against them. The relevant standards body would have to decide on precisely what to do for such cases, but the possibilities and justification for higher fees for timing are clear. Such fees will not affect the decision of buyer and seller at all. There are many alternatives to consider. The time segment for determining the trend line, and hence the increases over basic fees for timing could be clock-related (like every hour, day, or week), trading volume-related (when a certain amount of trading and number of trades has executed since the beginning of the segment), or mixed (the segment end is delayed until the sooner of a specified trading volume plus number of trades is reached or a total specified elapsed time). Could interday change points in trend lines, such as the three marked “I” on Fig. 3, be so well defined that no one would object to a computer algorithm deciding them? The authors have developed a proprietary resolution of this question, information on which is available on request. Finally let us consider the most important case, the emergency. A weak currency


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

is under attack and falling rapidly, say 50% in two days with respect to stable currencies, and perhaps accompanied by some rapid gyrations in price during the two days. This situation illustrates the best case for large fees. Large fees could slow the attack and at the same time contribute substantially to funds for humanitarian aid and/or financial rescue of the country under attack. Fees might be allowed to rise as high as 1%, a thousand times larger than the basic fee, not enough perhaps to seriously affect the profits of the successful bear raider, but perhaps enough to wipe out the profits of a less successful one. The attack could still occur but it would undoubtedly be (a) weaker and (b) spread out over more time. Both effects, (a) and (b) are of great benefit to the stability of the currency markets and the economies of countries. These large fees would act somewhat like the more familiar exchange “circuit breakers”. But circuit breakers imply a suspension of trading, not only hard to police in FXE, but also, more important, circuit breakers would totally violate the goals of the system of Table 1, to never stop or delay a willing seller and a willing buyer from executing a trade at a mutually agreeable price. Large fees would slow down, but not totally stop such trades. The emergency fees would kick in at some moment when the slope of the best fitting line in a time segment up to that moment reached some high level and would rise in reasonable increments as the slope went step-wise higher. At each step of slope increment, the emergency fee rates would also be higher if the gross amount of trading were higher in the same time segment. The fee size, as a percent of the trade, would be selected as the fee calculated from a fixed matrix of fee versus the two variables, slope and amount, maximized for all values of slope and amount on the matrix schedule, and redetermined for all initial times of the time segment. The total range of emergency fees might be something like 0.01% to 1% of the transaction value (just above the timing fee range) and would be charged only to the party buying the B instrument. This is logically correct as the amount of B required to trade for a fixed amount of A would be increasing rapidly with time and a substantial amount of trading had taken place. Another point of view, a mistaken one, is that the fee component for market timing is only justified in the emergency case. If that were so, the fee would be thought of as a “crisis” fee or tax. If at normal times there was only a basic fee for size, then the decision to impose “crisis” fees would have to be announced to participants and become widely known. Whether or not there was a decision to impose crisis fees (when and on what instruments and in what amounts) would be known or speculated upon by all players in the FXE world. This interaction between the system, the participants, and the relevant standards body decisions would itself play a role in the crisis with a net effect of perhaps curbing it or perhaps aggravating it. In any case the system owners would not have as much control as can be obtained by having fees depending on market timing as an entirely normal fee and, once accepted, uncontroversial in daily operation. The much larger fees received under crisis conditions would help system revenue and might either multiply it several fold or effectively abort bear market raids. If fee schedules were wisely chosen, some of each would take place and that might be optimal. We are not trying to stop heavy trading.

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777


Why have we not considered negative fees? Why have we only considered fee decreases for small size, but not for timing or emergencies. The reason is that the trade reporting system should be viewed as a service of benefit to everybody in the market. The basic trade fee is the reasonable cost of complying with regulations or of accepting other inducements in order to make the system work properly. There should be no reward to compensate individual players for poor market timing. That would not be fair to their counterparties who, de facto, have good market timing. Another point that we recommend is that the rules by which the fees are chosen are made publicly and widely known, with one exception. Some key choices in the parameters of the timing fee should be revealed publicly only as they apply to trades that have already taken place. To do otherwise would be to invite game playing irrelevant to the operation of the market. Some could argue that the basic fee should be doubled to say, $20 a trade, or tripled, or set even higher. They could reason that the system would be so valuable that the owners should reap even greater rewards. It is true that the system might still work with substantially higher fixed fees, in which case others might argue that no variable fees are required. Proponents of both viewpoints might argue together that the system would work with much higher basic fees, most players would still gladly use it, and it would be even more successful financially. All of these ideas would in effect be a system design substitution on the issue of stability. Instead of a surgical scalpel we would have a meat axe. Such a high fee system would produce little in the way of stability. It would be unfair to many classes of participants or potential participants and would not pass the test of good public policy. The meltdowns of markets witnessed in the last year would continue without any real understanding of where they would occur next. We hope the relevant standards bodies makes a sincere effort to appreciate these points and then makes the right and wise decision. The emergency and timing fees encourage dealers to narrow their spreads and, along with trade reporting and the latest technology will help make the owner’s marketplace more efficient for customers and superior to competitive marketplaces. There is one more important step to take to assure that marketmakers or dealers operate so that this FXTRS SM is without doubt the best system there can be. Dealers who have extensive relations with major FXE non-bank customers and are extremely well financed, can at times afford to take considerable risks to be competitive with other dealers not so well situated. It is important to provide an incentive to induce marketmakers to hold spreads narrower than market risk justifies, so narrow that they would on average lose money. These losses (in effect fees) will then have to be made up by a bonus scheme with penalties for spreads being wider than those of competitors. Marketmakers can only be expected to be competitive with others in their class, in terms of available risk capital and experience in the business. The FXTRS SM requires agreements to obtain competitors in each class as authorized marketmakers in each traded item. These market-makers agree not to make any trades outside of the system, to report their trades promptly and accurately and to keep audited, high


A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

quality records of their activities, meeting system standards, and available to standards committee inspection. When an authorized marketmaker’s bid is demonstrably too high or offer too low at the time of an execution by an authorized market maker, the computer assigns penalty points to the marketmaker whose bids and offers were not as good as the transacting marketmaker, and leads to a reduction in the marketmaker’s bonus, under certain conditions. In the absence of penalties, the bonus is high enough to assure the marketmaker will receive a fair profit. The opportunity for central banks to finance the risk capital of authorized marketmakers as part of the agreement should not be overlooked and surely could accelerate the creation of the best marketplace in the world for the currency of each central bank and full control on the part of the central bank to achieve best-marketplace status when it seeks to protect its currency.

7. Conclusions The advantages of the FXTRS sm include: 1) enhancing stability in currency markets via a technologically state-of-the-art trading system, which will be widely adopted voluntarily; 2) technology that will be a de facto, global standard for currency trading, reporting and supervision without legislation or cumbersome international bureaucracy; 3) provision of central banks with superior information and reporting, allowing levels of prudential supervision of currency markets which will be beneficial to all market players, 4) a “ticker-tape” for currency trades, and 5) new revenues to the user-groups and vendors.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

Henderson H, Kay AF, Introducing competition to global currency markets. Futures 28(4):305–24. Australian Financial Review. Henderson’s The Breaking Point, p. 1–9, December 4, 1998. BBC Online Network, London, U.K. October 30, 1998. U.S. Treasury Press Release. Declaration of G-7 Finance Ministers and Central Bank Governors, October 30, 1998. Greenspan A. The structure of the international financial system. Annual meeting of the Securities Industry, Boca Raton, FL, November 5, 1998. The Economist, January 23, 1999, p. 69, Argentina. Building a Win Win World, Berrett Koehler, San Francisco, (1996, 1997). The Economist. When countries go bust. October 3, 1998, p. 88. Business Week, January 25, 1999, p. 126. Henderson H. Building A Win Win World, Ch. 9. Information: the World’s New Currency Isn’t Scarce. Berrett-Koehler, San Francisco (1996, 1997). Australian Treasurer, Official Press Release, December 31, 1998, www.treasury.gov.au Ul Haq M, Kaul I, Grunberg I, editors. The Tobin Tax: coping with financial volatility. London: Oxford University Press, 1996. Soete L, Weel B. Cybertax. Futures 1999;30(9):853–71. The Economist. Global capitalism making it work. Invited essay by Jeffrey Sachs, September 12, 1998, 23–5.

A.F. Kay, H. Henderson / Futures 31 (1999) 759–777

[15] [16] [17] [18] [19] [20]


Business Week, February 8, 1999, pp. 64–77. The Economist. Global Finance section, January 30, 1999. Challenge to the South, South Commission, Oxford University Press (1990). Soros G. The crisis of global capitalism. New York: Public Affairs, 1998. The Economist, October 10, 1998, p. 18. Eichengreen B. Toward a New International Financial Architecture, Institute for International Economics, Washington, D.C., 1999. A useful overview of most current proposals.