It is important that discussion of a CTT should concentrate on what has been seriously advocated.

Recent discussion of the tax has supposed that it is collected at a very low rate: in some cases as high as 0.25% but mostly at 0.1% or less. There is wide agreement that the rate should be low enough to have no significant effect on international trade in goods and services or long-term investment. The lower the rate, the less will the "tax-base" (the volume of exchange-transactions) be reduced; so that revenue will not decrease proportionately to the decrease in the rate. But by the same token, of course, a very low rate is unlikely in itself to have much effect on capital movements. Investigators of the tax have recognized the possibility that raising it above some critical level may reduce the tax-base so much that it actually reduces the tax revenue that can be raised. So there may be a conflict between revenue-raising and reduction of capital movements.

There is no question of a world authority to impose the CTT. It will be imposed independently by national governments through their own legal processes or not at all.

It will also be collected by national governments. No one is suggesting an international tax department.

So what is proposed is really a set of national taxes, but imposed, it is hoped, in a coordinated way.

Most discussion supposes that it would be highly desirable that the tax should be collected across the world at a uniform rate. This would avoid various actual or possible difficulties. But it is not clearly and absolutely necessary that it should be collected by every national authority or at a uniform rate.

The attraction of the tax is of course the possibility that it might provide resources for global purposes that are otherwise likely to be seriously underfunded, such as anti-poverty programmes and peacekeeping. But there is no necessity that it should be devoted entirely to global purposes. One specific proposal widely supported is that the governments imposing and collecting it, even those of the industrialized countries, should keep for themselves a share of what they collect, so that there would be a direct incentive to participate. The less the cost of collection turns out to be, the lower the share that might reasonably be agreed for the countries collecting it. However the tax is administered, it is very likely that a uniform rate would allow a hugely disproportionate share to be collected by a handful of the major-currency countries. The shares retained by the collecting countries would have to be set with that in mind. There is no presumption that the part devoted to global purposes should necessarily be administered by any one international agency, such as the UN Secretariat or the IMF. But it is clear that institutions for administering that part would need to be set up or delegated by agreement.

1. As a source of revenue, for global purposes or otherwise, a CTT seems capable of raising a large amount of revenue with little distorting effect on "real" transactions. With annual foreign-exchange transactions of the order of $450 trillion [BIS, 1999, p.117, records daily figures for 1998 which, multiplied by 240 for the number of working days in a year, gives a figure of $476 trillion; there could easily be a fall of at least $30 trillion arising from the formation of the Euro in 1999, but other elements in the total would probably have continued rising], it is clear that, even if the imposition of a 0.1% levy reduced the value of the transactions by 67%, the amount raised would still be $150 billion, about three times the amount of official development assistance going to developing countries from the OECD countries and multilaterals in 1998. A 0.05% levy, with a 67% resulting reduction of the value of the transactions on which the tax was based (an assumption that could be regarded as implying an elasticity of unity in the response of volume to transaction-costs if pre-tax transaction-costs were 0.1% of transaction-value [Felix & Sau, 1996, p.242]), would yield $75 billion. If such estimates are realistic and it is accepted that the effect of the tax in reducing foreign-exchange transactions other than those directly connected with trade and investment on the part of "non-financial" customers was on the whole good or at worst harmless, the tax might be claimed to cause less distortion to economically useful activities than other feasible ways of raising the same amount of revenue globally.

2. There is also the political advantage that the burden of the tax on most of the world’s public, though real, would probably be highly diffused and not readily evident to persons not directly involved in the financial markets.

3. The CTT has been claimed to be a way of checking excessive short-term capital movements internationally. This is because it is likely to check exchanges of currency with short-term intentions much more than those with long-term intentions. It will thus reduce short-term currency speculation, which, evidence suggests, in contrast with speculation that has longer horizons, is on the whole destabilizing [Frankel, 1996, pp.54-5, 57-9].

4. For this reason, the argument goes, the CTT will allow particular governments more autonomy in policy. Even with otherwise-free capital movement, they will have greater scope as a result of the CTT for running interest-rate policy according to the needs of the domestic economy rather than having their own interest-rate rigidly limited by interest-rates and other conditions elsewhere. They will also find it easier to maintain stable (fixed or pegged or targeted) exchange-rates without fear of speculative attacks. And they can run their tax and spending policies with less fear of what "the markets" may think.

5. Finally, it has been suggested that an advantage of a CTT administered by a number of national authorities is that it would imply continual monitoring of foreign-exchange movements; and this could provide early warning of crises.

Arguments against a CTT and qualifications to the case for it

1. The burden of the tax on "real" international transactions may well be several times as high as the very low tax rate suggests, since some of the foreign-exchange transactions not directly connected to "real" trade and long-term investment may be ancillary to it, as in providing liquidity for the currency traders or hedging for the traders in goods and services. So the incidence of the tax on them may further increase the "spreads" between buying and selling prices of currencies for "real" transactions. Kenen [1996, p.121] gives reason for supposing that on 1992 figures the total additional transaction-costs might be 3 to 4 times as high as the tax. [But suppose that a factor of 4 would be enough to allow for this. This would suggest that a 0.1% tax could not imply an effective rate on trade in goods of more than 0.4%, and that an effective burden no higher than 0.1% on trade in goods could be achieved by a tax of 0.025%, which could still raise considerable revenue (and not proportionately less than a 0.1% tax). It is also worth noting that transactions costs on foreign-exchange were considerably higher 20 years ago than they are now without being recognized as a serious impediment to trade.]

2. There would be potentially serious problems over enforcement. Several years ago it was supposed [Kenen, 1996] that effective enforcement would require cooperation from at least the US, Japan, the European Union, Hong Kong, Singapore, Switzerland, Canada, and Australia, in order to cover the countries whose resident banks undertook significant foreign-exchange trading. Taxing in some of these countries and not others would shift the trading to those that were free of tax. Even with that degree of cooperation there were doubts. There are a number of countries that might make suitable sites for currency trading, and the markets might simply migrate if the existing locations became too expensive. [Kenen proposed that, in order to nip this in the bud, penal rates of tax should be imposed on transactions with dealers who dealt outside the taxed sites.] It was also suspected that currency trading might find ways of avoiding the banks that constituted the recognized markets; and that the internet might make this easier than before.

3. There are also serious political obstacles to match the apparent political convenience of the tax. Operators in the financial markets will generally have strong interests and ideological commitments against the tax, and they have ways of making their views prevail with central banks and governments. The tax can also be misrepresented as a threat to national autonomy by those suspicious of international cooperation.

4. Doubt may also be cast on the CTT’s alleged advantages as a means of mitigating the instability in international capital markets and its effects and of increasing governments’ autonomy. (a) There is no more than a very broad presumption that the tax will always make the right distinctions, discouraging destabilizing speculation based on following leaders and rumours, while not discouraging speculation based on "fundamentals" (such as "real" events and monetary expansion or contraction), which will tend to smooth exchange-rate movements. (b) A tax will raise the cost not only of speculation but also of hedging, which is used to mitigate the risks arising from instability of exchange-rates. [On the other hand, if it succeeds in reducing the percieved risk of crises, the cost of hedging may fall.] (c) In any case, a tax at the low levels supposed will not have much relevance in currency crises such as those of the 1990s, where the percentage gains and losses at stake to operators far exceed such tax rates. (d) Similarly taxes at the levels contemplated will have very little effect on particular countries’ independence in policy. The fact that national authorities have freedom to move exchange-rates or interest-rates either way by say 0.1 of a percentage point more than they would otherwise have been able will not mean much.

5. The outcome of the imposition and collection of the tax will be grossly unfair unless a large proportion of what is collected is redistributed. The burden of a CTT imposed uniformly across the world at the rates contemplated might well be largely hidden, but it will be real and its distribution across countries is likely to be broadly proportional to their involvement in "real" international transactions (which means incidentally that, in relation to income, it will be higher among the more open economies). By contrast, however the tax is administered, the great bulk of the revenue will be collected by a few rich countries. Unless much of this revenue is redirected to poorer countries according to need or reserved for genuinely global purposes, a CTT may be judged guilty of taxing the poor for the benefit of the rich.

6. As a minor point perhaps, it has been argued that the tax may encourage consolidation among those currency-dealers that constitute the markets, namely the banks that make a business out of providing currencies on request for others. Much of the world’s currency-trading, it is claimed, is done by these institutions so that they can maintain the necessary liquidity in the various currencies, that is to say, so that they can always have enough of these currencies available when they are needed. The more costly such transactions are for these banks, the greater the advantage to them of large size. The larger the amount of currency trading that an institution does for its customers, the less the stock of currencies that it needs to hold in proportion to the amount it trades for its customers, and hence the less trading (proportionately) it needs to do simply to maintain adequate liquidity in all currencies. So, as the cost of this trading to preserve liquidity goes up with a tax upon it, the advantage of large scale for the dealers increases. This might reduce the number of institutions providing exchange services. Alternatively, Frankel [1996] has raised the possibility that a sufficiently high tax might alter the whole character of the world foreign-exchange market, making it more like the stock exchanges, with the importance of dealers greatly reduced and final customers trading directly with each other through brokers in organized and centralized conditions. Such a change might not necessarily be bad in terms of efficiency, but it would greatly reduce the volume of transactions and so reduce the revenue-potential of the CTT. [These responses are of course entirely hypothetical; and whether they happen or not may not depend on the existence of a 0.1% tax.]

Before the East Asian events of 1997 and Spahn’s paper, it might understandably have been argued by opponents of the CTT that

• short-term speculative movements were not unanimously identified as a serious world problem (the Mexican crisis of 1994-5 might have been regarded as a one-off due to special circumstances, and the European Exchange Rate Mechanism crises of 1992 as desirable adjustments that actually benefited the countries which then devalued or left the Mechanism);
• there were uncertainties over the appropriate rate of a CTT, especially given the various purposes it was supposed to fulfil---low enough not to disturb trade and long-term investment, high enough to have some discouraging effect on short-term capital flows, and at the same time not so high that it discourages currency-market activity unduly for revenue-raising purposes;
• in any case a CTT in the form and at the rates discussed might do little to avert currency crises;
• whether the tax could effectively be collected without (or even with) effective cooperation by virtually every government in the world was at least doubtful;
• traders in the financial markets, who were feared and revered by governments as never before, were hostile; and
• misrepresentation of the idea had led to a populist backlash against it among certain politicians.

But then came the East Asian, Russian, and Brazilian speculative currency crises,

• making abundantly clear that the Mexican events had not been unique;
• acutely reviving doubt on whether a completely free international capital market was really always for the best; and perhaps
• making politicians less happy about their bondage to fears of what the currency-traders might think, and less willing to accept the financial markets as the repository of benevolent wisdom.

At the same time, by late 1999:

• Spahn’s proposal appeared as a plausible way of dealing with the newly-revealed danger and suggested that the CTT can be made highly relevant to currency instability.
• Schmidt’s argument, so far unchallenged, was that there is now an effective method by which a CTT can reliably and cheaply be collected.

None of this overcomes the political obstacles arising from misrepresentation of the CTT or prejudice against any move involving international cooperation. But the new ideas and events may well be seen as demolishing or seriously weakening each of the informed and rational arguments against it.

A CTT could now be reliably collected through the system of foreign-exchange settlement

At the Antwerp Consultation, Rodney Schmidt argued, in a so-far-unpublished paper [Schmidt, 1999; later updated by Schmidt, 2001], that recent changes make it not only possible but relatively easy in a technical sense for a CTT to be collected reliably through the system of interbank settlement of foreign-exchange trades---in other words through the arrangements under which banks credit and debit in their customers’ accounts transactions involving exchange of currencies, and the domestic payment systems and cross-border netting systems that settle the claims of these banks against each other.

Because Schmidt’s thesis is very recent and had not (before mid-2001) been published in an academic journal, it had not at the time of the consultation been exposed to all the possibilities of criticism that we should expect a novel argument to face before it could be regarded as well established. But it rests essentially on a description of the institutions of foreign-exchange settlement as they now stand. These are presented with considerable detail that carries conviction. If his institutional description is correct, the key conclusions that he draws about the possibilities of enforcing a CTT seem to follow. Before and after the Antwerp presentation, his paper was shown to several highly reputable economists with interests in the broad area of study. None that has commented has professed to know enough about the institutions to be able to endorse or to fault his account of them. But no reason has been given to doubt that, if his account of the institutions is broadly correct, his conclusions about the enforceability of the tax follow.

A reasonable judgment seems to be that he is very probably right in his main contention: a CTT could be reliably and consistently applied universally if the governments of the world, or even the four to six of them responsible for the main vehicle-currencies, had the will to do so. He also argues convincingly that many countries---the important exceptions being those four to six that are issuers of the main vehicle-currencies----could apply a CTT unilaterally to currency transactions settled within their banking systems without significantly diverting the activity taxed so as to put it outside their jurisdictions.

For many of those who will read this report, that is the key point: an impressive, and so far (early 2001) unchallenged if not generally accepted, case has been made for saying that (i) a CTT could be easily and reliably applied to the whole foreign-exchange market given general international consent or even the cooperation of only the four to six main currency authorities; and that (ii) many countries could apply a CTT unilaterally to transaction in their own currencies, without significantly diverting the activities taxed to other jurisdictions.

A brief summary of Schmidt’s case is as follows.

Though foreign-exchange markets are informal, dispersed, mobile, and uncontrolled, the systems of foreign-exchange settlement are by contrast formal, centralized, and regulated, and have become increasingly so.

This has come about recently because information technology has made it possible for settlement risk (the risk to one party in a currency trade that the other will not pay up) to be eliminated by simultaneous settlement of the two sides of any transaction, because many countries now have the institutions (domestic large-value payment systems) that allow this to be done, and because central banks have insisted that this facility be used by offshore netting systems.

This in turn has made it necessary for the information systems used by banks and central banks for purposes of foreign-exchange settlement to be linked globally, and for details to be kept within them of all gross transactions for exchange of currencies. So not only can modern domestic payment systems identify all gross foreign-exchange transactions settled within them (as those that have no domestic counterpart payments denominated in the same currency), but it is also true that cross-border netting systems have access to information enabling them to identify each gross foreign-exchange transaction involved in the amounts that they net between banks. Central banks may, and on the whole do, require netting systems to use these gross-transaction-identification-and-monitoring facilities as a condition of allowing them the access they need to domestic payment systems. Moreover it is planned that, at some time in the near future, the Continuous Linked Settlement (CLS) Bank will come into operation. This will enable the two sides of every wholesale interbank transaction between currencies to be settled simultaneously and, for this purpose it will have access to information identifying all gross foreign-exchange transactions.

Information technology used jointly has in effect created a single system of information management on the basis of which simultaneous settlement of currency transactions will soon be universally possible and universally required.

The use of this technology is already paid for by fees related to the gross transactions covered.

So it would appear to be easy for the domestic payments systems and netting systems---and soon the CLS Bank on the national authorities’ behalf---to make, for each transaction, an additional charge ad-valorem (that is, as a proportion of the amount of the transaction), the proceeds of which could be transmitted to the bank’s national authorities as payment of a CTT. This can be done electronically and automatically, and records can be and are kept for verification by central banks and regulators.

Central banks or national governments have the power to require such charges to be made---both technologically, through the necessary links between settlement institutions, and by international agreement (the Lamfalussy Accords).

Thus, given intergovernment cooperation, it should be possible, and fairly easy, to impose and collect a CTT on all wholesale foreign-exchange transactions settled between banks. Since information on all gross transactions exists and is available to central banks, there should be no difficulty over checking that the right amounts have been paid.

What about foreign-exchange trading through exchange of securities? Those who undertake these exchanges have reason to want simultaneous settlement in the same way as those trading through bank deposits. So securities exchanges also operate a simultaneous-settlement system ("delivery versus payment") in which all transactions are necessarily recorded and can correspondingly be taxed.

And how could the tax system deal with forward, futures, and option contracts in currencies? So far these will be adequately covered if the CTT is collected if and when the amounts are actually settled in currency. If in future there are in the foreign-exchange markets (as there are in others) "contracts for differences", in which only the difference between say an expected and an actual price is transferred in cash, it can be made a requirement that the notional principal amounts be recorded so that they can form the base on which the CTT is assessed.

One of the difficulties foreseen, when it was assumed that a CTT would have to be collected at the dealing-sites where trades were agreed, was that, if any one country did not cooperate in imposing the tax, foreign-exchange trading would move to that country and the tax would be avoided. Kenen [1996] pointed out that there were costs involved in moving trading facilities and that these would be balanced against the costs of the tax in determining whether trading would move site or not. However, he assumed, as did others, that it would be necessary for all countries that were now the sites of significant currency markets to impose the tax if it were not to be widely avoided.

The position would be different if the CTT were collected through the settlement system. Transactions involving any currency would need to be settled ultimately in the banking system of the country issuing that currency, either directly or through an offshore netting system (regulated by the domestic banking system as a condition of access). (Each transaction between currencies would be settled in two banking systems.) Payment for a trade or income-payment or investment transaction between two currency areas would involve settlements in (at least) the banking systems of the two areas, and these would persist in spite of the imposition of a CTT. But the overwhelming bulk of the currency transactions that take place occur between a few major currencies. Almost all, even of the transactions of "non-financial" customers, directly connected to trade or long-term investment, take place through the medium of vehicle-currencies, that is major currencies such as the Dollar, Mark/Euro, Yen and Pound. Trade in goods or services between say Brazil and South Africa would rarely involve a direct exchange of Reais for Rand but normally an exchange of each for a vehicle currency. The volume of transactions in Reais and Rand would be largely unaffected by whether Brazil or South Africa imposed a CTT through its banks or not, since they would be mainly directly determined and required by Brazilian and South African trade and (inward or outward) long-term investment and other purposes directly bearing on Brazil or South Africa for which for the relevant currency would be essential. By contrast, it is quite likely that, if say the European Central Bank imposed a CTT through the settlement system and the US did not, many of those using the Euro as a vehicle currency might shift to the Dollar, thus reducing the business for EURO-area banks and also avoiding the tax.

Moreover, most of the foreign-exchange transactions that occur are not directly related to trade, income-payments or investment. In most of them, the ultimate customers needing an exchange of currencies for these activities are not involved on either side of the transaction. It is evident from the huge volume of the world’s foreign-exchange transactions that a large part of them is not directly tied to the making of such real transactions at all, but is involved in (a) financial institutions’ adjusting their liquidity positions in various currencies; (b) hedging against exchange risk; (c) arbitrage to take advantage of differences in currency prices between markets; and (d) speculation. Most of these transactions take place between the few major currencies that are widely used as vehicle currencies, and again it seems quite likely that the imposition of a CTT by say Euroland and not by the US would shift transactions somewhat from the Euro to the Dollar and so reduce the amount taxable.

But, as we have seen, Brazil or South Africa (or India or Turkey) could probably apply a CTT unilaterally through its banking system with little fear of diversion of the taxed activities. Most of the transactions made into or out of Reais or Rand will be necessarily made into or out of Reais or Rand because they go with trade or investment (or hedging or speculation) tied to Brazil or South Africa. Each of these countries, or any others apart from the countries of the main vehicle currencies, would have a much smaller share of world foreign-exchange transactions than its share in "real" international transactions (trade, income-payments, and investment). But it is probably also the case that, just because a much higher proportion of their currency transactions follows directly from real transactions by their residents than is the case with transactions in all currencies taken together, the volume of their currency transactions is likely to contract less proportionately in response to the tax than the volume of world currency transactions. Their currency transactions are likely in other words to be more "essential", and the demand for them therefore less elastic to the burden on trading in them imposed by the tax.

Even so, the amount of revenue that could be raised by any one of these minor players unilaterally is not likely to be large in proportion to its GNP. Suppose for the sake of example that the currency transactions in a currency that are settled through its banking system each year are four times the sum of the two sides of its balance of payments and that all are taxed at 0.05%. The tax would amount to 0.2% of both sides of the balance of payments taken together, a small addition to revenue at best.

Serious revenue will depend on taxing transactions in the major vehicle currencies; and the authorities responsible for these currencies will probably judge that they have to act together or not at all.

At the same time, the fact that the overwhelming bulk of transactions (virtually all) are either between pairs from among the handful of major currencies that act as vehicle-currencies or between one of these currencies and one outside this circle has important implications for the possibilities of enforcement through the settlement system. Almost every transaction involves, on at least one side, the US Dollar, the currencies being merged in the Euro, the Yen, the UK Pound, or the Swiss Franc. (These constituted, in 1998 [BIS, 1999, p.117] 88% of the currencies involved in exchanges on either side, and were certainly involved in a much higher proportion of exchanges on one side or the other.) Including the Australian Dollar might raise the Proportion even higher. So, provided the few authorities issuing these currencies combine to impose a CTT, it can (initially at least) be enforced almost universally. These authorities will have a chance to tax virtually every transaction. If their aim is that each shall be taxed at 0.1% in total, they (and any others participating) can tax at 0.05% on each side of a transaction with a currency of a cooperating authority, and at 0.1% on the one side accessible to them of a transaction with a currency of a non-cooperating authority.

Provided the authorities issuing other currencies would be able by agreement to keep a certain part of what they collected if they joined the system, they would in these circumstances have every incentive to join. It would be important to give them this incentive in order to prevent the gradual development of vehicle-currencies outside the system.

A CTT offers a way of blocking selectively the movements involved in speculative currency crises: Spahn’s device

Paul-Bernd Spahn [1996] has proposed the device of a two-tier CTT to meet the two separate objectives: raising revenue and influencing capital movements. There might be a very low, uniform rate for the sake of collecting revenue. But arrangements could also be in place to collect a much higher, penal rate temporarily on transactions in any currency if its speed of exchange-rate movement indicated that it was likely to be subject to speculative currency flight.

The higher rate of tax would be agreed to come into effect automatically on certain objective criteria that could be taken to indicate that a currency’s exchange-rate was changing too rapidly. (The key criterion might be that the exchange-rate at which a transaction was to take place was more than a certain percentage below a specified "moving average" of its recent past values or below a declared target rate.)

The higher rate would have to be so high (on the two sides of the transaction together) that a speculative trade of currency would not be worthwhile to either party even if a rapid fall in relative value of say 50% in one of the currencies was expected by one of the parties.

If investors are confident that the higher rate will be applied in the circumstances announced, it need never actually be brought into operation. There will simply be no trades in the targeted currency at exchange-rates that would precipitate the higher rate of tax. Those involved will know that a sudden speculative run out of the currency will be impossible because participating in it will be in no individual’s perceived interest. Any dealer-bank will have been fully aware that it will be liable if it trades outside the range of rates permissible on that day.

There is no reason that the device should hinder normal and desirable exchange-rate adjustment. "Normal and desirable exchange-rate adjustment" is taken here to mean adjustment based on the reality and expectation of such "real" events as changes in relative productivity and "monetary" events such as changes in the relative volumes of different currencies created within the various domestic monetary systems. Reasonably assessed, such changes (unless themselves sudden and drastic), and expectations of them, lead to gradual movements in exchange-rates, not drastic rises and falls overnight. The purpose of the higher-tier CTT is to prevent panic (or contrived) runs from currencies by removing the expectations that lead to panic, and to do this by removing any rational individual reason for panic selling. Only extremely fast changes (how fast being determined by the exact conditions laid down) will be inhibited by the device. They will be slowed, or perhaps, if there is no good "real" or "monetary" reason for them, they will not take place at all.

Provided any government may decide to "opt out" of the possibility of its currency’s being subject to the higher-tier mechanism, the device will not in any way impair any country’s autonomy in exchange-rate policy. Any who want a completely free, market-determined rate may have it. Those who want to maintain a fixed or pegged or targeted rate will if anything be supported by the device.

If the mechanism for collecting a CTT is in place, the "higher-tier rate" designed to forestall speculative currency flights might be applied even if the lower-tier rate were zero, that is to say, even if the CTT were not being used for revenue-raising purposes at all. In other words, the urgent stabilizing function of the CTT might be brought into operation without its revenue-raising function. Or the device would allow cautious experimentation with a lower-tier rate that could start very low (say at 0.01%) and gradually be raised (with the level of revenue in mind) as the response of the volume of transactions to the tax rate became clear---while at the same time provision for the higher tier brought the key stabilizing function to bear in full force at once.

In principle the higher-tier mechanism could be applied by a country unilaterally to defend its own currency against speculative panic selling---provided only that it had the institutions (discussed in the previous section) that allowed watertight taxing of dealings in its currency. The country’s authorities could apply the tax to one side only of any transaction involving sale of its currency, but, provided the rate was high enough, this would do as well for the purpose as taxing both sides. However, awareness that the world was supporting the policy might create greater conviction in the markets that it would be successful. And this might increase the likelihood that (desirably) the threat of the tax would not actually have to be carried out.

1. There might, as in the Malaysian response to the 1997 crisis, be conventional selective capital-export controls. But the Spahn-CTT device is superior, in that, once in readiness, it need never actually be used; that consequently it need not block any transaction undertaken for purposes other than speculative; and that, because of its automatic character, it requires no arbitrary, bureaucratic decisions, and is free of the risk and suspicion of corrupt abuse. In other words, it is a transparent, market-conforming device.

2. An interest-free-deposit requirement on capital imports might be applied, as was done in Chile from 1991, with the idea of discouraging the entry of capital without long-term intentions. Studies suggested that this device did in fact shift the pattern of capital inflows toward the longer term as intended [Eichengreen, 1999, pp.51-5]. Chile certainly did not suffer from a speculative run in the 1990s as did Indonesia and Brazil and the rest, but it is impossible to say for sure whether the capital-import policy was crucial to this outcome (Chile also engaged over this period in financial-sector reform); if so, whether it would have had the same effect in the countries that did so suffer; and also whether a reasonably long prior period over which a country had been subject to this policy would be necessary if the threat of a speculative run were to be eliminated. Incidentally such a medium-term requirement for reducing short-term capital flows could be implemented, possibly in more watertight fashion, through a CTT administered as discussed above---with whatever subsequent remissions for various kinds of transaction.

3. A further possibility in principle is that there should be a joint promise from the outside world of quick and massive intervention in the market for a threatened currency. But at the moment such a promise looks politically less likely to be given than the international promise of a higher-tier CTT in similar circumstances, and less likely to be believed if it were given. And of course the many countries that can apply a CTT unilaterally have thereby a much more sure expedient.

• Some defence against speculative currency crises in industrializing middle-income countries is urgently needed.
• There is no guarantee that institutional reform of the affected countries’ banking and financial systems will quickly enough provide the defence required. It is likely to be at best a long-term process, and the hundreds of millions of people who might be affected by the next bout of crises can not wait ten or twenty years.
• A two-tier CTT has the potential for providing a watertight defence against speculative currency crises without impairing desirable market-adjustment or national autonomy in policy.
• The device is consistent with an extremely low lower-tier rate, which might be initially desirable in the absence of firm knowledge of the market’s response, and in fact it does not necessarily require a lower-tier (revenue-raising) CTT actually to be collected.
• It might be made to work by unilateral policy on the part of many of the potentially threatened countries themselves, without joint-international action, especially if it were applied through the settlement system.
• Nevertheless the knowledge that it was being applied internationally would enhance its credibility, which may be important.
• It is superior to conventional selective capital controls.
• It might be made unnecessary by Chilean-type capital-import disincentives applied in good time and to all kinds of financial instrument, but there is no guarantee of this, and incidentally they too might be implemented through a CTT at an approporiate rate.
• It seems much more likely to be politically possible in the near future than a credible promise that, in face of a threat of over-rapid exchange-rate change, there will be coordinated currency-market intervention by the rest of the world.

Contents