Back to top of the page

Back to the List of Participants with Research Papers

Back to Global Dialogue

Back to the Portal of the Global Community organization

FINANCING SUSTAINABLE DEVELOPMENT USING A FEASIBLE TOBIN TAX

(University of Antwerp – Universitary Faculties Saint Ignatius Antwerp-UFSIA)
Department of Applied Economics

Summary

This paper attempts at making a strong case in favour of the introduction of a specific type of tobin tax as a powerful instrument of the promotion of sustainable development, both directly as well as indirectly. Indirectly, it can discourage financial speculation and currency crises with their devastating effects on countries; directly, as a tax, the proceeds of it can be used as an alternative source of sustainable development finance in order to promote the establishment of international public goods. The original Tobin tax proposal can be made into a feasible instrument by engineering it as a two-tier tax (the so-called Spahn version of the Tobin tax), with tax collection through the international settlements system. A concrete scheme along these lines is proposed.

JEL Classification Nos: F3, G2.
Keywords: sustainable finance, Tobin tax, currency speculation

March 2000

Nota Bene: This document is a revised and updated version of a CIDSE background paper. The analysis and proposals presented in this paper do not necessarily reflect the views of all members of CIDSE, Caritas Internationalis and J&P Europe networks. Any comment on this paper is welcome.

INTRODUCTION

Globalisation of the world economic system is proceeding at a very rapid pace, and is generally promoted as being welfare-improving. This phenomenon is also present in the arena of international finance. In this area, however, the presumed virtues of globalisation are far from being materialised. Until now, no orderly or stable financial system has been implemented, as recent currency crises, such as the Mexican peso crisis in 1994-95, and, more recently, the currency crises in a number of South-east Asian countries, Brazil and Russia, painfully demonstrate. Furthermore, the current financial system does not succeed in channelling sufficient funds to finance crucial world problems such as adequate social development in poor countries.

The line of thought proposed here to cure these shortcomings is to use a tax instrument; a straightforward mechanism designed to tax the currently undertaxed (international) financial flows. More specifically, this proposal calls for the implementation of a tax that is levied on international currency transactions, i.e. a 'Currency Transaction Tax' (CTT). Over the past two decades, several proposals have been put forward on this subject. The most well-known of these, although not the most feasible, is the original proposal which was launched in 1972 by Nobel Prize for Economics winner, James Tobin, which called for an internati-onally uniform tax to be payable every time a currency was converted. Since then, most proposals of this nature have generally been described as "Tobin(-type)" tax proposals.

In the 1990s, the idea of taxing international currency operations, and more particularly the Tobin tax proposal, became part of a surge of interest in new international governance, with proposals including: strengthening the role of the UN, the creation of an Economic Security Council, the supervision of international banking, and a range of international taxes such as a tax on energy consumption or air travel. In 1994, the annual Human Development Report, published by the United Nations Development Program (UNDP) focused on and propagated many of these international taxes as part of a new design for development cooperation, and concluded that a Tobin-tax proposal seemed to be the most easily implemented. It received a lot of attention at the World Summit for Social Development in Copenhagen and at the 50th anniversary celebrations of the UN, both held in 1995. The issue was also considered during the preparatory process for the G7 summit at Halifax, but was at that time thought to have too many technical complications, and, more significantly, was, in fact, politically unwelcome. This political resentment was felt most strongly in the US, where the 'Prohibition on United Nations Taxation Act of 1996', designed to prohibit UN officials from developing or promoting Tobin-tax or other global tax proposals, was introduced to the US Congress. This proposition successfully prevented at least UN agencies and officials, who were at that time central to discussions on global taxes (see e.g. ul Haq e.a. [1996]), from formally discussing Tobin-tax proposals any further.

From the viewpoint of sustainable development, a very strong case in favour of introducing such a tax on financial transactions can be made, especially from a sustainable development viewpoint, because of two main reasons:

1. It offers a just mechanism to discourage excessive financial speculation in general, and, in particular, one specific type of highly-undesirable financial speculation, namely that of massive currency speculation that can trigger a currency crisis, with all its negative social effects. As such, it indirectly promotes development to be sustainable.

2. Such a tax instrument would generate revenue that could be used as one possible source of finance to help meet some of the world's global economic and political challenges, such as the promotion of adequate social development in poor countries, in other words, promote sustainable development directly. Therefore, it would secure not only the financing of necessary investment in safety nets to cure current negative social effects, but would also, and more importantly, reduce the vulnerability of the poor to the detrimental effects of possible future crises. Moreover, the successful implementation of an international financial transaction tax could act as a precedent in favour of intervention (including international taxation) with respect to other global public goods.

As such, the proposal could 'kill' two birds with one stone.

A CTT can work, in practice, if it engineered carefully, taking into account some of the indeed valid technical counterarguments to the (initial) Tobin tax proposal. A workable variant would draw upon the work of Professor Bernd Spahn of Frankfurt/Main University, who launched his proposal in 1995. Additionally, according to Smith [1999], the technical problem of collection can be overcome by using the international settlement system of transactions as a vehicle. As a consequence, essentially, only the issue of mobilising the necessary political will to implement it remains. The momentum created by the Asian currency crisis could be used to push forward for such a modest but realistic proposal. Putting this on the agenda now would also revive international discussion, and prevent further censorship from prevailing.

The paper has five sections. Sections one and two briefly sketch the nature of the problem as it relates to sustainable development and how a Tobin tax could intervene: section 1 deals with the potential negative impact on sustainable development of massive (currency) speculation, and how a CTT could help by throwing some sand in the wheels; section 2 deals with the lack of sustainable finance, in light of the massive need on it to establish a number of international public goods. Section three introduces a feasible variant of the Tobin tax, starting from the problems with the original proposal. Section four discusses some issues of implementation and suggests a concrete proposal along the lines traced out in the previous section. Section five briefly considers some further consequences.

1. The dramatic effects of excessive currency speculation

With currency trading, the daily volume of operations dwarfs any other financial market. In the last detailed market survey conducted by the Bank of International Settlements (BIS) in 1998, the global daily volume of currency transactions was estimated at about US$ 1.97 trillion. An estimated 80-90% of the transactions have no direct link with the necessities of the end-user demanding foreign exchange for international trade or payment purposes.

With respect to the time period of the transactions, the BIS statistics show that about 80% of all transactions involve round-trips (i.e. a purchase operation followed by a resale operation) within 7 days or less, and more than 40% within 2 days or less. This indicates that the majority of currency trading is short-term. Part of it is operations by large institutional investors and financial institutions routinely engaging in short-term arbitrage transactions, trading away price differences for the same currency in different markets. These operations are non-speculative by nature, but are important for the efficient fixing of prices in financial markets, and for the securing of world-wide liquidity.

Apart from that, agents engage in purely speculative (currency) transactions. In principle, large-scale speculation is triggered because the underlying 'fundamental' economic and political indicators worsen or necessary reforms are not carried out. Under a fixed exchange rate system, usually the central bank of a country tries to defend the current parity by selling its foreign exchange reserves for local currency and thereby matching the increased supply of local currency by increased demand, or by increasing domestic interest rates that increase the attractiveness of holding local currency. Such a stand-alone defence of a central bank cannot succeed for very long against the market. Alternatively, a country could try to (re)install controls on international transactions to make a currency attack more difficult; this might be more effective.

Indeed, it can be argued that intervention in currency markets should not be promoted when it goes against a 'fundamental' trend and is only used to enable countries to continue to pursue unsound economic policies, e.g. the kind of 'crony capitalism' pursued by some countries. However, intervention is justified because of the following phenomena existing in currency operations:

1. The existence of destabilising trading, i.e. so-called 'noise trading' (as opposed to trading on fundamentals). Noise-traders act on price dynamics only, driven by misinformation such as technical analysis or 'rumours'; behaviour that may drive prices away from their fundamental value. Band-wagon or herding effects, where everyone starts to mimic the action of a few leaders may, in principle, worsen the case. One of the more mythical players in this respect are the so-called 'hedge funds', that are sometimes held responsible for triggering a crisis because of the sheer volume of their operations and the demonstration effect they have on other investors, turning a currency crisis into a self-fulfilling prophecy. This also raises the issue of market manipulation by powerful agents.

2. Theories on speculative runs clearly show that speculative attacks take off too early, before a clear (current) fundamental worsening of the value of the currency is witnessed. It is driven by expectations, and it starts as soon as speculators think the attack has some chance of succeeding . As such, countries may not be given time enough to change their policies for the better. The purpose of effective intervention should be to give them time to readjust.

3. These speculative transactions are not just (private) 'zero-sum games', where one party gains what the counterparty in the transaction loses. Because of their potential to trigger a financial crisis, these 'games' can have large social costs:

· on the countries involved, especially on their most vulnerable groups. Box 1 highlights the social impact of the current crisis in Asia

· by 'contagion' on other countries that are not directly involved and which provokes a global chain reaction of financial 'panics and crashes' (the so-called 'systemic risk'). In the 1995 peso crisis, strong effects were transmitted to countries such as Argentina. The Asian currency crisis had direct contagion effects on countries such as the Philippines and Singapore, and ultimately, because of the real economic linkages, to the whole world. The direct costs are huge: the Asian currency crisis has lowered current world growth projection for 1998 alone by about 1%; since world GDP is about US$ 30,000 billion, total costs can be estimated at US$ 300 billion at least. World-wide, the ILO (International Labour Organisation), in its 1998 World Employment Report, estimated that unemployment increased by 10 million people solely due to the Asian financial crisis.

The increased potential for destabilising currency attacks is caused by the world-wide liberalisation of international capital flows, including the trend towards complete abolition of capital controls; it is seen as the virtuous twin sister of liberalised trade flows. This has facilitated global financial speculative behaviour: large sums of money can move largely uncontrolled (and untaxed) around the globe in search of the highest possible return in the shortest amount of time. There is, however, a growing consensus that challenges this positive welfare effect of increased liberalisation. More precisely, a strong case against full liberalisation can be made in the context of stabilising developing countries, which need to manage exchange rates and have relatively underdeveloped financial markets, and where the cost of failed stabilisation is high.

It is essential that an effective sanctioning mechanism, e.g. through taxation, is developed in order to eliminate this behaviour that is driven by individual short-term profit. This is the case not only from an economic viewpoint, but especially from a social justice perspective.

In order to meet this goal, additional action is required because current national and international policies are insufficient in obtaining the goals. So far, the current politically-preferred option is to maintain international financial order by a mix of strengthened requirements for the provision of information, stricter controls over all financial institutions (and not merely over banks), combined with an officially-financed 'safety net' of (largely IMF) loans. However, this approach has not prevented new currency crises, and it does not attack the speculative nature of financial operations. Instead, it looks more like "officially subsidised speculation, generating private profits and socialised losses, that have to be covered by government funding." (Raffer [1998], p.531).

The 1998 Russian currency crisis is a good example of failure to react to such a crises. Although this is not just a currency crisis, the ineffectiveness of current (IMF) policy with respect to currency crises is clearly evident here. Despite the country receiving an IMF loan of US$ 22 billion in July 1998, on ugust 1998, the Russian government decided to devalue the rouble, and allow it to float freely afterwards, with the value of the rouble rapidly depreciating. One slice of the IMF loan, about US$ 4.8 billion, seems to have been fully exhausted by Russia in an unsuccessful attempt to defend the rouble parity by supportive buying of roubles using the dollar (Financial Times [20/8/98, p.1]).

We wil argue that a transactions levy on financial flows, such as a CTT, could indeed be effective here, as it discourages and/or punishes undesired speculative behaviour. Stabilising capital flows by some sort of global tax instrument could be preferable to current policy, especially when this would also create additional revenue. However, matters are complicated because not every international financial transaction, even of a speculative nature, is undesirable. Some of these operations lead to more efficiently-functioning markets, such as arbitrage operations, or provide a desired social service, such as allowing end-customers to hedge themselves against financial risks. To give one example, arbitrage transactions can be classified as stabilising. With currencies, such as the dollar or deutschmark, margins can be as little as 3 to 5 basis points (i.e. 0.03% to 0.05%). This means that even a small tax of the same magnitude effectively wipes out this type of operation.

Ideally, an instrument that penalises speculation, e.g. by taxing, should try to discriminate against, and reduce as much as possible, the harmful effects of the system while maintaining its benefits. In particular, it should (at least) avoid 'excessive' swings that lead to a currency crisis with devastating social effects.

BOX 1: The social impact of the East-Asian currency crisis An early assessment of the social impact of the East-Asian currency crisis in three countries: Indonesia, the Republic of Korea, and Thailand. * unemployment: open unemployment has increased significantly relative to the pre-crisis year 1996. In Indonesia, the projected official unemployment rate will increase from about 4% in 1996 to between 9 and 12% at the end of 1998. In the Republic of Korea, it grew from 2% in 1996 to 4.7% in February 1998 and a projected 6.2% at the end of 1998. In Thailand, an increase from 1.5% in 1996 to about 6% at the end of 1998 is projected. Additionally, underemployment, which is much more difficult to measure, is said to be have increased considerably. In the Republic of Korea in particular, it put an end to the psychological myth of secured lifetime employment. More than occasional cases of suicide, due to loss of employment, have been reported. Most of those losing their jobs have little choice but to become active in the informal sector. * real wage levels: the strong increase in real wages of the past few years was abruptly halted in 1997 and turned negative in 1998. As an example, in the Republic of Korea, real wage growth which was on average 6.5% during 1995-96, was reduced to 2.7% during the third quarter of 1997, and -2.3% during the last quarter of 1997. It is particularly the bonus element in wages that accounts for this decline. In Indonesia, it is estimated that real wage levels will decrease by more than 15% * change in poverty levels: hard statistical material on the number of people falling below the poverty level as a result of the crisis is lacking so far. However, some indirect figures give some indication: in a report to its Development Committee, the World Bank estimates that an average drop of economic activity by about 10 per cent in Indonesia, Thailand, Malaysia and the Philippines, will result in a doubling of the number of poor people to more than 90 million (Financial Times [28/9/98, p.4]). More specific figures are becoming available on the country level: regarding the crisis in the Republic of Korea, at least an additional 100.000 people are likely to become recipients of the state's 'livelihood protection scheme'. In Indonesia, it is believed that the crisis would add "a substantial number" of people to the 22 million already living below the poverty line. As such, the crisis threatens to reverse all the achievements in poverty reduction of the last two decades. * price increases due to inflation and import curbs: especially since the last quarter of 1997, increases in basic food prices have started to accelerate, exceeding the rate of increase of overall consumer prices. The price of some basic drugs (which have to be imported) has tripled since the beginning of 1998. In Indonesia, the rice equivalent of the daily minimum wage (Rp.5800) fell from 6.28kg in January 1997 to 4.76kg at the end of December 1997. The import curbs have also hit the availability of basic food, such as rice in Indonesia. As a consequence, the president, Habibie has asked the people to fast twice a week to save on rice. * fragility of social safety nets: due to the absence of a meaningful official social safety net, loss of employment also leads to a large decline in social safety nets, since social security is by and large enjoyed by virtue of formal sector employment. The pressure on existing official schemes will be intensified, in a time where pressure on the overall government budget will not make it easy to ensure that adequate resources are available to meet the intensified demand. In Indonesia for instance, demand for subsidised public health-care service is estimated to double in 1998 to 68% of the population. * increased violence and unrest: increased cases of violence within the family have been reported. Also ethnical tension and violence is increasing considerably; pressure is particularly put on the migrant labour force. The purchasing power decrease of income has also increased illegal activities, prostitution and petty crime. Source: ILO, The Social Impact of the Asian Financial Crisis. Technical Report for the High-Level Tripartite Meeting on Social Responses to the Financial Crisis in East and South-East Asia (Bangkok, 22-24 April 1998); ILO, World Employment Outlook 1998.

2. Lack of available finance to meet global challenges

Many critical problems of today's world have global consequences: solving problems such as environmental degradation, poverty, migration, terrorism, organised crime, and the spread of diseases such as aids, need the global, concerted action of sovereign states, including the will to bear the necessary financial costs of it.

Meeting these challenges, providing the world with what the 1995 World Summit for Social Development in Copenhagen for example calls 'global public (good)s' , requires that the existing instruments for international redistribution of resources, such as development aid, which are predominantly national instruments, are complemented by truly global sources of finance .The main reason for moving to additional global sources of finance is the following: the basic concern is that (some) sovereign states default on their contributi-on to global effort, in the expectation that others will act first and save them the trouble and cost (i.e. to 'free ride'), and that this leads to collective inertia on these matters of global interest. This is typically the case for resources that are decided at the national level, and then redistributed at the global level.

Official Development Assistance (ODA), as a means of reaching the 'global commons' objective of curing world-wide poverty and adequate social development, is a clear example of this. According to the UNDP's Human Development Report 1997, attaining this goal, i.e. providing universal access to basic social services and transfers to alleviate income poverty, is affordable from a purely financial viewpoint: in 1994, the cost of it was estimated at about US$ 80 billion a year over 10 years to 2005 (Human Development Report 1997, box 6.4, p.112); this amounts to less than 0.5% of world income.

The 1997 UNDP Human Development Report reports that an amount of $ 40 billion per year would necessary for 5 years to achieve adequate coverage of basic needs in all developing countries. This figure refers to 1994 estimates compiled from different sources, based on available data from the early 1990s. The amount covers expenditures for basic education ($6 billion), basic health and nutrition (13 billion), reproductive health for all women (12 billion), and safe water and sanitation (9 billion).

Very recently, these estimates were updated within the framework of implementing the 20/20 Initiative, in a joint effort of UNDP, UNESCO, UNFPA, UNICEF, WHO and the World Bank (see UNDP et al., 1999, annex I). The aim was to estimate the cost of universal access to basic social services in developing countries, within the areas of basic public health (including nutrition), essential clinical services, reproductive health care and family planning, low-cost water and sanitation and universal primary education. These updated estimates suggest that about $ 206 to $ 216 billion are needed annually, over a five-year period, to achieve universal access to these basic services . Since current annual expenditure on basic social services in developing countries can be estimated at around $ 136 billion, the UNDP et al [1999] calculations estimate the necessary additional five-year effort to be about $ 70-80 billion annually, about twice as high as the earlier estimates. The total additional amount needed to a achieve minimum levels of human development in all developing countries is therefore, according to these estimates, in the $ 350-400 billion range.

Despite repeated calls to and promises by the richer countries to redistribute at least 0.7% of their annual GNP on ODA, realised ODA-flows do not match their targets: for OECD-countries, the current average is 0.3%, which amounts to about US$ 40 billion annually. On world basis, total ODA amounts to about 0.25% of GNP.

Moreover, the benefits of globalisation will not automatically result in more resources being made available to poor countries. For the world as a whole, the benefits of globalisation should clearly exceed the costs. For example, it is estimated that, during 1995-2001, the results of the Uruguay Round of the GATT (General Agreement on Tariffs and Trade) could increase global income by an estimated US$212-510 billion, due to gains from greater efficiency and higher rates of return on capital, as well as from the expansion of trade. However, without an explicit policy of redistribution of the overall gains, they will not result in a win-win situation; rather, it will represent a "complex balance sheet of winners and losers, where the losses tend to be concentrated on a group of countries that can least afford them" (Human Development Report 1997, p.82). It is estimated that, during the same 1995-2001 period, the least developed countries stand to lose up to US$ 600 million a year, and Sub-Saharan Africa up to US$ 1.2 billion.

Without explicit action, i.e. a redistribution of overall gains, a winners-all situation will not be accomplished. An efficient and effective instrument of redistribution can be a truly global source of finance, such as a global tax instrument. Because it generates revenue, a tax instrument, levied on a global tax base or levied more particularly on the cause that hampers attaining the global common, is more appropriate, provided it can be installed in a sustainable way.

The more general observations above are clearly illustrated by what is happening now in countries that are struck by a currency crisis. It is difficult to mobilise sufficient finance to cure the social costs of such a crisis and provide for adequate social safety nets. And, more important, in the longer term, only sufficient investment in social development will adequately reduce the vulnerable position of the poor with respect to such crisis situations in the future; here, global sources of finance can provide the necessary additional resources for this.

3. How can a Tobin-type tax help?

3.1. The simple Tobin CTT proposal

A Tobin-type tax proposal calls for a tax that would be payable every time a currency is converted. The original proposal by James Tobin, launched in 1972, calls for a internati-onally uniform tax (set at 1% in the original proposal) on all spot conversions of one currency into another, proporti-onal to the size of the transaction. Tobin's proposal has a large intuitive appeal, since it kills two birds with one stone:

Despite this intuitive appeal, the proposal was never seriously considered in the major international decision-making fora; not even when it re-emerged in periods of financial crises. Apart from the political aversion against tax measures that might in future lead to threatening national tax sovereignty by levying some sort of 'international taxes' (even if it is basically a national tax that is partly redistributed internationally), which is clearly very strong in some major countries, the proposal is highly criticised due to a number of technical problems. The left panel of Box 2 gives an overview of these problems. Clearly, some of these problems can be countered, but some, most of them linked to the specific original Tobin proposal are valid, but can be overcome using modificatons to the original proposal. The right panel of the box indicates useful modifications, to be discussed in more detail in the next two sections 2.2 and 2.3.

BOX 2: Technical arguments against the original Tobin(-type) tax and counterarguments and/or proposed solutions

Criticism. 1. Enforcement problems because of tax evasion: agents will try to shift operations to other instruments, other territories or other intermediaries that are not in the tax base. · With respect to the territorial aspect, it would require world-wide application, also in current 'tax haven' territories. · With respect to instruments, it necessitates the inclusion in the tax base of derivative exchange rate transactions, such as forwards, futures, swaps and options, and maybe also assets other than currencies, that might be considered as alternatives. For these, it would not always be easy to determine an appropriate tax rate. · Also wrt. types of institutions, it is not very clear which type of intermediaries should be taxed, and who should be explicitly exempt.	Counterarguments/modifications. 1'. Using a small base tax (under the Spahn variant, see 3.2.) will make this roundtrip too costly. · Universal application is preferred but not necessary as each transaction has two sides; when collection is done on a settlement’s basis (see 3.3.), also tax havens are included. · Tax evasion through using other asset instruments such as securities carries also a cost, since these instruments tend not to be full substitutes for the currency transaction. Using a small base tax (under the Spahn variant, see 3.2.) will make this roundtrip too costly. · If the base tax rate is small, it does not seem appropriate to waiver many parties, with the exception of central banks, BIS, …
2. It is impossible to clearly distinguish between desirable and undesirable operations. So, the tax would re-introduce distortions and increase the cost of 'normal' currency transactions. It could be that some valuable market segments would be killed completely due to the tax.	2'. Correct. A modification that solves this problem is using a two-tier system (see 3.2.)
3. Furthermore, it is not at all clear that 'undesirable' speculative behaviour is mainly short-term, nor is it clear that the short-term should be penalised relatively more.--	3’.Correct. The effect can be reduced by prefering a low base tax rate, and make high rates contingent on the circumstances, not on the maturity of the deal( i.e. a two-tier system; see 3.2.)
4. It is not clear at all that the tax would result in lowering exchange rate volatility	4. Correct. Less liquidity through a tax might even increase price swings. But at a low base tax rate, liquidity will not be affected much.
5. the determination of the exact magnitude of the tax rate: if imposed at a high rate, the tax would seriously impair also the more desirable operations of financial markets or would induce massive evasion efforts. Yet, if imposed at a low rate, it would not deter speculators .	5’. Correct. A modification that solves this problem is using a two-tier system (see 3.2.)
6. Moreover, it might provoke a never-ending rat race of agents finding new loopholes to evade the tax and regulators, and tax authorities closing them.	6’. Possibly correct. But using a low base tax (see 3.2.) reduces the incentive to evade. And collecting through settlement makes it much harder to evade.
7. You can not introduce and collect a universal tax in a fully-decentralised market such as the foreign exchange market.	7’. While dealing is indeed fully decentralised, settlement of trades becomes highly centralised (see 3.3.)

3.2. One realistic way out: a two-tier CTT à la Spahn

A basic flaw of the original proposal is linked to determining the exact tax rare: if imposed at a high rate, the tax would seriously impair also the more desirable operations of financial markets or would induce massive evasion efforts. Yet, if imposed at a low rate, it would not deter speculators, or, as Davidson [1997] puts it: grains of sand in the wheels is not enough where boulders are required. Fortunately, there is a solution to the basic flaw in the original Tobin tax just mentioned. A commonly-known solution to this type of problem is to match targets and instruments by installing a separate instrument for each goal to be attained. Applied in this context, it would ask for an instrument that, on the one hand, creates a fairly high revenue because of the sheer magnitude of the tax base involved, without inducing widespread evasive behaviour and without creating additional distortions, and, on the other hand, an instrument that is activated only in very specific circumstances, i.e. when large national social costs and systemic effects are incurred.

A CTT proposal of this nature is being suggested by Bernd Spahn [1996] , i.e. a two-tier Tobin tax, levied as a national tax but introduced through an international agreement, with a minimal-rate transaction tax on all transactions (the 'basic tax'), and a high tax rate (an exchange 'surcharge') that, as an anti-speculation device, would be triggered only during periods of exchange rate turbulence and on the basis of well-established quantitative criteria.

· The minimal nominal charge of one or two basis points (i.e. 0.01 or 0.02%) of the basic tax would not create a significant distortion, would not create massive evasion efforts and, as such, would establish a considerable revenue base. In addition to the revenue aspect, such a tax would act as a monitoring device, facilitating the follow-up of movements in the market.

· The surcharge would aim to tax, at a prohibitively high rate, the occurrence of 'excessive' exchange rate volatility. Spahn suggests a mechanism that is very similar to the (old) European Monetary System's mechanism for achieving exchange rate stability, with the surcharge being switched on whenever the trading price for a currency passes a predetermined threshold. This threshold would be determined by a changing target rate (a 'crawling peg', determined by a moving average) plus a safety margin (defined as a percentage of the target rate, creating a band). When the currency fluctuates within the band, no surcharge is levied - only large-scale speculative behaviour would induce the actual exchange rate to break out of the band and trigger the tax mechanism. As such, the mechanism rather discourages speculative behaviour because of the tax threat; ideally, the revenue from the surcharge is zero. Unlike the EMS however, exchange rates would be kept within the target range through taxation rather than through central bank intervention (which typically uses interest rate increases or depletion of international reserves).

The essence of the surcharge mechanism can be easily clarified when illustrated graphically, as below (figure 1). The exchange rate parity changes along time as a result of actual exchange rate changes, using a moving average technique. Around the parity, a lower and upper bound is determined in proportion to the target rate, creating a band area. As long as the actual exchange rate moves within the band, only a small tax is levied. Only when the actual exchange rate jumps out of the band (as a result of a large exchange rate change, possibly as a result of a speculative attack), is the difference between the actual and the bound (as illustrated by the shaded area in the graph) taxed at a high proportional rate.

Figure 1: The Spahn tax surcharge mechanism