Abstract

The European Union is experimenting with new non-binding policy instruments in business taxation, namely a voluntary code of conduct amongst member states against harmful tax competition. This paper discusses the context within which the code has emerged and provides an explanation of the political aspects of this instrument. It also discusses what has been achieved by the code and raises the question of its contested legitimacy. The conclusion is that the code has generated a process of cognitive convergence in terms of policy-makers' beliefs. Further, it has contributed to the institutionalisation of the tax policy paradigm based on the concept of harmful tax competition. In terms of policy change the evidence is still preliminary. The code is embedded within wider policy initiatives (namely, the tax package proposed by the Commission, the fiscal components of state aid policy, and the OECD forum on harmful tax practices). Its success is therefore contingent on the future of these wider initiatives.

As defined by the Lisbon Council of March 2000, the aim of the open coordination method is to produce convergence by diffusing best practice. The idea is to use the EU as a transfer platform rather than a law-making system. The main instruments of the method are guidelines and timetables, benchmarking, indicators and scoreboards. The method is based on mutual learning processes of peer review (note: INSERT CROSS-REFERENCE TO THE INTRODUCTION ONCE THE INTRODUCTION IS DONE).

Is direct tax policy amenable to the open coordination method and if so how? This policy poses serious challenges to the European Union (EU). There are at least five hurdles. First, both sovereignty and legal arguments - unanimity and the lack of a specific treaty base for direct taxation (in contrast with the relatively precise articles of the Treaty of Rome on indirect taxation) - produce a very narrow path for direct tax proposals. For example, the idea of peer review of domestic tax systems (peer review being a cornerstone of the open coordination method) was considered nothing less than political anathema by several governments up until recently.

Second, although the idea of direct tax harmonisation or coordination has been aired since the 1960s, there is disagreement about the content of a Community policy in this area. This is because it is difficult to see the end result. Contrast this with the Euro, where a simple concept (that is, one single currency for the EU) can be easily grasped by politicians and citizens. But the question of 'what could a possible fiscal system for the EU look like?' does not lead to simple answers. The open coordination method implies that the main EU policy goals have been set (Wincott 2001). Indeed, the method stipulates that convergence should be 'towards the main EU goals' (Lisbon conclusions, p.12, para.37). If the content of the EU tax goals cannot be specified, convergence becomes an elusive notion.

The problem is compounded by the lack of desirable models to be imitated and diffused throughout the Union. The lack of agreement on what is good or best practice in terms of tax policy is the third hurdle which has historically hindered progress in this area. Tax systems developed around different ideas about the proper role and size of the state (Steinmo 1993). Although there has been some convergence of policy-makers' beliefs around certain desirable properties of tax systems (McLure and Zodrow 1994), there is no shared notion of 'best' tax models in EU policy circles. Turning again to the comparison with the Euro, this means that there is no tax functional equivalent (in terms of national tax models) to the German model of monetary policy in the Economic and Monetary Union process (Radaelli 2000).

Finally, there are two hurdles revealed by theoretical analysis. The lack of theoretical and empirical uncertainty on who bears corporate taxes in open economies with relatively efficient financial markets makes the formulation of tax policy difficult. For example, every time the Commission has proposed measures to coordinate the taxation of savings, some economists and politicians have observed that in open markets these measures may end up in higher taxation of labour because companies will get the same return on capital whatever tax is introduced (Devereux 2000, Bond et al. 2000).

The last hurdle is that cooperation in taxation is generally self-destroying (Genschel and Plumper 1997). If n-1 countries decide to cooperate, the advantages of the country remaining outside cooperation increase. Consequently, any form of EU tax coordination is at risk if countries outside the Union pursue aggressive tax policies. By contrast, in areas such as standard-setting (think of telecommunication standards) if a sufficiently large number of countries decide to cooperate and agree to a standard, they will provide a critical mass for convergence. Typically, a large majority of countries will opt for the same standard if they want to trade in international markets.

Thus, not only is the political and strategic context a very daunting one for the adoption of the open coordination method (lack of best practice, sovereignty as an obstacle to peer review, no Economic and Monetary Union-like timetables for convergence around a clearly specified 'final system' etc.), it is challenging for the formulation of direct tax policy tout court. The aim in this paper is to show how, notwithstanding these constraints, direct tax policy has reached momentum and new policy instruments - notably, a code of conduct in business taxation - have been adopted. The major theme in this paper is to discuss whether the code of conduct can be considered an example of the open coordination method. As the code is a non-binding instrument, a second (but minor) theme will be the comparison between the code and the voluntary agreements and codes used in other policy areas, notably EU environmental policy (Jordan et al. 2001).

What are the arguments presented here? One argument in this paper is that the tax code is consistent with the basic features of the open coordination method, although it was introduced before Lisbon and no reference to it was made by the Council's group working on the code. In a sense, one can talk of open coordination method in disguise. However, it will be shown that the logic of the tax code is completely different from the one operating in the new environmental policy instruments. The latter respond to the problems of (a) finding alternatives to command and control regulation and (b) integrating economic and social actors (eminently industry) in the policy process. The tax code instead responds to the problems of (a) political necessity and (b) giving a prominent position to the goal of guarding the revenue base, even at the cost of leaving the business community somewhat outside the policy process.

Another argument presented in this paper is that the code is embedded in a wider context and therefore cannot be assessed without a consideration of contiguous and interdependent initiatives to crack down on harmful tax competition. The future of soft new policy instruments in EU tax policy is therefore contingent on the progress made in the traditional law-making process. Soft and hard approaches are inexorably intertwined. Having said that, the major result achieved by the code so far is the cognitive convergence of EU tax policy-makers.

The paper is organised as follows. Section 2 presents the historical context and illustrates how the current tax policy emerged in the second half of the 1990s. Section 3 provides a discussion of the political aspects surrounding the use of a non-binding agreement such as the code of conduct. Section 4 discusses what has been achieved so far. Specifically, it raises the question of whether cognitive convergence (the main result achieved by the code) has also produced policy change. This Section has also the aim of raising the issue of legitimacy. Section 5 presents the conclusions.

It is useful to situate the code in its historical context. What other techniques of coordination were used before the code was introduced? What type of learning process led to the shift to soft law?

Up until the mid-1990s, the Commission tried to overcome the hurdles described above by making reference to the single market. The argument of the Commission was that the completion of the single market required the elimination of domestic taxes which created distortions, typically double taxation of companies operating in two or more member states (Commission 1990). This argument was rooted in the paradigm of tax neutrality and economic efficiency. The Commission insisted on the implications of the single market for tax policy. Economists argued that the efficiency of the single market needed the elimination of international double taxation (Devereux and Pearson 1989). They endorsed the general thrust of the Commission's corporate tax initiatives. Companies flanked the efforts of the Commission. They were well represented in a high-level committee set up by the Commission (the so-called Ruding Committee, Commission 1992) with the aim of proposing a blueprint for the future of EU business taxation. At that time, the Commission proposed directives, although at least in one case (Commission recommendation on the taxation of non residents 21 December 1993) Brussels opted for a recommendation to member states. Overall, the focus was eminently on traditional policy instruments and the main goal of the Commission was to get a number of directives through the EU tax policy process.

The result of this strategy was somewhat poor. Two directives were approved in 1990. In the same meeting, the Council also approved a convention (instead of a directive, as the Commission had originally suggested) on arbitration in transfer pricing disputes. But nothing else happened, and the proposals for tax directives dusted on the shelf or were withdrawn by the Commission. No peer review of the domestic tax systems took place. More generally, the Council did not pay political attention to the argument of the Commission. The idea of eliminating distorting taxes - whatever the technical appeal of the argument might have been - was not very attractive to ministers of finance.

Since 1996 the Commission has changed its strategy to overcome the obstacles to tax coordination. The new strategy is based on the necessity to crack down on harmful tax competition. This strategy chimes with the long-term interest of revenue authorities, that is, to guard the revenue base. The response of member states was more positive than in the past. Political determination to curb certain forms of tax competition was achieved. Once the ball started rolling, the problem became to select the policy instruments suitable to the new tasks. Initially, the Commission, the European Parliament and some delegations thought of a directive on company taxation and a directive on the minimum taxation of savings in the Community.

Thus, at the beginning of the new anti-harmful tax competition crusade, there was no deliberate intention to introduce new policy instruments. The debate on taxation was not focused on the merits of non-binding instruments as alternative to traditional regulation. There was no discussion about bringing the economic and social actors in the tax policy process. All this makes the emergence of the tax code very different from the introduction of codes of conduct in environmental policy. As mentioned above, a number of actors were still confident in traditional policy instruments and wanted a directive on business taxation.

Even Luxembourg, a not-so-keen country on tax coordination, would have preferred a directive on company taxation rather than a non-binding instrument. But the explanation of this preference for a directive has something to do with the position of this country in the global competition for capital. EU member states have a different relative specialisation in tax competition. Ireland (Dublin Docks) and Belgium (coordination centres), for example, are very attractive locations for foreign companies. They were the natural target of a possible directive against harmful tax competition in company taxation. By contrast, Luxembourg has developed a relative specialisation in attracting savings. In other words, Ireland and Belgium are very competitive in corporate taxation, whereas Luxembourg is more competitive in attracting portfolio income. Hence the position of a country like Luxembourg, which saw the proposal for a directive on business taxation as the natural counterbalance to the losses to be incurred as a result of a directive on saving taxation.

Shortly, for a deal to be struck in the Council, conflicts (between different countries with different structures of pay-offs) had to be settled. Innovation in tax policy instruments emerged as a political expedient when it became clear that agreement on a business tax directive was impossible to achieve. Necessity - rather than a reorientation of regulatory approaches - led the Council towards the code of conduct. The Council (a) dropped the idea of a directive on company taxation and (b) composed the conflicts in a deal struck in December 1997. The ECOFIN Council agreement (OJ C 2, 6.1.1998) is based on a three-piece tax package and a fourth element concerning fiscal state aid. The tax package includes the code of conduct, a proposal for a directive on the minimum taxation of EU non-residents, and a proposal for a directive against the double taxation of multinationals. These three elements are formally bundled, which means that lack of agreement on one element makes agreement on the other two elements impossible. Let us now take a look at what's in the package.

The first element of the 1997 agreement is a voluntary code of conduct on business taxation (to be discussed below). The second component of the deal is the commitment to ensure a minimum of effective taxation of savings within the Community. In 1997, the Council requested the Commission to come up with a proposal for a directive and set a few points around which the proposal should be fleshed out. Following this invitation, the Commission presented a proposal for exchange of information or a 20 per cent withholding tax on interests paid to non-resident EU citizens in May 1998. This is the proposal that raised many objections in the City of London, worried by the possibility that capital markets would react negatively to the European tax on savings, and migrate elsewhere (Baron 1999). Further to an unsuccessful attempt to mitigate the worries of the British delegation in Helsinki (December 1999), the EU debate has veered towards exchange of information as the best option. This would mean the abolition of bank secrecy for EU non-residents. The Feira EU summit (20 June 2000) recognised the possibility to opt for the coexistence model (that is, withholding tax or exchange of information) for a limited period of seven years, after which all countries will run a regime of automatic exchange of information among tax authorities.

At Feira, the EU leaders dropped the May 1998 proposal for the taxation of savings, but set the coordinates for a new one. They agreed on a process that should lead to the approval (unanimity voting applies) of a three-piece package (a directive on the taxation of EU non-residents’ savings, a directive on interest and royalties and the full implementation of the code of conduct) by 'no later than 31 December 2002' provided that (in the area of the taxation of savings) ‘third countries’ implement ‘equivalent’ measures and that the dependent or associated territories of EU countries (the Channel Islands, Isle of Man, and the territories in the Caribbean) enact the same measures as the EU member states. By 2009, all member states should switch to exchange of information as the rule for the taxation of EU non-residents’ savings. At Feira, the European Council endorsed only a timetable, but, in doing so, it generated momentum for tax coordination. It also established the principle that exchange of information (rather than withholding taxes on non-residents’ savings) is the main target of the EU. Another important step was taken at the ECOFIN Council meeting of 27 November 2000, when the French Presidency secured an interim deal for those countries that - within the time limitations set at Feira - want to use withholding taxes rather than exchange of information. Austria, Belgium and Luxembourg will levy a withholding tax at 15 per cent for three years. After that, the tax rate will go up to 20 per cent. A regime for the so-called grandfathering of Eurobonds was also agreed. Thus some coordinates of the proposal for the taxation of savings are in place. But for the proposal to become directive, it will be necessary to demonstrate that the 'third countries' listed at Feira enact equivalent measures.

The third element is the proposal for a directive on cross-border payments of interests and royalties. This is a typical single-market tax measure. Conceptually, it has nothing to do with the EU fight against harmful tax competition. As such, the proposal should have been approved long time ago. However, the three elements of the package are still linked together. As averred, this means that the success of the code depends on the possibility to finalise the directive on savings and the directive on interest and royalty payments at some point in time next year. Luxembourg and Austria have made this link explicit when they wrote in the ECOFIN's minutes (27 November 2000) that they will agree to the directive on savings only if the Council reaches a binding decision on the roll-back of harmful company tax measures. The soft law elements of the code are therefore linked to the 'hard' components of the tax package. Additionally, the reference to 'binding decisions' made by Austria and Luxembourg introduces a somewhat 'hard' component in the code.

The final element concerns fiscal aids. This is an element of the 1997 agreement, although technically it falls outside the tax package that has yet to be finalised. What is the reason behind the inclusion of state aid in the 1997 agreement? Simply put, special tax regimes have too often been built under the rubric of legitimate state aid policy, thus circumventing the scrutiny of the tax Directorate of the Commission. The essential point here is the connection between tax policy and state aid policy. In a press release (23 February 2000), Monti stated that the DG for competition 'will examine all the relevant cases of fiscal state aids in business taxation, so as to allow the Commission to comply fully and promptly with its own institutional obligations’. The connection between state aids and taxation requires an examination of the progress made in the code of conduct. It is to the code that we now turn.

The code of conduct defines harmful tax competition. It covers the fifteen countries of the EU and - the point is extremely sensitive in tax policy - their dependent territories. The code - as defined by the 1997 Council agreement - includes general provisions for standstill and rollback of harmful tax regimes according to a five-year timetable. Roll-back can take two forms: (a) elimination of the tax measure or (b) elimination of the harmful components within the tax regimes without suppressing the regime itself.

As argued above, the choice of a voluntary code was the result of political necessity. Most governments felt that a directive would erode political sovereignty and would be too difficult to manage. By contrast, the notion of a non-binding instrument hinging exclusively on political determination calmed political apprehension.

When the code was introduced, the then Commissioner for the single market and tax policy did not start any comprehensive review of non-binding instruments in EU policy. Although at that time non-binding instruments were already in use in other policy areas, the Commissioner felt that there was neither time nor staff available for an inquiry on what these instruments could achieve and their transferability to tax policy. Consequently, the problem of how to make the code work was very much a learning by doing exercise for the Commission and the member states.

A Council group (the so-called Primarolo group, dubbed after its chair, British MP Dawn Primarolo) was established to manage the code. The challenges for the Primarolo group were formidable. To begin with, there was the problem of experimenting with the first example of soft law in corporate taxation. As already mentioned, this experimentation was taking place without the benefit of a 'lesson-drawing' exercise looking at new policy instruments in other EU policies. Neither was is possible to situate the code in a wider corpus of EU corporate tax directives: the law and jurisprudence on EU direct corporate tax law was limited (Farmer and Lyal 1994). The option of co-regulation, that is, to develop a non-binding instrument in the shadow of an existing and coherent body of regulatory laws, was therefore impossible.

The other challenge was political. For the first time in history, fifteen high-level tax policy-makers were sitting with the Commission around a table to undertake a peer-review of potentially harmful tax measures. To see an Irish delegate commenting on the potential harm created by the Belgian coordination centres was a sea-change in the tax history of the EU - an history dominated by the notion of sovereignty in their domestic direct tax legislation. How to find agreement by dint of peer review in sensitive areas such as special tax regimes for foreign multinationals, the tax systems of the Channel Islands, the legislation of the Trieste (Italy) financial services and insurance centre was a bit of a political mystery.

The third challenge was cognitive. The 1997 Council's agreement defined in general terms harmful tax competition. The issue to tackle was then 'what happens when the criteria hit the road of real-world decisions (on what is harmful and what is not)'? Note that there is no scientific consensus on the theoretical definition and the empirical consistency of harmful tax competition (Basinger and Hallerberg 2000; Ceps 2000; Dehejia and Genschel 1999; Gordon and Bovenberg 1996). The criteria identified by the Council do not derive from a list on which a representative panel of economists could agree. Quite the opposite indeed, some economists would simply argue that there is tax competition, but it is a nonsense to discriminate between harmful and non-harmful competition (see Devereux's statement in House of Lords 1999).

Political considerations tend to predominate because it is difficult to pin down what 'harmful' tax competition is. The Council group managing the code has to balance technical arguments (often open to different interpretations) with political arguments. For example, one of the criteria used by the code to define harmful tax competition is that a tax measure affects the location of business activity in the Community. But taxation is not the primary consideration in decisions about location. How can one demonstrate that location is affected by tax considerations and not by the quality of the work-force, transport costs, or other variables? Although economic analysis produces some useful answers to this question, there is plenty of scope for political decisions within the group.

On balance, the criteria of the code (see box 1) are the product of contingent political agreement. The Primarolo group set out to produce convergence on the practical meaning of the criteria. In addition, the code asks member states to commit themselves to re-examining existing tax laws and established practices having regard to the principles underlying the code' (par.D of the code) and to 'inform each other of existing and proposed tax measures which may fall within the scope of the code' (par.E of the code). Shortly, for the code to deliver, a re-orientation of tax policy beliefs had to take place. The question is therefore whether cognitive convergence has now taken place or not? In order to address this question, we now turn to empirical evidence on how the Primarolo Group managed the code.

The Primarolo group worked on a long list of 271 potential tax measures in member states and dependent territories. After a series of meetings and peer review sessions, the group reported on the implementation of the code to the Council on 29 November 1999 (SN 4901/99). The Helsinki summit (December 1999), however, was unable to agree on the Primarolo report. The important political point is that while the criteria listed in the Code of Conduct were discussed and formally agreed by all Member States on 1 December 1997, the 1999 Group's report was not. Consequently, reluctant member states can always rely on this lack of open endorsement (of the Primarolo report) in they do not want to implement the code in full. The November 1999 report listed 66 harmful tax regimes. But the report reflected either the unanimous opinion of the members of the group or the various opinions expressed in the course of the discussion. The reservations entered by some delegations to the conclusions of the Primarolo group speak volumes on the degree of agreement on what needs to be done in the near future. Whenever Council documents refer to the "code of conduct group's deliberations" , they indeed refer to broad consensus but not necessarily unanimity.

This ambiguity - in terms of the extent of substantive agreement among the 15 countries - can affect the implementation of the code. For example, in a progress report presented to the Council in November 2000, although all delegations agreed on the principles of standstill and rollback, one delegation objected to the specific criteria used to implement the principles. Nevertheless, the code of conduct group was able to meet every month since the European Council of Feira. Paradoxically, the group seems able to proceed in detailed analyses of the harmful tax regimes and the criteria for rollback, although the 15 countries have not resolved some fundamental issues on what they really want to achieve. The outcome is that at every meeting of the group there is some area of agreement, but if one scratches below the surface one finds that there are at least two or three views on what the real implications of the agreement are!

The ambiguity on the level of substantive agreement is increased by the fact that the code is still a component of the tax package to be agreed by the end of 2002. Until then, member states have all sorts of 'exit' options if they do not want to abide by the decisions of the group.

Before we turn to an assessment of the code, it is useful to observe that this instrument is embedded in a panoply of initiatives against harmful tax competition. In terms of policy dynamics, the future of the code - to use a metaphor - overlaps with three policy circles. The circle which is most immediate to the code is the one of the tax policy package. As averred, if the Council cannot agree on a directive of savings, some member states will be ready to terminate their contribution to the code policy process - or at least they will urge a new discussion of roll-back. In this sense the code is contained in the circle of the tax package. Consequently, the 'autonomy' (in terms of what can be achieved autonomously by the instrument) of new policy instruments such as the code is severely constrained. Put differently, if the 'old-style' directive on savings cannot be finalised by the Council, the 'new' non-binding instrument will not go far.

There is a second policy circle (that is, fiscal aid policy) which overlaps with the code. This is not a matter of concentric circles, but one of political interplay. As noted above, the Council gave a broad mandate to the Commission (in December 1997) to re-assess those fiscal aids possibly in breach of the fair tax competition paradigm outlined in the code. Now, there is a big difference between a non-binding instrument, the code of conduct, and state aids, where the Commission has considerable power. However, politically the 'arenas' of the code and state aids are connected.

The Commission gains considerable leverage from these nested arenas. When the 'decisional speed' in the code of conduct arena decreases (because governments disagree on how to make progress with the implementation of certain steps), the Commission gets ready for action in terms of competition policy. When the process in the area of the code (and more generally the tax package) re-starts, DG Competition seems inclined to a prudent wait and see policy. For example, during Autumn 2000, when the chances of making progress on the code and the tax bundle appeared low, DG Competition 'loaded the gun' by starting preliminary investigations on selected fiscal aids. Formal state aid procedures were not open however, to keep pressure on the national delegations negotiating in the code arena. The gun was loaded, but the trigger was not pulled. Soon after an agreement on the tax package was achieved (27 November 2000), DG Competition manifested no intention to open formal procedures against the fiscal aids object of the preliminary investigation.

Clearly, this is a political mechanism of threats, sanctions, and rewards. Member States are rewarded for their 'good behaviour' at the code of conduct table by putting an hold on state aid procedures. Some delegations have sought to take advantage of this connection between code and fiscal aids by stating the following: in the presence of compliance with the recommendations of the Primarolo group, a member state should be openly rewarded by DG Competition's promise not to start state aid procedures. This is clearly unacceptable: the Commission cannot make this type of deals. However, it remains true that the Commission is clearly taking into account progress and compliance in the code of conduct group in the formulation of its state aid policy. Concluding on this point, the link between state aids and the code advantages the Commission by giving it some political leverage in terms of threat, rewards, and sanctions.

The third policy circle is outside the EU. It refers to the OECD forum against harmful tax practices (OECD 1988). Although the EU code of conduct and the OECD initiative against harmful tax practices share different goals and cover different types of economic activity (see the analysis contained in Ceps 2000), politically the EU policy against harmful tax competition benefits from a similar orientation at the OECD level. Further to the OECD (1998) report on harmful tax competition - which contained 19 detailed recommendations to combat unfair tax practices both within the OECD and in tax havens outside the organisation - a second report was published in June 2000 (OECD 2000). This report identifies 47 tax regimes in OECD countries which are 'potentially harmful'. Unsurprisingly, the list includes the Belgian coordination centres and the Irish international financial services centres. Interestingly, the OECD has made the cautious political choice not to say whether these regimes are harmful or not. They are 'potentially harmful' - the OECD is currently assessing whether they are actually harmful or not between now and the year 2003. The point is that the determination of governments to roll-back of the harmful measures identified by the Primarolo group can be increased to a great extent by acknowledgement at the OECD table that those measures contradict OECD tax policy. If the OECD project flops, there will be yet another reason in national capitals to show less enthusiasm in complying with the recommendations of the Primarolo group. Overall, the OECD table represents an additional potential element of peer pressure.

So far the main achievement of the code refers to the cognitive aspects of tax policy. For the first time in history, the members of the EU have accepted the idea that reciprocal discussion of domestic tax regimes is acceptable, legitimate and even useful. Up until a few years ago, the idea of peer review of special tax regimes (some of which were characterised by low transparency) was utterly inconceivable. This is a remarkable result, although - as observed above - there is still disagreement on how to implement certain elements of the 1999 report of the Primarolo group. The code has been a very successful laboratory for trust and mutual understanding. It has also contributed to the institutionalisation of the theme of harmful tax competition. A theme - it should be noted - that is far from being fully accepted by social scientists, hence a theme in which there is no clear message coming from communities of experts.

Has cognitive convergence produced policy change then? It is too early to provide an answer to this question because the tax package is still the object of negotiation. Although the Council group has been able to meet regularly and to take decisions, the political determination to roll-back harmful tax regimes may vanish in the absence of agreement on the tax package. Being a non-binding measure, no legal action can be taken against a country unwilling to scrap its own tax regimes. In this connection, changes in governments may also represent a problem for the code, although up until now there has been no policy 'U-turn' on the code following national elections.

There is evidence that, notwithstanding this state of uncertainty, the code is already generating some policy change. For example, recent changes in the Netherlands' intermediate royalty and interest companies, advance pricing agreements and advance ruling practices have been linked to the intention of the Dutch government to comply with the criteria listed by the code. Should further evidence point in this direction, one should conclude that the code has acquired a life of its own - that is, the power to create domestic policy change independently of the approval of the whole tax package. At the moment, however, this conclusion would be premature.

Roll-back is the real acid test of the code. However, progress has already been made in other areas such as notification. Governments have been willing to submit proposals for new tax policy regimes. This has increased the transparency and the commitment to peer review in EU taxation. Aggressive, 'beggar-thy-neighbour' tax regimes have become less popular in the EU. It would be politically difficult to propose now the same type of regimes which were so popular up until the mid-1990s. This is evidence that the policy beliefs enshrined in the code are not irrelevant. However, this is not the end of tax competition. One type of competition may be less popular than in the past, but governments may turn to other types of competition. And it remains to be seen whether competition limited to a handful of special tax regimes (this is the type covered by the code) is less harmful than 'across-the-board' tax competition (the point is explored by Keen 1999).

Looking beyond the domain of tax policy change, the relationship between the code and legitimacy has been problematic. The Council group has met in secrecy. The agenda, the minutes and the results of the review group have not been published -- or have been published with considerable delay. There have been no hearings of the business community. Yet companies are the main economic agent involved in special tax regimes. They need to know whether they are putting their money in regimes which are about to be scraped or not.

Not only has the Primarolo group angered multinationals, but the lack of transparency has ruffled the feathers of MPs. In 1999, some parliaments debated the issue of harmful tax competition (French Senate, 1999; House of Commons Hansard Debates, 5 July 1999) without having access to the results of the Primarolo group. In London, at the House of Commons, two MPs expressed their disappointment with the following words:

According to some politicians, secrecy has made the code ‘nothing but a PR disaster’ (Lord Desai, see House of Lords, 1999: 163). The Commission and the members of the Primarolo group have defended secrecy on the basis that open discussion of national measures would have provoked havoc and uncertainty. This argument could have been reasonable – had the review group concluded its work in a few weeks. But secrecy over a long period has created uncertainty, a few panic attacks, and, unsurprisingly, more than one episode of leakage. Additionally, secrecy has deprived the review group of valuable feedback from professional organisations, tax experts and the business interests affected by the code. Finally, lack of transparency is exactly what the current Presidency of the EU and the President of the Commission Romano Prodi want to eradicate from the EU culture.

The political cost of preserving consensus among member states and calming their apprehension at the code's table has been high in terms of legitimacy. So much so that the most recent communication of the Commission (2001) seems to re-direct future initiatives towards the elimination of the tax obstacles faced by taxpayers (citizens and multinationals) doing business in the single market. This would balance the emphasis of the code on the elimination of harmful tax competition (Ceps 2001).

The code of conduct on business taxation fits in rather well with the approach and the content of the open coordination method. Voluntary (as opposed to binding) agreement, peer review, and timetables make the code consistent with the thrust of the method. The prominent role of the Council is another political similarity between the code and the method, although the question of 'who is in charge' of the open coordination method is more complicated than this (Dehousse 2001, Wincott 2001). To be true, the adoption of the code in 1997 was also an important victory of the Commission. The latter had tried to mobilise member states against harmful tax competition for a while. Consequently, this issue cannot be treated in terms of power dichotomy - either the Council, or the Commission.

Turning to best practice, at first glance it seems that the code has nothing to say on this component of the open coordination method. The criteria identified by the code do not define best practice directly. However, by highlighting the harmful dimension of tax competition, they show indirectly what good practice in tax policy is. In a sense, the criteria provide a functional equivalent to best practice. The latter does not make sense in EU tax policy - as argued in Section 1, there is no idea of what the 'best' or 'good' fiscal system should be. To determine 'worst' practice is therefore the closest EU business tax policy can get to the open coordination method's emphasis on best practice. On balance, the code can be considered an example of the method, although EU tax policy-makers have not mentioned this connection. In this sense, it is an instance of open coordination method in disguise.

Compared to other voluntary policy instruments (Jordan et al. 2001) used in environmental policy, the code has a completely different logic. Typically, codes of conduct have been introduced to make regulation more flexible and suitable to the needs of economic and social actors. By contrast, in the code on business taxation economic actors feel excluded. The code is more the response to the problem of calming governments' political apprehension over EU tax coordination than a way to bring the society back in the EU policy process.

Another distinctive element of the code is that it does not operate in the shadow of a relatively mature body of EU tax legislation. The code operates however in the shadow of other tax policy initiatives, both at the EU and the OECD levels. It cannot be assessed on its own merits. Thus, the future of the open coordination method in direct tax policy hinges on the progress of more traditional elements of EU legislation.

Baron, R. (1999) Serious damage: The impact of the withholding tax on the city of London, London: Centre for Policy Studies.

Basinger, S. and M. Hallerberg (2000) Remodeling the competition for capital. How domestic politics erases the race-to-the-bottom, Paper delivered to the annual APSA meeting, Washington, 31 August - 3 September 2000.

Bond, S., L. Chennels, M. Devereux, M. Gammie and E. Troup (2000) Corporate Tax Harmonisation in Europe: A Guide to the Debate, IFS Publications: London.

CEPS, Centre for European Policy Studies (2000) The Future of Tax Policy in the EU. From harmful tax competition to EU corporate tax reform, Brussels: CEPS Publications.

CEPS, Centre for European Policy Studies (2001, forthcoming) EU Corporate Tax Reform, Brussels, CEPS Publications.

Commission (1990) Guidelines on company taxation, Communication, 20 April.

Commission (1992) Report of the committee of independent experts on company taxation (Ruding Committee Report), Luxembourg: Office for Official Publications.

Commission (2001) Tax policy in the European Union. Priorities for the years ahead, Communication, COM(2001)260 final, Brussels, 23 May 2001.

Dehejia, V. H., and P. Genschel (1999) 'Tax Competition in the European Union' Politics and Society, 27, 3, 403-430.

Dehousse, R. (2000) Du bon usage de la méthode ouverte de coordination, Notre Europe, typescript, Paris, 13 October 2000.

Devereux, M.P. (2000) Recent policy initiatives in the international taxation of capital, Paper delivered to the TRN (Tax Research Network) conference, The University of Birmingham, 11 September 2000.

Devereux, M. and M. Pearson (1989) Corporate tax harmonisation and economic efficiency, London, IFS Report Series no.35.

Farmer, P. and R. Lyal (1994) EC Tax Law, Clarendon Press: Oxford.

French Senate (1999) [Rapporteur général: Philippe Marini. Direction: Jacques La Cacheux, Observatoire français des conjonctures économiques (OFCE)] La concurrence fiscale en Europe : une contribution au débat, Rapport d'information 483 (98-99).

Genschel, P. and T. Plumper (1997) 'Regulatory competition and international cooperation', Journal of European Public Policy, 4(4): 626-42.

Gordon, R.H. and L. Bovenberg (1996) 'Why is capital so immobile internationally? Possible explanations and implications for capital income taxation', American Economic Review, 86(5): 1057-75.

House of Lords (1999) Taxes in the EU: Can Cooperation and Coordination Coexist? London: The Stationery Office.

Jordan, A., Wurzel, R., Zito,A. and L. Brückner (2001) European govenance and the transfer of 'new' environmental policy instruments in the European Union, Paper prepared for the ECSA conference, Madison, Wisconsin, 30 May - 2 June 2001.

Keen, M. (1999) 'Aspects of harmful tax competition' in Bordignon, M. and D. da Empoli (Eds.) Concorrenza fiscale in un'economia internazionale integrata, Franco Angeli: Milan, 51-63.

McLure, C.E. and G.R. Zodrow (1994) 'The study and practice of income tax policy' in in J.M. Quigley and E. Smolensky (eds.) Modern Public Finance, MIT Press: Cambridge (MA), 165-209.

OECD (1998) Harmful Tax Competition. An Emerging Global Issue, Paris: OECD Publications.

OECD (2000) Towards Global Tax Co-operation: Progress in Identifying and Eliminating Harmful Tax Practices Paris: OECD.

Radaelli, C.M. (1997) The Politics of Corporate Taxation in the European Union. Knowledge and International Policy Agendas, Routledge: London.

Radaelli, C.M. (2000) 'Policy transfer in the European Union', Governance, 13(1) January: 25-43.

Steinmo, S. (1993) Taxation and Democracy, Yale University Press: New Haven.

Wincott, D. (2001) Beyond EU social regulation? Luxembourg, Lisbon, and new social policy instruments, Paper prepared for the ECSA conference, Madison, Wisconsin, 30 May - 2 June 2001.

BOX 1 - HOW THE CODE OF CONDUCT DEFINES HARMFUL TAX COMPETITION

1. Without prejudice to the respective spheres of competence of the Member States and the Community, this code of conduct, which covers business taxation, concerns those measures which affect, or may affect, in a significant way the location of business activity in the Community. Business activity in this respect also includes all activities carried out within a group of companies. The tax measures covered by the code include both laws or regulations and administrative practices.
2. Within the scope specified in Paragraph A, tax measures which provide for a significantly lower effective level of taxation, including zero taxation, than those levels which generally apply in the Member State in question are to be regarded as potentially harmful and therefore covered by this code. Such a level of taxation may operate by virtue of the nominal tax rate, the tax base or any other relevant factor.

Source: Official Journal C 2, 6.1.1998