; Contents
Learning Center
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>



  • pg 1

Peter Holmes and Paul Seabright

This study is about what the European Union can reasonably hope to achieve in
the field of industrial policy. Much has been written about what the state can
and should do in a modern industrial economy to influence the outcome of
competition between producers of goods and services, but although we shall
make reference to a small part of this vast literature it is not the main focus of
our concern. The European Union does not and could not have carte blanche to
fashion an ideal industrial policy out of nothing and according to the latest
wisdom. It is constrained by its constitution, its framework of law and the
operation of its institutions, as well as by various political limits to the likely
actions and decisions of its Member States. We shall explore the way in which
these constraints alter what is possible in the domain of industrial policy. In
many respects they make an EU policy less coherent and effective than it would
be in their absence. But the strategic perspective of game theory in recent years
has warned us to be wary of the conclusion that constraints on one's actions
must necessarily be harmful. Pre-commitment by the state may be valuable in
influencing the actions of other agents in the economy. We shall therefore also
explore ways in which the limitations on industrial policy in the EU, although in
some cases taking their present form inadvertently and as a by-product of a
complex and myopic political evolution, may even enhance its effectiveness in
certain other respects.
   Our task is therefore twofold: first and primarily, an exercise in positive
political economy, to explain how the various EU policies that affect industry
arise out of comprehensible actions by the main actors in EU policy making,
who face constraints inherited from the past. Secondly, a normative exercise, to
assess the extent to which these policies fall short of what might be achieved by
policies it would be reasonable to expect from the EU at its current stage of
    In some respects these two tasks will appear to be in tension, in that we shall
point to a large number of confusions and inconsistencies to which EU policies
currently give rise, while remaining (in the eyes of some readers) surprisingly
disinclined to damn the whole ensemble as a mess. To understand our argument
it is important to appreciate the extent to which the EU’s current stage of
evolution constrains the degree of coherence in industrial intervention that it is
reasonable to expect. Even at the level of objectives, it is inadequate and
potentially misleading to consider only aggregate welfare considerations, as
though the EU had attained the political maturity that would enable
compromises to be reached between economic interest even to the imperfect
degree this is possible in a modern industrial state. For example, if the EU had
all the attributes of a conventional state – even one as decentralized as the U.S. –
it would be reasonable to pose the question: what policy towards industry should
be adopted in order to maximize economic growth of the area as a whole (or
some comparable objective function, such as sustainable growth)?
    The EU is different in many ways from the U.S. It cannot be conceived as an
actor with a single objective function. Even in the U.S. it is often unwise to
abstract from the differences between Congress and the Administration, but in
the EU analogous differences are crucial. The EU is composed of multiple
principals and actors whose aims cannot be seen as conflated into a singe
compromise. There are major divisions of power and interest between EU and
national agencies and between different actors at the EU level.
    The nature of constraints put on the actions of decision-makers is critical to
the outcomes chosen. The EU is a rule-based system, with something in
common with the U.S. in that respect, and is much more so than the individual
Member States that make it up. This means that policies are legitimated in the
eyes of those they affect by the process that generates them. If there was ever a
time in which the Commission was able to see itself as a technocratic body
proposing technically correct decisions according to criteria that are independent
of the interests of the EU’s component bodies, this has now been largely
superseded by a decision making process in which the outcomes of actions are
known to be hard to predict. The ultimate arbiter, the Court of Justice (ECJ),
will judge decisions not by comparing the Commission or the Council’s view of
the facts with its own but principally by asking if a decision has been taken
according to the rules of the game, including the implicit ones.
    The decision in any one case will depend therefore on the prescribed
procedures for that category of action. This is particularly important for
industrial policy. As we shall see there is no such thing as industrial policy per
se, but a range of policies affecting industry. The nature of policies adopted will
depend on the policy process the decision goes through, such as whether we are
dealing with legislation that has to be adopted under unanimity or one of the
forms of majority voting.
   Certain kinds of policy are legally excluded under normal decision rules
(Article 236 of the Rome Treaty allows the Council of Ministers acting
unanimously to take any measures it wishes to further the aims of the Treaty).
Most subsidy policies are therefore ruled out. As we shall see, the system
requires that many objectives that could be dealt with by industrial policy are in
fact dealt with by the exercise of discretion in trade and competition policy, each
of which has different rules.
   One further feature of the EU system that is worth noting is that the rules and
procedures, though seemingly arbitrary at times and capable of generating
apparently irrational solutions, have all been chosen by the Member States
relatively recently. It is not an arbitrary set of rules externally imposed on
Member States. This framework of rules can be interpreted precisely as a set of
self-chosen constraints designed by the Member States to pre-commit
themselves to reciprocal market opening and the avoidance of mutually self-
defeating intervention in the market.
   The main constraints on EU policy we consider can be summarized under the
following headings:

     fiscal (the EU has extremely limited powers of taxation, and the
      overwhelming proportion of its expenditure is pre-committed to
      agriculture and the Social Funds)1;

     legal (actions by the EU institutions are subject to review by the Court of
      Justice and tend in consequence to be determined as much by procedural
      as by substantive criteria);

     constitutional (the Council and the Commission control different
      components of policy, and the powers of the EU to undertake industrial
      policy in the traditional sense of the term are limited);

     political (voting rules in the Council and Commission determine
      possibilities for trading agreements in certain areas of policy against
      agreements elsewhere; current practice and expectations likewise
      determine what agreements Member States are likely to endorse).

1   The European Coal and Steel Community is an exception to this, in that it has powers to
    impose duties that can then be spent on programs in these sectors.
We shall conclude that these constraints make for an industrial policy that is on
balance less interventionist, and less coordinated in its interventions, than if one
were to compare with a hypothetical case of the EU having a single government.
There are more hurdles that have to be cleared before a policy can be accepted.
This has the effect, in the long run, of diminishing the influence of the EU on its
own economy both for good and for ill. In the short run this phenomenon
creates a form of conservatism, so that policies once in place are not easy to
   In arguing for this conclusion it will be important to state clearly both what
we define industrial policy to be and the nature of the problems with which
industrial policy in the EU needs, and is perceived to need, to respond. For the
purposes of this study we shall use the term “industrial policy” fairly broadly, to
cover all policies decided at the EU level that primarily affect the industrial
sector of the economy. The term encompasses competition policy, trade policy,
technology policy and those regulatory policies that are not directly linked to
competition - in short, all policies that influence industrial outcomes, whether or
not they fall under the traditional rubric of “industrial policy” as it is familiarly
known. It is time to consider who are the agents in such a situation, what are
their goals, the precise strategies that are available to them and the constraints
they face.

Actors, Goals, Strategies, and Constraints

So who are the actors in industrial policy at the EU level? When we speak of the
EU taking actions, to whom are we referring and what are the presumed aims of
action by this agent? In effect the actions we describe in this study are ones that
fall within the mandate of the Commission or the Council. When analyzing the
motives of these agents from a positive point of view, we shall need to bear in
mind that in the Council the agents are Member States, and the actions of the
Council therefore reflect the interests of the Member States (governments) as
these are mediated by the rules of voting and the habits of political bargaining.
The Commission should not be thought of as representing a single “European”
interest group, since differences of perspective between its individual
Directorates-General (DG) are both well known and important; however, it does
not represent a compromise of national interests in quite the same way as the
Council. When we analyze industrial policy from a normative perspective, we
shall in contrast think in terms of benefits to European citizens as a whole in
some sense. Even if we recognize as we must that the agents of industrial policy
are not the citizens themselves but delegated political representatives and civil
servants acting on their behalf, we can still ask to what extent their actions are
conducive to the interests of the citizens they represent.
    However, there is one important difference between a normative analysis of
industrial policy in the European Union and one undertaken in an ordinary
nation-state. Economists are used to the idea of using the enhancement of
economic efficiency as the dominant criterion for evaluating industrial policy.
This is not because efficiency is the only criterion relevant for economic policy
in general but because of the presumption (given theoretical backing by the
Diamond-Mirrlees principle of optimum taxation) that distributive questions
should normally be considered the province of the tax and benefit system, and
industrial policy should aim at removing obstacles to productive efficiency. In a
political institution like the EU which enjoys very limited powers of fiscal
redistribution, such a presumption is no longer warranted. Efficiency therefore,
though an important - perhaps the most important - criterion for evaluating
industrial policy, cannot be the only one; other more explicitly distributive
criteria (such as the EU emphasis on “cohesion”) both can and should play a
So what might European industrial policy be reasonably thought to be trying to
achieve? What efficiency or distributive benefits could it realistically hope to
produce? This is not a question about the benefits of an industrial policy tout
court; it is about the benefits of a specifically European-level approach to the
involvement of the state in industry, rather than one left to the initiative of
individual Member States (whose own continuing involvement can be more or
less taken for granted).
   The alternative to involvement by the European Union in the regulation of
industry is not and has never been laissez-faire: it is the involvement of the
governments of its Member States, in competition with each other or engaged in
ad hoc intergovernmental co-operation, and restrained only by the rules of the
world trading system. Recognition of this fact was integral to the philosophy of
the Single Market program launched by the Member States of the European
Community in the mid 1980's. Although this was not explicit in the Cockfield
White Paper, Emerson et al. (1988) make it clear that the program was based on
an acceptance of the view that there were important economic and other gains to
be reaped from an increase in the intensity of competition between firms in the
Community. There was an important area of exception to this general principle
insofar as co-operation between firms in certain knowledge-intensive fields was
to be encouraged, but in general more rather than less competition was
acknowledged to be desirable. At the same time it was no less explicitly
accepted that, in order to realize the gains, Member States should engage in less
competition and more co-operation among themselves – whether in the field of
standard setting, competition policy or public intervention in the economy more
generally. Again there were areas of exception (notably those highlighted in the
debate over subsidiarity) but broadly speaking collaboration rather than
competition was the aim for governments, in sharp distinction to the aim for
   These two prescriptions were intimately related to each other: a degree of
collaboration between governments was viewed as a necessary condition for
bringing about greater competition between firms, since in the absence of co-
operation governments would tend to shield their own firms from competition.
Indeed, any kinds of market failure – not just inadequate competition – were
thought to require a co-operative solution; governments pursuing independent
objectives would inevitably bring about an outcome that was worse for them all.
   Why was it thought that firms and governments needed different
prescriptions in this way? Was it because firms and governments were thought
to be radically different kinds of creature, so that the competition, which was
good for one, would be harmful to the other? Or was it simply an empirical
judgement that competition had gone far enough for governments, but hardly far
enough for firms? In all the massive literature on European integration it is hard
to discern a clear answer to this question. The consequence is not just
intellectual confusion but sometimes also severe confusion in policy. EU policy
towards state aid, for example, has had to develop a coherent response to the
many actions that member state governments take in support of the competitive
efforts of their national firms, but it has had to do so in the absence of any clear
idea of what differentiates actions taken by governments in relation to firms,
from similar actions taken by firms in relation to each other. If a private Spanish
bank makes a loan to a Spanish firm, it has no need to notify the European
Commission, let alone submit to detailed inquiries about whether the terms on
which it does so are different from the terms on which it has made a loan to
some other firm of different nationality. If one of Spain’s regional governments
does the same thing its actions may very well be illegal under European law. In
assessing the acceptability of such intervention the Commission must decide
whether it distorts competition, but it must do so without any clear theory of
why and when governments’ involvement in the competitive process is more to
be feared than that of firms.
   Different intellectual traditions have viewed state involvement in the process
of competition between firms in strikingly different ways. First there are
traditions that view with grave distrust state intervention that favors particular
firms (whether directly in the form of subsidies, or indirectly through trade
policy or a favorable application of regulatory policy). In the language of
traditional anti-trust this has been seen as preventing a “level playing field,” and
therefore intrinsically unfair as well as probably inefficient. In the perspective
of micro-economic trade theory the inefficiency has been stressed more than the
unfairness. In recent years, it is true, the concept of the level playing field has
been re-examined. Recent findings, for example, that price discrimination can
have beneficial welfare consequences, especially when it enables markets to be
served that would not be under uniform pricing have tended somewhat to prise
apart the fairness and efficiency objections. Nevertheless, hostility from both
trade theory and anti-trust analysis to discriminatory state intervention of any
kind remains strong. It has, if anything, been reinforced by the recent literature
on strategic trade policy, which while offering a case for intervention from the
point of view of individual countries, has stressed that national strategic trade
policy interventions are typically bad for global welfare, and has underlined the
information requirements for such policies to bring more benefits than costs.
   A quite different tradition is in public finance, where there has long been a
relatively benign view of the process whereby local or regional governments
compete in the provision of local public goods and services, as in the famous
Tiebout (1956) model. Here there are positive benefits to inter-governmental
competition, since regions and localities can differentiate the kinds of good or
service they provide to match them more closely to the differing preferences of
populations than could be achieved by any centralized policy. Such competition
can of course give rise to externalities between governments, but in the absence
of reasons to the contrary this tradition has generally presumed that competing
provision is desirable. So long as the provision consists of consumer goods and
services this presumption is not strictly inconsistent with that of the anti-trust
tradition. But in fact it is readily apparent that many of the goods and services
(in the broad sense of the term, including a regulatory environment) that are
provided by governments consist of intermediate inputs into the productive
activities of firms. As a result inconsistencies abound: why should it be
distortionary for a government to provide preferential credit to a particular firm,
but quite acceptable for it to provide high quality infrastructure in the form of
airport facilities or access roads?
   One possible reconciliation of these two perspectives might arise from
distinguishing generic from specific interventions. A generic intervention might
be defined as one that affects the strategies open to any firms that fulfil a set of
general criteria that are grounded in an account of the appropriate state response
to market failure. A specific intervention affects the strategies of particular
identified firms, either assisting or penalizing them in their efforts to compete
with others. It is tempting to think that the interventions to which the anti-trust
and trade theory perspectives are hostile are particular interventions, while those
to which the public finance perspective is receptive are generic ones.
Preferential credit is bad - according to this view - because it is particular,
whereas infrastructure is acceptable because it is generic: any firm in the
appropriate location may benefit from it.
   A little reflection, however, suggests that this neat resolution of the problem
is hardly satisfactory. Interventions that appear generic almost always benefit
certain firms more than others, and when these firms are in competition with
each other the intervention “tilts the playing field” as surely as if the
intervention were particular. The fundamental question remains whether the
interventions concerned are anti-competitive or not, and this is a matter to be
resolved by substantial economic argument rather than taxonomy.
   Many activities of governments consist not of the provision of goods and
services directly to citizen-consumers, but of intermediate inputs to the
productive activities of firms; this means that a coherent account of when such
activities may be anti-competitive (in the sense of having negative welfare
effects overall including spillover effects on other producers) arises as a natural
corollary to a more general theory of the competitive consequences of vertical
relations between upstream and downstream production. Can we understand
under what general circumstances the actions of an upstream entity (be it a firm
or a government) in its relations with a downstream firm likely to be anti-
competitive? And secondly, are there particular reasons to think that these
actions are more likely to meet the criteria for being anti-competitive when the
upstream entities are governments rather than other firms?
   To answer this second question it is necessary to explore in greater detail the
ways in which it is reasonable to suppose that governments in their actions to
provide goods and services (including regulatory services) behave in a
systematically different way from firms. Four ways in particular suggest
1. The objectives of governments may differ systematically from those of firms,
    and may be less oriented towards economic efficiency. In particular, the
    political system will fail to deliver public goods and services in a
    productively efficient manner. Nevertheless, two caveats should be
    entered. First, there is some dispute about the extent to which even firms
    achieve productive efficiency, especially when they enjoy market power.
    Secondly, even if there is a systematic difference between firms and
    governments in this respect, it is still important to know why and how it
    makes government actions intrinsically more anti-competitive than those of
    firms, and why there is any reason to think that the solution to the
    inefficiencies of competition between national governments should be the
    intervention of yet another government, this time at a supra-national level.
2. Because governments occupy different territories, costs linked to distance
    (and therefore to market access) will create intrinsic market power when
    they compete with each other. Two private banks may lend to firms that
    compete very closely, but two governments competing through rival
    approaches to regulation, or through the provision of state aids are likely
    to face more muted competition.
3. Governments may (through their control of the tax system) have deeper
    pockets than those of firms. This fact may tempt them to engage in
    strategic manipulation of entry and exit decisions in particular markets by
    exploiting financial constraints faced by competitor firms.
4. Finally, the upstream goods and services provided by governments may be
    much more often naturally monopolistic than those provided by firms
    (though this has undoubtedly become less true since privatization of
    naturally monopolistic industries began in the 1980’s in parts of the EU).
    In particular, many of the natural activities of government are subject to
    significant scale economies, which means that the marginal tax revenue
    yielded by those most mobile factors of production usually significantly
    exceeds the marginal cost of providing them with access to public goods
    and services. This in turn has the implication that competition between
    governments to attract mobile factors can significantly distort not only tax
    systems but also regulatory activity, as the literature on fiscal competition
    has recognized (see Oates and Schwab, 1988).
Accordingly, therefore, an account of industrial policy in a political system such
as the European Union has to include not only the actions taken by the Union to
influence directly the activities of firms, but also the actions it takes to influence
the activities of Member States that in turn influence the activities of firms. This
means that the array of instruments we must include in our analysis of industrial
policy is remarkably wide. And when we consider an instrument employed to
bring about some particular desired outcome (improved environmental
protection, for instance), we must ask not simply whether such instruments are
appropriate for the desired end but more specifically whether such instruments
are more likely to be deployed effectively to that end by the EU’s own
institutions than by the Member States themselves.
   To summarize, EU industrial policy is best understood and evaluated not as
seeking to improve upon laissez-faire competition, but as moderating the
outcome of a large range of potential economic interventions that would
otherwise be undertaken by Member States. It is the fact that these
interventions, if uncoordinated, have the potential to be positively damaging to
the interests of these Member States, that makes the task of policy-making at EU
level so challenging and so difficult.

Strategies and Constraints
What then are the precise instruments available to the EU, and the constraints
upon their utilization? Bourgeois and Demaret (1995) provide a valuable
discussion of this question. In Part One (“Principles”) of the EC Treaty, Articles
3(b), 3(g) and 3(l) refer respectively to “a common commercial policy,” “a
system ensuring that competition in the internal market is not distorted,” and
“the strengthening of the competitiveness of Community industry.” This last,
which is the closest reference to industrial policy as such, is given substance by
a new provision introduced in the Maastricht Treaty, namely Article 130.
Unanimity in the Council is required to make use of this provision, a constraint
that significantly limits both the likelihood and the nature of any policy
interventions undertaken. The text refers to co-ordination of action by Member
States rather than collective action at an EU level and it also rules against
measures which might “distort competition.” Bourgeois and Demaret show
convincingly that from a legal and constitutional point of view, such measures
that the Council may introduce with this purpose in mind must be clearly
consistent with competition policy, which therefore emerges as a dominant and
regulative principle determining the limits of application of the other two
(arguably with less restrictive force in the field of trade policy). Remarkably,
they point out that the more specific terms of Article 130 if anything appear to
narrow the scope for industrial interventions which might otherwise have been
justifiable with unanimity under the catch all Article 235.
    There is thus a powerful institutional safeguard against interventionist
policies so long as even a single member state willing to use a veto is opposed.
It is very difficult to speak of an industrial policy per se in the sense that this has
been used in the past, i.e. the support for particular firms or industries as such.
France argued strongly for inclusion of an industrial component in the
Maastricht Treaty, and even under conservative governments there has been a
consistent voice from Paris urging a re-balancing of what is seen as the
dominance of “competition” as opposed to “industrial policy objectives.” The
terms in which industrial policy is defined in the Maastricht Treaty permit very
little traditional interventionism: the language of the Treaty, which echoes the
White Paper of 1990, emphasizes the need for non sector specific policies and
the language indicates that within any industrial policy the concepts which
govern competition policy must also be held to be paramount. Moreover by
giving the EU competence in industrial policy, political if not exactly legal
ammunition is provided for those who wish to see individual national initiatives
   The Maastricht Treaty industry chapter therefore is based on four principles
(Marchipont, 1994, p. 21) which are reproduced in the Commission’s 1994
White Paper (European Commission, (94) 319):
1.   a policy aimed at competitiveness is envisaged, not a traditional French-
     style industrial policy per se;
2.   the policy must not be sector specific;
3.   co-ordination and concentration rather than a common policy is envisaged;
4.   the policy is to be based on competitive markets.
Paradoxically, as we shall see, there is much more room for industrial policy by
the selective discretion by which competition policy and trade policy are applied
than there is under the heading of industrial policy per se. It is arguable that the
relative scale of national and EU level interventions in the field of technology
policy, for example, consign this to a very secondary role. The Commission’s
own figures indicate that Community Framework program funding for R&D and
Eureka funding together account for only about 5% of total public support for
R&D inside the EU.
   The way the merger regulation is applied, the rules on R&D joint ventures
and the actual application to cases of this and Article 85(3) powers in general,
the use of regulatory powers in telecommunications, sectorally differentiated
application of state aids policy: these are all ways in which competition policy is
a powerful framework for industrial policy. At the same time trade policy –
especially in the form of anti-dumping policy – has created an implicit industrial
policy for certain sectors. What is common to all these areas is that they are
ones in which the Commission has discretion apply rules governing the behavior
of firms or of governments without having to seek prior consensus between
Member States in Council.
   If the EU were a unitary state, one might expect to see industry, trade and
competition policies targeted fairly precisely according to the nature of the
market failures each instrument is best designed to correct, (Gual, 1995).
     competition policy: eliminating distortions due to market power or
      barriers to entry exercised or created inside the EU by firms or
     trade policy: eliminating distortions coming from the exercise of market
      power by firms or governments outside the EU;
     industrial policy: internalising externalities where private or national
      transactions costs of doing so are too high; broader regulatory and
      harmonisation issues where asymmetries of information or coordination
      problems are at issue.
In fact what we observe – as we discuss below – is a compromise allocation of
instruments to targets determined by
1. the allocation of rule-making powers to institutions (specifically to the
     Council or the Commission);
2. the procedural rules determining how such decisions may be taken;
3. the scope for judicial review of the decisions taken.
One might expect that the powers that have been ceded to the Commission, or
subject to relatively undemanding voting requirements in Council, are those for
which the exercise of discretion can be relatively tightly constrained by judicial
review (specifically the application of competition rules). Areas in which much
more discretionary judgement is required (such as the evaluation of
externalities) have by and large been retained under the control of Member
States. Although there is some element of truth in this view, it is tempered by
the fact that the constraints imposed by the possibility of judicial review are
always limited, not only by the scope of procedural criteria to determine
substantive judgements, but also by the difficulties faced by affected parties in

Table 1: Needs a title
                        MEMBER STATES                      COMMUNITY
                   Individually   Ad hoc               Council   Commission
Competition           U.K.                            Regulations         MAIN
                   Monopolies &                        approved;          Bloc
                     Merger                            Advisory       exemptions;
                   Commission                         committee       85(1),85(3)
                      BKA                                                chooses
Trade             Residual powers      BLEU          Legal power,      delegated
                                                       majority       powers, e.g.
                                                     but consensus   anti-dumping;
                                                        elusive        negotiates
Industry and          MAIN          Eureka, Airbus    Unanimity       Legislative
regulation           ACTORS                              under         initiation;
                                                      Maastricht;    Coordination
                                                     QMV for 100a
making use of the judicial review process. This is particularly significant in
respect of trade policy, as will be seen below.
   Table 1 is a simple matrix representation of the main respects in which EU
action can be taken in each of the three fields under consideration. We discuss
more detailed aspects in section 3.

Fields of Intervention

Competition policy can be said to have a dominant role among the instruments
of industrial policy in the EU in at least two senses. First, as we have seen the
language of the Treaty requires both trade policy and policy towards
competitiveness to be compatible with competition policy. Secondly, the fact
that the Commission has the power to take decisions in this field removes one of
the intrinsic obstacles to coherent policy making that characterize the other
domains. Although even the Commission is to some extent a coalition of
interests, it has the capacity to behave more like a unitary agent in this field than
in others.
    So much for theory, at least: in practice the Community’s competition policy
has been influenced to different degrees by the extent to which it can reconcile
tensions between different aims of policy, by the degree to which it can resist
the lobbying pressures of Member States and sometimes of firms themselves,
and by the constraints of the jurisprudence according to which it operates.
Roughly speaking the overall coherence of policy has been greatest in the field
of merger control, least in the control of state aids and the liberalization of the
so-called “excluded sectors,” with the remainder of anti-trust policy (control of
abuse of dominance and regulation of agreements between firms) occupying
something of an intermediate position. As we have emphasized, this should not
be taken to imply that Community action is most “successful” in the areas where
it is also most coherent, since the value added of Community involvement
should be understood in terms of the policies that would otherwise have been
undertaken by Member States on their own.
    Merger policy is perhaps the simplest to evaluate, and is also highly revealing
of the aims that Community policy has been established to pursue. When the
Council passed the Merger Regulation in late 1989 this represented the final
achievement of an effort to establish a Community merger policy that had taken
over a decade and a half to come to fruition. Although Commission initiatives
to review mergers during the previous two or three years may have accelerated
this process (as when the Commission imposed conditions on the British
Airways take-over of British Caledonian in 1986), the dominant explanation for
the development was undoubtedly a growing political consensus among
Member States that the objectives of the Single Market Program required
Community-level policing of merger activity. This was partly because of a
general belief (as expressed in Emerson et al., 1988) that the objectives of the
SEM risked being undermined if individual firms were allowed to erect market
entry barriers to replace the remaining trade barriers which it was the purpose of
the SEM program to remove. It was also due to the fact that capital market
liberalization made it likely that corporate restructuring via mergers and
acquisitions would increasingly supplement competition purely in product
markets: in that sense a policy specifically directed towards mergers was more
necessary than it had been before.
    The actual terms of the Regulation reveal that, in return for persuading the
more reluctant Member States to cede sovereignty over an important area of
policy, the Council found it necessary to circumscribe quite tightly the criteria
according to which this sovereignty could be exercised. First, mergers that
create or strengthen a dominant position are to be forbidden, and there is no
formal scope for offsetting any efficiency gains of a merger against an adverse
judgement on competition grounds. Secondly, in determining the criteria for
allocating jurisdiction between the Commission and the competition authorities
of Member States, the Regulation adopted a formula based both on the absolute
size of the firms involved and on the magnitude of cross-border spillovers - a
formula that represents in many ways the most systematic and impressive
application yet of the principles of subsidiarity to the allocation of power within
the EU. In practice, of course, the boundaries of jurisdiction were bound to be
controversial, and there have been a number of disputes over particular cases.
More importantly, there has been some reason to doubt whether the actual
application of merger policy has been as austerely tied to dominance criteria as
the text of the Regulation requires. The Commission’s record in evaluating
mergers has not been particularly restrictive, and Neven et al. (1993) have
suggested some reasons why the Commission remains vulnerable to pressures to
approve deals that have managed to appeal to Member States on other than
dominance grounds. They also suggest that, paradoxically, the validity of
arguments purporting to outweigh dominance might be more rigorously assessed
if there were explicit scope within the Regulation for them to be taken into
    The fact that it was the Single Market Program that expanded the activity of
the Commission in competition policy reminds us that the goals of competition
policy cannot simply be summed up as the promotion of economic efficiency in
the neo-classical economists’ sense of the term. Market integration is and has
always been a distinct and independent goal. In the first Report on Competition
Policy in 1971 the Commission states that “competition policy has to pursue
objectives specifically related to the creation and the good functioning of the
Common Market,” an aim that is restated in the 9th report in 1979 as one of the
major goals of competition policy. Nowhere is this clearer than in the
Commission’s policy towards vertical restraints, where block exemptions for
various kinds of exclusive and selective distribution agreement have refused to
countenance any provisions that block parallel imports between Member States;
indeed, the case law of the last few years suggests there is no surer way to bring
down the wrath of the Commission on a distribution agreement than to seek to
prevent cross-border trade (see Seabright, 1996). From an efficiency point of
view such a rigid emphasis is bizarre: selective and exclusive distribution
agreements are all about segmenting markets, and are usually justified in terms
of giving retailers various kinds of incentive to undertake complementary
investments. There is no intrinsic justification for allowing segmentation that
happens not to lie along national boundaries, while punishing segmentation that
does. The only way to rationalize such a policy is in terms of the value of
market integration independently of narrow efficiency concerns. Even here
there is some reason to doubt whether the policy achieves its own aims: ensuring
the possibility of parallel imports prevents price discrimination between
markets, and it is well known that in the absence of the ability to undertake price
discrimination certain markets may simply not be served. So it is ironic that
certain vertical restraints that might contribute to helping firms break into new
markets in other Member States are made harder by a policy that appeals to the
desideratum of market integration. But from the point of view of the present
argument what the case of vertical restraints illustrates is this: some of the
incoherence in competition policy arises not from the fragmentation of agency
between different institutions in the EU, but from the fact that the goals even a
relatively unitary agency such as the Commission may pursue are themselves
multiple and in tension.
   More generally, the application of Article 85 of the Treaty (which covers
both horizontal and vertical agreements between firms) illustrates two broader
points. First, the importance of jurisprudential constraints. As many
commentators have pointed out (see Bright, 1994, for example), owing to earlier
decisions of the ECJ - especially the much-cited Consten & Grundig decision
from the early 1960s - a great many agreements between firms that have no
adverse effect on competition nevertheless fall under the anathema of Article
85(1) and have to be retrospectively legitimated through the application of
Article 85(3). It is as though the fishermen of DG 4 catch in their nets vast
numbers of fish they subsequently return to the sea. The mechanism of block
exemptions moderates but does not adequately undo the adverse impact of this
procedure, particularly since block exemptions - with the “white list” and “black
list” of clauses - create a strong bias towards form-based rather than effects-
based evaluation of agreements. One result of this judicial inheritance is that the
growing activism of the Commission in competition matters has been perceived
quite differently in the Article 85 field from the field of mergers. The assertion
of a Community role in mergers has been broadly perceived as having a
deregulatory impact, as leading to less state intervention than would otherwise
take place; in the Article 85 sphere the business community has quite the
opposite perception.
    The second – and related – point illustrated by the case law under Article 85
is that the Commission enjoys quite wide discretion in allowing agreements that
have already been deemed anti-competitive. There is a sense in which industrial
policy implemented under the umbrella of competition policy is given wide rein
under Article 85 for a quite opposite reason from the one that explains its
appearance in the field of mergers. In merger policy industrial policy is ruled
out ab initio, so it has a tendency to creep in by the back door, unannounced and
unmonitored. Under Article 85 so many agreements have to be salvaged by an
appeal to industrial policy considerations that it is hard to know how
systematically the justificatory criteria are applied. Certainly, agreements that
make an appeal to technology or R&D considerations have a high chance of
being waved through by the Commission. This is an explicit part of policy: the
16th Report on Competition Policy in 1986 stated that “under the condition that
some level of effective competition is maintained, it [competition policy] allows
for agreements between firms when these are likely to promote technological or
economic progress, an in particular those that favor R&D or technology
transfers.” Furthermore, the language of Article 130 was actually prefigured in
Commission Competition Reports as early as 1987, where more competition is
said to promote “a greater competitiveness of EC firms both inside the EC, and
in third country markets;” a point reiterated more strongly in 1992 where it is
stated that “the Commission is willing to facilitate ...cooperative operations that
allow firms to adapt and improve their competitiveness in a global context.”
None of this is to imply that the Commission is wrong to act in this way, merely
that there is a risk of significant inconsistency between the application of
different instruments of policy, the more so as the distinction between certain
arrangements that are dealt with under Article 85 and those that fall under the
Merger Regulation can seem highly arcane (see Neven et al., 1993, pp. 84-90).
    Although pointing out inconsistencies is perhaps easy sport in this field, the
two areas of competition policy we have so far cited are ones in which
procedural criteria nevertheless visibly constrain the substance of policy. The
same cannot be said for policy toward market liberalization in the excluded
sectors, nor for policy towards state aids. Here the Commission faces much
more explicitly the political limitations on its power of maneuver with respect to
Member States, since it is usually the actions of Member States themselves that
are under the scrutiny of competition policy. In both of these domains the
actions in question involve the granting of rents to favored firms by the Member
States’ governments, in the former case by the protection of markets from
competition and in the latter by explicit subsidies funded through the tax system.
The two kinds of case differ only in that, when state aids are granted, the
economic distortion in consumption necessary to fund the rents is spread over
the entire tax base, whereas when specific markets are protected the distortion is
concentrated in the market in question. Conventional analysis suggests the
damage done by specific market protection is greater (at least in deadweight loss
terms) than that due to state aids, although it may be less transparent; however,
the additional damage is suffered directly by domestic citizens, so the case for a
specific Community-level involvement is no stronger than in the case of state
aids. The rationale for such involvement is essentially identical in the two cases:
firms whose activities in the market are underwritten by governments (whether
through subsidies or protection is immaterial) impose an externality on others,
by making it harder for more efficient but less publicly-favored firms to survive
and prosper. In addition to the externality argument there is also one based on
the commitment potential of a centralized policy: governments may find that the
presence of such a policy is the only credible way for them to resist domestic
lobbying by powerful firms (see Winters, 1988; Gatsios and Seabright, 1989).
   The goals of Community-level policy in these two fields have been much less
precisely spelled out than in the more conventional areas of anti-trust activity.
This is partly because, as we argued in the introduction, neither the traditions of
public policy nor of academia have been able to clarify exactly what an optimal
relation between firms and governments in a world of rival nation-states would
look like. The aims of EU policy, as expressed in the rhetoric of successive
Competition Policy Reports and of decisions and other public statements, have
been expressed in terms of the need for more competition and less state aid:
without knowing precisely where the optimum lies one may feel confident about
the direction we need to move from here. The lack of clarity over the goals of
policy may also owe something to the highly politicized nature of decisions, to
which we referred above: if you are doubtful about your ability to take
decisions that antagonize powerful interest groups it may be prudent not to spell
out in too much detail exactly what the goals of your policy are, and to rely
instead on your ability to construct a rationale after the event. However, the lack
of clarity has its costs, particularly where competition policy finds itself also
trying to take account of cohesion objectives. For example, Ashcroft and
Wishlade (1995), in their study of regional investment aids, argue that “The
revealed practice of the Commission suggests that an appropriate award level is
one that satisfies the regional cohesion objectives without damage to sectoral
competitiveness and economic efficiency within the EU... if evidence is found of
likely damage to competitiveness then it will be prepared to rule that an award
of state aid should not be made. But in practice this rarely seems to happen
because when certain elements of competitive damage are found the
Commission appears frequently to be able to find other more positive
competitive effects which are used to negate the initial finding of damage... the
process gives the impression of being largely ad hoc.”
    The comparative lack of coherence of liberalization and state aids policies in
comparison to those of traditional anti-trust should not, however, be taken to
imply that EU activity in these fields has been less desirable or beneficial than in
the others. Once again it is important to be realistic about the counterfactual
comparison: what would Member States do if entirely unconstrained by the EU?
It is quite conceivable that a completely uncoordinated set of anti-trust policies,
though less efficient and coherent than the current regime, would inflict
significantly less damage in the long run on the EU’s economies than a complete
lack of co-ordination in the liberalization and state aids spheres. Such a
comparison must inevitably be speculative, and should in any case not imply
that efforts to improve the coherence of EU policies can be relaxed.
Trade Policy
The Rome Treaty laid down that there should be a Common Commercial Policy
(CCP) by 1969 (Article 113). A Customs Union was created in the sense that
tariffs were aligned but national non-tariff barriers formally existed until 1993
and some informal measures may still survive.
    In trade policy, discretion is much wider than in competition policy. The
ultimate decision taking body - the Council - is free to adopt any objectives it
wishes that it can reconcile with its obligations to GATT/WTO. Nevertheless,
there are still quite strict legal rules, such that the borderline between “policy”
and application of rules is fuzzy. Bourgeois and Demaret (1995) argue that the
principles of competition policy do set a legal limit on the extent to which trade
policy (or any other policy towards industry): so “competition may not be
eliminated in a substantial part of the Common Market” (p. 84). Despite this
trade policy can fill a gap that other components of industrial policy cannot do.
In both procedure and substance it is subject to fewer constraints than
competition policy and industrial policy in terms of Article 130. And yet it
would be wrong to suppose that either the Commission or the Council can run in
a freely mercantilist direction.
    The EU component of trade policy should be seen as replacing whatever
measures Member States would have imposed in its absence. Critics of the EU
have invariably observed that policy measures are generally in a protectionist
direction. This is potentially misleading however: the EU trade regime is
undoubtedly more protectionist than welfare economics would recommend, but
it is debatable whether it is clearly more protectionist than what would otherwise
be adopted by the Member States.
    Articles 9 and 10 of the Treaty state that duty paid third country goods must
be given free circulation throughout the Common Market. But Article 115 of
the Rome Treaty allowed the Commission to authorize measures taken by
Member States to control the free flow of third country goods in exceptional
cases during the transition to the full common market (scheduled for 1969).
Article 115 continued to be used after 1969 and was an indirect target of the
White Paper.
    Non-tariff measures were not and are still not finally harmonized on an EU
wide basis. There were two types of gap in the CCP. In some areas the
Community acknowledged that there were formal derogations from the principle
of a CCP. The most obvious examples were the Multi-Fibre Agreement where
national quotas were agreed at GATT, a number of individual quotas
grandfathered into the Rome Treaty (notably the Italian quota on Japanese cars)
and a number of national restrictions on imports from state trading nations. For
these measures, the Commission was prepared until 1993 allow the use of article
115. At the same time there were a number of unofficial national measures,
such as the U.K. and French restrictions on Japanese cars (the former an
industry-to-industry VER which violated almost every relevant rule of GATT
and the Rome Treaty, the latter an administrative measure also almost certainly
illegal). In the case of unofficial measures the Commission would not allow
Article 115 to be used.
    An unspoken aim of the 1992 plan was to generate a truly common
commercial policy. The 1995 Cockfield White Paper acknowledged that the
removal of all internal frontier controls would make traditional customs checks
unusable as a means of allowing segmentation of markets where Member States
had quota regimes of varying degrees of tightness. It fell short of pointing out
the obvious conclusion that after 1993 there had to be a common external trade
regime if the internal market was to make sense, and it avoided all hints that a
collective decision would have to be taken on the degree of liberalism involved.
    The Commission aimed to end the use of Article 115 after 1993 and in fact
succeeded in refusing all applications for 115 measures (except a temporary
application for bananas) after 1993. However Article 115 was revised rather
than deleted by Maastricht: the new version requires pre-authorization of any
national controls on intra-community trade in third country goods, authorization
which the Commission is determined not to give.
    The trade policy of the EU is a striking example of the effects of “federalism”
on decision taking: Article 113 of the Rome Treaty gives ultimate decision
making power by qualified majority vote (QMV) to the Council of Ministers,
but the Commission’s right to propose together with additional powers of
implementation transferred by delegation have led to a gradual transfer of
authority to the Commission. In principle the Council has the final say on trade
measures by QMV, and on anti-dumping by simple majority of states. Formal
use of measures – like any change in policy – is restricted by the difficulty of
getting through all the hurdles involved. In the case of anti-dumping policy, the
firms involved must establish legal standing for a complaint, which is a modest
but not a trivial matter. They must then persuade trade policy officials that all
conditions for action are satisfied and the officials must not believe there is a
strong overall Community interest against action.                Then the Trade
Commissioner must approve; then the full Commission must agree to propose
the measures to the Council of Ministers, which must approve. Finally the
Court must agree that procedural rules have not been violated. What makes the
system appear protectionist is that we only see evidence of trade policy action
once a proposal has passed most of the earlier filters. The Commission will put
off complaints that it cannot support or persuade the Council of Ministers to
    In terms of incentives to intervene, the Commission has a clear mandate to
take overall European Union welfare into account, and risks being subjected to
criticism if it seen to back particular interests that do not command general
political support. Meanwhile within the Council of Ministers a new tendency is
emerging. It is harder to persuade Member States to vote for protection for
industries not present in their own territory. Usually evidence for this is not
visible, but we have recently seen examples of measures being proposed by the
Commission turned down by the Council, for no recorded reason other than that
the MS’s did not think it in their interest to offer protection.
    The complexity of this can be illustrated by looking at the way anti-dumping
policy works. Here we have a system in which an essentially economic
instrument is used in a political manner on the basis of tight legal rules.
    Although intended to curb “unfair competition,” anti-dumping differs from
competition policy in procedure and substance. Despite the superficial
resemblance to the criteria for predatory pricing, the legal definition of dumping
since the CATT 1947 covers all cases where export prices are lower than home
prices. The CATT allows signatories to impose anti-dumping duties whenever
this occurs and if it can be shown that injury to domestic producers has been
caused or is threatened.
    The legal rationale for this in cases where foreign producers are not in a
position to create and abuse a position of monopolistic dominance may be
termed the “market asymmetry principle:” If non-EU firms enjoy some form of
sheltered market condition at home which EU firms cannot profit from and
which allows the foreign firms to injure EU firms without reciprocity being
possible, then we can apply anti-dumping duties under WTO and EU law.
Economists like to point out that it may not always be in our interests to do so,
and that if there is an economic rationale for anti-dumping it can only lie in a
sub-set of the cases in retaliation is legally sanctioned. The kind of behavior
referred to above includes much of what would be covered by strategic trade
policy; this is in many cases likely to be globally inefficient, although it is not
always evident that retaliation may make things better. There may, it is true, be
circumstances in which a credible threat of anti-dumping action can deter
strategic policy and behavior, even when actual use of the instrument ex post
may be unwise (see Dornbusch (1990)).                Realistically, though, such
circumstances will be rare.
    However, the EU is unusual among jurisdictions in including in its
legislation, most forcefully since 1994, a “Community Interest” test, in addition
to the need to establish dumping injury and causation. The quasi-federal nature
of the EU system reveals itself at its most complex here. “Community Interest”
is ultimately a political concept and it would be impossible to define it in any
more than loose terms (as Article 21 of the anti-dumping regulation does –
listing the factors that must be weighed up without establishing exact priorities).
In principle it is the only element in an anti-dumping decision which allows for
political judgement to enter. Up to the time of writing virtually no explicit use
has been made of this provision to refuse measures (though in the case of Gum
Rosin, it is evident that an overwhelming majority of Member States told the
Commission that since they did not produce the product it could not be in their
interest to protect it).
    Responsibility for defining Community Interest is split between the Member
States, the Commission and ultimately the Court. The decision process is one of
a series of filters. The Commission receives an anti-dumping complaint, after an
initial investigation and hearing the advice of the Member States, the
Commission acting alone is empowered to adopt Provisional anti-dumping
duties. It publishes its decision, which may be legally challenged. Then after
the full investigation and in the light of any new arguments that may be put to it
by parties with standing to do so, the Commission proposes definitive duties to
the Council of Ministers. The Council votes by simple majority to accept or
reject these (prior to 1994 it was by QMV). Interested parties - who since 1994
include consumer groups as well as users - can challenge any decision in the
Court of First Instance if they can show a flaw in the reasoning of the Council or
the Commission. There is thus a succession of filters to pass, the anti-dumping
unit, the Trade Commissioner (who may well take a different view from his
staff), the full Commission, the Council, and finally the Courts (CFI and
eventually ECJ). Although the outcome of this procedure is frequently very
protectionist, the policy underlying it is much less so: we see at the end of the
decision making process only those measures that have been adopted and not the
cases where action has been averted. The Commission is usually reluctant to
propose measures if it does not think a majority can be found in the Council, so
is inclined to use informal means to suppress complaints it thinks stand no
chance of success. Given the degree of criticism of EU anti-dumping it is worth
recalling that according to the Commission's Annual Report on Anti-dumping
and Anti-Subsidy Activities barely 1% of imports are directly affected, although
of course the threat of anti-dumping affects a broader class.
   The pattern of anti-dumping actions shows a number of distinct clusters of
product, notably electronics, chemicals, labor-intensive manufactures and a
number of raw material products. Commission officials argue that this pattern is
the result of the spontaneous action of firms. The EU electronics industry has
chosen to file regular anti-dumping complaints and the Commission has
responded; the Commission has not chosen this sector as the focus for an
explicit industrial policy. Formally speaking the Commission cannot move on
trade policy without either a complaint from business or a mandate from the
Council; existing legislation offers only a small margin of maneuver.
   The Commission cannot easily plan to use trade policy pro-actively to
support industrial policy aims. One senior official has commented privately
“The only policy to use anti-dumping against Asian electronics firms is that of
the European companies themselves.” Nevertheless this gave the Commission
in the mid-1980’s an opportunity to put in place something resembling a
consistent policy, the consequences of which have not necessarily been what
was originally, intended. The one absolutely clear consequence of anti-dumping
measures has been a flood of inward investment by Japanese and Korean
consumer electronics firms, and also semi-conductor firms.
   The EU producers of consumer electronics show little sign of benefit
(Philips) or are quitting (Nokia, Thomson). On the other hand, since prices
continue to fall, the costs to consumers are hard to measure in the absence of a
well-specified counterfactual. The main impact on the EU economy is therefore
the DFI itself. Is this good or bad? In a simple neo-classical model tariff
jumping investment is likely to lead to resources misallocation, but these
conclusions do not necessarily carry over to more realistic models that
incorporate endogenous and firm specific comparative advantages, rents shared
with workers and spillover effects.
   The Commission’s view of these developments is paradoxical: it denies that
it has deliberately sought to use anti-dumping to attract direct foreign
investment, but argues that inward investment has been a beneficial
consequence. Officials argue that the aim is not to protect the complainant firms
but to simply curtail “unfair” competition. If the outcome of the actions is a
growth of foreign owned business in the EU that, they argue, is good.
   Many officials dislike the term “Industrial policy” and yet such arguments
have appeared in the reasoning behind anti-dumping decisions: the strategic
industrial case for sustaining EU production of consumer electronics, semi-
conductors and photocopiers has been made in this context. The photocopier
case illustrates the inconsistencies in the Commission's approach. In 1995 the
question arose whether or not to keep measures imposed in the 1980’s, since
when massive DFI had occurred, leaving the only major complainant Xerox (a
U.S. owned firm) with a very small share of EU production. The Commission
agreed that for the purpose of standing to make a complaint Xerox and two
smaller firms represented “Community industry,” but when it came to decide on
community interest the Commission argued that this should be seen as
encompassing all producers including the Japanese. The existence of duties
sustained their investment and consequently supported the existence of a large
scale component industry (indeed in the case of color televisions the
Commission argued that the promotion of a TV tube industry was the main
reason for wishing to retain EU production of TV sets). The problem with such
arguments is not that they are manifestly wrong, but that the reasoning is
opaque, since the instrument is being used for purposes wholly other than those
for which it was designed. The “Community Interest” is too often window-
dressing at the end. There are however signs that this is changing: pressures are
mounting for an explicit treatment of the issues that are now implicit and more
economic reasoning to be included in anti-dumping decisions, the logic of which
would then be subject to test in the courts.

Trade Policy in cars
Before 1993 several Member States had their own separate national rules
governing car imports and many authors attribute the marked discrepancies in
car prices across the EU to the resulting segmentation (for details see Holmes
and Smith 1995; also Bourgeois and Demaret, 1995). The Commission was
willing to authorize use of Article 115 only for the recognized Italian measures.
Much ink was wasted over what local content on cars was required before the
output of DFI plants could be given free circulation under EC law: Articles 9
and 10 make it clear that any duty paid Japanese car made anywhere in the
world has rights of free circulation.
   In 1991 the Commission agreed with MITI that all national measures in the
EU legal or not would be abolished in 1993 and instead MITI and the
Commission would negotiate six monthly on car imports. After the year 2000
normal WTO rules - including those governing anti-dumping - would apply to
the car industry.
   Meanwhile there would be six monthly negotiations between the
Commission and MITI over imports from Japan, with quotas being agreed for
each national market separately. Much further ink was then wasted over
whether transplant output was included in the deal. Different Member States
announced different interpretations. Transplant output cannot legally be
restricted under the Treaty, but of course the state of the overall EU market
including transplant output was at the back of everyone’s minds in the six
monthly negotiations - wholly superseding the actual numbers for the year 2000
announced in 1991.
   The agreement has no basis in EU law and compliance rests solely on MITI’s
willingness to enforce export restrictions and the willingness of Japanese firms
to “exercise moderation.” There was no definitive text that the Council would
agree. In principle therefore EU trade must be GATT compatible after 2000.
Any new import restrictions rely on securing agreement by a majority of
Member States approval by WTO. However in 1995 the Commission and
Council agreed to continue the block exemption for distribution of cars within
the EU, effectively perpetuating segmentation of markets, and facilitating the
administration of the quota regime.
   In this case EU trade policy became a substitute for industrial policy, and
competition policy discretion has had to be stretched to the limits. But the
involvement of the EU almost certainly liberalized the outcome.
       End box

    Overall, therefore, anti-dumping policy, while clearly capable of generating
outcomes with a defensible economic logic, appears a very indirect way to
achieve goals that would be better pursued directly. Commission officials
privately acknowledge this, but point out two disadvantages of subsidies as an
alternative. Firstly production subsidies would be restricted by WTO rules even
if the Rome Treaty was interpreted to allow them. Secondly it is observed, as
Krugman (1987, p. #) argues in his “sadder but wiser case for free trade in a
world whose politics are as imperfect as its markets,” that optimal subsidies are
as rare as optimal tariffs. The Commission’s state aid regime has its limitations.
It has been suggested that in the case of protectionist policies there is always a
legal actor entitled and eager to challenge the imposition of a doubtful measure
in court.
    In the two other industrial domains where trade policy has been active - cars
and textiles/clothing - we must surely conclude that while policies in both cases
have been unacceptably protectionist from the point of view of economic
efficiency, there is reason to argue that the “value added” of the Community
dimension has been to restrain the degree of protectionism that would otherwise
have occurred.
    Overall, therefore, the best case for EU trade policy may be to adapt Hilaire
Belloc’s advice to a nervous child: “Always keep tight hold of Nurse, for fear of
finding someone worse.” EU trade policy is far from being either equitable or
efficient, but a sober consideration of what might be implemented by Member
States in its absence casts it in a less unfavorable light than such an observation
might imply.
Technology Policy and Other Categories of Industrial Policy
The remaining aspects of policy towards industry are those involving
Community expenditure and those governing non-competition regulatory
matters (mainly standards policies). We shall focus on the technology aspects
here because this is where most of the action has taken place, even in terms of
standards, though it must be acknowledged that there are many other areas of
policy that affect industry (environmental protection legislation, for example).
   As we have noted, competence in technology policy as opposed to industrial
policy in general was given to the Community by the Single European Act, after
a series of initiatives had been launched by Commissioner Davignon in the early
1980’s. The Maastricht Treaty as we noted appeared to extend the competence
of community institutions, but in fact the application of Article 130 is highly
restricted. Measures must be adopted by unanimity, they must be in pursuit of
pre-existing Treaty aims, and they must conform to the principles of competition
   The SEA and Maastricht Treaty place a heavy emphasis on the technological
aspects of industrial policy, though they also refer to the possibility of
adjustment to assist adaptation to world competition.
   Before evaluating the logic of the EU’s approach, we need to ask: why have
a technology policy at the EU level at all? There is a major debate about the
usefulness of technology policy at any level of government. The two major
arguments in favor appeal to externalities and to co-ordination problems.
Despite some skepticism, a reasonable consensus exists among economists that
technological innovation tends to create positive spillover effects, which will
lead to under-investment in R&D. (Some models indicate that competitive
market conditions may produce too much R&D through duplication; too much
research but not enough innovation.) The degree of spillover is clearly likely to
be greater, the more difficulty there is in appropriating R&D results. So a prima
facie case exists for supporting basic R&D more than product specific
   What is less obvious is the precise geographical nature of spillovers, and
consequently the appropriate level of government for any particular intervention.
In addition, the information requirements for clearly discriminating between
different types of R&D activity may be very demanding.
   There is much to be said for the view that most R&D spillover activity occurs
through the operations of fairly localized industrial networking structures (cf.
Best, 1990) and that much of the spillover is accessible to the entire world
technical community. There is thus room for local and regional initiatives. The
respective roles of national and EU actors are less clear. To justify action at an
EU level we are looking for market failures that can be identified and
counteracted by public policy, but which in terms of the subsidiarity principle
cannot be handled by voluntary co-ordination by states.
   When we look at actual EU technology policy we can see a very clear tension
between the wish of the Commission to enlarge the scope of policy established
at an EU level and Member States who are mindful of the futility of zero-sum
intervention at national level, but equally concerned to avoid ceding unnecessary
power to the Commission. To this we must add the on-going tension between
DG 3/DG 13 (pre1994) and DG 4.
   One outcome of this is that the Member States of the EU very deliberately
chose to place a significant fraction of total collaborative R&D activity outside
the framework of the EC, using the Eureka program and such ad hoc
arrangements as Airbus. The Commission can participate in Eureka programs
but only as one partner. Advocates of the Eureka approach argue that it has the
benefit of avoiding the problems of the juste retour. EU spending rules would
allow the Community to pre-commit to fixed expenditure totals for industrial
policy (as Krugman has advocated in the U.S.), but they are not well adapted to
ensuring an optimal allocation of spending.

Mac and HDTV
In 1985 the Japanese proposed a new worldwide TV standard (MUSE) to
replace PAL NTSC and the abortive SECAM standard once intended as a
springboard for French ambitions. Driven by Philips and Thomson and the
French government (though restrained by some other Member States) the
Commission - especially DG 13 - felt obliged to act. It alerted the firms to the
challenge they faced and orchestrated a program of development for a European
standard Mac. Eureka aid (including Commission cash) and EC regulatory
instruments were used, but in the end Mac had to be abandoned. Critics argued
the policy was outdated interventionism. A closer reading suggests a more
balanced judgement. Any new TV standard must involve equipment makers,
broadcasters and regulators, if it is to take off: there are market failures here.
   Mac could only have succeeded if the broadcasters had been more closely
involved and if the project had been realized before new digital techniques
   But the balance is not wholly unfavorable: EU firms have a strong position in
new TV technology (though mostly in the U.S.).
   Lessons have been drawn: the Commission is now focusing on its role as a
potential regulator in digital TV looking at competition issues, e.g. for set top
boxes, etc. The basic policy problem remains the same: how to avoid global
market domination by non-EU players.

   The efforts of the Commission have been focused mainly on the promotion of
collaboration schemes for the creation of new technologies, while trying to
preserve competition in the product market. The effort has been driven both by
industrial/technology policy, and by relatively liberal rules under the
competition regime for technological collaboration. Some would even argue
that the technology component in competition policy on R&D joint ventures has
gone too far (Frazer, 1993).
   The EU’s efforts to promote collaboration in high technology have clearly
been more successful in promoting collaboration than in generating direct
results. The electronics industry does not appear to have responded to the
subsidies and collaborative projects of the 1980’s: the revival of production in
the hands of non-EU based firms may have as much to do with trade policy as
anything else.
   It does not follow that there is no role for EU intervention. The role of
Community policy might perhaps best be conceived as attempting primarily to
reduce the transaction costs of privately beneficial collaborative activities (with
a subsidiary aim of contributing marginal funding towards projects of general
interest with external benefits). Such a catalytic role towards collaboration can
be seen as solving a co-ordination problem, in which private firms and member
state governments seeking to internalize externalities might do so in a number of
mutually beneficial ways, but can benefit from an agency that systematically
examines collaboration possibilities and can make recommendations.
Sometimes the number of parties potentially involved is small and their degree
of information about the prospects for collaboration is high, as in the case of
aerospace.2 In other cases many more parties are concerned: product
standardization work, efforts to harmonize public procurement rules, and trans-
European Networks are cases in point. There is a plethora of models which
show how the market cannot be relied on to deliver the right results in standards.
Ironically the aim of the Mac program was precisely to accelerate the premature
lock-in that writers such as David have warned is a perverse result of market
forces. It could be argued that a better policy would have been to subsidize

2   Neven and Seabright (1995) suggest that external benefits to the EU from lower prices
    for aircraft have been fairly small, and that Airbus will have been justified primarily if it
    turns out to be profitable (of which there is a reasonable prospect). This in turn implies
    that the number of member states, which had an interest in collaboration, was not large:
    the EU as honest broker was redundant. This contrasts with the areas of product
    standards or trans-European networks, where potentially many more member states are
firms to carry on research in alternative technologies once they looked too
committed to Mac. Another paradox is that the money provided for Mac may
well have allowed Philips and Thomson to keep up the investment in digital
technology in the U.S. and enhanced PAL in the EU!
   Many of the lessons of the 1980’s have been drawn in the new approach to
the Information Society. As Foray et al. (1995) show there has been a major
switch in emphasis towards the use of competition policy instruments and
regulation of technology to facilitate rather than to impose solutions. Europe’s
success in the GSM mobile phone field can be identified as an example where
“market friendly co-ordination” has been able to generate positive benefits for
both producers and users. In the new digital media and telecom field the
Commission has preferred to try and structure markets in such a way that profit
rather than subsidies will be a lure. Use of the competition instrument breaks
the policy making log-jam but requires the Commission to forecast what the
effects of mergers and strategic alliance are going to be.
   We can therefore see that after many years of trial and error, technology
policy is increasingly dominated by a competition-based logic, but given the
nature of EU competition policy, its stress on pluralism, the justification of R&D
collaboration and the emphasis on market integration, it is not unreasonable to
say that to a large measure we have technology policy within competition
policy, not competition policy instead of technology policy (cf. Taddei and
Coriat, chapter.7).

What can we say about the nature of the rules and constraints governing
industrial policy in the EU? Institutionally the rules of competition policy are
hierarchically privileged in two senses. Firstly the Treaties give a certain
priority to the principles enshrined in competition law. Secondly decision
making made in the framework of competition law is much easier to accomplish
because the Commission (subject to the Court) is the sole actor. The
competition rules are not entirely oriented towards efficiency goals but their tilt
is definitely in this direction; however, as we have seen there is significant
institutional scope for non-competition criteria to influence what are ostensibly
arguments constructed purely on competition grounds. On the other hand
policies explicitly involving a high degree of intervention are not ruled out by
the Treaties. But as a general rule the more intervention is involved, the stricter
are the procedural requirements for action. A highly interventionist policy could
be adopted by the EU - but only if it secured unanimous support from the
Council on a proposal from a majority of Commissioners. There are several
examples where the Community has adopted policies that have no conceivable
efficiency rationale, but these have been in the presence of overwhelming
political backing which no technocratic criteria could be expected to block. If
we adopt the view that interventionist policies towards industry are sometimes
justifiable but that economic actors have an incentive to be opportunistic in the
information they supply, the broad structure of rules governing industrial
intervention in the EU, which involves it more as a regulator than a spender, can
be seen to have a reasonably compelling (if baroque) rationale.

Ashcroft, B./Wishlade, F. (xxxx): Methodology for Evaluating Regional Investment Aids,,
   European Politics Research Centre, University of Strathclyde.
Bourgeois, J./Demaret, P. (1995): The Working of EC Policies on Competition, Industry and
   Trade: A Legal Analysis, in Buigues et al. al list all and source pp.65-114.
Buigues, P./Jacquemin, A./Sapir, A. (eds.) (1995): European Policies on Competition, Trade
   and Industry: Conflict and Complementarities. Aldershot, Edward Elgar.
Cawson, A/Holmes P. (1995): Technology Policy and Competition Issues in the Transition to
   Advanced Television Services in Europe. Journal of European Public Policy, Winter,
Deacon, D Comments on Foray et al. pp 322-328. in Buigues et al. al list all and source
Dornbusch, R. (1990): Policy Options for Freer Trade: the case for Bilateralism. In: R.Z.
   Lawrence/C.L.Schultze (eds.): An American Trade Strategy: Options for the 1990s..
   Washington, D.C.: Brookings. pp. 106-133.
Emerson, M. et al. (list all names) (1987): The Economics of 1992: The E.C. Commissions’s
   Assessment of the Economic Effects of Completing the Internal Market. Oxford: Oxford
   University Press.
European Commission, An Industrial Competitiveness Policy for the EU, Com (94) 319
Foray, D./Rutsaert, P./Soete, L. (1995): Telecommunications services. In: Buiges et al al list
   all and source: pp.. 268-321.
Frazer,T (1994): Vorsprung durch Technik: In: T.Frazer/M.Waterson (eds): Competition
   Law and Economics. London: Harvester Wheatsheaf, pp. ###.
Gatsios, K./Seabright, P. (1989): Regulation in the European Community. Oxford Review of
   Economic Policy, 5 (2), pp. 37-60.
Gual, J. (1995): The Three Common Policies an economic analysis. In: Buigues et al list all
   and source1995 pp. 3-48.
Holmes, P./Smith, A. (1995): The Automobile Industry. In: P. Buiges/A. Jacquemin/A.
   Sapir/ E. Elgar (eds.): European Policies on Competition Trade and Industry. location &
   publisher, pp. 125-159.
Krugman P. (1987): Is “free trade passé”. Journal of Economic Perspectives, 2(1), pp. 131-
Krugman, P. (1992): The Age of Diminished Expectations. Cambridge, MA.: MIT.
Marchipont J. (1994): La Stratégie Industrielle de l’Union Européenne pendant et après
   l’achèvement du marche unique. Nancy: Doctoral Thesis, University of Nancy.
Neven, D./Nuttall, R./Seabright, P. (1993): Merger in Daylight. Centre for Economic Policy
   Research (CEPR).
Padoa-Schioppa, T. (1987): Efficiency Stability and Equity. Oxford: Oxford University
Seabright, P. (1996): European Union Policy Towards Vertical Restraints: A Proposal and
   an Evaluation. Cambridge: University of Cambridge, mimeo.
Taddei D./Coriat, B. (1993): Made in France. Livre de Poche, Paris, France.
Winters, L.A. (1988): Completing the Internal Market: Some Notes on Trade Policy.
   European Economic Review, 32, pp. 1477-1499.

To top