A service of the

Download article as PDF

This study analyses institutional features national authorities should have in order to apply common policies effectively. It shows that the effective implementation of EU law and policy can be achieved without increased control. The article suggests that benchmarking of the results of the application of common rules by the national authorities and comparative assessment of their performance is more efficient.

This article examines the role of the institutions of partner countries in successful integration schemes. The theory of economic integration has largely been preoccupied with three questions:

  • What are the benefits and costs of integration and under which conditions can integrating economies become better off than they would by acting unilaterally?
  • What common instruments and institutions are necessary for the integrating economies to function together?
  • Which countries have economies with the right level of development, factors of production and product mix which make them best suited to become partners?1

The theory has always recognised the inherent tension in the process of integration. On the one hand, there are benefits to be had from increased trade and investment between partner countries and from cooperation in setting common policies. On the other, there are also advantages in being able to act alone to further national interests. Moreover, even if cooperation were to be preferred in theory over unilateral action, casual observation suggests that not all countries would be suitable partners. Their policy preferences may be too far apart.

The first attempts to answer these three questions were made in the 1950s, and over the past 60 years or so the literature has grown exponentially. Considering the large number of books and articles on why, how and which countries should integrate, it is rather surprising that there is hardly any reference to the kind of institutions that partner countries should have. I do not mean common or supranational institutions but rather national institutions which are responsible for implementing at the national level the common rules and policies decided by partner countries at the supranational level.

It is not that the literature has ignored institutional issues; on the contrary, there are many books and articles on the powers and instruments that should be vested in common (i.e. supranational) institutions.2 Indeed, much attention has been given to the type of institutions that a regional bloc should establish in order, first, to facilitate the adoption of common rules and, second, to incentivise or force, if necessary, partner countries to respect those common rules.

One can also find comments in the literature on the need for national institutions to apply common rules correctly.3 But there is no systematic analysis of the type of institutions that have that capability. The design, structure and functioning of the institutions of partner countries have been neglected in integration theory.

On this issue, practice in the EU is far ahead of theory. As of the early 1990s, the European Commission began inserting in its legislative proposals provisions on the organisational structure of national institutions that were to be responsible for the application of EU law. In effect, the Commission was trying to create “allies” within Member States who would function as agents supporting integration.4

Moreover, the Commission adopted a tougher position towards the then candidate countries of Central and Eastern Europe by insisting that they demonstrate the existence of administrative capacity to apply the acquis communautaire. The Commission’s attitude was partly shaped by the fact that those countries had to undertake wholesale institutional reform to remove the vestiges of communism. The integrity of their public administration was in doubt. But the Commission’s position was also influenced by the increasing realisation that national administrations did matter to the uniform and coherent application of EU law.

The importance of well-functioning national institutions for successful integration has been made abundantly clear by the ongoing economic crisis. A case in point is the misreporting and falsification of statistical information by Greece in order to gain entry into the eurozone.

The formulation of economic policies at the EU level depends on timely and reliable statistics. Eurostat, which is the statistical authority of the EU and is a department of the European Commission, does not collect any data of its own. It is the national authorities which collect data and compile statistics for use at the EU level. For this reason, once the consequences of the Greek misreporting of data in 2004 were realised, the Council requested the establishment of minimum standards for independence, integrity and accountability for national statistical authorities.

It is now well understood that the quality of national institutions does have an effect on the quality of EU policymaking. It is, therefore, pertinent to consider whether the direction of causality also runs the other way – whether the quality of national institutions affects, perhaps decisively, the quality of the application of EU policies. It then becomes necessary to ask what kind of national institutions a bloc like the EU should want its members to have. Naturally, the EU or any other bloc should desire its members to have institutions that can ensure the success of the integration process. In other words, they should be capable of correctly applying the rules and the policies of the bloc.

The purpose of this article is to consider what institutional features and capacity national authorities should have in order to apply common rules effectively. In any sporting competition, no one would want to team up with someone who is eager but does not have the right skills. So why should countries team up with others which are eager but do not have the right institutional skills or capacity? Team members have to be reliable because they need to trust each other. What features make institutions reliable so that countries can trust each other?

The Nature of the Problem of Effective Policy Implementation

The implementation of a policy is a never-ending process. It involves constant monitoring of the effects of implementing actions and the adjustment, if necessary, of implementing instruments. To understand why policy implementation is more likely to follow a circular rather than linear path, it is necessary to appreciate the “root problem” of policies. The word “root” has two meanings here. First, it indicates a problem which is at the core of any policy formulation and application. Second, it is a pictorial representation of the nature of that problem. It very much looks like the root systems of plants, which spread in various, sometimes unexpected, directions.

The root problem has two aspects. First, a policy objective can normally be achieved through multiple instruments. This necessitates selecting the instruments according to certain desirable features such as efficiency or cost, rather than simply choosing them at random.

For example, the economic integration of migrants can be achieved through training courses or through placement at selected jobs. The protection of energy users from abusive business practices can be achieved through price regulation or market liberalisation and the entry of more providers. The choice of the right instrument is dependent on factors such as the availability of information on costs, the need to incentivise market operators to invest and offer cost-efficient services, and the administrative costs of supervision.

The second aspect of the root problem is the mirror image of the first. Different instruments may have additional effects beyond the desired ones. These side effects can be costly and counterproductive.

For example, as is now well understood, price regulation of energy utilities can keep prices at an affordable level for consumers by preventing energy suppliers from charging above cost. This, however, comes at the expense of failing to induce sufficient investment that can lead to lower prices in the long term. Companies can be incentivised to make long-term investments only if there is the prospect of sufficient profit, which implies that prices may have to remain considerably above costs in the short term.

Figure 1 depicts the two aspects of the root problem. The important point here is not that there is a magic prescription that can eliminate the root problem. Rather, the root problem requires constant monitoring of results, assessment of feedback and, if necessary, admission of failure to reach pre-determined targets. Failure should then lead to adjustment. In a world of informational imperfections, policymakers can never be sure that what appears feasible can indeed be achieved. Furthermore, if it is eventually achieved, they can never be sure that it is indeed the best result that could potentially be achieved. The act of implementation itself generates information which is useful and has to be fed back into the design of the implementation process. However, since the information that the act of implementation generates is partly shaped by the efforts of the implementing or enforcing authority, one can never be certain about the extent to which such information reflects the objective state of the world or the subjective efforts of the authority. One way of making sense of the feedback is to compare what one authority achieves or how it performs with the performance of other similar authorities. Peer comparison is a valuable “reality check”. I will return to this issue in the penultimate section.

Figure 1
The Root Problem of Policy Formulation and Implementation
Nicolaides Fig-1.ai

A Model of Institutional Capacity to Apply Common Rules Correctly

The capacity to apply rules correctly, or at least as they are meant to be implemented, means at minimum that national institutions should be able to carry out their tasks without any hindrance or interference from other political authorities. Yet, Member States do interfere whenever it suits them to do so. It is not unusual for Member States to agree to something in Brussels and then fail to honour their commitments later on because it transpires not to be in the national interest to do so. It is also not unusual for Member States to be outvoted in the Council. There is then a strong incentive for them to find creative methods of non-compliance, thereby obstructing the performance of their own institutions. By contrast, it is in the interest of the EU that national authorities respect both the letter and spirit of EU rules. Let us call this institutional feature “autonomy”, meaning protection from political interference.

At the same time, no one should want institutions to apply rules blindly and without consideration of whether they produce the desired results. No policy is perfect. Policies have to be adjusted during their application. This is another form of learning by doing. Institutions or organisations that are empowered to implement policies should have the capacity (meaning ability or willingness) to assess the outcomes of their decisions, their policy record and their performance and to take corrective action if needed. At the same time, they must have an obligation to justify what they do. Let us call this feature “accountability”, meaning both an incentive to improve and an obligation to explain one’s performance.

The EU should, therefore, want its members to have institutions which are autonomous and accountable. Let us formalise these institutional features and examine their possible interrelationships. Assume there is a variable N which measures the degree of outside involvement in supervising the functioning of an institution or organisation. Variable N can also be thought of as a measure of the instructions given by outsiders. It is reasonable to conjecture that autonomy is inversely proportional to N, so that high autonomy means a low value of N.

By contrast, we can surmise that accountability is directly proportional to N, so that high accountability means a lot of outside supervision and control, i.e. a high value of N. In notational form, autonomy and accountability are functions of N such that

        Autonomy: T = f (N) with (dT/dN) < 0

        Accountability: A = g (N) with (dA/dN) > 0

Figure 2 illustrates the possible shape of the two functions.

Figure 2
Institutional Autonomy and Accountability as Functions of Outside Involvement and Supervision
Nicolaides Fig-2.ai

How much autonomy and accountability should institutions have? Too much or too little of either is not necessarily an optimum state of affairs. In fact the postulated relationships between outside involvement and autonomy, on the one hand, and accountability, on the other, indicate that too little outside involvement, while good for autonomy, is not good for accountability and vice versa. For example, excessive outside involvement and controls on what an institution can or cannot do may force that institution to account for every decision it makes, but this comes at a cost of having little discretion to adjust the policy instruments it manages. The question that naturally arises is whether there is an optimum value for the degree of outside involvement N.

To answer this question, we need a way to measure the costs and benefits of institutional autonomy and accountability. This would allow us to construct the best institutional design by choosing the value of N so as to maximise net benefits.

In notational form, the objective of the EU would be to impose such obligations on Member States so as to

                maximise BT(N) + BA(N)

where BT(N) are the net benefits from autonomy and BA(N) are the net benefits from accountability for any given value of N. The net benefits are the gross benefits minus any costs.

I assume that like most functions of benefits they reach a peak, after which decreasing returns occur so that the benefits decline as institutional effort increases.

Once the functions for the benefits can be calculated, we can derive N*, which denotes the optimum degree of outside involvement (i.e. control, supervision, auditing, etc.). A possible solution is shown in Figure 3.

The reader may think that this has little practical value. Admittedly, no one has yet managed to measure the benefits from autonomy and accountability on a cardinal scale. If N* cannot be identified, why bother?

Figure 3
Net Benefits from Autonomy and Accountability
Nicolaides Fig-3.ai

There are two answers to this important question. First, Figure 3 demonstrates the dangers of simplistic prescriptions. Consider for a moment the position of N*. A small move to its left would reduce the benefits from accountability but would increase those from autonomy. A small move to its right would reduce the benefits from autonomy but would increase those from accountability. A large move to either direction, which in political terms could be perceived as the necessary radical reform, would in fact reduce the benefits from both. If these curves reflect reality to some degree, they warn that it is easy to get it wrong, especially when the effects are difficult to quantify.

The second answer is that the usefulness of this exercise is in its demonstration of the likely impact on net benefits from an upward shift of BA(N). This is illustrated by B'A(N). It indicates an increase in the sum of net benefits. How can this be achieved? It is possible through appropriate organisational reform that can make an institution more accountable without intensifying the degree of outside involvement. That is, for any fixed value of N denoted by Ni ,

                [BT (Ni) + B'A(Ni)] > [BT (Ni) + BA(Ni]

If this can be achieved, then we would be reasonably assured that strengthening accountability mechanisms, ceteris paribus, raises social welfare unambiguously.

How to Increase Accountability Without Increased Control?

The theoretical considerations in the previous section suggest that there can be benefits from more accountability without increased control. The question therefore is how this can be achieved.

I dismiss at the outset the option of imposing penalties for poor performance. This option would require that the right policy outcome or the right policy response to all possible market eventualities could be defined in meaningful detail in advance. But this is wishful thinking. The application or enforcement of many policies is an incomplete contract between principals (politicians) and agents (civil servants of public officials). Politicians delegate responsibility to public administrations without being sure that civil servants will always act in the most efficient or effective manner.

Possibly the only way of raising accountability without more external control is “self-control”. In this context, self-control should be understood as the willingness or incentive to assess and explain one’s own performance which, at any rate, is a fundamental aspect of accountability. Given that policy outcomes cannot be defined ex ante with sufficient accuracy, such explanations take the form of hypotheses of what another person or institution could reasonably be expected to do under similar circumstances. This incentive to explain one’s own performance, therefore, can be reinforced by appropriate comparison with other persons or institutions operating in similar conditions. It follows that peer review or an assessment of comparative performance are methods of establishing what could be reasonably expected from institutions responsible for policy implementation.

Comparative assessment is incentive-compatible. It induces institutions to reveal information about their performance and to explain what may prevent them from reaching the same level of performance as their peers (or, conversely, to outline the effective methods they use that their peers could adopt). At any rate, the process of comparative assessment generates useful information.

As a result of the economic crisis, it appears that peer review and collective surveillance have been expanded and strengthened in the European Union. Admittedly, such developments do not necessarily prove that comparative assessment leads to more self-control and improved accountability. However, they indicate that it is practically possible to use methods other than legislation and direct control to improve policy implementation.


In this short article I have explained that the success of integration schemes depends on the effective implementation of common rules. Such rules are largely implemented by public authorities in partner countries rather than by common, i.e. supranational, institutions. Therefore, the integrity of the integration process depends critically on the quality of the application of those rules by national authorities.

It follows that the implementation of EU law and policy is improved not only by more direct control of the decisions of national authorities but also by benchmarking the results of their decisions and by comparatively assessing their performance. Thus, the effective implementation of EU law and policy is a collaborative project in the sense that national institutions learn from each other and collectively raise the quality of application of EU rules.

Phedon Nicolaides, European Institute of Public Administration, Maastricht, the Netherlands.

The author is grateful to Maria Geilmann for research assistance and useful comments on an earlier draft.

  • 1 See references in J. Pelkmans: European Integration, Harlow 2006, Pearson Education, 3rd edition.
  • 2 See references in G. Majone, (ed.): Regulating Europe, London 1996, Routledge; G. Majone: From the Positive to the Regulatory State: Causes and Consequences of Changes in the Mode of Governance, in: Journal of Public Policy, Vol. 17, No. 2, 1997; P. Nicolaides: A Theory of Regulatory Integration, in: Intereconomics, Vol. 41, No. 1, 2006, pp. 37-43 . H. Hofmann, A. Türk: The Development of Integrated Administration in the EU and its Consequences, in: European Law Journal, Vol. 13, No. 2, 2007, pp. 253-271.
  • 3 Most of this literature is actually concerned with identifying political and economic factors that determine transposition of EU directives and compliance by member states. For a review of the literature, see G. Falkner, M. Hartlapp, O. Treib: World of Compliance: Why Leading Approaches to European Union Implementation Are only ‘Sometimes-True Theories’, in: European Journal of Political Research, Vol. 46, No. 3, 2007, pp. 395-416; T. Borzel, T. Hofmann, D. Panke, C. Sprungk: Obstinate and Inefficient: Why Member States do not Comply with EU Law, in: Comparative Political Studies, Vol. 43, No. 11, 2010, pp. 1363-1390; M. Haverland, B. Steunenberg, F. van Waarden: Sectors at Different Speeds: Analysing Transposition Deficits in the European Union, in: Journal of Common Market Studies, Vol. 49, No. 2, 2010, pp. 265-291.
  • 4 P. Nicolaides, A. Geveke, A.M. den Teuling: Improving Policy Implementation in an Enlarged European Union: The Case of National Regulatory Authorities, Maastricht 2003, European Institute of Public Administration.

Download as PDF

DOI: 10.1007/s10272-012-0416-9