Understanding the Waves of Agencification and the Governance Problems They Have Raised in Central and Eastern European Countries

The goal of this paper is to provide tools to understand and analyse waves of “agencification�? in transition countries of Central and Eastern Europe. Agencification is a shorthand for the process of delegation and devolution, in which more autonomy, particularly in personnel and financial issues, is granted to public bodies, which either remain legally part of the state or acquire their own legal personality. It can also mean creating or moving functions to bodies, which are subsidiary or separate from ministries/departments (Gill, 2002). In transition countries, most public organisations inherited a legal personality already from communism, with consequences which are both procedural and substantive. Therefore, agencification in transition countries usually means the creation of new autonomous bodies for new functions or a significant increase in the autonomy of existing legally separate bodies either on an individual or a collective basis. The paper aims to contribute to knowledge and analysis of the process of agencification in Central and Eastern Europe by pointing out differences between transition countries and stabilised industrial democracies, but also similarities. It focuses on incentives of individual actors in the process and presents an analytical framework for a closer look at the process. After analysing its practical consequences, it makes several recommendations.


Introduction
The goal of this paper is to provide tools to understand and analyse waves of "agencification" in transition countries of Central and Eastern Europe. Agencification is a shorthand for the process of delegation and devolution, in which more autonomy, particularly in personnel and financial issues, is granted to public bodies, which either remain legally part of the state or acquire their own legal personality. It can also mean creating or moving functions to bodies, which are subsidiary or separate from ministries/departments (Gill, 2002). In transition countries, most public organisations inherited a legal personality already from communism, with consequences which are both procedural and substantive. Therefore, agencification in transition countries usually means the creation of new autonomous bodies for new functions or a significant increase in the autonomy of existing legally separate bodies either on an individual or a collective basis.
The paper aims to contribute to knowledge and analysis of the process of agencification in Central and Eastern Europe by pointing out differences between transition countries and stabilised industrial democracies, but also similarities. It focuses on incentives of individual actors in the process and presents an analytical framework for a closer look at the process. After analysing its practical consequences, it makes several recommendations.

Transition and agencification parallels and differences
Transition is usually seen as a unique process without a precedent or a parallel. This approach can be exemplified by the saying that it concerns the issue of how to make fish out of soup as opposed to the more tractable question of how to make fish into soup. It should be noted though that many processes that are part of transition (particularly privatisation, liberalisation, deregulation) have at least some parallel in what has been happening in many developed and developing countries since the late 1970s.
In the case of agencification in Central and Eastern Europe, as in the case of most other reforms, these views translate into two possible extreme approaches to analysis of the process: • "normalism", according to which issues faced in transition are not qualitatively different from issues faced by other countries and therefore there is much to be learned from other countries not only on desirable final results, but also on the process and sequencing.
• "exceptionalism", which can roughly be summarised as the view that the process of transition is so different from anything experienced before that any parallels are bound to be irrelevant and can even be damaging as they can easily mislead.
Some of the principal arguments for normalism can be grouped under three headings.
First of all, one can argue that individual actors/agents respond to identical incentives in a similar way everywhere in the world so the response of individuals and institutions to reforms should not differ substantially.
Secondly, the creation of autonomous agencies during the transition process has a counterpart in the process of transformation of government that has taken place in many OECD Member countries since the late 1970s. As already noted above, the public administration and public sector in these countries have been in a constant flux, including privatisation of some functions, decentralisation and devolution and reorganisation of the delivery of public services within public administration.
Thirdly, the reforms in nearly all countries were driven at least to some extent by a similar feeling that the public sector needed a change of culture. According to this perception, it had grown too large, all-encompassing and bureaucratic to deliver desirable public services in an effective and efficient manner. As a consequence, emphasis has been placed on refocusing public administration on its unique and core competencies, on significantly more extensive use of private sector techniques in management and organisation and on transparency of public administration as a means of delivering more accountability to the public and politicians.
It is easy to overstate similarities between the transition and reform processes in OECD Member countries during the 1980s and 1990s. While many specific processes are similar, if not identical, they take place in a legal and institutional environment so different that identical processes or reforms can lead to unexpected results as they generate perverse incentives in a transition country. This is reinforced by the overall speed and scope of change in transition, which is of a higher order of magnitude than any experienced by an OECD Member country including the "Thatcher revolution" or New Zealand reforms under Roger Douglas. It has rarely been the case that within five to 10 years, prices covering 80-90 % of GDP were liberalised; sectors of economy covering more than 80% of GDP privatised; a new legal framework for civil, commercial and other interactions introduced; the public administration in all its aspects overhauled; health care and education systems restructured and put on a new footing, etc. Since much of this change took place in several steps, the resulting speed of change and reforms are unparalleled.
Transition countries also differ from OECD Member countries by the fact that transition meant a significant decrease in the scope of publicly "guaranteed" services. As Schick (2002) notes, despite the 1980s and 1990s rhetoric of rolling back the state in many countries, the share of public expenditure in GDP has grown rather than been cut and the scope of services, which the government is guaranteeing (either producing them, paying for them or using regulation to provide them) has been relatively stable. The situation in transition countries has been markedly different. Not only did the level of government involvement in the economy and in the society decrease by any fiscal or financial measure, the role of state was redefined "down", sometimes repeatedly.
Partially as a consequence of this redefinition, but also due to the economic challenges of transition, the public sector in transition countries has experienced a long-lasting and profound decline in funding, prestige and stability. It was clear in early transition everywhere and it is still true for most transition countries that the extent of government involvement in society both in fiscal and other terms and the size of the public sector were incompatible with a market democracy and the actual wealth of these countries. To remedy this, public sector and government involvement in civil society have been repeatedly and relentlessly cut during the transition, but these cuts were smaller than the actual decrease in financial resources available to governments. As a consequence, not only is the public sector doing much less in transition countries than it used to do, it is doing so with proportionally smaller funding and resources. This, together with a very high level of uncertainty and decreasing prestige of the public sector due to new employment © OECD 2002 alternatives for the best staff, led to the already mentioned long-lasting decline in funding, prestige and stability. How to manage that decline has been one of the key tasks for managers at all levels in the public sector.
This leads to the fifth major argument for exceptionalism. While boundaries of what is and should be in the public sector were and are uncertain to some extent everywhere in the world, 1 in OECD Member countries this uncertainty concerns only a relatively small portion of all organisations. On the other hand, during communism, everything was public even though there were different organisational forms. As a consequence, the boundaries of public sector both in positive and normative terms are much more blurred, particularly in the minds of citizens and public sector workers.
Formal and informal norms, i.e. rules of the game, have a very important effect on how individuals and organisations act. However, uncertainty caused by only gradually established rule of law, arbitrariness of the judicial system and repeated changes to the legal framework can in itself have an important effect on people's behaviour. In such an environment, risk-aversion becomes very high and what can be perceived as an action involving risk can also be quite different from what is usual in most OECD Member countries.
The last major argument for exceptionalism is the challenge experienced by many transition countries with regard to administrative capacity and its building. Many of the states involved are new, their employees could not be called civil service in any meaningful way and they had to deal with war and other extreme situations. These factors were often combined. This is not to say that there are no countries in the world experiencing similar issues. On the contrary, many developing countries face the same challenges, but they certainly have not been on the agenda of OECD Member countries over the last 20-30 years as they grappled with the issue of devolving and delegating power to autonomous agencies.
On the whole, therefore, one can argue that the transition process in the area of public administration reform in general, and in the area of agencies in particular, is a change of a different order than public administration reforms in stabilised industrial democracies. This, however, does not mean that comparisons and lessons from these countries are irrelevant. On the contrary, if applied carefully, with an eye for differences in the environment, they can be very useful. This is the approach used in this paper.

Process of agencification and public sector reforms
Agencification, as defined in the introduction, has been taking place during the whole transition process in countries of Central and Eastern Europesince 1990 and is still continuing.
It needs to be understood as only one potential change of status for components of the public sector. Other options include decentralisation to local or regional self-governments, sale to a private investor, or transformation into a nongovernmental organisation. The choice of an alternative has profound political implications as it influences a distribution of power that control of an agency bestows. Is it going to be national politicians and managers (autonomous agency), local politicians (decentralisation), shareholders (privatisation), activists and elites (NGOisation)?
Therefore, it needs to be repeated that agencification concerns only changes for organisation, which remain within the wider public sector of central government.
To understand what agencification really means an analytical framework is also needed. It allows a closer look at what agencification really means. When discussing changes in the status of public sector organisations, the usual vocabulary contains words such as commercialisation, corporatisation and privatisation. While these terms are very useful in distinguishing various types of changes, they are not sufficiently nuanced to capture all intricacies of change in the status of public organisations.
For example, a corporation can be 51% owned by the state, but be effectively private if a strategic investor for the remaining 49% has signed a contract with the government, which gives it management control. On the other hand, government can control only 34% of a corporation and still effectively control it, if the remaining shares are diffused among many small shareholders. 2 A slightly more sophisticated analytical framework for defining separateness of agencies could proceed along several axes: • personnel axis; • budgetary axis; • customers axis; • governance axis.
The personnel axis would measure the level of autonomy concerning personnel issues. These could range from civil service rules through other public sector rules to general labour code. It would also have to take into account the fact that these rules differ in individual countries.
The budgetary axis would measure how integrated in the general public finance is the budget of a particular organisation. This can range from gross transfer through net transfer to no transfer at all, but also include ability to raise own revenue and loans and ability to dispose of property.
The third axis concerns customers. In reality, public agencies range from those that work only for "its" ministry (e.g. some policy institutes) through a mixture of public and private customers to those that work only for private customers (i.e. traffic control).
The final axis would concern governance arrangement, ranging from various arrangements in management, but also including systems for audit and oversight.
Agencification as a process of delegation and devolution means significant movement along any of these axes towards more autonomy and as such can mean many things. This is also the experience in most countries of Central and Eastern Europe and even though the situation differs from country to country, such a process was a dominant form of change in the central government -agency relationship in the following areas: • corporatisation of government businesses (i.e. of activities, which are consensually perceived as for-profit) and which are to remain as governmentcontrolled for various reasons; • higher education institutions; • research and research institutions; • artistic and cultural institutions; • publicly controlled media; • earmarked funds -from pensions to agriculture; • regulators.
In most of them, agencification involved more independence in management, quite often independence in personnel issues and in certain cases even creation of an independent source of financing. Even though there are exceptions, agencification generally took place in organisations where the principal customers were not governments or their elements, but individuals and corporations.

Initial conditions: Paradox of Communist autonomy
To begin an analysis of the process of agencification, one needs to take into account the initial conditions left by the Communist regimes. These conditions can be succinctly described as the paradox of Communist autonomy. This section explores the paradox and sets out some of its main consequences for the process of agencification in transition countries.
Communism (or Real Socialism as it was officially and, in a way, aptly called in countries themselves) meant a totalitarian system based on collective ownership of all means of production and a repressive and intrusive political system. Both outsiders and insiders often tend to see it as an environment with a very low-level of autonomy for individual actors in any area. Such a view tends towards perception of the whole Communist society as a centralised, vertically and horizontally integrated hierarchy, where the centre (e.g. central committee of a Communist party and its government apparatus) directed resources and activities of sectors, organisations and individuals.
This often leads both insiders and outsiders to see a similarity between such vertical structures of Communist state and vertically integrated ministries of OECD Member countries before the agency drive in some of them. Consequently, it is easy to perceive a similarity between the fall of communism and the beginning of reforms in some OECD Member countries to provide more autonomy to parts of the governmental machinery.
Such an approach ignores several important facts.
The officially tightly knit hierarchy contained thousands of organisations with legal autonomy and millions of individuals. While the system could rely to some extent on its ability for arbitrary use of power to resolve conflicts between interests in this hierarchy, arbitrary use of power in itself was an insufficient answer to daily routines of administration in a complex society. The Communist system lasted from 40 to 70 years in the industrialised economy in countries of Central and Eastern Europe. Citizens were provided with a welfare-to-cradle superstate (issues in quality, responsiveness and ability to generate wealth notwithstanding). No economic, political and social system would be able to last so long in these complex conditions unless it developed a relatively predictable system for conflict resolution between both individual and organisational interests.
As several authors, including Mlčoch (2000), argue, the hierarchy was also largely an illusion: "In a closed system of hierarchical management, the planning was a widespread social game based on a all-encompassing dichotomy between the real rules and the official ones. In reality, local "controlling" groups tried to maximise their share of a social pie in the inverted pyramid. Planning was an instrument and an ideological smokescreen to utilise a monopoly power over allocation of scarce resources, information and decision-making processes within the social reproduction process" (pp. 30-31). 3 In other words, the hierarchy involved not only functions undertaken by the public sector in OECD Member countries, but also the whole corporate sector (enterprises). The well-known problems of information flows and information asymmetries for public sectors of OECD Member countries were accentuated due to span-of-control problems and lack of accountability systems. Managers of organisations were the real masters of the system because of their unique position in the information flows and decision-making. Since the "public" and "corporate" sectors were treated equally under the Communist system -both were controlled by sectoral ministries and were subject to similar regulatory environment -this led enterprises to behave like the "civil service" and "civil servants" to behave like corporate managers.
The result was that the Communist system bestowed to its democratic heirs a system with a very high number of individual organisations both in the "corporate" and "public" sector with dedicated functions. These organisations were often based on a single-purpose agency principle. While the Communist system had precise rules for their budgeting, management and HR procedures, these organisations generally had no accountability systems in addition to their supervision by ministries. They did, however, have a number of legal relationships sometimes enforceable in courts with other elements of the government. Since the "public sector" organisations themselves and their ministries saw them on par with "corporations" (as it was unclear anyway what belongs where), this led to a high-level of both de iure and de facto autonomy.
During transition, this philosophy was reinforced by "a-need-to-survive" mentality of dramatic cuts in budget allocation, in which public organisations were encouraged and forced to raise revenue regardless of their mission and suitability for such an approach.
All of this led to a situation where the real heritage of communism is not a hierarchical, disciplined public sector, but a chaotic free-for-all, where organisations often have legally-defined autonomy, rights and responsibilities, their staff and particularly managers feel certain informal ownership rights and the distinction between public-and private-sector mentality is blurred or non-existent in the eyes of most actors.

Actors in the agencification process and their conflicting incentives
Any analysis of the process of agencification should include an analysis of the role of individual actors in the process and their incentives. Indeed, this paper uses incentives and expectations as its principal tool of analysis.
First of all, let us introduce groups of the most important actors: • managers of agencies and agency staff (in the case of already existing organisations) or managers and staff of those parts of the ministerial structure that might be structured separately; • ministers and political parties; • foreign actors (European Union, International Monetary Fund, World Bank, Organisation for Economic co-operation and Development, etc.); • other influential stakeholders (e.g. trade unions); • general public.
Each of these groups CAN have ex ante incentives to support agencification and increase in autonomy of a given agency based on its EXPECTATIONS. A stylised summary is provided below.

© OECD 2002
For managers and staff of agencies, the incentives are usually related to an increase in autonomy, a decrease in uncertainty and an increase in pay and other benefits. For staff, the status of autonomous agency can also mean release from many restrictive regulations valid in the rest of the government (procedural, decision-making, etc). The decrease in uncertainty is slightly paradoxical and related to frequent absence of a civil service system and frequent political changes in policy directions in many transition countries, particularly in the early to mid-1990s. Arm's-length relationship with the government can thus increase certainty in terms of employment, but also in terms of planning. Last, but definitely not least for both managers and staff, autonomous status is often related to freedom from governmental pay rules or, if not that, at least their relaxation.
For ministers and political parties, the motives for agencification are, in this stylised approach, based on expectations of decreasing responsibility, decreasing financial burden and increasing patronage. The distance between a ministry and an agency can be exploited as a tool for decreasing political responsibility for agency's actions and problems. The expectation of decreasing financial burden is often related to an increased motivation for an agency to improve its financial situation. This occurs by generating revenues and decreasing costs, as well as more frequent use of some sort of extrabudgetary means of financing for agencies. Expectation of increasing patronage is based on the fact that the number of positions, which are attractive for patronage purposes, usually increases in the case of autonomous agencies due to the existence of boards and an increase in the number of managerial positions.
Foreign actors can support creation of agencies in the expectation of improved performance and professionalism. This is related to the point made above with regard to incentives of staff and managers. Since foreign actors also often have very narrowly focused priorities (based on projects in which they are involved or due to donor specialisation), it is often useful for them to increase the autonomy of an agency or even create a new one, if such an agency can have a focus similar to the one of a foreign actor.
For the media and the general public, autonomous agencies frequently hold a promise of improving services and decreasing tax burden based on expectations already explored above: Increased professionalism and stability as well as efficiency gains.
However, as any observant reader has probably noted, many of these expectations are mutually exclusive, i.e. they can hardly all be true at the same time for the same agency.
It is hard to reconcile increasing pay and benefits for employees and managers with a decreasing financial burden for the public purse unless dramatic efficiency savings can be made. Increased independence and professionalism does not square well more patronage for politicians and political parties. Even though decreasing political responsibility and improvement in services are not mutually exclusive by definition, it is difficult to see why a decrease in accountability to the public should lead to improvement in services for the same public.
The consequences of the irreconcilable expectations will be explored in the following section.

Consequences of agencification in transition countries
This section explores where agencification occurred and what the results were in countries of Central and Eastern Europe.
First of all, one should remember that actual results and expected results differ and that it is the latter that ex ante drive agencification. Expectations and results differ as all expectations can only rarely be fulfilled due to their mutual conflicts discussed above.
Agencification usually occurs when there is a strong coalition with incentives based on positive expectations, and the potential opposition is weak and/or disorganised and/or misinformed.
A stylised example is the minister together with agency head, based on or using foreign pressure, driving through the reform over heads of indifferent or mildly sceptical staff and public. The "mass" stakeholders -public and staff -are very often co-opted or at least neutralised by promises of better pay or better service or general "depoliticisation". Of course, many variations on this coalition could be presented and analysed further, but the value of such an exercise would be very limited and country-dependent.
More importantly, what generalisations can one make about the process and results of the agencification process?
First of all, agencification in these countries was rarely, if ever, pursued within a systemic conceptual and legal framework. The latter would set out in advance the goals, the instruments, their relationships and the timetable across the board. Due to the low quality and capacity in public administration and the political classes, agencification usually occurred as a quick sectoral fix based on a combination of the incentives mentioned above.
The result is often paradoxical. On one hand, the legal, accountability and financial framework for autonomous agencies is patchy and is often taken over by specific solutions and exceptions. This very often creates unexpected results, perverse incentives and negative consequences. On the other hand, there has rarely been substantial institutional innovation and creation of complex new institutional frameworks. This paradox is due to the sectoral, quick-fix nature of agencification.

© OECD 2002
The need for a speedy specific solution creates incentives to tweak existing, easyto-use forms rather than try significant institutional innovation.
When institutional innovation is undertaken, it is frequently exercised by use of institutional transplants from other countries, most often in the context of foreign advice or aid. The problem plaguing such solutions is that when institutional transplant is imported, other elements setting up its legal, accountability and financial framework are often lacking (e.g. activity-based budgeting, financial control and audit mechanisms, general accountability mechanisms for executives within the civil service rules). This is to be expected as such frameworks are very complex and often invisible even (or particularly) to those who work in the given field in a country providing the model. As de Soto (2001) writes in a slightly different context about the role of property administration mechanisms such as land ownership registration in the process of economic development, experts from developed countries themselves usually do not know how the system came about and do not ask themselves these questions, since their own system is "natural". It is but only ex post when it is transplanted that the missing elements are uncovered, often in a costly manner.
Another conclusion about agencification in transition, which differentiates it from many developed countries, is the use of agencification as a means of raising the tax burden. This is most often done in one of the two following ways. One is the so-called stealth taxation, i.e. non-transparent and earmarked taxation. Its forms are numerous -from taxes on the general population in the form of the socalled compulsory insurance payments to taxes on a specific group such as taxes on the regulated subjects to finance the regulator. The other one is by creating or increasing user charges at least to cost-recovery level and reducing tax funding of public services.
Neither of the steps is usually compensated by an equal decrease of other taxes or increase in transfers to the public, so it can be seen as a politically expedient way of doing the difficult thing in politics -increasing taxes.
An unsurprising result of the previous two factors is that agencification is usually implemented without thinking through the consequences, particularly without trying to simulate the incentives of individual players in a dynamic system.
On the other hand, given all the issues analysed above, it is surprising that transition has been marked by an excessive faith in institutional solutions. Many policy-makers, but also observers confused institutional reform with a reform of the system. To be more specific, many countries created so-called independent insurance institutions in the areas of health and/or pensions and/or unemployment, based on German or other models and confused this with a health care or pension reform.
Given these issues, it is to be expected that agencification creates, in many specific instances, a predictable public and political backlash when increased autonomy is followed by problems. The sources of problems can be roughly divided into two groups: avoidable and hard-to-see. The first set is caused by flaws in the design of the agencification process that could have been foreseen ex ante. A good example are extrabudgetary funds. If they are not subject to the existing control systems (however imperfect), which regulate use of "normal" budgetary funds, such funds are often characterised by mismanagement, fraud or expenditure beyond approved limits.
These issues are exacerbated by problems that were hard to foresee. The reason why many rational beings do not mismanage or undertake corrupt actions when heading an agency is not necessarily the fear of jail. It is the detriment to reputation that a scandal or public criticism creates, even when no criminal acts can be proven. Transition as such weakens the mechanism and contributes to unreliable functioning of accountability mechanisms. It is characterised by many upheavals and the so-called unrepeated games, so the value of reputation is much lower and potential benefits of misbehaving much higher.
The second problem comes about because the dynamic transition environment requires constant revision of the informal "social contract" between politicians and agencies. However, once substantial autonomy is granted, agencies become players in their own right that can and do resist changes they dislike. This gives rise to frequent public and political conflicts. This is related to the issue of policy entrenchment. Agencies generally have more information and expertise than ministries on policy issues related to their work. 4 The cause is often the policy weakness of ministries rather than particular strength of agencies. Larger agencies also have more resources for public relations, work with media and stakeholders. As a result, the expected split of policy/ execution tasks between ministries and agencies does not de facto apply as agencies are very often informally or even formally responsible for preparation of policy and legal changes in their area that are officially submitted by the ministry in question. Agencies can and usually do block changes they dislike using their relationships with stakeholders as well as the media.

Recommendations
What recommendations can one make on the basis of these findings?
Autonomy for an agency is meaningful if a clear formal or informal contract can be written between politicians as repositors of a public mandate and an agency. In other words, it should not be an agency's mission to decide what its mandate is. To create an agency for a given area without a clear set of goals just to get rid of a thorny political issue is a recipe for problems.

© OECD 2002
The second recommendation is a need for consistency in the creation of an agency. If a financial market regulator is being created, it is not truly relevant what the governing arrangements for a similar agency in the United Kingdom are, but how would any arrangements fit within a particular legal system as well as informal rules in the country where it is being created. Internal consistency means that the actual rules governing an agency's activities and incentives are mutually consistent in supporting efficient, effective and accountable functioning of an organisation.
Institutional solutions not only need to be internally consistent, they also need to be robust. Robustness, roughly speaking, means that a system will continue to work relatively successfully even if much goes wrong in the systems underpinning it, e.g. under budgetary duress or if important elements of the outside system change (civil service rules, public procurement rules, etc.). This is due to the extremely dynamic environment of transition. The number of changes taking place is so high and many of the changes are so significant that there is usually a number of unexpected effects on organisations stemming from interaction of various changes and "second-round" effects. The political system is also highly unstable. If the framework for an agency cannot successfully deal with these factors, an agency can be easily crippled or made dysfunctional.
In this context, there is a particular need for a robust system of financing. It is easy for part of the public sector to get lost in the enormous fiscal pressures transition countries have been facing. If an agency is at arm's-length with the government, the likelihood that it will be ignored so that the reconciliation of pressures will be done at its expense is significant. The exception here is if the basis of financing is robust and immune to immediate pressures. In other words, unless certain conditions are met, agencies can be destroyed by their own autonomy.
For reasons explained above, it is also recommended that each reform pushing for the creation of autonomous agencies or for an increase in autonomy of existing agencies should be required to answer a simple question about accountability: How is it going to create real accountability in a transition environment where the cost of reputation is low, formal mechanisms often slow and the wrongdoing usually not of provably criminal nature? If it cannot convincingly be answered, agencification might not be an answer.
One should also generally prefer market-based solutions if meaningful markets can be created. However, this recommendation needs to be balanced with transaction costs of a market-and contract-based relationship. Particularly in transition countries, one needs to ask whether effective contract enforcement as an important element in consideration of transaction costs exists? This ties in with the overall issue of institutional capacity. In countries with underdeveloped legal regimes, systems-based on command-and-control principles require a less sophisticated set of institutions and management tools and can be therefore sometimes preferable.

© OECD 2002
A related issue is whether formal accountability mechanisms cannot be complemented by a very powerful informal accountability mechanism -customer choice. If at least a significant part of customers has a choice in selecting their service provider, this can alleviate some of the accountability problems.
However, accountability is the ex post mechanism to guarantee efficiency, effectiveness and quality. The ex ante mechanism is the process of management/board selection. Anecdotal evidence suggests that the quality of the process for making that choice has significant consequences for the functioning of a public organisation, possibly even more significant than in more stabilised environments.
Finally, a reader convinced by our explanation of why agencification has been taking place in many transition countries will agree that it is often driven not by a need to grant some elements of the public sector a clear mission and autonomy to achieve it, but by piece-meal sectoral efforts to increase professionalism in parts of the public sector. Agencification measures are then a selective attempt at public sector reform, including reform of financial and governance aspects. If professionalism is the issue though, one might ask if transition countries would not be better served in using that energy for an overall reform of the institutional framework for the public sector and particularly on civil service reform.
The following box presents some of the reforms undertaken in this area in the Slovak Republic between 1999 and 2002.

Box 1. Audit reforms of public organisations in the Slovak Republic
This box discusses reforms of the public administration in the Slovak Republic undertaken between 1999 and 2002 relevant to the agencification issue. They were undertaken within the so-called "Audit of financing and activities of central authorities of state administration and organisations under their jurisdiction", conducted based on the Government Resolution No. 985/1999, Article C.26, in which the Slovak Government gave the Deputy Prime Minister for Economic Affairs Ivan Mikloš the task to procure such an audit. Based on the audit report submitted by the Deputy Prime Minister, the Slovak Government approved, in August 2000, a set of 53 tasks to implement the report recommendations. The tasks can be distinguished into several groups: • 19 tasks for all ministers aimed at implementing some of the recommendations across the board; • 13 tasks for the Deputy Prime Minister to prepare specific reform proposals (policy papers, government resolutions, laws) on most important cross-cutting systemic issues; • 21 tasks for specific ministers to prepare specific reform proposals (policy papers, government resolutions, laws) on systemic issues in their own areas.

Conclusion
This paper looked at the process of "agencification" in transition countries of Central and Eastern Europe. Its focus was on incentives of individual actors in the process and their practical consequences, acknowledging initial conditions and Box 1. Audit reforms of public organisations in the Slovak Republic (cont.) In the area of autonomous agencies, it should be noted that the audit implementation included specifically: • change in organisational form and/or ownership for more than 50 organisations. This was a systemic, across-the-board approach, but allowed flexibility as ministries could choose what form the change would take, be it privatisation, corporatisation or transformation into an NGO. • change in the relationship with remaining departmental agencies, which introduced an output-based contractual relationship between agencies and ministries; • integration of contract management of agencies into a framework of "programme" (activity-based) budgeting being introduced in the Slovak Republic; • improvements in transparency of departmental agencies to the ministries, the public and media. The principal tools were introduction of compulsory public annual reports from 2000, along with a list of information requirements to be fulfilled in the reports as well as compulsory public hearings of the annual reports; • unification of the legal framework for state-owned enterprises and corporations, with obligations regarding transparency, corporate governance and management selection; • elements in civil service reform to increase individual accountability such as individual evaluation and strengthening of its consequences; • creation of a more precise legal framework for market-based activities of public organisations, which would emphasise the end of cross-subsidisation and a need to compete on equal footing with the private sector; • creation of a unified framework for calculation of user charges in the case of administrative services where the government has a legal monopoly; • creation of legal tools to transform public agencies into a private law NGOs where the government can preserve its influence; • creation of a unifying conceptual framework for public law administrations; • reform of advisory bodies of the government and ministries, leading to halving their numbers and streamlining their operations. path dependency they induced. It proceeded on the assumption that the transition process in the area of public administration reform in general and in the area of agencies in particular is a change of a different order than public administration reforms in stabilised industrial democracies. Comparisons with and lessons from the latter can be very useful though, if applied carefully, with an eye for differences in the environment.
One of the key differences is the initial conditions left by the Communist system. However, the real heritage of communism is not a hierarchical, disciplined public sector, but a chaotic free-for-all, where organisations often have legallydefined autonomy, rights and responsibilities, their staff and particularly managers feel certain informal ownership rights and the distinction between public and private sector mentality is blurred or non-existent in the eyes of most actors. This has significant consequences for the process and results of agencification.
Agencification, defined as creation of new autonomous bodies for new functions or a significant increase in the autonomy of existing bodies either on an individual or a collective basis, has been generally pushed by a coalition of actors united in their ex ante positive expectations of results of such a move. However, many of these expectations are mutually exclusive, i.e. they can hardly all be true at the same for the same agency.
Agencification was rarely, if ever, pursued within a systemic conceptual and legal framework. It usually occurred as a quick sectoral fix based on any of the incentive combinations mentioned above. Such an approach resulted in an abundance of specific solutions and exceptions, very often creating unexpected results, perverse incentives and negative consequences, but very little substantial institutional innovation and creation of complex new institutional frameworks. When institutional innovation is undertaken, it is frequently exercised by the use of institutional transplants from other countries, most often in the context of foreign advice or aid. The problem plaguing such solutions is that when institutional transplant is imported, other elements creating its legal, accountability and financial framework are often lacking. Surprisingly enough, transition has been marked by an excessive faith in institutional solutions, where many policy-makers and observers confused institutional reform with a reform of the system.
Politically, agencification was often used a as means of discreet increase in the tax burden via so-called stealth taxation or increased user charge for services. Neither of the steps is usually compensated by an equal decrease of other taxes or increase in transfers to the public.
Frequently, the predictable public and political backlash occurs when agencification autonomy is followed by problems. Reasons are twofold: ex ante detectable flaws, but also in-built resistance to change, which comes from establishing a separate powerful actor -an autonomous agency. Consequently, one can speak of policy entrenchment, where agencies de facto take over policy-making in a given area and can successfully block changes they dislike.
Based on these findings, the paper makes several recommendations. Autonomy for an agency is meaningful if a clear formal or informal contract between politicians as repositors of a public mandate and an agency can be written. Such a contract needs to be internally consistent: actual rules governing an agency's activities and incentives need to be mutually consistent in supporting efficient, effective and accountable functioning of an organisation. Institutional solutions also need to be robust, meaning that a system will continue to work relatively successfully even if much goes wrong in systems underpinning it, e.g. under budgetary duress or if important elements of the outside system change. In the transition context, there is a particular need for a robust system of financing.
Each reform pushing for the creation of autonomous agencies or for an increase in autonomy of existing agencies should be required to answer a simple question: How is it going to create real accountability in a transition environment where the cost of reputation is low, formal mechanisms often slow and the wrongdoing usually not of provably criminal nature? One particular constraint is that in countries with underdeveloped legal regimes, systems based on command-andcontrol principles require a less sophisticated set of institutions and management tools and can be therefore sometimes preferable. Since accountability is the ex post mechanism to guarantee efficiency, effectiveness and quality, the ex ante mechanism -the process of management/board selection -should be given equal attention.
The final point made by this paper is that agencification in many transition countries was usually NOT driven by a need to grant some elements of the public sector a clear mission and autonomy to achieve it, but by a piece-meal sectoral efforts to increase professionalism in parts of the public sector. Under such circumstances, it is worth considering whether energy involved in sectoral agencification efforts would not bring more effect if spent on an overall reform of the institutional framework for the public sector and particularly on civil service reform.

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and which came into force on 30th September 1961, the Organisation for Economic Co-operation and Development (OECD) shall promote policies designed: -to achieve the highest sustainable economic growth and employment and a rising standard of living in Member countries, while maintaining financial stability, and thus to contribute to the development of the world economy; -to contribute to sound economic expansion in Member as well as non-member countries in the process of economic development; and -to contribute to the expansion of world trade on a multilateral, nondiscriminatory basis in accordance with international obligations.