Advanced Public Financial Management Reforms in South East Europe
  • 1 0000000404811396 Monetary Fund

Contributor Notes

Author’s E-Mail Address:

This paper aims to clarify possible systemic bottlenecks to the introduction of advanced PFM reforms in the SEE countries. It relates key fiscal developments to PFM reform processes over the last 15 years. PFM reform strategies must be realistic, with clear objectives and timetables, and with strong country ownership. Among the advanced reforms, some aspects of medium-term budgeting seems to be somewhat less challenging than performance-oriented budgeting, and it could be rational to make sure that there is solid progress in this area first. When developing performance budgets, countries should consider focusing initial efforts on the areas that are most suitable for performance management, such as education and health.


This paper aims to clarify possible systemic bottlenecks to the introduction of advanced PFM reforms in the SEE countries. It relates key fiscal developments to PFM reform processes over the last 15 years. PFM reform strategies must be realistic, with clear objectives and timetables, and with strong country ownership. Among the advanced reforms, some aspects of medium-term budgeting seems to be somewhat less challenging than performance-oriented budgeting, and it could be rational to make sure that there is solid progress in this area first. When developing performance budgets, countries should consider focusing initial efforts on the areas that are most suitable for performance management, such as education and health.

I. Introduction and overview

A. Introduction

The countries in South East Europe (SEE) have achieved significant improvements in public financial management (PFM) in the last 15 years. This group includes Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Kosovo, Macedonia, Moldova, Montenegro, Romania, Serbia, and Slovenia. The PFM improvements have provided crucial support for broader economic and political reforms, and all the countries in the region have come a long way in their transition to market economies. Slovenia became an EU member in 2004, Bulgaria and Romania did so in 2007, and several other SEE countries are actively pursuing membership. For all the countries, consolidation of ongoing PFM reforms and further improvements in PFM systems are important for realizing key economic and political goals.

Progress has been uneven in different reform areas. In some important areas, most of the countries have made significant steps towards good international practices. One example is the consolidation of government funds into treasury single account systems. In other areas, progress is more mixed. This is largely related to differences in countries’ capacities and levels of development. However, some of the reforms have proved to be very difficult for all the countries, even the more sophisticated ones.

None of the SEE countries have been able to introduce complete, functioning medium-term and performance-oriented budgets. For most of the countries, these reforms are important elements in the overall PFM reform strategy, and considerable resources have been spent on these efforts. However, in most cases the results have fallen short of expectations.

This paper aims to identify bottlenecks to the introduction of advanced PFM reforms in the SEE countries and proposes measures to overcome these obstacles. For each area, the objectives of the reform are outlined and the achievements so far are discussed. Possible systemic problems and bottlenecks are identified, and options for solving these are proposed. The focus is on medium-term and performance-oriented budgeting, but the findings in the paper should also be relevant for other types of advanced reforms, such as the integration of cash and debt management.

The second section of this paper gives an overview of PFM objectives and the status of PFM reforms in SEE countries, as well as key fiscal developments over the last 15 years. Section III discusses medium-term budgeting in more detail and Section IV discusses performance-oriented budgeting. Section V summarizes the findings of the paper and makes recommendations on strengthening advanced PFM reform processes.

II. Fiscal institutions and outcomes 1990–2006

A. PFM Objectives

Table 1 lists four major objectives for public financial management, and some key fiscal indicators related to each objective. When a country improves its PFM system, and thereby its capacity to pursue the main objectives, this should lead to improvements and changes in these fiscal indicators over time. However, fiscal developments are related to many different factors and there may be significant time lags between the introduction of institutional changes and subsequent improvements in fiscal indicators.

Table 1:

Main Public Financial Management Objectives and Indicators

article image

B. PFM Reform Agendas in SEE Countries

All the SEE countries are pursuing comprehensive PFM reform agendas. Most of these reforms contribute towards meeting the four major objectives listed in Table 1, but there is considerable variation in the more detailed objectives and priorities. Table 2 provides an overview of 10 important reform areas for the SEE countries. These are divided into basic reforms, which are necessary to achieve a minimum level of financial management capacity, and advanced reforms, which will enable countries to move towards international good practices. The objectives and benchmarks are consistent with broader benchmarking frameworks for PFM systems, such as the PEFA approach.2

Table 2:

Selected Public Financial Management Objectives and Benchmarks

article image

Table 2 does not attempt to provide a complete or generally applicable framework for assessing basic and advanced PFM reforms. The 10 reforms that are included in the table formed the basis for the IMF’s regional technical assistance program to the SEE countries from 2005–08, and were selected on the basis of proposals from the recipient countries and IMF country teams. The definition of reforms as basic or advanced was based on subjective judgments of the specific situation in these countries in 2005. The detailed formulation of such reforms, may vary significantly over countries and regions, and over time. The definition will also depend on the level of aggregation: modernization of a budget classification include basic components (economic, functional and organizational codes) as well as more advanced ones (program and funding codes).

C. Status of PFM Reforms in SEE Countries

Some SEE countries are near to completing many basic PFM reforms. For instance, Bulgaria, Croatia, Moldova and Slovenia have made good progress in this regard. Several other countries have accelerated their reform efforts in recent years, but have not yet reached the same level of maturity. Figure 1 gives a short summary of the status of public financial management capacities in each SEE country, related to the 10 objectives listed above.

Figure 1:
Figure 1:

Remaining PFM Reform Needs in SEE Countries 3

Citation: IMF Working Papers 2009, 102; 10.5089/9781451872491.001.A001

There has been less progress in implementing advanced PFM reforms. Figure 1 indicates that most of the countries need substantial or very substantial improvements in order to meet the benchmarks in many of the five areas defined as advanced reforms. The two areas where the needs for further improvements are highest are medium-term budgeting and performance-oriented budgeting.

D. Fiscal Objectives and Outcomes

This sub-section presents some key fiscal outcomes in the SEE countries, for 1995, 2000, and 2006, and relates these outcomes to the objectives and benchmarks listed in Table 2.

Aggregate fiscal control

Table 3a provides data for general government expenditures in SEE countries as a share of GDP, whereas 3b describes overall fiscal balances. The numbers in the tables refer to 1994, 2000, and 2006, or the closest years available.

Table 3a:

SEE Expenditure Levels

article image
Table 3b:

SEE Fiscal Balances

article image
Source: IMF staff reports, national authorities and World Bank staff estimates.1Data for consolidated central government.2Preliminary Outcome.3Serbia and Montenegro.

The data in Tables 3a and 3b indicate that most SEE countries have improved fiscal control during this period. In most SEE countries government expenditures as a share of GDP have been reduced or been fairly stable during this period. However, the shares increased in Croatia, Moldova and Serbia. More significantly, fiscal balances have improved significantly over the period, in particular from 2000-06. Most SEE countries now have modest deficits or surpluses. This is, of course, strongly related to positive international and domestic market conditions, but it also suggests that countries have succeeded in establishing a stronger degree of aggregate fiscal control.4 Data for public debt, which are not included here, also indicated that most SEE countries have been successful in strengthening their fiscal position.

Financial discipline

Table 4 provides indicators for financial discipline in six SEE countries. It is based on the PEFA assessment framework. Countries’ ability to implement the budget as planned is a key indicator of financial discipline. Their ability to contain and to manage arrears is another important indicator. The first indicator set compares total budget estimates with actual expenditures, for the last three years before the PEFA assessment was undertaken. The next set describes variations in sectoral spending compared to the budget. The third set provides information about arrears.

Table 4:

Indicators for Financial Discipline in Six SEE Countries

article image
Source: PEFA assessments.

2004 budget expressed in commitment terms. 2005 budget put on a cash basis.

data considered uncertain because not audited for the last four Fys.

data on the stock of arrears may not be complete.

The table indicates that budget estimates still are inaccurate in many SEE countries, and that there still are arrears, despite some improvements in recent years. We do not have comparable PEFA assessments for previous years, but there is strong anecdotal evidence that financial discipline was very weak in many countries in the 1990s. A review of IMF staff reports from this period provides several examples of much more significant discrepancies between initial budget estimates and final accounts data in many countries during this period, regarding overall spending levels as well as composition, and also much higher levels of arrears than today.

Efficient resource allocation

Table 5a and 5b compare the functional breakdown of expenditures in Albania and in Bulgaria, for 1995, 2000, and 2006, whereas Table 5c indicates the regional shares of GDP for public investment in 2001, 2003, and 2006.

Table 5a:

Albania Functional Breakdown of Expenditures

article image
Source: IMF staff reports.
Table 5b:

Bulgaria Functional Breakdown of Expenditures

article image
Source: IMF staff reports.
Table 5c:

Government Investments as Share of GDP in SEE Countries

article image
Source: IMF staff reports.

The tables indicate that the composition of public spending has changed considerably over the last 15 years. In Albania and Bulgaria, fewer resources are directed to general public administration and to defense, and more is spent on health, education and social protection. This pattern is also prevalent in many other SEE countries. In addition, most countries have realized a modest shift of resources from current spending to capital spending, although there have been reductions in others. These changes indicate that countries are making conscious efforts to realign public spending in line with political and economic priorities, and that the PFM systems are capable of facilitating this realignment. However, we know from country work that some SEE countries want to increase capital spending further, but that they do not have the capacity to do this, and that the new EU member states are struggling to absorb the available EU funding.

Cost-effectiveness of public spending

Cost-effectiveness implies that sector objectives, defined as outputs or outcomes, are achieved with as little use of resources as possible. It is challenging to achieve this, and also to assess whether it has been achieved. This is usually only possible in countries with advanced performance-oriented budgeting system. However, we can use more basic data to make initial assessments about whether there have been improvements in outcomes and whether costs have been reduced. The following two tables provide two examples from the health sector. Figure 2 presents data for infant mortality, which is an internationally accepted outcome indicator for the quality of health systems. Figure 3 presents data for average length of stay in hospitals, which is a common indicator for the productivity of hospital systems.

Figure 2.
Figure 2.

Infant Mortality

Citation: IMF Working Papers 2009, 102; 10.5089/9781451872491.001.A001

Figure 3.
Figure 3.

Average Length of Stay in Hospitals

Citation: IMF Working Papers 2009, 102; 10.5089/9781451872491.001.A001

Source: WHO/Europe, European HFA Database, November 2007.

The tables indicate that there have been improvements in health outcomes, as well as in hospital productivity, in most SEE countries over the last 25 years. In some countries these improvements have occurred at a steady pace over the whole period, but in many SEE countries the onset of the transition process in the 1990s led to visible set-backs both in outcomes and productivity, before performance recovered and continued to improve. For some SEE countries, health outcomes and productivity are still considerably lower than for most EU member states. One of the objectives of improved budget systems, as discussed in other sections of this paper, is to provide effective support to the continuation of important sectoral reforms: for instance, in the health sector. Country studies of case payment systems in Slovenia and Bulgaria indicate that these budget reforms have enhanced cost-effectiveness, but further analysis is required to draw clearer conclusions.

E. Assessment

As demonstrated in the preceding subsections, SEE countries have achieved considerable improvements in PFM systems and in fiscal outcomes over the last 15 years. While the PFM reforms are only one of many influences on the countries’ fiscal situation, they have had important impacts in many cases. Further improvements in fiscal outcomes will be influenced by further PFM reforms. While there is still a need to address some weaknesses in basic PFM systems, for instance to eliminate arrears and enhance basic budget realism, the focus is gradually shifting to more advanced reforms.

In particular, countries are putting more emphasis on the efficiency of spending and the cost-effectiveness of public services through medium-term and performance-oriented budgeting. The following sections will provide an assessment of the progress so far in these areas, and make suggestions for further improvements.

III. Medium-term budgeting

A. Overview

Most budget decisions have impacts well beyond the budget year. To make rational and efficient budgetary decisions, decision-makers must be aware of and consider the medium-and long-term implications. In line with this, many countries around the world have taken steps to extend the time horizon of their budget systems. The design and focus of these reforms, and even the terminology, varies between countries. This paper uses the term medium-term budgeting (MTB) as a broad, generic term that covers the many different approaches.5

The potential benefits of medium-term budgeting are well known. A well-designed and well-managed framework for medium-term budgeting will improve fiscal control, financial discipline, allocative efficiency and cost-effectiveness of service delivery; through greater clarity of policy objectives, more predictability in budget allocation, increased comprehensiveness and validity of budget information and enhanced accountability and transparency in the use of resources.

In practice, however, efforts to introduce medium-term budgeting have failed to realize the potential gains in many countries. Sometimes the medium-term budget process has ended up as a ritualistic, resource-consuming effort of little practical value. In some countries medium-term spending proposals are not effectively reconciled with the resources available during the budget period. As a result, the spending plans become mere wish lists and have limited impact on annual budget preparation or on medium-term resource allocation. Inefficient coordination with national or sectoral strategies, for instance public investment programs, is another common problem, as is the failure to ensure that the basic requirements for effective annual budgeting is met before attempting to extend the budget period.

So far, there has been little assessment of the impacts of MTB in transition and developing economies. The most comprehensive study [World Bank 2002] covers 13 countries in Africa. According to this study, only one of the 13 countries (Uganda) had been able to establish several key components of a functioning MTB system. The study found some limited empirical evidence that MTBs are associated with reallocation to subsets of priority sectors, but no clear evidence of improved macroeconomic balances, greater budgetary predictability or efficiency gains in sectoral spending. This is not surprising, given the limited levels of implementation of the new approach.

Among the transition economies in Central and Eastern Europe, Slovakia is the only country that has introduced all key elements of a functioning medium-term budgeting system. According to Slovak authorities, this system has strengthened long-term planning and simplified annual budget preparation. Russia introduced a three-year budget appropriation system for 2008-10, but the impacts of this are not yet clear.

B. Phases of Medium-Term Budgeting

A complete medium-term budget system includes a number of different steps. Some of these steps are easier to introduce than others. Most countries have taken a phased approach to the introduction process, which often takes several years. Usually, the different steps in the MTB process also develop over time. For instance, multiyear spending ceilings may initially be indicative and related to broad functions, but over time they evolve to become more authoritative ceilings for individual budget organizations.

A medium-term macrofiscal framework is perhaps the most essential building block for a medium-term budgeting system. In most countries, the extension of the macro-economic framework to cover at least three years and to include broader sets of fiscal parameters, is the first step towards MTB. To ensure that estimates are realistic, the macrofiscal framework must be updated regularly and be fully reconciled with budgetary decisions. For instance, decisions regarding new tax policies during the budget process should feed back into an update of the macrofiscal framework.

The next step is often to introduce multiyear expenditure ceilings. The purpose of such ceilings are to enhance predictability for line ministries, while containing aggregate expenditures at an acceptable level. Initially, expenditure ceilings tend to be fairly vague. They are often related to broad functional categories and are only indicative. In order to meet the objectives of enhancing predictability and fiscal control, the ceilings will usually need to be related to specific organizational entities, which can be held accountable for not exceeding the limits, and to have a formal status in the budget process. Unless the budget ceilings for the years beyond the budget year are the starting points for negotiating the budgets for the following years, they will tend to have limited impact.6

Thereafter, line ministries can be asked to develop medium-term spending plans that fit within the established expenditure ceilings. These spending plans should be based on relevant government and sector strategies. They should identify the costs of different programs, policies and activities within the ministry’s area of responsibility, to allow the responsible minister to prioritize among the different policy options within the sector, and the cabinet to choose between proposals for new policies and activities in different sectors. The quality of a sector spending program will generally depend on whether the country in question has introduced performance-oriented budgeting, which is discussed in a subsequent section of this paper.

To enable rational medium-term budget discussion, it is essential that sector spending programs separate the costs of existing policies from new spending initiatives. Some countries have developed advanced forward estimates systems to facilitate this process. With such systems in place, the roll-over of the budget for existing activities becomes a largely technical issue, and budget discussions can focus on new policies and spending proposals. Without this separation, budget discussions will tend to become very general, focusing on how much the total level of resources given to a specific sector needs to change compared to last year’s budget. It will be difficult to link this discussion to specific policy objectives in any transparent way.

An MTB provides an important basis for the coordination of current and capital spending. This is a weak spot in many budget systems, and budgets often fail to recognize the costs required for maintenance of existing capital and operational costs for new capital projects. At the early stages of MTB introduction, these costs can be captured by lump-sum allocations, whereas advanced MTBs can provide a detailed breakdown of the costs related to individual capital projects.

Over time, MTBs will change the character of the annual budget negotiation process. Traditionally, budget negotiations focus only on the budget year. With the introduction of MTB, the scope of the negotiations can be gradually extended to reflect impacts of policy decisions beyond the budget year. In an advanced MTB system, budget negotiations should effectively cover the whole MTB period. The extension of the negotiation period expands the scope for possible solutions and makes it easier to reach agreements.

Similarly, the coordination between the annual budget and the MTB documents will increase over time. In the early stages of an MTB reform process, the documents will tend to be separate and inconsistent. The MTB is prepared at an early stage of the budget process, and budget parameters are changed in connection with the annual budget without being reflected in an updated MTB. This means that the MTB figures may become largely irrelevant for following year’s budget preparation. Because this undermines the credibility of the MTB process, countries will at some stage need to begin updating the MTB during the budget year to ensure consistency. In an advanced MTB system, the MTB and the annual budget will often be a single, unified document, which is updated regularly throughout the preparation process.

C. Status of Medium-Term Budgeting in SEE Countries

All the SEE countries have taken some steps towards medium-term budgeting, but most of them are at a fairly early stage in this process. Box 1 gives a brief summary of the MTB systems in these countries, and Table 6 assesses the maturity of the systems compared to the seven steps described in the previous subsection. For each of the steps, the table indicates whether there has been no significant progress (score 0) or whether the country is at an initial, interim or final stage of implementation (scores 1–3). Appendix I explains the benchmarks that have been used in estimating the scores.

Table 6.

Medium-Term Budget Systems in SEE Countries

article image

Table 6 indicates that SEE countries have generally made quite limited progress in introducing MTBs. However, some countries seem to be more advanced than others. Slovenia scores fairly high on many parameters. One reason for this is that Slovenia has a two-year budget system, which meets many of the same objectives as a more traditional, three-year MTB, but not all of them. Macedonia also has some high scores, but these are to some extent based on recent changes in regulations that are yet to be fully tested.

Medium-Term Budgeting in South East European Countries

Albania introduced a medium-term budget program (MTBP) in 2001 and gradually improved it from year to year. Budget ceilings are set for three years. The ceiling for the first year is now largely binding, whereas out-year ceilings are indicative. The MTBP is approved by government and sent to Parliament for information. Neither the MTBP nor the annual budget specifies consistency between initial budget ceilings, medium-term budget allocations and final budget allocations, or explains changes in out-year budget ceilings from one year to the next.

In Bosnia and Herzegovina the state and entity MOFs prepare medium-term budget frameworks on a three-year, rolling basis. The entities have prepared MTBFs for some years, and the first MTBF for the state was prepared for 2006. Line ministries are required to prepare three-year estimates within predetermined ceilings (for the budget year in Republika Srpska and State, and for 3 years in Federation). Annual and medium-term budget ceilings are indicative, in particular at the state level. Consistency from year to year is not transparently monitored and reported.

In Bulgaria the fiscal framework is well developed and linked to the EU convergence program. Ministries present medium-term budget estimates before cabinet sets budget ceilings. 2008 regulations provided specific mechanism for linking the medium-term budget for one year explicitly to previous years’ medium-term budget, but implementation was incomplete.

In Croatia a multi-annual fiscal framework was submitted for the first time with the 2005 budget. This includes indicative three-year ceilings by budget user. Ministries prepare three-year budget estimates. There is no general costing of sector strategies, but ministries should prepare development programs for capital spending and transfers, and present the fiscal implications of legislative changes. The medium-term fiscal framework is updated each year, but there is no explanation of changes in spending ceilings compared to last year’s framework.

A medium-term expenditure framework was introduced in Kosovo for 2006-08. Expenditure estimates are broken down by five main groups, not by ministry or function. For 2007 there were ex-ante ceilings for current expenditure, but not for capital. For 2008 there were no ex ante ceilings for ministries’ budget proposals, so the MTB had limited impact on budget decisions.

The 2005 Budget Law provides the legal basis for medium-term budgeting in Macedonia: the government determines a set of strategic priorities, while budget users prepare three-year strategic plans. The MOF prepares a three-year fiscal strategy and then proposes spending limits for the next fiscal year. For 2007-09, the MOF set binding spending ceilings for 2007, ceilings for 2008 and 2009 were indicative. Line ministry strategic plans are required to be consistent with ceilings for the whole planning period. Fiscal framework for 2008–10 was submitted to parliament prior to the draft budget bill.

In Moldova in 2007 the medium-term expenditure framework was developed for 2008-10 period. The MTEF provides indicative ceilings for the MTEF period, with separate ceilings for current and capital spending. It also provides a set of policy proposals, which should be consistent with the National Development Plan. According to the MOF, the policy proposals and medium-term spending plans are fairly consistent with the budget ceilings in the MTEF. The MTEF is approved by the government and submitted to Parliament for information. It forms a basis for the annual budget preparation, but there are deviations between the MTEF and the annual budget figures. The MTEF is updated each year, but there is not yet a systematic tracking of changes in budget estimates between years.

In Montenegro the fiscal framework has been established. Budget preparation for the 2007 and 2008 budgets were based on a bottom-up approach. MOF plans to introduce a top-down procedure with budgets within specific ceilings for the 2009 budget. A first step towards medium-term budget was made in 2006 with the implementation of the new capital budget process.

In Romania the medium-term fiscal framework is quite well developed, and the capacity for macrofiscal estimates and analysis is being further strengthened. According to the 2002 Public Finance Law, the government should establish medium-term spending ceilings for the budget organizations, and the latter should provide medium-term budget proposals in accordance with these ceilings. In practice, this approach has not been fully implemented. Ceilings have been set too late to effectively discipline the budget process, and have been based on increments to last year’s budget. The medium-term estimates have been sent to Parliament for information, but the projections have had limited impact on annual budgetary allocations.

In Serbia, medium-term fiscal forecasts are provided. There are annual, indicative spending ceilings, and some indicative medium-term ceilings provided by functional classification. Ceilings are based on history and technical coefficients, but there are no explicit links between sector strategies and budget allocations

In Slovenia the MOF prepares a medium-term fiscal framework. There is a two-year budget process, with authoritative ceilings, but no ceilings for year 3 and 4 of the medium-term fiscal framework. At the beginning of the budget process, allowed levels of increase in development programs for years 3, 4 and 5 are determined. The line ministries prepare development programs that include four-year budget estimates. These are related to sector strategies, and primarily include capital spending. The budget is based on programs, but spending estimates are not stringently linked to specific policies and activities.

All the countries prepare comprehensive macrofiscal frameworks, which are generally updated at least twice a year. This is the strongest element of their existing MTB systems. Many of the countries are EU member or candidate countries, or (recent) IMF program countries, which ensures that macrofiscal projections and the related budget estimates are kept up to date.

Many of the countries prepare multiyear expenditure ceilings for the budget process, but in most cases these are too general to have any strong disciplining effects. In most of these countries the ceilings are only indicative, either by definition or because they are allocated to broad functional areas and not to organizations that can be held accountable for meeting the ceilings. In most SEE countries, the out-year ceilings are usually changed from one year to the next without any explicit explanation, so they are of limited value for planning.

There is considerable variation in the scope and quality of sector spending programs. In some countries, such as Albania and Macedonia, these have been key components of the MTB reforms. In Albania, the sector spending plans are seen as key input to national strategies, whereas in Macedonia there has been no comprehensive national strategy so far. Moldova initially developed their national development strategy largely independent of sector spending plans for the MTB process, but harmonization improved in the second iteration of the national plan.

Some countries have also made progress in using the MTB to improve coordination of current and capital spending. Although there is still much progress to be made, most have at least some indication of maintenance and/or operational costs in their MTBs.

For the remaining benchmarks, progress has been very limited. Countries have great difficulties in separating the costs of existing policies from new spending proposals. This is not surprising, given that most budgets still focus on financing institutions, not on implementing policies. None of the SEE countries have been able to develop comprehensive forward estimates systems. Because of this, MTB discussions tend to be fairly general discussions about total funding levels and not about specific policy priorities. In general, the MTBs have had little impact on the specific budget negotiations, which continue to have a one year focus. The MTB remains a separate document in all countries but Slovenia, and only one or two countries update the MTBs to reflect budgetary decisions.

D. Implementation Challenges and Possible Solutions

The objectives of the MTB reforms are not always clearly defined and understood or commonly agreed upon. The first phases of the MTB process, in particular the establishment of medium-term macrofiscal frameworks and indicative spending ceilings, have generally been agreed upon and understood by all the stakeholders. However, there seems to be considerable ambiguity about how to take the next steps. Many of the SEE countries do not have clear strategies and action plans in this regard, and even where such plans are in place, they are not always well understood and agreed among the stakeholders.

It is not obvious that the MTB model that has been applied in most SEE countries is the most appropriate and realistic. All the countries have received extensive advice on these reforms, and the models proposed have largely been based on recent practices in a small number of very advanced countries. However, in these countries themselves the new practices developed gradually. It could be that a slower and more organic development process would also be effective in the SEE countries. It is interesting to note that that the country that has had most success in this regard, Slovenia, has followed an evolutionary development path and that the (partial) MTB system in place is quite different from the general model recommended to the other countries.

The MTB models and reform strategies in SEE countries do not always reflect their different PFM priorities. One of the attractive features of an MTB system is that it can contribute towards all the four main PFM objectives mentioned in Section I. However, the details and the balance between different components can and should vary considerably between different countries. For countries that already have MTB, some emphasize that forward estimates should enable line ministries to plan their activities more effectively, others may be more focused on setting binding spending limits to contain expenditure.7 Such differences in focus are not clearly specified in SEE countries’ MTB plans.

In most SEE countries, the MTBs remain separate from the annual budget process. As long as this fragmentation persists, it seems unlikely that the MTB process will have significant impacts on the actual budget allocations. The fact that only Slovenia presents the annual budget and the medium-term budget estimates in the same document is quite striking, and suggests that more efforts could be put into creating a unified budget process and budget presentation in other countries. Instead of two separate documents, the budget document itself could include forward estimates, initially perhaps only for one additional year and only for parts of the budget, as is the case in Slovenia.

The efforts to introduce MTB may have taken the focus away from critical weaknesses in the annual budget process. In many SEE countries, there continue to be significant discrepancies between spending ceilings and subsequent budget requests. Budget discussions tend to be very general, have little policy focus, and be repeated from year to year. One improvement in many OECD budget systems in the 1970s and 1980s was the introduction of a separate, initial phase in the budget process that focused on “rolling over” the budget for existing activities from one year to the next. This exercise would be done prior to the first budget meeting of the cabinet, and would help identify the fiscal space for new activities. The “rolling-over” process also has its risks, however. Unless it includes a critical review of existing activities, it may result in the automatic retention of all these activities. None of the SEE countries have introduced this step; the rationale sometimes given is that the forward estimate system in the MTB should take care of this.

The authority of the out-year estimates is very weak in most countries. Given that even the budget year ceiling fails to discipline budget requests in many countries, it is not surprising that estimates for subsequent years are given little attention. However, this effectively undermines a major premise for the whole reform. Unless the out-year estimates are managed and maintained in a credible manner, the process effectively collapses to a traditional one-year budget process with some ritualistic estimates added on.

Any MTB reform will be complex and time-consuming, and will require considerable staff training and considerable capacity building. Even if countries sort out the challenges mentioned above, additional efforts will be required to successfully implement MTB. In general, the SEE countries seem to be well aware of the need for capacity building in finance ministries. However, in many countries it could be beneficial to put additional efforts into strengthening institutions and training staff in line ministries.

Finally, a note of caution in interpreting the survey results: reform success must be seen in light of objectives. If a country’s main objective for MTB actually is to establish a credible macrofiscal framework and to provide some indicative guidance on the budget process, many of the current MTB systems are actually reasonably successful. However, if countries are aiming for more complete systems of MTB, in line with stated objectives, there is obviously a long way to go. In many countries, there seems to be significant discrepancies between the official, stated objectives and the objectives that are actively pursued. This may be because countries accept certain objectives as a result of pressure from donors and advisors, but then do little to actually implement these.

Tentative summary of recommendations

  • Establish a clear strategy and timetable for MTB reforms, and ensure that this strategy is well understood and generally agreed among all stakeholders.

  • Avoid attempts to replicate MTB systems from other countries, focusing instead on developing approaches that are consistent with the countries’ specific PFM priorities and institutional capabilities.

  • Measures to improve the annual budget process, for instance by clearer phasing of budget decisions and developing the capacity for costing of policy proposals, are often prerequisites for successful MTB reform.

  • The MTB and the annual budget should be prepared through a unified development and negotiation process, should be consistent and presented in a single document.

  • The out-year estimates must be the starting point for subsequent years’ budget preparation. As a minimum, the MOF should provide detailed explanations for any differences between budget ceilings and previous out-year estimates.

  • MTB reforms will require considerable capacity building in line ministries.

IV. Performance-oriented budgeting

A. Overview

As for MTB, the potential benefits of a performance-oriented budgeting system are well established and well understood. Better information about outcomes and effectiveness helps allocate budget resources to the programs where the benefits are highest, and helps address equity concerns. Monitoring of cost-effectiveness helps reduce the costs of delivering specific government services and public goods. Some countries have also found that a performance-oriented budget helps achieve fiscal control in the long run, because the mechanism provides a well-founded basis for budget realignment and reduces pressures to allocate resources to less productive programs.

However, whereas the benefits are well understood, there is still considerable uncertainty about how to go about implementing a performance-oriented budget system. Most countries have encountered significant difficulties and have found that it can take many years to move through the different steps. Many OECD countries established program classification structures in the late 1960s or early 1970s. A few of them have recently reached advanced stages of the process, and many have not come this far. Not surprisingly, efforts to introduce performance-oriented budgeting in transition and developing countries have often met great difficulties.

For performance-oriented budgeting system to become effective, agencies must be given some freedom in determining how to meet their stated objectives and managing their budgets. For this, good reporting practices as well as appropriate internal control must be in place to prevent and detect fraud and error. In this system, the control over the agencies is not based primarily on the inputs it uses, but on review of the activities carried out and results achieved. It is important to emphasize that more freedom implies more responsibility and accountability. Agencies should not be given more authority if there are no appropriate controls and accountability mechanisms.

Most SEE countries have taken initial steps to strengthen the performance orientation of their budget systems. Some countries have mainly focused on establishing basic program budget structures, whereas other countries also have developed indicator sets related to the programs, and begun using the performance information for budgetary decision-making. In all the countries, the processes have met a number of obstacles, and implementation has generally been significantly slower than initially expected.

B. Phases of Performance-Oriented Budgeting

A complete POB system includes a number of different steps. The exact phasing may vary, but countries tend to follow fairly similar development paths. Each step involves a number of complexities and challenges, and will in most cases depend on successful completion of more basic reform steps. For instance, it is not possible to develop an effective performance (output or outcome) budget in the absence of a well-functioning input budget.

The first step is usually to introduce a programmatic budget classification. This involves dividing the budget into programs that aim at delivering specific goals and objectives (outputs and outcomes). In many countries programs are related to the high-level functional budget classification (COFOG). For instance, the COFOG function health care may be divided into programs aimed at addressing specific health problems.

A second step is usually to define indicators that describe the inputs to each program and the related activities. What actions are planned for the achievement of objectives and which resources are required to carry out these activities? Some inputs may be quite generic, staff resources for instance, whereas others are program-specific. Activities also tend to be program-specific.

Outcome and output indicators should be defined at the beginning of a program and set for the whole duration of the program. For many programs the focus is on output indicators, since outcomes are usually difficult to measure. There may be long delays between outputs and outcomes or it may be unclear to what extent an outcome can be attributed to a particular organization. Very often an outcome depends on the work of several different organizations and external factors. Intermediate outcome and output indicators are used for monitoring progress in program implementation, as well as for revision and adjustment of indicators.

Once the program structure and related indicators have been established, it is important to ensure that all activities contribute to the established objectives. Governments will often be involved in a number of activities that have a historical rationale, but limited impact on current program objectives. It is important to review such activities to assess whether they should be discontinued, or restructured to give a clearer contribution to specific objectives. It may also be necessary to revisit the initial program definitions.

Performance indicators are used to measure effectiveness, efficiency and economy. These indicators are generally measures of the relationship between input indicators and outcome or output indicators. For performance indicators to be valid and reliable, it is essential that the programs themselves, as well as input, activity, output and outcome indicators are well-defined.

Once performance information is available, it can be used to support management decisions regarding the allocation of funds. At this stage, performance information is usually only one of the factors influencing the funding decisions, and there is usually no one-to-one relationship between performance and funding. The decision mechanisms can be more or less formalized and stringent. However, in order to use performance information consistently, the budget process should include explicit mechanisms for the collection of estimates and results for budget program performance.

In some countries, a final, advanced step is to create stringent funding decision rules based on performance indicators. Such mechanisms include purchaser-provider systems, where the central government purchases specific services according to pre-determined prices, for instance health care services, from agencies and other providers. In those few countries that have established such systems, they are often combined with more discretionary funding mechanisms. Systems that include direct performance-funding links are very challenging to implement, and are not necessarily part of all performance budget reforms. Some countries have decided not to pursue such mechanisms at all.

C. Status of Performance-Oriented Budgeting in SEE Countries

Most SEE countries have also taken initial steps towards performance-oriented budgeting. We carried out an initial country survey in 2006 to identify to what extent and in which areas countries have introduced performance budgeting mechanisms in their budget process. The results were updated in 2008. Box 2 gives a brief summary of the country systems in these countries, and Table 7 assesses the systems compared to the steps described above. The table indicates whether there has been no significant progress (score 0) or whether the country is at an initial, interim or final stage of implementation (scores 1–3). Appendix II explains the benchmarks that have been used in estimating the scores.

Table 7.

Performance-Oriented Budget Systems in SEE Countries

article image

Table 7 demonstrates that most SEE countries are at an early stage of developing and implementing performance budgeting. A few countries are introducing elements of output oriented budgeting, but many are still struggling with the preparation of programs. None of the countries, except for Bulgaria and Slovenia in the Health Sector, have developed any mechanisms where the funds provided are directly related to the results achieved.

Almost all the countries have introduced programmatic budget classifications, but their quality varies considerably. In many countries, programs are designed to primarily describe an agency’s activities, rather than its operational objectives. Because of this, there is often significant overlap between programmatic and organization classifications. In many places, the assumption seems to have been that there should be at least one program for each existing organization, and at least one subprogram for each department in this organization.

Performance-Oriented Budgeting in South East European Countries

In Albania, ministries are required to define missions and programs and to describe a policy for each program, to set policy objectives for three years, to set policy standards, and to provide program expenditure and investment plans. Programs and main indicators are included in the MTBP document. So far, most program data are activity-related. There is no structured collection of data on performance or efficiency, or systematic use of this for decision-making.

In Bosnia and Herzegovina budget organizations present spending programs with the budget, but there is no formalized program classification structure. There is no systematic collection or use of performance indicators in the budget process.

In Bulgaria, the National Assembly, the Council of Ministries, all ministries and all state agencies have developed a program structure for their activities, including performance information, i.e. output and outcome indicators with target values. Budgets are appropriated according to programs and economic classifications, but accounting systems do not allow proper tracking by program. Budget regulations provide for systematic production and collection of performance data.

In Croatia a program structure has been established. The ongoing efforts to introduce programmatic budgeting have had limited impact so far. There is no clear link between the government’s priorities and budget allocations and there is little emphasis on spending efficiency. Work to establish performance indicators started in 2008.

In Kosovo program structure exists but is largely organizational. Budget documents give some references to broad objectives and activity targets, but do not define specific performance or efficiency indicators. There is no systematic reporting on performance.

In Macedonia program structure is established. Since 2005, all budget users are required to prepare strategic plans and to submit them with the budget circular. So far there are no performance indicators. To separate costs of existing and new activities, sector budget estimates could be based on specific activities, not just programs.

In Moldova a program structure now covers most ministries. Programs are based on different conceptual approaches and are often activity descriptions. The development of sector and ministry strategies will provide inputs to the budget program structure, which needs to be further developed.

In Montenegro a program structure has been introduced for several budget units, but budgeting has not yet changed for these units. Program classification is not introduced in the Treasury accounting system.

Romania has established a program structure for all institutions and entities. For the 2007 budget, the MOPF issued an updated budget circular, requiring ministries to provide information on strategies, outcomes, efficiency and performance information. However, compliance was mixed and the quality of data generally unsatisfactory. For the 2008 budget, the requirements were compromised due to the reorganization of government. There are no clear links between results and budget decisions.

In Serbia a program structure for five pilot ministries was established in the 2007 budget. There are plans to extend to others subsequently. Introduction of performance indicators and performance orientation is a long term effort.

In Slovenia a program structure is established. The budget contains a number of activity indicators, but there is limited focus on results and performance information is not systematically used in the budget process. An MOF task force is developing example performance budgets for some agencies and programs. There is a purchaser-provider system for hospital financing. Due to phased introduction, much hospital financing is still based on traditional cost recovery, but fees for service become more important each year.

In addition, all the countries have developed some activity and input indicators. This is fairly easy to do, partly because programs tend to be based on organizations. Accordingly, many activity and input indicators primarily describe the traditional activities of the agencies.

Three countries (Bulgaria, Macedonia and Slovenia) have developed reasonably comprehensive indicators for outputs, but there are few outcome indicators. In many countries the term “outputs” is used, but the indicators related to this term in many cases describe activities, not outputs. For instance, for a financial inspection service there may be indicators for the numbers of controls it carries out (an activity) but not for the number of errors that it discovers (an output) or for the improvement in financial management quality (an outcome).

Due to the lack of well-defined output and outcome objectives and the organizational focus of the budget programs, realignment between the different elements in the result chain is often lacking or unclear. It is often difficult to see whether and how current activities contribute to stated objectives. Even in countries with comprehensive program structures, some programs lack clearly stated objectives and indicators. Where objectives are defined, it may still be very unclear how the different activities contribute to achieving these.

The lack of a clear result chain, with missing elements and unclear relationships, also makes it difficult to define efficiency indicators and to systematically assess program efficiency. There are some ad-hoc attempts to identify efficiency indicators in a few countries, but this is not done in a comprehensive or systematic manner in any of the SEE countries.

There is anecdotal evidence that performance information may have some impact on decisions in a few countries. Line ministries are in principle required to include performance information in their budget submissions and justifications, but the observance and enforcement of such provisions is often weak. Even when some information is available, it is in many cases ignored at the time of decision-making.

In two countries, Bulgaria and Slovenia, payments from the Health Insurance Fund to hospitals are largely based on case payment schemes, based on sets of prices for specific health services (Box 3). These are so far the only examples of advanced performance-based payment and budget schemes in the region. In both countries, the health authorities report that these schemes have had positive impacts on productivity and resource allocation in the hospital sector, but there are also challenges related to the price-setting mechanisms, to avoid incentives for suboptimization and misreporting of activities.

D. Implementation Challenges and Possible Solutions

In all SEE countries, progress in implementing program and performance-oriented budgeting has been slower than expected. There are different reasons from country to country, but common obstacles encountered include a capacity shortages, especially in line ministries, a reluctance of staff to adopt changes, underestimation of time and resources required, and in many cases, the existence of more pressing priorities or more basic reforms to be put in place. It is apparent from Table 7 that there is still a long way to go until a performance-oriented approach is fully integrated into the budget process.

As for medium-term budgeting, there is considerable uncertainty about the purpose of ongoing program and performance budgeting reforms in many SEE countries. A few countries have concrete and comprehensive reform programs in this area. However, in most countries this is very much an ad hoc effort. In many ways, the performance budgeting reforms themselves suffer from the same weaknesses as the budget programs that are defined through the performance reforms: unclear objectives, excessive activity focus and unclear result chains.

Many SEE countries have found it difficult to establish a clear operational understanding of performance budgeting concepts among the budget organizations. As mentioned, it is very common that outputs are confused with activities. It seems even more difficult to establish a clear and agreed understanding of concepts related to outcomes and efficiency. These concepts are in themselves quite complex, and may be difficult to operationalize even in countries with a more established performance focus in their budget systems. For budget systems with a strong input and activity focus, a shift to an outcome focus represents a dramatic change in administrative culture and priorities, and may be almost insurmountable.

One generic challenge for performance-based budget systems is that officials are reluctant to take responsibility for results that they do not control. This reluctance is understandable, but it is a major factor in the continued focus on activities, instead of outputs and outcomes. In organizations with a strong legalistic tradition, combined with a culture of risk aversion, it becomes particularly difficult to shift focus. This seems to be an issue in many SEE countries.

It is important to establish a clear understanding of how and to what extent managers will be held accountable for results. For instance, while outcome indicators are useful for measuring the progress of a government program in the longer term, it is generally too crude a measure to use directly for the annual performance evaluation of a manager. The outcome indicators will have to be supplemented with indicators related to other parts of the result chain, to get a clear sense of the effectiveness of the organization and its manager. This performance management framework should also recognize that managers need to take risks in some areas, and that in some cases these risks will lead to negative results.

A common misconception, and another reason that budget organizations and managers are reluctant to embrace performance-oriented budgeting, is the belief that budgets will be allocated mechanistically on the basis of performance compared to targets. This is generally not the case. There is a potential to define fairly direct performance-payment links in some sectors, for instance for hospital services. However, in most sectors performance information is only a part of the basis for decision-making. Even decisions about whether under-performance should lead to decreased or increased financing will usually require an analysis of the reasons for this underperformance.

Action plans for the introduction of performance budgeting have, in many cases, been unrealistic. These plans have sometimes been developed to try to demonstrate concrete impacts during the action plan period, which often is limited to one to two years. Some countries have ended up with several consecutive short-term plans for performance budgeting. Together, these plans may cover several years, but because of the emphasis on short-term results little has been achieved. A better alternative would be to have a realistic long-term strategy, which could cover a five-to-ten-year period and help realize some real gains during this period.

Some countries have embarked on performance budget reforms prematurely. Because performance budgeting adds many new dimensions and significant complexity to a traditional input budgeting process, it is difficult to implement. An absolute minimum requirement for introducing elements of output or outcome budgeting should therefore be the existence of a well-functioning input budget system. There are no examples, neither in the SEE region nor internationally, of countries that have been able to make a rapid transition from an ineffective input budget system to a well-functioning performance budget.

Successful introduction of performance budgeting requires parallel reforms in public sector management and human resource practices. Some countries have attempted to strengthen output control, while retaining traditional input controls. It is very difficult for managers to improve their efficiency or effectiveness, if they do not have the authority to take decisions regarding allocation of resources or deployment of staff. Because of this, they will naturally be reluctant to accept any increased responsibility for the performance of their organizations.

Effective mechanisms for accountability and control are also critical. When managers are given increased autonomy, it is essential to have well-functioning accounting and reporting mechanisms, to ensure that they are held accountable for their results and for abiding by regulatory provisions, in a transparent fashion. Some OECD countries have seen accounting reform, including the introduction of accrual accounting, as an important pre-requisite for performance budgeting and increased managerial accountability.

In most SEE countries, performance-oriented budget reforms are broad, but quite shallow. As indicated in Table 7, many countries have established budget programs for all parts of the budget, and defined fairly comprehensive input and activity indicators. It seems very difficult to get beyond this stage and get closer to the actual performance of budget programs. Because of this hurdle, program managers and consultants may end up going back to their initial program definitions and indicators and spend the reform project resources on perpetual iterations around these concepts.

There could be benefits to a narrower, but deeper approach. Instead of covering all budget programs, further reform efforts could focus on a select number of budget programs that are particularly well suited for performance budgeting, and try to develop a complete performance management framework for these programs, including the budget decision stage. This could focus on programs where the outputs are well defined, easy to observe and measure, and where there either is little variability in quality, or quality can be readily monitored. This approach would provide important learning opportunities, and it could create some models that could then be emulated in other sectors. The financing schemes for hospital services in Bulgaria and Slovenia are interesting examples of this narrow and deep approach to performance budgeting.8

In many countries, the authorities indicate that any movement towards increased performance-orientation in the budget system is positive, even if the short-term effects are very modest. There is some merit to this argument. Due to the many cultural, administrative and technical challenges discussed above, the introduction of performance budgeting will require considerable learning over a long period of time, and it is useful to start this learning process. However, this objective can probably be met more effectively with a long-term program for staff training and gradual systemic change, than through a series of unrealistic, short-term attempts at premature introduction of performance budgets that ultimately will fail.

Tentative summary of recommendations

  • Countries should have well-functioning input budget systems before embarking on any output or outcome budget reforms.

  • There must also be adequate arrangements for accounting, reporting, internal control and internal audit, to ensure that managers can be held effectively accountable for their decisions.

  • Performance budget reforms must be combined with reforms in public administration and human resource management, to ensure that managers have the necessary authority and autonomy.

  • The purpose of performance budgeting reforms should be clearly defined and well understood by all stakeholders.

  • Countries that have made initial progress in introducing program budgets, and where the pre-requisites for more extensive performance budgeting are in place, could base further efforts on developing complete performance budget mechanisms for select sectors, rather than trying to cover all sectors at the same time..

  • Performance budget reforms will always take several years, and should be guided by comprehensive long-term strategies and action plans.

  • The basic concepts and approaches must be clearly defined and commonly agreed upon among all the different stakeholders, including the line ministries and agencies.

  • There should be a clear definition of how and to what extent managers will be held responsible for the results of their organizations. In most cases, annual performance evaluations should be based on elements from different parts of the result chain, not just on the final outcomes.

There should also be a clear framework in place for how performance information will influence future funding decisions. This framework will vary between different budget programs. For most programs it will be important to combine performance indicators with specific analysis and judgment, but for a few programs it may be possible to establish direct performance-funding links.

V. Conclusions

SEE countries have achieved very important improvements in many areas of public financial management over the last 15 years. These improvements have had positive impacts on fiscal outcomes and on the transparency and quality of public expenditures. The priority has generally been to establish basic building blocks for modern public financial management, and improvements in fiscal control and financial discipline have been impressive.

Many SEE countries have also been pursuing more advanced public financial management reforms: medium-term budgeting and performance budgeting, in particular. Naturally, these are at an earlier stage of implementation than the basic reforms. Many countries have encountered bottlenecks, and the reforms efforts have in some cases stalled or are proceeding at a very low pace. As a result, these reforms have so far had limited impacts. Sector reforms have led to improvements in outcomes and productivity, for instance in the health sector, but the general public financial management systems have not yet made any significant contributions to these types of improvements.

When assessing progress in advanced PFM reforms, it is important to keep in mind the potential disconnect between formally stated reform objectives, and the actual expectations of these reforms. Reforms have often been driven by external stakeholders, with limited country ownership. Not surprisingly, this has sometimes led to unrealistic objectives and timetables, and actual implementation has lagged considerably behind.

SEE countries have often been quite successful in introducing some of the mechanics of medium-term and performance-oriented budgeting. Procedures have been put in place, data produced and documents prepared. Most SEE countries now have a document called a medium-term budget, as well as a programmatic budget structure with some indicators linked to the programs.

The weaknesses tend to lie in the quality, validity and usefulness of the new processes, data sets and documents. Although countries may have a notional medium-term budget, the contents are often updated in an ad-hoc manner, with limited explanation and little documentation, and the MTB has little impact on annual budget decisions, particularly beyond the budget year. Budget programs often coincide with existing organizational structures, and objectives and indicators largely reflect existing activities. Many reform attempts have degenerated to ritualistic exercises, with limited impacts on organizational cultures and behaviors.

Generally, the top-down dimensions of the advanced PFM reforms have been the most successful so far. Most SEE countries now have quite well-functioning medium-term, macrofiscal frameworks. In some countries, there has also been considerable progress in defining broad government budget priorities, and in determining high-level budget program structures. These mechanisms are typically under direct control by one or two central bodies, such as the Ministry of Finance and the Cabinet office, and these bodies are motivated and quite well resourced to handle the reformed mechanisms.

In contrast, the bottom-up dimension of the advanced PFM reforms tend to be much weaker. Sector strategies continue to be weak, and very uneven across sectors. Effective reconciliation of strategies with fiscal envelopes is lacking in most cases. The quality of sector objectives, activity costing and performance indicators is variable. These activities are by their nature strongly decentralized, require extensive information and training efforts across the whole government, and the possibility of establishing effective centralized control over the reforms are limited.

Although there are many similarities between the two types of advanced PFM reforms discussed in this paper, there has been somewhat more progress in the area of medium-term budgeting than in performance-oriented budgeting. One reason is that the incentives are stronger. Medium-term budgeting is seen as an important element of ensuring adequate fiscal control and improving broad resource allocation, which have been clear priorities in all SEE countries, whereas budget performance-orientation tends to be seen as less urgent. At the same time, the MTB models are clearer. Looking beyond the SEE region, the number of advanced countries that have reasonably effective medium-term budget systems is somewhat higher than the ones that have reached an advanced stage of functional, performance-oriented budgeting.

If one agrees that progress in advanced PFM reforms in SEE countries has been modest so far, the obvious question is what to do with this. There are at least three fundamental options:

  • Continue with the current approach, accepting that results will only be visible in the long term;

  • Increase resource allocations and try to accelerate implementation;

  • Abandon some reforms and allocate the resources to the reform activities with the highest and/or most immediate payoff.

The responses to this question will clearly be different across different countries and sectors. Often, the most appropriate response could be a combination of the three approaches mentioned above.

Firstly, it is essential to have realistic reform strategies with clear objectives and timetables, and with strong country ownership. A realistic reform strategy needs to reflect the fact that fundamental changes in political and administrative cultures may take 10 or 20 years to realize. Such changes will often require series of inter-connected reforms in several areas. Most OECD countries have learnt, from experience, that efforts to improve performance orientation in the budget need to be combined with fundamental changes in management approaches, human resource practices, accounting and fiscal reporting; in order to get beyond the mechanistic and ritualistic stage.

Secondly, it is very difficult to address all issues at the same time, in particular as reform efforts become more challenging. Countries could benefit significantly from avoiding too many parallel reforms, and overly broad reforms.

  • If it is correct that medium-term budgeting is somewhat less challenging than performance-oriented budgeting, it could be rational to make sure that there is solid progress in this area first, before embarking on ambitious performance budget reforms.

  • When developing performance budgets, it could make sense to focus efforts on areas where the likelihood of success is highest, instead of allocating resources across the budget. Education and health are generally suitable for performance budgeting, and constitute large parts of many budgets. As mentioned, the health sector is the only sector in which any SEE country has made tangible progress in introducing advanced performance budgeting mechanisms.

Thirdly, decisions in this area should be informed by results, not just hopes and ambitions. If ongoing reform efforts fail to produce the planned results, countries should not automatically assume that an extension or a follow-up project will help resolve the problems. If ongoing reform efforts fail to produce the planned results, countries should respond by either cancelling or radically restructuring reform programs rather than automatically extending them as some countries have done.

Appendix I: Categorization of Medium-Term Budget Systems

article image

Appendix II: Categorization of Performance-Oriented Budget Systems

article image

Eivind Tandberg was the IMF’s regional advisor in the SEE countries from 2005–08. Mia Pavesic-Skerlep works as an interpreter for the IMF in Ljubljana. The authors are greatly indebted to representatives of the SEE finance ministries, who have provided detailed comments to previous drafts of this paper, and to colleagues in the IMF and other international organizations, in particular Richard Allen, Michel Lazare, Ian Lienert, and Tej Prakash.


Public Expenditure and Financial Accountability (PEFA) assessments are based on 31 main indicators with more than 100 subindicators. Most of the benchmarks in Table 2 are reflected in PEFA subindicators.


The shading in the figure indicates areas where there is still need for some improvement (light grey), substantial improvement (dark grey) or very substantial improvement (black) to achieve the benchmarks in Table 2. The fields without any shading reflect areas where countries have completed or are near to meeting the benchmark in question.


There may still be macroeconomic reasons for some of the countries to maintain an even tighter fiscal stance, for instance current account imbalances.


Some of the literature uses this term in a more narrow sense, to describe budget systems with multi-year appropriations.


This paper uses the term ‘binding’ budget ceiling when the out-year ceilings are the starting point for negotiations. It does not require that the out-year ceiling is the final budget for that year.


For instance, in the Australian MTB system, the MOF is involved in discussing line ministry forward estimates at a detailed level, whereas in the Swedish system the MOF is mainly concerned with the aggregate spending limits and does not get involved in the detailed sector estimates.


In both countries, the case payment schemes were a result of reforms within the health sector, and the Finance Ministries were not very active in promoting the new mechanisms.

Advanced Public Financial Management Reforms in South East Europe
Author: Mia Pavesic-Skerlep and Mr. Eivind Tandberg