Skip to main content

Medium-Term Budget Frameworks

Go Search
New Atlas
Atlas, A-Z
Atlas Maps
MPP/MPA Programs
Core Topics
Illustrative Courses
Topic Encyclopedia
Concept Dictionary
Career Tips
Best Practices Project


Practice Advice on Budgeting and Financial Management

Medium-Term Budget Frameworks (OECD and IMF) 

Summary Advice (OECD): The OECD advises that governments should develop medium-term budget frameworks because they form the basis for achieving fiscal consolidation.

Summary Advice (IMF): The most appropriate model of a MTBF (medium term budget framework) for a particular country or context depends on the point at which policymakers strike the balance between the competing objectives of multiyear budget planning. Aggregate expenditure ceilings tend to promote multiyear expenditure discipline more effectively, ministerial ceilings are more effective at facilitating multiyear expenditure planning, and forward estimates provide increased certainty that specific policy outcomes will be achieved.

Main Points (OECD): The frameworks should clearly state the government’s medium-term fiscal objectives in terms of high-level targets such as the level of aggregate revenue, expenditure, deficit/surplus and debt. They then need to operationalise these high-level targets by establishing hard budget constraints for individual ministries and programmes over a number of years. This lends stability and credibility to the government’s fiscal objectives. The OECD has identified seven key institutional features that play a key role in order to effectively control public expenditure. The establishment of medium-term budget frameworks is one of these features. 

The OECD provides the following advice on medium-term budget frameworks: 

  • By their very nature, high-level fiscal targets are set in a medium-term context. They aim to achieve a certain fiscal outcome over a number of years. Budgets are however enacted for a time period of one year, and are notorious for their short-term focus. This shot-term time horizon is often criticised for impeding effective expenditure management; decisions on resource allocation are said to be made on an ad hoc or piecemeal basis with the implications of past and present decisions beyond the next year being neglected. Medium-term budget frameworks aim to bridge this gap. Their successful implementation has been nothing short of a “cultural revolution” in government.
  • The level of detail of such frameworks varies from country to country, but they generally mirror the format of the budget, i.e. they are at the same level of detail as the annual budget. This means that a formal framework or hard budget constraint exists for each and every appropriation, most often for three years beyond the current fiscal year. These are rolling frameworks that are presented with the budget each year. This has increased the effectiveness of planning and eased the annual budget process. These frameworks are not, however, enacted into legislation; they are planning documents that reflect the political commitment to fiscal discipline.
  • Thes are living documents. The fact that a three-year budget framework is in place does not mean that no changes can be made to the document. In fact, shifting appropriations within ministries is key to successful fiscal discipline. It is imperative, however, that all such changes be clearly depicted and explained, i.e. whether the changes are the results of changed economic circumstances or new policy decisions. Most countries publish detailed reconciliations between year-1 in the previous year’s framework and the current budget proposal. 
  • The frameworks also serve to deter expenditures by illuminating the budget implications of decisions in next year’s budget whose expenditure may not be fully reflected in the budget. This can refer to: the operating costs of various capital projects being launched; programmes that come into effect late in the budget year thus not exposing their full costs in the initial year; programmes whose spending implications may not be fully reflected under the circumstance prevailing during the budget year but will become more actual in out-years. These are all classic examples of budgeting games in member countries, which the medium-term frameworks aim to end. 
  • Medium-term frameworks enable agency managers to be in a better position to plan their operations as they have some indicative level of funding beyond the next budget. This is especially relevant when resources are being reduced. Many downsizing options involve more than one year in order to reap the full benefits. Prior to the advent of medium-term frameworks, such options were often not considered as the time horizon only extended to the next budget year.

Problems with medium-term budget frameworks 

The OECD recognizes that medium-term frameworks are themselves not without their own problems. The United Kingdom was a pioneer in the area of multi-year budget forecasts in the 1960s and 1970s and encountered significant problems. These problems can be divided into three groups; most of these problems have been experienced by other member countries as well. 

  • There was a tendency to overestimate the growth potential of the economy when making the multi-year budget forecasts. This made excessive resources available in the forecast period and created upward pressure on expenditure.
  • Ministries and departments viewed resource allocations in the forecast period as an entitlement. This made subsequent downward revisions difficult, even when it became clear that the basis on which the allocations were made was not correct.
  • The multi-year budget forecasts were made in real terms rather than in nominal terms. In the 1970s, when economic growth subsided and inflation accelerated rapidly, expenditure forecasts were adjusted automatically for increases in prices while revenues suffered. This created further pressure on public finances.

These experiences caused many to view medium-term budget frameworks with some suspicion. It must, however, be observed that the early medium-term budget frameworks took place in an environment of rapid expenditure growth. Regardless, action has been taken to rectify the specific problems identified above. First, member countries are systematically making use of more “prudent” economic assumptions in order to avoid having excessive resources made available. This has tended to eliminate the second problem identified above. Third, medium-term budget frameworks are now invariably made in nominal terms, not real terms.

Main Points (IMF):

An MTBF is a set of institutional arrangements for prioritizing, presenting, and managing revenue and expenditure in a multiyear perspective. Such a framework enables governments to demonstrate the impact of current and proposed policies over the course of several years, signal or set future budget priorities, and ultimately achieve better control of public expenditure. An MTBF, therefore, does not refer solely to the actual numerical multiyear revenue and expenditure projections and restrictions presented alongside a given budget. Rather, an MTBF comprises all the systems, rules, and procedures that ensure the government’s fiscal plans are drawn up with a view to their impact over several years.

Countries introduce MTBFs for a number of reasons. Looking beyond the annual budget cycle can improve public financial management outcomes by:

  • Ensuring better control over the evolution of the aggregate fiscal position,
  • Promoting a more effective allocation of expenditure between sectors and priorities, and
  • Encouraging more efficient use of resources by budget managers. 

These objectives are, to some extent, mutually reinforcing. Strengthening the emphasis on achieving sustainable public finances creates incentives to prioritize policies so that the most important ones receive funding. Producing detailed multiannual budget scenarios improves the administration’s understanding of the cost drivers in different sectors and of the resource requirements of different policies, enabling government to set more credible multiyear ministerial expenditure plans, which, in turn, enable ministries to plan their activities in a more efficient way. The remainder of this section examines how effective MTBFs contribute to the realization of each of these objectives.

The tension between these objectives arises because the future demands on government are inherently uncertain. All governments find it difficult to commit themselves credibly to a comprehensive and detailed multiyear expenditure plan. Although a well-designed MTBF reduces uncertainty about future spending requirements, unforeseen pressures will inevitably arise. These pressures will have to be accommodated by increasing total expenditure, revising the allocation of expenditure, or leaving a large proportion of expenditure unallocated over the plan period, or through a combination of these measures.

This tension between competing objectives is reflected in the design of different MTBF models. Three broad approaches to medium-term budget planning can be identified:

  1. No MTBF. No multiyear expenditure and revenue estimates are presented alongside and on the same basis as the annual budget. These countries may produce aggregate fiscal or budgetary projections (such as required by stability or convergence programs); however, these documents are not integral to the budget documentation and do not constitute an ex ante framework for budget preparation.
  2. Indicative MTBF. The multiyear expenditure and revenue estimates presented with the annual budget are intended to reflect the future costs of current policies and decisions but are not intended to bind future policies and decisions. These medium-term revenue and expenditure estimates are reset every year, without any reconciliation with the estimates presented in the previous year. Because they can be revised without any consequence, indicative frameworks are often comprehensive in their coverage and provide considerable detail about the composition of expenditure by economic, administrative, or program category. However, they provide relatively little certainty about future expenditure at either the aggregate or the detailed level.
  3. Binding MTBF. The multiyear expenditure and revenue estimates presented with the annual budget are intended to both reflect the future costs of current policies and bind future policy changes. However, as discussed below, the nature, categorization, level of detail, coverage, and frequency of policy revisions of the medium-term commitment vary substantially across countries. Binding MTBFs can be further subdivided into three distinct models based on the point at which they strike the balance between the three objectives highlighted above.
    1. The fixed aggregate ceiling approach fixes a binding limit on all or most central government expenditure for two or more years, and is not revised during that period. Given the primacy attached to ensuring that this aggregate ceiling is respected, these models do not set binding multiyear limits on expenditure categories within the overall ceiling but leave this to the discretion of the annual budgeting process. As such, this type of model is characterized by a higher degree of comprehensiveness and control at the aggregate level, but maintaining flexibility to revise and reallocate at the more detailed level.
    2. The fixed ministerial ceiling approach fixes binding multiyear expenditure limits for each of the 25–30 central government line ministries. Given the challenges and risks associated with fixing expenditure allocation at this more detailed level, these models tend to revise the ceilings more frequently and cover less central government expenditure. Thus, this type of model is characterized by a high degree of fixity and specificity but a lower degree of comprehensiveness.
    3. The forward estimates approach employed by Australia sets multiyear expenditure estimates for each of the central government’s 217 programs (or outcomes, as they are described in the budget). Unlike in the fixed ministerial ceiling model, these estimates are subject to revision twice a year. However, revisions are permitted only (1) to realign budget allocations to recognized changes in external parameters, such as a change in inflation or a change in underlying program volumes, or (2) to reflect discretionary policy decisions approved by the cabinet. In the absence of any approved changes, the previous years’ estimates remain and eventually become the budget appropriation. Thus, this type of model allows for a high degree of comprehensiveness and specificity, focuses on policy delivery, and opens up the possibility for revisions and reallocations within both expenditure areas and the aggregate amount. However, the lack of a fixed ceiling leads to a lower degree of certainty about total expenditure.

Although more and more countries have adopted MTBFs, their performance against the various objectives set for them has been mixed. One reason for this mixed performance record is that medium-term budgeting is demanding and requires a level of stability and predictability that typically develops only over time. Specifically, four preconditions are critical to the successful introduction of an MTBF:

  • A credible and predictable annual budget;
  • Accurate medium-term macroeconomic and demographic projections;
  • Established fiscal objectives and rules; and
  • A comprehensive, unified, top-down budget process.

Countries that have introduced medium-term budgeting have done so in different institutional contexts, facing different challenges, and pursuing different budget objectives. Accordingly, successful MTBFs differ in many respects. However, successful models have four key features in common:

  • Multiyear expenditure limits that define the nature, level, and terms of the restrictions being placed on future budget decisions;
  • Expenditure prioritization mechanisms that ensure that expenditure is allocated within those multiyear restrictions in a manner that reflects government policy priorities;
  • Forward-looking expenditure controls through which the consistency of updated medium-term expenditure projections with approved medium-term expenditure plans is monitored and enforced; and
  • Dynamic accountability arrangements through which adherence to stated medium-term objectives can be assessed by parliament and the general public over time.

Enforcing the credibility of the MTBF requires regularly updated multiyear expenditure projections, inclusion of adequate safety margins, firm control over multiyear expenditure commitments, and clear rules about the carryover of unspent appropriations. The credibility of a country’s medium-term budget plans depends on government’s ability to present its annual budget and final accounts in a manner consistent with those plans, transparently account for any deviations, and hold budget actors responsible for any unjustified deviations.

Binding MTBF models are more effective than indicative MTBFs in promoting aggregate fiscal discipline and enabling multiyear expenditure planning. Some indirect evidence suggests that binding MTBFs also induce greater focus on the effectiveness of expenditure. There is less evidence that the adoption of an MTBF facilitates the reallocation of expenditure between sectors. Finally, binding MTBFs also appear to provide a more effective platform for planning, communicating, and delivering both fiscal stimulus and fiscal consolidation plans. Anecdotal evidence suggests that, in the wake of the 2008 global economic crisis, countries with binding MTBFs were better able to convince markets of the credibility of their fiscal consolidation plans.

Source: OECD (2003). Budget Reform in OECD Member Countries: Common Trends at (accessed 14 October, 2012) and IMF (2013) "Public Financial Management and Its Emerging Architecture" edited by Cangiano, M., Curristine, T., and Lazare, M. in Chapter 4: "Medium-Term Budget Frameworks in Advanced Economies: Objectives, Design and Performance" authored by Harris, J., Hughes, R., Ljungman, G., and Sateriale, C. (accessed April 26 2014).

Page Created By: Matthew Seddon on 13 November 2012. Updated by Ian Clark on 2 January 2013. Updated by Matthew Seddon on 26 April 2014. The content presented on this page is drawn directly from the source(s) cited above, and consists of direct quotations or close paraphrases. This material does not necessarily reflect the official view of the publishing organization.

Important Notices
© University of Toronto 2008
School of Public Policy and Governance