By Richard Hughes and Richard Allen
On Thursday, 2 October 2008 the IMF’s Fiscal Affairs Department hosted a half-day seminar on Medium-term Expenditure Frameworks: Lessons from Success and Failure in Advanced and Developing Countries. The seminar was the first in a series that FAD will be hosting on the subject. It was chaired by Marco Cangiano (IMF, Fiscal Affairs Department) and divided into three sessions (Download mtef_seminar_agenda.doc ):
- A presentation by Richard Hughes (IMF, Fiscal Affairs Department) the lessons from his experience with MTEFs in advanced countries (Download rhughes_mtef.PPT);
- A presentation by Dr. Salvatore Schiavo-Campo (World Bank) on the lessons from experience with MTEFs in developing countries (Download rsc_mtef_presentation.doc) ; and
- A panel discussion on how to make MTEFs more credible.
This posting provides a summary of the presentations and discussion.
SESSION 1: LESSON FROM ADVANCED COUNTRY EXPERIENCE
In the first session, Richard Hughes presented the initial findings of his work on the lessons from advanced country experience with MTEFs (Download rhughes_mtef.PPT). The presentation started by highlighted the fundamental tension in the design of any multi-year budgeting framework between the competing objectives of:
- maintaining aggregate fiscal discipline in the face of macroeconomic and other shocks;
- facilitating a more strategic allocation of expenditure between competing sectoral priorities; and
- providing greater certainty to budget-holders about their future resources to encourage more efficient inter-temporal planning of expenditure.
The tension between these objectives is reflected in the range of MTEF models one finds in advanced countries which vary in terms of (i) the categories of expenditure they cover; (ii) the amount of detail they provide beneath the overall expenditure envelope; and (iii) the level of discipline they impose over future budgets.
Lessons from Advanced Country Experience
The second part of the presentation looked at 5 key lessons one could draw from advanced country experience with multi-year expenditure planning for the design and implementation of MTEFs in low and middle income countries:
Lesson # 1. Model Design: Countries need to make choices between the three competing objectives above in designing their MTEFs. Relentless pursuit of ever greater detail and ever wider coverage will eventually undermine the overall discipline imposed by the MTEF.
Lesson # 2. Political Commitment: Collective commitment at the highest level of decision-making to the disciplines that an MTEF imposes is critical to its credibility and durability. It is no coincidence that most of the successful advanced country MTEFs were introduced following a period of fiscal crisis (and often a change government). The memory of that crisis and the painful fiscal adjustment it required helped to engender a sense of collective commitment around the cabinet table to fiscal discipline and to elevate concerns about maintaining overarching fiscal credibility above narrower sectoral concerns.
Lesson # 3. Budgetary Foundations: A credible annual budget is the cornerstone of a successful MTEF. For medium-term expenditure projections to play a meaningful role in shaping future resource allocation decisions, they have to be rooted in a clear understanding of how much is being spent in the current financial year on a given activity.
Lesson # 4: Underpinning Institutions: If an MTEF is to be more than a run of numbers in a document then they need to be supported by set of rules and procedures that lend those numbers credibility. These procedures include an expenditure planning and arbitration process, accountability arrangements, rigorously enforced adjustment mechanisms, and medium-term commitment controls. A strong, capable ministry of finance within the cabinet and a strong executive in parliament are critical to ensuring that these “rules of the game” are both clearly defined and consistently enforced.
Lesson # 5: Cautious Forecasting: All countries that have had success with MTEFs built considerable caution into their fiscal projections to ensure they were able to deal with contingencies. This caution reveals itself in the form of lower expenditure, higher receipts, and a lower budget deficit than initially forecast. These "pleasant surprises" were critical to meeting the costs of unexpected contingencies during the period for which expenditure plans were fixed.
Implications for MTEF Design in Low and Middle Income Countries
The presentation concluded by looking at the implications of advanced country experience for the design and implementation of MTEFs in low and middle income countries (L/MICs) and raised a set of questions about the advice the IFIs and donor communities are currently providing in this area.
Discussion
Discussion of the first presentation was lead by Eric Mottu (IMF, European Department) and Tom Rumbaugh (IMF, Asia-Pacific Department) and focused on:
- Annual budget vs. MTEF: Discussants echoed the emphasis placed on a credible annual budget as the foundation for a meaningful MTEF. However, some in the audience argued that placing expenditure forecasting in a medium-term perspective may be critical to improving forecasts for the year ahead as it helps to highlight unrealistic or unsustainable assumptions. In the end, it was agreed that while including estimates of expenditure (and revenue) beyond the budget year had value, the amount of time and effort being put into these estimates had gone well beyond the point of diminishing margin returns in many if not most countries. Furthermore, in few countries did these forward projections constitute a meaningful “framework” for the preparation of future budgets. Finally, in some lower income countries where personnel were already thin on the ground, the (often donor-driven) focus on developing ever more detailed and elaborate multi-year projections risked becoming a dangerous distraction from the more urgent and fundamental task of improving the credibility of the annual budget.
- Pros and cons of cautious forecasts: A number of those in the audience questioned whether persistent underforecasting of receipts contributed to or detracted from the credibility of a government’s medium-term expenditure plans over time. While such a strategy might help to “grease the wheels” in the first few years under the new framework, over a longer period the expectation of continued positive “surprises” can encourage or force governments to resort to ad hoc measures or mini-budgets which undermine incentives for medium-term planning. It was for this reason that a number of countries, including the Netherlands, have removed the explicit downward bias in the economic assumptions used to construct their multi-year spending plans. On balance, most agreed that it was better to err on the side of prudence when projecting revenues, especially in a system where expenditure is fixed on a multiyear basis, but most also felt that large and persistent underforecasting of receipts was likely to be counterproductive. One way of splitting the difference (discussed in greater detail in a subsequent seminar on the fiscal management of resource revenues) is to use a reasonable amount of caution in forecasting revenues but to be clear ex ante also about how any surplus receipts might be divided between debt reduction, tax cuts and spending increases.
- Should MTEFs include aid flows?: A number of discussants argued for the continuation of efforts to bring aid and other extrabudgetary flows into MTEFs despite their volatility. Particularly in low income countries where such flows account for a large share of public expenditure it was argued that their exclusion from the MTEF process would render the exercise meaningless from the point of view of expenditure prioritization. Furthermore, their inclusion within the MTEF was seen as a critical mechanism for getting donors to provide more considered and reliable estimates of future commitments. However, it was important that the commitment controls discussed in the presentation extended to line ministries’ discussions with donors which too often resulted in fait accompli presented to the Ministry of Finance in the form of promises to match multi-year commitments from donors with domestic resources. Tunisia was regarded as a good example from the point of view of disciplined and coordinated interaction with donors.
SESSION TWO: LESSONS FROM DEVELOPING COUNTRIES
In the second session, Dr. Salvatore Schiavo-Campo presented his paper on Medium-term Expenditure Frameworks in Developing Countries (Download rsc_mtef_paper.doc) which has been blogged on this site previously (click here for the link). The ensuing discussion was led by Chris Lane (IMF, Africa Department) and Luis Cubeddu (IMF, Western Hemisphere Department) based on their respective experience of MTEFs in Africa and Latin America.
Discussion
Chris Lane observed that one of the objectives of establishing an MTEF was to rationalize the number of national planning and budgeting initiatives and documents. However, in some countries, this objective was overlooked and MTEFs had actually led to a proliferation of budget/planning instruments including the annual budget, the national development plan, the PRSP, the PIP, etc. This had the effect of shifting the locus of decision making out of the Ministry of Finance and further weakening the medium-term discipline that the MTEF was supposed to engender. He wondered whether in countries where attempts to construct comprehensive MTEFs had been a disappointment, it might be better to start again with sectoral MTEFs, such as the one Rwanda has developed in the health sector, which were often based on more accurate forecasting models and better information about aid flows.
Luis Cubbedu asked how far and how quickly could one could push the institutional revolution that MTEFs required. It was clear that well-functioning MTEF required good macro-fiscal forecasting units, IT systems that provide regular and comprehensive information about budget execution, capacity to accurately model the future cost of existing and new policies, and strong institutions of external validation accountability. These institutions were likely to emerge in developing countries only over time. He also noted that none of the speakers had pointed to any clear evidence that MTEFs had improved the productivity and efficiency of expenditure which was a major selling point of MTEFs when they were introduced in both advanced and developing countries.
SESSION 3: HOW CAN WE MAKE MTEFs MORE CREDIBLE?
The final session of the morning was a panel discussion chaired by Richard Allen (IMF, Fiscal Affairs Department) on what experience in advanced and developing countries teaches us about how we can make MTEFs more credible. The panel of discussants comprised Jim Brumby (World Bank), Mario Sanginés (Inter-American Development Bank), Professor Allen Schick (University of Maryland), and Michael Stevens (World Bank).
Jim Brumby started off with the observation that according to the most recent OECD Survey of Budget Practices, some 90 percent of countries already operate within some sort of medium-term framework. However, this term appeared to cover a variety of practices. A useful distinction could be drawn between:
- medium-term fiscal frameworks (MTFFs),
- medium-term budget frameworks (MTBFs); and
- medium-term expenditure frameworks (MTEFs).
Correct sequencing of the three was critical with MTFF being a basic requirement followed by an MTBF or MTEF. In addition, Rino Schiavo-Campo’s presentation proposed a further distinction between “forecasting” MTEFs, “planning” MTEFs, and “programmatic” MTEFs.
Recent World Bank studies, including work by Jim Brumby and the work that Mike Stevens’ has been leading for the Bank’s recent Quality Assurance Group, had revealed significant shortcomings in the way that MTEFs had been designed and implemented in many developing countries. The IFIs and development institutions themselves bear some responsibility for the mixed success in implementing MTEFs and for generating unrealistic expectations about what MTEFs could and would achieve in practice.
Mario Sanginés underlined concerns about country ownership and too many MTEFs being donor-driven “projects” rather than lasting institutional reforms. In many cases national authorities support the establishment of an MTEF not out of a genuine belief in the value of the instrument but as a means of contracting with donors to secure aid financing. As a result, MTEFs had often been designed and implemented mechanically without due attention to a country’s institutions, and the need to establish a close relationship between the MTEF and fiscal policy.
The MTEF model developed in many low-income countries was often more appropriate to OECD countries than the needs of countries with low capacity and weak institutions. In addition, insufficient attention had been paid to sequencing the introduction of MTEFs with improvements in country capacity. In many cases, an country’s established an MTEF unit that was isolated from the budget office and the treasury department, and had little real impact on the budget process, or policy decision-making. As a result, linkages with the annual budget had been ignored and key issues such as reconciling MTEF projections with the government’s medium-term fiscal strategy and the annual budget had not received sufficient attention.
Referring back to his classic 1998 presentation on “Look Before You Leapfrog” (sometimes referred to as “Getting the Basics Right”), Allen Schick commented that, in many countries, the focus on developing an MTEF was an advanced reform that could distract the authorities’ attention from putting in place the fundamental building blocks of a sound budget process. Second, he warned that the message that MTEFs should be kept “simple” could be misinterpreted. By emphasizing the importance of expenditure aggregates, rather than detailed line items and the key process of monthly cash allocations, the emphasis on establishing an MTEF could reduce the effectiveness of control that finance ministries exercised over public spending (“The West was created because of line item budgeting”).
There was some discussion of whether MTEFs should be established on a top-down or bottom-up basis. In the latter case, the initial focus would be on developing medium-term frameworks for key sectors such as education, health and defense, and pilot projects could be set up in these sectors before launching a full-blown MTEF. In many emerging markets and low income countries, sectors such as education, health and defense have already developed relatively sophisticated methods and data for analyzing and programming budget allocations and investment decisions over the medium-term. Through expertise gained and the demonstration effect, such methods and data, and possibly the use of sectoral pilots, can be useful in informing and facilitating wider macro-level exercises. Nevertheless, it was stressed that such sectoral initiatives did not constitute medium-term expenditure framework in the conventional sense since they lack an overall fiscal policy perspective and a direct link with the national budget process.
The discussants broadly agreed that, although the implementation of MTEFs had been flawed in many emerging markets and developing countries, the basic MTEF concept was sound. Reversing the policy of encouraging such countries to develop MTEFs was probably not feasible or even desirable at this late stage. But the efforts of the IFIs, and other development partners, should be refocused on:
- building the basic capacity of the budget department of finance ministries in the core areas of budget operations and processes;
- strengthening the capacity of finance ministries in macro-fiscal analysis and forecasting;
- strengthening the role of the IFIs in supporting policies that stabilize public finances in the medium term, including the use of fiscal rules and oversight institutions where appropriate; and
- integrate the MTEF with basic budget and treasury processes and improve the transparency of fiscal institutions and the budget process.
