Executive Summary

Ronald McGill, UNCDF

This book[1] introduces the challenge of building performance budgeting (PB) into public expenditure management (PEM), particularly at local government (LG) level in least developed countries (LDCs). The fact that it deals with LG is hard enough. This is because of the dominance of international experience at central government levels and the paucity of practice in the local tiers of government. The fact that it looks at not only developing countries but also, the poorest of them – the LDCs – might beg the question of “what possible relevance” can such a locational challenge have? The first answer is that UNCDF is the only UN agency mandated to invest capital in LDCs. Secondly, UNCDF strives to pursue policy experiments to reveal innovative solutions to the local development challenge, ultimately, to achieve a discernible impact on the poverty spectrum; at present, captured through the prism of the Millennium Development Goals (MDGs). The PB challenge is one such experiment.

[2]The book therefore advocates, through practice, in favour of meeting the challenge of PB in the LDCs. It does so by reviewing the concepts and context for PB in local government. It goes on to cite the achievements of PB in Tanzania (where this author helped to pioneer PB development). It presents the challenges of PB implementation in two UNCDF decentralisation programmes; Mozambique and Yemen. It introduces gender responsive budgeting (GRB); increasingly achieved through PB. Finally, it sets out methods for local government practitioners (and their advisers) to make progress in implementing PB in LG.

Concepts and Context

PB in local government

The demand for infrastructure and services confronts every local government (LG) in the developing world. The weakness of that local government compounds the enormity of the challenge. The fundamental importance of access to infrastructure and services, as a means of supporting both economic development and to impact on various parts of the poverty spectrum, is now accepted in common parlance. From the LG perspective, the challenge is three-fold:

A process of planning and measurement, to support targeted infrastructure and service delivery, is being practiced in the developed world, is being applied increasingly in transition economies and has been explored at the national levels of public expenditure in developing countries. The process is known as performance budgeting (PB). The first prerequisite for its success is in understanding the functions of government and how they fit each other.

In government reform, the agenda for change boils down to (a) an understanding of the current and desired functions of government and (b) the translation of the desired functions into the:

The shorthand for this is the institutional development (ID) agenda. The ID process, to achieve locally understood and determined reform, involves care in the facilitation of the change itself.[4]

The central point is that the ID agenda of context, structures and processes, to perform the functions, should be mutually inclusive. This optimises the potential for the efficient and effective delivery of infrastructure and services. If one part of the agenda is ‘out of step’, or does not “fit”, it distorts the potential for full performance from the particular institution being reformed. For UNCDF, this starting point is fundamental when it attempts to lever institutional development and supporting capacity building, through its incentive-based availability of capital investment.[5]

The current debates on the Millennium Development Goals (MDGs), their supporting national poverty reduction strategy papers (PRSPs) and public expenditure management (PEM) decisions, within a medium-term expenditure framework (MTEF), are dominated by the national or macro-picture. There is little evidence of substantive progress towards the translation of these macro-concerns at the local level. This localizing challenge is compounded by the sector-wide approach to planning (SWAP) for and financing of service delivery; a vertical logic. This works against the principle of integrated development planning, where all players (all funders) are contributing to a common or integrated development strategy in a given location; a city; a district; thus, a horizontal, even spatial, logic.

LGs are being given increased responsibility for ensuring the delivery of basic infrastructure and services in developing countries. That increased responsibility is often hampered by the institutional environment; e.g. service delivery responsibilities are decentralized without commensurate resources to perform these decentralized functions. In short, policy debates on PEM, infrastructure and service delivery (ISD) tend to focus on central government systems. While this is an essential prerequisite, the impact on LGs and the importance of developing policy solutions with planning and budgeting tools at the local level, is currently, not being given enough attention.[6]

It can be argued that there are four sets of decentralization variables in the policy and practical environment of PB. These concern:

Each must be understood in order to secure success in the PB implementation challenge. Within that context, a number of principles need to be accepted. It has been suggested that there are seven workable principles governing the implementation of PB, presented in the form of a checklist.[7]

1. PB fails at the first hurdle if the shift from input to output-based budgeting is not accepted and practiced.
   
   
2. PB is conceptually redundant without a strategic context to condition the resource allocating process.
   
3. The strategic context for PB is being satisfied increasingly through public annual reporting in terms of outcomes (wider societal impact) and outputs (organisationally specific, directly attributable achievements).
   
4. PB assumes that the real test is of resource allocation against future intentions (plan), tempered by recent performance (review).
   
5. PB requires all priorities to be in ranked sequence so that difficult choices are impossible to avoid.
   
6. PB’s key unit of planning and budgeting analysis is the programme. However, PB has to reconcile the programme structure with the organisational structure it represents.
   
7. PB measures the economy, efficiency and effectiveness of the infrastructure and services delivered by or on behalf of the organization.

This checklist is a foundation for local understanding. The bottom-line however is to recognise clearly the suggested difference between performance budgeting and programme budgeting. Performance budgeting is the over-arching process of (a) planning and targeted budgeting, to achieve (b) infrastructure and service delivery, that is (c) measurable and accountable. Programme budgeting is a technique or instrument, embedded in this process. This point is elaborated later. 

Making PB work

All local government (LG) is locked into the annual planning, budgeting, implementation and review cycle. In turn, it relates to and expresses the annual implications of its strategic context: government’s strategy for poverty reduction (PRSPs) and the medium-term expenditure framework (MTEF). This is then converted, through the annual plan and budget, into infrastructure and service delivery (ISD). One method for achieving government annual decisions and actions, in a PB format, is an organisation’s annual report, plan and budget (ARPB).

When the ARPB cycle is working fully, it always spans three years; last year’s performance, this year’s implementation and next year’s plan. These come together annually, at community level, at the start of the cycle. The desired result from ARPB is therefore publicly accountable, community supported, reviewing, planning and budgeting, to achieve targeted infrastructure and service provision.

A normal annual report starts with a general statement from the organisation’s leadership. This persists, in something like the following order:

• a statement by the organizational leader, including a review of the key performance highlights for last year and development thrust for next year (Part 1 of ARPB);
• an executive summary of the entire document (Part 2 ARPB).

The analytical core of ARPB; for performance budgeting, is:

• Strategic performance framework - the 3 year perspective (Part 3 of ARPB);
• Performance last year (Part 4 of ARPB);
• Proposals for next year (Part 5 of ARPB);
• Resource requirements for next year - the budget bid (Part 6 of ARPB).

The desired result of ARPB is publicly accountable and community supported, targeted infrastructure and service provision. This is to have a direct impact on particular client groups (short-term) and general socio-economic conditions (medium term). This is both in terms of the Millennium Development Goals (MDGs) and local development.  Thus, there are three levels of planning and budgeting analysis in PB:

1. The strategic framework and objectives analysis.
2. Programme and target analysis.
3. Activities and input analysis.

The strategic framework moves from baseline data to quantified objectives. This must be completed. It is the context for level 2. Programme and target analysis is the core of the performance budget. It involves the conversion of each objective into 3-year targets. It requires the prioritisation of these targets. It needs the conversion of these 3-year targets into annual targets, with an initial input classification – whether capital, operations or capacity. The result of all this is the context for level 3. Input analysis moves from annual targets to all activities and inputs, including costs and expenditure codes. Input analysis is meant to be a verification of the annual target cost estimates.  For the purposes if this text, the focus is on levels 1 and 2 only.

Governance

Performance budgeting is a means to good governance. Here, good governance focuses on the relationship, ultimately, between local government and its various communities. The idealised view is that this relationship is a partnership of equals. That is to say, both groups (local government and the local communities being governed) should have equal access to information upon which they can participate and make proposals concerning local development, its implementation and results.

More specifically, PB’s ultimate test is, perhaps, two-fold. First is the successful and publicly supported delivery of infrastructure and services. Here, local community participation is embedded in the annual cycle of reviewing last year, starting implementation this year and planning for next year’s proposals – captured in a public organisation’s annual report, plan and budget (ARPB). Secondly is the underlying intention to achieve a much more open and transparent system of planning for, implementing and reviewing expenditure to achieve results. If conducted properly, this is the basis for highlighting poor performance. It can also be a start to eroding corrupt practices.

Local communities must be central to the review of performance. Communities must test progress (completion on time, to specification, at the correct location), to solve the original problem that was agreed locally. This is particularly where a new public asset (an access road; a clinic) has been constructed. Where a service is being delivered, the client communities need to judge the immediate impact - (satisfaction of demand, occupancy rate, use of the facility; in short, has the original problem been solved?).

Local communities should be involved in the procurement process for new projects, particularly where the projects are directly community-based. With larger-scale projects, community representatives should have a role in a supervising group. Such groups must also check progress in implementation.

Communities are the key to locally relevant problem identification and subsequent solutions through ‘raw project’ definitions. These projects should be related to one another, ideally in the form of a map (with assets, opportunities and problems). They should be sequenced in priority. The process of prioritisation is central to participation and the wider notion of accountability. This is because PB requires all priorities to be in ranked sequence so that difficult choices are impossible to avoid.

The sum of all this work is each local council’s annual report, plan and budget (ARPB). This is the aggregation of accountability. ARPB is the method to achieve PB, in terms of both planning and reporting. ARPB should be freely available; being supported by things like summary leaflets and radio broadcasts. Thus, PB compels accountability through community participation and public reporting. In short, it is a legitimate approach to participatory budgeting.

More widely, PB encourages accountability of a local council that delivers (or ensures the delivery of) a public service. The ARPB tells the public what the organisation’s strategic framework is. It goes on to review last year’s performance of declared targets. The strategic framework and last year’s performance condition what should be proposed for next year. Thus, the organization is compelled (a) to tell the public what it intends to do and how much each target will cost, and (b), at the end of the year, whether the targets have been achieved and at what actual cost.  The difference between (a) and (b), publicly declared through the ARPB, makes the organization accountable. The accountability is enhanced by the ability to measure. This is both at the annual level and in every third year, strategically. It all adds up to an integrated assessment of strategic effectiveness, operational efficiency and budgeting with economy.

Having outlined UNCDF’s entry point to PB, through and on behalf of LG in LDCs, it is necessary to understand the converse: the national dimension and the classification of revenue and expenditure to support it. This is through the Government Financial Statistics (GFS) framework. This framework is being implemented in every country that UNCDF is piloting PB at the local level.

Government Financial Statistics (GFS) [8]

For possible sustained success, UNCDF’s PB policy experiment (as a journey of local development and reflective iterations) is grounded in the need to be meshed with GFS. This is because the budgetary coding and chart of accounts are the most fundamental building blocks of a government’s public finance management system. They are the means by which a government informs itself, and reports to its parliament and population. Decisions related to the choice of classifications systems can therefore have very far-reaching consequences. They can either enhance, or inhibit, the government’s ability to plan, oversee, allocate, prioritize, manage, control, account, audit, and report on the collection and use of public funds.

There is general agreement that classifications fall into three distinct areas. These are (a) administrative (or institutional), (b) functional and operational, and (c) economic or object/input. These three areas serve different purposes:

• The administrative classification identifies WHO is responsible for collecting/spending public resources, from the general (overall) responsibility down to the individual operational unit or project;
• the functional classification, and the related classification systems aimed at identifying operations of government, identify WHAT the government is doing and how these relate to government priorities, objectives and long term plans; and
• the economic classification identifies HOW the funds are to be collected or spent, i.e. which tax or administrative fee is being applied or what mix of inputs is required in order to deliver a particular public service.

Developing classifications under the GFS framework 

A sound administrative (institutional) classification will cover all government entities to a reasonably detailed level of organization, to match accountability needs. This will reflect the hierarchical nature of relations within government, allowing for progressive consolidation of budgets and accounts.

Functional classification serves a different need; that of identifying the activities of government, in a more sophisticated manner than the traditional sectoral breakdown. In general, the functional classification is statistical in nature. It is used to determine how much, or how little, government is active in given areas such as education, health, agriculture, tourism, justice, and so on. By itself, the functional classification will not determine specific directions being taken within a function, nor the outputs that can be expected from government intervention there.

The analysis and presentation of government spending, using the functional classification, increases the clarity of government operations. Its introduction and use can help improve the allocation of resources according to policy priorities. This is a first direct link to performance budgeting. In some cases, it can also lead to the identification and reduction or elimination of duplications in government activities. This suggests the functional determinants of resulting structures. 

The economic classification allows GFS to provide a clear and detailed framework through which the multitude of different transactions of government can take place. Its immediate purpose is to improve the timely, accurate and comprehensive reporting of government economic and financial activities. This is to guide macro-fiscal analysis, as well as budget formulation, management and evaluation. One of the key outputs of an economic classification system is the ability to produce a meaningful balance sheet.

Linking policy to budgetary allocations – the programme

One of the most difficult tasks in budgetary management is to ensure that the budget adequately reflects government policies and priorities. Traditional budgets have been focused on an incremental input approach. This has resulted in a lack of clarity in resource allocation and the growth of bureaucracies. The lack of clarity is that resource inputs are disconnected form the delivery of services. Concerning bureaucracies; they are primarily intent on defending past gains and are often reluctant to entertain any reduction in their budgetary allocations, even when their original purpose has become obsolete. Therefore, the system does not facilitate the reallocation of resources towards evolving needs. Recent efforts in Africa have focused primarily on the development of MTEF as a tool for resource allocation. In practice there continues to be a disconnection between the resulting MTEFs and the adopted budgets.

One of the drawbacks of the MTEF approach practiced in Africa has been the setting aside of salaries and wages from the resource allocation decisions.[9] This has reduced the effectiveness of reallocation decisions and disconnected staff (as an input) from the goals of MTEF (as the output). Yet, the MTEF approach is essential to establish the medium term nature of public spending; the essential strategic framework of performance budgeting. 

In a resource restricted environment, which all LDC governments face, it is not possible to implement all policy priorities in any one year. Even if this was possible, practical realities, such as the time taken to build a road or train a teacher, simply will not allow it. Furthermore, many budgetary allocations are semi-permanent in nature (such as the cost of running schools or the maintenance of roads) and must be ring-fenced if output (i.e. service delivery) is to be maintained.[10]  Budgetary management structures focused on outputs must therefore be conceived in a medium term context. Programme design should have this in mind, as most programmes, once established, will have a medium to long term existence, although their specific outputs, and in some cases objectives, will be adjusted from year to year.[11]

Finally, it is important to recognize that programmes require administration to support them, which are an integral and essential part of their cost. For example, the administration of teachers, which may include licensing, employment and assignment management, and other matters of common interest to teachers, and the education policy making units in a Ministry of Education, are essential components of the education program. Without these, the program may fail to meet its objectives.[12]

More recent forms of performance budgeting, which have been developed over approximately the last fifteen years, aim to establish a tighter relationship between funding and performance, generally but not exclusively as a means of putting greater pressure on agencies to improve their delivery, economically, efficiently and effectively. The main approaches used to build such a tighter relationship are:

• The use of unit costs for a given output in estimating expenditure requirements;
• Funding ‘incentives’ based upon performance;
• Purchaser-provider models; and
• Budget linked performance targets.Most contemporary performance budgeting systems use as their starting point the results-based program classification of expenditure.

Funding incentives for agency performanceAn influential theme in contemporary discussions of performance budgeting is that of financial ‘incentives’ for agency performance. Such incentives can take either or both of two forms:

• Additional funding for good performance: the proposition here is that agencies performing well should be given increased budgets because the prospect of increased budgets will motivate them to further improve their performance.  And conversely, poorly performing agencies would be motivated to improve their operations by the knowledge that, if they do not improve performance, their budgets would be cut.
• Retention of savings from efficiency: an agency producing it outputs efficiently would be able to keep the savings, even though its aggregate funding would not increase. The ‘purchaser-provider’ model discussed in the next section represent perhaps the most important application of this. The issue of what they can spend their savings on, however, needs to be carefully considered, especially in the context of limited overall resources.

The idea in both cases is to emulate the motivational power of profit and loss in the private sector, or of performance pay at the individual level, so as to boost agency performance.If such agency-level funding incentives are to produce better performance, they must be capable of motivating the individuals who comprise the agencies concerned. For this reason, funding incentives to agencies are often accompanied by provision to permit agencies to use their ‘profits’ to provide financial performance bonuses to their staff.

In short, international experience at the national level suggests that:

• The appropriate starting point in the introduction of performance budgeting is the introduction of a program budgeting system in which expenditure is classified by objective[13], and steadily improving the performance information which is used as an input into the budget process to facilitate better decisions about how limited budgetary resources should be allocated between competing purposes;
• Over time, as the necessary information is collected, the selective use of budget estimation based upon unit cost measures can be added to the basic program budgeting model – but only for appropriate types of output (principally for standardized outputs produced in large volume, like school education); and
• Budget-linked target-setting should not be rushed into, but should be attempted only when central budget decision-makers have acquired the necessary informational base.

This first section has reviewed the UNCDF institutional context for PB at the local level and the IMF-based GFS context for PB, dominated by national experience. From UNCDF’s perspective, the successful interface between PB (as a policy-budgeting tool), and GFS (as a revenue-expenditure classification) is fundamental. Tanzanian practice shows progress in both PB and GFS.

TANZANIAN PRACTICE

Strategic planning[14] 

Elements of strategic planning in Tanzania’s public service were first introduced during the Civil Service Reform Programme (CSRP) in 1995/98 under the Organization and Efficiency (O&E) reviews.[15] These reviews looked at ministries’ and independent departments’ internal structures and processes to improve organizational performance.

At that time organizational performance was conceived a requiring each government office to prepare an “Annual Report and Service Improvement Plan – (ARSIP)”. This embodied a strategic framework including a three-year planning horizon and annual performance targets. The ARSIPs were initially piloted in 7 Ministries in 1998/1999, to be the basis for introducing performance budgeting in central government. ARSIPs were to be the mechanism to implement publicly declared service delivery targets as contained in the Annual Plan through to a specific budget for each target.

 Strategic planning experience in Tanzania is still relatively new. Its various iterations since 1995-99 culminated only about a year ago, when the process of installing a full performance management system (PMS) was completed.  Yet recent reviews in various ministries suggest that the institutional development challenge and its supporting capacity building has much yet to do, based on the following conclusions:

• The quality of some ministry plans is poor and planning is not taken too seriously.
• Ministry plans are not consistent. Institutions maintain multiple plans, one for each audience or promoter of the plan.
• Ministry plans are not used as the foundation for implementation.
• Little, if any, formal ministry performance monitoring and performance reporting occurs.
• Too many processes have been introduced at the same time, which are beyond the absorption capacity of the ministries.

There are a number of reasons to explain these findings. 

Poor linkage between strategic planning and performance budgeting. While achievements have been made in terms of instituting strategic planning in the public service, implementation of the plans have, to some extent, been frustrated because the strategic planning process is not linked to the performance budgeting process. Currently, there are problems of terminologies used and linkage between strategic plans and MTEF formats. Thus, some ministries have found it difficult to implement their strategic plans.

Clear guidelines. The guidelines that were used by consultants and Government facilitators in the installation of PMS were not detailed enough. As a result, consultants tended to deviate from the templates that were provided by the President’s Office – Public Sector Management (PO-PSM). Also, ministries found it difficult to sustain the introduced changes when there were no detailed guidelines to refer to.

Commitment and ownership. Commitment of staff at all levels is important and an imperative for any undertaking. The participation of some of the ministries seems, however, to have been severely hampered by the inability of top management to be committed to the process. There is some evidence that, even for the staff who were involved, they were not always available. This reduced their ability to absorb and internalize the process, thus undermining general sustainability.

Capacity building. Most of the concepts under the PMS are still very new to most of the public servants. Intensive training and re-training is required to ensure that the new culture of result-oriented work is embedded in the service. The emphasis should be given in developing positive values and attitudes that will ensure that the knowledge acquired is used to enhance the performance of the employees. A universal principle is that of building incentives to perform.[16] 

Countries that are thinking of introducing strategic planning or performance management systems might consider the following lessons from Tanzania’s experience:

• Start with a simplified framework and concentrate on key processes.
• Guidelines and training material should be in place before embarking on this type of undertaking. They are useful for quality assurance and sustainability through on-the-job-training.
• Before introducing the processes, it is important to ensure that there is an institutional “fit” of systems and which institution is responsible of which process. For example, is the planning process aligned to the budget framework?
• Commitment at all levels will ensure ownership and success in implementation. There is need to have incentives and sanctions understood by all institutions
• Capacity building – having competent and skilled people is vital for understanding and implementing the new changes.
• Reward system – a reward system is self-evidently important.

Performance budgeting[17]  

Performance budgeting (PB) was introduced to the government system to convert the incremental input based budgeting to output or target based budgeting. This technique of budgeting was initiated as a logical extension of the work done by the Organization and Efficiency (O & E) component of the Civil Service Reform Programmes (CSRP), taking account of the established institutional roles in government.  Having finished defining organizational structures, the O and E component of CSRP started to address the issue of improved efficiency and accountability in service delivery.  It was established that improved organizational efficiency, effectiveness and accountability could be achieved through the process of Annual Performance Reporting and Service Improvement Planning (ARSIP).  Performance budgeting was thus recommended as a tool that could operationalise the whole process.

This new approach was expected to improve budget performance as measured by output indicators, unit costs and the measurable deliverable quantity of services for a given allocation of budget resources. After the introduction of each ministry’s Strategic Plan and Performance Management System (PMS) – which was results oriented – performance budgeting provided the necessary supportive framework.  Performance improvement (through economical, efficient and effective target achievements), was to be reflected in the budget allocation process.

It was generally accepted that PB was trying to yield practical benefits. These included:

• To enhance efficiency in service delivery;
• To enhance management accountability through performance monitoring and review;
• To improve resource allocation by linking resources allocation to specific quantitative and monitorable targets;
• To ensure consistency of resource allocation with institutional perspectives in a strategic medium term framework; and
• To facilitate a holistic approach to budgeting that would enhance budget integration.

All work to date was based on the principle of “re-introducing planning to budgeting” through PB. Towards the end of this phase of PB development work came the recognition of the over-arching imperative of the IMF-conceived system of government financial statistics (GFS) – outlined above.

The need for reclassification was felt because of the weakness in the coding system of the Government budget. The structure of the old classification was inadequate to provide for meaningful economic analysis of a performance-based budget. This required a consistent and unified classification to assess results at target level. Further to that, the government accounting system was being computerized. It was therefore seen as prudent to unify the system of budgeting and take advantage of the accounting reform initiative. There was also need to adopt common international reporting standards.

Reclassification of the budget in terms of the Government Finance Statistics Manual (GFSM 1986) started in 1999.  The work was carried out progressively, starting with the economic classification of Recurrent Budgets of Ministries, Departments and Regions. After gaining necessary experience, the Councils were later taken on board. Thus by year 2000/2001, the whole of the recurrent budget was processed using GFS item codes.  Subsequently, the development budget was reclassified and reflected in the year 2003/2004 Budget Estimate Book (Volume IV).

The following were the main steps taken to implement the GFS classification:

• A structure of the GFS classification was worked out by the Ministry of Finance (MoF) with the assistance of a consultant;
• An initial GFS item code structure was prepared;
• A bridge table, showing the link between old items and the corresponding new GFS items or sub-items, was prepared;
• Budget officers were trained on the technical aspects of GFS;
• The GFS item code list was revised and used as reference source in coding the Budget;
• Ministries were trained on GFS classifications, to equip them with necessary skills; and
• The recurrent budget for year 2000/2001 was processed using GFS codes; the  development budget was taken on board subsequently.

PB has been fairly successfully implemented both in central and local government.  As already explained, with effect from Fiscal year 2000/01, all Ministries, Independent Departments and Regions, including Local Government Authorities prepared performance/output based budgets.  Within the institutional budget presentation, one can trace quantifiable/qualitative milestones/targets and activities which are ‘SMART’ (Measurable, Achievable, Realistic and Time bound). Other achievements include, increased accountability, transparency and efficiency in resource allocation.

The GFS economic classification, based on GFS Manual 1986, has been completed. All Budget books Volumes I – IV are coded in line with GFS requirements.

Further, Annexes to Volume III and IV, that capture transfers to Local Government Authorities are prepared in GFS Codes.  Annex to Vol. III indicates details of expenditure allocations to education, health, water, agriculture, roads and general activities, while volume IV provides the details of development projects.   In both annexes, resource allocations are linked to predetermined objectives, SMART targets and activities.

To facilitate analysis of performance, segment 2 reference codes are provided in the computerized budget data entry which is composed of:

• Objective
• Target
• Target type
• Activity
• GFS item

The introduction of performance budgeting was a very important step in the process of improving budgetary allocation and implementation. Resource allocation is now linked to specific, quantitative and monitorable targets. Management accountability has been enhanced through performance monitoring and reviews. Various steps were taken to introduce performance budgeting in the Government Budget system.  Now, all budgets i.e. recurrent and development, are performance based.

Similarly, the adoption of the GFS economic classification has improved the coding of the budget estimates. Analysis of the impact of government expenditure in the economy can be done more effectively than before.  The Government took other initiatives to improve accountability and efficiency in the management of the public resources and in putting in place the institutional framework for stakeholder participation in the budget process. The public expenditure review (PER) process has continued to be an important forum for public sector participation in expenditure management issues.

The PER process is also one of the key initiatives in enhancing efficiency in public expenditure management. Inputs from PER are incorporated into the sector plans and budgets.  The MTEF, which came out of the PER process, is based on performance budgeting. It focuses on enhancing predictability and efficiency in public expenditure management in the context of a three-year timeframe.

The second phase of implementation of Poverty Reduction Strategy (NSGRP) has underscored the importance of performance based budgeting.  Originally, resource allocation ceilings were generally fixed at sector or vote level. With MKUKUTA (PRSP II), the focus is now on outcomes/expected performance.

Increasingly therefore, performance budgeting in Tanzania is becoming a reality. It is being embraced in many fiscal reforms for both the central and local governments.   However, more achievements seem to have been registered in terms of budget planning. The biggest challenges still remain with regard to performance monitoring and assessment. Development of appropriate software and capacity improvement for the key actors in Central Government and LGs is needed so that performance results from the budget system can be traced and accounted for.  The need for an innovative approach to this organizational information system’s challenge is paramount.

This second section has reviewed (already) mature PB and GFS practice in Tanzania. Next, it is necessary to reflect on fledgling UNCDF practice in Mozambique and Yemen; both of which are committed to the PB policy experiment, through agreements with their respective decentralisation programme’s counterpart ministry. 


MOZAMBIQUAN AND YEMENI FLEDGLING PB DEVELOPMENTS

Mozambique[18]

The Mozambique team is still fighting to secure universal acceptance to the integration of planning and budgeting at the decentralised level.

A little over a decade ago, decentralised planning and finance hardly existed in Mozambique. Those that advocated it were questioned over its viability and practicality. Indeed only a few took the exercise seriously. Now, the Ministry of Planning and Development (MPD) is currently preparing a national strategy for the decentralised planning and finance programme which will guide its development and evolution over the next ten years.

Despite this progress, some cautionary notes are necessary in relation to the policy context, the organisational structures and the planning and budgeting processes which underpin and determine the success of institutional reform and development. Whilst there has been some attempt to reform and restructure sub-national tiers of Government and introduce participatory approaches, outside of the municipalization programme, the changes introduced at provincial and district level represent a deconcentration of administrative responsibilities only. Moreover, under the Local Government Act (LOLE), Districts can make budget proposals but cannot approve their budget. District and provincial budgets are aggregated at national level into a single “State Budget” (Orçamento do Estado) which is approved by parliament. Thus the decentralisation process in Mozambique has not yet reached the stage where districts (or even provinces) have any substantial autonomy or discretionary power.

Concerning PB; whilst the pre-conditions are being created for the introduction of performance budgeting at district level, it is still very much an “add-on” to existing processes rather providing a new underlying concept for the local planning and budgeting process. Moreover, at central level, despite progress on GFS through the (gradual) introduction of SISTSAFE, there is still some degree of scepticism as to whether the essential preconditions for performance budgeting are actually in place. With regard to the policy context, then, some significant challenges remain.

Organisationally, the foremost challenge relates to human and institutional capacity of sub-national tiers of government to adapt to, promote and implement the changes envisaged in the legislation; procedures that have recently been introduced. In the public sector reform programme, this is referred to as “change management”. LOLE and SISTAFE (the GFS reforms) in particular will require the restructuring of district government in order for it to assume some new responsibilities. This will include the need to establish a district planning and finance department and a district treasury; both of which to date have not existed. More fundamentally, the range of skills and experience required for the implementation the new legislation are rare at district level. This will require significant investment in the public service training system especially in the areas of management, planning and budgeting, execution and auditing. Also, a change of “mind set” will be required to think in terms of the “District Government” as a homogenous unit rather than a collection of sector representatives (an issue that pervades Yemen as well).

Greater attention, too, needs to be given to performance monitoring and evaluation. There are concerns, for instance, over the adequacy of the annual monitoring and reporting process (an integral part of the PB cycle). This concerns whether the information systems and analytical and report writing skills are sufficient for sound assessments of performance. There is also a question of promoting a greater degree of downward accountability, through the introduction of appropriate communication techniques. This is to ensure that those who participate in the planning process are adequately informed of the outcome of their involvement. If not, there is a risk of undermining the planning process (another issue shared with Yemen).

These challenges are not peculiar to sub-national tiers of government. At national level the lack of an adequate financial management framework, systems and audits creates serious problems of inefficiency, accountability and transparency in the collection and application of funds. The causal relationship between planning and budgeting is still weak. A single, integrated, harmonised planning and budgeting system, linking medium term strategic planning and fiscal objectives with annual operational plans and budgets has still not been fully achieved at national level. In contrast, it has been, through the UNCDF-led programme at district and provincial levels in Nampula Province. However, sub-national planning priorities and budgets are still poorly integrated into the national planning and budgeting system.  Clearly, there is a long way to go. Much depends on the institutional development framework and its progress.

Yemen[19]

The Yemen team has put great emphasis on the levels 1 and 2 analysis required for PB, through the ARPB format.[20] 

The experience of the decentralisation and local development programme (DLDSP) in Yemen indicates a number of things. First, a sound infrastructure for accessing, archiving and analysing information on the extent of development, service coverage and quality in a district is one of the critical prerequisites for monitoring performance and the introduction of PB. Such infrastructure should also function as the foundation for the integrated planning and budgeting process. This infrastructure must include effective and sustainable information accessing mechanisms. This must be grounded in operational participatory systems that ensure the equitable engagement of all segments of a district’s population; it must be conducive to local participation. It must also include effective forums for such participation to take place and facilitate substantive communication between elected council members and targeted communities. Information generated through this infrastructure and channelled through these forums should be processed through efficient information systems that enable local authorities to decipher any information category. This is to (a) support their decision making activities as they allocate limited district resources and (b) evaluate the efficiency of implementation and effectiveness of their interventions.

Capacity at the district level to convert quantitative data into a usable and easy to understand analytical report, which summarizes the state of development of a district, must be developed. This is in order to ensure that all stakeholders can be equally informed and fully participate in decision making processes to allocate district resources. The quality of such reports and the accuracy with which they convey a district’s state of development will have a direct impact on the quality of a district’s plans and its effectiveness to address its development challenge.

The DLDSP’s experience also indicates that the aggressive piloting of methodologies and concepts of integrated planning and performance budgeting at the district level is critical. It ensures the laying of a necessary foundation for a coherent approach for setting development objectives and implementing programs and projects that accurately contribute to their achievement. Such methodologies are displacing set ways of thinking that are grounded in sector (silo) planning approaches and line-item budgeting. These contribute to the implementation of fragmented interventions that are inaccurate and difficult to track in terms of impact and implementation efficiency.  However, this PB pilot is facing a series of challenges. These stem from resistance from line ministry branches at the district level (classic deconcentration) interested in maintaining decision making control over the allocation of sector specific investments.  Resistance also stems from cognitive inflexibility among district level officials and council members. This curtails the smooth transition from sector thinking (and its ‘laundry-list’ approach to addressing development need) to an integrated and program based approach instead

However, the most significant challenge facing the program in piloting an improved PEM and developing the foundation for PB, stems from severe limitations in recurrent budget allocations for district authorities by central government. These limitations prevent the full activation of council members to mobilize community participation. They also curtail the ability of district administrative departments to perform their role in carrying out basic support functions. These include information and financial management, revenue collection and the provision of technical support and supervision over project implementation activities, and the management and upkeep of district service assets.  Furthermore, these limitations have severely affected the ability of line ministry branches to perform their sector specific role of guiding and supervising service delivery and ensuring the gradual achievement of sector targets and global development goals. 

To this point, this Executive Summary has reviewed concepts, mature and fledgling practice in PB. Participatory budgeting is defined in the conceptual first section; as a sub-set of the PB experiment. Gender responsive budgeting (GRB) is (or can be argued to be) another sub-set of performance budgeting. It is a view that is certainly shared by others, conceptually and practically. The next section outlines the GRB challenge in relation to PB.

GENDER RESPONSIVE BUDGETING [21] 

Concept

Gender-responsive budgeting (GRB) is a very broad and fast developing field. There are many evolving conceptual frameworks and methods and there are varying forms and levels of intervention. Interestingly, the developing countries are leaders in the field, ahead of Europe and the United States. Developing country initiatives have tended to focus on national budgets, while some practical applications at the regional as well as local / municipal levels are found in many European countries. GRB initiatives (GRBI), although they have tended to apply similar, widely disseminated frameworks and tools, are also diverse in terms of the institutional actors involved; where the initiative and impetus comes from, whether from public administration, parliamentarians or NGOs.

The majority GRBIs focus more on gender analysis of existing budgets, albeit as entry point to advocate for gender-responsive budgeting. In most of the cases, these initiatives fail to get beyond the sensitisation and training workshops. When they do go further, they stumble on the follow-through, mainly because of the difficulty of sustained and on-going technical support relating to how to practise it.

With very few exceptions, GRBIs have not engaged with performance budgeting (PB) and the broader class of results-based budgeting (RBB). At the conceptual level, the paper by Rhonda Sharp[22], one of the pioneers of gender budgeting in Australia, is the only major think-piece on the subject. The focus is mostly at the national level. There have however been some local level initiatives combined with national level initiatives.

The purpose of gender-responsive budgeting (GRB) is to make states accountable to their commitments to gender equality and sustainable development. GRB involves actions to change policy and budgetary processes so that fiscal measures regarding revenue and expenditure take into account the disparities in income, resources, assets, decision-making between women and men. Specifically the concern is for public expenditure to be more oriented to the differentiated needs, priorities and perspectives of women and men. GRB uses gender analysis to uncover the disparities among women and among men, which provide a rationale for public expenditure targeted to these groups. This is in order to address both gender equality and poverty reduction objectives. GRBR uses a results-based approach. It seeks to draw the chain of results, linking public expenditure, physical investment and service provision, actual access and utilisation of these services and the outcomes in terms of capacities, livelihoods and well-being of women and men.

There are diverse reasons why governments are compelled to take action to make policy planning and budgeting more gender-responsive. There are the global rights and policy frameworks that bind states to achieve gender equality and address women’s disadvantage as well as to achieve development goals.[23] The national policy frameworks are increasingly results-focused and purpose-driven. This is now common-place through the poverty reduction strategy papers (PRSPs). They target the achievement of the Millennium Development Goals within a specific time frame. There are overarching strategies for poverty and inequality reduction. The focus of public action extends to impacts on the population and cannot just be limited to outputs; the delivery of public services, the application of rules and regulation. These routine missions of government agencies and the sectoral programmes are now to be more closely aligned to the overarching frameworks in PRSPs and their resulting medium term expenditure frameworks.

In a policy framework characterised by privatisation and liberalisation, the rationale for public expenditure is limited to what the market cannot do in terms of efficient allocation of resources.  A central tenet of public expenditure management (PEM) is that the grounds for public intervention rest on the existence of market imperfections and failure: public goods, externalities, missing markets as well as inequalities. All questions of allocation, distribution and redistribution, which are at the heart of fiscal policy, have specific gender dimensions and are underpinned by gender relations.

Non-market services are a powerful rationale for a gender-responsive approach to managing public expenditure. Not taking non-market services into account can lead to a significant misallocation of resources as the following sections will analyse further. These non-market (unpaid) services of the care economy[24], are performed mostly by women in providing for human beings and sustaining the social fabric. In fact, they underpin the operation of marketed goods and services but are hidden resources. They are not reflected in market transactions, economic data and state processes and instruments. They provide a link between economic and social dimensions. State intervention has a significant influence on the care economy and can regulate the relationship between the care economy and the market economy.

The GRB to PB interface

Gender-responsive budgeting is both political and technical. It covers the entire policy cycle. It is not just concerned with the budget as a specific instrument. Thus, it sees the budget as the financial translation of the policies and objectives of the government. This is an excellent point and therefore, worth reinforcing, both for GRB and PB:

The process of budget setting is the opportunity for strategic issues to be identified and acted upon. This is the policy development process. Experienced officials know that expenditure is policy; policy is expenditure. They are so intermeshed that any either/or answer about causation is foolish.[25]

GRBR is therefore predicated on a closer alignment of policy planning and budgeting as an integrated and therefore, not as separate and poorly articulated processes.

GRBR starts from the development outcomes that are needed and works towards the budgetary implications. It then compares the actual budgetary outcomes with the development outcomes in order to improve effectiveness and progressively attain goals. There are several broad steps in the gender-responsive policy cycle.

1. Start with a situation analysis which yields information on the socio economic and environmental conditions governing the lives of women and men;
2. Compare this information against development norms and standards such as human rights conventions and goals such as MDGs as well as targets;
3. Respond with gender-responsive appropriate policies and intervention strategies, through  a more inclusive process of identifying options and making choices;
4. Translate these into medium term expenditure and annual budgetary measures and actions;
5. Monitor the performance results; and
6. Feed the monitoring information back into better performance, in order to achieve the desired outcomes.

Steps 4 and 5 concern planning, programming and budgeting. It is in this area that the interface between gender-responsive budgeting for results and performance budgeting is the closest. PB typically involves the specification of a chain of results - from inputs to outputs to impacts. Policy, planning and budgeting are therefore intended to be integrated. The key point is the linking of allocations to a strategic framework, such as a plan. Performance can then be measured in relation not just of inputs (how much has been actually spent compared to how much has been allocated) but in relation to what the allocation has been spent on and for whom. This can be specified at the level of outputs, but also at the level of each point in the chain of results. Performance can be in terms of impact, where the effectiveness of the output in leading to impact is in question, or in terms of output, where the efficiency of translating inputs into outputs through activities is the performance measure.

The synergies between GRB and PB have been recognised by leading exponents of the GRB field; Rhonda Sharp and Debbie Budlender. Sharp, who pioneered gender budgeting in South Australia before 1995, has written a UNIFEM publication that describes performance budgeting and the scope it provides for GRB. She suggests that the three E’s of performance budgeting; the measures of economy, efficiency and effectiveness, be supplemented by a fourth E, equity, to make performance budgeting more responsive to the gender equality and equity concerns central to GRB.[26]

Debbie Budlender of the South African NGO, CASE, the Community Agency for Social Enquiry, has been the key resource person and champion of the Women’s Budget Initiative in South Africa. In a booklet, again commissioned by UNIFEM, about gender budgeting with a rights-based focus[27], she introduces the section on some basic budgeting concepts thus: “The ideas in this booklet will work best in a country that is using some form of performance-budgeting”. If not she argues, it would be necessary to delve into the logical framework of the programmes that are being budgeted. That means looking at the way in which the chain of results between inputs, activities, outputs and impacts have been worked out.

Experience with PB shows that it is easier to focus on outputs, whereas it is more important and meaningful to focus on impacts.[28] It is certainly an improvement over ‘line’ and input-budgeting and incremental budgeting. The importance of PB to GRB success is evident here. However, UNCDF’s challenge is to try and make PB – ideally incorporating GRB – work within its decentralisation programmes in the LDCs.

MAKING IT WORK

From UNCDF’s perspective, “making it work” means three things: developing an understanding of PB, from a LG practitioner’s point of view; building on lessons to date; and establishing the practical components. This Executive Summary has sought to establish a basic understanding, from an institutional development perspective. It has attempted to review some PB implementation lessons to date, from LDCs. What remains are the practical components. The full PB book, Achieving Results, presents material on “taking things forward”. This includes a basic training programme to support the production of a performance budget, through its modality, an ARPB. It illustrates an example of the results of such training; a draft ARPB.

Finally, it presents a generic PB manual, to capture all the practical steps to date. The manual is sub-titled a “bottom-up” and “output-based” approach to planning and budgeting. The “bottom-up” Part 1, builds upon seasoned UNCDF practice in the deployment of local development funds (LDF) to achieve community involvement in the project cycle. Part 2 builds upon (and is designed to encourage more) experimentation in the PB challenge.

This Executive Summary is therefore concluded with three diagrams: an illustration of the external logic of PB (Figure 1 [pdf]); a presentation of the internal structure of PB (Figure 2 [pdf]); and finally, a generic “bottom-up” and “output-based” planning and budgeting cycle (Figure 3 [pdf]). Understanding all three, in the context of this text and the book from which it comes, may go some way to helping meet the challenges of implementing PB in the LDCs. In so doing, we may see passive local councils being transformed into active development agencies,  concerned with all sources of revenue (local; from government; from donors) and expenditure, committed to Achieving Results.