Publication date
December 1999

List of abbreviations and acronyms








Executive summary






The conceptual framework






Structural context of the budget process






Weaknesses in public expenditure management












Budget outcomes

The budget process













The institutional setting of the budget process










The limits to legislative oversight of the budget

The ascendancy of the sector ministries

Sub-national government’s peripheral role in the budget

The incremental nature of the budget

Why the executed budget deviates from the approved budget

The consequences of external aid

Civil society involvement in the budget process










Interests and capacities of key actors







The legislature and political parties

The executive

Non-state actors

External actors








Case study: HIV/AIDS treatment and the budget





Evolution of policy on HIV/AIDS treatment and its financing

The key actors and their incentives


The sources of future change















Steps in the budget cycle

Planning and budgeting in the Ministry of Health

IMF targets, “fictitious” projects and off-budgets






List of figures

















Relationship between actors, institutions, budget process and budget


Annual growth rate of GDP, 1985-2003

Government finances in percentage of GDP, 1997-2003

Percentage share of PARPA priority sectors in total government

expenditure (including interest payments)

Government expenditure per capita and poverty headcount by provinces,


Expenditure per capita by provinces, average 2002-2003

Normative relationship between planning and budget instruments

External grants, 2001-2003

Budgeting and resource flows between MPF, donors, sector ministries and

provincial directorates

Cumulative percentage of district health directorates receiving first budget

transfers, by date, 2002

Internal and external budget resources, 2003

Percentage of households possessing radio and TV sets, 2002/03

Percentage of staff with university degrees, 2004

Distribution of DNPO’s national technical staff by level of education, 2003

HIV prevalence and deaths from AIDS, 1998-2010


















List of tables



Election results, 1994, 1999 and 2004


















































Associação dos Antigos Combatentes (Veterans Association)

Assembleia da República (Assembly of the Republic)

European Parliamentarians for Africa

Country Assistance Plan (of DFID)

Comissão Consultativa de Trabalho (Labour Consultative Commission)

Cenarário de Despesa e Financiamento de Médio Prazo (Medium Term Expenditure

and Financing Framework), of MISAU

Cenário Fiscal de Médio Prazo (Medium Term Fiscal Framework)

Conta Geral do Estado (General State Accounts)

Central de Medicamentos e Artigos Médicos (Drugs and Medical Materials Centre)

Conselho Nacional de Combate à SIDA (National Council to Combat AIDS)

Comunidade dos Países de Língua Portuguesa (Community of Portuguese Speaking


Conta única do tesouro (single treasury account)

Development Assistance Committee, of OECD

Departamento de Análise Económica e Social (Department of Economic and Social

Analysis), of DNPO

Direcção de Administração e Finanças (Directorate of Administration and Finance),

in line ministries

Direcção de Administração e Gestão (Directorate of Administration and

Management), of MISAU

Direcção Distrital de Saúde (District Health Directorate)

Department for International Development, of the UK Government

Direcção Nacional da Contabilidade Pública (National Directorate of Public

Accounts, in MPF)

Direcção Nacional do Plano e Orçamento (National Directorate of Planning and

Budget), of MPF

Direcção Nacional do Tesouro (National Treasury Directorate), of MPF

Departamento do Orçamento (Budget Department), of DNPO

Direcção Provincial de Apoio e Controlo (Provincial Directorate of Support and


Direcção de Planificação e Cooperação (Directorate of Planning and Cooperation), of


Departamento de Programação Macro (Macro Programming Department), of DNPO

Direcção Provincial do Plano e Finanças (Provincial Directorate of Planning and


Expenditure Tracking and Service Delivery Survey, conducted in health sector in


Fundo de Apoio ao Sector da Educação (Education Sector Support Fund)

Fundo de Compensação Autárquica (Autarchic Compensation Fund), for recurrent

expenditure subsidies provided by the OE to autarquias

Fundo Comum Provincial (Provincial Common Fund), of MISAU

Fundo de Iniciativa Local (Local Initiative Fund), for OE resources provided to

autarquias for investment

Fortalecimento da Planificação e Orçamentação Sectoriais (technical assistance

project financed by DFID in DNPO)

Frente de Libertação de Moçambique (Mozambique Liberation Front)

General budget support

Gross domestic product

Gabinete de Estudos (Office of Studies), of MPF

Government of Mozambique

Group of 15 GBS donors

Group of civil society organizations involved in monitoring poverty issues

Highly active anti-retroviral therapy

Highly Indebted Poor Countries (debt relief initiative)

Human Immunodeficiency Virus/Acquired Immune Deficiency Syndrome

Inspecção Geral das Finanças (General Inspectorate of Finance, in MPF)

















































International Monetary Fund

Ministério da Agricultura e Desenvolvimento Rural (Ministry of Agriculture and

Rural Development)

Ministério da Administração Estatal (Ministry of State Administration)

Ministério da Educação (Ministry of Education)

Ministério da Saúde (Ministry of Health)

Memorandum of understanding

Ministério do Plano e Finanças (Ministry of Planning and Finance)


Non-governmental organization

Overseas Development Institute, London

Orçamento do Estado (State Budget)

Organization for Economic Cooperation and Development

Organização da Juventude Moçambicana (Mozambican Youth Organization)

Organização da Mulher Moçambicana (Mozambican Women’s Organization)

Organização de Trabalhadores Moçambicanos-Central Sindical (Organization of

Mozambican Workers-Trade Union Centre)

Plano Annual de Actividades e Orçamento (Annual Plan of Activities and Budget),


Performance Assessment Framework

Programme Aid Partners (GBS donors)

Plano de Acção para a Redução da Pobreza Absoluta (Action Plan for the Reduction

of Absolute Poverty)

Plano Estratégico da Educação (Education Strategic Plan)

Public expenditure management

Public expenditure review

Plano Económico e Social (Economic and Social Plan)

Plano Estratégico Sector Saúde (Health Sector Strategic Plan)

Plano Operacional Annual (Annual Operation Plan), of MISAU

Poverty Reduction and Growth Facility, of IMF

Programa Nacional de Desenvolvimento Agrícola (National Programme for

Agricultural Development)

Fundo Comum de Apoio ao Sector da Saúde (Common Fund for Support to the

Health Sector)

Poverty Reduction Support Credit, of World Bank

Poverty Reduction Strategy Paper

Programa Trienal de Investimento Público (Triennial Public Investment Programme)

Resistência Nacional de Moçambique (Mozambique National Resistance)

RENAMO-União Eleitoral (RENAMO-Electoral Union)

Review of Codes and Standards, of IMF

Sector Expenditure Review

Sistema de Administração Financeira do Estado

Sexually transmitted infections

Sector wide approach

Technical assistance

United Nations

United Nations Development Programme

United States Agency for International Development

Unidade Técnica da Reforma da Administração Financeira do Estado (Technical Unit

for the Reform of State Financial Administration)

Unidade Técnica para a Reforma do Sector Público (Technical Unit for Reform of

the Public Sector)








This book resulted from research commissioned by the British Government’s Department for

International Development (DFID), which, as a major donor to Mozambique, wanted to

deepen its understanding of the nature of the budget process in Mozambique. It encouraged

the authors to adopt a broad political economy approach, rather than the narrow technical

approach often found in studies on the budget process, and to give particular attention to the

institutional framework in which the budget is formulated, approved and executed and the

interests and capacities of the various actors involved in the process.

The authors greatly appreciated the encouragement and support they received from the DFID

Office in Maputo, in particular the Representative, Eamon Cassidy, and the DFID staff

members directly responsible for overseeing the study, Maja de Vibe and Alicia Herbert. A

workshop involving DFID staff and the authors, in June 2004, helped to develop the

conceptual framework for the study and a second workshop, in October 2004, provided an

opportunity for in-depth discussion of the draft text, making many useful suggestions that

helped to refine the analysis and sharpen the conclusions.

However, we must stress that the study was conducted in a completely independent manner

and that the resulting book reflects the views of the authors alone, and so does not represent in

any way the official policy of DFID.

We wish also to acknowledge the contributions of two researchers who assisted our work by

preparing short texts and participating in some of the initial discussions. Aaron Schneider, of

the Institute of Development Studies, University of Sussex, provided invaluable

methodological advice for the conceptual framework of the study, while David Jackson, a

consultant of Wise Owls Organisation, London, with many years of working and research

experience in Mozambique, contributed material on the role of parliament and the political


We were engaged to conduct this research through our respective consultancy companies,

Oxford Policy Management (OPM) and Analítica-RJT. We would like to thank OPM for

facilitating the implementation of the study and in particular Stephen Jones, Director of

OPM’s Economic Policy Programme, for providing access to a number of similar studies

conducted by OPM in other parts of the world, which provided valuable methodological

pointers for our work.

The research involved extensive discussions with a wide range of participants in the budget

process, as well as many observers of the process, in civil society and among donors. The

logistics of setting up these meetings were handled admirably by Leticia Fernandes of DFID

and Kátia Taela, an assistant provided to the team by Analítica-RJT.

Among officials of the Ministry of Planning and Finance, we are particularly grateful for the

information and views we received from José Sulemane, National Director of Planning and

Budget, Pedro Couto, Director of the Office of Studies (Gabinete de Estudos), Domingos

Lambo, Deputy National Director of Planning and Budget, and Amílcar Tivane, Jaime

Manjate, Bruce Byiers and Sam Jones, technical staff members of the National Directorate of

Planning and Budget.

As the research gave special attention to planning and budgeting in the health sector and

included a case study on HIV/AIDS, we also spent much time in discussions with senior

officials of the Ministry of Health and the Conselho Nacional do Combate ao SIDA. We

would like to acknowledge in particular the assistance we received in the Ministry of Health

from Ernesto Mazivila, the acting Director of Planning and Cooperation, Gertrudes

Machatine, the Director of Administration and Management, Avertino Barreto, the Deputy

National Director of Health, Rui Bastos, of the Central Hospital of Maputo, Stijn Broecke,




technical staff member in the Directorate of Planning and Cooperation, Francisco Saúte,

coordinator of epimediological surveillance for sexually transmitted infections, HIV and

AIDS, and Carla Silva, head of the Repartição das Doenças Não Transmissíveis. Useful

information was also provided by Enoque Ngomane and Lauren Wojtyla at the Conselho

Nacional de Combate ao SIDA.

As the research extended beyond the executive to parliament, we attached particular

importance to understanding how the budget process is regarded by deputies, particularly in

parliament’s Planning and Budget Commission. We were fortunate to receive detailed

briefings from both the President and the Vice-President of the Commission, Virgínia Videira

(FRELIMO) and Abel Mabunda (RENAMO-UE), as well as from a third deputy, Máximo

Dias, Secretary-General of MONAMO. We are very grateful for their information and


We also held extensive discussions with a wide range of civil society organizations, with a

view to understanding whether and how these organizations play a role, through dialogue with

the government or other means, in influencing the budget process. We would like to thank in

particular José Negrão, Coordinator of the G-20 network of civil society organizations,

Boaventura Mondlane, member of the Executive Secretariat of the Organização dos

Trabalhadores Moçambicanos-Central Sindical, Kekobad Patel, Vice-President of the

Associação Industrial de Moçambique, Paulo Fumane and Otília Pacule of the Confederação

das Associações Económicas, and Miquelina Menezes of the Associação Moçambicana de

Economistas. Important discussions were also held with leaders of NGOs working on

HIV/AIDS, including Kerry Sherr and Helen Smuts of Health Alliance International, George

Jagoe of the Clinton HIV/AIDS Initiative, Graça Neves of Kindlimuka, David Melody and

Vera da Silva of Médecins sans Frontières, and Giovanni Guidotti of Santo Egidio.

These discussions with civil society organizations extended to editors and journalists in the

print and broadcast media, among them Simão Anguilaze, Director of Information at

Televisão de Moçambique (TVM), Maria Esperança Macovele, presenter of the TVM

programme «Semanário Económico», Jeremias Langa, Director of Information of STV,

Manuel Veterano, Chairman of the Board of Rádio Moçambique, Boaventura Mandlate,

Director of Information of Rádio Moçambique, Fernando Lima, Chairman of the Board of

MediaCoop, and Fernando Gonçalves, Editor of Savana.

Finally, in view of the importance of external assistance in financing Mozambique’s budget,

we talked at length to officials of a number of donor agencies. Apart from DFID officials, we

received briefings from Perry Perone, Representative of the International Monetary Fund in

Mozambique, Humberto Cossa and António Franco at the World Bank office, and Telma

Loforte and Rogers Dhliwayo at Swiss Cooperation. We were also fortunate to receive

comments on the draft text from António Franco and Telma Loforte, for which we are

sincerely grateful.

We hope we have done justice to the advice and inputs of all those whom we consulted, while

absolving them of any responsibility for the arguments and opinions expressed in the final


Tony Hodges and Roberto Tibana

Maputo, December 2004








This book is about the nature of the budget process in a highly aid-dependent developing

country with weak institutions. It argues that, while external aid helped to rebuild

Mozambique after a long and devastating civil war and has contributed to economic growth, it

has also had perverse side-effects, fragmenting government planning, budgeting and

management and weakening national ownership of policymaking. Since Mozambique also

has a weak civil society and a weak parliament that is not yet able to act as an effective check

on the executive, high aid dependence means that the budget process essentially involves only

two actors, the executive and foreign donors. Accountability to donors is much stronger than

it is to Mozambican society.

Mozambique is a country with a large structural dependence on external aid. Because of the

low level of development of the economy, the tax base is very small and internal revenue

accounts for only about half of total government expenditure (49 per cent in 2003). At about

15 per cent, the aid/GDP ratio is twice as high as the average for sub-Saharan Africa.

The high levels of aid, most of it in grant form, have made it possible to bring about a

substantial real increase in government expenditure over the past decade, financing a large

expansion of public infrastructure and services, while also making it easier for the

government to restore overall fiscal stability – although there was a partial reversal in 20002003. However, high levels of aid may have reduced the incentive to use resources more

effectively and efficiently. The government has done little to improve the allocation of

resources between and within sectors or to shift the pattern of territorial distribution to redress

historical disparities.


Institutional setting of the budget

The budget process takes place in an institutional framework of constraints and incentives,

which limit what the actors in the process can do and motivate them to behave in certain

ways. These are the rules which determine how the “budget game” is played. In Mozambique,

three key features of the budget game stand out: the fragmented, incremental nature of the

budget process within government; the role of the donors as the main interlocutors of the

government; and the weakness of internal demand, through parliament or from civil society,

for improved management of public finances.

Fragmented, incremental nature of the budget process. Within the executive, planning and

resource allocation are highly dispersed, mainly due to the diversity of funding sources. There

is not a single unified budget, in which the Ministry of Planning and Finance (MPF) plays a

dominant role, but rather a plurality of partially overlapping budget systems. In large sectors

such as education, health, agriculture and public works, donor assistance, much of it offbudget, often outweighs the resources made available through the State Budget (OE). Even

within the line ministries, power over resources is dispersed, due to donor funding of projects

managed by different directorates and departments.

As far as the OE itself is concerned, budget formulation has been driven only to a very limited

extent by policy objectives and priorities. The only targets that guide budget formulation are

those agreed with the IMF on overall fiscal aggregates and, since the adoption of the Action

Plan for the Reduction of Absolute Poverty (PARPA) in 2001, a target of allocating 65 per

cent of government expenditure (minus interest payments) to the plan’s six priority sectors.

Apart from this, MPF does not apply clear criteria for allocating expenditure between and

within sectors and provinces. In practice, MPF simply applies an incremental approach to

expenditure allocations, in so far as rising levels of resources permit increased spending.

Recurrent expenditure allocations largely reflect the existing location of public administration

and services, while most investment decisions are perceived to be pre-determined by donors.




The government does not use its Medium Term Fiscal Framework (Cenário Fiscal de Médio

Prazo) to allocate resources in a more prioritized, policy-driven manner and the input-based

nature of the budget encourages an incremental approach because it provides no clear way of

linking resource allocation to planned activities, outputs and outcomes. Furthermore, the use

of policy-driven resource distribution criteria is hindered in practice by the lack of

consolidated data on the totality of available resources, due to the large amount of resources

(mainly from donors) that are fully or partially off-budget.

Recent reforms, initiated by the Law on the State Financial Administration System (Law n.º

9/2002), have focused mainly on budget execution weaknesses, notably in public accounts

and the treasury system. By contrast, the government has given little attention to overcoming

the weaknesses in budget formulation. In fact, there has been little demand from the top of

government for measures to strengthen the links between policy objectives, plans and

expenditure allocation. Possibly this reflects a desire to avoid having to make hard choices

about trade-offs. An incremental approach to budgeting is simpler, because it is less

contentious. The generally laisser faire style that has characterized government leadership in

Mozambique in recent years would lend weight to this view.

The consequences of external aid. While technical or capacity constraints partly explain the

weaknesses described above, they are reinforced by the incentives created by external aid.

Traditional project assistance has fragmented government planning, budgeting and

management. Since projects are negotiated directly between donors and line ministries, or

individual directorates within them, and the associated funds are often off-treasury (if not

entirely off-budget), this form of aid financing weakens the authority of MPF vis-à-vis the

line ministries and undermines executive accountability to parliament. Furthermore, the sector

ministries’ direct access to resources from donors reduces the incentive for the government as

a whole to engage in hard debates about resource allocation. Ministers rationally prefer to

direct their efforts towards maximizing access to the funds available from donors at sectoral

level, which they see this as the easiest way to achieve their priorities, and this obviates the

need to engage in political debates over development strategy and run the risk of infighting

over prioritization.

To overcome the disadvantages of traditional project assistance, many donors have begun to

provide some of their assistance to Mozambique in the form of “programme support”. This

takes the form either of untied common funds within individual line ministries, linked to

sector strategic plans, or general budget support (GBS). The sectoral common funds have

helped to reduce the fragmentation caused by project financing within line ministries. GBS is

untied budgetary assistance to the government as a whole. It adds to the resources available

for allocation in the “normal” budget process and increases MPF’s influence vis-à-vis the

sector ministries.

For donors, the high fiduciary risk associated with GBS is being offset by the opportunity to

engage in an enhanced policy dialogue with the government and to link future GBS

commitments to government performance. GBS should provide a strong incentive to the

government, and MPF in particular, to “take command” of budget policy, allocating resources

in a more rational, policy-driven manner than in the past. However, two major constraints

limit the extent to which this is really likely to happen: first, the government’s weak capacity

to improve budget management; and, second, the fact that GBS still accounts for only about

two fifths of aid to Mozambique.

Weak internal demand. The strong influence of donors contrasts with the weakness of the

internal pressure on the government from Mozambican society. Although parliament plays a

formal role in debating and approving the OE, within the multi-party system introduced by

the 1990 constitution, its ability to act as a real check on the executive in the budget process

has been quite limited in practice. Besides capacity limitations and the polarized nature of

inter-party relations in parliament, these constraints include: first, the poor presentation of the

budget document, which is difficult for deputies to understand and analyse; second, the fact




that tax rates are not set as part of the budget legislation submitted to parliament; third, the

wide latitude given by law to the executive to reallocate expenditure after parliament’s

approval of the budget; fourth, the off-budget problem, which limits parliament’s oversight of

the budget; fifth, the long delays in submitting the audited Conta Geral do Estado (CGE) to

parliament; and sixth, the failure of the executive to involve parliament in the debate,

approval and monitoring of the PARPA and other medium term planning instruments, which

should provide the strategic framework for the annual budget.

Despite the more positive environment for civil society since the end of the one-party system,

including the establishment of various frameworks for a formal exchange of views between

civil society organizations and the government, budget issues have not become a focus of

interest for civil society organizations such as business associations, trade unions and NGOs.

Media coverage of budget issues has also been weak and there is no evidence that government

budget policy is influenced directly by the media. More generally, the development of civil

society organizations has been held back by deep-rooted structural constraints, including the

predominantly peasant nature of the rural population, which is atomized and unorganized, the

small size of the formal economy (and thus the weakness of both the trade unions and private

sector bodies) and the low levels of education and literacy, which limit access to information

and ideas, particularly through the print media.


Interests and capacities of key actors

The way in which the institutional setting affects behaviour depends on the interests and

capacities of the actors. Incentives, for example, interact with the interests of certain actors,

stimulating them to behave in certain ways. On the other hand, weak capacity limits the extent

to which actors can respond to incentives or find ways of dealing with constraints.

The executive. The main interest of the politicians involved in overseeing the budget process

is to ensure the continued functioning of the state apparatus and the payment of salaries to the

countries’ civil servants. They also need to avoid the risk of public service disruption creating

an impression of administrative incompetence that could damage the public image of the

government. In so far as a growing internal resource base supplemented by donor support

makes it possible to expand services, this also serves the politicians’ interest, as they need to

be seen to be responding to the population’s concerns and needs. However, as noted above,

there is little motivation at the highest political levels to think strategically about resource

allocation or to introduce structural changes in the composition of expenditure, either

territorially or sectorally.

On the capacity side, the low supply of university graduates and strong competition in the

labour market make it difficult for the public sector in Mozambique, including both MPF and

the line ministries, to recruit and retain highly qualified personnel. MPF’s capacity is

constrained by a lack of personnel with relevant specializations, which prevents it from

playing a “challenge” function vis-à-vis the sector ministries, and by the organizational

separation of planning and budget functions, which reinforces the technical divorce between

plans and resource allocation. This organizational divide is replicated in the line ministries.

More generally, low salary levels and the informal tolerance of income-diversification

activities among government staff have resulted in large numbers of civil servants,

particularly those with the best qualifications, engaging in consultancy and other remunerative

activities alongside their official jobs, with obvious costs for productivity and conflicts of

interest. These problems are compounded by an antiquated system of human resources

administration, which primarily rewards formal qualifications and years of service, rather than

on-the-job performance, resulting in weak staff incentives.

Parliament and political parties. There is not really a FRELIMO budget policy distinct from

that of the state. The content of the budget comes from the executive, rather than the party.

However, the approach to some key budget issues reflects the ruling party’s general




ideological stand: for example the taboo on discussing issues of territorial equity in resource

allocation and service delivery reflects FRELIMO’s stance on national unity, which leaves no

place for acknowledging sectional interests.

As for RENAMO, it is “against” FRELIMO, but short on distinctive policies. It boycotted

parliament for a year after the 1999 elections, which it claimed had been rigged, and

thereafter failed to put forward alternative policies, contributing little to the budget debate in

parliament. This ineffectiveness may be a consequence of the highly personalized nature of

the party and the fact that, as a past presidential candidate, the party leader does not hold a

seat in parliament. There is no RENAMO shadow finance minister.

The way the budget is debated in parliament also reflects characteristics common to almost all

deputies, irrespective of the parties to which they belong. The deputies generally have neither

the academic nor professional background to analyze complex budget issues. Furthermore,

parliament’s general secretariat does not have the technical personnel that would be needed to

provide research services to the commissions and deputies. Apart from their own salaries and

benefits, deputies are interested mainly in micro issues, such as individual projects in their

constituencies, and sensitive political issues such as delays in the payment of salaries, rather

than broader fiscal questions or major expenditure allocation issues.

None of the parties has more than a token capacity for policy development, particularly in

technical areas such as public finance, although FRELIMO can call on ministers and party

members in the higher echelons of the civil service to contribute to policy discussions and the

preparation of documents, giving it a huge advantage compared to the other parties.

RENAMO is disadvantaged both by its lack of direct government experience and by its

inability to draw on the expertise of senior civil servants.

Non-state actors. Several common factors account for the lack of interest in budget issues

among civil society organizations. First, the low level of development of the formal sector of

the economy means that the tax base is narrow, undermining the notion of a tax/services

“compact”, and the resulting lack of public interest in the budget is accentuated by the fact

that tax rates are set separately from the budget legislation. Second, the fact that much public

expenditure is financed not by the government but by donors makes the budget seem less

important than it would normally be. Third, the small size of the private sector, along with the

practice of patronage as a tool for political survival, results in a highly concentrated and

politically connected elite, which has little interest in developing and articulating alternative

policies. In some cases, CSOs may be co-opted by the government or not wish to venture into

potentially “hot” political issues.

On the capacity side, few civil society organizations have the technical capacity to analyze

budget issues. Private sector bodies do have the resources needed to hire technical expertise,

but they have chosen not to give priority to budget issues. As for the trade unions, they have

suffered an erosion of membership in the past decade, limiting their ability to finance

research, despite the benefits of the recognition of trade union freedoms in 1990. Finally,

while several media organizations report on economic news, their coverage of the budget is

limited and weak. In this respect, the media houses are simply reflecting the low public

demand, even among niche audiences. In addition, their inability to provide salaries that

would attract and retain university graduates is a severe capacity constraint on coverage of

complex technical subjects like the budget.

Donors. Donors are heterogeneous in their motives, strategies, rules and procedures.

Fiduciary risk, restrictive internal rules and/or a desire to “show the flag” have prevented

some donors from evolving from traditional project aid to common funds and GBS. On the

other hand, many donors have been influenced by the new thinking on aid effectiveness,

which has emphasized the importance of partnership, based on government commitment and

leadership, and the use of government systems and procedures, supported by necessary

reforms and capacity building. These donors have accepted the high level of fiduciary risk

associated with such aid modalities, because the benefits potentially outweigh the costs. A




major incentive for many has been the opportunity to engage in high-level dialogue with the

government. Donors have also moved some common funds and projects on-budget. However,

much external aid remains off-budget and most donors are likely to continue using a broad

range of aid modalities to balance different interests and hedge against risks.


HIV/AIDS treatment and the budget

Government policy on HIV/AIDS provides an example of how the institutional setting and

the interests and capacities of key actors affect budgeting in a specific area of high policy

importance. Mozambique has one of the highest HIV prevalence rates in the world (13.8 per

cent), so the budgetary implications of an effective national response are potentially immense.

Until very recently, the strategy for combating HIV/AIDS in Mozambique centred on

prevention and failed to recognise the link between prevention and treatment. This narrow

approach reflected a lack of awareness of the rapid advances in treatment, concerns about the

capacity of the National Health Service (SNS) to deliver a treatment programme and, above

all, the presumption (in the government and among donors) that Mozambique was too poor to

afford such a programme. Policy began to shift in 2001, when the Ministry of Health

(MISAU) decided to provide anti-retroviral treatment to patients who could afford to pay. By

2002, policy was evolving towards a major scaling up of access to highly active anti-retroviral

therapy (HAART) for people living with HIV/AIDS. This culminated in the development of

the Plano Estratégico Nacional de Combate às DTS/HIV/SIDA – Sector Saúde 2004-2008,

which is costed at almost $500 million and aims to scale up the provision of HAART from

just under 8,000 patients in 2004 to more than 130,000 in 2008, backed up by strengthening

of the capacity of the health system to provide testing, counseling and treatment.

Two external developments created an incentive for scaling up: first, the steep fall in drug

prices, which by 2003 had dramatically cut the cost of treatment, and second, the emergence

of major international initiatives to fund integrated HIV/AIDS programmes, including

treatment. These opportunities were seized because there were also strong interests involved.

First, the epidemic affected the politically influential elite as much as anyone else. Second, a

sense of duty goaded some health professionals to help design the policy instruments and

technical guidelines. Third, the civil service needed an institutionalized treatment policy, as

ministries with large work-forces, such as MISAU itself, risked losing thousands of trained

staff in the absence of treatment.

However, despite the fact that 1.4 million people are infected with HIV, there has been almost

no public pressure for treatment. National NGOs working on HIV/AIDS or representing

people living with the disease have had little impact because of their small membership and

weak capacity, which in turn reflect the structural constraints on civil society development

generally. It is also striking that parliament and the political parties have been silent on the

issue of HIV/AIDS treatment, due to the weakness of public pressure, deputies’ own limited

capacity, the executive’s desire to keep issues like HIV/AIDS policy out of the parliamentary

arena and the opposition’s lack of vision and weak leadership on the issue.

In summary, apart from some inputs from health professionals and specialized international

NGOs, the development of policy on HIV/AIDS treatment and its financing has remained

almost entirely within the executive, while being heavily influenced by the changes in the

international context and the discussions with donors. At the highest political level of

government, there was no felt need to engage in trade-offs in resource use to meet the costs of

this major new area of service delivery, as the financing available from donors removed the

need to reorient the allocation of internal resources or the budget as a whole. In short, scaling

up HAART has been a reaction to the incentives created by the emergence of international

donor initiatives, allied with the steep fall in drugs prices, not domestic pressure strong

enough to lead the government to allocate significant internal resources to the fight-back.





The almost complete absence of a domestic “demand function” outside government for

improvements in the budget is a manifestation of deep-rooted structural features of the

Mozambican situation that will change only gradually as the country achieves higher levels of

economic and social development. Given this reality, as well as the government’s heavy

dependence on external aid, it is perhaps not surprising that the most important dialogue on

budget policy and performance is now external, between the government and donors.

Since more than half of aid is channeled through projects and common funds, in the line

ministries, there is little incentive for the government to attempt to allocate global resources

rationally. Rather than engage in potentially damaging political battles over resource

allocation, ministers prefer to negotiate additional resources directly with donors. In short, the

preponderance of non-GBS aid flows has had the effect of absolving the government of the

need to exercise leadership in resource allocation. The incentives would only change radically

if GBS’s share of total aid rose to much higher levels, but this is unlikely as most donors have

opted for a mixed financing strategy to balance the advantages of non-earmarked aid against

high levels of fiduciary risk and the disadvantages of “losing control” over the setting of

priorities and allocation of resources. Some donors have opted not to contribute to GBS at all.

While the share of GBS may not increase enough to transform the incentive structure

decisively, other steps may also help move incentives in that direction. One is the shift from

project financing to common funds at sectoral level, which is already encouraging line

ministries to assume greater leadership of the policy, planning and budgeting processes in

their own sectors. Second, some attempts are being made to reduce the proportion of external

assistance that is off-budget, so that the government has fuller information on the resources at

its disposal and how these resources are used.

However, even with these changes in aid modalities, there is no guarantee that the

government will manage its resources effectively and efficiently in pursuit of clearly

articulated pro-poor goals. First, for many years it will continue to face severe capacity

constraints, particularly in human resources, due to its difficulty in competing for skilled

personnel in the labour market. Moreover, the government does not yet have the systems and

tools needed to manage resources in a prioritized, policy-driven way, such as an effective

medium term fiscal framework and a programme-based budgeting methodology. These

constraints highlight the importance of capacity development, going beyond the short-term

technical assistance favoured by most donors, as well as an extension of the reforms in public

financial management beyond their current focus on the accounting and treasury dimensions

of budget execution to encompass the methodologies and procedures that determine the

content of the budget.

More fundamentally, however, if internal pressure on the government remains weak, capacity

development will not be enough, as commitment to pro-poor priorities is likely to be

tempered by the more narrow “predatory” interests of the leading families that constitute the

politico-business elite. To some extent donors can act as a “proxy” restraint on the elite in the

absence of strong internal checks and balances. Nonetheless, there are limitations to this – and

some inherent contradictions. Much more important in the long run will be the development

of internal demand for improved budget policy and performance.

A crucial development will be the emergence of a more effective parliament, with resources

for proper scrutiny of the executive. Even more important, at a deeper societal level, will be

the empowerment of the mass of the population, including the poor, so that they have the

capacity to obtain access to information and ideas, organize in support of their rights and give

voice to their concerns, and can thereby influence government policy and performance.

Ultimately, however, such positive developments are likely to result mainly from structural

changes -- reductions in poverty, illiteracy and vulnerability – as the country reaches higher

levels of economic and social development.












This book is a case study about the nature of the budget process in a highly aid-dependent

developing country with weak institutions. In Mozambique, aid from donors finances just

over half of public expenditure. The basic hypothesis is that, despite the political

democratization introduced by the 1990 Constitution, the Mozambican government’s

accountability to the country’s citizens remains quite weak. Meanwhile, donors have come to

exert enormous influence, affecting the incentive structure in a way that fragments the role of

the executive in policy-making, planning and budgeting. Only recently and to a limited extent

has this fragmentation been reduced by donors shifting part of their aid to sector-wide

programmes and more recently to general budget support.

Actors, institutions, process and outcomes. The analysis employs a conceptual framework

that focuses on the nature and role of institutions and actors in the budget process, thereby

affecting budget outcomes. By budget outcomes we mean the actual results of the executed

budget, while the budget process concerns the way in which the budget is formulated and

executed. Institutions are the ‘rules of the game’ within which the budget process takes place

and actors are the various agents directly or indirectly involved in or influencing the budget

process and its outcomes.

One way to address outcome issues would be to apply Musgrave’s categories of stabilization,

allocation and distribution. Stabilization refers to fiscal aggregates that shape the size of

government and rates of interest and inflation. Allocation refers to the pattern of expenditure

between and within government bodies, sectors or territorial units, and distribution refers to

the social impact of expenditure on different groups and strata of society [Musgrave, 1959].

However, we do not attempt to make an in-depth assessment of budget outcomes in

Mozambique. Other studies, such as the public expenditure review and related sector

expenditure reviews conducted in Mozambique in recent years, have explored these questions

more fully.1 Instead, our main focus is on the budget process, its institutional framework and

the actors involved. On outcomes, we will simply highlight a few key features of

Mozambique’s recent fiscal performance as a starting point for the analysis of the budget


In an ideal situation, budgeting would be linked to the policy framework and planning. The

government would set certain objectives and devise multi-year programs to achieve them,

within the constraints of a projected resource envelope. The annual budget would translate

these long-term programmes into costed annual activities to be implemented by the agencies

of government. The outputs and impact of these activities could then be measured to check

the performance of government in fulfilling annual budgets and meeting long-term goals

[Premchand, 1993].

In Mozambique, a set of long and short-term instruments for planning and budgeting would

appear to perform the roles described in this ideal construct. Yet, in practice, the instruments

of long-term planning are almost completely disconnected from annual plans and budgets,

and even annual budgets are incompletely implemented. In addition, the state budget only

partially covers the actions of government, with much expenditure falling outside the budget

and conducted almost completely out of view of those who approve the budget. In this book

we will explore the reasons why planning and budgeting takes place in a confusing and



Stage 2 of the public expenditure review, conducted jointly by the World Bank and the government,

analysed the pattern of inter- and intra-sectoral expenditure and its implications for efficiency,

effectiveness and equity [World Bank, 2003]. Key inputs into Stage 2 included the sector expenditure

reviews carried out in education [MPF and MINED, 2003], health [MPF and MISAU, 2004], water and

roads, as well as an expenditure tracking and service delivery survey conducted in the health sector in

2002 [MPF and MISAU, 2003].




disconnected way and why significant resources continue to be received and spent outside the

official budget channels.

The analysis applies a political economy approach, highlighting the roles of actors and

institutions. We hypothesize that the way actors operate in and interact with institutions

shapes the problems that arise in the budget process and influences budgetary outcomes. In

stating this, we do not make the claim that everything is wrong with public finance in

Mozambique. Some things are right. Still, a focus on the actors and institutions responsible

for process problems can help identify interventions that would lead to improvements.

Our approach is action oriented. We focus on the actors that possess a degree of agency

during the budget process. To be actors, they must be able, at least conceptually, to choose

among various options during the budget process [Giddens, 1994] and act on these options in

such a way as to influence the budget, either directly or indirectly. We concentrate therefore

on the limited set of players that intervene directly in the budget process or influence it from

the outside. The main categories of actors we consider are the executive (the Ministry of

Planning and Finance, the line ministries, the Council of Ministers), the legislature and

political parties, other non-state actors, such as the media, trade unions, professional

associations, NGOs and the private sector, and, given the importance of donor assistance,

external actors.

Through their actions, actors pursue interests. These cannot simply be reduced to material

interests determined by economic structures. Material interests are certainly present and they

are frequently expressed in zero-sum competition over the distribution of resources. Yet, in

budgeting, not all competition is zero-sum, and positive-sum gains, such as economic growth,

offer opportunities for cooperation. Further, even in the resource-focused world of budgeting,

non-material interests such as political power, cultural cohesion or professional ethics can be

surprisingly relevant. In budgeting, both material and non-material interests are important to

understanding behaviour.

In addition to interests, we are also concerned with actors’ capacities, which can take two

forms, technical and political. Technical capacity refers to the skills, information and

resources necessary to analyse budget issues and engage in the budgetary process. Budget

documents are incredibly complex and the data provided in multiple tables and annexes are

difficult to read and understand. Multi-year plans introduce even more complexity. Few actors

possess the technical capacity to prepare and assess these estimates, and those with technical

capacity frequently dictate much of the budgetary debate [Wildavsky and Caiden, 1980].

The second dimension of capacity is political capacity. Political capacity refers to the power

to influence others by persuasion, negotiation, force or other means. There is a wide range of

types and degrees of political capacity across budgetary actors. Some actors who should be

important simply lack the political influence to have much impact.

We also recognise that there is more to agency than simply pursuing interests with different

capacities. Actors operate within institutional boundaries that set the ‘rules of the game’ for

interaction. These rules define what options exist at any given moment and include or exclude

actors. In some circumstances, institutions are merely epiphenomenal expressions of the

interests of one actor or another. Where this is the case, it makes more sense to focus on the

actors themselves. In other circumstances, however, institutions can have relative autonomy

[March and Olsen, 1989]. They provide incentives to actors that make certain actions and

interests more appealing. At the same time, institutions also present constraints that limit

agency by ruling out certain options. Actors may like to influence processes and outcomes in

one direction, but the incentives and constraints presented by institutions may lead them in

another. Process and outcomes are sometimes biased by institutions and cannot always be

reduced to interests and capacities.

One more conceptual remark is relevant about institutions at this point. Institutional rules of

the game can be informal or formal. Formal rules are usually written, either legally or




bureaucratically, and are monitored and enforced with sanctions that are public and often

backed by state power, although this is not always the case in relatively weak states such as

Mozambique. Informal rules operate on the margins and “under the table”, and actors may

know they exist while rarely articulating them. To understand what really goes on in

budgeting, it is necessary to consider both formal and informal institutions. At times they

operate in a contradictory way, as when a formal law prohibits bribery, but everyone knows

that bribes are an informal requirement for securing government contracts. Alternatively,

informal rules can reinforce formal ones, as when parliamentary legislation would not pass

were it not for the party discipline that senior leaders impose on their deputies in parliament.

The characterization of actors and institutions will help us understand the factors that shape

process problems. To make the framework complete it must go beyond a static design to

include a dynamic approach. It must address the ways in which a change in one variable

brings about changes in others. In static form, actors operate within institutions to shape

budgetary processes and then budgetary processes shape fiscal outcomes. A dynamic

approach asks what happens when there is a change in one of the elements in this set of


The concepts, connections and feedbacks are displayed in Figure 1.1. One can read from left

to right how the actors and institutions interact to produce process problems and these unfold

in outcome weaknesses. The loops that feed back in the other direction demonstrate the

complexities of the causal connections.

Figure 1.1


Relationship between actors, institutions, budget process and budget outcomes



























planning and budget





Overall fiscal



Off-budget accounts









Sectoral and




The first step in the chain links actors and institutions to processes. As actors respond to or

are constrained by their institutional setting, they shape the way budgetary processes unfold.

Changes that will be important to trace at one level will be the emergence of new actors or the

entrance of previously excluded actors. Equally important, actors may experience an increase

or decrease in their political or technical capacity or simply change their interests over time.

In addition, institutions can shift in a number of ways, as a result of sudden changes, such as

new legislation or external pressures, or through more gradual evolution. Behind these

institutional changes may lie deeper structural changes, such as urbanization, improving

literacy and levels of education, and changes in the economic and social structure.

The second step in the chain links processes to outcomes. As budget process problems are

addressed, they will trickle through to fiscal outcomes. For example, efforts to create closer

links between budgeting and planning should have repercussions in fiscal outcomes, although

it will be interesting to evaluate whether these repercussions are the ones intended. Patching

one kind of process problem may open up others of a more serious kind.

The causal flows between all these elements are complex, with many different elements

interacting together and various feedback loops further complicating attempts to understand

the sources and outcomes of change. To cite just one hypothetical example, one set of actors,

the donors, might introduce changes in the modalities of aid, such as a large-scale shift from

project aid to general budget support. This would radically improve budget coverage (a

process change), which in turn would change the institutional framework of incentives and

constraints governing the relationship between the Ministry of Planning and Finance and the

sectoral ministries, to the benefit of the former.

This conceptual framework has laid out some initial concepts: actors, institutions, process and

outcomes. Actors are agents in the budget process that can be characterized by their interests

and capacities. Institutions are the formal and informal rules of budget interaction that present

the actors with incentives and constraints. Problems in budget processes are manifested by

failures to formulate and execute budgets with clearly designated government outputs and

social outcomes. Fiscal weaknesses are manifested in unsustainable deficits, ineffective and

inefficient expenditure patterns and undesirable distributional impacts. Together, these

concepts form the basis of our political economy approach to the budget. The causal links

between them flow basically from actors and institutions to processes and outcomes, but are

inherently complex because of the large number of factors that may be at play and the

feedback effects that can reverse the direction of causality.

Focus on expenditure. For reasons of practicality, given the wide range of issues that would

have to be addressed in a comprehensive study of the public financial management system,

from budgeting through execution to auditing, it was decided to limit the scope of inquiry by

focusing mainly on the determinants of the pattern of expenditure. The study therefore

concentrates on budget formulation, including the links to policy and planning, although some

issues concerning execution (including budget revisions and readjustments during the course

of execution) are considered in so far as these affect the actual, as opposed to the approved,

allocation of expenditure.

The overall research question was therefore formulated as follows: ‘How is the pattern of

government expenditure affected by the institutional framework of the budget process (the

“rules of the game” constituted by formal and informal constraints and incentives) and the

interests and capacities of state, non-state and external actors?’ This requires the identification

of the actors, their roles in the budget process, the interests and capacities that determine their

engagement in this process, the source and nature of demand for changes in the process and

the constraints and incentives that the institutional set-up creates for the actors to behave in

one or another way to shape the budget process and its outcomes.

In order to make the analysis more concrete, it was decided to complement the general

analysis with a case study that would illustrate how institutions and actors affect the budget

process and outcomes in a specific thematic area. The theme selected was the roll-out and




scaling-up of highly active anti-retroviral therapy (HAART) for HIV/AIDS patients, a subject

selected due to its importance in a country with one of the highest HIV prevalence rates in the

world and its potentially huge budgetary implications. The following research question was

set for the case study: ‘How are decisions regarding public expenditure on anti-retroviral

treatment for HIV/AIDS patients affected by the institutional framework (of constraints and

incentives) and the interests and capacities of state, non-state and external actors?’

Sources of information. Due to its political economy approach, the study had to go beyond

the technical focus of most studies of the budget system in Mozambique, which have been

driven mainly by the requirements of donors engaged in providing budget support or assisting

technical improvements of the system.2 The research had to take into account the findings of

these technical studies, but broaden the scope of analysis to address the broader dimensions of

the institutional setting and the interests and capacities of the actors involved directly or

indirectly in the budget process. This involved the consultation of literature in a range of

relevant subject areas and interviews with numerous actors, in MPF, the line ministries,

parliament and the political parties, the donor community and civil society. The documents

and other literature consulted in the course of the study are listed in the bibliography.

Regarding the line ministries, special attention was given to the Ministry of Health, due to the

health sector’s importance as a priority sector for poverty reduction and its large share in

overall government expenditure (13 per cent in 2003), as well as the Ministry’s importance as

a source of information for the case study.

Structure of the book. Following this introductory chapter, Chapter 2 places Mozambique’s

budget in context, notably by highlighting some of its structural features, which reflect the

low stage of development of the economy and widespread poverty. Chapter 3 then highlights

some key weaknesses in budget outcomes and in the budget process in Mozambique, as a

starting point for the analysis.

Chapter 4 discusses the institutional framework of the budget process, analyzing in turn how

it is affected by the political system, the role of parliament and the various components of the

executive, the scale and modalities of external aid and the nature of civil society and its

interaction with the state. This institutional framework provides the setting in which to

analyse the interests and capacity of different actors involved in the budget process, in

Chapter 5. This starts with the legislature and the political parties, and then covers in turn the

executive, the donors and a series of non-state actors, including the media, trade unions,

professional associations, NGOs and the private sector.

The case study, on government expenditure on anti-retroviral treatment for HIV/AIDS

patients, is presented in Chapter 6. Finally, Chapter 7 draws a number of conclusions from the

study, highlighting the factors that might bring about future changes in the budget process in





These earlier studies have included a Country Financial Accountability Assessment, conducted by the

World Bank and the IMF in 2001, two stages of a public expenditure review (PER), conducted jointly

by the Government and the World Bank in 2000-2003, fiduciary risk assessments commissioned by the

donors providing general budget support to Mozambique, the IMF’s Report on Observance of

Standards and Codes and a series of studies carried out in 2002-2004 by the National Directorate of

Planning and Budget (DNPO) in the Ministry of Planning and Finance (MPF) with technical assistance

financed by the UK’s Department for International Development (DFID). Many of these studies are

included in the bibliography.












The budget process in Mozambique cannot be properly understood without taking into

account basic structural features of the economy, society and polity, which in turn reflect the

country’s historical background.

Historical background. As a Portuguese colony, Mozambique’s development was stunted.

Even at the very end of the colonial period, the population was overwhelmingly rural and the

economy was predominantly agrarian, despite the emergence of some industrial activities and

the development of modern cities in the final decades of Portuguese rule. Educational levels

were very low and most of the population was illiterate. Portuguese settlers, who numbered

about 300,000, controlled the modern sectors of the economy and occupied most of the

skilled jobs in both the private sector and the public administration. Furthermore, the

dictatorial Salazarist regime in Lisbon, which considered Mozambique an integral part of

Portugal, with the status of a ‘province’, made no preparations for independence, failing to

build the institutions needed for a stable transition. Instead, it repressed all stirrings of anticolonial resistance, eventually driving nationalists to take up arms.

The liberation war that began in 1964, under the leadership of the Frente de Libertação de

Moçambique (FRELIMO), culminated in independence in 1975, following the overthrow of

the dictatorship in Portugal the previous year. Power was handed over to FRELIMO without

elections, establishing the foundations for a one-party system of government that was to last

for a decade and a half. Meanwhile, more than 95 per cent of the settlers left the country in a

sudden mass exodus, throwing the economy into crisis.

FRELIMO, with very limited technical capacity, tried to overcome the crisis by nationalizing

most private companies, banks and commercial farms, especially those abandoned by the

departing settlers, and then attempting to manage the economy through a system of

centralized planning. At its third congress in 1977, FRELIMO formally adopted MarxismLeninism as its guiding ideology. In the Cold War context of the time, these moves were

supplemented by a close alliance with the USSR and other Eastern bloc countries, which had

assisted FRELIMO’s liberation war.

The defeat of colonialism and the emergence of a nationalist government committed to

socialism and allied to the USSR prompted a hostile response from the white minority regime

in power in neighbouring Rhodesia and later from South Africa. They exploited internal

discontent within Mozambique, fuelled by political repression, economic decline, ethnic

rivalries and misguided attempts at social engineering, by sponsoring an armed revolt

spearheaded by the Resistência Nacional de Moçambique (RENAMO). This plunged

Mozambique into a brutal and destructive civil war, which continued for 16 years, forcing

more than a quarter of the country’s population to flee abroad as refugees.

Changes in the international context (the demise of the white minority rule in southern Africa

and the collapse of the Soviet Union), along with the stalemate in the civil war and the related


ITEP webstie (archive)

Record updated on : 04 July 2012
Record id : 9783