eotN.353G Guinea-BissauIntegrated Poverty andSocial Assessment (IPSA) Volume I 34553-GW No. Report Report No. 34553-GW Guinea- Integrated Poverty and Social Assessment (IPSA)

Transition Authorized Authorized Disclosure Disclosure Public Public from Post Conflict to Long-Term Development: Policy Considerations for Reducing Poverty (In Two Volumes) Volume I: Main Report

May 26, 2006 PREM 4 Africa Region Public Disclosure Authorized Authorized Disclosure Disclosure Public Public Public Disclosure Authorized Authorized Disclosure Disclosure Public Public

Document of the World Bank Public Disclosure Authorized Authorized Disclosure Disclosure Public Public

MOE Ministry of Education MOH Ministry of Health MOL Ministry of Labor NGOs Non Governmental Organizations NRRP National Reconciliation and Rehabilitation Program OHADA Organization for Harmonization of Business Law in Africa PAIGC Partido African0 para a IndependCncia da GuinC-Bissau (African Party for the Independence of Guinea- Bissau) PDDAA Programme de'taille pour le de'veloppernent de 1 'agriculture Africaine PEASA Programa de Ernergencia do Apoio ri Seguranca Alimentar PER Public Expenditure Review PNDS Programme National de DCveloppement de la SantC PNDS National Health Development Plan PRGF Poverty Reduction and Growth Facility PRS Partido de Renovaqgo Social (Social Renovation Party) PRSP Poverty Reduction Strategy Paper PSR Public Sector Reform QS Qualitative Survey SMP Staff Monitored Program SSR Security Sector Reform TEC Tarifa Externa Comum (Common External Tariff) TIPS Trade and Investment Promotion Support TOT1 Terms of Trade TR Transition Rate UF User Fees UNCDF United Nation Capital Development Fund UNDP United Nation Development Fund UNOGBIS United Nation Office in Guinea-Bissau USAID United States Agency for International Development's USD United States Dollars VAR Vector Auto Regressive WAEMU West African Economic and Monetary Union

Vice President: Gobind T. Nankani Country Director: Madani M. Tall Sector Manager: Robert R. Blake Task Team Leaders: Boubacar-SidBarry Estanislao Gacitua-Mario

... 111

Table of Contents

ACKNOWLEDGEMENTS ...... VI11 EXECUTIVE SUMMARY ...... X VOLUME 1: TRANSITIONFROM POST CONFLICT TO LONG-TERMDEVELOPMENT: POLICY CONSIDERATIONS ...... x VOLUME 11: CONFLICT, LIVELIHOODSAND POVERTY IN GUINEA-BISSAU XXIV 1. INTRODUCTION TO VOLUME I...... 1 A. BACKGROUNDAND LINKWITH THE PRSP ...... B. DATAAVAILABILITY AND EXISTING STUDIES , ...... C. THEANALYSIS ...... 2. INSTITUTIONAL CONTEXT OF THE POST-CONFLICT PERIOD ...... 6 A. A LONGHISTORY OF POLITICAL INSTABILITY...... B. A HISTORYOF WEAK GOVERNANCEAND ACC ...... c. FACTORSBEHIND THE WEAK GOVERNANCEAND INSTITUTIONAL DEVELOPMENT...... D. SECURITY SECTOR, LANDREGULATION AND CONFLICT ...... E. CONCLUSION AND POLICY CONSIDERATIONS ...... 23 3. POVERTY ANALYSIS ...... 24

...... 43 D. CONFLICTAND POVERTY......

4. MACROECONOMIC AND FISCAL POLICIES FOR PROMOTING SOCIOPOLITICAL STABILITY AND ECONOMIC GROWTH...... 54 A. RECENTMACROECONOMIC PERFORMANCE ...... 55 B. SHORT TERM FISCALPOLICY FORSOCIOPOLITICAL STABILITY c. MEDIUMAND LONGTERM POLICY FOR MACROECONOMICSTABILIZATION AND GROWTH... c. FACTORSTHAT MAY CONSTRAIN IMPLEMENTATION OF D. CONCLUSION AND POLICY CONSIDERATIONS ...... 70 5. MAKING AGRICULTURE MORE PRODUCTIVE AND SUSTAINABLE FOR POVERTY REDUCTRION ...... 71 A. MAINOPPORTUNITIES AND CONSTRAINTS IN AGRICULTURE ...... B. TRANSACTIONCOSTS AND POVERTY REDUCTIONIN THE CASHEW NUTSECTOR. C. SUSTAINING POVERTYREDUCTION AND GROWTHIN AGRICULTURE...... 84 D. CONCLUSION AND POLICY CONSIDERATIONS ...... 85 6. MAKING BASIC EDUCATION AND HEALTHCARE MORE AFFORDABLE FOR THE POOR ...... 86 A. EFA AND EDUCATIONSERVICE ...... 86 B. UF AND HEALTHSERVICE ...... 103 C. A COMMUNITY-BASEDSERVICE DELIVERYFOR REDUCING POVERTY ...... 1 1 1 D. CONCLUSION AND POLICY CONSIDERATIONS ...... 113 REFERENCES ...... 114 ANNEXES ...... 122 ANNEX1. VAR MODEL...... 122 ANNEX2. PUBLIC SECTOR DOWNSIZINGAND SEPARATION PACKAGES ...... 129

iv List of Tables

Table 1.1: Population Distribution (percent) by Age and Sex in 2005 ...... 2 Table 2.1: Selected ICGR Indicators of Cohesion for Guinea.Bissau. 2005 ...... 12 Table 2.2: Institutions That Are Trusted the Most by the Population in 2005 ...... 13 Table 3.1: Poverty Decomposition by Region in 2002 (in percentage) ...... 25 Table 3.2: Subjective Perceptions of Well-Being in 2005 ...... 28 Table 3.3: Main Source of Inequality in 2005 ...... 30 Table 3.4: Poverty by Age in 2002 ...... 31 Table 3.5: Poverty by Demographic Characteristics in 2002 (in percentage) ...... 31 Table 3.6: Educational Attainment and Poverty in 2002 (in percentage) ...... 32 Table 3.7: Poverty and Sector of Employment in 2002 (in percentage) ...... 32 Table 3.8: Poverty by Distance from Basic Services - Bissau Region in2002 (in percentage) ...... 33 Table 3.9: Poverty by Distance from Basic Services - Rest of the Country in 2002 (in percentage) ... 33 Table 3.10: Access to Energy for Cooking and Lighting in 2002 (in percentage) ...... 34 Table 3.11: Poverty Decomposition by Recipients of Prenatal Care in 2002 (in percentage) ...... 34 Table 3.12: Poverty by Type of Healthcare Provider Used (in percentage) ...... 34 Table 3.13: Revenue Shares by Economic Status in 2002 (in percentage) ...... 36 Table 3.14: Revenue Shares by Region in 2002(in percentage) ...... 36 Table 3.15: Impact of Various Characteristics on Household Welfare in 2002 ...... 42 Table 3.16: Impact of the Conflict on Poverty ...... 51 Table 3.17: Number of Years Required to Achieve Cumulative Growth and Poverty Targets ...... 52 Table 3.18: Millennium Development Goals and draft PRSP Targets ...... 52 Table 3.19: Perceptions on the Security Situations and Sources of Conflict in 2005 ...... 53 Table 4.1: Comparative Military Data for Selected African Countries in 2003 ...... 63 Table 4.2: Impact of a Reduction of Wage spending on Growth in Guinea-Bissau, 2005-08 ...... 68 Table 4.3: Efficiency in Tax Collection and Medium-Term Fiscal Sustainability, 2005-08 ...... 68 Table 5.1: Cashew Sales by Tercile in 2005 ...... 80 Table 5.2: Impact of change in cashew prices on poverty, Guinea-Bissau ...... 82 Table 5.3: Stochastic Relationship Between Households’ Non-agricultural and Agricultural Incomes ...... 82 Table 6.1: Educational Status of Primary School Students by gender and economic status in 2002 (in percentage) ...... 90 Table 6.2: Educational Status of Primary School Students by Region in 2002 (in percentage) ...... 90 Table 6.3: Adult illiteracy Rates ( percent) for population above 15 years of age) ...... 91 Table 6.4: Reasons for Dissatisfaction by Region in 2002 (in percentage) ...... 92 Table 6.5: Reasons for Dissatisfaction by Type of School in 2002 (in percentage) ...... 92 Table 6.6: Public Education Spending by Level, 2002-04 ...... 94 Table 6.7: Health Providers by Location and Economic Status (in percentage) ...... 105 Table 6.8: Health Providers by Region (in percentage) ...... 105 Table 6.9: Selected Health Indicators in 2002 (in percentage) ...... 106 Table 6.10: Malaria Patients who Sought Professional Treatment in 2002 (in percentage) ...... 106 Table 6.11: Health Indicators in 2002 (in percentage) ...... 107 Table 6.12: Persons who consulted a Health Provider and were not satisfied by the Reason of Dissatisfaction in 2002 (in percentage) ...... 107 Table 6.13: Persons Sick who did not Consult a Health Provider and the Reason Why ( percent) .. 108 Table 6.14: UF Collected by Region in 2004 ...... 109 Table 6.15: Utilization of UF from Medical Consultations in 2004 (in percentage of total revenue collected) ...... 109 Table 6.16: Utilization of UF from Drug Sales in 2004 (in percentage of total revenue collected) .... 109

V Annex Tables

Table A1 1: Impact on GDPK of a shock to the cyclical component of GEK and TOTI (formulation) ...... 123 Table A1 2: Impact on GEK of a shock to the cyclical component of GEK and TOTI (formulation) ...... 124 Table A1 3: VAR Impulse Response Function for TOT (effect of one standard deviation innovation in TOTI) ...... 125 Table A1 4: VAR Impulse-Response Function for GEK (effect of one standard deviation innovation in GEK) ...... 125 Table A1 5: Normalized Impulse-Response Coefficients for GDP Growth...... 126 Table A1 6: Normalized Impulse-Response Coefficients for Changes in Government Spending ..... 126 Table A1 7: Impact on GDP of a shock to the cyclical component of GEK and TOTI (Simulation Results) ...... 127 Table A1 8: Impact on GEK of a shock to the cyclical component of GEK and TOTI (Simulation Results) ...... 128 Table A2 1: Average compensation per Separated employee (thousands of CFAF) 1/ ...... 131

List of Boxes:

Box 2.1: Land Regulation in Guinea-Bissau ...... 21 Box 2.2: Status of the Cadastre ...... 22 Box 3.1: 2002 ILAP Survey - Design, Data Coverage and Construction of Poverty Line...... 26 Box 3.2: 2005 QS - Methodology and Data Coverage and Analysis ...... 27 Box 3.3: Perceptions of Food Security: Meeting the Food Needs of the Household ...... 38 Box 4.1: Summary of the Main Findings of the 2001/02 Diagnostic Studies of the Public Sector ...... 61 Box 4.2: Objectives of a SSR in Guinea-Bissau ...... 63 Box 5.1: Improving the Business Environment for the Development of the Cashew Chain ...... 83 Box 5.2: TIPS Initiative...... 85 Box 6.1: Structure of the Education System in Guinea-Bissau ...... 88 Box 6.2: MainFactors Influencing Dropout Rates in Primary Schools ...... 91 Box 6.3: Designing a Social Fund ...... 112

List of Figures

Figure 2.1: Selected Governance and Accountability Indicators. 2004 ...... 11 Figure 2.2: Governance Indicators in 2002 ...... 14 Figure 3.1: Perceptions on the Evolution of the Economic Situation of Households by Quintile in 2002 ...... 36 Figure 3.2: Problem Satisfying the Food Needs of Households by Quintile in 2002 ...... 37 Figure 3.3: Impact of the 1998 Conflict on Per Capita GDP ...... 50 Figure 4.1: Budget and External Accounts Balances 1996-04 (in percentage of GDP) ...... 55 Figure 4.2: Real GDP and Cashew Production Volumes, 1999-03 (change in percentage) ...... 56 Figure 4.3: External Transfers, 1999-04 ...... 57 Figure 4.4: Changes in Exports and Cashew Nut prices, 2000-04 (in annual percentage) ...... 58 Figure 4.5: Revenue and Recurrent expenditure, 1999-04 ...... 59 Figure 4.6: Evolution of the Payroll and the Wage Bill, 1999-2004 ...... 60 Figure 5.1 Trends in Cashew Prices, 1979-03 ...... 75 Figure 5.2: Cashew Yield and Poverty...... 77 Figure 6.1: Literacy Status of the Head ofthe Household in 2002 ...... 87 Figure 6.2: Access to Primary Education by Region in 2002 ...... 87 Figure 6.3: Main Reasons for not Attending School in 2002 ...... 90 Figure 6.4: Access, Usage and Satisfaction with Education in 2002 ...... 90 Figure 6.5a: Unit Cost per Student in EBE, 2002-2004 ...... 95

vi Figure 6.5b: Unit Cost Per Teacher in EBE. 2002-2004 ...... 95 Figure 6.6: Enrollment Rates in Basic Education, 10997-2003 (in percentage) ...... 95 Figure 6.7: Gross Intake Rates (GIR), 2004-2004 in EBE ...... 96 Figure 6.8: Transition Rate in Primary& Secondary Education 2002-04 ...... 100 Figure 6.9: Incidence and Coverage of Public Spending on Education in 2002 ...... 102

Annexes:

VAR Model ...... 122 Public Sector Downsizing and Separation Packages ...... 129

vii ACKNOWLEDGEMENTS

This Integrated Poverty and Social Assessment (IPSA) is organized into two volumes. The current volume, the main report, provides the core poverty and social analysis. The second volume offers a series of selected issues derived from the core themes and analyzes the impact of conflict and vulnerability on growth and poverty. Volume two was prepared with financial support from the ESSF Trust Fund.

IPSA is based on the findings of three missions to Guinea-Bissau headed by Boubacar-Sid Barry and Estanislao Gacitua-Mario. The first one was a preparatory mission which took place in December 2004. During this mission, preliminary contacts were established with the authorities to explain the objectives of the work to be carried out, and the identification of international and local consultants to carry out the work. A local IPSA team was established at the National Institute ofResearch and Census (INEP) which also provided local consultancy services for the realization of the Quality Survey. The coordinator of the local team was Mr. Mamad6 Jao (Director General, INEP). The second mission that took place in February 2005 worked in close collaboration with the local team and the team of international consultants to train the local consultants in the use of the survey questionnaire, and to launch the QS. The third mission that took place in April 2005 worked with a team ofthe National Institute of Statistics (INEC) to process the 2002 household survey data. The INEC team was composed of Messrs. Adulai Jalo (Economist), and Bessa Vitor da Silva (Demographer). The mission also worked with the INEP team to analyze the QS data. The INEP team included MessrsMmes. Mamadh Jao (Director General) and Fernando Leonard0 Cardoso (Researcher), and Catarina Gomes Viegas (Researcher) and Deolinda Mendes Silva, Bucar Indjai and Iaia Biai (local consultants). Security sector-related issues were discussed with Mr. Pedro Correia, Coordinator of the Demobilization Reinsertion and Reintegration Program (DRRP) of Guinea-Bissau. During the mission, the team discussed preliminary results with the authorities in a workshop that was attended by officials in ----the Minister of Finance, and chaired by the Minister ofEconomy.

The other members of IPSA’S core team are Quentin T. Wodon, Edward Creppy, Hakon Nordang, Sigrun Aasland and Anne Marie Bodo. Other Bank staff that- participated in the missions were Christian Fauliau, Miki Takahashi, Carmen Pereira and Gilbert Pouho. Contributions and useful suggestions and comments were also received from Michael Drabble, Laura Rose, Harry Snoek, Maurizia TOVO,Abdoulaye Seck, Ndiame Diop, Clarence Tsimpo, Ilda Lourenqo-Lindell, and Humberto Lopez. Faye Harbottle and Glaucia Reis Ferreira provided logistic and language assistance during document preparation. The peer reviewers are Marcel0 R. Andrade, Louise J. Cord, and Anis A. Dani. The report was prepared under the supervision ofRobert R. Blake (Sector Manager, AFTP4), who offered conceptual guidance, provided critical analytical advice- and ensured quality control and management support. Madani M. Tall (Country Director AFC14) and Iradj A. Alikhani supported the process and provided usehl comments.

Messrs. Yuta Harago and Takayoshi Fukuyo, international consultants from the Japan Brazil Network (JBN) were contracted to collaborate with INEP in the implementation of the QS. Senior officials from the Ministries of Finances, Economy,

... Vlll Education, Health, Public Administration and Labor, the BCEAO, the INEC, the INEP, the UNDP, UNICEF and other representatives of the main donors based in the country cooperated with the missions and provided relevant inputs for the analysis. The Bank missions benefited from the exchange ofinformation with IMF and African Development Bank missions that were also visiting the country at he same period. The report benefited from generous support from a TFESSD grant managed by Louise Fox and Quentin Wodon in the Africa PREM front office.

In July 2005, Mr. Yuta Harago suddenly passed away in an unfortunate accident. We want to acknowledge his work and express our sincere regrets.

ix EXECUTIVE SUMMARY

IPSA is organized in two volumes. The current volume, the main report, provides the core poverty and social analysis. The second volume offers a series of selected issues derived from the core themes and aims at analyzing the impact of conflict and vulnerability on growth and poverty.

VOLUME 1: TRANSITIONFROM POST CONFLICT TO LONG-TERMDEVELOPMENT: POLICY CONSIDERATIONS

1. Introduction to Volume I

Following the 1998 military conflict, instability has become a major constraint for. good governance, poverty reduction and growth in Guinea-Bissau. According to the 2002 light household survey (ILAP), about two-thirds of the country’s population lived in poverty, with three-fourths of the poor residing in rural areas and practicing agriculture. In this context, ensuring sociopolitical stability in the short-term, and promoting fiscal adjustment and growth in the medium-term should represent a crucial element of the authority’s Poverty Reduction Strategy Paper (PRSP). As a core element of the Bank’s analytical work, the Integrated Poverty and Social Assessment (IPSA) is intended to provide support to the reform process currently taking place in the country.

2. Institutional Context of the Post-Conflict Period

A. A Long History of Political Instability. Following a decade-long struggle with , Guinea-Bissau became independent in 1974. After two decades of a centrally-planned economic system marked by recurrent episodes of military-related instability, the country held its first multi-party legislative and presidential elections in 1994. An army uprising in 1998 triggered a bloody civil war that displaced over 300,000- members ofthe civilian population. Failed governance, breakdown of the rule oflaw, and limited accountability and transparency of public sector management were among the causes ofthe conflict that lasted nearly a year.

Following the conflict, attention was turned to the task of reconciliation and rehabilitation. A Government of National Union (GNU) prepared a National Reconciliation and Rehabilitation Program (NRRP), and organized free and fair parliamentary elections in 1999. However, the implementation of the NRRP was undermined by the lack of political consensus, as a new constitution approved by the Parliament in April 2001 was not endorsed by the President. The resulting ambiguity undermined the rule of law, and a prolonged friction developed between the executive and the legislative branches of the Government. Following several military-related incidents during 2001-02, and delays in holding parliamentary elections as initially scheduled, the military led a bloodless coup in September 2003. A civilian transition Government was immediately appointed, and charged with the task of rebuilding the administration and preparing legislative and presidential elections.

X The transition Government prepared the Emergency Economic Management Plan (EEMP) in end-2003, and successhlly organized free and fair parliamentary elections in early-2004. A new Government was appointed which endorsed the EEMP, prepared a draft PRSP and a budget for 2004. Budgetary management has improved during the second half of 2004, and free, fair and transparent presidential elections were organized in June-July, 2005.

B. A History of Weak Governance and Accountability. In Guinea-Bissau, boundaries between the public and private sectors are often ill-defined. This contributes to erode the public’s confidence in the government, and thus citizens’ willingness to pay taxes or support the Government’s reform policies. The financial accountability system remains relatively weak, and mass disclosure of information on public accounts remains limited. Also, Guinea-Bissau’s rule of law and government effectiveness indicators, and the overall risk rating of the International Country Risk Guide (ICRG) are among the lowest in Sub-Saharan Africa. The public’s perceptions of different institutions indicate that local schools and health posts are the most trustworthy institutions. The judiciary, the police, religious organizations and traditional authorities benefit from much lower levels of trust. Non Governmental Organization (NGOs) and other Civil Society Organizations (CSOs)/Community Based Organizations (CBOs) often supplement the often weak presence ofthe Government in rural areas.

C. Factors behind Weak Governance and Institutional Development. The current weak governance situation dates back to the colonial era. The colonial State promoted limited presence of the Government in rural areas. Traditional authorities, therefore, played a leading institutional role in local communities during the colonial rule. The post-independent State largely inherited the weak institutional legacy ofthe colonial era. Traditional authorities, particularly in rural areas, continued to wield power in the communities, either by dominating the State-sanctioned village committees or by sidelining them. In the end, the village committee system gradually became defunct, leaving the State with little presence at the village level. At the same time social and political leaders have established strong paternalistic relationships with their constituencies without building broader coalitions or strengthening the legitimacy of the existing state institutions. In recent years, some progress has been made in social service delivery, notably in education and health, thanks to the active role played by civil society organizations and the support ofthe international community.

D. Security Sector, Land Regulation and Conflict. The armed and security forces reportedly lack adequate dwelling places and sanitation services. Moreover, the. army recruits are predominantly from one ethnic group, which represents a major risk of ethnic tension. These important issues would need to be urgently addressed through a comprehensive security sector reform (SSR) in the broad context of a PSR program. Finally, the absence of a formal regulatory framework for the attribution and use of land represents another major source of inequality and civil conflict particularly at the local level, which will need to be urgently addressed through the official adoption of the new Land Law in place since 1998.

xi E. Conclusion and Policy Considerations. At this juncture of Guinea- Bissau transition from post-conflict to long-term development, it is of outmost importance to enhance political stability in order to mitigate risks of conflict as experienced in the past. It is therefore critical to address security and land issues, the two main factors ofconflict in the country.

In the short-term, improving living conditions in military barracks will be critical for preserving political stability. In that vein, key immediate actions under the SSR would include the rehabilitation of military dormitories, and sanitation and food services in military caserns. In the medium-term, the SSR would need to address issues related to downsizing the armed forces, reforming the pension system, and balancing the ethnic composition ofthe security sector.

With regard to land reform, the Government would need to officially adopt and implement the new Land Law in order to provide a legal and regulatory framework for the resolution of conflict stemming from land use. Finally, updating the cadastre will be critical for the implementation ofthe land reform.

3. Poverty Analysis

A. Poverty Profile. The incidence ofpoverty, as measured by the headcount poverty index, remains relatively high. At the household level, 58.9 percent of the population lived in poverty in 2002, with 16.8 percent leaving in extreme poverty. The poverty incidence is higher for larger households, and for men compared to women. This' phenomenon is partly explained by the different family structures (monogamous vs. polygamous) and their distribution across urban and rural areas, as well as by the fact that women generally engage in various income generating activities, such as the processing of cashew nuts, to supplement the income ofthe household. Based on the findings ofthe QS completed in 2005, poverty is believed to have increased since the last household survey. Nearly 77 percent of households believe that their livelihood situation worsened in 2004 compared to the period before the conflict. Similarly, 54 percent believe that their situation has worsened in 2004 compared to 2003. Simulations based on these findings and initial conditions in 2002 indicate that the incidence of poverty is likely to have increased to 62.4 percent in 2004.

B. Determinants of Poverty. The poverty rate among household heads within the age cohort of 35-55 years is higher (58 percent) than the rate among households headed by people who are less than 35 years old (48 percent). As expected, the highest incidence ofpoverty occurs in households headed by persons aged 55 or more (71 percent), who are likely to have a higher level of family responsibilities due to the relatively large household size. Surprisingly, the poverty headcount shows that the incidence of poverty is higher among men (61 percent) than women (51 percent) as pointed out in the poverty profiling. The incidence of poverty among household heads with tertiary education is 29 percent, compared to 65 percent for heads with no level of education. Not surprisingly, poverty incidence is higher among heads working in agriculture (56 percent) compared to Government and private sector workers.

xii 1. Socioeconomic characteristics of poverty. With regard to social services and poverty, the poverty incidence among those who live less than 15 minutes from' potable water, primary school and health facilities, in Bissau is lower than those who live more than 15 minutes from these services. In the rest of the country, however, the only services with a similar pattern of distribution are secondary schools and health facilities. About 68 percent of the households in Bissau do not have access to energy and electricity. In the rest of the country, the proportion of households without electricity is much higher at 95 percent (65 percent of households who use firewood for cooking are below the poverty line whereas only 12 percent of those who use electricity or gas for cooking are poor). In terms of access to healthcare, the poverty incidence is very high' (about 83 percent) among women who have had a live birth in the last 12 months without prenatal care compared to those who had it (63 percent). Also, the poverty rate among those who use neighborhood midwiveshurses as the main healthcare provider is highest at 88 percent, whereas only 39 percent of individuals who used private clinics lived below the poverty line.

With respect to income and expenditure and poverty, about 40 percent of household spending are allocated to the purchase of food, while salaries and the sale of agricultural products are the main sources of income. The overall economic situation of households and their communities at large seem to have deteriorated compared to the year prior to the survey year. This situation seems to have aggravated the ability of household heads to adequately feed their families. However, women headship of the household is believed to have contributed in improving food security.

2. Determinants of welfare. In rural areas, for each additional infant and child the expected level of consumption of a household is likely to decrease respectively' by 8 and 10 percentage points. In urban areas, where the cost of living is likely to be higher, the decrease in the expected level of consumption is higher especially for infants and children. In female headed households in rural areas, consumption per capita and the accumulation ofassets is likely to be higher than households headed by men.

The marital status of household heads seems to have a positive impact on consumption per capita and on the accumulation of assets. In rural areas, the expected level of consumption, for household heads who are single, married (monogamous) and' those who are married (polygamous) is respectively 19 percent, 15 percent and 16 percent higher than the consumption levels of households headed by those who are divorced or widowed. In urban areas, however, only households headed by those who are single are likely to increase consumption per capita. In fact, the magnitude of the effect is pretty strong at 35 percent, compared to heads that are divorced or widowed. In rural areas, consumption increases by 12 percent for heads with secondary education and by 35 percent for heads with tertiary education compared to heads with no level ofeducation. In urban areas, the effect is even stronger at 24 percent for secondary school educated heads' and 53 percent for heads with tertiary education.

It is also evident from the results that the impact ofeducation on the accumulation of assets is high both in urban areas as well as in rural areas. In urban areas, for example, asset holdings of households increase by 8 percent, 21 percent and 48 percent,

xiii respectively, for heads with primary, secondary and tertiary education, whereas in rural areas, the increases in assets are 4 percent, 16 percent and 43 percent, respectively. The impact of education on income is limited to urban areas where the income level of household heads with tertiary education is 70 percent higher than the income of heads with no level ofeducation.

For household heads that live in rural areas and are employed in sectors like Transport/Commerce and Educatioflealth, income per capita is likely to increase by 47 percent compared to heads working in the agricultural sector. This finding may confirm the existence oflow productivity in agriculture compared to other sectors. In urban areas, however, the only sector which seems to have a positive impact on income is Transport/Commerce, which increases income by 46 percent. Moreover, rural household heads who are employed in Transport/Commerce and in the fields of Educatioflealth are likely to increase the asset holdings oftheir households by 24 percent and 15 percent, respectively. It is worth noting that in these areas, working in the Industry and Construction sectors can also enhance the accumulation of assets. Surprisingly, there is likely to be a 17 percent reduction in consumption per capita for households whose heads work in the healtldEducation sector in rural areas. This phenomenon may be partly explained by the frequent salary arrears and payment delays, which may adversely affect consumption patterns. Urban household heads, who work in the private sector, are likely to increase income per capita by 42 percent compared to their counterparts who work in government. This finding not only illustrates the potential of the private sector for reducing poverty, but also confirms the likely presence of low productivity in the public sector.

With regard to migration, rural households whose heads have migrated (at least for a month in the last 12 month have experienced a small increase in the level of asset accumulation compared to households whose heads have never migrated. In terms of consumption per capita, the impact is observed in both rural and urban areas. In urban areas, for example, temporary migration is likely to increase the level ofconsumption per capita by 16 percent higher than in similarly located households without a head who has ever migrated.

C. Livelihoods of the Poor. In urban areas, the informal sector plays a. significant role in livelihood strategies. Informalization of livelihoods, while characteristic across Guinea-Bissau, is a particular feature ofthe livelihoods ofthe urban poor. More than 90 percent of the urban families recently surveyed in urban areas, lives essentially from activities in the informal sector. In this setting, women have historically played an important role in small trading, domestic work, dressmaking, weaving, extraction of cashew juice/wine etc. Informality is also a common feature of the labor market. During periods ofharvests, some people, particularly younger women and youth, seasonally migrate towards the rural areas of the country to offer their labor to farmers,. particularly cashew producers. Salary earning activities represent men’s main livelihood means. For urban men, casual work is increasingly the main activity and livelihood strategy. This represents a serious cause of vulnerability as formal wage opportunities become limited.

xiv In rural areas, subsistence agriculture represents the main livelihood strategy in rural areas. Rice remains the most widely grown staple crop in the country. Cashew (including production and sale of cashew wine) has played a particularly important role for many households, providing an important source of income. Livelihood strategies of the rural poor tend to differ across regions and are also influenced by the social characteristics of households and individual household members (such as ethnicity, family structure, gender, religion, age, etc.). Agriculture and cattle are the main livelihood means in the Northern and Eastern regions and part of the South (rice), while fishing is common in the South and the islands.

D. Conflict and Poverty. Evidence suggests that Guinea-Bissau’s economy would have performed better if there had been no conflict. The estimation results indicate that the country’s Gross Domestic Product (GDP) per capita would have been more than 40 percent higher in 2004 than it actually was if there had been no conflict. As a point of comparison, in Rwanda where the human, social and economic costs of the Genocide have been staggering, the estimates suggest that per capita GDP today would be between 25 and 30 percent higher if the Genocide had not taken place, which suggests a smaller economic impact than in Guinea-Bissau. Similarly, the results indicate that the poverty incidence would have been lower if the conflict had not occurred. Today, about two- thirds of the population is poor, i.e. with a level of consumption per equivalent adult’ below the US$ 2 poverty line in purchasing power parity terms, with 21.6 percent being extreme poor. If the conflict had not taken place -- so that per capita GDP today would be 42.3 percent higher than actually observed -- poverty would be much lower. The results suggest that the share of the population in poverty today could be around 43.0 percent, instead of 65.7 percent, while the share of the population in extreme poverty could be at 7.6 percent, instead of21.6 percent.

Under these circumstances, it will be difficult for Guinea-Bissau to achieve the’ Millennium Development Goal (MDG) for poverty reduction by 2015. If the real GDP were to grow on average at 5 percent per year in the long run, the poverty incidence will be reduced to 45.7 percent by 2015. However, rural poverty will remain relatively high at 50.6 percent. If the long run growth performance is less than 5 percent, the poverty rate is expected to remain above 50 percent, with rural poverty remaining around 60 percent. Based on these results, it is safe to argue that if the conflict had not occurred, the country would have made significant progress toward achieving the MDG target for poverty reduction by 2005. In order to address this challenge, the Government would need to- diversify the sources of growth and implement policies designed to make public spending more pro-poor, in the context ofthe PRSP Iorder to alleviate poverty.

With regard to the impact of security and land-related conflict on poverty, 50 percent ofhouseholds consider that the security situation has improved over the last year. However, the majority of households in the bottom tercile (the very poor) believe that their situation has deteriorated. Land remains a major source of conflict. Apart from internal family issues, most local conflict situations seem to arise fiom issues related to‘ land, water, and livestock, especially in rural areas. There is evidence that poor households tend to rely mostly on traditional authorities in order to solve conflicts, while the better offrely more on the police.

xv E. Conclusion and Policy Considerations. Guinea-Bissau is a country with very limited poverty data. With the data currently available it is almost impossible to analyze in detail the factors that cause poverty to change over time (growth versus distributional effects), and/or to carry out poverty impact analyses in key sectors such as cashew, rice, etc. As the implementation, monitoring and evaluation of the PRSP will start soon, it is important that such analyses are carried out in order to make the PRSP a" more result-oriented framework. In this regard, it is critical to develop a coherent and reliable poverty database over time. The scope ofthe next survey should be broadened to include: (i)detailed information about households' income and expenditure and price information of key commodities such as cashew nuts and rice production etc.; (ii) education and health service delivery; and (iii)conflict and land issues.

4. Macroeconomic and Fiscal Policies for Promoting Sociopolitical Stability and Economic Growth

A. Recent Macroeconomic Performance. Guinea-Bissau performed in a generally satisfactory manner during 1987-98. Following a decade of a ruinous State- controlled economic policy, the country embarked on a comprehensive reform program starting in 1987, and agreed upon a three-year Enhanced Structural Adjustment Facility (ESAF) arrangement with the Fund in 1995. Broad liberalization of the economy contributed to reduce large macroeconomic imbalances. Between 1996 and 1997, the public sector primary surplus increased from 3 to 5 percent, and the external current account deficit (excluding official transfers) improved from 18 to 14 percent. Inflation- decelerated from a record 50 percent in end-1997 to 8 percent in early 1998, owing to restrained public sector demand and regained stability from becoming a member of WAEMU. Domestic production, particularly in agriculture, accelerated with real GDP growth reaching more than 5 percent in average during 1987-97.

The disruption of the economy was significant during the conflict in 1998. Real GDP contracted by about 28 percent in 1998 due to a sharp decline of production activities in all sectors. The armed conflict adversely affected the country's fiscal and balance of payments positions. The external current account deficit (excluding official transfers) remained high at nearly 16 percent of GDP and the current primary budget balance, moved from a surplus of about 5 percent of GDP in 1997 to a deficit of about 7 percent of GDP in 1998. As a result, domestic arrears and external debt service arrears piled up to reach nearly 15 percent ofGDP.

After the war, real annual GDP grew by nearly 8 percent during 1999-2000. Simultaneously, Government revenue reached 18 percent of GDP, and the overall budget- deficit (excluding grants) increased from 14 to 25 percent of GDP between 1999 and 2000, as a result of increased post-conflict reconstruction activities. During the period through December 2000, progress continued to be encouraging on the macroeconomic front, and the country completed its interim PRSP, reached the decision point under the HIPC Initiative, and agreed upon a three-year PRGF program with the Fund for the period 2001-2003. Early in 2001, the macroeconomic program was found to be substantially off target owing to large unbudgeted expenditures mainly on defense, and financed by credit from the banking system and promissory notes. Consequently, the first

xvi review of the PRGF by the Fund was not concluded as scheduled in 2001, and the program was declared off track. After two unsuccessful short-term staff-monitored programs (SMP), efforts to revive the program were abandoned by mid-2002, and the arrangement expired at end-2003 without the completion of a review. Subsequently, economic performance deteriorated and real GDP decelerated at the rate of 7 percent by end-2002. In 2003, the economy remained sluggish with real GDP growing by only 0.6- percent, and fiscal management further deteriorated with a substantial diversion of resources to expenditures taking place outside the legally established budgetary procedures. In 2004, the Government embarked on the EEMP with the support of the international community. Since then, public finance management has started to improve and the economy has begun to recover with real GDP growing at more than 2 percent by end-2004.

B. Short Term Fiscal Policy for Sociopolitical Stability. In order to- preserve stability in the current polarized sociopolitical context, the authorities would need to take immediate measures aimed at improving living conditions in military barracks, ensuring the payment of public sector salaries, and delivering basic social services to the population. It is unlikely that the Government will be able to achieve these objectives by its own efforts. Education and health programs, for example, are highly dependant on external donor financing. The Government is, therefore, likely to be unable to deliver basic social services, as domestic resources are just sufficient to pay salaries.

C. Medium and Long-Term Fiscal Policy for Macroeconomic Stabilization and Growth. In the medium-term, the authorities would need to implement structural reforms aimed at restoring fiscal sustainability through a major fiscal consolidation effort. The latter should preferably be carried out in the context of a broad public sector reform (PSR) program aimed at bringing the non education and health public sector wage bill to a sustainable level, and reforming the security sector. The success of the PSR will depend on the capacity of the private sector to create new job opportunities as an alternative to public sector employment. In order to revive the private- sector, the authorities would need to accelerate the divestiture program, improve the business climate, and rehabilitate utilities, road, and port infrastructure and microfinance services.

In order to enhance public sector productivity and growth it is urgent to downsize the Government’s wage bill (in sectors other than education and health) at sustainable levels, in the context of mediurn-term civil service reform. For efficiency and fairness purposes, separation packages may be needed to overcome resistance to public sector downsizing.

In the case of Guinea-Bissau, evidence suggests that if real Government consumption (namely wage spending) were reduced by 2 percent (5 percent in nominal terms corresponding to 1 percentage point of GDP) annually over the period 2006-2008, real GDP would increase by about 1.6 and 2.9 percentage points in 2007 and 2008, respectively, compared to baseline growth targets for the same years. Although these results are consistent with empirical findings, they may be overestimated in the case of-

xvii Guinea-Bissau, given the weak institutional capacity of the State and other binding constraints to private sector growth.

Public sector downsizing will, however, not be sufficient for achieving medium- term fiscal sustainability. In order to achieve a current surplus in the medium-term, the Government will need to pursue sound expenditure management and improve the efficiency ofrevenue collection by reinforcing and sustaining the ongoing reforms in key departments of the Ministries of Finance and the Ministry of Fisheries. The proposed SSR should also address the size, the pension system, and the ethnic composition of the armed forces over the mediudlong-term. The sustainability of Guinea-Bissau’s external debt remains a major concern. Medium-term debt sustainability depends on continued fiscal adjustment and support from the international community.

Finally, persistent internal political tensions, weak institutional and governance capacity of the State and vulnerability of cashew nuts to external price fluctuations represent a major limitation to the implementation ofreforms.

D. Conclusion and Policy Considerations. It is imperative that the new administration ensures the continuity of institutions and commits to the reform program put in place since end-2003. In that vein, the authorities will need to adopt a comprehensible short and medium term reform strategy. In the short-term, the Government will need to pursue priority spending programs such as wage and salary payments and the delivery of social services to civilians and the military. In the medium- term, the authorities will need to: (i)accelerate ongoing structural reforms aimed at promoting the development of the private sector; (ii)carry out a broad fiscal consolidation reform with the objective of reducing the size of the wage bill, improving the efficiency of revenue collection and promoting fiscal sustainability; (iii)reforming the security sector; and (iv) normalizing relations with external creditors. Strengthening the government, political stability and institutional capacity, are critical in implementing these reforms.

5. Making Agriculture More Productive and Sustainable for Poverty Reduction

Agriculture has been for a long time the backbone of Guinea-Bissau’s economy. The sector represents more than half of the GDP, employs four-fifths of the labor force, and contributes to more than 90 percent of the country’s export earnings. Over the past three decades, the production of cashew nuts has taken a leading role in the country’s economy. However, high transaction costs associated with the poor business environment and weak physical and administrative infrastructure, contributes to limit income generation for smallholders in the cashew sector.

A. Main Opportunities and Constraints in Agriculture. Guinea-Bissau has a largely untapped potential for agriculture. The country is endowed with fertile soils, abundant water and favorable climate. It is the sixth largest producer of cashew nuts (6 percent of the world cashew production) after India, Vietnam, Brazil, Ivory Coast, and Tanzania. Guinea-Bissau’s weather is also favorable to the production of other cereals and a large variety of fruits. The country has large and valuable fish resources. The

xviii fishery sector provides about half of the Government’s budgetary revenues though fishing license payments. Livestock represents about 32 percent of agricultural GDP and is produced mainly in the Eastern part of the country. However, the country is a net importer offiesh meat, mainly poultry.

Growth in agriculture is constrained by the lack ofinvestment, unclear land tenure rights and inappropriate technical assistance support. As in the other economic sectors, public investment in agriculture has stalled in recent years due to the reallocation of public resources to defense and security programs. Similarly, private investment has suffered fiom the lack ofadequate micro financing mechanisms. Guinea-Bissau produces and exports raw cashews. This situation places the sector in a vulnerable condition vis-a- vis price fluctuations on the international markets. The absence of a coherent rice-cashew strategy represents a major threat to rice production and food security. The lack of a strategy fostering the articulation of commercial and export agriculture, with small farm subsistence and local market food security oriented agriculture, has threatened food security and hampered the development ofvalue added particularly to the cashew system. The farmers are naturally inclined to engage in commercial crop production of more guaranteed value. Even in that case, the lack ofroad infiastructure in rural areas limits the access ofrural production to local and regional markets, and often contributes to increase agricultural transaction costs.

B. Transaction Costs and Poverty Reduction in the Cashew Nuts Sector. The cost of doing business in the cashew sector remains relatively high. As mentioned in the first two chapters, instability and weak governance remain major disincentives for the much needed domestic and foreign investment in the sector. This situation is further complicated by changing regulations, unclear and cumbersome administrative requirements for existing and new businesses, and dysfunctional legal enforcement and inadequate utilities and seaport systems. Moreover, bureaucratic regulations, the absence of coordination between producers, and the lack of supervision on the Government side also lead to higher transaction costs, which translate into lower farm gate prices for farmers. Furthermore, the current export tax on cashew nuts (6 percent), relative to other exports (2 percent), is likely to have a disproportionate effect on rural incomes in the short-term.

Transaction costs represent an implicit tax on cashew (and other exportable) producers. Reducing these costs is likely to increase the farm gate price for cashew producers, and therefore improve their consumption of goods and services. In the case of Guinea-Bissau, evidence suggests that a 15 percent increase in farm-gate prices for cashew nuts could result in an increase in consumption for the extreme poor of 9.5 percent, and 3.3 percent for the poor. Moreover, an increase in the export tax on unprocessed cashew nuts of about 10 percent could potentially (with full pass through to farmers) generate a reduction in their farm-gate price of close to 15 percent (this assumes that farmers get 70 percent of the export price). The finding also suggests that there are significant second round poverty reduction effects of cashew on other non-agricultural activities. A 1 percentage point increase in households’ agricultural income is likely to increase non-agricultural incomes by 0.4 percentage point. The labor market is the main

xix transmission channel of these spillover effects because cashew farming is a labor intensive activity which employs mostly unskilled laborers.

C. Sustaining Poverty Reduction and Growth in Agriculture. The monoculture ofraw cashew poses a serious threat to long-term agricultural development. Although rice is the main staple food in the country, its farming has sharply declined in favor of cashew over the past three decades. As a result the country has relied heavily on rice imports to meet the domestic demand for cereal. The gap filled by rice imports represents about 25 percent of the trade balance. The rice deficit has also been filled by barter transactions between rice and raw cashew nuts starting in 1984. Bringing rice production to pre-independence levels would require the introduction of high-yield rice seeds, and the rehabilitation ofrural roads.

Guinea-Bissau produces mainly raw cashews, which provide less value added and remain highly vulnerable to exogenous climate and price hazards. The processing of cashew will contribute increased agricultural value added and create more jobs. Guinea- Bissau could potentially have a sizeable processing industry that would create thousands ofjobs and earn a great deal of value added that is currently lost by not processing raw cashew. However, as mentioned earlier, the development of the cashew processing industry depends on the ability to attract investment, of which most will have to come from outside the country.

For the rural sector to ensure food safety and create new jobs, policy makers would need to adopt a coherent agrarian development strategy in the context ofthe PRSP, which would aim at rehabilitating and encouraging rice production, the diversification of agricultural production and promoting the processing of raw cashew into exportable cashew kernels. The development of a variety of agricultural products and the diversification of agricultural exports, including fruits and vegetables, forestry, and small animals would be a priority. This is particularly important at this moment as the Government is readying to implement major civil service and SSR. These reforms would revive the private sector and contribute to the creation of new jobs to absorb part of the labor force to be discharged from the public payroll.

D. Conclusion and Policy Considerations. Despite Guinea-Bissau's large potential in agriculture, there are high transaction costs stemming from administrative' barriers, weak infrastructure and lack of diversification. For the rural sector to create labor intensive jobs, policy makers would need to adopt a coherent agrarian development strategy in the context ofthe PRSP, which would aim at promoting: (i)the processing of raw cashew into exportable cashew kernels; and (ii)the export of tropical fruits. This is particularly important at this moment as the Government is readying to implement major PSR. These reforms will contribute to reviving the private sector by creating the much needed new jobs to absorb part of the labor force to be discharged from the public- payroll.

xx 6. Making Basic Education and Healthcare More Affordable for the Poor

A. EFA and Basic Education Service. The rising concern with the quality of education, especially in Sub-Saharan Afr-ican countries was reflected in the protocols oE the World Conference on Education for All (EFA) in Jomptien (Thailand) in 1990 and at the World Education Forum in Dakar (April, 2000), where the international community observed that ten years after Jomptien, not much progress had been made, especially in the quality of education in Africa. In this context, Guinea-Bissau adopted the Dakar Framework for Action, EFA in 2000.

1. Situation of the Education Sector. Guinea-Bissau is characterized by low adult literacy rates with striking differences based on gender, the area of residency and. other socio-economic considerations. Enrollment in primary education is also influenced by factors such as location, gender, and economic status. Enrollment rates tend to be higher in Bissau - where households are closer to primary schools -- compared to other regions. Gender differences in enrollment rates for primary education are also quite noticeable. Enrollment for boys in primary school is estimated at 66 percent, which is on average 9 percentage points higher than that of girls. In Bissau, about 2 percent of children enrolled in primary school dropped out and about 44 percent never attended. The main reasons for the drop out are issues related to child labor and early marriage (at the national level, about 32 and 29 percent of children are not attending school for reasons associated with child labor and early marriage, respectively).

2. EFA and Basic Education Service Delivery and Utilization. The EFA initiative has contributed to increase public spending on basis education. During 2002-04, there was a consistent increase in per capita public spending (unit cost) as a ratio of GDP per capita for primary school students as well as for teachers in the same cycle. The unit cost on students in nominal terms increased from 5.6 percent of GDP per capita in 2002 to 6.4 in 2004. Similarly, teachers’ unit cost increased from 1.7 percent to 1.9 percent during the same period. The question, however, remains about the internal efficiency of EFA or its ability to retain pupils in the school system and ensure an orderly transition in the secondary education system. In fact, many children drop out of school before they have achieved the necessary skills to function as literate individuals.

The primary and secondary levels of education appear to be the largest recipients ofpublic spending, while allocations to tertiary (excluding spending on student studying abroad) and pre-school education are considerably smaller. This finding seems to indicate that while resource allocation in the primary and secondary education is not a concern, at least at the macro level, there is a need to improve the efficiency of spending in the various programs within each level. There is also a need to create a bridging or transition mechanism between the primary and secondary levels. For example, providing income subsidies to families with school-age children and better training and higher salaries for teachers may contribute to limit dropout rates, improve the quality on education, and ensure an orderly transition from primary to the secondary levels ofeducation.

3. EFA and Equitability of public resource Allocation to Schooling. The analysis of the 2002 survey data shows that basic education has the best targeting

xxi coverage compared to the other levels of education: about 60 percent of primary school allocations going to the poor, it is the only level of education that is best-targeted to alleviate poverty.

B. User Fees (UF) and Health Service. Guinea-Bissau prepared the PNDS, which covered the periods 1997-02 and 2003-08, and adhered to the Bamako Initiative’ (BI) in the early 1990s. In 2002, the Government adopted a law for the implementation of UF in public sector health centers and hospitals.

1. Situation of the Health Sector Prior to the Adoption of UF. The demand for health services remains high, particularly in rural areas. According to the 2002 survey data, the majority ofpatients who sought modern professional health care in Bissau and other regions visited private clinics and regional health centers, respectively. Contrary to expectation, traditional healing centers do not seem to be patronized by a significant number of people. ‘The poverty decomposition of those who use these healthcare providers shows that the highest percentage ofpoor people and the extreme poor use the services of non professional health providers, and traditional healing facilities and pharmacies. In the particular case of malaria, more than 50 percent of patients sought treatment from non-modem healthcare providers. The non frequentation ofmodem health services may partly be explained by the lack of income. However, that of the rich may signal other factors such as the lack ofmalaria drugs and other malaria-related health care supplies in modern public health care facilities.

Access to and satisfaction with health service is generally limited. In 2002, about 37 percent of the poorest households had access to health services compared to 46 percent for the wealthiest. Consequently, the rate ofusage ofmedical services was only 9 percent for the poor, as opposed to 19 percent for the rich. At the same time, a large proportion of households were dissatisfied with public health services. For example, more than 80 percent of the residents of Gabu had a bad experience with their medical facilities, and the issues most ofthem were concerned with were the cost oftreatment and- the availability of drugs.

2. Implementation of UF and Impact on Health Service Delivery and Utilization. In recent years, public sector health expenditures have declined in Guinea- Bissau to the point where nearly all Government funding went to salaries. In the early 1990~~the country adhered to the BI. However, many patients were informally paying for health services ‘under the table’. Recognizing the difficulty to cover the recurrent cost of the health system, the authorities recently passed a law introducing UF. The law on UF- was a means of legalizing this practice and providing more transparency and efficiency in the collection and use of UF. The new system was considered a viable solution because Guinea-Bissau has already successfblly implemented the BI (fees for drugs). It is important to understand the dynamic underway in Guinea-Bissau, because UF have always discriminated against the poor.

UF are being collected and managed at the facility level in a fairly transparent manner. Since 2002, Guinea-Bissau’s new cost-recovery system includes the recovery ,of- drug fees (also covered under the BI), and charges for outgoing visits, lab tests, and other

xxii medical visits. So far, the resources are being collected and managed in a relatively transparent manner. Revenues at the level ofthe facilities are managed by a management committee composed of the facility’s staff and representatives of the civil society. The UF system is being carried out in all the 127 health facilities at the national level. In 2004, the amounts collected per region were within the range of CFA 6 to 33 million (about US$ 12,000 to 66,000). The total cost recovered in 2004 represented about 5 percent of the health sector budget. On average 30 to 60 percent of the resources collected were primarily used to pay for drugs and other maintenance costs. The remainder of the resources served for the payment of incentives to the personnel (20-30 percent) and the constitution of reserves (10-40 percent). In limited cases, 10 percent of the resources were transferred to the Regional Health Department.

Although the initiative is recent, overall, it has contributed to improve the availability of drugs and the access to health services at affordable costs. Improved birth attendance and immunization coverage have contributed to significantly decrease child mortality. and in recent years to improved prenatal health care for women. However, as turn-out rates are gradually increasing, health facilities would need to be adequately equipped in order to meet the surplus demand. It is therefore critical that the authorities accelerate the implementation of the ongoing reform program in the context of the ‘Programme National de DCveloppement de la SantC (PNDS)’, train additional medical staff and build housing facilities for medical personnel in rural areas.

C. Conclusion and Policy Considerations. While the EFA initiative has contributed to increase enrollments in primary education, there remain high risks about the retention ofpupils in school. Moreover, there are uncertainties about the sustainability ofthe EFA because major external supports are ending, while the Government is still not ready to auto finance the program. With the ongoing support of the Bank and other donors, the health sector would have the much needed boost, if the reform agenda agreed upon in the context ofthe PNDS is carried out.

xxiii VOLUME 11: CONFLICT, LIVELIHOODSAND POVERTY IN GUINEA-BISSAU

1. Conflict, Livelihoods and Poverty in Guinea-Bissau: A Brief Overview

This second volume of IPSA seeks to analyze the impact of conflict and vulnerability on growth and poverty in Guinea-Bissau. As stressed throughout the first volume, conflict and political instability have considerably weakened the country’s productive tissue. This situation increased the degree of vulnerability of the population - mainly in rural areas where most of economic activities take place - and hence impeded economic growth. Consequently, per capita incomes deteriorated significantly, thereby increasing poverty. This volume provides a background analysis of a number ofthemes derived from the first volume. The remaining five chapters are organized as follows.

2. Conflict, Growth and Poverty in Guinea-Bissau

Conflicts have had a devastating impact on the economies of affected countries in Sub- Saharan Africa (World Bank, 2000a). In Guinea-Bissau, conflicts and political instability have been the main constraints for growth and poverty reduction over the past three decades. This first chapter measures the impact of the conflict on per capita GDP in Guinea-Bissau. This is done using a technique for the identification and correction of outliers in time series that was developed by Tsay (1986, 1988) and Gomez, Maravall and Peiia (1997), and applied to the assessment ofthe impact of conflicts on GDP per capita in Sub-Saharan African countries including Guinea-Bissau by Lopez and Wodon (2005a, 2005b). The estimates suggest that GDP per capita would be roughly 42-43 percent higher today if the conflict had not occurred. In turn, assuming that the conflict has had no impact on inequality (we don’t have any data on these potential impacts), we find that about one third of the population now in poverty can be said to be poor because of the conflict itself.

3. Institutions, Social Networks and Conflicts in Guinea-Bissau: Results from a 2005 Survey

Guinea-Bissau has suffered from a major conflict in 1998 and from persistent political instability since independence. This second chapter provides results from a recent survey that gives insights into the opinions of the population among others about changes in well-being over time, trust in various institutions, sources of conflicts at the local level, and ways to deal with conflicts. There is a clear perception among the population that there has been a deterioration in well-being as a result ofthe conflict, as well as a lack of improvement since then at least for the poor who are highly vulnerable. The data suggest an increase in the lack of security after the conflict, or again at least no clear sign of. improvement. The population has little trust in national institutions such as the army, the police, the judicial system, and the central government. Local conflicts often emerge due to the competition for scarce productive resources, but poorer households deal with these conflicts in a different way than wealthier households.

xxiv 4. Poverty and its Determinants in Guinea-Bissau

Given the poor economic performance of Guinea-Bissau over the last few years, including a severe recession towards the end of 2002, poverty is likely to be high and to have risen in recent years even versus its already high post conflict level. This third chapter estimates the share ofthe population in poverty in 2002, predicts how it may have evolved since then, and assesses the levels of growth that will be required in order to reduce poverty measures in the future. It also provides a brief poverty profile and an analysis of the determinants of poverty using the 2002 nationally representative survey which was recently made available for analysis. Geographic location, demographic structure (both household size and headship), employment (both in terms of sector and type), education, and migration all have potentially large effects on the consumption level and thereby poverty ofhouseholds.

5. Livelihoods in Guinea-Bissau

Standards of living have historically been low, and the conflict that took place in 1998 had a very large negative impact on GDP per capita and thereby on poverty, as discussed in the first chapter. This fourth provides an analysis of the modes of livelihood of Guinea-Bissau’s population. The country is poor and highly dependent on agriculture and extractive activities such as fishing. Yet beyond such general statements, it is useful to describe in qualitative terms the main activities that take place both in rural areas at the regional level, and in urban areas through mostly informal employment. This is done on the basis of field work carried in the country in early 2005. In addition, the paper provides a brief quantitative empirical analysis of the main income sources of households, and a more detailed analysis of two key sectors of the economy that could form the basis of a pro-poor growth strategy: the cashew nuts sector, and the fishing sector.

6. Cashew Production, Taxation, and Poverty in Guinea-Bissau

Agriculture is the engine of Guinea-Bissau’s economy. The sector relies mainly on cashew nuts, rice and subsistence production of food crops. Cashews represent 90 percent ofthe country’s exports and the principal source of income in rural areas. Unfortunately,” cumbersome administrative arrangements, weak legal systems and an absence of credit often lead to high transaction costs for cashew buyers and exporters, which contribute to decrease the farm gate price of the raw nuts. This fifth chapter provides a review of the cashew sector in Guinea-Bissau, as well as estimates of the likely impact of changes in farm-gate prices and export taxes on poverty among cashew producers and in the country as a whole. The chapter also notes that over the last three decades, the production ofrice has significantly decreased in favor of cashew farming. This situation represents a threat to food security. For the rural sector to ensure food security and create new jobs, policy- makers would need to adopt a coherent agrarian development strategy in the context of the PRSP, which would aim at rehabilitating and encouraging rice production, and also promoting the processing of raw cashew into exportable cashew kernels, in order to generate more value added in the cashew sector.

xxv 1. INTRODUCTION TO VOLUME I

A. BACKGROUNDAND LINK WITH THE PRSP

1.1 Demographic Context. Guinea-Bissau is about the size of Maryland, located on the Atlantic coast of West Africa. Current estimates’ suggest that the total population is about 1.5 million inhabitants, of which 66 percent is rural. Population dynamics and trends (migration and growth) are difficult to describe given the lack of actualized data and the conflict that the country experienced in the late 1990~.~Population growth rate has remained almost constant at 2.4 percent from 1970 to 2000, and estimates for the period 2000-2004 suggest a growth rate of 2.0 percent (U.S. Census Bureau 2005). The population is unevenly distributed across regions. Bissau concentrates about 25 percent of the population, followed by Oio (14.6 percent), Bafata (13.8 percent), Cacheu (13.8 percent), +Quinara(4 percent) and Boloma-Bijagos (2.3 percent) (UN 2001). Over the last 15 years there has been a strong growth of the urban population (5.1 percent per year).3 From 1975 (at independence) to 2002, the urban population grew from 16 percent to 44 percent (UN 2001). Rural-urban migration has speeded up over the last decade, concentrating the population on four main areas Bissau-Biombo, Cacheu, Bafata- Bamdadinca and Buba. Slum dwellers as a percentage of urban population represent about 93 percent (SDStats, 2001).4

1.2 Guinea-Bissau has more than 20 different ethnic groups, with no group having an absolute majority. The five largest ethnic groups are: Balanta (28 percent), Fula (23 percent), Mandinga (13 percent), Manjaco (11 percent), and the Pepel (7 percent). The other 15+ ethnic groups represent about 18 percent of the population. The urbdrural distribution ofthe groups suggests that Balanta and Pepel are more concentrated in rural. areas, whereas the Manjaco, Fula and Mandinga tend to reside in urban areas. Most of the population is Muslim (40 percent) - mainly drawn from the Eastern groups such as the Fula and Mandinga - or Animist (37 percent), followed by Catholics (17 percent), and other religions (6 percent).

1.3 Over 60 percent ofthe population is less than 24 years old (see Table 1.1) and 41 percent is under 15 years old.5 The estimated population under 15 years old in 2002 was 47.1 percent, which is the 6th highest in the world after Uganda, Niger, Mali, Burkina. Faso and Yemen. Recent data from a micro household survey in Guinea Bissau6 suggest

’ The last population census was conducted in 1991. Current estimates are based on census data and electoral cadastre. Figures by U.S. Census Bureau (2004) Some 300,000 to 350,000 people were displaced as a result of the 1998 conflict. Most of them have returned to their homes. According to the International Organization for Migration (IOM) the demobilization, reinsertion and reintegration program is in its last phase. To date, among the 7,182 beneficiaries, 2,406 have already been reintegrated and 2,031 are expected to be reintegrated by June 2005. (UN 31/03/2005 Security Council 5157” Meeting (PM) Press Release SC/8350). This growth has been fueled by the lack of employment opportunities in rural areas and, in the late 1990s, due to the displacement. caused by the war. The rapid urban growth has dramatically increased urban density and pressure over an almost inexistent urban infrastructure and public services (UN 2001 and UNDP 2001). US Census Bureau, 2005 (htto://www.census.eov/cgi-biniiuc/idbsum.ol?ctv-PU). The survey was carried out between March and May 2005. The sample size was 446 households. The sample was selected following a purposive sampling strategy covering all regions of the country (North, East, South, Islands and Bissau). In each region provinces were selected. In each province were selected three localities (1 urban and 2 rural) on the basis of ethnic, ruravurban and similar trends. The dependency ratio is 0.91, above the average for Sub-Saharan (0.88) and Low Income countries (0.71).

Table 1.1: Population Distribution (percent) by Age and Sex in 2005

Datafrom US Census Bureau Datafrom INEP/JBN/WB Age Cohort Total Male Female Total Male Female 0-14 41.5 42.7 40.5 39.6 41.5 37.0 15-24 20.2 20.6 19.6 22.5 22.6 22.3 25-59 33.5 32.7 34.4 32.5 29.9 35.3 60+ 4.8 4.1 5.5 5.4 5.4 5.4 Source: US Census Bureau and the 2005 QS 1.4 Total fertility rate (births per women) estimated at 7.1 (UNDP 2003), is the third highest in the world (after Niger and Angola), well above the average for Sub-Saharan Africa (5.4) and least developed countries (5.1). Women start their reproductive life very early. On average women enter into a family union at age 16 and men at 20. One out of five women 15 to 19 has a child (UNFPA 2003).7 Prevalence ofcontraceptive use among women 15 to 49 is low, reaching only 4 percent (UNFPA 2003).

1.5 Economic Context and Link with the PRSP. The country is one of the poorest- in the world, ranking 172nd out of 177 nations on the United Nations Human Development Index of 2004. Guinea-Bissau’s state of development is characterized by low per capita income (US$140 in 2003)8 and high illiteracy rates (above 60 percent). According to the 2002 light household survey (ILAP),about two-thirds of the country’s population lived in poverty, with three-fourths of the poor residing in rural areas and practicing agriculture. Following the 1998 military conflict, instability has become a major constraint for good governance, growth and poverty reduction in Guinea-Bissau. After a slight recovery of 1999-00, the country’s macroeconomic and fiscal situation sharply deteriorated throughout 2001-03 as a result of lingering political instability and ineffective public policies. In this context, restoring sociopolitical stability, pursuing reforms and adjusting the macro economy are critical for enhancing poverty reduction and growth. The country completed its interim Poverty Reduction Strategy Paper (PSRP) in 2000, and subsequently launched the preparation of the full PRSP which has been repeatedly delayed. The Government prepared a draft PRSP in 2004, which is expected to be finalized soon. As a core element of the Bank’s analytical work, IPSA is intended to support the Government’s reform agenda in the context ofthe authorities’ PRSP.

agro-ecological criterion. In Bissau (SAB), 2 neighborhoods (Hhfia and Belem) were selected. In tum, in each of the 16 communities selected, a random sample of approximately 30 households was selected. ’According to UNFPA (2004), the rate of teenage pregnancies in Guinea Bissau ranks 8” in the world, after Niger, Congo DRC, Angola, Liberia, Somalia, Sierra Leone and Uganda. The average rate for least developed countries is 124/1000, Western Africa 119/1000; all Africa 107/1000. Understandably, fertility is high because of the early age of first pregnancies and very low use of contraception.

WDI, World Bank (2004).

-2- B. DATAAVAILABILITY AND EXISTING STUDIES

1.6 IPSA relies heavily on the data ofthe 2002 Light Household Survey’ (ILAP), and the 2005 QS. The latter was conducted during March-May, 2005 as part of the preparation ofIPSA. It collected qualitative information on households’ livelihood means and perception ofpoverty and the quality ofbasic education and health service delivery. Besides survey data, the report also relies on macroeconomic information from the fiscal and national accounts data produced by the authorities and the Fund. The macro data are final for past years till 2004, preliminary for 2005, and projectedsimulated for 2006-08. Other data have been provided by the Ministries of Agriculture, Education and Health. IPSA takes stock of existing analyses such as the Civil Society Assessment Tool (CSAT) prepared by the Bank in 2004, and the Government study” on the cashew sector carried out with the support of the Bank and the United Nation Food and Agriculture Organization (FAO) in 2004.

1.7 IPSA’S poverty analysis has been significantly constrained by household data limitations. For example, the 2002 ILAP survey, which is the core data of the analysis, does not provide detailed information on households’ income and expenditure in key agricultural activities such as cashew and rice production. Moreover, the light design of the survey does not allow comparability with the 1991 Full Fledged Household Survey (ICOF) data. As a result, attempts to perform poverty dynamic analysis have been abandoned. This fact underscores the importance of broadening the scope of the next household survey and aligning the methodology with that of the 1991 survey to allow performing dynamic poverty analyses.

C. THEANALYSIS

1.8 Rationale. Guinea-Bissau is a Low Income Country Under Stress (LICUS) with weak institutional and governance capacity, and widespread poverty. Over the past years, persistent political instability contributed to weaken the State’s capacity to promote economic growth and deliver social services to the population. In this context, there is an urgent need to: (i)reinstate the basic hnctions ofthe State in the short-term (payment of public sector salaries, delivery of education and health services, and rehabilitation of minimum water and electricity services); and (ii)carry out a major fiscal adjustment reform and promote the diversification of agriculture in the mediudlong-term. While pursuing these objectives, the Government will need to demonstrate strong commitment to reform so as to mobilize the much needed external support. Finally, given the trust that exists in CSOsKBOs, the Government would gain in involving these organizations in the- delivery ofsocial services.

1.9 Objective and Content. IPSA is the first poverty assessment for Guinea-Bissau since the 1998 conflict. The main objective of the report is to build knowledge on the interconnections between socio-economic and institutional factors, instability, poverty and growth. The analysis seeks to inform macro level policy dialogue and to improve the

The 2002 ILAP survey was sponsored by the Dutch Government and the Bank. 10 Jeager and Lynn (2004). effectiveness of development interventions at both policy and program level by focusing on policies aimed at removing systemic differences in opportunities between social groups. It also seeks to mitigate risks and constraints to development arising from such practices and enhancing macroeconomic and fiscal stability and social service delivery, while enhancing sociopolitical stability. The remainder of this first volume is structured in five main chapters dealing with the following issues: (i)institutional context of the post-conflict period; (ii)poverty analysis; (iii)macroeconomic and fiscal policies for' promoting sociopolitical stability and growth; (iv) making agriculture more productive and sustainable for poverty reduction; and (v) making education and health care more affordable for the poor. The rationale behind the core issues and how they fit in the structure ofthe report is briefly outlined below.

1.10 Political Economy and Institutional Context of the Post-Conflict Period (chapter ' 2). This chapter takes stock of the political economy of the country that may help understand its current state of institutional fragility and political and macroeconomic- instability. It analyzes the interconnections between the country's history of conflict and instability and its sociopolitical development after the 1998 civil conflict. Particular attention is paid to security and land issues, which represent potential risks ofconflict and instability.

1.11 Poverty Analysis (chapter 3). This chapter presents an in-depth analysis of the socioeconomic profile and determinants ofpoverty using the 2002 ILAP and the 2005 QS data. The findings suggest that the incidence of poverty remains relatively high and- concentrated in rural areas. About two-thirds of the population is living below the national poverty line. Moreover, wage earners (in the public and the private sectors), literate people and those with easy access to infrastructure, utilities and basis social services bear a lower poverty incidence compared to others. The finding also confirmed that vulnerability is relatively high in rural areas, and the country lost significant growth opportunities as a result ofthe conflict.

1.12 Macroeconomic and Fiscal Policies for Promoting Sociopolitical Stability and Economic Growth (chapter 4). This chapter reviews recent economic performance, and explores policy options for preserving sociopolitical stability in the short-term and promoting fiscal adjustment and macroeconomic stabilization and growth in the mediudlong-term. As policy reform in a fragile institutional context is by definition politically difficult and requires a piecemeal approach, the report stresses the need to: (i) pursue essential expenditure programs in order to preserve social and political stability in the short-term (paying salaries, delivering social services, improving living conditions in military barracks, rehabilitating basic utility services); and (ii)implement a major fiscal. consolidation (reducing the wage bill in non-productive segments of the public administration in the context of a PSR program) and structural reforms in the security sector (reducing the size, balancing the ethnic composition of the army) in the medium- term.

1.13 There are indications that the reduction of wage expenditures is likely to spur growth. However, this effect may be overstated given the country's weak institutional capacity. The findings also suggest that fiscal consolidation relying solely on a reduction'

-4- of the wage bill (compatible with the economy’s absorption capacity) will not be sufficient to restore medium-term fiscal sustainability. The report therefore recommends that additional efforts be made to improve the efficiency of revenue collection, while pursuing sound expenditure control.

1.14 Making Agriculture More Productive and Sustainable for Poverty Reduction (chapter 5). This chapter reviews the main challenges and opportunities in agriculture, and assesses the impact of lower transactions costs on income generation and poverty reduction in the cashew sector. A particular attention is paid to the need to diversify agricultural output through the processing of cashew nuts and the promotion of tropical fruit production and export.

1.15 Making Education and Health Care More Affordable for the Poor (chapter 6). This chapter assesses the effect ofthe Government’s free basic EFA and health UF (UF) programs on social service delivery in the public sector. In the area of basic education, the report shows that while the EFA initiative has contributed to improve public spending and enrollment rates, there are uncertainties about the sustainability ofthe initiative. With regard to the UF system, the report reviews the cost recovery system and assesses the effect ofUF on health outputs since its inception in 2002. Overall, the UF system is being implemented and managed fairly well. Resources recovered are being used to renew drug inventories, and pay salaries and incentives to the personnel ofthe health facilities.

1.16 Finally, in order to improve education and health services -- in the current weak State capacity -- the authorities should consider establishing social development funds to support these initiatives. Given the relative strength of the civil society in delivering social services, the management of these funds could be entrusted to CSOs and NGOs with the supervision ofthe authorities.

-5- 2. INSTITUTIONAL CONTEXT OF THE POST-CONFLICT PERIOD

Since the independence of Guinea-Bissau in the early 1970s, conflict and political instability have represented the main obstacle to good governance, poverty reduction and growth. The armed forces have been ojien at the center of this process, and still remain a major factor of instability as evidenced by recent coups and mutinies. Despite democratic elections in 1999/2000, 2004 and 2005, the orderly transition to constitutional rules remains elusive due to frequent interventions of the military in political afairs. The military allegedly lack adequate dwelling place and sanitation conditions. Moreover, the predominance of one ethnic group in the army represents a major risk of ethnic conflict. Preventing further military-related instability would require addressing these issues in the context of a comprehensive SSR. Beside military-related conflicts, the absence of a formal regulatory framework for the attribution and use of land represents another major source of civil conflicts.

2.1 Conflicts and political instability have been the main constraints for sustainable institutional development in Guinea-Bissau over the past three decades. Rather than being massive and pervasive, conflicts and instability have become cyclical events confined mainly in the capital. While often not causing significant physical destruction, the continued instability, however, has been detrimental to institution building and has not produced an environment conducive to investment, growth and poverty reduction. As many as six major military-related incidents concerning political power have taken place since independence in 1974. Thereafter, the country experienced two decades of a State- controlled system before organizing its first democratic elections in 1994. The fledgling democratic process was disturbed by the 1998 civil conflict and the state of political instability that prevailed after the war. In late 1999, early 2000, the second set of democratic elections brought to power a civilian Government. However, the democratic process stalled as the elected Government failed to restore political stability and good governance. Following several cabinet reshuffles, alleged coup attempts, and the dissolution of the Parliament during 2001-2002, the military staged a bloodless coup d’ktat in September 2003, and a civilian transition Government was charged with the task of restoring order and rebuilding the public administration. Since mid-2004, the authorities made significant progress in restoring fiscal discipline despite the highly polarized and still fragile political environment. Guinea-Bissau remains also vulnerable to social risks of instability related to land use and ownership as the country lacks an adequate legal and regulatory framework for land. The remainder of the chapter will successively: (i)review the institutional context of the post-conflict period; (ii)analyze governance issues; (iii)discuss the main factors behind the current governance problems; and (iv) discuss the issue of conflict prevention through a reform of the security sector and land regulations.

-6- A. A LONGHISTORY OF POLITICAL INSTABILITY

2.2 Guinea-Bissau has experienced a long history of conflict and political. Following a decade-long freedom war with Portugal, Guinea-Bissau became independent in 1974. The government of the first President, Luis Cabral, was later overthrown in a bloodless coup led by Prime Minister and former Armed Forces Commander Joao Bemardo (Nino) Vieira in 1980. There were several coup attempts against the Vieira government in 1983, 1985, and 1993. In 1994, the country's first multi-party legislative and presidential elections were held. President Vieira of the ruling Party for the Independence of Guinea-Bissau and Cape-Verde (PAIGC) was re-elected with a slight margin over Kumba Yala of the Party of Social Restoration (PRS). An army uprising against the Vieira Government in June 1998 triggered a bloody civil war that displaced over 300,000 members of the civilian population. Failed governance, breakdown of the rule of law, and limited accountability and transparency of public sector management were among the causes ofthe conflict that lasted nearly a year.

2.3 Following the conflict, attention was turned to the task of reconciliation and rehabilitation. By June 1999 the situation, although still fragile, calmed down and improved steadily. A Government of National Union (GNU) was put in place and charged with the task of preparing the NRRP and organizing elections. Free and fair parliamentary elections were held on November 28, 1999, and a presidential runoff vote took place on January 16, 2000. Opposition parties scored a landslide victory in both elections. Elected President Koumba Yala of the PRS took office on February 17, 2000. Shortly after, a Cabinet of broad coalition including members of various parties represented in the National Assembly was appointed.

2.4 The task ofnational reconciliation and reconstruction was undermined by the lack of political consensus. Following the elections, the Parliament approved a new constitution in April 2001, which was neither vetoed nor promulgated by the President. The resulting ambiguity undermined the rule of law, and a prolonged friction developed between the executive and the legislative branches ofthe Government over the choice of the new cabinet chief. As a result, the country practically operated without an effective government for several weeks, and the announcement of yet another foiled coup attempt. in mid-April 2001 demonstrated once again the fragility of the political and democratic processes. The prevailing political instability, which affected all institutions, was no more conducive to economic recovery and post-conflict reconstruction. In December 2001, the Government allegedly thwarted a coup attempt by army officers.

2.5 Delays in holding parliamentary elections led to a military coup and the appointment of a transition Government in September 2003. Between December 2001 and December 2002, President Yala carried out two cabinet reshuffles, dissolved the, Parliament, and called for legislative elections, which were repeatedly postponed. The ensuing discontent among the population led the army to seize power in a bloodless manner in September 2003. President Yala announced his "voluntary" resignation and was placed under house arrest. The military, led by the Chief of Defense General

-7- Verrisimo Seabra,” quickly restored power to civilian control, and a civilian caretaker government was appointed with the objective ofrestoring the rule of law and reasserting control over public finances. In this context, the businessman Henrique Rosa was sworn in as Interim President, with the support of almost all political parties and the civil society. By end-2003, the newly appointed Government prepared the EEMP and a budget for 2004 with the support ofthe international community.

2.6 Parliamentary elections were successfully organized in early-2004 and the post-election Government endorsed the EEMP reform agenda. Legislative elections were held in March 2004 and an orderly transition to a Government elected on the basis of a broad reform platform took place in May 2004. The transition government, led by the leader of the PAIGC Carlos Gomes Junior, was appointed with the objective of ensuring the functioning ofthe public administration and preparing presidential elections by mid-2005. Despite the difficult economic situation, the Government prepared a draft ofthe country’s PRSP and significantly improved budgetary discipline during the second half of 2004. The Government also endorsed the reform program agreed with the international community under the EEMP, prepared the 2005 budget and successfilly organized free, fair and transparent presidential elections won by the independent candidate and former President Joao Bemardo (Nino) Vieira.12

2.7 The recent presidential elections that took place in June and July 2005 brought to an end the transitional arrangements that had been put in pace since the ousting of former President, Kumba Yala, in September 2003. They were also meant to put an end to the post-conflict transition that began after the civil war of 1998-9 and bring a return to constitutional and democratic rule in Guinea-Bissau following years of political instability and administrative chaos. While the elections themselves were free, fair and transparent, the presidential campaigns and the immediate period following the elections were marred by fear and uncertainty amongst the country’s population. President Vieira, however, aware that the legacy ofhis rule from 1980 to 1999 had made his reelection this year controversial and divisive, has so far tried to counter these fears by continuing his electoral campaign platform ofreconciliation and national unity, promising to uphold the law, modernize the economy, defend freedom of expression and reform the my.He has also confirmed that he is willing to cooperate with the current government and Prime Minister, despite the ruling party’s previous attempt to have his victory annulled. Vieira also seems to have gained the support of the military, with Armed Forces Chief of Staff, Baptista Tagme Na Wai, and Navy Chief of Staff, Bubo Na Tchuto and Aniceto Na Fla, all pledging their support to the country’s democratically elected civilian leaders.

2.8 Nevertheless, the return to electoral democratic rule still leaves many issues unresolved which will continue to plague the political situation in Guinea-Bissau. Ultimately, many of the causes underlying the political instability and insecurity which has defined Guinea-Bissau during the last 7-8 years still remain.

” General Seabre was killed along with other officers during the October 6,2004 mutiny. ’* The first round poll opposed seven candidates including three former Presidents: Malam Bacai Sanha of the PAIGC, independent candidate Joao Bemardo (Nino) Vieira, and Koumba Yala of the PRS. The second round poll opposed Sanha and Vieira. The latter was declared winner in the second round of votes with 52 percent of the votes, with a strong backing from Yala’s supporters. Eight team of international observers, including an 80-strong mission from the European Union (EU), issued a joint statement saying the first and second round polls had been “free, fair and transparent”.

-8- 2.9 An immediate cause for concern relates to the current constellation of, and competition for, power between the executive and the legislative. Although an immediate dispute between Nino and the majority ruling-party PAIGC over the electoral outcome has been overcome, tensions between the President and the Prime Minister, Carlos Gomes J6nior will remain [as a political analyst recently remarked, “with Gomes seeking to limit Nino’s powers and prerogatives, and Nino seekin to undermine the premier’s position within the PAIGC - the party Nino once This situation has been further complicated during the electoral campaign by an increasingly visible split within the PAIGC, between on the one hand, its leadership, many of whom refuse to recognize Vieira’s victory and insisting the official PAIGC candidate, Malam Bacai Sanha, was the rightful winner, and, on the other hand, other party members - mainly drawn from the “old guard” who are still committed to their old party leader.

2.10 The political situation remains highly volatile in the aftermath of the presidential elections. In end-October 2005, President Vieira dismissed the PAIGC Government of Prime Minister Carlos Gomes Junior. In early-November, the President swore in a five-party coalition government led by PAIGC dissident Aristides Gomes. The PAIGC continues to claim the legitimacy of the new cabinet while several parliamentarians are believed to have defected from its rank. A session of the National Assembly to be held in December 2005 is expected to review the Government’s program, and determine whether it has the support ofParliament.

2.1 1 Despite the President’s pledges to build national unity, this fierce competition over power reflects a lack of unity between the different political groups in the country. Political power in Guinea-Bissau is concentrated in an executive subject to limited checks and balances from weak democratic institutions. The state also retains main control over access to the majority of resources in the country. As a result, the stakes are high for the presidency and key political positions, undermining efforts at building national unity in a country with more than 20 different ethnic groups, rife poverty and great disparities between people. The lack of mechanisms for social and political accountability and representation extending beyond the capital of Bissau to the regions where the majority ofthe population resides further exacerbates this problem.

2.12 Recurrent military interventions on the political scene, as experienced in recent past, still remain a threat to stability and democracy. So far, the military have kept their promise not to intervene in the political process during and after the presidential elections. However, the widespread poverty in the barracks, an inflated, unpaid myoften sharing popular discontent, and the highly polarized political environment remain ingredients for future interventions of the military. Moreover, Guinea-Bissau’s army remains highly ethnicized as the majority of soldiers are from the BaZante ethnic group. In order to prevent further instability and conflict, there is a critical need for reforming the security sector as discussed below in section C ofthis chapter.

2.13 Despite the recent elections, there is much to be done before Guinea-Bissau will complete its post-conflict transition. Perhaps most importantly, the state continues to lack

l3hid.

-9- legitimacy and credibility amongst most members of the population. This is due to a variety of factors, particularly its continued inability to provide basic services and infrastructure, particularly outside of the capital of Bissau, and the perceived persistence of corruption and mismanagement in the public sector. This suggests that the volatile nature of the process of political transition in Guinea-Bissau is at its roots merely a reflection of a much deeper problem of governance, which will be discussed in greater detail below in section B.

B. A HISTORYOF WEAK GOVERNANCEAND ACCOUNTABILITY

2.14 The relative lack of transparency poses a major risk of corruption in public affairs. In Guinea-Bissau, boundaries between the public and private sectors are often ill- defined. There is clearly an overlap between political, economic and civic spheres of power. For example, many of those engaged in national politics, including some high- ranked civil servants, are also from the private sector and together control large segments ofthe economy. While this may be merely a reflection ofthe size of the country, and the scarcity of capacity, it is important to understand how this overlap of personas and responsibilities relates to ethnic differences, the question of the autonomy of the State, and the division ofthe public and the private spheres. While this is not necessarily in and of itself a problem, it underscores the need to establish a clearer separation ofpowers, as well as improved systems oftransparency and accountability.

2.15 The lack of confidence in the Government may contribute to decrease the public’s incentive to pay taxes and support the Government’s policies. If the popular confidence in the government continues to erode, citizens’ willingness to pay taxes or support the Government’s reform policies will falter as well. Public perception of different institutions confirms that local level institutions are more trustworthy than central or local governments, the military or political parties. Moreover, taxpayers will be less willing to “contribute” if there is the presumption that the Government will not use the resources collected for due purposes. The alleged Government response to this relative absence outside of Bissau (and even in poorer areas within Bissau), has been to ’ suggest that rather than being a provider of services, the State should instead become a promoter and a coordinator ofservices de1i~ery.l~This approach risks being unresponsive to the needs ofthe poor and if implemented as a strategy could provide an excuse as well as a justification for the relative absence ofthe State.

2.16 Despite recent progress, the financial accountability system remains relatively weak. Following the 2004 parliamentary elections, accountability mechanisms have somewhat improved as evidenced by the timely preparation ofthe 2004 and 2005 budgets by the executive power in end-2003 and end-2004, respectively, and their approval by the Parliament. Moreover, the current executive power is not interfering in legislative and judiciary matters, as it was the case during Yala presidency. However, major areas of accountability and auditing systems require hrther improvement. The financial accountability system of Guinea-Bissau is composed of the General Inspection of Finance (IGF), which is a financial control unit in charge of the internal audit; and the

14 Speech by Minister of Social Solidarity, CSAT Launch Workshop, December 8,2004.

- 10- Court ofAccounts (CC), which is an independent body in charge ofthe ex-ante control of public contracts and ex-post control ofpublic accounts. The IGF is not hlly autonomous and does not have the means to ensure that its recommendations are implemented. The external audit system remains weak due to the lack of qualified judges and the weak technical capacity of the staff. The post-transition authorities would need to pursue the timely preparation of annual budgets, and ensure adequate and timely auditing of public accounts.

2.17 Despite improved access to media, mass disclosure of information on public accounts remains limited. Access of civil society to general public information has somewhat improved with the increase of programs broadcasted in the local dialect (Creole) in Bissau and the regions. There are more than two dozen radio stations and several newspapers throughout the country. Except for public radio and television, all the other media are relatively independent from the political power. Despite this proliferation ofmass media, information on public accounts (resources allocated to line ministries and reports on budget execution) are still not widely made available to the public. The continued political instability and the limited institutional capacity of the public administration are among the factor behind this problem. In 2001, the Government succeeded in producing the biannually budget execution report to the Parliament. It is critical that such effort be resumed and the information made available to the public.

2.18 Recurrent instability has contributed to Figure 2.1: Selected Governance and Accountability Indicators, 2004 undermine the rule of law and Government Control of effectiveness. While Corruption in some indicators of Rule of Law governance and Regulatory accountability Quality Guinea Bissau Government performs close to the Effectlveness average for Sub- Political Stability

Saharan countries, Voice and rule of law and Accountability I government 0 10 20 30 40 50 effectiveness are Dl Sub Saharan Afrlca Liberia Niger Guinea Bissau among the lowest in the region (see Figure 2.1). Most state institutions have been incapable of generating or commanding long term loyalties. In part this is explained by their lack of capacity to deliver services and benefits. In addition, social and political leaders have established strong paternalistic relationships with their constituencies without reaching out or bridging to other social groups, and thereby without building broader coalitions or strengthening the legitimacy of the existing state institutions. In the absence of strong state institutions, the population has strengthened their local networks and has learned how to count on their own forces to search for solutions to their problems. Local

- 11 - institutions (from neighborhood associations to the local schools and the traditional authorities) therefore perform many functions (mutual help, leisure, collective security, conflict resolution etc.), some ofwhich could be handled by the state. While at the local level communities work together to resolve some immediate problems, the capacity to work together beyond the local dimension and to resolve differences in Guinea Bissau is poor. Political risks and ethnic tensions are building up. Over the past 10 years there has been an increasing politization of ethnicity, which has been conveniently used by self- serving leaders seeking to increase their legitimacy and power. As shown in Table 2.1 below, overall risk rating (ICGR 2005) for Guinea Bissau is below the average for Sub-- Saharan countries.

Table 2.1: Selected ICGR Indicators of Cohesion for Guinea-Bissau, 2005 Average Government Internal Military Religious Ethnic Risk Stability Conflicts Politics Tensions Tensions Rating Guinea Bissau 5.5 7.0 1.5 5.0 3.0 52.5

Sub-Saharan Africa 8.9 8.5 2.5 4.1 3.2 56.2 I Somalia (Worst) 7.0 4.5 1.o 3.0 2.0 26.0 Botswana (Best) 11.0 10.5 6.0 5.0 4.5 78.5 Source: ICGR 2005

2.19 The public’s perceptions of different institutions confirm that local level institutions are considered more trustworthy than central or local governments, the military or political parties. The analysis of the 2005 survey’s data shows that local schools and health posts are perceived as the most trustworthy institutions (see Table 2.2). Trust in schools and health posts is explained by the services provided by these institutions and the level of participation that the community has in parent associations and health post teams. This trust appears to be particularly strong in rural areas, where the local school and the health post are the only institutions that are present and have some legitimacy for the community. The judiciary, the police, religious organizations and traditional authorities benefit from much lower levels of trust. It is worth noting that there are differences between terciles: among the poor, for example, trust in local institutions dealing with education and health as well as religious issues is even stronger relatively speaking (Le., as compared to trust in other institutions) than among the better off. Indeed, at least some ofthe better offtend to have higher levels oftrust than the very poor in NGOs, the justice system, and local government. In all three terciles, very few find that the police, the army, political parties and the central government are very trustworthy. There are also some differences between urban and rural areas as well: for example, trust in traditional authorities is higher among the rural population, while trust on NGOs is more prevalent among urban dwellers. It is interesting to highlight these differences because they are related to the different roles these institutions have. In urban settings trust in NGOs appears to be related to the role NGOs have in service delivery and- access to goods and services (such as concession ofcredit to the members, etc.) Trust in traditional authorities, while higher in rural areas, still is low. This is partly explained by the fact that traditional authorities are perceived as controlling access to information and mediating the relationship with outsiders. At the same time, in rural as well as urban settings, there are multiple local, community based organizations that inspire trust and

- 12- convoke people to participate and engage in activities that play an important role in the survival ofthe community and the families.

Table 2.2: Institutions That Are Trusted the Most by the Population in 2005 Tercile Area Total 1 2 3 Urban Rural lstmost trusted organization (%) Religious organizations 12.21 5.65 9.3 8.78 9.32 9.11 Political parties 1.53 0 1.55 1.35 0.85 1.04 Local government 0 1.61 5.43 4.73 0.85 2.34 NGOs 6.87 0 17.83 15.54 3.81 8.33 Traditional authorities 2.29 4.03 1.55 0 4.24 2.60 Schools 42.75 37.10 31.01 31.76 40.25 36.98 Health centers 29.01 43.55 26.36 3 1.08 33.9 32.81 Amy 0.76 0 0 0.68 0 0.26 Police 2.29 1.61 0.78 0.68 2.12 1.56 Justice 0.76 4.84 6.2 4.73 3.39 3.91 Central government 1.53 1.61 0 0.68 1.27 1.04 All 100 100 100 100 100 100 ZUd most trusted organization (%) Religious organizations 7.27 3.54 13.4 11.5 5.8 7.8 1 Political parties 1.82 0.88 4.12 1.77 2.42 2.19 Local government 1.82 2.65 4.12 4.42 1.93 2.81 NGOs 0.91 0.88 6.19 4.42 1.45 2.50 Traditional authorities 6.36 5.3 1 2.06 0.88 6.76 4.69 Schools 30.00 38.94 23.71 28.32 32.85 3 1.25 Health centers 45.45 35.40 36.08 36.28 40.58 39.06 Amy 1.82 0 0 0 0.97 0.63 Police 3.64 5.31 5.15 7.08 3.38 4.69 Justice 0 6.19 4.12 5.31 2.42 3.44 Central government 0.91 0.88 1.03 0 1.45 0.94 All 100 100 100 100 100 100 Source: Staff estimation from 2005 QS.

2.20 The weak presence of the Government has been largely compensated by the active role played by civil society organizations. According to the Civil Society Assessment Tool (CSAT) beneficiary assessment, 86 percent of the population turns not to the State but to civil society organizations, to respond to their basic needs. In rural areas, this number tends to be above 95 percent. Moreover, expectations of the government as a service provider are quasi nonexistent, even among members of the intellectual and organizational elite in Bissau.” Since the early 199Os, the number of NGOs operating. in Guinea-Bissau has increased significantly. Their various areas of intervention included developmental activities aiming mainly to complement the State in social service delivery. While citizens generally consider NGO engagement a positive

l5This should not be understood as the incapacity of the authorities to address these needs in a timely and sustainable manner. The underlying causes of such incapacity reside in the difficult economic situation, and the lack of a well defined poverty reduction strategy.

-13- contribution, their motives and willingness to involve communities is sometimes questioned. Moreover, NGO-sponsored activities may only be short-term, and are often unpredictable. l6NGOs, on their part, are heavily dependent on scarce donor resources,. and often lack incentives to specialize, as well as financial ability to commit to long term investments. While this demonstrates an expression of valuable social capital at the community level, the absence ofthe State is strongly felt.

Figure 2.2: Governance Indicators in 2002

GUINEA-BISSAU (200 I I

iovernnent Effectiveness

?egulatory Quality

-Ik Zule of Lau

Control of Corruption

1 I I e 25 5e 75 Conparison betueen 2804, 2882, 1998 (top-botton order) Country’s Percentile Rank (8-18B)

Sourcel 0. Kaufmnn, A. Kraav, and tl. Harlruiri 20061 Gowmance tlattw. IVI Gguwnanc= Indicators for 19964604 ~httri/~.~rldbank.e*q~i/qou.*nancc/pub~/sounrtt~rrt.htnlI

2.21 To sum-up, existing indicators suggest that there is weak sense of public institutions and nation state building in Guinea Bissau. The accountability of formal institutions is rather limited and cohesion is poor. Social capital functions at the local level to establish networks that contribute to achieve discrete goals ofvarious population groups, helping in the provision of services and the support of livelihoods. Within those networks accountability is higher and reflects the paternalistic nature ofthe relationships between local leadership and the community or groups for which these leaders act as protectors and hinges or intermediaries with larger social institutions. However, at the national level, public institutions are not effective and have apparently failed to gain public credibility. At the same time there is an increased fracture or disarticulation between national level institutions and leaderships and local communities and their authorities.

l6See MEP (2004), Civil Society Assessment Tool (CSAT) beneficiary assessment.

- 14- c. FACTORSBEHIND THE WEAK GOVERNANCEAND INSTITUTIONAL DEVELOPMENT

2.22 The current weak governance problems are deeply rooted in the colonial past. Guinea-Bissau has been plagued by a history of institutional weakness which preceded, and has to a large extent influenced, developments during the post-colonial period. The institutions and governance regimes put in place during colonial rule were primarily created for the purposes of economic extraction, with a minimum of emphasis and investments going into the development of a productive base and provision of social services and infra~tructure.’~Rural areas were left to be governed by traditional authorities, while a restricted concept ofcitizenship was introduced in the ‘modern’ urban sector. l8Unlike other African colonies, (the name of Guinea-Bissau prior to independence), however, was not as successful at subjugating or co-opting traditional authorities in the rural areas. This was arguably the result of a variety of factors including Portugal’s own relative economic and military constrains as a colonial power, as well as the ability of the rural civil society to partially disengage from the colonial State.lg State-society relations in Portuguese Guinea were, hence, at least in rural areas, characterized by a mixture of exclusion and neglect on the part of the State, and strategies of avoidance and withdrawal on the part of large elements of Guinea-Bissau’s society.

2.23 A culture of informality developed during the colonial era. The vibrancy of the informal economy during colonial rule showed the inability of the colonial State in regulating key sectors ofthe society.20 The informal economy during colonial rule, which is to a large extent similar to the current situation, dominated trading activities in the country and with neighboring countries. In response to reportedly exploitative and disadvantageous terms of trade in the regulated formal economy, commercial traders (cornerciantes and djilas) and producers reduced production, migrated and/or smuggled their goods instead of selling them through the colonial State.21As a result, the majority of the production and trade of agricultural products bypassed the colonial government regulation.22 Subsequently, a large part ofrural agriculture therefore remained outside the official economy, hence, leading to large tax evasions.23

2.24 Nowadays, traditional authorities continue to play a leading institutional role in local communities at the expense of the State. After independence, the one-party

See Pelissier (1989), Forrest (2003), Rudebeck (1988; 1990; 1997; 2001), Lourenco-Lindell(2002: 51)

Rudebeck (2001: 90-1). The development of the colonial State of Portuguese Guinea followed to some extent Mahmood Mamdani’s thesis of the ‘bifurcated’ State (Mamdani, 1996: 16-8), where the population in rural areas were treated as subjects rather than citizens, who were subject to customary rule, rather than populations in the urban areas, that were granted limited forms of citizenship and were controlled by the modem, colonial State. Mamdani argues that this bifurcated State continues to influence State development and citizenship in postcolonial Africa. A similar argument has been applied in the context of Portuguese Guinea by Cardoso (1992) and Lourenpo-Lindell (2002: 55-59), using the distinction between ‘non-civilized’ and ‘civilized’, or ‘non- assimilated’ and ‘assimilated’ citizens, echoing the distinction between ‘subjects’ and ‘citizens’. Rudebeck (2005: 149), Forrest (2003). 2o Forrest (2003) argues that the prevalence of the informal economy in Portuguese Guinea, amongst other things, provides evidence of the significance of ‘non-state social spheres’ during colonial rule. 21 Lourenpo-Lindell(2002: 52). 22 Forrest (2003: 225-6). Castro (1970: 347) estimated that in the 1950s, the colonial government was only able to acquire about one- fifth or one-fourth of total production. 23 Forrest (op.cit.: 222-3).

-15- State attempted to occupy a more central role in the rural areas. Traditional political structures seen as strategic to colonial domination were sidelined. At the community level, elected local councils - village (tabanca) or neighborhood committees, in rural and urban areas, respectively - were put in its place. As a result, some of the powers of the traditional structures were, at least at a juridical level, withdrawn and transferred to the local structures of the modern State.24 Despite these attempts to undermine the existing local authorities, the traditional structures in most cases continued to function, creating in effect a parallel system of competing instances of governance at the local level. The new structures imposed by the modern State, however, often did not end up serving their intended purpose of bridging the gap between society and the State.25 Downward accountability, from the State to the village-level, was often lacking, and few mechanisms were put in place to transmit the views and opinions of village committees up to the central level. Moreover, the village committees did not enjoy the same legitimacy as existing structures of power, and so were often undermined. Undeniably, traditional authorities, particularly in rural areas, continued to wield power in the' communities, either by ending up dominating the State-sanctioned village committees or by sidelining them.26In the end, the village committee system gradually became defunct, leaving the State with little presence at the village level. Therefore, in rural Guinea- Bissau, much like during the colonial era, the regulation and distribution of land and other key resources, the administration ofjustice, and the main provider ofsecurity at the local level, often remains in the hands of traditional authorities, with the State having little or no influence. This has also clearly undermined the State's ability to collect taxes on a regular basis, further undermining the fiscal situation ofthe State and its ability and' capacity to provide services at the community and village levels. Conversely, without receiving any services from the State, the obligation on the part of the citizens to pay taxes to the State was also undermined [it is crucial that attempts are made to bring greater clarity and predictability to the different roles and responsibilities of the competing sources of authority, particularly at the local level]. The law on local autarchy provides some legal foundations for formalizing this relationship, but in the absence of a thorough effort to develop a framework for implementation that aims to integrate customary law and authority alongside the frameworks of the modem State, little can be- expected from such initiatives. As will be explored in greater detail in Box 2.1 of section D below, recently there have been some promising and innovative attempts, particularly as regards the regulation of land, that aim to clarify, harmonize and formalize the relationships between modern and customary forms of land ownership, use and distribution. In so doing, however, it is crucial that attention is also paid to questions of access and distribution to ensure that potential sources ofinequity at the communal levels are not thereby institutionalized.

2.25 The post-independence State largely inherited the weak institutional legacy of the colonial era. Although the successive post-colonial governments attempted to extend the grasp of the State to the local levels, overall, the situation at present has not changed dramatically compared to the one during the colonial rule. In Guinea-Bissau, the

24 LourenFo-Lindell (2002: 61). 25 Forrest (2003: 209). 26 Forrest (2003: 207-9); Rudebeck (1998; 1990; 1997).

- 16- State continues to be largely absent outside ofthe capital Bissau (as evidenced by a weak Government Effectiveness Index in 2002, see Figure 2.1 below). At best, there is an administrative presence ofthe State at regional and sector levels, leaving community and village-levels institutions unattended. In recent years, some progress has been made in social service delivery, notably in the education and health sectors thanks to the support of the Bank and other development partners of Guinea-Bissau (bilateral and multilateral donors, NGOs, and CSOs/CBOs).

D. SECURITY SECTOR, LANDREGULATION AND CONFLICT

2.26 The military has been at the forefront of all major conflicts and instability in recent years. The role of the military in triggering and sustaining instability reflects a vacuum left by Government fragility. In the absence of resilient democratic institutions, the military represents a viable tool for access to power and resources, and is easily manipulated for political gains. This pattern is illustrated in almost all the major destabilizing events that the country experienced since independence. In 1998, the military turned out in support of their Commander-in-chief who was seen as unjustly treated by an alleged corrupt government. In 2003, governance failures prompted the military to oust the president and turn power over to civilians, while continuing to play a critical political role in the back scene. In 2004, the military played a key role in the mutiny that resulted in the killing of the Army Chief of Staff, but many believe civilian opposition politicians were pulling the strings. Moreover, the military suffers from ethnic imbalance, and poor living conditions in the barracks. While it is rarely a major source of conflict in itself, the military remains highly subject to political and ethnic manipulation until these problems are addressed.

2.27 Widespread poverty and ethnic problems within the army are major risks of military-related conflict. Soldiers allegedly lack adequate food supply and dwelling conditions. In order to prevent military-related conflicts, the authorities will need to improve living conditions in military barracks, reduce the size, reform the pension system, and balance the ethnic composition ofthe armed forces through a comprehensive SSR program. This issue will be further addressed in the discussion on short and- medium-term policies in the fourth chapter.

2.28 Control over and access to land is a major source of vulnerability and conflict in urban as well as rural areas. Although Guinea-Bissau’s new Land Law27 was enacted in 1998, the law has not been officially approved yet (see Box 2.1 at the end of this section).28 Without a proper regulatory framework, access to land for dwelling, water, and transhumance represents a major source of conflict. Land remains a basic source of livelihood for the majority of the population. Women, youth, and migrant- groups are potentially vulnerable in access to and control over land. Rural land is managed by traditional authorities, the group historically controlling the land (donas de terra). Consequently, groups that are not indigenous to the area and more recent migrants do not have access to the same assets, land in particular, in which they can make long-

27 ~awn~5198 28 Pursuant to Section 1 of Article 52 of the new Land Law.

-17- term investments. Specific ethnic groups such as the Balante in the south and refugees- from Casamance in the north29 are the most vulnerable social groups as far as access to land is concerned.

2.29 In many cases issues related to access to land and other productive resources involve a family component due to differences in access between genders. Specifically, access to land and labor in Guinea-Bissau, like in most other African countries, requires membership and belonging to social institutions, in particular membership in autochthonous kinship and domestic groups.3o Critical productive resources are vested within these groups, and, in most cases, these resources are principally controlled by more senior men. Marriage is the foundation of these groups, and is the institution through which men are allocated land. Women gain use rights over land through the land allocated to their husband^.^' They are also primarily able to call upon and mobilize labor, through the institution of marriage. Unmarried women or women in female-headed households, (unmarried) youth, and ‘outsiders’, i.e. those that have not descended from the community founders, are therefore also restricted in their access to productive resources. The access to productive resources and participation in- social life by non-autochthones depends on the acceptance and goodwill of senior, male ‘patron’ autochthones - frequently the Chefe de Tabanca- and normally can only occur by marrying into a local, autochthonous family. The ‘client’ outsider would be expected to return the favor. in a number ofways, although these relationships are never equal and work in favor ofthe more powerful patron who has greater bargaining power. Fieldwork carried for this study picked up on the problem of access to land by non-natives (non- autochthonous) - including immigrants from the Casamance region of to the north of the country but also migrants internal to Guinea-Bissau - who as a result were. either left without the ability to farm (in some cases), or gained limited and conditional access to land.

2.30 Those that gained access to land often had to abide by a range of conditionalities, such as being limited to cultivating rice, corn or cassava, and not being allowed to use the land for hticulture and cashew plantations. Access rights were also only temporary, meaning that non-native households had to re-negotiate the access to and use of land on an annual basis. Land received was also often oflesser quality, which in some parts ofthe. country (e.g. the less fertile East) could mean chronic food shortages. Women’s access to land and other productive resources was also seen to be affected by this. As women only gain access to land through their husbands, which creates a certain level ofdependency of women on men, women that had been abandoned, divorced or widowed would in some cases sometimes find themselves without access to land. While women were in most cases still provided with land, though mainly on a temporary basis, the more serious problem in these cases was their access to and ability to mobilize labor to help work and

29 The UNHCR estimates a total refugee population in Guinea Bissau around 7,600 in 2003, including all nationalities.2003 Global Refugee Trends, UNHCR, June 15 2004 30 Autochthonous groups refer to those descended from the original founders of the community. For information on the larger African context, see Berry, 1993. ” In rare cases, women can also own land in their own right. Widowed women might inherit land in trust for their sons which is then allocated to the sons upon marriage. Women that are single or divorced, and widows who have lost all their sons, may be forced to retum to their families and might be given a small area to farm. Some widowed or otherwise single women have to negotiate access to land every season.

- 18- cultivate the land. The extent of this problem seemed to vary according to ethnicity and religion, with women in the East seeming to be left more vulnerable if single, widowed or abandoned, than in other areas.

2.31 The issue of access to land is of immediate importance given the current national and legislative debate on reforming the access regime to land. While the new law makes a grand leap forward in recognizing traditional authorities’ role in managing and distributing communal land, the discussion here highlights the importance of also addressing the specific needs ofwomen, youth and migrants in accessing land. Moreover, the regulations also need to take seriously the many latent yet intensifying conflicts over access to land, particularly those between livestock breeders and agriculturalists, as well as between traditional communities and private investors (including ponteiros). Conflicts over land in urban areas are also becoming more serious, most of which have been between urban settlers and the Municipal Council of Bissau (CMB). The CMB has taken on the responsibility of managing land in Bissau, and claims ownership of much land currently being inhabited by poor dwellers, particularly in the outskirts of the city of Bissau. The local government has encroached extensively also on customary land in Bissau and nearby Biombo, literally expropriating the Pepel (native owners of these lands) of large shares of their land. An area in Alto Bandim, for example, which was originally used by local farmers for cashew production, but which was expropriated by the municipal government without any compensation and later sold to private persons or turned into luxury residences, such as the recently built Ministers’ Quarters on Alto Bandim.32 Perhaps most importantly, efforts need to be made to ensure the implementation capacity of the new law and its regulations, without which the land law will have little impact on the ground.

2.32 The availability of labor in rural areas seems to be decreasing. Fieldwork documented a so-called “rural exodus”, primarily of youth escaping rural areas in search of further education and better life opportunities in urban towns and cities. In addition to the thousands of uneducated youth this adds to the informal and unemployed labor market in Bissau, this exodus is also leaving rural areas without the required productive capacity to farm the lands and ensure sustainable rural livelihoods. Again, however, the households with large proportions of elderly and female members are as a result at a particular disadvantage. Economic development within a context of agrarian transformation and industrialization represents risks of vulnerability and exclusion in- access to land. Moreover, climatic trends may escalate land vulnerabilities and conflicts in the near future. Although rural land in Guinea Bissau is still abundant, saltiness ofland and reduced rainfall, along with fast growing cashew farming, may create further tensions unless land rules are clearly regulated and disseminated. Land related tensions are therefore likely to persist in the future if adequate legal and regulatory mechanisms are not put in place.

2.33 Conflicts over land continue to disrupt agricultural and other production- activities, and undermine livelihoods for the poor. Access to water and transhumance areas are major sources ofviolence and disruptions in agricultural production. Four types

32 Lourenpo-Lindell, 2002

- 19- of latent conflicts have been observed in rural areas. First, access to water is reported as one ofthe principal source of conflict. Second, pastoralist stockbreeders migrate to more fertile agricultural land where their livestock may cause destruction in planted fields. Since land distribution is generally managed by traditional authorities, such inter- community conflicts between sedentary farmers and migrating stockbreeders may be more difficult to resolve. New regulations include requirements that private landowners as well as local Management Committees designate corridors of passage of livestock. Experience from other countries particularly points to the importance of access to water within these corridors. Third, conflicts have already been observed between traditional land management and use, and external investors for commercialization of agricultural product, particularly cashew. In these cases, ceding land by government authorities toward private use clashes with customary land management in communities. These types" of conflicts may increase in potential as commercialization of cashew plantations expands. Finally, land for urban settlement is another major source ofconflict which will need thorough attention in the context ofthe land regulation system.

2.34 Guinea-Bissau needs adequate legislation for aligning the roles of traditional authorities and the State in management of land. A regulation to the 1998 Land Law is currently under elaboration (see Box 2.1). The legislation seeks to align the responsibilities of the authorities and the State in management of land by giving legal- authority to traditional leaders, and involving them in the local commissions to be set up to govern land distribution and taxing. The division of responsibilities and capacities between the four levels (national, regional, sector, and section) of the commissions to be established will be important to avoid duplication and conflicts, particularly between commercial use and local traditional use. Information and dissemination of the new regulation will be critical to its impact and effective enforcement. The regulation includes provisions for taxing of land, and the local commissions with traditional leaders being- responsible for tax collection. Finalizing this legislation will be critical for land management and conflict prevention in Guinea-Bissau

- 20 - Box 2.1: Land Regulation in Guinea-Bissau

Under the colonial administration, two types of property rights had been conceptualized under Decree 43893 of ,1961: the land right of the State and that of the Copmunities called reserved areas. The reserved areas had covered the areas actually under cultivation and residential areas, yet excluded the areas under fallow and extraction activities (fishing and forestry). The Land Law 4/75 adopted after the independence in 1975 had transferred land property to the State.

The new Land Law (Law nQ5/98) adopted in 1998 has articulated three principal objectives: (i) guarantee land rights to local communities for economic use; (ii)acknowledge customary land regime by representative institutions; and (iii)Stimulate land investment by creating market value. Contrary to the previous legislation, the new law has authorized and introduced permanent or temporary land concessions both in urban and rural areas. In addition, the law has created a tax mechanism to increase the land use efficiency aimed at dismantling large-scale farms (Zutifundios) unable to demonstrate income generating capacities. Furthermore, it has permitted the formation of an administrative commission of land whose objective is to safeguard the law enforcement and consorted interventions in the land use. The land in Guinea-Bissau has been declared a property of the State constituting a common property of all the people (Article 2Q, Section 1). Yet, all citizens have a right to private use of land, and the State can grant the right of private use to individuals and collective units, national and foreign entity for economic and social development objectives (Article 4Q, Section 1). Such land rights for private use are recognized either by: a) customary use (us0 consuetudincirio), as is envisaged land use by traditional local communities, and b) concession (Article 4O, Section 3). The present law therefore recognizes land use rights based on customary use practiced by traditional local communities. This is an important instrument to recognize and guarantee traditional rights in a modern legislation context. The private land use right under customary use is a permanent title and can be exercised in rural and urban land including reserved areas for local communities (Article 16). The land under the customary use covers cultivated and inhabited zones as well as unexplored zones and resources attributed to the local community residents through their representatives. The administration and distribution of the land under customary use by the local community residents is to be exercised according to the custom and practices of each community, and any omitted part shall be governed by this law (Article 17, Section 1, 2). The law clearly stipulates in Article 17, Section 3 that the State recognizes the residents’ right for community administration and exploration of the land, forest and other natural resources according to the local custom and practices. The private land use right under customary use is guaranteed by the State independent of written contract or register, obliging the Land Commission to promote and constantly update the registration services (Article 17, Section 4). This right shall be transferred by successive inheritance and is transferable to other resident(s) within the local community free of charge, according to the local practices independent of endorsement by contract and register (Article 18). The transfer of the private land use right under customary use to a third party other than those referred to in the Article 18 shall be made in agreement with the State and the local communities and follow the process to be applied in the concession. In such cases the population of the local communities retains the right to freely and directly negotiate the terms of transfer in accordance with the conditions specified in the Article 23 (Article 19). The private land use right under customary use can be converted into the private land use right under concession (Article 21). The private land use right under concession is granted by administrative contract and has a duration of 90 years to be automatically renewed if not revoked (Article 22). This right is transferable by contract or inheritance upon request for official authorization. The process of grant, transfer of and conversion into the private land use right under concession involves tax obligations, and the right is subject to annual land tax payment (Article 38). The land tax revenues shall be distributed in the following manner; a) 60 % to the public treasury, b) 20 % to local communities, c) 10 % to regional and sectoral administrative authorities, and c) 10 % to the Land Commissions.

- 21 - 2.35 Finally, there is a critical need to resume work on the development of the cadastre. As indicated in Box 2.2 below, prior to the conflict the authorities initiated work on an integrated cadastre system. The task was financed by the EU under the VI1 FED program, and planned for execution during the period of April 1996 - March 1999. The objective of the project was to develop a land tenure system for regulating land management in rural areas. The initiative was stopped in June 1998 when the conflict broke out. It is therefore critical to update the topographical map to a digital topological structured map database. The new database was created from existing 1:50,000 paper maps using the satellite imagery for update.

Box 2.2: Status of the Cadastre

Prior to the conflict, the Ministry of Plan and Cooperation was working with a private Danish firm, ‘Consultant Within’ (COWI), to design an integrated cadastre and land management system. The goal was to develop a land tenure structure and to improve land tenure security and legitimacy for the rural population. The project included four main activities: (i)rehabilitation ofthe geodetic network (review of the present network, planning, construction ofbeacons, surveying, and evaluation); (ii)organization of the cadastre (institutional development and regional organization, improvement of the physical infrastructure and training); (iii)improvement of village delineation (development of a new system for village delineation and methodology for improved land tenure security for small-scale farmers and subsequent implementation in different areas); and (iv) improvement of land inspectiodtaxation base (review and updating ofland concession processes, including transfer to database, field inspections, and implementation ofthe new land law and land taxation system).

During the project, it became clear that a new cartographic basis was needed. Earth Observation Satellite (SPOT) images were used as a new comprehensive basis with a digital output and printouts at scale 1:50,000 to fit into the existing topographic map series used as cadastral maps. The project also contributed to the analysis and design ofthe new Land Law approved by Parliament in March 1998.

A community participatory approach was introduced as a way to improve the understanding of the customary and local land tenure structures and to integrate the rural population in the land titling activity. Costhenefit aspects were focused on the land titling activities, with a view to sustainability. Privatization of the surveying fieldwork as well as the institutional framework of the Directorate was analyzed.

The staff at the Directorate of the National Services for Cadastre and Cartography was trained in rapid and participatory rural appraisal, low-accuracy Global Positioning System (GPS), static GPS, traditional surveying methods, software for surveying, database, mapping, accounting, reporting, and English.

m:COWVKAMPSAX h~://www.kampsax.co~projects/div04proj/Countes/divO4~GUMEABISSAU.asp

- 22 - E. CONCLUSION AND POLICY CONSIDERATIONS

2.36 At this juncture of Guinea-Bissau transition from post-conflict to long-term development, it is ofoutmost importance to enhance political stability in order to mitigate risks ofconflict as experienced in the past. In this context it is critical to address security and land issues, the two main factors of conflict in the country.

2.37 In the short-term, improving living conditions in military barracks will be critical for preserving political stability. In that vein, key immediate actions under the SSR would include the rehabilitation of military dormitories, and sanitation and food services in military caserns.

2.38 In the medium-term, the SSR would need to address issues related to downsizing the armed forces, reforming the pension system, and balancing the ethnic composition of the security sector. With regard to land reform, the Government would need to officially adopt and implement the new Land Law in order to provide a legal and regulatory framework for the resolution of conflict stemming from land use. Finally, updating the cadastre will be critical for the implementation ofthe land reform.

- 23 - 3. POVERTY ANALYSIS

Over the last two decades, the incidence of poverty has remained relatively high in Guinea-Bissau, and adversely afected economic performance, governance and the development of institutions. About two-thirds of the population lived in poverty in 2002. Poverty is characterized by limited access to education, health, water and energy services. People living in rural areas and practicing agriculture bear the highest poverty incidence (63 percent), compared to those living in urban areas and earning salary incomes (50 percent). Moreover, the poverty incidence is higher for men compared to women. The analysis also provides evidence that agriculture represents the main livelihood strategy for the poor. Finally, estimations based on empirical evidence suggest that the efect of the conflict has signijicantly undermined prospects of achieving the MDG objective of halving the poverty rate by 2015.

3.1 . The 1994 poverty assessment based on the 1991 ICOF data estimated that nearly four-fifths of the population was living in poverty in the early 1990s. During the 199Os, the Government implemented a package of reforms, which contributed to improved economic performance. The latter may explain the slight decrease in the poverty incidence between 1991 and 2002 as shown in Section A below. If the expansionary trend was not disrupted by the conflict it is believed that poverty would have decreased at a higher rate, hence increasing chances for achieving the MDG target for poverty. The main objective of this chapter is to analyze socioeconomic characteristics and determinants ofpoverty in Guinea-Bissau. The chapter is structured in the five sections as follows: (i)poverty profile; (ii)determinants ofpoverty; (iii)livelihoods ofthe poor; (iv) conflict and poverty; and (iv) conclusion and recommendations.

A. POVERTY PROFILE33

1. Monetary Poverty in 2002

3.2 The methodology used in the report is consistent with the authorities’ PRSP. The quantitative analysis is based on the 2002 ILAP survey conducted by the National Statistical Institute (INEC).34The survey’s sample size is 3216 households, ofwhom 672 live in Bissau, and 2544 in the rest of the country. The methodology for estimating poverty is non-standard, due to weaknesses in the data (limited questionnaire on food consumption, among others). Instead of computing a poverty line based on the cost of basic needs method, we follow the PRSP as well as previous work35 in assessing the value of the international US$1 and US$2 poverty line in local currency units. It turns out that the extreme poverty line corresponding to the US$1 threshold is CFAF 108,000

33 All mentions to poverty and poverty incidence, in this section, refer to the poverty headcount index. 34 The survey was actually implemented between March 2001 and April 2002. 35 Syllah (2004)

- 24 - per person per year. The poverty line corresponding to the US$2 threshold is therefore CFAF 216,000 (see Box 3.1). Again following the method used for the preparation ofthe PRSP, the indicator of well being on which the poverty measures are based is the consumption level per equivalent adult. Household members above 15 years of age are considered as adults, while younger members count for one half ofan adult.

3.3 Nearly two-thirds of the population was living in poverty in 2002. The annual consumption per capita recorded in the survey is CFAF 164,061 and the consumption per equivalent adult is CFAF 211,927. As shown in Table 3.1, the individual level poverty measures suggest a headcount of poverty of 65.7 percent and a headcount of extreme poverty of 21.6 percent. When population weights are taken into account (i.e., when the poverty rate is computed at the household level), the incidence ofpoverty is estimated at 58.9 percent, while that ofextreme poverty is 16.8 percent.36

3.4 Poverty decomposition by region reveals that coastal regions have lower poverty levels compared to landlocked areas. There are noticeable differences in the poverty measures by region. As shown in Table 3.1, the coastal towns of Bissau and to a lower extent Biombo/Bolama and Cacheu have the lowest levels ofpoverty, whereas the interior areas ofBafata and especially Oio have much higher levels ofpoverty. While this finding seems to contradict conventional wisdom, it is likely to be an indication of the pervasive effect ofthe region’s heavy reliance on the production ofraw cashew which is a low value-added activity. As discussed latter in chapter five, cashew kernel processing could represent a good alternative for creating more jobs and reducing the incidence of poverty. Yet in comparison to some other sub-Saharan Afr-ican countries, the differences in poverty between regions are in fact somewhat limited. That is, poverty seems to be widespread everywhere in the country and it also affects many different segments of the population (see the fourth chapter ofVolume 2 for further discussion). Table 3.1: Poverty Decomposition by Region in 2002 (in percentage) ’’ GIN1 Poverty Extreme Poverty Region index Po P1 p2 Po p1 p2 National 36 ” 65.7 25.7 12.9 21.6 5.5 2.2 Bissau 37 52.6 17.5 7.6 9.8 2.0 0.6 Bafata 73.2 30.5 16.0 27.2 7.5 3.1 Gabu 67.0 25.7 12.8 20.4 5.6 2.5 Cacheu 64.7 27.9 115.2 28.4 7.8 3.1 Oio 80.4 34.8 118.6 35.3 9.0 3.6 BiomboBolama 63.9 22.4 10.0 14.5 2.8 0.8 QuinaraITombali 70.3 27.4 13.9 23.1 6.5 2.5 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002. ” PO,Pi and P2 are the usual poverty headcount, poverty gap and poverty square measures. 12 The GMI index was estimated at 40.3 percent in 1994.

36 The relative importance of the individual-level poverty compared to the household-level can be explained by the fact the household head is more likely to eam an income as opposed to each individual member of the household.

- 25 - Box 3.1: 2002 ILAP Survey - Design, Data Coverage and Construction of Poverty Line

Design. The ILAP data collection was carried out from March to May 2002. The survey was based on the Questionnaire on Basic Welfare Indicators (QUIBB), designed by the Bank in collaboration with the United Nations agencies specialized in development. The 2002 QUIBB was a light survey as opposed to 1991 full fledged household income/consumption survey (ICOF). It was carried out in a relatively short timeframe during which data processing and consistency check is performed in a coherent manner. QUIBB is based on a master sample of 3,216 households shared between the autonomous sector of Bissau (672 households) and the other areas of the country (2,544 households). The sample selection relied on ‘Census Districts’ (Districts de Recensements, DRs) as established in the General Census ofPopulation of 1991. The work ofupdate of the DRs located in Bissau was made in December 2001 within the framework of the WAEMU’s Harmonized Consumer Price Index project. The DRs for the other regions were updated based on information gathered during the survey. The full sample is then divided into seven (7) strata: Bafata, Gabu, Cacheu, Oio Biomba/Bulama, QuinardTombali and the autonomous sector of Bissau. The sample selection was thus a two-stage selection process. The fxst stage of selection was the master sample. The second stage of selection was a sub-sample of 12 households for each stratum. They were selected systematically with equal probability and, moreover, equally allocated among the 7 regions ofthe country.

Data Coverage. QUIBB was designed to obtain the relevant indicators necessary to profile the well- being ofpersons and households in Guinea-Bissau, and to provide reliable estimates ofthese indicators at the national level, for Bissau and the other regions in the country. QUIBB data used various expenditure characteristics of households (housing, property, equipment, agricultural land, livestock, cattle, etc.) to measure the standard of living of the populations, and analyze the degree of satisfaction of the populations with regard to the delivery of basic social services in the areas of infrastructure, education and health. However it did not include disaggregated information on production, expenditure and price data for key agricultural commodities (cashew nuts, fish products, etc.).

Poverty Line. The report used the poverty line constructed by INEC for the purpose ofthe authorities’ PRSP work. The construction of this line required the definition of a welfare indicator using adult equivalent expenditures. The latter included the following expenditure items: food and non-food consumption, imputed rent and transfers. These expenditures were then converted into adult equivalent values based on the so-called Oxford Equivalence Scale, which assigns a value of 1 to people aged 15 years old or more (adults), and a value of 0.5 to children (less than 15). The poverty line was constructed as follows: First, the adult equivalent expenditures were adjusted for purchasing power parity (PPP) factors based on the US inflation rate. Second, the international poverty line of $1 and $2 PPP(l985) were used to determine the $1 and $2 PPP(1999) using the Guinea-Bissau PPP conversion rate for 1999 available in the World Development Indicators (WDI) 2001. Finally, the $1 and $2 PPP(2002) poverty lines were constructed by updating the 1999 poverty lines using the Harmonized Consumer Price Index ofthe WAEMU.

2. Subjective Perceptions of Well-Being and Safety Nets in 2005

3.5 This section analyses people’s subjective perceptions of poverty, and the role of social networks in Guinea-Bissau. It uses the 2005 QS data to construct an asset index, which is used to classify the households according to their level ofwealth (see Box 3.2; a detailed account of the methodology used is presented in the third and fifth chapters of Volume 2). On the basis of this wealth index, households were ranked in three terciles: The poorest households are in the first tercile, and the richest are in the third, with each of the three terciles accounting for one third of the households. Roughly speaking, the first

- 26 - tercile can be seen as representing the very poor, while the bottom two terciles account for the poor.

Box 3.2: 2005 QS - Methodology and Data Coverage and Analysis

Methodology. For the purpose of the QS, the country was sampled in 5 regional zones, Bissau and Byombo being one zone, along with the East, Islands, North, and South. Community-level fieldworks targeted at least 16 research sites (villages/neighborhoods). In each regional zone, thee communities were selected except in Bissau and Byombo where four communities were retained. The research sites were selected based on a purposive cluster sampling drawing on a mix of pre-identified criteria, including: Rural/urban characteristics Social diversitylethnic group Predominant livelihoods/agro-ecological zones

Levels ofaccess to infrastructure and services x The survey was conducted by two international consultants and four local consultants organized in two teams. Each team covered 2 regional zones: one team for the North and the East; and the other team for the South and the Islands. The research sites in Bissau and Byombo were shared between the two teams. Interview guides and questionnaires were developed for the rapid survey, and the semi- structured interviews camed out in the context offocus group discussions.

Data Coverage and Analysis. The QS covered a total of446 households and collected qualitative data on perceptions of livelihood means, institutions, conflict and social service delivery. The information collected was primarily used to estimate descriptive poverty statistics for the aforementioned areas.

3.6 The majority of people believe that their poverty situation has worsened following the war. Clearly, an overwhelming majority of households declare that they are worst off today as compared to before the war, with few variations between terciles (see Table 3.2). In addition, a majority also believes that they are worse off today than one year ago, but here the proportion is especially high among the poor or extreme poor. This suggest that the continuous state of instability may be more detrimental for poorer than wealthier households. The last four items in Table 3.2 are measures of subjective poverty and information on the number of meals per day, whether households feel that they are lacking food, and if they do, when this occurs. These data suggest that households that have been classified in the bottom terciles do indeed see themselves as comparatively poorer that those in the top tercile, they have fewer meals per day on average, and they are also more likely to lack food (this affects a very large share of the poor), especially during the rainy season.

3.7 Social networks often seem to provide crucial sources of support for poor’ households, particularly during times of stress and coping. A minimal social protection is offered to many households by ‘safety nets’ that poor households build among them~elves.~~There exist a range of social networks, the most common ofwhich are those based on kinship and family, village-membership, religion, and neighborhood ties. Kinship ties are often important sources of food and money, both for consumption purposes, as well as for market activities. Neighbors in urban areas living within the same house or compound, provide a similar role. Also in urban areas, relations of assistance

37 Lourenpo-Lindell, 2002.

- 27 - develop at the market place in a variety of ways between sellers in the same trade who assisted each other, sometimes pooling resources and contacts. Some traders count on credit from suppliers. Abotas, the rotating savings groups referred to earlier is another key resource. Another kind of cooperation consisted of organized groups among day workers (surni), where men share job opportunities and redistribute incomes. At the village-level, during the last two decades there have been revivals of traditional social clubs, mandjuandades, which provides recreational and some level of economic assistance to its members.38 Finally, some households participate in local religious groupings, Catholic, Evangelical and Islamic congregations. Some of these, however, may be more important in terms of providing personal acquaintances and contacts for cooperation in other social settings, such as the market, than in terms of the volume of material assistance they provide to needy members.

Table 3.2: Subjective Perceptions of Well-Being in 2005 Tercile Area Total 1 2 3 Urban Rural Change in well-being versus before the war Got better 14.18 12.06 15.79 12.18 15.06 13.98 Same 9.22 9.93 7.52 7.05 10.04 8.92 Got worst 75.89 77.30 76.69 80.77 74.13 76.63 Not applicable 0.71 0.71 0 0 0.77 0.48 Total 100 100 100 100 100 100 Change in well-being versus last year Got better 28.47 27.66 40.14 33.73 3 1.03 32.08 Same 18.06 15.60 9.86 10.84 16.86 14.52 Got worst 53.47 56.74 50.00 55.42 52.11 53.40 . Total 100 100 100 100 100 100 Household situation versus its community Very poor 33.83 29.69 16.92 23.33 29.05 26.85 Poor 4.51 3.13 4.62 4.00 4.15 4.09 Medium 45.11 57.03 50.77 50.67 5 1.04 50.9 Rich 15.79 9.38 26.92 21.33 14.94 17.39 the most rich 0.75 0.78 0.77 0.67 0.83 0.77 Total 100 100 100 100 100 100 How many meals do you take in a day? One 43.97 29.41 26.09 25.31 38.34 33.25 Two 37.59 43.38 44.20 46.30 38.74 41.69 Three or more 18.44 27.21 29.71 28.40 22.92 25.06 Total 100 100 100 100 100 100 Lack of food Yes 93.71 95.74 73.24 79.52 92.69 87.56 No 6.29 4.26 26.76 20.48 7.3 1 12.44 Total 100 100 100 100 100 100 Ifyes when? End of dry season 0 4.51 2 1.57 2.52 2.19 Rainy season 60.61 55.64 45 42.52 60.92 54.52 Beginning of rainy season 7.58 6.77 5 2.36 8.82 6.58 Beginning and end of rainy season 3.03 0.75 11 8.66 2.10 4.38 Dry season 7.58 8.27 8 8.66 7.56 7.95 . Rainy season and dry season 3.79 0.75 1 1.57 2.10 1.92 Othershowresponse 1.52 0 1 0.79 0.84 0.82 All the time 15.91 23.31 27 33.86 15.13 21.64 Total 100 100 100 100 100 100 Source: Staff estimation from 2005 QS.

38 These informal, communal groups, which may include both women and men, are structured according to age. Within a tabanca a new group is established about every three years, and is often named after the major event occuning in the tabanca that year. In Cai6, the groups go under the name wan; the latest group was established there in 2004 and is called uran telemobel named so as this was the year cell-phones became popular in Guinea-Bissau.

- 28 - 3.8 Nevertheless, these safety nets are limited and often exclude the poorest of the most vulnerable households. Households and individuals vary in their access to social networks and are therefore not equally protected against shocks and unexpected events. The more vulnerable of individuals and groups are those that are more isolated, with a very small number of sources which they can rely on for assistance. Also, many households or individuals do not seem to be in a position to combine different sorts of these ‘safety nets’. Some remain very reliant on one type ofnetwork, seemingly unable to select those networks they prefer and to break free from those they deem oppressive and disadvantageous. Other networks are based on hierarchical and inegalitarian patron-client relations that while providing some source of security to the poorer client, the relationship may still be abusive. Hence, the existing social networks are not necessarily a panacea to solve the welfare problems ofthe poor.

3.9 Vulnerability may also be exacerbated during periods of general economic downturn - the time when social assistance is often needed the most. Economic hardship seems to erode the collaborative efforts among the poor, as well as the material basis of their support networks. As a result, many are forced to rely on their own resources, as they perceive other members oftheir networks as being as badly off as they are. The redistribution groups among casual workers, for example, have over the last couple of years become less and less reliable, due to a continued influx of unemployed youth into the labor market, a reduction in employment opportunities, and declining incomes. In a situation of increasing impoverishment, the assistance provided by religious groups is also put under pressure. And, as noted earlier, participation in the abotas also becomes difficult for those unable to comply with regular contributions in times of uncertain and irregular incomes. As such, it may seem that those that rely exclusively on market-based networks - while having more flexibility and choice - may end up being more vulnerable than those that also, or even only, have access to kinship- based sources of support, which may be more tolerant of the inability to reciprocate during long periods oftime. Given this, it would be worth asking the question ofwhether it may be likely, that those who due to their social status had limited access to productive resources - single, widowed or abandoned women, (unmarried youth) and non-native migrants - may also have lesser access to kinship based networks.

3.10 High inequality and social diversity are potential threats to the formation or maintenance of social capital. The country comprises some 20 different ethnic groups, underscoring the fact that social capital formation may not always be durable and sustainable, particularly in urban settings where inequality and diversity is at greater display. When asked about potential sources of inequality in the country, households mentioned that religion, age and gender were more important factors at the source of existing inequalities than ethnicity (see Table 3.3). Geographic location played a less important role in the mindset ofthe population, even though it can be demonstrated that it actually plays an important role in explaining differences in well-being between households (see Chapters 3 and 4 ofthe for further discussion).

- 29 - Table 3.3: Main Source of Inequality in 2005 Tercile Area Total 1 2 3 Urban Rural Ethnicity A lot 4.65 7.41 14.39 14.10 5.42 8.84 A little 25.58 31.11 34.85 35.90 27.08 30.56 No 69.77 61.48 50.76 50 67.50 60.61 Total 100 100 100 100 100 100 Religion A lot 13.99 12.86 34.3 1 31.87 13.08 20.24 A little 25.17 28.57 22.63 21.88 27.69 25.48 No 60.84 58.57 43.07 46.25 59.23 54.29 Total 100 100 100 100 100 100 Age A lot 35.1 1 19.26 26.52 21.15 30.58 26.88 A little 14.50 22.22 30.30 30.77 16.94 22.36 No 50.38 58.52 43.18 48.08 52.48 50.75 Total 100 100 100 100 100 100 Gender A lot 10.85 14.81 20.74 20.63 12.13 15.54 A little 34.1 1 25.93 36.30 32.50 3 1.80 32.08 No 55.04 59.26 42.96 46.88 56.07 52.38 . Total 100 100 100 100 100 100 Geographic Location A lot 3.88 10.53 16.79 18.83 5.02 10.43 A little 34.1 1 27.07 32.82 31.82 30.96 31.30 No 62.02 62.41 50.38 49.35 64.02 58.27 Total 100 100 100 100 100 100 Source: Authors’ estimation from 2005 QS. 3.11 The fragility of the community-based safety net system; as noted above, underscores the need for the authorities to seriously address the issue ofsocial protection of the most vulnerable. The inclusion of these issues under the third pillar of the authorities’ draft PRSP represents a positive move, and would need to be articulated and sustained over the medium-term and in the next generations of PRSP to follow.

B. DETERMINANTSOF POVERTY

1. Socioeconomic Determinants of Poverty in 2002

3.12 The poverty incidence is lower among people within the age group 31-45‘ relative to the others. The poverty rate among people within the age cohort of 31-45 years is lower (62.8 percent) than the rate among people who are more or less than 45 years old (more 65 percent in average) (see Table 3.4). As expected, the highest incidence ofpoverty occurs among people aged 66 or more (75.6 percent), who are likely to have a higher level of family responsibilities due to the relatively large household size and high dependency ratios (as evidenced by a relatively high poverty incidence among people aged 31 or less). Senior people are expected to have accumulated wealth from years ofemployment and normally would be financially better offthan relatively younger persons. However, in Guinea-Bissau as in other parts ofAfrica, social security is not well developed and workers are unable to save because of a lifetime of poor employment conditions. With these factors combined with relatively high dependency ratios, it is not surprising that the incidence ofpoverty is liigher among people in the 66 plus age group.

- 30 - Table 3.4: Poverty by Age in 2002 Poverty (in percentage) Household Size (in number of individual) Age Category PO P1 p2 45 65.5 25.4 12.7 9.6 16-30 65.7 25.5 12.7 9.3 3 1-45 62.8 24.3 12.1 8.7 46-65 69.3 27.9 14.4 9.0 Xi6 75.6 31.8 17.2 9.3 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

3.13 While the overall poverty incidence is even across gender, the decomposition by gender age groups yields mixed results. The national poverty is about two-thirds for men and women alike. However, the incidence ofpoverty is higher for women below 31 years and above 65 years compared to men (up to 3 percentage points of difference in the poverty headcount measures). By contrast, women tend to be better off compared to men within the age cohort 31-65 (2 to 10 percentage points of difference in the poverty headcount measures) (see Table 3.5). A number of reasons could explain this unusual phenomenon. Traditionally, the formal sector of employment, particularly the public sector, has been dominated by male workers. Women, on the other hand, have typically been most active in the informal sector. The groups of small traders, mostly women that resisted the restrictive policies of the colonial and early post-independence State expanded greatly with the liberalization ofthe economy in the late eightiedearly nineties. In addition, the hardships imposed by adjustment and the contraction ofthe public sector have increasingly served as incentives for women to engage in income generating activities to supplement the income of the household. Moreover, an increasingly larger number ofwomen are involved in the processing ofcashew nuts, which account for more. than 90 percent of Guinea-Bissau's export earnings, and is a major source of income for households. According to the Ministry of Agriculture, the planting and processing of cashew provides employment for 82 percent of the rural workforce, of which women represent 49 percent.

Table 3.5: Poverty by Demographic Characteristics in 2002 (in percentage) Demographic Characteristics Poverty Ext. Poverty Household Size Gender Male Female Male Female Male Female . Age Category 45 65.3 65.6 21.4 21.2 9.6 9.7 16-30 64.4 67.2 20.3 21.8 9.2 9.4 3 1-45 67.3 57.5 23.1 17.2 9.1 8.2 46-65 70.7 68.1 27.5 20.8 9.2 8.8 ;66 73.5 77.6 31.6 33.1 9.5 9.2 Milieu Bissau 52.6 52.6 10.5 9.0 8.3 8.3 Rest ofCountry 70.5 70.1 26.0 25.4 9.7 9.7 Region Bissau 52.6 52.6 10.5 9.0 6.9 6.4 Bafata 74.2 72.1 27.1 27.2 10.8 10.5 Gabu 66.8 67.2 19.6 21.2 10.9 10.4' Cacheu 64.1 65.4 27.9 29.0 9.1 9.0 Oio 80.5 80.3 37.0 33.5 9.8 9.6 BiombolBolama 64.1 63.7 14.6 14.3 8.6 8.9 QuinardTombali 70.9 69.7 23.6 22.7 9.5 9.2 National 65.9 65.6 22.0 21.1 9.4 9.3 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAF') Data, 2002.

-31 - 3.14 'Low education limits the income generation capacity of the poor. Empirical evidence suggests that the lower the education level, the higher the probability of being poor. In Guinea-Bissau, the incidence ofpoverty among people with tertiary education is 32 percent, a rate which systematically increases to 71 percent among people with no. level of education (see Table 3.6). Low education has a pervasive effect on poverty because in addition to its impact on income, it has significant spillover effects on other socio-economic factors, such as the health status of children, reproductive behavior, infant and child mortality and em~loyment.~~

Table 3.6: Educational Attainment and Poverty in 2002 (in percentage) Poverty Educational Attainment Quintile Distribution Headcount .

Rest of Female Male Bissau Country Poorest 2"d 3rd 4th Wealthiest No Level 71.6 54.8 27.2 69.7 68.6 62.4 60.1 55.6 47.1 71.1 Primary 16.3 26.0 28.7 22.7 23.5 25.8 24.2 24.8 23.8 65.3 Secondary 10.3 15.6 34.1 7.0 6.3 10.2 13.9 16.4 22.1 49.4 Superior 1.5 2.8 7.4 0.6 1.1 0.8 1.2 2.8 5.7 32.2 Other 0.3 0.9 2.5 0.0 0.5 0.8 0.6 0.4 1.3 56.2 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

3.15 Poverty incidence remains very high in agriculture. Agriculture provides employment for over 55 percent ofworkers at the national level. As expected, 74 percent of the people working in the agriculture sector live outside Bissau and mostly belong to the poorest quintile. Bissau, on the other hand, provides employment for a higher percentage of people working in administration. Not surprisingly, individuals living in households where the head works in the agriculture sector have a higher poverty incidence (56 percent) compared to individuals in households where the head works in other sectors (see Table 3.7).

3.16 Poverty incidence is lower among Government and private sector workers. The majority of government employees are based in urban areas and there seems to be a positive relationship between the economic status of households and employment in the government sector. The low rate of poverty observed among private sector employees could be attributed not only to better pay in the private sector but also the stability ofthe pay or the better education endowment ofprivate sector workers.

Table 3.7: Poverty and Sector of Employment in 2002 (in percentage) Quintiles Poverty All Urban Rural Female Male Poorest 2"* 3'* 4' Wealthiest Headcount Agriculture 55.82 6.00 74.38 40.92 57.98 70.60 57.92 56.23 53.04 42.08 63.1 Industry 5.14 6.44 4.66 16.88 3.44 1.48 6.91 5.95 4.85 6.39 51.9 B.T.P. 6.65 10.8 5.10 0.18 7.59 8.32 6.1 6.5 5.80 6.59 58.6 Transport 1.37 3.37 0.62 0.50 1.50 0.97 0.39 1.81 1.47 2.16 44.6 Commerce 9.86 20.09 6.04 23.97 7.81 5.51 8.95 9.03 10.52 15.04 47.9 Other Serv. 6.58 15.99 3.07 4.29 6.91 3.21 8.19 6.27 7.44 7.65 51.9 Educ./Health 1.25 2.52 0.78 1.96 1.15 1.65 0.86 0.61 1.48 1.68 46.1- Publ. Adm. 13.33 34.8 5.33 11.3 13.63 8.27 10.69 13.6 15.41 18.40 50.4 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

39 Creppy (2003)

- 32 - 3.17 Difficult access to basic social services is associated with higher poverty incidence. In the capital Bissau, the headcount poverty among those who live less than 15 minutes from potable water, primary school and health facilities (45.3, 44.9, 42.7 percent, respectively), is lower than those who live more than 15 minutes from these services (57.1, 48.2, 47.0 percent, respectively) (see Table 3.8).40 In the rest of the country, however, the only services with a similar pattern of distribution are secondary- schools and health facilities. In these regions, especially, a higher proportion of the poor lives closer to the source ofwater and primary schools (see Tables 3.9).

Table 3.8: Poverty by Distance from Basic Services - Bissau Region in2002 (in percentage) Poverty rate in percentage for those who have access to: Distance Secondary

(Minutes) Potable Water Food Market Primary School School Health Facility I 0-14 45.3 46.4 44.9 48.2 42.7 15-29 57.1 45.1 48.2 45.0 47.0 Total 45.8 45.8 45.8 45.8 45.8 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

Table 3.9: Poverty by Distance from Basic Services - Rest of the Country in 2002 (in percentage) Poverty rate in percentage for those who have access to: Distance Primary Secondary ,- (Minutes) Potable Water Food Market School School Health Facility 0-14 65.7 64.5 65.8 61.4 62.7 15-29 54.5 64.1 62.3 64.5 64.7 Total 64.2 64.2 64.2 64.2 64.2 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

3.18 Access of the poor to electricity supplies for cooking and lighting is limited and often unreliable. About 68 percent of the households in Bissau do not have electricity and in the rest of the country, the proportion of households without electricity- is much higher at 95 percent (see Table 3.10). It is also evident that the availability of electricity in the home is a function of economic well-being as the proportion of households with electricity increases from 5 percent in the lowest quintile to 26 percent in the highest quintile. It is therefore not surprising that five out of every eight households without electricity are below the poverty line. The inadequate and irregular supply of electricity and other forms of modem energy have resulted in the dependence on inefficient traditional energy sources.41 The household survey shows that the dominant fuel for cooking or lighting in most households is firewood and charcoal or other biomass such as dung and crop waste, especially among the poor. For example, 65 percent of households that use firewood for cooking are below the poverty line whereas only 12 percent ofhouseholds using electricity or gas for cooking are poor.

40 Access to basic services is measured by the average distance (in miles) between households and basic services like potable water, food markets, schools and health facilities. 41 The majority ofpower plants and transmission facilities built during the colonial era have completely deteriorated due to little investment, poor maintenance and a civil war that left a large portion of the generation and distribution infrastructure damaged or destroyed.

- 33 - Table 3.10: Access to Energy for Cooking and Lighting in 2002 (in percentage) Rest of Ext. the Quintiles Poverty Poverty Access to Electricity Bissau Country Poorest 2 3 4 Wealthiest Yes 31.74 4.54 4.60 6.09 8.43 10.93 26.30 33.33 6.70 No 68.26 95.46 95.40 93.91 91.57 89.07 73.70 62.40 18.20 For Cooking Firewood 8.16 93.84 88.26 77.01 69.43 67.62 50.68 65.30 21.10 Charcoal 88.30 6.02 11.38 22.37 29.40 31.91 46.75 44.50 6.80 Electricity/Gas 2.13 0.09 0 0 0.27 0.23 2.28 11.60 0 Other 1.42 0.06 0.36 0.63 0.90 0.23 0.29 77.60 18.60 For Lighting Petrol/Gas oil 12.06 69.87 73.32 58.83 54.22 49.50 38.19 66.10 22.60 Firewood 73.58 13.91 13.9 27.76 33.85 34.15 40.80 51.20 8.80 SoladGenerator 13.83 2.54 1.85 2.25 3.05 5.50 13.14 29.60 3.50 Other 0.53 13.69 10.94 11.17 8.88 10.85 7.86 60.90 18.30 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

3.19 Access of the poor to modern healthcare remains very limited. The poverty. incidence is very high (about 83 percent) among women who have had a live birth in the last 12 months without prenatal care compared to those who had it (63 percent) (see Table 3.1 1). Also, the poverty rate among those who use neighborhood midwiveshurses as the main healthcare provider is highest at 88 percent, whereas only 39 percent of individuals who used private clinics lived below the poverty line. It is noteworthy that among those who use these neighborhood healthcare providers 63 percent are considered as extremely poor whereas only 3 percent of individuals who use private clinics are extremely poor (see Table 3.12). These results suggest that there is an unmet need for affordable healthcare facilities, which is probably the reason why such a disproportionately large number of poor people rely on neighborhood midwiveshurses as the primary source of healthcare.

Table 3.11: Poverty Decomposition by Recipients of Prenatal Care in 2002 (in percentage) Prenatal care Po PI p2 No 82.5 35.8 19.7 Yes 62.7 24.1 12.1

Total 63.8 24.8 12.5 m:Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

Table 3.12: Poverty by Type of Healthcare Provider Used (in percentage) Poverty Extreme Poverty Type of Healthcare Provider Po PI pz Po PI P2 Private hospital 59.2 22.0 10.5 16.3 4.0 1.4 Public hospital 51.2 16.2 7.7 12.6 3.6 1.4 Community health center 58.5 21.6 10.7 16.6 4.6 1.8 Private clinic 38.7 10.6 4.2 2.9 0.6 0.2 Traditional herbalist 66.8 25.4 10.4 11.9 0.4 0.0 Neighborhood midwife/nurse 87.5 48.4 29.1 63.3 18.6 5.9 ONG clinic 65.9 21.6 14.8 27.0 7.0 2.9 Pharmacist 63.8 27.0 14.1 15.1 6.0 2.9 Total 57.6 21.2 10.3 15.9 4.2 1.6 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

-34- 3.20 Food and auto-consumed energy account for significant shares of household spending. Across quintiles, about 40 percent of household expenditures are allocated to the purchase of food. Expenditures on auto-consumed4* food and auto-consumed energy, however, seem to be influenced by the economic status of the household. The values of auto-consumed food and auto-consumed energy decrease as the wealth status of the household increases. For example, auto-consumed energy in households in the lowest quintile is estimated at 28 percent oftotal expenditures but systematically decreases to 11 percent in the top quintile. Considering the acute energy situation in Guinea-Bissau, it is not surprising that the value of auto-consumed energy is so high in the poorest households. As discussed above, the unreliable supply of electricity requires that households look for alternative sources of energy. A small number ofhouseholds that are economically well-off purchase generators and solar panels to supplement the supply of electricity. The vast majority of households, particularly the poor ones, which cannot afford these alternative sources of energy, are forced to use firewood and charcoal, which they often collect, or prepare themselves, to satisfy their energy needs. The share of expenditures on food is higher in Bissau than in the other regions. As expected, the. regions of Oio, and BiomboBolama, which are among the poorest in the country, have the highest shares of expenditures on auto-consumed food, whereas Oio and QuinardTombali have the highest shares of auto-consumed energy (such as collection of firewood and other combustible substances).

3.21 Salaries from private and public sectors and income from the sale of agricultural products are the main sources of income. These sources of income are however sensitive to the economic status of the household as well as the region of. residency. Expectedly, the share of salaries in household income is lowest in the poorest quintile (about 19 percent) and highest in the top quintile (about 43 percent) (see Table 3.13). Conversely, revenue from the sale of agricultural products is highest in the lowest quintile and decreases as wealth increases. Overall, the share of total revenue obtained from private and public sector salaries as well as from non-agricultural products is significantly higher in Bissau than in any other region. The regions of Bafata, QuinardTombali, and Oio, derive the highest shares of income from agricultural products. Even though all these regions are rice producers, rice farming in Bafata is more modernized than the other regions. Whereas the other regions use mainly traditional and labor intensive methods, Bafata is a beneficiary of a Chinese pilot program which uses a combination of animals and tractors for rice cultivation. Hence the 52 percent share of household income from agricultural products in Bafata is mainly from the sale of rice (see Table 3.14).

3.22 The overall economic situation of households and their communities at large seem to have deteriorated compared to the year prior to the survey year. The majority ofhousehold heads, across quintiles, indicated that the economic situation ofthe household was worse than the previous year (see Figure 3.1). Similar response was obtained when the question was referred to the perceptions on the economic situation of the community.

42 Auto-consumption in this particular survey is defined as the production of an item for the consumption of the household. The intended purpose of producing the item does not necessarily have to be considered as a revenue-generatingactivity.

- 35 - Table 3.13: Revenue Shares by Economic Status in 2002 (in percentage) Sources\ Quintiles Poorest 2 3 4 Wealthiest Local Drinks 3.67 3.72 2.65 2.90 1.57 Livestock 2.88 2.04 4.12 2.25 1.34 Animal Products 0.71 0.19 0.11 0.09 0.17 Fish Products 2.57 6.84 3.38 3.31 2.73 Firewood 2.18 0.58 0.95 0.42 0.50 Salary - Agric. Sector 2.86 1.74 1.29 1.26 0.47 Salary - Private Sector 8.06 17.14 21.19 19.42 29.28 Salary - Public Sector 8.00 12.82 19.97 13.92 13.06 Rentals 13.07 7.99 9.33 9.60 7.62

Pensions 0.66 1.10 0.75 0.44 1.93 ~ Transfers 5.86 7.67 6.74 13.84 12.49 Non-Agriculture 7.31 6.09 5.49 7.89 16.12 Agricultural Products 42.18 32.07 24.04 24.67 12.72 T&l 100 100 100 100 100 m:Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

Table 3.14: Revenue Shares by Region in 2002(in percentage) Sources\Regions Bissau Bafata Gabu Cacheu Oio Biomboh3olama QuinardTombali Local Drinks 1.34 1.32 0.29 6.32 2.23 7.08 3.75 Livestock 0.08 7.04 9.93 0.91 8.68 1.62 1.81 Animal Products 0.00 0.45 1.15 0.47 0.26 0.27 0.10 Fish Products 0.61 0.61 0.18 8.86 6.32 17.59 5.72 Firewood 0.58 0.14 0.28 0.46 3.05 0.52 0.39 Salary - Agric. Sector 0.66 0.35 1.83 2.18 1.54 1.44 2.29 Salary - Private Sector 33.15 9.58 21.59 5.19 9.15 4.64 15.48 Salary - Public Sector 20.92 2.78 4.52 4.78 5.05 5.84 6.27 Rentals 6.40 6.56 3.63 26.73 4.42 13.03 7.70 Pensions 2.00 0.20 0.45 0.02 0.56 0.48 0.39 Transfers 13.03 13.81 9.69 6.85 6.94 4.93 6.80 Non-Agriculture 14.77 4.67 15.89 4.12 4.27 7.76 4.01 Agricultural Products 6.47 52.49 30.57 33.10 47.53 34.81 45.30 Total 100 100 100 100 100 100 100 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

Figure 3.1: Perceptions on the Evolution of the Economic Situation of Households by Quintile in 2002

Wealthiest

4th Quintile

3rd Quintile

2nd Quintile

Poorest

National

0 20 40 60 80 100

Worse Same fl Bener

Source: Staff estimated based on the 2002 ILAP survey.

3.23 The deterioration of the economic situation also seems to have aggravated the ability of household heads to adequately feed their families. That is, in the year prior to the survey year, a higher percentage of household heads were unable to meet the

- 36 - food needs of their households. With the exception of the wealthiest quintile, the percentage of household heads, in all other quintiles, that had difficulty, satisfying the food needs oftheir households in the previous year, was higher than the national average (see Figure 3.2). When the data is analyzed by region, it is evident that more than half of the households in regions like Bafata, Gabu and Oio, were often confronted, in the previous year, with the problem ofproviding adequate food for their households.

Figure 3.2: Problem Satisfying the Food Needs of Households by Quintile in 2002

Wealthiest

4t h 3r tn d Poorest National

0 2 4 6 8 10 0 0 0 0 0 Ofte 0 Sometimes NevedRarely n w:Staff estimated based on the 2002 ILAP survey.

3.24 Recent data shows that a third of the households have only one meal a day and just one in four households have 3 meals per day. Regarding the quality of the meals, rice is the main staple and fish is the primary source ofprotein. While fish is widely available, the meals in one out of every ten households did not have any protein source during the week before the survey (60 percent ofthe households did not include any other animal protein). Most households interviewed in 2005 (87 percent) indicate that over the last year they went without what they considered enough food to feed their families. Of those households that indicated not having enough food, half of them manifested that they experienced deprivation for at least two to three months, mostly during the dry season.

3.25 Women headship of the household and migration are believed to have contributed in improving food security. The participation of more women in a myriad ofpetty revenue generating activities -- both domestically and in the context ofmigration -- seems to provide them with the ability to better assure the food security of their households ihan iheir male counierparis who are mosiiy engaged in formai aciivities, like working for the government where the payment of salaries is highly unreliable (see Box 3.3 for a discussion on food security). Anecdotal evidence suggests that the characteristics of migration may be different in Bissau from the rest of the country. Residents of Bissau typically travel to Portugal or Brazil, where some of them angage in different types of economic activities to raise money for their families back in Guinea- Bissau. Most of the temporary migrants from rural areas, however, typically travel to neighboring countries, especially Senegal (some with agricultural products sale at relatively higher prices in Senegal) where they clean houses, wash clothes (in relatively affluent homes), work as tailors and laborers at construction sites to raise money, which

- 37 - they invest in farming, especially rice cultivation when they return to Guinea-Bissau, several months later. To further explain the economic well-being of households headed by women, it is worth noting that most ofthe people, who engage in these cross-boarder activities, are women, some heads ofhouseholds.

Box 3.3: Perceptions of Food Security: Meeting the Food Needs of the Household

A simple regression analysis, based on the 2002 household survey data, is used to shed more light on the impact of various household characteristics on meeting the food needs of the household, when controlling for other characteristics. The results, in the table below, indicate that the migration status of the head, the sector of employment and access to basic services are major determinants of the problem of meeting the food needs of a household. Gender of the Household Head: In rural areas (rest of the country), the female headship of a household diminishes the likelihood that the food needs of the household will not be met by 23 percentage points compared to households headed by men, which is the reference variable. This result is a further confirmation of earlier findings that female headed. households are economically better-off than male-headed households. Migration Status of the Head: Heads that have migrated overseas or domestically, at least once, have a lower probability of having difficulties meeting the food needs of their households. The impact of the migration status of the head is higher in urban areas (Bissau) where households are 36 percent less likely to have this problem than in rural areas (rest of the country) where households are 24 percent less likely to fail to meet the food needs of the household. Sector of Employment: The results suggest that the employment of the head in sectors like TransportKommerce, EducatiodHealth and Administration is likely to improve the chances of meeting the food needs of the household. These effects are particularly high in urban (Bissau) areas and to a lesser extent, in rural (rest of the country) for heads employed in TransportKommerce who are 73 percent and 30 percent respectively less likely to have dificulties feeding their households compared to heads working in Agriculture, paradoxically. Access to Basic Services: The proximity of households to basic services presents some interesting results. Households located 15 minutes or less to Public Transportation in urban (Bissau) and rural (rest of the country) areas are associated with a lower probability that the heads would have difficulty meeting the food needs of the household. This effect is especially strong in urban areas where the likelihood that the food needs of the household will not be met is diminished by 86 percentage points compared to a household in the same area of residency but located more than 15 minutes from public transportation. Similarly, access to secondary schools and health facilities also make it less likely for heads of households to have problems feeding their households. These effects are only observed in rural areas, however.

Impact of Selected Factors on the Problem of Meeting Food Needs in 2002 Full Sample Urban Rural Gender : Household Head is Female NS NS -0.23 8. Migration Status of Head Migrated (at least once) -0.24 -0.36 -0.20 Sector of Employment (Head) Constructionlhdustry NS NS NS TranspotVCommerce -0.42 -0.73 -0.30 EducatiodHealtWSeices -0.37 -0.64M Administrations -0.40 --0.80 NS Access to Basic Services (in minutes) Potable Water (4.5) 0.33 NS 0.52 Food Market (45) 0.42 0.43 0.52 Public Transportation (45) -0.29 -0.86 -0.28 Primary School (45) --0.25 NS NS

Secondary School (45) NS NS -0.35~ Health Facility (45) NS NS -0.32 World Bank staff estimates. NS: Not significant. Displayed mflicienu are significant with a 10% level of confidence. Undalinsd cocflicients are significant with a 5% Iwd of confidence. Bold cocfficienu are significant with a I%INSI of confidence. Omitted categwiea: Gender of Head (Male); Migration Status of Head (NN~Migrated); Sector of En@oymsnt of Head (Ag~iculhlrs);Access to Basic Services (30 minutes or me).

- 38 - 2. Determinants of Welfare

3.26 Poverty profiles are informative but they also have limits. The main drawback is that they cannot be used to assess the determinants ofpoverty. For example, the fact that some household group is poor (say, agricultural workers) may be due in large part to other characteristics of this group (say, the educational level ofthe group's members). In order to provide more insights into the determinants or correlates ofpoverty, a regression analysis is carried out in table 3.15. A regression analysis of the determinants of well- being with three dependent variables is used: the logarithm ofthe per capita consumption ofhouseholds, the logarithm of the income per capita ofhouseholds, and the logarithm of an assets index that aims to capture the wealth of household^.^^ This section discusses the findings of such analysis for Guinea-Bissau using the 2002 light household survey data. To assess the impact of various characteristics on the dependent variables of interest, the analysis relies on linear regressions. Separate regressions are provided for the full sample" and, where possible, for rural and urban samples of the data. The key determinants of households identified in this analysis reflect the information available in the survey database, and, therefore, do not pretend to represent a perfect household model for Guinea-Bissau.

3.27 In rural areas, and more so in urban areas, a unit increase in the number of infants and children is likely to cause a decrease in per capita consumption. An increase in the number of, infants, children or adults in a household is likely to cause a* decrease in per capita consumption. To some extent, this is an automatic result since the indicator ofwell-being depends directly on household size. Note that the negative impact of having more adults is not statistically significant for per capita consumption in urban areas, perhaps because of better employment opportunities there so that adult members provide through work for their consumption needs. Note also that the impacts on income per capita and on assets are much weaker. In fact, a unit increase in the number of adults is expected to have a positive effect on asset accumulation (this variable is not 2- normalized by household size).

3.28 In rural areas, female headed households tend to be better off. In female headed households in rural areas, consumption per capita and the accumulation of assets is likely to be higher than in households headed by men, all other things being equal. For example, female headed households in rural areas have on average consumption levels 20 percent higher than households headed by male counterparts. A similar finding is observed for the accumulation of assets, albeit to a smaller degree. This result is a clear indication that female headed households are not necessarily worse off than households. headed by males.

3.29 In rural areas, households headed by people who are between the ages of 15 and 35 and those who are between the ages of 35 and 55 are likely to increase the

43 The main regressors of the model include: (a) family structure variables and their square (number of infants, children, and adults), @) the characteristics of the household head (the gender of the head, the age group the head belongs to, the marital status of the head, and the migration status of the head, the head's level of education, hisher employment status, hisher sector of activity, whether he/she. works in the public or private sector); (c) the region where the household is located; and (d) household access to basic services such as potable water, schools, public transportation, food market and health facilities.

-39- asset holdings of the household by about 4 percent higher than those heads who are 55 and older. The marital status of household heads seems to have a positive impact on consumption per capita and on the accumulation of assets. In rural areas, the expected level of consumption, for household heads who are single, married (monogamous) and those who are married (polygamous) is respectively 19 percent, 15 percent and 16 percent higher than the consumption levels of households headed by those who are divorced or widowed.44 In urban areas, however, only households headed by those who are single are likely to increase consumption per capita. In fact, the magnitude of the effect is pretty strong at 35 percent, compared to heads that are divorced or widowed. The results also suggest that rural households with heads who are married (monogamous), have a slightly higher level of asset holdings than heads who live in those areas and are divorced or widowed. There is no evidence that the marital status of heads has any impact on the expected levels ofincome.

3.30 As the level of education of household heads increases, so is the impact it has' on consumption per capita, income per capita and the level of asset does accumulation. In rural areas, consumption increases by 12 percent for heads with secondary education and by 35 percent for heads with tertiary education as compared to heads with no education at all. In urban areas the effect is even stronger at 24 percent for secondary school educated heads and 53 percent for heads with tertiary education. The impact of education on the accumulation of assets is also high in both urban and rural areas for assets. In urban areas for example, asset holdings increase by 8 percent, 21 percent and 48 percent respectively for heads with primary, secondary and tertiary' education, whereas in rural areas, the gains are 4 percent, 16 percent and 43 percent respectively. The impact of education on income appears to be limited to urban areas where incomes for households with heads with tertiary education are 70 percent higher than incomes for heads with no education at all.

3.31 For household heads living in rural areas and employed in sectors like Transport/Commerce and EducationDIealth, income per capita is likely to increase by 47 percent compared to heads working in the agricultural sector. This finding- confirms the presence oflow productivity in agriculture relative to the other sectors ofthe economy. In urban areas, however, the only sector which seems to have a positive impact on income is TransportKommerce, which increases income by 46 percent. Rural household heads who are employed in Transport/Commerce and in the fields of EducatiodHealth are likely to increase the asset holdings of their households by 24 percent and 15 percent respectively. It is worth noting that in these areas, working in the Industry and Construction sectors can also enhance the accumulation of assets, to a much smaller degree. Contrary to our expectation, however, there is likely to be a 17 percent reduction in consumption per capita for households whose heads work in the healtWEducation sector in rural areas. This phenomenon may be due to the effect of frequent salary payment delays and arrears in these sectors as experienced in the past, which may adversely affect consumption patterns. Urban household heads, who work in

44 In the case of Guinea-Bissau, as women tend to be better off compared to men, polygamous household are likely to bear higher income and consumption levels and therefore less poverty incidence.

- 40 - the private sector, are likely to increase income per capita by 42 percent compared to their counterparts who work in g~vernment.~’

3.32 Proximity of households to infrastructure positively affects their welfare situation. Another set of variables in the regression which clearly matters for standards of living relates to geographic location. Living nearby a school, a health facility, and means of public transportation tend to be positively associated to consumption, income and assets. By contrast, those households who live near food markets seem to be poorer,- because these tend to take place in poorer neighborhoods. What is more important is the fact that controlling for all the characteristics discussed above, households who live in Bissau, Bafata, Gabu, Cacheu, Biombo/Bolama, and QuinardTombali are better of than households living in the reference category of Oio, the poorest region in the country. For example, living in Bissau is associated with a 79 percent increase in income’compared to residing in Oio. More generally, the gains observed for Bissau explain much of the migration patterns observed in the country, and the fact that Bissau itself has relatively high levels ofpoverty probably in part because ofmigration.

3.33 The availability of public transport within 15 minutes of households seems to have the most positive impact on consumption and the accumulation of assets especially in urban areas. For example, the results suggest a 39 percent increase in the assets for urban households with easy access to public transportation compared to households located in the same areas but not as close to public transportation. Primary schools situated close to households in urban areas are associated with a 72 percent increase in income but a 22 percent decrease in assets. Secondary schools located 15 minutes from households appear to have an incremental effect only on the accumulation of assets in rural areas. Accessibility to health facilities is likely to increase consumption in both rural and urban areas as well as have a positive effect on income per capita in rural areas.

3.34 Rural households whose heads have migrated (at least for a month in the last 12 months) have experienced a small increase in the level of asset accumulation compared to households whose heads have never migrated. Temporary migration- serves an important livelihood purpose in rural areas. In addition, those with kin or family in urban areas with sufficient means often send their children to live with their relatives in the towns and cities to enable them to continue with their studies. Some ethnic groups also engage in seasonal migration from one rural area to another, either in pursuit of wage-labor, for trading and commerce purposes, to return to their “native lands” to engage in rice production, or to lead their livestock through transhumance towards more fertile lands. Others have emigrated from the Casamance region of Senegal and have settled indefinitely in the northern areas of Guinea-Bissau. Others again,’ particularly from the northern and eastern areas of Guinea-Bissau have permanently emigrated into neighboring countries or to Europe (primarily Portugal and France).

45 The results do not provide any evidence of the impact of the type of employment on any of the two other dependent variables.

-41 - Table 3.15: Impact of Various Characteristics on Household Welfare in 2002 Consumption Income Wealth (per capita) (per capita) (level of assets) Full Full Full Sample Urban Rural Sample Urban Rural Sample Urban Rural Family Structure Numbkr of infants (under- 5s) in household -0.12 -0.18 -0.08 -0.13 N.S. -0.15 -0.024 -0.06 N.S Square of number of infants in household 0.01 0.02 NS NS NS 0.025 NS NS NS Number of children (aged 5- 14) in household -0.13 -0.20 -0.10 -0.19 -0.19 -0.19 NS NS NS Square of number of children in household 0.01 0.02 0.01 0.02 NS 0.02 NS NS NS Number of adults (aged 15- 64) in household -0.12 NS -0.16 NS NS NS 0.024 0.05 NS Square of number of adults in household 0.01 NS 0.01 NS NS NS NS NS NS Age of Household Head Age group of Head (<35) NS NS 0.08 NS NS NS -0.04 NS 0.05 Age group of Head (>=35 -- <55) NS NS NS NS NS NS -0.04 NS 0.05 Gender of Head Household Head is Female 0.13 NS 0.19 NS NS NS -.os NS 0.07 Marital Status of Head Single 0.26 0.35 -0.19 NS NS NS NS NS NS Married (Monogamous) 0.01 NS 0.15 NS NS NS NS NS -0.07 Married (polygamous) -0.13 NS 0.16 NS NS NS NS NS NS Migration Status of Head Migrated (at least once) 0.10 0.16 0.10 NS NS NS 0.05 NS 0.05 Education of the Head Primary Education NS NS NS NS NS NS 0.06 0.08 0.04 Secondary Education 0.17 0.24 -0.12 NS NS NS 0.19 0.21 0.16 Tertiary Education 0.51 0.53 0.35 0.55 0.70 NS 0.48 0.48 0.43 Sector of Employment (Head) Constructionhdustry NS NS NS NS NS NS 0.09 NS 0.09 TransportKommerce NS NS NS 0.60 -0.46 0.47 0.21 -0.15 0.24 Education/Health/Services NS NS --0.17 0.54 NS -0.48 0.17 NS 0.17 Administration NS NS NS 0.32 NS NS 0.10 NS NS Type of Employment (Head) Private Sector NS NS NS NS 0.42 NS NS NS NS Business Owner NS NS NS NS NS NS NS NS NS Access to Basic Services (in minutes) Potable Water (45) -0.16 NS NS NS NS Ns NS NS 0.10 Food Market (45) -0.1 1 NS NS -0.51 --0.50 -0.50 -0.05 NS NS Pub. Transportation (45) 0.12 NS -0.12 NS NS NS 0.06 0.39 -0.06 Primary School (45) NS NS NS -0.25 -0.72 NS NS -0.22 0.05 Secondary School (<15) NS NS NS NS NS NS 0.09 NS 0.15 Health Facility (45) 0.16 0.25 -0.12 NS NS 0.24 NS NS NS Regional Dummies Bissau 0.35 0.79 0.43 Bafata 0.24 NS 0.13 Gabu 0.27 --0.34 0.13 Cacheu 0.15 0.29 0.05 BiomboBolama 0.32 0.40 NS QuinardTombali 0.18 NS -0.06 m:World Bank staff estimates. P Not significant. Displayed coefficients ar gnificant with a 10 percent level ofconfidc :e. Underlined coefficients are significant with a 5 percent level of cc dence. Bold coefficients are significant H a 1 percent level of confidence. Omitted categories: Age Group of Ho hold Head (55t); Gender of Head (Male) arital Status of Head (DivorcedlSeparatedidowed); Migration Status of Head (Never Migrated); Education 01 :ad (Never Attended); Sector ofEmployment of Head (Agriculture); Type ofEmployment (Government Employee); Access to Basic Services (30 minutes or more); Regional Dummies (Oio).

- 42 - c. LIVELIHOODSOF THE POOR

1. Livelihoods of the Urban Poor46

3.35 While the poverty rate on an aggregate level tends to be less in urban areas, increasingly in Guinea-Bissau, poverty and vulnerability is becoming an urban issue, with many urban households, for parts of the year, living on only one meal each day - a phenomenon referred to in Guinea-Bissau as urn tiro - the “one shot”. Moreover, there are indications that many urban dwellers are more vulnerable to particular shocks than rural populations. Even though their financial incomes tend to be higher, many lack assets (including access to kinship and social networks, access to land) which constrains’ their ability to cope with unexpected stress and uncertainty. There are multiple causes contributing to this situation, including high levels of unemployment, low wages, non- payment of salaries, poor access to basic social services and infrastructures, etc. Independently ofthe place ofresidence (rural, urban or peri-urban), most families cannot survive with barely one income source, making the most common feature of livelihoods in Guinea-Bissau the diversification of strategies to make ends meet. A string of activities by various household members is thus required. The survival of families is guaranteed by a combination of several income generating activities (formal and informal).

3.36 The informal sector plays a significant role in livelihood strategies of urban dwellers. As has been described in the context of the waning influence of the state, particularly in rural areas, Guinea-Bissau has been experiencing an extensive process of informalization, particularly when it comes to the livelihood strategies available to the majority of the population. Informalization of livelihoods, while characteristic across Guinea-Bissau, is a particular feature of the livelihoods of the urban poor. More than 90‘ percent of the urban families recently surveyed in urban areas, lives essentially of activities in the informal sector ( ‘ffazer bida ”, making a living in Creole-Portuguese), or of a mix of informal and formal (mainly public sector) activities. Trade and commerce, most of which is informal, play a dominant role as livelihoods strategies of urban households. Women have historically played an important role as traders in Guinea- Bissau. These so-called bideiras, many of whom come from the coastal ethnic groups (Manjaco, Mancanha, Pepel, etc.), mainly trade in foodstuffs, such as locally produced vegetables, fruits and alcoholic drinks (such as cashew wine), roasted groundnuts, and- fish. Many also cultivate small horticultural plots for consumption and income. In Bissau, where the land traditionally belongs to the Pepel ethnic group, this often occurs on land that is either borrowed or rented during the dry-season when the land is not being used for rice cultivation. As many of the local products are seasonal, however, sellers often need to complement these activities by engaging in other strategies when these products are out of season, such as the collection and sale of firewood and charcoal. Many women also engage in the provision of personal services, including doing domestic work for others, dressmaking, weaving (etc.).

46 In addition to drawing on the recent MEP/JBN/World Bank (2005) survey, this section also relies on the recent survey of informal livelihood strategies provided by Louren$o-LindelI(2002), as well as diagnostic reports on informal livelihoods published by MEP (Paulo and Jao, 1996; Cardoso and Imbali, 1996; Monteiro, 2001).

- 43 - 3.. 3.37 Informality is also a common feature of the labor market. During periods of harvests, some people, particularly younger women and youth, seasonally migrate towards the rural areas of the country to offer their labor to farmers, particularly cashew producers (an activity referred to locally as pirmi). During this period, they set up- temporary residence in the interior of the country, and live and work with relatives or friends. Others simply work alone, offering their services to small cashew producers. Remuneration is either in cash or in-kind, such as cashew, which is then frequently traded for rice for consumption. Many of the women also extract cashew juice which is later transported to the cities and sold as wine, providing a major source of income during the cashew season (April to July).

3.38 Women more than men often rely on informal activities as major livelihood- strategy. Women’s role as traders goes back to pre-colonial times. Particularly for Bissau, women have been pivotal in ensuring the regular supply of fresh foods. Small garden agriculture, essentially practiced by women, is a very important activity at the urban level (existing estimated by UNDP suggest that more than 2 000 plots were identified in the green belt of the city of Bissau). The principal products are vegetables for internal consumption that are sold in the local markets. Among urban households, some 30 percent practice small garden agriculture, which during the dry season covers 70 to 80 percent of the demand for fresh vegetables; while in the rainy season would reach up to 60 percent.

3.39 The participation of women in market activities has particularly intensified following the process of market liberalization that was initiated in the mid-1980s. According to some studies of informal activities in Bissau, this may have resulted from an attempt to compensate for households’ declining incomes throughout this period as well as due to women’s increasing expenses as the state retreats from the delivery of social services.47 As came out of the urban focus groups, women, have in some cases become the main breadwinners in the household - which according to some men made women the de facto heads ofthe hou~ehold.~~This is also to some extent born out in the findings from the 2002 household data which suggest that female heads of household tend to be better off than male heads of household. According to INEP, this is increasingly leading to a profound transformation of the notion of the ‘family’ in particularly urban areas of Guinea-Bissau. This “inversion” of gender roles, they suggest, is on the one hand leading to increasing independence and autonomy for some women within the household, but which on the other hand has led to situations of instability. within the family, sometimes resulting in separation and divorce, putting further strains on the capability of households to cope.49 Nevertheless, while potentially better off than men, women are often more vulnerable given that they are the ones that are responsible for most family expenses, such as clothing, education, and health expenses.

47 Gomes and Duarte, 1996: 98, 100; Lourenqo-Lindell, 2002: 77.

48 MEP, 2005: 17.

49 MEP, 2005: 28. This latter point, moreover, corresponds with the findings from the 2002 Light Household Survey that suggested that divorced or widowed heads of household in Guinea-Bissau are generally worse off.

- 44 - 3.40 Salary earning activities represent men’s main livelihood mean. For urban men, casual work (surne) is increasingly the main activity and livelihood strategy as formal wage opportunities become limited. Most urban men indicate that they are “on call”, waiting for temporary jobs - typically in construction (incl. bricklaying), loading work (often at the dock), and transportation (taxis, private buses) - to appear on the market. While there particular activities pursued depend on the location (neighborhood),. ethnic group and skills of the individuals, most of the men end up unemployed and peddling in the streets. The situation has worsened over the last decade, as the number of informal wage-seekers and operators in urban areas has been swelled by an exodus of youth from rural areas in search ofa better life. Even children are increasingly turning to the informal sector to assist households in making ends meet.50 Some, though mainly men, have jobs in the formal sector mostly as teachers or health workers and other low- wage public sector employees, including employment by the military or police. A majority of these, however, aim to complement low and irregular payment of salaries by- also turning to the informal sector, and often juggle multiple activities to ensure their livelihoods.

3.41 In addition to the female traders, the bideiras, referred to above, some men, particularly those drawn from the Eastern ethnic groups (Fula, Mandinga, etc.) also play a historically important role as traders, reflecting long indigenous trade traditions in Guinea-Bissau and the broader ~ub-region.~~These, who go by the name of djilas, are mobile traders who tend to concentrate in the trade of imported foodstuffs and rice, and who in recent years have become dominant in Bissau. In urban settings, these men have recently also been joined by women also drawn from the Eastern groups. Together with the bideiras, the djilas are part of“a marketing system [,. .] in which traders are linked by shifting and flexible ties into a decentralized and unbounded network that connects markets located in different areas.”52 These markets include lumos, so-called weekly and rotating markets spread around the country. Some of these traders may be considered “formal” in that they are registered and hold trade licenses, but due to their constant geographical mobility and the relatively small size of their enterprises, most of them easily escape registration and the various taxes and regulations ofthe state.53 While some ofthese traders have been able to build prosperous trade careers, often through engaging in vertical business networks headed by import/export firms, a large share of those households that rely on trading and commerce as a primary livelihood strategy seem confined today in low-income activities. Many indeed, live from hand to mouth, and are barely surviving.

50 Cardoso and Imbali (1996: 229) have also documented an increasing level of street children recently in Bissau. In recent years, particularly in the years after the war, there has also been an increasingly visible presence in Bissau (especially in the large Bandim market) of merchants and traders from neighboring countries and others in the sub-region (particularly, Senegal, Mali, Guinea-Conakry, Mauritania, and Nigeria), though their presence is also noted in other regions primarily along the ancient trade routes through the Eastem provinces of the country.

52 Lourengo-Lindell, 2002: 76.

53 While some of these do pay a daily fee to the municipal council, and are therefore supposedly “registered” and, hence, “formal” traders, according to a survey conducted by Lourenqo-Lindell (2002: 74) most of the ones that rely on these strategies for their livelihoods do not. In any case, for all intents and purposes, even those that do pay the daily fee were involved in activities that could otherwise be described as “informal” - activities where participants lacked a written work contract, if employed, and which offered highly irregular incomes.

- 45 - 3.42 In peri-urban areas outside of the Bissau metropolitan area, as well as in primary and secondary towns in the regions (where the regional and sector state administrations are based), livelihoods are hrther intertwined with rural / agricultural activities. Indeed, in the peri-urban town of Bubaque, on the island of the same name, most households engage in fishing and exploitation of forest resources (palm oil and wine). Public employment is scarce, and only a small minority engages in wage-labor in the small tourism industry on the island. Hence, the main difference between the livelihood strategies of poor households in peri-urban areas with rural households would be that some of these households are getting part of their income from household members working in the formal wage sector. Nevertheless, the overwhelming majority ofthe peri- urban population lives off agriculture as in any Tabanca in the rural areas. What characterizes both those households that live primarily off agricultural activities and those that rely on income flows from public work and/or services, is that neither make enough from any one activity to guarantee their survival. Therefore, most households are forced to pursue several income and consumption streams from several different household members to ensure their livelihoods throughout the year.

2. Livelihoods of the Rural Poor

3.43 In rural areas, small traditional producers who rely mainly on subsistence farming make up the majority of the rural population. Only 9 percent of the national land, or 27 percent of arable land, is occupied by larger farmers (ponteiro~).~~Approximately 90 percent of the production of the estimated 90,000 smallholder households goes towards internal consumption; small farmers also cultivate as much as 75-85 percent of the country’s annual cashew production.

3.44 Subsistence agriculture represents the main livelihood strategy in rural areas. What characterizes rural livelihoods, thus, is their dependence on subsistence agriculture. Rice remains the most widely grown staple crop in the country, although in the East and amongst the Fula and Mandinga corn is the preferred staple crop. Specialization in any one activity is minimal, however, and diversification ofstrategies is the norm - dictated by the necessity to reduce vulnerabilities and minimize risks in order to ensure food and livelihoods security. Cultivation of different products with different cycles and demands is considered an important way of diversifying livelihoods. Cashew (including production and sale of cashew wine) has played a particularly important role for many households, providing an important source ofincome - as well as consumption, if traded for rice - during parts of the dry season (April until July). Other important sources of both consumption and income come from the cultivation of peanuts, cassava, and beans, etc. Horticulture (tomatoes, onions, etc) and fruticulture (mango, bananas in particular), and exploration of forest products (wood, charcoal, straw, extraction ofpalm- oil and wine, etc.) are also pursued, partly as a complement to ricekorn for consumption purposes, but increasingly as a source of income for households. Some also breed chicken and hens, goats and other small animals as risk insurance and to ensure food security. While fishing has traditionally been considered a marginal activity in Guinea-

54 Large farmers @onteiros) with land right concessions assigned by the government possess about 2,200 properties, of which 1,200 are currently not under regular production. An average property is 136 hectares, ranging from 20 to 3,000 hectares.

- 46 - Bissau, during times of poor harvests, or during the regular hunger season, fishing also serves as an important coping strategy for those communities based along the coast and many rivers.55 As noted above for urban areas, migration also serves an important livelihood purpose in rural areas. Those with kin or family in urban areas with sufficient means, often send their children to live with their relatives in the towns and cities to enable them to continue with their studies. Some ethnic groups also engage in seasonal migration from one rural area to another, either in pursuit ofwage-labor, for trading and commerce purposes, to return to their “native lands” to engage in rice production, or to. lead their livestock through transhumance towards more fertile lands.

3.45 In the Northern region, cashew farming and small trading are the main livelihood strategies. Livelihood strategies of the rural poor tend to differ across regions and are also influenced by the social characteristics of households and individual household members (such as ethnicity, family structure, gender, religion, age, etc.). For example, in the sector of Cai6, in Northern Guinea Bissau, the main ethnic group is the Manjacos who mainly live off agriculture, particularly cashew production. In addition to. the Manjacos, the sector is also populated by migrating Fulas, Mandingas, Pepel, and Mauritanians. Among the male Fulas and the Mauritanians the principal activity is small trade and commerce, while the Pepel that mainly occupy the coastal areas primarily engage in seasonal fishing during the dry season and many return to the region of Biombo during the rainy season to engage in agricultural activities. In terms of the gendered division of labor, men tend to be more involved in the clearing and planting of agricultural lands, while women - in addition to tending to the domestic chores - are more involved in the transportation and sales of the crops. Women from the non-Islamic- ethnic groups, much like in urban areas, are also more actively engaged in the sale and resale of foodstuffs.

3.46 In the Eastern region, rice and cattle production are the main livelihood means. The East, where the Fula and Mandinga ethnic groups make up the majority of the population, the men tend to work on corn production in the highlands, while women tend to cultivate rice in the lowlands (bolanhas). While cotton was for many years the main cash-crop in the East, supported by large investments in the area by the para-statal. SAGB (the Cotton Society of Guinea-Bissau), in recent years cotton plantations have been abandoned also here in favor of cashew c~ltivation.~~Households in the East also rely on livestock breeding (mainly cattle), though this is mainly used to ensure livelihood security rather than for purposes of consumption. Livestock serve as a particularly important resource ofcollateral, which is used to invest in agricultural production, and in some cases also for the emigration of household members. For those households that engage in the latter, remittances also become an important source of income. The East, moreover, has for centuries been the center of the main trading route in Guinea-Bissau. The lumos there, especially in the city of Gab6, attract mobile trades from all of the

55 Despite the vast fishery potential in Guinea-Bissau, only 10% of fish caught annually is by domestic artisanal fishers. Indeed, fishing in Guinea-Bissau has traditionally been a marginal activity, mainly to complement agricultural activities. For those households where fishing is one of the livelihood strategies, fishing is on average practiced some 4-6 days a month, over 8-10 months, for a total of less than 50 days a year. According to the FA0 Fishery Country Profile (2001), in 1996 fishing was the main activity only of approximately 18% of those that engage in fishing, while agriculture was the main activity for 78%. 56 According to the focus groups, most of the small farmers abandoned cotton farming after the civil war, in 1999.

- 47 - countries in the sub-region (Guinea-Bissau, Senegal, Gambia, and Guinea-Conakry, amongst others). The Zumos, however, are most active during the dry season.

3.47 The East, and to some extent parts of the North, suffer to some extent from drought during the dry season, however, with important negative consequences in terms of food security in these regions. In addition to the impacts on food security, this has also led to increasing tensions and at times local conflicts over access to water and naturally irrigated lands, particularly between cattle-breeders and agriculturalists. In search for watering holes for their cattle, many livestock breeders engage in transhumance towards across the Eastern regions towards the North during the worst parts ofthe drought, and in the process often trespass on the lands of agriculturalist^.^^

3.48 In the Southern region, fishing and rice farming are the main livelihood means. Households in the fertile South, despite the many problems in terms of access to services and infrastructure described earlier, are well endowed with natural resources, including fish, and therefore in most cases manage to avoid the food security dilemmas of the East. Indeed, the South is where a majority of the locally produced rice and food in Guinea-Bissau is cultivated. Low-land (bolanha) rice cultivation in paddy fields is the primary activity for most households, followed by the production of hits - pineapple, banana, oranges, etc - as well as exploration of forest products (extraction of palm oil and wine) and horticulture. Fishing, while rarely a primary activity for households, is a strategy often used to supplement consumption and incomes, which becomes particularly important in times of poor rice harvests and during the dry season. Women typically engage in the sale of fish and production of wine from honey. Amongst the Balante, women also typically engage in the production and sale of congutu, a snack made from rice and peanuts. Given the poor road and transport conditions in the South, however, households and communities have a hard time marketing and selling their produce.' Indeed, as a result of the relative isolation of many communities, the local economy is mainly based on bartering and exchange, with rice serving as a common currency. Unable to access markets, the communities are rarely compensated well for their produce. The constraints in accessing markets, and transforming and conserving their products, lead each year lead to the rotting and wastage of tons ofproducts, particularly fruits and horticultural produce. Due to the geographical isolation, especially ofthese communities, many of the households aim for agricultural self-sufficiency, But lacking electricity in most Tabancas, transformation and conservation of the produce is scarce, making it' difficult for most households to engage in much consumption-smoothing behavior.

3.49 In the islands, fishing, extraction, craft work and tourism represent the main livelihood strategies. In the islands, unlike on the mainland, fishing serves as a primary livelihood strategy. Nevertheless, households still pursue a mix of livelihood strategies, combining fishing with agricultural activities and the exploration of forest products. Fishing, extraction ofpalm oil and wine, and clearing and planting agricultural fields, is generally the main activities pursued by men. Women sell the fish, and produce- horticultural products for consumption and cash. Young males also take part in fishing and extraction, as well as the sculpting ofwood statues etc for the tourism industry, but in

57 In one of the villages surveyed, Mambonco, in the region of Oio, respondents referred frequent local conflicts between the Mandinga agriculturalists and the Fula cattle-breeders in the area.

-48- peri-urban areas may also occasionally provide casual labor for tourists and hotels in the area. For the most part, the tourism industry and commerce on the islands are dominated by foreigners, the former by the French, and the latter by mobile traders from neighboring countries. The latter also work in the artisanal fishing industry.

3.50 The main constraints to the different livelihood strategies were - in addition to the low level of access to basic services mentioned in a previous section - the difficulty in marketing any ofthe goods, and the lack oftransformation and conservation facilities for their products. As a result, people have few incentives to work beyond what they need to feed themselves. Across the country, households face a range of constraints in improving their livelihoods. Agricultural activities are mainly non-mechanized, relying in most areas on manual labor only.58Many households, furthermore, lack basic tools and implements, such as plows and rice husking tools, etc. Access to markets, credit and inputs (such as seeds, fertilizers) is low across the country, something which is exacerbated by the devastated transport and energy infrastructure, the limited activities of the private sector particularly in rural and remote areas, as well as a retreating and reduced state with limited capacity and resources to provide agricultural extension services to the rural poor.59 The lack of certainty, or “guarantees”, that results from this, leads to low and suboptimal investments, in addition, of course, to subsistence level harvests and‘ consumption. While a lot of these constraints and obstacles could be overcome with a minimal level of investments, a lot of these investments presume a present and capable state, as well as a functioning and vibrant private sector. As discussed earlier, however, the general vacuum left by the state, particularly in the regions, has left the provision of most basic services to traditional and informal authorities and organizations.

D. CONFLICT AND POVERTY

3.51 This section seeks to analyze the impact of conflict on poverty, and the main sources of conflict in Guinea-Bissau (the methodology used in the analysis is further explained in the third chapter ofVolume 2).60

3.52 Evidence suggests that Guinea-Bissau’s economy would have performed better if there had been no conflict. The estimation results indicate that the country’s GDP per capita would have been more than 40 percent higher in 2004 than it actually was if there had been no conflict (see Figure 3.3). As a point ofcomparison, in Rwanda where- the human, social and economic costs ofthe genocide have been staggering, the estimates suggest that per capita GDP today would be between 25 and 30 percent higher if the genocide had not taken placeY6lwhich suggests a smaller economic impact than in Guinea-Bissau. Much of the difference between the impacts observed in Guinea-Bissau and Rwanda stems from the fact that in Rwanda, two outliers were identified. First, in

58 In the East, the use of animals for traction and plowing is more common than in other parts ofthe country

59 After independence, the Departamento de Pesquisa Agraria (DEPA) served as an agricultural research and extension agency with a - relatively significant presence in the regions. Over time, and particularly following market liberalization in the mid to late 1980s, DEPA reduced its field presence and shifted its emphasis towards agricultural research. DEPA was ultimately superseded by IMPA (???) which focuses exclusively on research.

6o Also see Lopez and Wodon (1995) who conducted similar work for Rwanda. 61 Lopez and Wodon (2005).

- 49 - 1994, the year of the Genocide, there was a loss of GDP of about 40 percent. Yet the following year, there was a rebound (another additive, but this time positive outlier) so that part of the negative impact of the conflict was offset. Such a rebound was not observed in Guinea-Bissau, leading to a larger economic impact of the conflict, even though the loss in human life was substantially less severe. It is also important to note that in Guinea-Bissau as in Rwanda, losses in GDP were observed, but the conflicts did not seem to have a negative impact on the growth rate of GDP per capita itself. These results could be interpreted as indication that these conflicts generated a reduction in the stock values of key economic variables such as human and physical capital, without changing the returns to these stocks.

Figure 3.3: Impact of the 1998 Conflict on Per Capita GDP

140 Counterfactual

130 ~

a3 120 a*" 110 8 100 90 ga

80 I b....

60

3.53 Similarly, the results indicate that the poverty incidence would have been lower if the conflict had not occurred. Today, as shown in Table 3.15 below, and following the methodology used in the PRSP for poverty measurement, about two-thirds of the population is poor, i.e. with a level of consumption per equivalent adult below the US$ 2 poverty line in purchasing power parity terms, with 21.6 percent being extreme poor. If the conflict had not taken place -- so that per capita GDP today would be 42.3 percent higher than actually observed -- poverty would be much lower (the results not provided here are similar after correcting for terms of trade). The results suggest that the share of the population in poverty today could be around 43.0 percent, instead of 65.7 percent, while the share of the population in extreme poverty could be at 7.6 percent, instead of21.6 percent. Note that the percentage reduction in poverty following the 42.3 percent reduction in GDP per capita suggest that the elasticity of poverty reduction to growth is roughly around one (it is lower for some poverty measures, and higher for others), which is a reasonable estimate according to international experience. The corresponding elasticities for the extreme poverty measures are larger, as expected (it can

- 50 - be shown that these elasticities are a function ofthe poverty lines - the higher the poverty line, the lower the elasticity). Overall, one in three persons is in poverty today in Guinea- Bissau due to the conflict, with an even higher share for extreme poverty measures.

Table 3.16: Impact of the Conflict on Poverty Poverty Poverty Poverty reduction Poverty reduction observed without the in percentage points in percent versus in 2002 survey conflict (2)-(1) baseline (1) (2) [(2)-(l)]/(l) Poverty Headcount index 65.7 43.0 -22.7 -34.6% Poverty gap 25.7 13.2 -12.5 -48.6% Squared poverty gap 12.9 5.7 -7.2 -55.8% Extreme poverty Headcount index Poverty gap 5.5 1.8 -3.7 -67.3% Squared poverty gap 2.2 0.7 -1.5 -68.2%

3.54 Yet, security remains a major concern mainly among the poor. Only half of the households consider that the security situation has improved over the last year (see Table 3.19). The share ofthe population that believes that the situation has deteriorated is higher among the bottom tercile. Data on the sources oflocal conflicts are also revealing.

3.55 Land also remains a major source of conflict. Apart from internal family issues, most local conflict situations seem to arise from issues related to land, water, and livestock, especially in rural areas. There is evidence that poor households tend to rely mostly on traditional authorities in order to solve conflicts, while the better off rely more on the police.

3.56 Under the current circumstances, it would be difficult for Guinea-Bissau to- achieve the Millennium Development Goal (MDG) for poverty reduction by 2015. In the projections in Table 3.17, it assumed that future poverty is a function of the growth in per capita GDP.62 The results provide the number of years that would be required with various annual GDP growth rates to achieve cumulative per capita growth rates of respectively 25 percent, 50 percent, 75 percent, and 100 percent. For simplicity, a single population growth rate was used of2.5 percent per year, which is below the current level, but takes into account the fact that population growth will decline over time (clearly, in the long-run, for some of the large number of years required in the table, the rate of population growth would decrease much more, but the exercise is provided just for giving an idea of what could be achieved in the next 10 years or so, by the 2015 deadline for reaching the MDGs in 2015). The results suggest that if the economy were to grow at 4 percent per year, it would take 11 years to achieve a cumulative growth rate of 25 percent, which would in turn reduce the share of the population in poverty from the estimated level of 72.3 percent in 2005 to 58.8 percent by 2020. To give another example, assuming an annual growth rate of 8 percent per year of the economy, the

62 . First, it is assumed that growth in consumption per adult equivalent at the household level as measured in household surveys follows closely GDP per capita growth, as measured in the National Accounts, which is distributional neutral.

- 51 - cumulative growth rate would reach close to 75 percent by about 2015 (as shown in the table, the number of years required is 9.6). This would in turn lead to a share of the population in poverty of about 36.4 percent. While these simulations should be considered with caution, due to the underlying strong assumptions, they do provide an idea of the magnitude of the challenges ahead. The message from the table is somewhat disheartening, as even with much higher growth rates than were observed in the recent past, Guinea-Bissau is not likely to reduce its level of poverty well below the level that would prevail today if the conflict had not taken place (this level is estimated at 43.0 percent for the headcount index (see Table 3.16). Based on these results, it is safe to argue that if the conflict had not occurred, the country would have made significant progress toward achieving the MDG target for poverty reduction by 2005 (see Table 3.18). The Government would therefore need to diversify the sources of growth and implement policies designed to make public spending more pro-poor, in the context of the PRSP Iorder to alleviate poverty.

Share of the Number of years required to reach target population 4% 5% 6% 7%grOwth 8% in poverty (%) growth growth growth growth Baseline estimate in 2005 72.3 Per capita GDP growth 25% 58.8 11.3 7.5 5.7 4.6 3.8 Per capita GDP growth 50% 47.5 20.5 13.7 10.3 8.3 7.0 Per capita GDP growth 75% 36.4 28.3 18.9 14.3 11.5 9.6 Doubling of per capita GDP 27.9 35.0 23.4 17.7 14.2 11.9 -

MDG Target 2015 Current (PRSP if different) MDG 1: Eradicate extreme Poverty and hunger Population living below the poverty line of 2$/day (2002) 65.7% ’* 40%” Prevalence of underweight children 25% MDG 2: Achieve universal primary education Net enrollment ratio in primary education (2001/02) 80% (gross) 100% , , Primary completion rate (2002) 30% 100% MDG 3: Promote gender equality and empower women Ratio of girls to boys in primary school (2002) 41% Ratio of literate female (F) to male (M) ages 15-24 (2002) 81%(F) 41%(M) MDG 4: Reduce child mortality Under 5 mortality (per 1000) 214 70 Infant mortality (per 1000 live births) 130 48 Immunization (children under 12 months), including measles 67%

MDG 5: Improve maternal health , Maternal mortality (per 100,000 births) 349 175 MDG 6: Combat HN/AIDS, malaria and other diseases Prevalence of HIV/AIDS 3 MDG 7: Ensure environmental sustainability and access to water and 56% 100% sanitation: Access to improved water Source: World Bank Gfthe date indicated in parentheses in the first column. ”Poverty rate computed at the individual level (the household level poverty is 58.9 percent) ’’ Set as half ofthe poverty rate in 1991.

- 52 - Table 3.19: Perceptions on the Security Situations and Sources of Conflict in 2005

Tercile Area Total 1 2 3 Urban Rural ge in security situation Improvement 45. 42.2 57.75 54.82 44.66 48.6 No difference 9.72 11.97 11.27 12.65 9.92 10.98 Deterioration 42.36 44.37 30.28 31.33 43.89 39.02 No data 2.08 1.41 0.70 1.20 1.53 1.40 Total 100 100 100 100 100 IO0 Y Sources of conflict (%) Land 14.60 21.80 17.42 18.95 17.27 17.91 Water 14.60 13.53 16.67 9.15 18.47 14.93 Family 20.44 12.03 32.58 31.37 15.66 21.64 Livestock 34.31 27.07 17.42 18.95 30.92 26.37 Fishing 0 1.so 2.27 1.96 0 80 1.24 Extraction activities 0 0 0.76 0.65 0 0.25 Participation to community live 1.46 1so 7.58 8.50 0.40 3.48 Religion 0.73 Iso 0 0.65 0.80 0.75 Wealth 0.73 1so 0 1.31 0.40 0.75 Others 13.14 19.55 5.3 1 8.49 15.26 12.69 Total 100 100 100 100 100 Whom do you turn to in case of conflict? Police 23.57 15.33 64.54 64.24 15.42 34.69 Traditional authority 56.43 52.55 17.73 20.00 56.52 42.1 I Police &traditional authority 9.29 3.65 5.67 3.64 7.91 6 22 Family 5.00 16.79 4 26 3.64 11.86 8.61 Others 5.71 11.68 7.80 8.48 8.29 8.37 Total 100 100 100 100 100 100 m:Staff Estimation based on the 2005 QS.

E. CONCLUSION AND POLICY CONSIDERATIONS

3.57 Guinea-Bissau is a country with very limited poverty data. With the data currently available it is almost impossible to analyze in detail the factors that cause poverty to change over time (growth versus distributional effects), andor to carry out poverty impact analyses in key sectors such as cashew, rice, etc. Currently the authorities’ PRSP is being finalized. As the implementation, monitoring and evaluation of the PRSP will start soon, it is important that such analyses are carried out in order to make the PRSP a more result-oriented framework. In this regard, it is critical to develop a coherent and reliable poverty database over time. One way to achieving this objective is to carry out a new household survey similar to the 1991 ICOF survey, so as to allow comparability and poverty dynamics analysis. Moreover, the scope ofthe next household surveys should be broadened to include: (i)detailed information about households’ income and expenditure- and price information of key commodities such as cashew nuts and rice production etc.; (ii)education and health service delivery; and (iii)conflict and land issues.

- 53 - 4. MACROECONOMIC AND FISCAL POLICIES FOR PROMOTING SOCIOPOLITICAL STABILITY AND ECONOMIC GROWTH

Prior to the 1998 military conflict, Guinea-Bissau’s economy performed in a satisfactory manner thanks to the broad economic liberalization carried out by the authorities in the early nineties. The conflict shattered the country’s socioeconomic fabric and caused a massive destruction to private capital. Following the conflict, lingering political instability did not allow economic recovery to materialize. It is imperative that the new administration ensures the continuity of institutions and commits to the reform program put in place since end-2003 by the outgoing transition government. In that vein, the authorities will need to adopt a comprehensive short and medium term fiscal strategy. In the short-term, the Government will need to pursue priority spending programs such as wage and salary payments, the delivery of social services for civilians and in the security sector. In the medium-term, the authorities will need to accelerate ongoing structural reforms aimed at promoting the development of the private sector; and carry out a broad fiscal consolidation reform with the objective of reducing the size of the wage bill in the context of a PSR program. Evidence suggests that a reduction of the Guinea-Bissau’s wage bill equivalent to 1 percentage point of GDP is likely to improve growth by about 2 percent on average during 2006-08. The results also suggest that a reduction of the wage bill within the limits of the economy’s absorption capaciw, alone, is not suficient to achieve medium- term fiscal sustainability. Achieving a current surplus in the medium-term will require the combination of expenditure contraction and structural reforms aimed at improving the eficiency of tax collection.

4.1 As mentioned in the second chapter, Guinea-Bissau’s history of conflict and political instability has limited the country’s economic performance during the past three decades. After independence and nearly two-decade long State-ruled economic policies, the economy was gradually liberalized in the 1980s and macroeconomic performance improved during the 1990s despite a difficult political environment. The 1998 armed conflict interrupted the reform agenda and destroyed the country’s social economic fabric. After the conflict, growth resumed during the two subsequent years (1999-00); however, political instability continued to hinder performance throughout the period 2001-03. Starting in 2004, the authorities took a set of measures that contributed to improving fiscal discipline and growth. Given that the sociopolitical situation remains unsettled, public policies aimed at preserving and consolidating peace and social stability should represent the Government’s highest policy priorities in the short term. In order to achieve this objective, the authorities would need to ensure timely and regular payment of salaries for the coming twelve to eighteen months starting in 2006.63 At the same time, the Government would need to strengthen institutional capacity and improve expenditure control and revenue collection. This would allow limiting pressure on the budget and

63 The remainder of 2005 budget is expected to be fully financed provided that there is continuity of governance and reform after the presidential elections.

- 54 - paving the way for a major fiscal consolidation in the medium term. In this context, fiscal consolidation would allow reducing wage costs to sustainable levels and restoring macroeconomic stability and growth in the medium and long term. The remainder ofthis chapter will successively discuss: (i)recent economic performance; (ii)short term fiscal policies aimed at preserving social stability while enhancing institutional capacity and fiscal management; and (iii)medium and long term fiscal consolidation aimed at promoting macroeconomic stability, poverty reduction and growth.

A. RECENTMACROECONOMIC PERFORMANCE

4.2 Guinea-Bissau performed in a generally satisfactory manner during Figure 4.1: Budget and External Accounts Balances, 1987-98. Following a decade of a ruinous 1996-04 (in percentage of GDP) State-controlled economic policy, Guinea- Bissau embarked on a comprehensive reform program starting in 1987.64 With the support of the Bank, the authorities 10 0 0 conducted a poverty assessment in 1994, -10 -20 which established that nearly four-fifths -20 -30.. -40 ofthe population lived in poverty. In light 1996 1997 1998 1999 2000 2001 2002 2003 2004 of the positive macroeconomic outcomes

and the urgency of fighting poverty, the Current primary balance (left axis) authorities agreed upon a three-year II Overall balance excl. grants (left axis) Enhanced Structural Adjustment Facility ...e..- Ext. cur. Acc. Bal. excl. off. Tr. (left axis) (ESAF) arrangement with the Fund in 1995. Broad liberalization ofthe economy contributed to reducing large macroeconomic imbalances. Until end- 1997, the country performed in a generally satisfactory manner. Government primary surplus increased from 3 percent in 1996 to 5 percent in 1997. At the same time, the external account deficit (excluding official transfers) improved from 18 percent to 14 percent (see Figure 4.1, above). Inflation -- measured as the change in the consumer price index (CPI) -- decelerated from a record SO percent in end-1997 to 8 percent in early 1998, owing to a restrained public sector demand and the regained stability associated with the adhesion to the CFA franc. Domestic production, particularly in agriculture, accelerated with real GDP growth reaching more than 5 percent on average during 1994-97. Exports expanded substantially, especially due to better prices for cashew nuts, the predominant source of foreign exchange earnings. The private sector was emerging and playing a more dynamic role in the economy. To address poor social indicators, medium-term development programs for the education and health sectors were approved focusing on improving basic social service delivery. In this context, the third-year ESAF arrangement was successfully completed in early 1998, prior to the outbreak ofthe military conflict.

4.3 The disruption of the economy was significant during the conflict in 1998. Real GDP in 1998 contracted by about 28 percent, due to a sharp decline of production

As evidenced by successful implementation of the Bank-supported Second Structural Adjustment Credit, SACII (Credit 2019- GUB), approved in 1989 and closed in 1994.

- 55 - activities in all sectors. Agriculture production declined by about 17 percent, particularly due to significant deceleration in the production of cashew nuts and cereals. Most ofthe formal industrial, trade and service activities were interrupted in the last seven months of 1998. Closing ofbanks and nonpayment ofmost wages suppressed the impact ofthe war on prices. The armed conflict adversely affected the country's fiscal and balance of payments positions. Marketing of cashew nuts, the main source ofpublic revenue,65 was severely disrupted and merchandise exports and imports (in US dollar terms) decreased by about 45 and 30 percent, respectively, over the period 1997-98. As a result, the external current account deficit (excluding official transfers) remained high at nearly 16 percent of GDP and the current primary budget balance, moved from a surplus of about 5 percent of GDP in 1997 to a deficit of about 7 percent of GDP in 1998 (see Figure 4.1, above). After June 1998, external debt service arrears piled up, equivalent to almost 10 percent of GDP. Arrears on salaries, goods and services, pensions, and transfers were also built up representing 5 percent ofGDP.

4.4 The economy bounced back during 1999-00 due to the return to Figure 4.2: Real GDP andcashew Production Volumes, 1999-03 (change in peace and good agricultural percentage) performance. During the two years following the conflict, economic 10.0 performance improved as a result of enhanced security situation and gradual 0.0 resettlements of displaced populations. 1998,1999'2000 2001 h6f2003t 40 Nearly all of the 300,000 to 350,000 -10.0 persons initially displaced by the conflict -20.0 in 1998 returned to their homes. \I '-. ,' to Agricultural production, which -30.0 represents nearly 60 percent of GDP, grew on average by 6 percent in real -Real GDP Growth Rate (left axis) term per annum during the period 1999- - - - - Cashew Production (right axis) 00. Cashew nut production, the leading agricultural sub-sector, averaged 70,000 metric tons over the indicated period (compared to 50,000 metric tons in 1997). Performance in agriculture had positive spillover effects on the economy as a whole (as evidenced by the trends ofreal GDP growth and cashew nut production growth shown in Figure 4.2, above). Services grew on average by 15 percent in real terms, largely due to the close interrelationship between transportation activities and the boom in the cashew sector. Industrial activities suffered from heavy damage in infrastructure and electricity supply.66 Industrial growth first decelerated by 1 percent in 1999 and then picked up by 10 percent in the following year, thanks to donor support to Government's reconstruction program.

4.5 Real annual GDP grew by nearly 8 percent in average during 1999-00. At the same time, Government revenue reached 18 percent of GDP, and the overall budget

Tax and non-tax revenues, excluding grants, fell by the equivalent of IOpercentage points of GDP to about 5 percent of GDP, due to significant shortfalls in proceeds from fishing licenses and taxes on international trade. 66 It has been estimated that private capital losses due t property destruction and lost of business are estimated at about US90 million. Electricity production plummeted from 50 to 20 million of kilowatts-hour, from 1997 to 2000.

- 56 - deficit (excluding grants) increased from 14 to 25 percent of GDP between 1999 and 2000, as a result of increased post-conflict reconstruction activities. Late in 1999, the Government and the Fund agreed upon a macroeconomic framework for the period 2000- 200 1, with the following objectives: consolidate and sustain economic performance; reduce inflation; and limit fiscal and external gaps to financially sustainable levels. During the period through December 2000, progress continued to be encouraging on the macroeconomic front, and the country completed its interim PRSP, reached the decision point under the HIPC Initiative, and agreed on a three-year PRGF program with the Fund for the period 200 1-2003.

4.6 Inadequate public policies prevented the conclusion of the Fund’s first PRGF review in 2001 which resulted in balance of payment difficulties. Early in 2001, the macroeconomic program was found to be substantially off track owing to large unbudgeted expenditures mainly on defense, financed by credit from the banking system and promissory notes. Late in 2001 a string of political crises added to the worsening security situation at the northern border with Senegal in the Casamance region. In this context, military crackdowns by the authorities led to a significant expenditure overrun. As a result the primary fiscal deficit reached 3.7 percent of GDP in 2001 compared to the targeted surplus of 0.8 percent under the PRGF program. Consequently, the first review of the PRGF by the Fund was not concluded as scheduled in 2001, and the program was declared off-track. After two unsuccessful short-term staff-monitored programs (SMP), efforts to revive the program were abandoned by mid-2002, and the Figure 4.3: External Transfers, 1999-04 arrangement expired at end-2003 without the completion of a review. In view of the 45T 40 disappointing developments, the IMF 35 established that it would be unlikely that 30 Guinea-Bissau would be able to meet PRGF 25 standards in the near future. 20 15 4.7 In these conditions donor support IO (equivalent to 4 percent of GDP) did not 5 materialize as expected, causing severe 0 h 0 0 0 0 0 balance of payment difficulties. External 0N 0N 0N 0N h N budgetary supports decreased from US$12 million in 2000 to US$1 million in 2001 (see BOP support in % of offwial transf (left axis) Fishing and other % of official transf (right axis Figure 4.3). The balance of payment and - licenses m OfficialTransfers m $minion (left axis) fiscal difficulties were further fueled by a record 37 percent decline in world market prices ofcashew nuts in 2001 (see Figure 4.4). Consequently, export volume growth plummeted from 26 percent in 1999 to minus 1 percent in 2002, and Government budgetary revenues as a share of GDP decreased from 19 to 15 percent during the same period. In this context, the external account and budget deficits (excluding external grants) reached 13 and 26 percent of GDP, respectively, by end-200 1.

- 57 - 4.8 The economy became recessionary in 2001 and further Figure 4.4: Changes in Exports and Cashew stagnated through 2002-03. Real GDP Nut prices, 2000-04 (in annual percentage) grew at a rate of only 0.2 percent by end- 2001, and contracted by 7.2 percent in end-2002. The main contributing factors 10.0 to the economic contraction in 2002 were 0.0 15.0 the 15 percent fall of the production of ;;002 2003 2004 10.0 -20.0 cashew nuts, and lower than expected 5.0 -30.0 'A agricultural performance, due to -40.0 10.0 unfavorable weather conditions, and the - ._- Cashew Prices (left axis) continued suspension of donor-funded -Exports volume (right axis) Revenue excl. grants policy lending representing about 7 - percent of GDP. The fiscal situation continued to be precarious due to the sharp decline in economic activity and the implementation of a number of policy- decisions outside the emergency financial management framework agreed with the IMF and the Bank.

4.9 Economic performance remained sluggish in 2003. Although the agriculture sector grew by about 5 percent, real GDP grew by only 0.6 percent due to a significant contraction in the other sectors of the economy. Simultaneously, fiscal management further deteriorated with a substantial diversion ofresources to expenditures taking place outside the legally established budgetary procedures. Perhaps the most deleterious action in public finance management was the decision by President Yala to increase by more than ten-fold the salaries of the military in 2003. Consequently, the public sector wage bill increased from 8 percent ofGDP in 2003 to about 11 percent in 2004.67

4.10 Economic performance started to recover in 2004 thanks to enhanced fiscal management and increased donor support to the EEMP. During the second half of 2004, the transition Government formed after the September 2003 military coup took decisive measures aimed at restoring fiscal discipline. The treasury committee,- established in 2003, was reinstated and tasked with implementing a strict cash flow management system. As a result, budgetary revenues increased from 15.5 percent in 2003 to 17.5 percent by end-2004. It is worth noting that this performance was largely attributable to improved tax collection in the informal sector and illegal fishing activities. The Government, however, continued to rely on external assistance for the implementation of the EEMP in 2004. The economy started to recover in 2004 with real GDP growth reaching 2.2 percent by end-2004. While growth is not expected to change substantially in 2005 due to capacity constrains and continued political uncertainties, performance is expected to improve in the medium term if adequate fiscal and capacity building reforms are carried out.

67 Most public sector wages remained unpaid in the first nine months of 2003 prior to the military coup.

- 58 - B. SHORT TERMFISCAL POLICY FORSOCIOPOLITICAL STABILITY

4.11 In order to enhance sociopolitical stability, the authorities would need to improve living conditions in military barracks and ensure the payment of public sector salaries, and the delivery of basic social services. As part of the SSR, the new administration would need to take urgent measures aimed at rehabilitating the country’s main military barracks, and providing adequate food and sanitation services to the security sector personnel. The Government would also need to ensure regular payment of salaries in the public sector. As the private sector is still recovering from the damage of the conflict, the public sector represents the main source of employment in Guinea- Bissau. Hence, many households depend directly or indirectly on the income ofa relative who earns a salary in the public sector. Ensuring the regular payment of salaries in the civil service and the army would therefore contribute to enhance sociopolitical stability. Another challenge facing the Government in the short term is the delivery of basic education and heath services (school materials and fees, and drugs and other medical supplies) in urban and most importantly in rural areas where the majority ofthe poor live. This includes the provision of free school materials; the continuation of the free tuition program in basic education; and the provision of drugs at affordable prices. Finally, the rehabilitation of minimum electricity and water services should also be placed on the authorities’ short-term agenda (large scale rehabilitation work may be then be carried out in the mediudlong-term).

4.12 The implementation of these ~~ Figure 4.5: Revenue and Recurrent urgent short-term actions will only be expenditure, 1999-04 possible if the international community continues to support Guinea-Bissau. It 35.0 is clear that the Government will not be 30.0 able to implement these priority 25.0 immediate actions on its own. The 20.0 Government is likely to face more 15.0 10.0 challenges in delivering basic social 5.0 services. In the past five years, only about 0.0 5 percent of the Government’s budget 1999 2000 2001 2002 2003 2004 went to education and health. These Recurrent primary spending resources mainly served to pay salaries, Revenue (includes all grants) while school and health non-salary 0 Revenue excl. grants spending were mainly financed by external donors. The Government’s current fiscal balance has been negative over most of the post-conflict era (see Figure 4.5). The Government, therefore, relied mostly on external assistance to balance its current account. As the country is still recovering from the structural damage of the conflict, this trend is likely to persist at least for a couple of years. Budgetary support -- in the context of the IMF’s emergency post-conflict assistance (EPCA), the Bank’s Economic Management Reform Grant (EMRG), and the European Union’s budget assistance program -- are expected to provide more fiscal space to the Government for the payment ofpublic sector salaries in end-2005 and during 2006. It is crucial that such support is sustained for some time, while the Government is preparing to engage in fiscal consolidation reform over the medium term.

- 59 - c. MEDIUMAND LONGTERM POLICY FOR MACROECONOMICSTABILIZATION AND GROWTH

4.13 In the medium-term, the authorities would need to restore fiscal Figure 4.6: Evolution of the Payroll and the Wage discipline through a major fiscal Bill, 1999-2004 consolidation effort. As mentioned earlier, Guinea-Bissau’s current fiscal balance has been negative for the past five years (average deficit of 6 percent of GDP during 2000-04). Several factors are 20 T T 15.0 - 5.010.0 believed to have contributed to this 18 outcome. First, the public sector wage bill 16 0.0 appears as the main contributing element 1997 1998 1999 2000 2001 2002 2003 2004 among the factors under the Government’s control. Following the conflict there was a Payroll excl. parastatals, in 1000 people (left axis) need for rebuilding the public - - - - Wage bill in ‘YOof GDP (right axis) administration as a response to post- conflict reconstruction. Consequently, the wage bill increased from 3 percent of GDP before the conflict in 1997 to 5 percent of GDP in 1999, and stabilized around 7 percent of GDP on average over the 2000-02. In 2003, President Yala increased significantly the salary ofthe military (8-fold) in order to gain their support. As a result, the wage bill rose to nearly 12 percent ofGDP by end-2004 (see Figure 4.6).

4.14 Second, weak capacity for revenue collection and expenditure control has also contributed to the current weak fiscal stance. The personnel of the customs, the main revenue collection agency, lack adequate material and training for implementing custom tariffs and rules of the WAEMU. Similarly, major expenditure management agencies such as the procurement department of the Ministry of Finances (MOF) and the human resource management department of the Ministry of Labor (MOL) will need to be reinforced for better management of public procurement and payroll systems, respectively.

- 60 - Box 4.1: Summary of the Main Findings of the 2001/02 Diagnostic Studies of the Public Sector

Enforce the 1994 Administrative Law. Guinea-Bissau’s public administration is in principle governed by the 1994 Law. In practice, however, it continues to be regulated by the 1973 Law, which does not clearly make the distinction between administrative and career positions and remains dubious.

Redefine staff ranking systems based on career structure and educational levels. Currently, the civil service uses a hierarchical structure combining career (or rank) and administrative (or position) systems. The different levels ofrank and position are defined by the letters ofthe alphabet (from “A” to “N” for administrative staff 78 percent of the payroll; and from “0” to “Z” excluding “W for career staff 22 percent ofthe payroll). As recommended by the public sector reform study carried out in 2001, the current system would gain in defining a limited number of categories in each of the career and administrative groups based on the country’s different educational levels and career structure.

Define a performance evaluation mechanism aligned with the remuneration and advancement system. The 1973 Law was amended to introduce an article on performance evaluation. However, the performance evaluation system currently in use remains subjective and does not include a mechanism which allows defining ex-ante individual result agreements for a given period. Moreover, the current performance evaluation mechanism is not fully aligned with the advancement and remuneration system. The current advancement and salary increase mechanisms are not transparent and take little account of seniority and performance.

Strengthen human resource and payroll management capacities of the Ministry of Labor. The capacity ofthe Ministry ofLabor and Civil Service to manage human resources and the payroll remains very weak. In particular, the General Direction in charge of the management of human resources does not have adequate tools for centralizing and harmonizing the management of the payroll system (civil and military).

Reduce the wage. The wage bill ofthe public sector grew fiom 5 percent in 1999 to 11 percent in 2004 (CFA 6.9 and 16.2 billion in nominal terms, respectively). The Central Government payroll is estimated to have increased from about 17,300 to 19,600 public servants during the same period (the salary increase of 8-fold granted to the military by Yala in 2003 has led to an overall increase in the wage bill ofabout 2-fold by end-2004). Reducing the wage bill would require netting out possible ghost workers from the payroll, encouraging early retirement and enforcing the mandatory retirement age in the civil service and the military. The 1994 Law, established the mandatory retirement age in the civil service and the army at 60 years. This represents a relatively long professional career, as the average public servant in Guinea-Bissau, is approximately 40 years old, and has a live expectancy rate higher than the national average of45-50 years.

Harmonize the pension system in the army. Guinea-Bissau’s current pension system is based on the ‘pay-as-you-go’ principle. The system has the advantage of being simple because no specific pension fund is required. Retirement benefits represent, therefore, a debt on the Government. The current system provides a safeguard mechanism against the misuse ofpre-established retirement funds as it has been the case in other countries. The main caveat ofthe current system is most likely the fact that it is not filly enforced, mainly in the army. Moreover, benefits granted to high ranked officers are 10 times higher than those granted to rank and file.

- 61 - 4.15 The proposed fiscal consolidation reform should be carried out in the context of a comprehensive PSR program.68As expenditure contraction will primarily target wage expenditures (in sectors other than education and health), it is important to place the fiscal consolidation reform in the context of a broad PSR program including a civil service reform and a SSR. The most recent civil service reform in Guinea-Bissau dates- back in the late eightiedearly nineties. In the early 2000s, the Government carried out diagnostic studies of the public sector with the help of the European Union and The Netherlands (see Box 4.1, above). These studies were intended to serve as ingredients for a reform of the civil service in Guinea-Bissau. The implementation of the reform has, however, been delayed due to the prevailing political instability. A civil service reform in Guinea-Bissau would ideally target three categories of civil servants: ‘ghost workers’; people still working while in retirement age; and the redundant staff. In the simulations carried out in Section 2, the potential saving through a reduction of wage spending - within the limits of the economy’s absorption capacity - is estimated at CFA2 billion6’ per annum. This amount is equivalent to the salary of about 6,000 civil servants per year at the rate ofthe minimum wage (CFA 28,000). The Ministry ofLabor has estimated that about 4,000 the number ofcivil servants (‘ghost workers’ not included) would need to be discharged in the context ofthe civil service reform. The real estimate will depend on the number of ghost workers, which can be determined on the basis of the civil service census carried out in end-2004.

4.16 The PSR program will also need to include structural reforms of the SSR. Overall, there are 5.04 soldiers per 1000 inhabitants in Guinea-Bissau compared to a regional average of 1.23, and the majority of the army staff is from the Balante ethnic group. On average, security spending represented 2.8 percent of GDP compared to only 1.7 percent in West Africa. Addressing the structural problems in the army would require a comprehensive SSR in the context of a broad PSR. In addition to the short-term actions mentioned in the previous section, the SSR should address the mediudlong-term structural reforms ofthe security sector (see Box 4.2): (i)reforming the pension system in- the army and encouraging early retirements (see Table 4.1);70 (ii)downsizing the army, which exceeds the average army size in the sub-region; and (iii)gradually balancing the ethnic composition of the military, currently dominated by the Balante ethnic group. Several development partners, including the Community of Portuguese Language Countries (CPLP), the United Nation Office in Guinea-Bissau (UNOGBIS), and the Economic Community ofWest African States (ECOWAS), have expressed willingness to support a SSR process in Guinea-Bissau.

The analysis in Box 4.1 is based on the findings of the study “ElCments de RCforme du Systbme Salarial et du Regime du Personnel” camed out by the Ministry of Public Administration and Labor in 2001. 69 Significantly higher contraction of the wage bill would not be technically feasible due to absorption capacity constrains. ’O The number of high ranked officers is estimated at around 25, mostly from the Balante ethnic group. Hence there are nearly as much officers per thousand soldiers as there are soldiers per thousand populations.

- 62 - Box 4.2: Objectives of a SSR in Guinea-Bissau

Reducing poverty in the barracks (short term). Currently there are 24 military barracks in Guinea-Bissau. Their locations are mainly based on the Portuguese military strategic points during the war of liberation. The barracks lack proper housing, sanitary facilities, and health services. Education levels among military are very low. While many stay despite the conditions for the minimal security it allows them to 'at least receive one meal a day'. Desertion is reportedly an increasing feature of the military forces, although no official numbers are available. The desertion is not prosecuted or sanctioned, and contributes to the proliferation of arms particularly in Bissau. Rehabilitating military barracks and ensuring adequate provision of food, dormitory and sanitary facilities should represent a critical component of a SSR. In this regard, the Government will need to prepare an action plan for the SSR in the annual action plan of the PRSP for 2006. This will facilitate discussions around the financing of the SSR, as many donors have already expressed willingness to participate in the initiative. Reducing the size of the army (medium term). There are uncertainties about the actual size of the armed forces. A recent public sector census did not include the police and military. A census of the armed forces would therefore represent a key element of the SSR. In 2003, Guinea-Bissau counted about 7,000 armed forces personnel, corresponding to 5 soldiers per 1000 inhabitants and a public expenditure share of nearly 3 percent of GDP (about CFAF6 billion, of which 5 billion are for salaries, and the remainder are for food and other military expenditures). Hence, the size of Guinea-Bissau's army (in terms of the number of soldiers per capita and public spending share of GDP) is relatively higher than the regional average of 1.2 soldier per 1000 and 1.7 percent of GDP, respectively (see Table 4.1, above). The Community of Portuguese Speaking Countries (CPLP) recommended downsizing the current number to around 3,000 troops. Reducing the size of the army would likely require a voluntary retirement scheme backed with a reformed pension system (which would address issues related to the reluctance to retire from the army). It would also require, mainly for discharged soldier still in the working age, the creation of an environment conducive to private sector-led employment (this aspect is hrther discussed in chapter two). Balancing ethnic composition (long term). The third objective would be a longer term challenge of balancing the ethnic composition ofthe army without inflating the size. Ex-President Yala accelerated Bulunte recruitments and appointments to the armed forces during his presidency, and thereby made the army problem also a potential ethnic problem. Attempts are being made to reintegrate officers considered arbitrarily demobilized in 2001, predominantly belonging to President Vieira side of the 98-99 conflict A significant difference from other post-conflict armies is its one-sided character, where creating a balance between ethnic groups and political affiliations without inflating the forces will inevitably cause great disgruntlement. Mandatory military service and nondiscriminatory recruitment procedures are possible ways to address the issue of ethnic imbalances in the army over time.

Table 4.1: Comparative Military Data for Selected African Countries in 2003 Personnel Expenditure Total Per 1000 Growth rate (%) per capita (US$) % of GDP Chad 30,000 3.14 5.89 2.1 Congo (RDC) 8,000 0.13 15 1.98 1.4 Cameroon 13,000 0.8 79 11.77 1.4 Benin 5,000 0.68 7 12.55 2.7 Burkina Faso 7,000 0.5 1 70 3.88 1.6 Cote d'Ivoire 8,000 0.46 -39 10.01 1.2 Gambia (The) 1,000 0.64 60 0.58 0.3 Guinea 10,000 1.08 -2 6.32 1.7 Guinea-Bissau 7,000 5.04 -15 6.05 2.8 Mali 7,000 0.58 51 4.27 1.3 Niger 5,000 0.44 141 1.91 1.1 Nigeria 76,000 0.55 -19 3.42 0.9 Senegal 10,000 0.92 -7 8.82 1.5 Average of WAEMU (*) 7,000 1.23 30 6.78 1.7 w:Nationmaster.com (*) ExcludingTogo.

- 63 - 1. Development of the Private Sector as Prerequisite for Fiscal Consolidation

4.17 The success of the proposed fiscal consolidation reform will depend on the capacity of the private sector to create new job opportunities as an alternative to public sector employment. In Guinea-Bissau, the bulk of the workforce in the formal sector is employed by the public sector. In 2002, the Government approved a private sector development strategy to support the privatization and divestiture of public enterprises; improve the business environment; and rehabilitate basic infrastructure. Progress in the implementation of the divestiture program, albeit slow, has been encouraging during the past three years. In 2002, the parastatal sector of Guinea-Bissau consisted of 44 enterprises, employing about 3,400 staff, Except for utilities, these enterprises were small and many had ceased activity. More than half of these firms have been now divested or liquidated, notably the ex-Banco International da GuinC Bissau (BIGB), the country’s sole commercial bank. Completing the divestiture ofthe remaining public enterprises would contribute to eliminate contingent liabilities on the Government budget and create new jobs.

4.18 The development of the private sector will require improving the business climate, rehabilitating rural infrastructure, and microfinance. Accelerating structural reforms will be needed for improving the business climate and enhancing job creation in the private sector. Guinea-Bissau is a member of the Organization for Harmonization of Business Law in Africa (OHADA).~~In 2002 and 2003, the Foreign Investment Advisory Services (FIAS) of the Bank carried out a series of studies which revealed that weak institutional capacity, and cumbersome tax and other administrative barriers are major- constraints for the development of the private sector in Guinea-Bissau. In 2004, the Government developed an action plan with the objective of reducing the tax burden on businesses; simplifying the tax system; and improving the institutional capacity of the legal system. So far, the implementation of the action plan is progressing well. Since 2004, the corporate income tax rate has been unified to a single rate, and reduced from 39 to 30 percent. Moreover, in 2005, the Government adopted a law to reduce and simplify business licensing requirements (alvara). The implementation ofthese measures remains, however, constrained by the lack of adequate technical capacity within the public administration.

4.19 The rehabilitation of telecommunication, water, energy, road, and port infrastructure is critically needed for the development ofprivate businesses. In the area of telecom, the Government, through the Bank-supported private sector program, has successfidly liberalized the telecom sector. Two private companies have been providing cellular phone services (more than 50,000 lines) in the country since end-2004. Rebuilding the main rural and urban road networks, rehabilitating the port and the water. and electricity company will represent the Government’s main challenges over the medium-term. As these areas of reform are mainly financed through external donor assistance, good governance and strong commitment to reforms will be key in attracting the much needed financing.

” OHADA covers the following areas: general commercial law; corporate law; laws concerning guarantees and collaterals; debt recovery and enforcement law; bankruptcy law; arbitration law; and accounting law.

- 64 - 4.20 The development of microfinance services will provide critical financing means to private businesses. Only one commercial bank is currently operating in Guinea-Bissau. The banking sector depth is very limited with a relative low level of monetization. The financial sector is also characterized by relatively high lending interest rates (4.75 percent), for small enterprises (12-18 percent), compared to relatively lower average market rates in the West African Economic and Monetary Union (WAEMU). Private sector credit represents only 2 percent of GDP. A small number of microfinance institutions operate in the country, mainly in the informal sector, without adequate legal and regulatory framework (2 microfinance institutions per 100,000 inhabitants compared to 7 per 100,000 in Senegal).72 A program for developing the microfinance sector for the period 2004-07 was prepared by the United Nation Development Fund (UNDP) and the United Nations Capital Development Fund (UNCDF) in 2003. However the country’s unstable political environment has not allowed making progress on this front. Given the importance ofmicrofinance in the country’s economy, the participation ofother donors in the UNDPRJNCDF program in the form of parallel financing would provide greater leverage to the initiative.

2. Fiscal Consolidation: A Necessary Step toward Fiscal Sustainability and Growth a. Rational for Fiscal Consolidation

4.21 Empirical evidence suggests that reducing nonproductive public spending can spur growth. It is well established that in low income countries with low public sector productivity, high budget deficits often lead to a high domestic demand which puts pressure on domestic price levels. This contributes to reduce consumption and investment opportunities for the private sector through the traditional crowding out effect. Fiscal consolidation -- the reduction of unproductive public spending (namely wages) -- has often resulted in sizeable fiscal su luses, more stable and decreasing debt to GDP ratios, and faster and sustainable growth? When the level ofthe public sector deficit is high and unsustainable, recent empirical studies suggest that expenditure contractions can yield significant expansionary effects.74

4.22 Several factors may contribute to explain the expansionary effect ofa reduction of wage related expenditure in Guinea-Bissau. First, in recent years the authorities have increasingly turned to domestic borrowing to service the budget deficit, in the absence of external assistance. The resulting crowding-out effect combined with the already fragile situation of the private sector have contributed to depress growth. Reducing wage spending in the context of a civil service reform will contribute to reorganize the

72 A subsidiary of the Compagnie Bancaire de I’Afrique Occidentale (CBAO) based in Dakar, has already realized an investment of . CFAl.2 billion CFA for starting activities in Bissau by end-2005 or early 2006. ” Baldacci et al (2003) show that increases in Government non-wage expenditures in Benin, The Gambia, Senegal (among others) were compatible with higher growth.

74 A study by the IMF using pooled information (time series and cross sectional data) on 39 countries (industrial and developing countries) over the period 1990-2001 established that, on average, a reduction of less than 0.5 percentage points of GDP in budget deficits led to a per capita income growth of 0.5 percent per year during the period. Similar conclusions were found for Brazil where the sharp fiscal consolidation policy carried out by the Cardozo Government in 1999 was crucial in restoring fiscal sustainability and growth.

- 65 - attribution of task in the civil service and thus improve public sector productivity. Subsequently, public sector budget deficit and indebtedness levels will improve, and the ensuing crowding-in effect would counteract the usual contractionary effect stemming from fiscal consolidation as posited by the traditional Keynesian multiplier theory.

4.23 Second, fiscal consolidation is expected to improve governance and public confidence in Government policies. Following the 1998 conflict the Government has accumulated a considerable stock of domestic arrears vis-&vis the private sector, which has been only partially cleared so far. Moreover, public policies in recent years have led to major budgetary slippages which undermined economic performance and public confidence in Government policies. A strong commitment to reform through a sustained fiscal consolidation policy is likely to restore public confidence in Government’s ability to create a favorable environment for growth. The quality ofthe fiscal reforms, the speed with which they are adopted, and the conviction that they will be carried out are fundamental in convincing businesses and households that economic conditions have improved in such a way that they can adjust their consumption and investment behaviors accordingly to boost aggregate demand and growth.

4.24 Third, fiscal consolidation is expected to generate stronger economic growth through expenditure composition effects, which could, for example, flow from an increase in the proportion of productive expenditure and the reduction of distortionary taxes. It is well known that private investment in particular appears to react strongly to fiscal measures that increase expected benefits, for example from the reduction of transaction costs. Increased spending in infrastructure, for example, would contribute to reduce transaction costs and increase productivity and growth in the private sector. b. Sequencing of Actions

4.25 Downsizing the public sector. In order to enhance public sector productivity and growth it is urgent to downsize the Government’s wage bill (in sectors other than education and health) at sustainable levels, in the context of a medium-term civil service reform. Prior to the conflict, the Government had already considered downsizing the public sector. The options discussed included the suppression of all unskilled personnel working on cleaning, maintenance and surveillance. These activities were to be subcontracted to private firms to be selected through competitive bidding. Following the conflict, the absence of adequate regulation of recruitments in the army and the civil service contributed to further increased wage spending. Moreover, weak monitoring of retirements, transfers and deaths within the public service contributed to increase ‘ghost workers’. The downsizing would therefore need to include active workers as well as ‘ghost workers’ if there are any.

4.26 For efficiency and fairness purposes, separation packages may be needed to overcome resistance to public sector downsizing (see Annex 2 for further discu~sion).7~ Ifno compensation were provided, many public sector workers would suffer as a result of

75 Separations packages have been used by governments all around the world often with support from external donors. Haltiwander and Singh (1999) in a cross-country study on developing countries estimated the average compensation package at US$2,400. However, the figures were as high as US$13,000 in neighboring Senegal.

- 66 - job separation. In this case they are likely to be exposed to low income and poverty. A World Bank study on Guinea-Bissau estimated the “just right” compensation package, using a forward-looking formula to determine the present value of the expected loss in income and social benefit.76 The results indicate that the average loss in earning for a public servant would range from CFAF672,OOO to CFAF745,OOO (roughly US$1,300).- The compensation package offered to redundant workers is generally a lump sum payment or a multiple of the base salary that takes into account seniority. However, the “right” compensation package would also need to take into account education levels and location factors (for example, the compensation for workers with secondary and technical education or living in remote areas may need to be higher than the one for those with primary education or living outside the capital city Bissau).77

4.27 Enhancing shared growth and fiscal sustainability. In the case of Guinea- Bissau, evidence -- based on simulations from a three-variate Vector Auto Regressive (VAR) model7* (see details on the model in Annex 1) -- suggests that if real Government consumption (namely wage spending) were reduced by 2 percent (5 percent in nominal terms corresponding to 1 percentage point of GDP) annually over the period 2006-2008, real GDP would increase by about 1.6 and 2.9 percentage points in 2007 and 2008, respectively, compared to baseline growth targets for the same years (see Table 4.2). Although these results are consistent with empirical findings, they may be overestimated in the case of Guinea-Bissau, given the weak institutional capacity of the State and other binding constraints to private sector growth. Assuming the aforementioned preconditions are met, the model predicts that if expenditure contraction were not pursued beyond 2008, the expansionary effect would falter and even reverse by 201 1.

4.28 Public sector downsizing will, however, not be sufficient for achieving medium- term fiscal sustainability. In order to achieve a current surplus in the medium-term, the Government will need to pursue sound expenditure management and improve the efficiency ofrevenue collection by reinforcing and sustaining the ongoing reforms in key departments of the Ministries of Finance and the Ministry of Fisheries. Activities will need to include well targeted capacity building programs in customs and fishing licensing, payroll and procurements department^.^^ Table 4.3 provides two series of simulations to illustrate the fiscal effect of higher revenue collection efforts.*’ In 2004, Guinea-Bissau’s revenue-to-GDP ratio reached 17 percent (of which only 8 percent were for tax revenue). This ratio is expected to improve to about 20 percent in 2005 thanks to tax and custom reforms currently being carried out by the Government. If the revenue collection effort during 2006-08 were kept the same as in 2005, the budget will yield a- current primary deficit of4.3 percent in average during 2006-08 (simulation 1).

76 Chong and Rama (2001). 77 Potential earnings are higher in Bissau than outside the city. Likewise, potential losses are higher outside Bissau due to the scarcity ofjob opportunities outside of the capital city. 78 The model uses annual data on real Government consumption, real GDP and terms of trade changes over a two-decade period for- Guinea-Bissau. 79 This does exclude a discretionary increase in tax rates. Any increase in tax rates is likely to have a contractionary effect as tytulated by the neoclassical economic theory. The simulation assumptions are footnoted to Table 4.3.

- 67 - Table 4.2: Impact of a Reduction of Wage spending on Growth in Guinea-Bissau, 2005-08 Years 2005 2006 2007 2008 Shock to real public spending (reduction in real wages and salaries) 0.0% -2.0% -2.0% -2.0% Shock to terms oftrade (- = depreciation) 0.0% -1.0% -1.0% -1.0% Short-term growth effect /1 0.0% 0.0% 1.6% 2.9% Short-term fiscal effects /2 0.0% -2.0% -1.2% -1.1% Source: Bank Staff Estimates /I Deviation from baseline real GDP growth targets as defined in IMF’s SMP /2 Deviation from baseline real government spending growth as defined in the SMP Source: Bank Staff Estimates.

Table 4.3: Efficiency in Tax Collection and Medium-Term Fiscal Sustainability, 2005-08 2005 Average 2006-08 Proj. Simulation 1 ” Simulation 2 (in billions of CFAF) GDP 153.7 166.0 166.0 Budgetary Revenue 29.3 31.6 39.8 Government current primary expenditure 37.5 35.5 35.5 ofwhich: Wage Bill 19.5 17.5 17.5 . Current primary balance -8.2 -3.9 4.3 (in percentage of GDP) GDP 100.0 100.0 100.0 Budgetary Revenue 19.1 19.1 24.0 Government current primary expenditure 24.4 21.4 21.4 of which: Wage Bill 12.7 10.5 10.5 Current primary balance -5.3 -2.3 2.6 Source: Bank Staff Estimates. /‘lation 1 assumes an average nominal GDP growth of 7 percent (4 percent in real terms based on the growth pattern predicted in Table 4.1), and a reduction of the wage bill of 5 percentpased on the a real decrease of 2 percent as predicted in the VAR model). ’’ Similation 2 assumes the same conditions as in simulation 1, and improved efficiency in tax collection leading to an increase of tax revenue equivalent to about 5 percentage points of GDP.

4.29 However, if the collection effort improves as a result ofincreased efficiency in tax collection, the budget is expected to generate a current primary surplus of 4.6 percent on average during the same period. The strength and sustainability ofthese outcomes would depend on the Government ability to accommodate these fiscal policies with sound monetary and structural reforms aimed at stabilizing the macro economy, and improving the business environment for the promotion and development ofthe private sector.

4.30 Reducing the size, reforming the pension system, and balancing the ethnic composition of the security sector. In addition to improving living conditions in military barracks in the short-term, the Government would need to reduce the size, reform the pension system, and balance the ethnic composition of the armed forces over the mediwdlong-term. The proposed SSR should be carried out in the context of the authorities’ broad PSR program.

4.3 1 With regard to reducing the size ofthe security sector, the authorities would, first, need to carry out a census of the armed forces. The census should be designed so as to

- 68 - allow determining the actual size of the security sector; providing information on the pension system for the military and ex-combatants; and providing information on the retirement system in the army. The next task would be to determine the ‘optimum size’ of the armed forces, based on a recent Government study (carried out in collaboration with the Portuguese Military Cooperation),81 and the PRSP’s medium-term public expenditure framework. Preliminary assessment suggests that both high ranked officers and rank and file staffs are potential candidates for downsizing.82 As in the case of the civil service reform, separation packages may be needed to compensate redundant military personnel, and mitigate risks ofinstability. Mandatory military services may also represent a way to downsize the active security personnel to the strict minimum.

4.32 With respect to the pension system in the army, there is a need to reform the system so as to provide greater incentive to retire. In the current system, the benefit package for high ranked officers is 10 times higher than those for rank and file staff. A retired officer earns on average CFAFFl80,OOO (about $360) per month, which includes the regular pension for officers (between CFAF28,OOO and CFAFF30,OOO) and other benefits (subsidies for electricity, water, telephone, gasoline and rent). A retired rank and file staff earns only the regular pension which is approximately half ofthe rate applied to an officer (between CFAF14,OOO and CFAF17,OOO). The relative “parsimony” of the pension system toward the rank and file may explain why people are reluctant to retire from the army (retirement rules in the army -- which are similar to those applied in the civil service - are not fully enforced, because the mandatory retirement age of60 years is not being observed).

4.33 With regard to the ethnic issue, the country’s army is in majority composed of people from the Balante ethnic group. If the ethnic composition ofthe armed forces were to remain the same, the military would continue to represent a potential candidate for political manipulation as mentioned earlier. In the long-term it would be important to adopt a recruitment system that promotes ethnic balance within the armed forces.

4.34 Addressing external debt issues and normalizing relations with creditors. The sustainability of Guinea-Bissau’s external debt remains a major concern. The country qualified for the decision point of the HIPC Initiative in December 2000, and started receiving interim debt relief from the Bank and other multilateral and bilateral donors (IMF, AfDB, Paris Club, etc.). As mentioned above, fiscal slippages in early 2001 derailed the donor-supported macroeconomic program, and interim debt relief by the Fund and the Paris Club creditors stopped after 2001 (only the Bank and the AfDB continue to provide HIPC interim debt relief). Subsequently, most external debt service has gone in arrears, and additional borrowing -- albeit enerally on concessional terms -- led to a firther increase in already high external debt. 8f In the medium-term’ a complete write-off of all external debt as called for by the recent G8 initiative would imply debt service savings for the Government of about CFAF 25 billion per year, and would substantially reduce the Government’s financing requirements. However, there would

*’ ‘Livro Banco’ da Defesa Nacional da Rep6blica da GuinC-Bissau (2002). There are 5 military personnel per 1,000 inhabitants in Guinea-Bissau compared to a regional average of 1. Moreover the MOF’s data indicate that there are Isenior officer for every 5 soldiers (2040 times above standard). 83 The total nominal debt stock was estimated at around US$915 million (730 percent ofNPV of debt-to-exports ratio) at end-2003.

- 69 - still remain a sizeable financing gap after debt relief. While some of this gap would presumably be filled by additional external support, some restructuring of Government expenditures would still be required. Finally, medium-term debt sustainability depends on continued fiscal adjustment efforts by the Government and support from the international community particularly during the period 2006-2007.

c. FACTORSTHAT MAY CONSTRAIN IMPLEMENTATION OF REFORMS

4.35 Political instability. Persistent internal political tensions represent a serious threat. to peace and stability. The political situation remains polarized following the recent change of the Prime Minister by the President. In this context, it is likely that the 2005 Donor Round as well as the finalization and implementation of the Government’s PRSP will be hrther delayed.

4.36 Weak institutional capacity. As mentioned in the second chapter, the weak institutional and governance capacity of the State represent a major limitation to the implementation of the aforementioned reform agenda. Provided that political stability is maintained, ongoing donor support to the authorities’ capacity building program would contribute to mitigate this risk in the medium-term.

4.37 Vulnerability of cashew nuts to external price fluctuations. Government’s revenue relies heavily on the taxation ofraw cashew export. Consequently, terms of trade losses stemming from the fall of the international prices of raw cashew nuts often negatively affect Government revenues. In such cases, the Government often faces difficulties to pay salaries and the debt service.

D. CONCLUSION AND POLICY CONSIDERATIONS

4.38 Achieving a sustainable macroeconomic and fiscal performance continues to represent a major challenge for Guinea-Bissau. This important task is complicated by the continued political instability and inadequate public policies, such as the increase of security sector salaries implemented in 2003. It is imperative that the new administration ensures the continuity of institutions and commits to the reform program put in place since end-2003. In that vein, the authorities will need to adopt a comprehensible short and medium term reform strategy. In the short-term, the Government will need to pursue priority spending programs such as wage and salary payments and the delivery of social services to civilians and the military. In the medium-term, the authorities will need to: (i) accelerate ongoing structural reforms aimed at promoting the development of the private sector; (ii)carry out a broad fiscal consolidation reform with the objective ofreducing the size of the wage bill, improving the efficiency ofrevenue collection and promoting fiscal sustainability; (iii)reforming the security sector; and (iv) normalizing relations with external creditors. Strengthening political stability and institutional capacity and government are key for implementing these reforms.

- 70 - 5. MAKINGAGRICULTURE MORE PRODUCTIVE AND SUSTAINABLE FOR POVERTY REDUCTRION

Agriculture is the engine of Guinea-Bissau’s economy. The sector relies mainly on subsistence production of food crop and the monoculture of cashew nuts. The latter represents the main export earning commodity and the principal source of income in rural areas. Its production represents more than 90 percent of country’s exports and employs about 80percent of the rural labor force. The presence of bureaucratic and cumbersome administrative arrangements and weak legal systems often lead to high transaction costs for cashew buyers and exporters, which contribute to decrease the fam gate price of the raw nuts. The reduction of these costs through improved investment climate and better organization among actors in the cashew chain would contribute to improve incomes and reduce poverty in rural areas. For the rural sector to create new jobs, policy makers would need to aim at promoting the processing of raw cashew into exportable cashew kernels, and exports of tropical fruits. This is particularly important at this moment as the Government is readying to implement a major PSR. Agricultural diversification would revive the private sector and contribute to create the much needed new jobs to absorb part of the labor force to be ? discharged from the public payroll. 5.1 Agricultureg4 has long been the backbone of Guinea-Bissau’s economy. The sector represents more than half of the GDP, employs more than four-fifths of the labor force, and contributes to more than 90 percent ofthe country’s export earnings.g5 Overall, two types of agricultural activities are practiced: subsistence agriculture based on the production of food crops and cashew nuts fanning. Agricultural production rests essentially on subsistence fanning. Small producers, estimated to be some 90,000 households,g6 represent the bulk of the rural population and generate about 90 percent of the production for internal consumption (rice is the major crop and staple food), and close to 80 percent ofthe export production (cashew). In addition to the traditional farmers (in Tabancas), there is a small number of “ponteiros” (2,000), that have access to important land concessions. Over the past three decades, the production of cashew nuts has taken a leading role in the country’s economy. Guinea-Bissau essentially produces raw cashew- nuts. The sector, therefore, remains highly undiversified and vulnerable to external price fluctuations. Moreover high transaction costs associated mainly with the poor business environment and weak physical and administrative infrastructure, contribute to limit income generation for smallholders in the cashew sector.

84 Includes cash and food cropping, fishing, livestock, and forestry 85 Agriculture is practiced by two types of producers. Small farmers comprised of approximately 90,000 households with an average farm size per household are about 2-3 hectares. Large farmers or ponteiros, with land right concessions assigned by the government, possess about 2,200 properties covering 27 percent of arable land or 9 percent of the national land, of which 1,200 are not under production. An average property is 136 hectares ranging from 20 to 3,000 hectares. Small farmers account for 75-85 percent of the cashew plantations, ranging from 2 to 3 hectares in property size. 86 FA0 2004 Nouveau Partenariat Pour Le Dkveloppement de L‘Afrique (NEPAD). Programme Dktaillk Pour le Dkveloppement de L’Agriculture Africaine (PDDAA) RCpublique De Guinke-Bissau TCP/GBS/2907

- 71 - 5.2 The ‘cashew boom’ has, however, had an adverse effect on rice production, which represents a major livelihood mean for the population. Increasing agricultural productivity in order to reduce poverty, will require the implementation of policies aiming at promoting the production of rice - which is the main livelihood mean of the poor - and the diversification ofthe cashew sub-sector. Both rice and cashew productions are subject to high transaction costs which represent a major constraint to increasing productivity and incomes in rural areas. The discussion in this chapter will, first, provide an analysis of the main opportunities and constrains in the agricultural sector, and then focus on analyzing transaction costs and their impact on productivity and income generation for smallholders in rice and cashew sectors.

A. MAINOPPORTUNITIES AND CONSTRAINTS IN AGRICULTURE

5.3 Guinea-Bissau has a largely untapped potential for agriculture. The country is endowed with fertile soils, abundant water and favorable climate. Agricultural production includes cashew nuts, rice and other cereals, fruits, fishing, livestock and forestry products. Exports include cashew nuts, fish and seafood, fruits, palm kernels, and timber. Cashew and rice remain the major cash crop and staple food, respectively.

5.4 The fast growing cashew sector remains the main source of income for the majority of the population. Guinea-Bissau is the sixth largest producer of cashew nuts (6 percent ofthe world cashew production) after India, Vietnam, Brazil, Ivory Coast, and Tanzania. As a labor intensive activity, the production of cashew nuts plays a vital and leading role in Guinea-Bissau’s economy. Started as a marginal activity with an export volume of about 2,500 metric tons in the 1970s, exports of cashew nuts have increased steadily during the past three decades, reaching over 80,000 metric tons by 2004. Cashew planting was first established in the Northern regions ofOhio and Biombo, and expanded. rapidly in the East and the South. About 5 percent of the country’s land or 175,000 hectares are devoted to cashew production, with average farm sizes of 1.6 and 14.2 ha for smallholders and large producers (ponteiros), respectively. This share continues to grow at the rate 4 percent per year.

5.5 Brush forest clearing and planting are the primary activities required for establishing a cashew plantation. These activities are generally carried out by unskilled laborers and can take approximately 30 person-days of labor at a cost of approximately. CFAF60 000 per hectare (US$1 10). When the fruits and nuts ripen in April through June, small farmers collect their own, while larger planters (“ponteiros”) hire workers, often women, to do the collection. Moreover, during the collection of the cashew nuts it is estimated that about CFAF125 (US25 cent) per Kg is paid as labor cost (approximately 50 percent of the average farm gate price per Kg). This does not include stolen nuts and whatever additional incentives and food the collectors receive. Under this system, if the collectors gather 35 Kg/day of nuts, they earn CFAF8,750 every four days, or CFAF2,188/day (about US$4.5), which is a very attractive rural wage rate in Guinea-- Bissau.

5.6 Guinea-Bissau possesses a large untapped potential in rice production. Rice (paddy) represents the primary staple food crop. As in many African counties, rice plays

- 72 - an important socioeconomic role in Guinea-Bissau. It is generally produced for consumption and food security, sale on the market, and batter transactions. Rice production has increased by more than 10 percent per year during the 1980s, largely because of improved economic incentives. However, supply persistently failed to meet demand, hence leading to increased imports of rice (80,000 tons per year on average representing about 85 percent of total cereal production). Rice production, traditionally practiced in the South, stalled following the boom of cashew in the 199Os, due mainly to increased reallocation ofagricultural land and labor force to cashew prod~ction.~~

5.7 Guinea-Bissau’s weather is favorable to the production of other cereals and a large variety of fruits. Other cereals (millet, sorghum and maize) play a major role in food security. Average annual production reached more than 50,000 metric tons per annum in recent years. Guinea-Bissau’s weather is favorable to the production of a large variety of hits including mango, palm oil and cola nuts. Although the production of other food and hit crops is mainly for subsistence, there is significant potential for industrial processing.

5.8 Fishing remains a strategic income generating activity in the economy. However, the fishing sector in Guinea-Bissau is very underdeveloped. Despite this, fishing serves as an extremely important livelihood and particularly coping strategy for- most coastaVriver villages. The country has large and valuable fish resources. Especially the inshore resources remain largely untapped and traditional fishing by the local population has a strong export potential. The fishery sector represents a big production potential making use of the country’s coastal line extending 274 km and a continental platform of 53,000 km2. An inventory study in 1998 indicated a total fishery biomass of 1.3 million tons, which represents an annual potential harvest of almost 300,000 tons. Traditional fishing catches have in recent years ranged between 20,000-25,000 tons per year. The later is reported to yield a potential of more than 50,000 tons. The country’s annual consumption offish is 25-28 Kgper capita. Current fish production is estimated to be just fewer than 70,000 tons of which 26,460 tons are derived from subsistence fishing.88The fishery sector represents an important opportunity for employment for the poor. Effective employment in this sector is less than 5,000 jobs, although much of the population in rural coastal areas is involved with traditional fishing, processing and trading. The sector is reported to have generated a minimum level of direct employment, as it engages a mere 609 industrial and about 2,000 traditional fishermen. Those involved with commercialization were also recorded modestly in 2002 (103 female and 37 male- intermediaries). According to the Ministry of Fisheries, revenues collected from (off- shore) fishing licenses from a range of large, mainly industrialized countries (EU, China, Italy, Japan), currently contributes about half of Government revenues. Given the heavy reliance on off-shore fishing licenses, issues of transparency of government fishing revenue flows, as well as mechanisms of financial and public accountability, are of great importance in order to avoid elite capture and further instability associated with the use of those funds.

’’ Guinea-Bissau was a net exporter of locally produced rice up to 1961. During 1940-1960 average rice exports reached 5,000 metric tons, and fell below 1,000 metric tons starting in the early 1960s. ’’ Most recent data are from 2003. Provided by ICAF’, Ministry of Fishing, in INEC (2005).

- 73 - 5.9 The industrial fishing fleet is authorized to fish anywhere in the country's Exclusive Economic Zone (EEZ), at least 12 miles off the coast, as artisanal fisheries have exclusive rights within the 12 mile zone. In 2000, 60 percent of the legal fishing fleet constituted chartered vessels, 17.5 percent from the EU, and 7.5 percent from China. Since then, the chartered fleet has declined dramatically as license fees have increased four-fold (to CFAF140,000/GRT), from 85 in 2000, to 57 in 2001, to 4 in 2002. There are indications that the number of licensed vessels have increased recently however to address the country's acute financial problems. Given the almost complete lack of surveillance and enforcement, it is doubtful that all formerly licensed chartered vessels have left the country's waters. Some 80 percent of the so-called "local" industrial fleet (apparently many foreign companies have made deals with local companies to fish under the Guinea Bissau flag) and foreign licensed vessels landed their catches in Dakar; a much smaller percentage in Abidjan and Las Palmas. Fish landings of foreign industrial vessels in Bissau are still rare, and mostly limited to those of Chinese vessels. Demand for and prices of fish in Guinea-Bissau are low in comparison to Senegal, in part because of a lack of fish processing capacity and because of price policies that depress prices for some industrially caught species in the local market. In addition, poor sector services, red tape and corruption and the absence of a standard quality control system that satisfies major importer markets, undermine efforts to bring fishing "on-shore" in Guinea-Bissau.

5.10 There is little tradition or culture of involvement in marine fisheries as exists in Senegal or Ghana. Artisanal fisheries in Guinea-Bissau are concentrated along the coast, islands, and rivers, and the catch found here constitute roughly 80 percent ofthe artisanal catch over two periods of the year: April to May (also coinciding with the drylhungry season), and from October to December. Currently about 3,500 artisanal fishers are reported to be operating in the country, ofwhich at least 50 percent are from Senegal and Guinea-Conakry. The number has been almost halved since the war, when it is thought that many ofthe foreign artisanal fishers abandoned Guinea-Bissau. Most ofthe artisanal catch is taken by foreign artisanal fishers and likely transported to neighboring countries. Only about 10 percent or so is reported to be taken by domestic fishers. This small- proportion is attributed to the fact that fishing in Guinea-Bissau has traditionally been a marginal activity - mainly to complement agricultural activities, or as a coping strategy - practiced some 4-6 days a month, over 8-10 months, for a total of less than 50 days a year. According to the FA0 Fishery Country Profile (2001), in 1996 fishing was the main activity of approximately 18 percent of the total numbers of fishers while agriculture occupied 78 percent. These fishers typically use small, traditional canoes and hand-lines or small gillnets.

5.11 As compared with industrial fishing, however, the bulk of the domestic sector benefits accruing to the private sector originate from artisanal fisheries, though primarily from local rather than foreign artisanal fishing. The Strategy Note states that the share of value added of gross production value is far higher in artisanal fisheries (75-90 percent) than in industrial fisheries (35-50 percent), and local processing and marketing create additional value added. Nevertheless, the local market for fish in Guinea-Bissau is not well developed yet and prices are generally low; virtually no export-oriented processing and marketing facilities exist that could offer higher prices, as exist in Senegal. Locsll' demand for high-value fish is limited as purchasing power is low.

- 74 - 5.12 Livestock, forestry and environment. Livestock represents about 32 percent of agricultural GDP and is produced mainly in the Eastern part of the country.89However, the country is a net importer of fresh meat, mainly poultry. Forestry covers 2 million hectares (5 percent of the country) with diverse ecosystems. The forest is however gradually being destroyed due to abusive use oftrees for domestic charcoal. The coastal zone of Guinea Bissau is internationally recognized for the richness ofits biodiversity as well by the number and the variety of birds. The country’s mangroves are the main reproduction site ofspecies various fish, shrimp and mammals.

5.13 Growth in agriculture is constrained by the lack of investments. As in the other economic sectors, public investments in agriculture have stalled in recent years due the reallocation of public resources to defense and security programs. Similarly, private investments suffered from. the lack of adequate micro financing mechanisms needed for boosting productivity.

5.14 Cashew production is not Figure 5.1 Trends in Cashew Prices, 1979-03 diversified and 1600 therefore yield little 1400 ! 0 value added and g 1200 remains vulnerable to 1000 price fluctuations. z.! Guinea-Bissau “l 800 produces and exports P raw cashew. This situation places the ; 11, sector in a vulnerable $ 200 condition vis-A-vis 50 price fluctuations on 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 the international markets (see Figure 5.1). Moreover, the agriculture sector would generate more value added if cashew nuts were processed and conditioned before being exported. Another main constraint facing the cashew sector is the absence of research and development, and vulgarization activities. In Guinea-Bissau, cashew trees are grown naturally without using advanced scientific methods, while in other countries research programs are carried out in order to improve the size ofthe nuts and their capacity to resist deadly diseases such as the Fungi Anthracnose and Oidium Anacardii, which have already affected plantations in Mozambique. So far, Guinea-Bissau has been fortunate as these infections have not spread in the country. There is evidence to suggest, however, that Guinea-Bissau may be increasingly affected by such environmental and climactic risks. During the 2005 harvest, the prices for cashew grown in Guinea-Bissau fell sharply due to the discovery of parasites attacking the country’s cashew trees. If these price fluctuations are passed on to the small-producers, they could have substantial poverty inducing effects.

89 The livestock is composed of 410,000 bovine; 240,000 sheep; 225,000 caprine and 67,000 porcine and 473,000 fowls.

- 75 - 5.15 The absence of a coherent rice-cashew strategy represents a major threat to rice production and food security. The farmers are naturally inclined to engage in commercial crop production of more guaranteed value. This is part of the reason why domestic rice production is stagnant in return for increased cashew production^.^^ The rapid expansion of cashew production -- although good for the economy -- represents a menace to the production of rice, the main staple food. Small farmers have increasingly adopted cashew plantation as a cash earning strategy. Consequently, subsistence production ofrice has been replaced by rice purchase enabled by cashew production and its barter for rice.g1 The advantage ofthis strategy is that it allows farmers to reduce their production cost, especially in terms of labor inputs for a household (high for rice compared to cashew). The other major disincentive for rice production is the pricing mechanisms. Farmers have been sensitive to the prices of rice on domestic market, on which local rice variety is not praised favorably enough to give incentives to engage surplus production beyond what would be necessary for subsistence needs for a family. Locally produced rice sometimes is hard to sell. The strategy of replacing subsistence crop with cash crop involves a major food security risk. Farmers need to survive within what rice harvest permits as it completes in September-October in the small valleys, or October-November in mangroves. They have to wait for the cashew nuts to ripen for harvest in MarcWApril through June when barter transaction with rice is possible. Farmers' subsistence rice stock often runs out before such transaction is possible, continuing to present a food security risk. In this context food insecurity is a major concern for the majority ofthe population.

5.16 Access to financial means remains limited for smallholders in agriculture. Microfinance is ill-developed and there is only one commercial bank in Guinea-Bissau and virtually all banking activities are concentrated in the capital city Bissau. Commercial loans are limited to a few small enterprises in the form of short-term agricultural working capital (crkdit de campagne). Access to credit is generally limited for the majority of the population due to lack of loan collateral. Improving access to microfinance credit for smallholders in the fast growing cashew sector, for example, will facilitate the diversification ofthe sector and contribute to job creation.

5.17 Lack of appropriate technical assistance for smallholders is a severe constraint to improved agricultural production and diversification. Technical assistance to small poor farmers is almost inexistent due to institutional constraints and lack of resources. Expanding the potential of the smallholder sector would require technical assistance support for the development oftheir production capacity.

5.18 The lack of rural road infrastructure represents a major constraint for the development of agriculture. Much of Guinea-Bissau's infrastructure was destroyed- during the 1998 conflict. The lack ofroad infrastructure in rural areas limits the access of rural production to local and regional markets, and often contributes to increase agricultural transaction costs.

Almeida 2005. 9' 71 percent of the households surveyed in the 2005 QS reported that they bartered cashew for rice, while 37 percent stated that they sold cashew in order to purchase rice. 3.

- 76 - B. TRANSACTIONCOSTS AND POVERTY REDUCTION IN THE CASHEW NUTSECTOR

5.19 Over the past three decades, cashew nut has had a significant poverty Figure 5.2: Cashew Yield and Poverty, reducing effect in Guinea-Bissau. 1992-04 Episodes of good cashew performance are often associated with higher per capita incomes and better economic performance (see Figure 5.2). However, there are significant transaction costs, which contribute to reduce the farm gate price for producers. These transaction costs generally derive from the absence or the lack of adequate institutional and regulatory arrangements, and high 1992 1994 1996 1998 2000 2002 2004 domestic and international transportation costs (due to poor rural Poverty rate (%), left axis road infrastructure and the small - -0- - Cashew yield (Kg/Ha), right axis number of freight ~ompanies).~~The discussion in this section will: (i)review the main determinants of transaction costs in the cashew sector; (ii)assess their impact on poverty reduction; and (iii)discuss the issue of sustaining poverty reduction and growth in agriculture.93

1. Analyzing the Main Determinants of High Transaction Cost in the Cashew Sector

5.20 The cost of doing business in the cashew sector remains relatively high. As mentioned in the first two chapters, instability and weak governance remain major disincentives to domestic and foreign investments necessary for the development of the cashew sector. There are numerous risks associated with changing regulations, unclear and cumbersome administrative requirements for existing and new businesses, dysfunctional legal enforcement and inadequate utilities and seaport systems. Addressing these issues would require improving the investment climate (simplifying and clarifying relevant laws and taxes, streamlining investment and export-related administrative procedures, rehabilitating the physical and administrative infrastructure ofthe port). Laws concerning cashew nut trading have changed often in recent history, including two changes in 2000 and another in 2001. The laws themselves are unclear, or the traders believe that enforcement may not always correspond to what the laws say. For example, it is currently not clear whether or not foreign enterprises are permitted to buy raw cashew nuts within the country as intermediaries or exporters. Not only does this reduce

92 The country does not have a railway system, and possesses only 10 percent of paved road for a network of 4,400 km. The road network in rural areas is generally unpaved and in bad condition. The country has a large river transportation network yet to be developed. The cost of freight to India for example is significantly higher in Guinea-Bissau than from other West African ports. A 20 feet container shipped from Bissau to India was quoted as US$1,400 in June 2004(rising to US$1,700 by late June) compared with US$600 from Abidjan. The higher costs are due to to shipping line charges for calling at Bissau port and the absence of competition in Guinea-Bissau’s freight market. 93 Sub-section 1 and 3 are based on a recent study on the cashew sector in Guinea-Bissau (Jaeger and Lynn, 2004).

- 77 - the competition for the nuts but the uncertainty is an unambiguous danger signal to investors, and needs to be cleared up. When enforcement is believed to be contrary to the law, there must be recourse through the courts, which requires an effective judicial. sys tem,

5.21 Bureaucratic regulations often lead to higher transaction costs, which translate into lower farm gate price for farmers. The farm gate price of the nuts was between 60-70 percent of the FOB value of the raw cashew nuts in 2004. The price doubled during the first weeks ofthe season and then fell back as the export buying price had moved sharply higher and then dropped again -- due to increased speculation on the side of the Indian exporters. While the motives for such speculation remain uncertain, high transport and excessive export costs associated with bureaucratic regulations and a particularly higher export tax on cashew (6 percent) compared to other exports (2 percent) are believed to be the main determinants. In this context, exporters not only reflect incurred transaction costs in the fann gate price during the season (mainly at the end when they have already constituted sufficient stocks), but they also reflect them in the reference price for the upcoming season. Subsequently, as international prices are exogenous, any adverse price fluctuation on the world markets would translate into a lower farm gate price for cashew producers (because exporters would respond to the price decrease by offering lower prices to farmers).

5.22 The absence of coordination among producers and the lack of supervision on the Government side also contributed to increase transaction costs for cashew producers. Currently, there is no private organization that encourages and sustains collaborative actions between all parties involved in the cashew chain. Since the primary beneficiaries of most developments in the cashew sector are private agents (farmers, processors, traders, workers, and marketers), it stands to reason that these private entities should have an interest in promoting the development ofthe sector. Private stakeholders in the cashew sector need to organize themselves under a structure that they create to promote their mutual interests (see Box 5.1).94 The composition and leadership of this structure should be chosen entirely by its private constituents, as should its agenda. Among other tasks the organization would: (i)formulate a strategic vision for the sector and liaise with the government in pursuit of this strategy; (ii)coordinate and provide logistical and informational support to various partner organizations involved in training, credit, research, extension, and other services; (iii)explore the possibility of deals between producers and buyers/exporters to pre-finance the cashew campaign; (iv) set benchmarks and norms for cashew production to protect the local production against competing producers and processors; (v) investigate opportunities for collaborative marketing through for example trade shows; (vi) collect and disseminate information on current international raw nut and kernel prices and trade data; and (vii) promote a Code ofPractice within the cashew sector.

5.23 There is, at the same time, a need for the Government to set up a small office- dedicated specifically to policies related to cashews. This office should address issues

94 During a recent workshop on the cashew sector, it was agreed that the Organization of Cashew Stakeholders (Clmara do Caju) would seek to promote collaboration within the cashew chain and support its members through its services.

- 78 - such as: (i)setting up a free zone or cashew park; (ii)providing tax and administrative incentives to investors; (iii)creating a consolidated administrative bureau for investors, covering corporate establishment, importation of equipment and materials, residency for expatriate personnel, and other regulations imposed on new and existing enterprises; (iv) creating a consolidated certification bureau for exports, including phytosanitary certification, certificates of origin, and other necessary clearances; (v) monitoring and quality control of cashew kernel exports, including laboratory testing and certification necessary for shipment; (vi) providing key information to potential investors; and (vii) applying existing research to improve the quality ofthe nuts.

5.24 Finally, inadequate taxation may lead to lower return for farmers.95Exports of cashew nuts are currently taxed at the rate of 6 percent (since 1989 this rate has been gradually reduced from 34 percent, contributing to reduing the effective tax rate on cashew by 52 percent). The reduction was intended to boost raw cashew export.' However, the authorities may decide in the future to promote cashew processing by raising rates in order to discourage raw cashew exports. Although the overall net effect of such policy may be positive in the long run, it is likely to have a disproportionate effect on rural incomes in the short-term (as kernel processing will take years to fully develop). It has been estimated that a 10 percent tax rise on raw cashew export, for example, would take about US$5 million away from farm gate earning^.'^ A 10 percent tax increase at FOB level is equivalent to a rise of at least 14 percent at farmer level. The international sales price will not change; rather the buying price from the farmer will be reduced to' maintain margins in the chain. In the absence of a processing capacity, a tax increase would penalize farmers and contribute to worsen poverty.

2. Assessing the Poverty Reducing Impact of Lower Transaction Costs in the Cashew Sector

5.25 Cashew nut production represents an important activity for the poor. Based on data of the 2005 QS presented in Table 5.1 below, slightly more than a third of the- households (37.4 percent) have positive sales, with a higher proportion in the bottom tercile (46.5 percent) than in the top tercile (26.1 percent). The proportion ofhouseholds with positive cashew sales reaches close to half among those household who declare having land, with smaller differences by tercile. While poorer households are more likely to produce and sell cashew, the total value of the sales is higher among wealthier households when the sample is restricted to those who sell cashew. However, the differences are not very large in the sample, ranging from an average sale value of about CFAF121,OOO in the bottom tercile versus 176,000 in the top tercile. This lack of differences may be due to the small scale of the survey, in which large producers (ponteiros) are not likely to be well represented. Said differently, the sample is likely to represent the situation of relatively small producers only, rather than the situation of all producers in the country. Another important result is the fact that barter sales (exchange of cashew for rice) are most common, especially among the poorest tercile, where they represent 55.9 percent of all transactions at the end of the season (and 73 percent of all

95 Exports of cashew nuts are taxed at the rate of 6% (since 1989 this rate has been gradually reduced from 34%). This measure contributed to reduce the effective tax rate on cashew by 52 percent. 96 If 80,000 tons are exported at US$650 per ton, IOpercent tax rise equates to US$5 million

- 79 - transactions year-through). In the top quintile, barter accounts for 39.3 percent of sales at the end of the season (59.4 percent year-through). This is probably because wealthier households have other means to purchase rice, and prefer to sell their production in cash rather than in kind.

Table 5.1: Cashew Sales by Tercile in 2005 Tercile Area Total 1 2 3 Urban Rural Positive sale (%) 46.5 39.4 26.1 28.9 42.8 37.4 Positive sale among those with land (%) 50.4 47.0 43.6 50.6 46.4 47.6 Sales value when positive (CFAF) 120,924 147,540 176,189 170,292 131,332 143,020 Time of sale Beginning of season 1.3% 4.3% 3.1% 2.5% 3.1% 2.9% Mid-point of season 3.1% 2.6% 1.5% 2.2% 2.6% 2.5% End of season 22.7% 23.9% 36.0% 29.3% 25.6% 26.9% Time ofbarter Beginning of season 5.3% 10.6% 12.2% 11.8% 7.7% 9.2% Mid-point of season 11.8% 7.1% 7.9% 7.1% 10.0% 9.0% End of season 55.9% 51.5% 39.3% 47.1% 5 1.O% 49.6% Source: Staff Estimation from 2005 QS.

5.26 Lower cashew transaction costs are likely to increase farm gate prices and alleviate poverty. Empirical evidence worldwide suggests that improved access to markets through improved business environment, better institutional, regulatory and tax arrangements, and adequate rural infrastructure may have a strong potential for reducing poverty. In the case of Guinea-Bissau, a combination of data from the 2002 ILAP survey and the 2005 QS is used to perform simulations regarding the impact of cashew policies on poverty and inequality under a certain number of assumptions. The first simulation assesses the expected impact on poverty of an increase in the prices paid to producers thanks, for example, to a higher price obtained by exporters or through a reduction in transaction and transport costs. The second simulation estimates the expected effect on the poor of a reduction or an increase in the tax to be paid on cashew exports (a full account of the simulation methodology is provided in the sixth chapter ofVolume 2).

5.27 Impact of changes in the farm-gate prices of cashew. In Guinea-Bissau, although farmers may receive a relatively large share ofthe export price, the latter itself is low due in part to the unavailability of financing in the country. As exporters must rely on. financing provided by Indian firms, on strict terms, farmers ultimately suffer from low farm gate prices. A 15 percent increase in farm-gate prices for cashew nuts could result in an increase in consumption for the extreme poor of 9.5 percent, and 3.3 percent for the poor. The results for the extreme poor are likely to be overestimated because the estimated share oftotal per capita consumption accounted for by cashew nuts is large. On the other hand, the very poor who are not cashew producers themselves, but work on farms as day laborers, could benefit from trickle down effects which are not accounted for here. Moreover, taking into account the measures of poverty and extreme poverty obtained in the 2002 ILAP survey, and estimates ofthe elasticity ofpoverty reduction to

- 80 - an increase in consumption (which are preliminary at this stage), the results indicate that among cashew producers, extreme poverty could be reduced by 3.14 percentage points and poverty by 1.81 percentage points under the umbrella of a 15 percent increase in cashew prices. These are substantial impacts that result from the important role that cashew revenues play among those households who produce and sell cashew. Finally, in order to assess the potential poverty impact on poverty in the country as a whole, one must take into account the fact that not all the poor or extreme poor are cashew producers. Even though only about half of the households are likely to be involved in cashew production and sales according to the 2005 survey, the net impact on national poverty remains fairly large, with anticipated reductions in the share of the poor or extreme poor of 0.78 and 1.46 percentage points respectively. However, these estimates do not take into account potential derived impacts through wage increases for cashew farm workers.

5.28 Poverty impact of taxing cashew exports. An increase in the export tax on unprocessed cashew nuts of about 10 percent could potentially (with full pass through to farmers) generate a reduction in their farm-gate price of close to 15 percent (this assumes that farmers get 70 percent ofthe export price). In other words, this would have the same but reverse effect on poverty as the estimates in Table 5.2 (poverty could increase by the amount estimated in Table 5.2). Alternatively, if the 6 percent export tax were eliminated, and if we were again to assume full pass through to farmers, farmers could benefit from an increase in consumption of 8.6 percent, which would have poverty impacts of about half the estimates presented in Table 5.2.

5.29 There are noticeable second round poverty reduction effects of cashew on other non agricultural activities. Econometric estimates using the 2002 ILAP datag7 indicates that 1 percentage point increase in households’ agricultural income would lead to about 0.4 percentage point increase in their non-agricultural income (see Table 5.3). The labor market is the main transmission channel of these spillover effects: cashew farming is a labor intensive activity which employs mostly unskilled laborers. Although these labor activities take place only during the cashew season (April through June), they represent an important source of income which in turn serves to jumpstart small scale‘ trading activities in numerous non-agricultural informal activities. Perhaps that is the reason why some argue that “cashew nut is to Guinea-Bissau what petroleum is to Saudi Arabia.”

97 OLS regression of the logarithm of households’ agricultural income on the logarithm of their non agricultural income and geographical location. i“

- 81 - Table 5.2: Impact of change in cashew prices on poverty, Guinea-Bissau Assumptions Impact Impact on poverty on extreme poverty Value of household cashew production (CFAF) 120,924 134,232 Household size (number of equivalent adults) 7.6 6.6 Per capita value of cashew production (CFAF) 15,911 20,338 Mean consumptiodequivalent adult among the poor (CFAF) 25152 9 1895 Gain in consumption with 15% rise in cashew price (CFAF) 2387 305 1 Headcount index of poverty/extreme poverty (%) 21.6 65.7 Elasticity ofpoverty reduction to consumption growth -1.53 -0.83 Change in consumption with 15% rise in cashew price (“A) 9.5% 3.3% Change in poverty among cashew producers (YO) -14.5% -2.8% Percentage point change in poverty among cashew producers -3.14 -1.81 Share of households producing cashew nuts (%) 46.5% 43.0% Impact on poverty/extreme poverty among all households -1.46 -0.78 Source: Staff Estimates from 2002 ILAF’ and 2005 surveys.

Table 5.3: Stochastic Relationship Between Households’ Non-agricultural and Agricultural Incomes Dependant variable: Log Per capita non-agricultural Coefficients income

Explanatory variables:

Log Per capita agricultural income 0.40 (4.91) Rural dummy -0.90 (-3.06) Constant 6.90 (7.99)

Source: Staff Estimates based on the 2002 household survey data. Figures in parentheses are the t-statistics at 95 percent confidence.

- 82 - Box 5.1: Improving the Business Environment for the Development of the Cashew Chain Prospective investors need to know the costs of establishing and operating a business (operating licenses, fees of various types, taxes on personal and corporate income, tariffs on goods and services, etc.) in order to make the decision to invest. The government, through the Department for the Promotion of Private Investment (Departamento pela PromoCBo do Investimento Privado, DPIP), should collect and publish all regulations, taxes, and fees relevant to foreign investment in general, and the cashew sector in particular. The information should preferably be made available in the form of a booklet published in Portuguese, French and English. Certain fees need to be simplified: for example, the transport of cashew nuts from the interior requires a permit valid for one month. The time limit adds to the unit cost of transport for every day that the lory is not carrying goods, possibly through mechanical breakdown, but it also stops smaller operators who might not transport cashew regularly. The permit adds to the administrative cost in requiring renewal during the season. This type of indirect tax is ultimately paid by the farmer through a lowered farm gate price. The government should consider tax incentives for new cashew enterprises. This last task might best be performed by a special government commission on cashew affairs, in collaboration with DPIP and the Organization of Cashew Stakeholders (CBmara do Caju). While simplifylng taxes and fees relating to industrial investment, the Government should simultaneously consolidate administrative requirements for corporate establishment, importation of equipment and materials, residency for expatriate personnel, and other regulations imposed on new and existing enterprises. Rather than having to submit paperwork to the MOF, Ministry of Economy, the Ministry of Labor, the Ministry of Commerce, the Ministry of Foreign Affairs, and other government offices, there should be one office where indusmal companies can accomplish the majority of their required administrative procedures at a single address. Such simplicity would be a strong incentive to investment and a sign that the Government is serious about creating favorable conditions for investors. When exporting food products such as cashew kernels, enterprises are required to have phytosanitary certificates, certificates of origin, and quality control certification. Certain importing counties may require others. There should be one office empowered to issue all of these, at a reasonable fee. The Republic of Guinea, for example, already has such an ofice in Conakry, entitled the Support Center for Export. Formalities. It is in the country’s own interest to facilitate exports and minimize obstacles, in order to generate external revenue. As the Government is already aware, there is an urgent need to attract international banks to Guinea-Bissau to furnish services such as secure depositories for business income, sources of cash for day-to-day operations, international transfers, currency conversion services, guarantees of liquidity to cover commitments, and trade finance. Aside from these practical considerations, the absence of internationally-recognized banks in the country is an unmistakable sign of instability and of additional overhead costs to potential investors. The Government needs to continue pursuing the location of new banks in the country. The current arrangement at the port, both physically and administratively are not satisfactory. The port represents an expensive bottleneck in the swift flow of the cashew nuts. Various intentions have been discussed for the port but a radical upgrade is now overdue. A USAID - TIPS report in 1997 highlighted six problem areas: (i)navigation difficulties (access to the port, dredging and maps); (ii) operational inefficiencies; (iii)lack of port equipment; (iv) insufficient port infrastructure (depth, mooring equipment, transit warehouse etc.); (v) cargo losses; and (vi) high direct and indirect freight costs. These problems should be addressed as a matter of some urgency. It is not clear how much progress has been made. Since the TIPS report the port has come under complete government control with GUIPORT (49% Portuguese owned) who managed the port, replaced in 1999 by a government agency APGB. A task force is required to thoroughly appraise the port with a view to developing this facility not only for the benefit of cashew exports but also for a diversified portfolio of trade. This task force should include representatives of exporters as well. As a means of insuring the safety of investments and the respect for contracts and an impartial enforcement of laws, the judicial system is another key element of the enabling environment for attracting investment. Unfortunately the judicial system in Guinea-Bissau is widely considered to be nonfunctional and subject to conuption. One foreign business operator said bluntly that he would sooner “kiss his money goodbye” than waste his time and more money trying to settle a dispute through the courts. This is the clearest possible indication that a reassuring legal environment for investment does not currently exist in’ Guinea-Bissau. Reform of the judicial system is a complex and involved process; reform programs have been funded in the past by donors, but evidently the desired effect has yet to be achieved. Special emphasis should be paid to the enforcement of civil law as it relates to contracts, torts, and property law. Employers should also be protected from the opportunistic tendency of employees to lodge baseless complaints and damage claims against them, when leaving a job for whatever reason. Labor laws adapted to the cashew processing industry will be a key element looked for by investors. The payment structure for cashew processing workers needs to be based on productivity, rather than on attendance. The Labor Law (Lei de Trabalho) should therefore be amended as necessary to allow for employment paid on the basis of productivity. Thus the law needs to contain the flexibility necessary for cashew processing employers to hire, reassign, or dismiss employees in ways that they deem necessary for the efficiency of the factories. Application to the cashew processing industry of traditional labor laws for fixed-wage workers and permanent employees would cause the factories to close due to non- profitability. However, workers must be protected from exploitation: piece rates should be fixed at levels that allow those with average efficiency to earn wages comparable to jobs of a similar skill or educational level that are paid at an hourly, daily, or monthly rate. If workers are assigned to tasks that are inherently less productive, such as shelling smaller nuts, the piece rate must be raised in compensation. If labor market forces are not sufficient to prevent exploitative practices, then the Organization of Cashew Producers (Cdmara do Cuju )should provide a forum for negotiation. Source: Jeager and Lynn (2004).

- 83 - c. SUSTAINING POVERTY REDUCTION AND GROWTHIN AGRICULTURE

5.30 The monoculture of raw cashew poses a serious threat to long-term agricultural development. Guinea-Bissau produces mainly raw cashew, which provide' less value added and remains highly vulnerable to exogenous climate and price hazards. Over the past three decades, the rapid growth of the cashew sector has contributed to suppress rice production for which the country was a net exporter prior to the introduction of cashew in the early seventies. Such expansion is occurring spontaneously due to higher returns to cashew farming relative to rice. Although rice is the main staple food in the country, high production costs contributed to promote the production of cashew nuts. As a result, farmers produced cashew in order to sell it or to exchange it for rice.98 The barter between cashew and rice is an increased phenomenon as rice production- has been stagnant or sometimes abandoned unable to satisfy domestic consumption needs. Such a trend has a significant implication in relation to food security, in case cashew production is adversely affected by market and natural forces. Cashew production is an easy process. Virtually no labor is required until nut harvesting after planting. This is also providing incentives to grow cashew as a cash earning strategy instead of growing rice. This means new areas are increasingly converted into cashew plantations, especially in the eastern and southern part of the country. Most of such newly established plantations have yet to mature for hit-bearing, so the production peak is expected to occur sometime in fbture years.

5.3 1 Diversifying agricultural production can spur growth and address the issue of food security. In the case of Guinea-Bissau, agricultural diversification can be achieved through moving towards more value added in cashew nuts, and taping the country's potential in the production of wide ranging tropical fruits. The processing of cashew will contribute to increase agricultural value added and create more jobs. In view of the failure of larger processing units in both Guinea-Bissau and other countries, the focus should be on promoting small labor intensive units as experienced in the mid 1990s under the auspices of the United States Agency for International Development's (US AID). Trade and Investment Promotion Support (TIPS) project (local processing -- roasting, drying, shelling, and packaging cashews for export -- can quadruple producer's earnings). Guinea-Bissau could potentially have a sizeable small-scale processing industry that would create thousands ofjobs and earn a great deal of value added that is currently lost by not processing raw cashew. However, as mentioned earlier, the development of the cashew processing industry depends on the ability to attract investment, of which most will have to come from outside the country. In addition to improving the general investment climate, described above, other factors that can contribute in attracting prospective investors include: (i)the availability of financing means through loans made available to investors on favorable terms for use in setting up cashew processing factories (the terms could include subsidized interest rates, deferred repayment, and reduced requirements for loan collateral); (ii)increasing kernel marketing

'* Terms of trade for bartering cashew for rice may deteriorate as more cashew nuts are needed to secure necessary food supply. Cashew Nut is vital as a source of cash income, and indirectly as food, for the country's peasant farmers. They sell the raw nuts to traders for up to 400 CFA francs (80 US cents) per kg, but often they simply barter a 50 kg bag of cashew nuts directly for a 50 kg of rice. The cashew-rice barter system was introduced in the country in the eighties as a mechanism for enhancing food security in the country. The system has instead contributed to substitute rice imports to domestic production.

- 84 - by participating in international food and nut trade fairs to exhibit the country’s products; and (iv) the establishment of a uniform quality control system for exported kernels. Finally, given the country’s tropical climate, export of mango and other fruits seems to offer opportunities that could be easier to cash in.

Box 5.2: TIPS Initiative

The TIPS initiative contributed to increasing the level of broad-based, market-oriented economic growth by assisting Guineans to remove barriers to that growth and stimulate private-sector trade and investment.

TIPS has assisted the government and the private sector in identifying and implementing needed changes in national laws and policies which have restricted economic growth, helping them to work together better toward a more responsive system of govemment. The project has also provided support to a nascent judicial system, helping to establish national small claims courts and strengthen the independence of the upper level judiciary fiom political influence, providing a framework for the rule oflaw to replace a legacy ofarbitrary legal decisions.

‘Procaju’ -- an association of cashew processors who graduated from the Center and are seeking to expand their markets -- is the result of work by TIPS with the private sector. TIPS has always stressed the need to assist organizations and associations to achieve their targets, avoiding top-down, paternalistic “solutions” that collapse once the aid is gone. Using this philosophy, TIPS helped ANAG, National Association ofAgricultural Producers (representing primarily small farmers), and the National Association of Business Women (MAE, representing over 10,000 women), to establish the^ Quinhamel Training Center in 1994. The Center has introduced cottage-industry cashew processing to Guinea-Bissau, having trained more than 400 people in processing and business skills. It has also spawned the inauguration of five more training centers elsewhere in the country, using graduates of the Quinhamel Center to train the staff and teachers for these new operations.

m:Steele Communications Fttp://www.steele.com/cashau/tips.html)

D. CONCLUSION AND POLICY CONSIDERATIONS

5.32 Despite Guinea-Bissau’s large potential in agriculture, there are high transaction costs stemming from administrative barriers, weak infrastructure and lack of diversification. For the rural sector to create labor intensive jobs, policy makers would need to adopt a coherent agrarian development strategy in the context ofthe PRSP, which would aim at promoting: (i)the processing of raw cashew into exportable cashew kernels; and (ii)the export oftropical fruits. This is particularly important at this moment as the Government is readying to implement major PSR. These reforms will contribute to reviving the private sector by creating much needed new jobs to absorb part of the labor force to be discharged fiom the public payroll.

- 85 - 6. MAKING BASIC EDUCATION AND HEALTHCARE MORE AFFORDABLE FOR THE POOR

The adoption of the Dakar Framework for Action, EFA by Guinea-Bissau in 2000 contributed to improving enrollment rates in basic education. Analysis of survival rates (SR) in the primay, however, suggests that the majority of pupils are likely to drop out Ji-om school before completing the primary education if appropriate actions are not taken to keep them in the schooling system. These actions could include: (5, the provision of monetary incentives to parents: (io, the construction of schools offering a six-year primary education cycle within walking distance of dwelling areas; and (iii) actions aimed at improving education quality. Although Guinea-Bissau has been implementing the BI since the early 1990s, patients have been paying high cost for healthcare. The introduction of a UF system in the early 2000s contributed to improve drug availability and outgoing visit services in health facilities. UF have also contributed to improve child immunization, which in turn significantly reduced infant mortality rates. At this time of diflcultfiscal situation for the Government, the sustainability of the EFA and UF initiatives highly depends on the continued support of the international community.

6.1 The massive destruction ofinfrastructure and population displacements caused by the 1998 conflict has considerably weakened Guinea-Bissau’s education and health systems. Bringing social service delivery to the level attained before the civil conflict -- by providing affordable basic education and health service to the poor -- has been the Government’s main challenge since 1999. Although this goal is yet to be fully achieved, the adoption of the EFA and the UF initiatives by Guinea-Bissau in the early 2000s has contributed to substantially improve primary school and the provision of drugs and consultation services in health centers. The objective of this chapter is to analyze the effect ofthese two important policy actions on social service delivery in Guinea-Bissau.

A. EFA AND EDUCATIONSERVICE

6.2 The rising concern with the quality of education, especially in Sub-Saharan African countries was reflected in the protocols of the World Conference on EFA in Jomptien (Thailand) in 1990. These concerns were again reiterated at the World Education Forum in Dakar (April, 2000), where the international community observed that ten years after Jomptien, not much progress had been made, especially in the quality of education in Africa. The analysis ofthe educational system in Guinea-Bissau provides a great deal of perspective on some of the fundamental factors affecting the quality of education in Africa. In view of achieving the Millennium Development Goals (MDGs) for education, Guinea-Bissau adopted the Dakar Framework for Action in 2000 and presented to the international community in early 2003, its EFA National Plan ofAction. This section seeks to analyze the effects of the EFA initiative on education service delivery in the public sector. The analysis focuses on the following points: (i)history of EFA; (ii)implementation ofEFA; (iii)impact ofEFA on service delivery and utilization; and (iv) conclusion and recommendations.

- 86 - 1. Situation of the Education Sector

6.3 Guinea-Bissau is characterized by low adult literacy rates with striking differences based on gender, the area of Figure 6.1: Literacy Status of the Heads of residency and other socio-economic Households in 2002 considerations. There are large disparities between the literacy status of residents of loo80 1 Bissau (urban) and residents of the rest of 60 the country (rural) as well as between male 40 and female heads ofhouseholds. About 80 20 percent of female household heads are 0 illiterate. In the region of Bissau, for example, 1 out of every 5 household heads has no education at all, and only two heads Not Literate Literate out of five have an education level beyond the completion of primary school. In the rest of the country, 70 percent of household heads have no education and only about 1 out of 15 has gone beyond primary schooling (see Figure 6.1, across). It is also clear from the data that the proportion of household heads with no level of education increases as the welfare status of the household decreases. In the poorest quintile, for example, about 70 percent of households have no level compared to 47 percent in the richest quintile (Box 6.1 below provides information about the structure ofthe education system in Guinea-Bissau).

6.4 Enrollment in primary education is influenced by factors such as location, Figure 6.2: Access to Primary gender, and economic status. Current Education by Region educational efforts are best captured by the responses given with regard to accessg9to 100 educational services and enrollment rates. 80 About 80 percent of households in Guinea-Bissau are located 30 minutes or less from a primary school, with 96 percent for Bissau and 79 percent on average for the rest of the country (see Figure 6.2 across). Improved access to primary education may help explain enrollment patterns in Bissau compared to Region other regions. Primary schools are accessible to more households in Bissau than in other regions. As a result, in 2000 the gross primary enrollment rate in Bissau was estimated at 102 percent compared to the national average of 84 percent."' Another important factor which seems to have had a positive influence on enrollments is the level ofsatisfaction with the quality ofeducation.

99 Access to primary school is defined for households with a primary school less than 30 minutes away.

IwEducation data collected by the Ministry of Education on the 200-01 school year. ,-

- 87 - There is clearly a relationship between satisfaction, usage and enrollment rates as shown in Figure 6.4 below: enrollment tends to be higher in regions where people are more satisfied with the education system (Bissau and Bafata).

Box 6.1: Structure of the Education System in Guinea-Bissau

Education Chart Flow

I Educational System of Guinea-Bissau I +c Formal Education

Secondary Secondary Tertiary - Education: Education: Education General Professional - and Technical

Secondary: 1st Cycle Vocational and Technical

2nd Cycle

Koranic and Community b Schools

Structure of the Education System

Formal education: The starting point for this track is an optional pre-school education for children between the ages of 3 to 6 years old. Pre-school is typically organized by the private sector and local communities with the Ministry of Education providing support, training and monitoring, especially in the area of curriculum development and teacher qualification. Pre-school is followed by basic education, which takes 6 years to complete. In Guinea-Bissau, the official age for admission into the first year ofprimary school is 7 years. The first four years ofbasic education, called “Enseignement de Base Elkmentaire (EBE)” is compulsory and free (since 2002). Subsequently, children then enter a complementary program - “Enseignment de Base Complkmentaire (EBC)” - which takes two years to complete. ...I.. *

- 88 - Box 6.1 continued

Community schools and koranic schools (Madrassas) are considered as part of basic education. The theoretical age group for children in the primary cycle is seven to twelve years old; but the actual age group of students in this cycle is believed to be 6 to about 18 years old. After six years of primary education, children are admitted to secondary schools which consist of two main cycles. The first cycle provides general education (Enseignement Secondaire General - ESG) for 3 years, after which students are admitted to the second cycle (Enseignement Secondaire ComplCmentaire - ESC) where they study for ' two additional years. Theoretically, students are expected to enter secondary school at the age of thirteen and complete it by the age of seventeen; but most of the students in this cycle are between the ages of 12 to 24, according to officials of the Ministry of Education. Students who choose not to go to traditional secondary schools can attend technical and professional training schools, which are designed to provide technical education in trades like automobile repair, building construction, carpentry, textile and clothing manufacturing, farming and animal husbandry. With the high drop out rate in traditional secondary schools, these institutions are very popular with the youth, a significant majority ofwhom see the training programs offered by these schools as an opportunity to learn a skill that would enable them to make a fast entry into the job market or establish their own businesses. In response to the high demand for technical. schools, the Ministry ofEducation is negotiating with the African Development Bank to establish three of these schools across the country - Cacheu (Bula), Oio (Farim) and Tombali (Bedanda). The creation of these schools and the skills-set that they are expected to offer, is likely to create employment opportunities, especially for the youth, a great majority of whom are currently out of school and unemployed. Finally, the capstone for the formal system is tertiary education, which provides university education and includes other institutes of higher learning such as teacher training colleges, nursing colleges, and physical training schools, among others.

Non-Formal Education: This system is designed mainly to provide literacy programs, principally in local languages, to those who never attended school. As is shown in this report, there is a significantly ' large segment of the population in Guinea-Bissau, which either never attended school or dropped out of school.

The educational system in Guinea-Bissau, which has been in existence since the country obtained independence from Portugal, is considered to have maintained salient features of the colonial system of education. Several attempts were therefore, made in the past to reform the system One of such reforms undertaken in 1989/90, called for a unified system of six classes to be completed in three phases, at the Basic Education Level. Each ofthese phases will consist oftwo classes. Promotion from one class to the other will be automatic but movement between phases is based upon a teacher's assessment of student achievements. This measure was adopted in order to reduce the repetition rate within a phase which is proven to have no pedagogic benefits, but on the contrary, increases the risk of dropping out. Implementation of this reform has been very slow. The curriculum content of the new unified system is not yet finalized. Many schools are still incomplete (i.e.: do not offer a six-grade primary education cycle). Consequently, even though the new system is theoretically considered to be the current system, in reality, the actual current system is a hybrid ofthe old and the new system depending on where the school is located. The failure to have a well-defined, successfully implemented and a nationally uniform system is likely to affect the performance ofthe educational system.

- 89 - 6.5 Gender differences in enrollment rates for primary school are also quite noticeable (see Table 6.1 below). Enrollment for boys in primary school - estimated at 66 percent -- is on average, about nine percentage points higher than that of girls."' In Bissau, about 2 percent of children enrolled in primary school dropped out and about 44- percent never attended. This implies that in the year ofthe survey, more than 46 percent of the children in that age cohort were not enrolled (never attended plus dropout rates). The main reasons reported in the survey for the drop out are issues related to work and marriage. At the national level, about 32 and 29 percent of children are not attending school for reasons that might be associated with child labor and early marriage (see Figure 6.3 and Box 6.2)).

Figure 63: Main Reasons for not Attending School in 2002 Figure 6.4: Access, Usage and Satisfaction with Education 50 1

100 80 -S I -b 60 5 40 i a 20 0 National Female Male Access Enrollment Satisfaction

E! Work Maniage W Bissau W Bafata 0 Oio

Table 6.1: Educational Status of Primary School Students by gender and economic status in 2002 (in percentage) Students aged 7-14 National Female Male Poorest 2 3 4 Wealthiest Enrolled 61.81 57.03 66.13 53.05 56.34 62.38 66.51 73.22 Dropped Out 1.55 1.89 1.24 1.44 1.8 1.89 1.48 1.07 Never Attended 36.64 41.08 32.64 45.52 41.86 35.73 32.01 25.7 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

Table 6.2: Educational Status of Primary School Students by Region in 2002 (in percentage) Biombo/ Quinard National Bissau Bafata Gabu Cacheu Oio Bolama Tombali Enrolled 61.81 86.12 49.73 49.08 71.87 34.82 60.26 60.49 Dropped Out 1.55 1.79 1.82 2.40 1.81 0.47 0.96 1.23 Never Attended 36.64 12.09 48.44 48.52 26.32 64.70 38.78 38.28 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002.

lo'Education data collected by the Ministry of Education on the 2003-04 school year estimated the gender parity in primary education at 1.21 (ratio of boys enrolled in primary education compare to girls). This figure is much higher than the data reported in the Guinea- Bissau Light Household Survey in 2002.

- 90 - Box 6.2: Main Factors Influencing Dropout Rates in Primary Schools

The two main factors that influence the dropout rate in primary schools are work (home or job) and marriage. Both of these factors are gender-based, and in a very subtle way, seem to be motivated by financial considerations.

Work (home job). A significantly higher proportion ofboys drop out of school either to help out in domestic chores or to help the family in income-generating activities. In poor countries such as Guinea-Bissau, most often, households require the participation of all members ofthe nuclear, and sometimes the extended, family for the economic sustenance of the household. Under these circumstances, the decision makers in a household may perceive that schooling for their children, instead of their economic contribution to the household, represents too high an opportunity cost. Household heads depend on their wealth and borrowing capacity to support their children’s education. In addition, the opportunity cost is further elevated when the children have to cover long distances to attend poorly equipped schools, most of which do not adequately prepare them for the job market, even if the opportunities for work existed. Legislation banning child labor is desirable, but it has no impact on the large number ofchildren who work with or for their parents.

Early Marriage. A higher percentage of women drop out of primary school because of early marriage. The practice of girls marrying at a young age is quite common in Guinea-Bissau as in other African countries, according to studies conducted by UNICEF. Parents choose to marry off their daughters early for a number ofreasons. Considering the high level ofpoverty in Guinea-Bissau, poor families may regard a young girl as an economic burden and her marriage as a survival strategy for her family. They may also think that early marriage offers protection for their daughter from the dangers of sexual assault, or more generally offers the care of a male guardian. Early marriage may also be seen as a strategy to avoid pregnancy outside of marriage, which often is an embarrassment to the family. Gender discrimination can also underpin early marriage. Girls may be married young to ensure obedience and subservience within their husband’s household and to maximize their childbearing. In addition to the potential for maternal incompetence, early marriage can have serious harmful consequences for children, including the denial of the girl’s education, health problems caused by premature pregnancies that may lead to higher rates of maternal and infant mortality and sexually transmitted diseases, including HIV/AIDS. Moreover, in societies where early marriage is common, children who refuse to marry or choose partners against the wishes of their parents often become the victims of abuse or are socially ostracized. In Guinea-Bissau, UNICEF is using daily broadcasts on television and radio informing girls and their families of the consequences of early mamage.

6.6 In spite of the progress made, adult illiteracy is about 58 percent, among the 10‘ highest in the world, with sharp differences between the male and female population (see table 6.3 below). Only five countries perform worse than Guinea Bissau on female literacy (Niger, Burkina Faso, Mali, Iraq and Benin). Data from the 2005 household survey suggest a similar trend with an overall illiteracy rate of 47.5 percent for all, population above 15 years old and 63.2 percent for the female population. About 70 percent of all illiterates live in rural areas. Illiteracy among rural population above 15 years old reaches 58 percent and 73 percent for rural females (compared with 37 percent ofthe rural adult males).

Table 6.3: Adult illiteracy Rates ( percent) for population above 15 years of age) 1980 1985 1990 1995 2000 2003 % illiterate adult Males 68.9 63.4 57.7 51.7 45.9 42.2 % illiterate adult Females 93.3 90.5 87.1 82.5 76.5 72.4 % adults (Total) 81.4 77.3 72.8 64.5 61.6 57.7 Source: UN Common Database (UNESCO estimates). http://globalis.gvu.unu.edu/indicator_detail.cfm?~dicatorID=l26&Country=GW

- 91 - 6.7 Dissatisfaction with public education in general was still relatively high in 2002. The classification of pupils by the type of school attended using the 2002 light household survey showed an overwhelming dissatisfaction with the public education system (see Table 6.4). As expected, 81 percent of the children in public schools seem more dissatisfied with conditions in those schools compared to 19 percent in private schools. The high dissatisfaction with the public school system was attributed to poor teaching, inadequate teaching personnel and lack of books and supplies. Remarkably, a higher proportion of public school students compared to private school students thought the cost of education was too high. The motivation behind this response is not exactly clear since private schools in Guinea-Bissau are generally more expensive than public schools. The likely reason is that the respondents may be alluding to the opportunity cost of public education, which may be higher considering what they are getting for their money’s worth. It may also suggest that those that attend private schools belong to socioeconomic groups that are relatively better off, and therefore consider the level of school fees to be less of a constraint.

Table 6.4: Reasons for Dissatisfaction by Region in 2002 (in percentage) Bissau Bafata Gabu Cacheu Oio BiomboBolama QuinardTombali Not Satisfied (%) 56.69 64.13 46.46 47.86 66.86 63.63 70.92 Lack ofbooks/supplies 33.33 48.13 23.61 31.90 48.95 40.99 42.73 Poor teaching 13.67 9.57 9.54 8.70 15.10 3.36 13.23 Inadequate teachers 8.22 23.6 10.59 23.98 21.65 19.67 15.03 Poor facilities 12.27 34.03 17.30 12.32 25.56 28.86 39.97 High fees 8.22 23.6 10.59 23.98 21.65 19.67 15.03 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

Table 6.5: Reasons for Dissatisfaction by Type of School in 2002 (in percentage) Lack of Poor Inadequate Poor High National Not satisfied books/supplies teaching teachers facilities fees Public 75.4 81.33 79.78 92.06 91.83 76.49 91.83 Private 24.6 18.67 20.22 7.94 8.17 23.51 8.17 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

- 92 - 2. EFA and Basic Education Service Delivery and Utilization in Guinea-Bissau

6.8 Guinea-Bissau adopted the Dakar Framework for Action, EFA at the World Education Forum held in Dakar (Senegal) in early 2000. The Forum agreed on six two main goals for EFA, of which Box 6.3: EFA and MDGs became Millennium Development Goals for education later in 2000 (see Box 6.3, 1. The Six Goals of the Dakar Forum across). During the same year, Guinea- Goal 1. Expanding and improving comprehensive early Bissau adopted its EFA National Action childhood care and education, especially for the most Plan, which aimed at providing vulnerable and disadvantaged children. Goal 2. Ensuring that by 2015 all children, particularly compulsory free and non discriminatory girls, children in difficult circumstances and those basic education. The initiative was part belonging to ethnic minorities, have access to and of the education sector program of the complete free and compulsory primary education of good quality. Government's Poverty Reduction Goal 3:Ensuring that the learning needs of all young Strategy (PRS). people and adults are met through equitable access to appropriate learning and life skills programs. 6.9 The EFA initiative has Goal 4. Achieving a 50 percent improvement in levels of adult literacy by 2015, especially for women, and contributed to increasing public equitable access to basic and continuing education for all spending on basis education. Since the adults. adoption of EFA, the Government has Goal 5. Eliminating gender disparities in primary and secondary education by 2005, and achieving gender constructed several hundred new low equality in education by 2015, with a focus on ensuring cost classrooms and eliminated tuition girls' full and equal access to and achievement in basic and book fees in public primary schools education of good quality. Goal 6. Improving all aspects of the quality of education in 2000. As a result, public expenditure and ensuring excellence of all so that recognized and on education improved from 1 to 3 measurable learning outcomes are achieved by all, percent of GDP from 2000 to 2002, and especially in literacy, numeracy and essential life skills. further increased to nearly 4 percent of Source: The Dakar Framework for Action. Education for GDP by end-2004. During 2002-04, All, para. 7, Paris, WESCO, 2000. public spending on basic education grew 2. MDGs Relating to Education for Guinea-Bissau on average by nearly 6 percent per annum (see Table 6.6 below). Primary, Goal. Achieve universal primary education. secondary and adult alphabetization Target. Reach an overall GER in basic education of 98 percent by 2015. programs were allocated nearly 80 Goal. Promote gender equality and empower women. percent of total spending on Target. Reach a GER of 69.7 percent for girls by 2005 education.'02 The remaining 20 percent and eliminate gender disparity in primary at all levels of education no later than 2015. supported training and other scholarship programs in tertiary education. The Source: UNESCO and Ministry of National Education of policy of supplementing attendance in Guinea-Bissau. local universities with training in foreign institutions ofhigher learning has been in place for a number of years,'03 and is designed to build capacity in areas where the local university is not equipped to provide capacity.

102 This refers to current expenditures only. Public resources were just enough to pay salaries, and other current spending. Therefore, little public investments have been made in education and health during the period under study. Even if the national budget for education has increased in real terms between 2000 and 2004, the execution rate has often been less than 50 percent of the budget voted by the Parliament. In fact, the public finance effectively allocated to education has not even been enough to cover recurrent salary expenditure. lo3Grants to study abroad have been curtailed significantly in the last few years and in most cases students receive their stipend irregularly.

- 93 - The question, however, is whether there is a mechanism to get those graduate students to come back home once they complete their studies. There is no data on the number of students who return but anecdotal evidence suggests that only a small proportion make it back to Guinea-Bissau, which raises the question as to the efficiency of the use of those resources, especially at a time when the educational system within the country is in dire need of funding. External donor assistance provided the bulk of the financing, with the World Bank-supported Basic Education Project, providing on average about 20 percent ofthe annual education sector budget since during 1998-2004.'04

Table 6.6: Public Education Spending by Level, 2002-04 Year 2002 2003 2004 (in million of CFAF francs) Total Current Expenditures on Education 4,322 4,703 5,049

Alphabetization 95 197 307 Pre-school 39 41 44 Primary Education 1611 1717 1806 Elementary (EBE) 1129 1207 1275 Complementary (EBC) 481 509 531 Secondary Education 1766 1873 1950 First Cycle 642 689 727 Second Cycle 674 682 680 Technical & Professional 449 500 542 Tertiary 810 873 940 Local University Education 172 203 237 Foreign University Education 637 669 703

(in percentage oftotal) Alphabetization 2 4 6 Pre-school 1 1 1 Primary 37 37 36 Secondary 41 40 39 Tertiary 19 19 19

Memorandum Item: Total Current Expenditures on Education (in % ofGDP) 3.0 3.4 3.5 GDP (in billion ofCFAF francs) 141.9 138.7 142.6 Source: Ministry of Education of Guinea-Bissau

6.10 During 2002-04, there was a consistent increase in per capita public spending (unit cost) as a ratio of GDP per capita for primary school students as well as for teachers in the same cycle. The unit cost on students in nominal terms increased from 5.6 percent of GDP per capita in 2002 to 6.4 percent of GDP per capita in 2004.'05 Similarly, teachers' unit cost increased from 1.7 percent to 1.9 percent during the same period (see Figures 6.5a and 6.5b).lo6

lo4The Basic Education Support credit (FIRKIDJA) in the amount of US$ 14.3 million was approved by the Bank in 1997 and was closed on June 30,2005. 'Os The corresponding real values of the unit cost followed the same trend as inflation (measured by the change in consumer price index (CPI)) remained low and stable at around 3 percent per annum over the period under consideration. lo'As previously indicated in footnote #65, the increase of public financing allocated to the education budget has not translated in additional funding for the education sector, because the budget's execution rate has been very low for the last few years.

- 94 - 99 Figure 6.Sb: Unit Cost Per Figure 6.5a: Unit Cost Per Student in EBQ Teacher in EBE, 2002-2004) 2002-2004

6.8 6.6 1 2 64 g 62 U 3 1.8 56 s 1.7 ," 58 8 56 1.6 54 1.5 52 2002 2003 2004 2005 5 I II 2002 2003 2004 2005 1

6.11 The Government's policy of subsidizing books and school fees under Flgure 6.6: Ehrollment Rates in Basic EFA contributed to improve enrollment Education, 1997-2003 (in percentage) and accessibility rates in basic ed~cation."~In Guinea-Bissau as in other 100.07 9 1.0 low-income countries, the lack of income is 80.0 the main constraint for attending school. 60.0 Hence, poor households often make 40.0 judgments about the relevance of schooling 20.0 before making the decisions about which 0.0 children to send to school. When Guinea- 1997 1999 2001 2003 Bissau started implementing its EFA national action plan in 2003, many public schools very rapidly experienced a surge in ~ enrollment numbers. Information provided by the Ministry of Education of Guinea- Bissau indicates that gross enrollment rates (GER) in basic education (computed as the average GER in the primary and elementary levels) rose steadily from 57.8 percent in 1997 to 91 percent in 2003 (see Figure 6.6, above).lo8 The increase was mainly attributable to record enrollments in primary school, with rates exceeding 100 percent during the period considered in some regions. In the case of Guinea-Bissau, it is not uncommon for GER in the primary to be higher than 100 percent. Several reasons could account for this seemingly anomalous situation. The first reason is that children who missed out in the past on education are being encouraged to catch-~p.'~~In this regard, sources at the Ministry of education indicate that, several NGOs have set up programs in various regions ofthe country to provide basic education for young adults who otherwise would have given up on education. Second, the elimination of tuition fees in primary education created a surge in demand for education while at the same time parents who

lo'The primary school GER is the proportion of the total number of pupils in primary 1-6, irrespective of age, expressed as' percentage of the population of 7-12 year olds. lo*The student population and the number of teachers in the basic education increased proportionately. This resulted in a roportionate increase in both populations, thus maintaining a constant student-teacher ratio during the period considered. The same may apply if children are allowed to enter school earlier than the prescribed age of enrollment.

-95- did not enroll their children in schools during the conflict were eager to put them in schools once the conflict was over. Third, significant amounts of repetition of school years by children, who are not achieving appropriate levels ofeducation, as is common in Guinea-Bissau, could also increase primary school enrollment rates above 100 percent. In the long term however, it is expected that these rates will decline and stabilize slightly around 100 percent as most ofthe over-aged pupils would be leaving the system, making the GER almost close to the net enrollment rate. Of course if poverty increases the opportunity cost ofschooling will probably increase which in return will push children to drop out early without completing the six-year primary education cycle. Indeed, given the high incidence of poverty in Guinea-Bissau, parents may decide to interrupt the education process of their children (mainly girls who are converted as domestic workers or prematurely wed). An analysis of child labor and schooling decisions ofhouseholds is provided in Box 6.4 below.

6.12 The accessibility to primary education measured by the gross intake Figure 6.7: Gross Intake Rates (GIR), 2002- rate (GIR) has exceeded 100 percent 04 in FBE during 2002-04 (see Figure 6.7).'10 A 1.60 GIR above 100 percent is an indication 1 An 1 of the presence of under-aged and/or over-aged children enlisted in the primary school age group. The high interest in primary education could be attributed to several reasons. First, the government's policy to provide free and 2002 2003 2004 compulsory basic education at this levei, may-have served as an incentive I GIR (year one) GIR (year four) for many parents who, otherwise, could have been discouraged by their inability to pay the tuition for their children. Second, the government in collaboration with other donors and NGOs, especially those based in rural areas, started offering food and beverages to parents who enrolled their children, especially female children in school. This incentive-based policy is likely to have made a significant and positive impact on the gross intake ratio. Third, some efforts mostly through donors financing' '' was done to build new classrooms, provide textbooks and some school materials and provide in-service teacher training, all ofwhich contributed to stimulate the demand for education and made basic education services slightly more attractive.

'loThe gross intake rate is a measure of the total number of new entrants in the first grade of primary education, irrespective of age, expressed as a percentage of the population of the official school age to primary education. 'I' The government-IDA financed project Firkidja, already mentioned previously, supported the delivery of education services between 2000 and 2005.

- 96 - Box 6.4: Child Labor and Schooling Decisions of Households in Guinea-Bissau

A parent’s decision to send a child to school, work or both is a time allocation decision, as both activities could be competing for the child’s time. How households decide over schooling and labor force participation of children is by all indications, one of the most important development concerns, since the impact of this decision on human capital accumulation is likely to have profound short and long-term implications not only at the household level, but also at the national level. An econometric specification that specifically takes this interdependency into consideration can be used to characterize the child’s schooling and work decisions jointly. A bivariate probit model tests the likelihood of children working and going to school conditional on such household characteristics as the (a) household size variables and their square (number of infants, children and adults); (b) characteristics of the household head, including the gender of the head, the age group the head belongs to, the marital status of the head, the migration status of the head, hisiher level of education; the employment status of the head, hisiher sector of activity, and whether helshe works in the public or private sector; (c) the region where the household is located; and (d) household access to basic services such as potable water, schools, public transportation, food market and health facilities. The results ofthe econometric estimation are provided below: Determinants of Child Labor and Employment in 2002 Full Sample Urban Rural In School Working In School Working In School Working Demographic Characteristic Number o? infants (under-5s) in household -0.16 NS NS NS -0.15 0.11 Square of number of infants in household NS -0.02 NS NS NS -0.02 Number of children (aged 5-14) in household 0.45 NS 0.65 -0.60 0.43 NS Square of number of children in household -0.04 -0.01 -0.05 -0.07 -0.04 0.01 Number of adults (aged 15-64) in household 0.13 0.23 0.44 NS 0.10 0.24 - Square of number of adults in household NS -0.02 --0.03 NS NS -0.02 Gender of Head Household Head is Female 0.41 NS 1.41 NS 0.42 NS Subjective Poverty Problem Feeding Household -0.28 0.21 -0.75 0.71 -0.30 0.27 Education of the Head Primary Education 0.44 --0.18 -0.67 NS 0.45 -0.26 Secondary Education 0.55 --0.33 1.01 NS 0.59 --0.41 Tertiary Education 0.81 -5.90 6.03 -5.50 NS -5.88 Sector of Employment (Head) ConstructiodIndustry NS --0.30 NS NS NS -0.31- TransportKommerce -0.27 -0.56 NS NS -0.34 -0.39 Education/Health/Serices -0.48 -0.51 NS Ns 0.60 -0.54 Administration 0.48 -0.60 NS NS -0.45 -0.54 Access to Basic Services (in minutes) Potable Water (<15) NS NS NS NS NS NS Food Market (45) NS -0.29 0.86 NS NS -0.31 Public Transportation (45) --0.27 NS -0.94 NS NS NS Primary School (<15) 0.91 --0.24 NS NS 1.06 -0.23 Secondary School (<15) NS NS --1.25 NS -0.53 -0.42 Health Facility (45) NS NS -1.02 NS NS 0.24 ~ Regional Dummies Bissau 0.95 -0.66 Bafata 0.47 -0.25 Gabu 0.37 -0.24 Cacheu 0.67 -0.43 Biombo/Bolama 0.42 NS QuinardTombali 0.53 NS Source: World Bank staff estimates. NS: Not significant. Displayed coefficients are significant with a 10% level of confidence. Underlined coefficients are significant with a 5% level of confidence. Bold coefficients are significant with a 1% level of confidence. To be continued ... I...

- 97 - Box 6.4 (continued)

The analysis reveals that the domestic responsibilities required in households with a high number of infant children are likely to drive child labor and decrease the probability of enrollment. The results show that outside the region of Bissau, a unit increase in the number of infants will significantly decrease the probability that the parents ofthat household will send their children to school.

Among the characteristics of the head of household, the results consistently show a strong preference for schooling in female headed households compared to households headed by males. If financial well-being is a major determinant of primary education in Guinea-Bissau, then this result is further proof that female heads of households are economically better-off than their male counterparts as shown in the second chapter. Other household characteristics which seem to play a major role in the decision of parents to provide primary education to children are the educational and employment status of the head of household, and whether or not the household head had a problem meeting the food requirements of the family. In each of the samples, particularly the sample for regions outside Bissau, the results show that the higher the educational attainment ofthe head, the stronger the preference for schooling and the deterrence from child labor. As pointed out by Grootaert and Patrinos (1999), when the educational status ofthe head serves as an indicator for the preference ofchild education, it may also work as an important policy variable to the extent that it is used as a targeting mechanism for identifying those children at higher risk ofnot going to school: In the sample for all regions outside Bissau, employment of the household head in the Transport/Commerce and EducatiodHealth sectors appears to have a negative and significant effect on child labor but a positive and significant impact on school enrollment. The effect of working in the construction industry is positive and not significant for schooling but negative and significant for child labor. Another remarkable result from the analysis is that household heads that had a problem meeting the food requirements of their households seem to have a strong and significant likelihood of not enrolling their children in school, preferring instead to make them engage in child labor. The impact of this effect is strong, particularly, in Bissau, where the cost of living can be extremely high for people who do not have a reliable source of revenue.

Surprisingly, the regression results do not present a clear picture on the impact of basic services on household decisions to educate their children or make them work. In the sample for Bissau, households situated 14 minutes or less from food markets and health facilities seem to have strong and significant preference for educating their children. Interestingly, proximity to secondary schools and public transportation seem to diminish the likelihood ofparents to send their children to school. Correspondingly, the data for Bissau does not provide any evidence on the impact of any of the infiastructural services on child employment. In the sample for the rest of the country, there is evidence to suggest that the location of primary schools near households is likely to have a strong, positive, and significant effect on the decision to send children to school as well as serve as a deterrent to child labor. There is no evidence that food markets will encourage schooling but proximity to them is likely to serve as a disincentive to child labor. Contrary to the effect in Bissau, health facilities situated in the rest of the country seem to have a positive effect on work. In the full sample, we also tested possible regional influences on household decisions to educate children instead of having them work to help sustain the household. The results indicate that households in all regions are more likely to provide education for children compared to children in the reference region of Oio. With respect to child labor, however, it appears that children in Oio have a better chance of not working than children in Bafata and Gabu.

- 98 - 6.13 The finding of the 2005 QS seem to suggest that EFA initiative has had- positive outcome. About 90 percent of parents acknowledged that the basic education service was 'good or satisfactory'. The contrast with the situation in 2002 showing that the majority of people surveyed were dissatisfied with the quality of education suggests that EFA has indeed had a positive effect on education services (at least on enrollment rates).

6.14 While the EFA program seems to have a positive impact on access and participation in primary education, there is little guaranty about an improvement of the internal efficiency of the education system. Access (intake) and participation (enrollment) define the populations having access to learning, but these indicators do not tell whether pupils will complete their primary education. That is, enrollment rates provide a general idea about the physical capacity of the educational system, in terms of its ability to accommodate students at a particular level of education, but it does not provide an indication of the performance of the education system. In Guinea-Bissau, the internal efficiency ofthe education system remains low while access has improved. Despite some improvement, especially in first and second grade, the average repetition rate in primary remains around 20 percent (it varies depending on grade level). The drop out rate is also an important measure of the education system internal efficiency. Students that repeat a grade once or twice are likely to drop out. In fact, many children drop out of school before they have achieved the necessary skills to function as literate individuals. To really measure the performance of the education system, one has to measure what children are learning and the level ofproficiency achieved as well as to assess the amount ofteaching taking place.

6.15 By the same token, the SR in grade four - which measures progression in the educational system - was estimated at 58 percent in early-2005 (an increase of 14 percentage points from 2002). This suggests that by 2005, only 58 percent of the cohort of children who had entered primary school four years earlier were retained through to grade 4. The increase in the SR is a good sign which could be attributed to the corresponding decrease in the drop-out rate over the reference period. Nevertheless, this result also suggests that 42 percent ofthe children in that cohort could not make it to the end of the first cycle of primary school.'12 If four years of primary school is to be considered as a prerequisite for sustainable literacy, (or more likely five or six years considering the quality ofeducation in Guinea-Bissau), then policy makers would have to redouble their .efforts to improve student retention rates, which is needed in order to increase the primary completion rate.

6.16 The considerable difference between the GIR at the first grade ofprimary school and the intake rate in the final grade of (the first cycle of) primary school, the net of repeaters is another illustration of the need to improve retention rates. In general, the completion of primary schooling remains a major issue. In many low-income countries, children are pushed out by costs, unfriendly classroom environments or the need to supplement family income. Although the SR to grade 5 increased in many countries

112 This phenomenon is also confirmed by studies on other developing countries, which have shown that high enrollment may be accompanied by low primary completion rates (UNICEF, Global Education Digest, 2004).

- 99 - during the 1990s (and in some cases rose by more than ten percentage points between 1998 and 2001), they fell short of 75 percent in thirty out of ninety-one countries with available data, and were below 66 percent in half of the sub-Saharan African countries. Substantial declines in many countries reflected deterioration in education quality (UNICEF, 2003). The point here is that, educational policy, through EFA, underlines that need for all children to start grade one, but it also called for all children to complete a primary education of good quality. Guinea-Bissau’s EFA National Plan of Action seems to have failed until now to retain students in the school system until the desired level of completion.’ l3

6.17 Despite the apparent success of Guinea-Bissau’s EFA initiative, the transition from the primary to the secondary levels remains dubious. The transition rate (TR) to secondary school is measured by the number ofstudents admitted, net ofrepeaters, to the first grade of secondary school in a given year, expressed as a percentage of the number of pupils enrolled in the final grade of 1 primary education in the previous year. In the Figure 6.8: Transition Rate in case of Guinea-Bissau, the TR was estimated Primary& Secondary Education 2002-04 at about 81 percent in 2002 but systematically 0.82 , dropped since then to about 75 percent in 0.81 early-2005 (see Figure 6.8). The persistent 0.8 decline in the TR could be a signal of 0.79 bridging problems between the two cycles or 0.78 levels of education, due either to deficiencies 0.77 in the examination system, or inadequate 0.76 admission capacity in the higher cycle or 0.75 level of education, or both. Data provided by 0.74 the Ministry ofEducation, show that the GIR I 2002 2003 2004 for the first year ofthe secondary school cycle averages 36 percent for the period of Transition rate (EBC-ESG) observation. The limited intake capacity could very well be the bottleneck restricting the movement of children from the primary to the secondary cycle. It could also be due to the high cost of secondary education or simply the general lack of interest in pursuing an education beyond the primary level, given the poor remunerative incentives in the formal labor market. This does not, however, rule out the possibility of other problems related to the administering of examinations or physical problems concerning the availability of adequate school facilities. While measures of access, participation or completion, are not measures of achievement, they do, however, provide very valuable information about the viability of an education system, especially those features about the system that are amenable to policy change. Considering the country’s fragile economic situation, the sustainability of the EFA initiative and the likelihood to make progress in achieving the MDGs in education would require continued support from the international community (see Box 6.6).

‘13 Due to the lack of data on students who have actually graduated, the Ministry of Education of Guinea-Bissau (MOE) uses a proxy indicator (for the rate of completion), which is calculated based on the number of children enrolled in the last year of primary education, net of repeaters.

- 100 - Box 6.6: Sustainability of the EFA Initiative

Until now, Guinea-Bissau's EFA National Action Plan has been su ported in large part by the international community through the Government National Education Program (r ). The fimding initially committed (now nearly disbursed) is far from being sufficient to cover the total financing needs ofthe EFA National Action Plan. The World Bank-supported Basic Education Project, which provided a significant amount of external financhg to education since 1998, was closed on June 30,2005. Initially, external support to EFA was meant to be short - lived, as it was expected that the Government will be able to self-finance the program after the country's economic situation had recovered. This was probably unrealistic. Unfortunately, almost six years after the war, restoring political stability and good governance - necessary conditions for private sector development and growth - remain a major challenge.

Given the country's difficult fiscal situation, it is unlikely that the Government would be able to pursue the funding of the EFA in the absence of further external assistance. In recent years, the Government's budget for education has been hardly sufficient to pay for salaries of teachers. This situation is likely to persist at least for some time as the fiscal stance is expected to remain difficult for the next two to three years. Until then, Guinea- Bissau faces the risk of experiencing high drop out rates because households would not be able to afford high education costs.

The issue of sustainability of EFA can be illustrated by comparing the estimated average unit cost of education per student for the representative household and the estimated average disposable income that the representative household can set aside to pay for education costs. The results of these estimations are provided in the table below. Estimation of the Income and the Unit Cost of Education for 2006 and beyond (in CFA francs) Estimation of the annual income per capita for the poor A. Total population (number of inhabitants) 1,500,000 B. Number ofpoor in the population (60% ofthe population) B=Ax60% 900000 C. Nominal GDP (in CFA) 142,600,000,000 D. Agricultural income (60% of GDP) D=Cx60% 85,560,000,000 E. Disposable income per capita for the poor (in CFA) E=D/B 95,067 El. Sustenance and other expenses E1=E/2 47,533 E2. Health care E2=E/4 23,767 E3. Education E3=E/4 23,767 Estimation of the potential annual unit cost of education for the poor F. Total Education current budget (in CFA) 5,050,000,000 G. Salaries (in CFA) 2,000,000,000 H. Potential total education costs for the poor H= F-G 3,050,000,000 I.potential unit cost education for the poor (average household size=7) I=W(B/7) 23,722 m:Bank Staff Estimates

It is assumed that: (i)agricultural GDP represents the total annual disposable income of the poor;") (ii) households are able to devote only one quarter of their disposable income to education costs; and (iii)in the absence of external assistance the Government will be able to pay for only teachers' salaries, while parents take care of the tuition and book fees.@) In these circumstances the share of disposable annual income per capita available for the household's education spending is estimated at CFA 23,767. Compared to the average unit cost of the household which is estimated at CFA 23,722, parents would be able to send only one child at a time to school per year. Knowing that the average household size in Guinea-Bissau is about 7 individuals, it is safe to argue that a lot of children will simply not attend school, and those already enrolled will drop out because parents would not afford to pay for the ensuing costs .(t)

Based on the results of this simple analytical framework, it is likely that EFA will not be sustainable beyond 2005 if the program relies solely on domestic financings.

") Financing of the Government's Education Program Since 1998 were as follows (US$ million): World Bank (14.5), Sweden (3.0), Afrizan Development Bank (18.0), Plan International (6.0), World Food Program (1.9) and UNICEF/FNUAF' (4.0). (') For the sake of simplicity it is also assumed those households do not pay taxes. @) Under-funding of the EFA program may lead to a decrease in the quality of education services in primary schools. In this context, rich households would rather opt for sending their kids to private schools, hence leaving the poor to support all the burden of the cost. (') Inflation is expected to remain under control at less than 3 percent.

- 101 - 3. EFA and Equitability of Public Resource Allocation to Schooling

6.18 This section uses a benefit incidence analysis to determine the distributional characteristics of public spending on education in Guinea-Bissau. Figure 6.9 Figure 6.9: Incidence and Coverage of Public is a visual presentation ofthe incidence Spending on Education in 2002 and coverage of public expenditure on the different levels of education in 1 .oo Guinea-Bissau (based on the 2002 survey 0.80 household data). Three 0 indicators are represented in the Figure: 0.60 the share of the poor that benefits from 3f public transfers (vertical axis), the share of the total income that reaches the poor (horizontal axis), and the size of the transfershncome (proportional to the size of the bubbles). Four types of o'200.00 0.001 0.40 0.60 0.80 1.00 0.20 programs (pre-school, primary, secondary and tertiary levels of Target education) are analyzed and the equitability of the distribution of the 0 Primary Q Secondary Tertiary 0 Pre-school resources for those programs, vary significantly.

6.19 Allocations to primary and secondary education are the largest public transfers to households. The Primary and secondary levels of education appear to be the largest recipients ofpublic spending, while allocations to tertiary (excluding spending on students studying abroad) and pre-school education are considerably smaller as indicated by the sizes ofthe bubbles. These findings seem to indicate that while resource allocation in the primary and secondary education is not a concern at least at the macro level, there is a need to improve the efficiency of spending in the various programs within each level. With regard to primary education, the results confirm the positive impact of the EFA initiative on spending in basic education as foreshadowed in the previous section. With respect to secondary education, the results seem to support earlier allegations regarding the need to create a bridging or transition mechanism with primary education through a system of incentives to parents, pupils and school teachers. Similar to the cases of Brazil and the other countries, providing income subsidies to families with school-age children on the condition that each child attends school is likely to limit dropout rates in the primary. At the same time, teachers in both levels of education should be adequately trained and remunerated if they are to delivery quality teaching services. It is estimated that at least 50 percent of teachers in both levels of primary education do not have the required qualifications to teach at that level. It is worth noting that there are two main schools in the country for training primary and secondary school teachers but the absence of any continuing education programs (i.e.: in-service teacher training) for certified teachers makes it impossible to train even certified teachers on proposed reforms in the school system. Public resources in both levels of education should therefore be allocated

- 102- in a manner allowing these incentive mechanisms to operate in order to limit dropout rates and ensure an orderly transition ofpupils from the primary and secondary levels.

6.20 In Guinea-Bissau, primary education and secondary education are potentially ‘good’ for the poor. The results suggest that while 62 percent of poor households receive benefits for primary education, public transfers from secondary and tertiary education respectively reach only about 50 percent and 37 percent of the poor.. Beyond the numbers, these results indicate how important the first two levels of education are crucial for reducing poverty in Guinea-Bissau. It is worth noting that while the adequacy oftransfers for both tertiary and pre-school education are about the same, in terms of the coverage of these programs, a higher percentage of the poor benefit from pre-school education.

6.21 In Guinea-Bissau, basic education appears to be the only level of education that is best-targeted to alleviate poverty. Pre-school and primary education have the best targeting incidence and coverage. Even though the adequacy ofpre-school education can be improved, it has, along with primary education, the best targeting incidence and coverage. The adequacy of public transfers for secondary education is high and it has a fairly good coverage but the targeting incidence is quite low. Tertiary education has a low coverage of the poor, a weak targeting incidence, and is inadequately funded from a poverty alleviation-perspective. As discussed in earlier chapter, about two-thirds of households in Guinea-Bissau are below the poverty line. A targeting mechanism which gives less than that amount ofthe benefits to the poor would not be considered pro-poor or well-targeted to alleviate poverty. The results show that with about 60 percent of primary school allocations going to the poor, it is the only level of education that is best- targeted to alleviate poverty. The share ofpublic transfers in pre-school education going to the poor does not benefit all households below the poverty line; nonetheless, with fifty five percent ofthose benefits received by the poor, one can argue that a significantly high proportion of their population share is targeted. The same can, however, not be said for tertiary education where the target incidence is low at only thirty percent. These findings underscore the important fact that primary education is the main gateway of access to formal education. As such, high spending programs in the primary leading to high enrollment rates are necessary but non sufficient conditions for the internal efficiency of the schooling system as indicated earlier.

B. UF AND HEALTHSERVICE

6.22 Good health is important for all, especially for the poor whose physical well- being is a critical determinant of economic well-being. The poor are, on average, more. prone to illness because they are often more exposed to disease-causing agents such as polluted water and inadequate sanitation; they often live in crowded conditions where infections spread easily; they have lower access to preventive health services and their poor nutritional status makes them more vulnerable to disease. Guinea-Bissau recently prepared a National Health Development Plan (PNDS), which established the rules and the priorities, and defined the strategies for the activities in the health sector. The first and second plans covered the periods 1997-02 and 2003-08, respectively. In addition to implementing preventive health measures, especially in the areas of maternal and infant

- 103 - care, another major objective of the PNDS is to make healthcare affordable and accessible to those with limited financial means. In the early 199Os, the country also adhered to the BI; however, the Government’s program did not get offthe ground and, as a result, people in Guinea-Bissau have had to continue paying the normal prices for services that many cannot afford. In 2003, the Government adopted a law for the implementation of the UF in public sector health centers and hospitals. This section seeks to assess the impact of the cost recovery system on public health service delivery in Guinea-Bissau. The remainder of the section will focus on the following points: (i) situation of the health sector prior to UF; and (ii)implementation of UF and impact on health service delivery and utilization. The analysis is based on the 2002 household survey data and macroeconomic data from 2002 to 2004.

1. Situation of the Health Sector Prior to the adoption of UF

6.23 As the implementation ofUF begun after 2002, the 2002 ILAP household survey data will be used to assess the situation prior to the implementation ofthe initiative. With regard to the BI, although it was adopted since the early 199Os, the implementation of the initiative has effectively started only after 2002. It is therefore unlikely that the BI had bore any significant positive outcome prior to 2002.

6.24 The demand for health services was particularly high in rural areas in 2002. The type ofhealthcare provider a patient consults for the treatment ofan illness can make a difference in how effective the treatment is and how rapid the period ofrecuperation is likely to be. In general, the majority of patients who sought modem professional health- care in Bissau in 2002, visited private clinics. In the other regions people’s preference went toward regional health centers. Findings of the survey indicate that hospitals and regional health centers played the most significant role in meeting the healthcare needs of the nation. The regional health centers are located in the main urban city of each region. They provide a wide range ofhealth care services such as maternal and child health care, general and other specialized consultations, basis surgical services, laboratory and pharmacy services. People and mainly the poor often visit these centers because they find a wide variety ofservices at a low cost.

6.25 At the national level, more than fifty percent of all those who consulted professional health providers in the four weeks preceding the survey went to regional health centers (Table 6.7). The use of these facilities decreases from the poorest to the wealthiest quintile, which explains why a higher percentage ofthe number ofpeople who use regional health centers was higher in the rest of the country than in Bissau. From Table 6.8, 81 percent ofthe residents of Oio, the poorest region in the country, relied on regional health centers for their healthcare needs. The poverty decomposition of those who used these healthcare providers shows that the highest percentage ofpoor people and the extreme poor use the services of private doctors/dentists. The providers of these services are often not qualified medical doctors but are nurses with several years of medical experience who consult patients at home, sometimes without the appropriate, license. The poverty and extreme poverty rates are also high among the users of traditional healing facilities and pharmacies.

- 104- 6.26 Contrary to expectation, traditional healing centers did not seem to be patronized by a significant number of people. The opinion of healthcare experts and other social researchers in Guinea-Bissau is unanimous in contradicting the survey results. According to these experts, and anecdotal evidence gathered in field interviews, the people who used these services cuts across different socio-economic classes and for most of these people, the decision to use these services is so personal that they may not necessarily want to share it with people outside their family circles.

Table 6.7: Health Providers by Location and Economic Status (in percentage) '

Quintiles Location and Economic Rest of Extreme Status National Bissau Country Poorest 2 3 4 Wealthiest Poverty Poverty Regional Health Center 50.79 45.36 52.66 51.66 54.48 53.59 48.77 47.60 59.2 16.3 Private Health Center 7.48 6.25 7.90 5.74 4.73 6.56 11.50 7.65 51.2 12.6 Hospital' 29.36 28.75 29.56 31.39 28.53 30.88 21.07 29.59 58.5 16.6 Private Clinic 6.13 15.36 2.96 0.58 5.28 3.80 6.36 11.05 38.7 2.9 Pharmacy 0.38 0.18 0.45 0 1.17 0.42 0.21 0.11 66.8 11.9 Private Doctor/Dentist 0.47 0.18 0.57 1.90 0.1 1 0.62 0.12 0.12 87.5 63.3 Traditional Healer 4.25 2.68 4.79 1.36 3.52 3.88 4.49 3.15 65.9 27.0 Other 1.15 1.25 1.11 1.36 2.18 0.26 1.48 0.68 63.8 15.1 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002 *Hospitals include regional and national hospitals

Table 6.8: Health Providers by Region (in percentage) Regions National Bissau Bafata Gabu Cacheu Oio Biom/Bol QuidTomb Regional Health Center 50.79 45.36 48.48 34.19 52.82 81.88 61.38 47.01 PrGate Health Center 7.48 6.25 2.55 4.87 13.61 3.57 13.43 7.15 Hospital' 29.36 28.75 34.1 50.85 21.91 5.61 21.14 38.42 Private Clinic 6.13 15.36 4.03 6.69 1.31 1.78 2.26 1.63 Pharmacy 0.38 0.18 0.42 0.17 0.92 0.30 0.32 0.32 Private Doctor/Dentist 0.47 0.18 1.03 0.26 1.42 0 0.11 0 Traditional Healer 4.25 2.68 8.27 2.46 5.51 6.55 0.89 4.54 Other 1.15 1.25 1.11 0.52 2.50 0.30 0.48 0.92 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002 'Hospitals include regional and national hospitals

6.27 In the particular case of malaria, a striking proportion of patients frequented non professional health providers or traditional healers. About 19 percent of the people surveyed were sick or injured and therefore needed medical help. Of this proportion more than half were victims of malaria. Surprisingly, the highest number of cases of the disease was in the top two quintiles and in the region ofBissau; groups that. are better disposed economically to prevent the contraction of the parasite (see Table 6.9). A possible reason to this unusual phenomenon is that people in these groups are in a better position to identify the symptoms of malaria compared to people in rural areas or in the poorest quintiles. The latter are likely to be ill-informed about the disease or are more liable to attribute the symptoms of malaria to other causes. At the national level, more than 60 percent ofmalaria patients consulted traditional healers (see Table 6.10). It

- 105 - appears that a higher percentage of malaria patients in the wealthiest quintile as well as residents of Bissau also sought this type of treatment. However, considering the deadly nature ofmalaria, it is disturbing that a much higher proportion ofthose afflicted with the disease did not seek professional help. The non frequentation of modern health services by the poor may be explained by the lack ofincome; however, that ofthe rich may signal other factors such as the absence of programs to inform the population about malaria prevention and treatment, and/or the lack of malaria drugs and other malaria-related health care supplies.

Table 6.9: Selected Health Indicators in 2002 (in percentage) Rest of National Bissau Country Poorest 2 3 4 Wealthiest Sick in last 4 weeks 19.13 16.84 19.93 16.24 17.34 18.01 20.32 23.74 Malaria 54.40 56.74 53.70 51.48 50.51 50.41 58.39 58.83 Diarrhea 8.5 1 7.12 8.92 8.88 7.50 10.69 8.31 7.50 Accident 2.21 1.22 2.90 3.35 3.11 2.42 1.97 0.83 Dental 2.16 2.46 2.07 2.91 1.54 2.07 2.03 2.29 Skin 3.59 3.50 3.62 4.04 3.95 3.71 3.35 3.15 Eve 3.17 2.33 3.41 3.66 2.53 3.04 3.04 3.50 E& Nose & Throat 2.40 1.55 2.65 2.36 3.14 1.80 2.76 2.05 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

Table 6.10: Malaria Patients who Sought Professional Treatment in 2002 (in percentage) Rest of National Bissau Country Poorest 2 3 4 Wealthiest Contracted Malaria 54.40 56.74 53.70 51.48 50.51 50.41 58.39 58.83 Consulted NPHP or TH' 64.39 69.41 62.83 54.81 63.39 61.34 66.42 7 1.03

National Bissau Bafata Gabu Cacheu Oio Biombo/Bo QuinardT Contracted Malaria 54.40 56.74 45.10 55.56 65.00 41.77 62.10 49.55 Consulted NPHP or TH' 64.39 69.41 66.21 64.51 62.68 42.27 69.69 68.86 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002 'NPHP: Non Professional Health Provider 'TH: Traditional Healer

6.28 Limited access to health services restricted the use of medical services. In 2002, the proportion of households with access to healthcare services, and the impact on the use ofmedical services varied by economic status and region. About 37 percent ofthe poorest households had access to health services compared to 46 percent for the wealthiest. Consequently, the rate ofusage ofmedical services was only 9 percent for the poor, as opposed to 19 percent for the rich. The improvement in access in the higher quintiles was reflected in increased usage of health facilities in those quintiles. In the regions, however, the correlation between accessibility and usage was not very clear. While access was higher in Bissau than in the rest ofthe country (55 percent compared to 38 percent in average), overall, usage remained relatively the same across all regions (12 to 14 percent) (except in Oio where the rate was 7 percent) (see Table 6.11, above). In 2002, the population of Guinea-Bissau in general was discriminated against in the usage ofmedical services.

- 106 - Table 6.11: Health Indicators in 2002 (in percentage)

Medical Services

Access Need Use Quintiles Poorest 36.69 16.24 9.02 2"d 34.97 17.34 11.14 3rd 38.36 18.01 1 1.43 4th 42.55 20.32 13.39 Wealthiest 45.70 23.74 16.88 Regions Bissau 54.46 16.84 12.22 Bafata 24.67 21.06 13.71 Gabu 41.90 19.69 12.92 Cacheu 41.49 23.23 14.66 Oio 24.03 16.40 7.48 Biombo/Bolama 38.91 21.45 14.10 QuinardTombali 37.90 18.10 12.64 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

6.29 The general level of dissatisfaction with the health care system was quite high in 2002, without significant differences across socioeconomic groups and the area of residency. As shown in Table 6.12 below, a large proportion of households were dissatisfied with public health services in 2002. More than 80 percent of the residents of Gabu, for example, had a bad experience with their medical facilities, and the two issues most of them were concerned with were the cost of treatment and the availability of drugs. With regard to the BI, this finding seems to confirm the ineffectiveness of the initiative prior to 2002.~~~

Table 6.12: Persons who consulted a Health Provider and were not satisfied by the Reason of Dissatisfaction in 2002 (in percentage) Dissatis- Facilities Long No trained No drugs Treatment fied not clean wait professionals Cost available Unsuccessful Other Quintiles Poorest 63.03 1.56 21.08 4.18 17.14 36.47 5.54 13.25 2"d 66.46 2.69 15.01 5.05 32.59 37.94 9.16 15.47 3rd 68.1 1 3.01 19.49 4.54 30.82 40.49 9.5 8.72 4th 65.98 2.95 16.34 3.29 28.99 42.15 8.94 7.62 Wealthiest 62.29 4.03 22.26 9.4 25.05 41.45 7.42 7.05 Regions Bissau 65.02 3.01 18.99 5.64 27.17 40.07 8.18 9.90 Bafata 65.00 2.32 24.46 3.57 25.54 34.1 1 7.68 8.21 Gabu 80.35 7.15 21.87 8.40 41.83 59.71 13.24 13.88 Cacheu 54.84 0.65 10.23 2.62 27.41 35.08 5.02 5.97 Oio 58.33 3.43 14.9 6.77 20.44 3 1.82 6.99 12.57 Biombo/Bolama 66.69 1.81 25.78 5.57 28.17 39.20 5.00 4.95 QuinardTombali 68.61 1.45 22.05 4.51 34.41 44.24 6.50 5.11 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAP) Data, 2002

114 Gabu was selected as pilot region for the implementationof the BI.

- 107 - 6.30 Of all people not consulting health facilities in the four weeks preceding the survey, the overwhelming majority (about 92 percent) thought there was no need. Residents ofBafata, Cacheu, and Oio, have a higher percentage ofrespondents who have not used medical care in the last four weeks because it is either too expensive or the facility is too far away (see Table 6.13). These findings clearly illustrate that prior to the adoption of the UF, the health system was confronted with numerous difficulties, which were hindering efficient delivery and utilization ofhealth services.

Table 6.13: Persons Sick who did not Consult a Health Provider and the Reason Why ( percent) Not Consulting NoNeed cost Distance Other Quintile Poorest 90.98 93.18 3.22 2.02 3.17 2"d 88.86 92.21 4.48 2.76 2.90 31d 88.57 9 1.40 4.83 2.97 3.04 4~ 86.61 91.61 4.08 2.53 3.63 Wealthiest 83.12 90.9 3.80 4.45 4.56 Region Bissau 87.78 94.76 2.29 0.22 3.06 Bafata 86.29 87.95 5.48 6.36 4.45 Gabu 87.08 92.85 3.70 1.31 3.46 Cacheu 85.34 90.09 4.54 5.32 4.11 Oio 92.52 89.33 7.48 4.65 3.05 Biombo/Bolama 85.90 91.43 4.31 3.69 2.59 QuinardTombali 87.36 94.3 1 1.73 1.56 3.83 Source: Staff Estimates based on the Guinea-Bissau Light Household Survey (ILAF') Data, 2002.

2. Implementation of UF and impact on health service delivery and utilization

6.3 1 In recent years, public sector health expenditures have declined in Guinea-Bissau to the point where nearly all Government fbnding went to salaries. Although Guinea-" Bissau has adhered to the BI since 1990, many patients were informally paying for health services 'under the table'. Recognizing the difficulty to cover the recurrent cost of the health system, the authorities passed a law introducing UF. The law on UF was a mean of legalizing this practice and providing more transparency and efficiency in the collection and use of UF. The new system was considered a viable solution as UF have always discriminated against the poor.

6.32 UF are being collected and managed at the facility level in a fairly. transparent manner. Since 2002, Guinea-Bissau's new cost-recovery system includes the recovery of drug fees (already covered under the BI), and charges for outgoing visits, lab tests, and other medical visits. So far, the resources are being collected and managed in a relatively transparent manner. Revenues at the level ofthe facilities are managed by a management committee composed of the facility's staff and representatives of the civil society. The UF system is being carried out in all the 127 health facilities at the national level. In 2004 for example, the amounts collected per region were within the range of CFAF 6 and 33 million (about US$ 12,000 and 66,000) (see Table 6.14). The total cost recovered in 2004 represented about 5 percent of the health sector budget. The resources collected were primarily used to pay for incentives to the medical personnel (20-30

- 108 - percent). The remainder of the resources served for the payment of drugs and other maintenance costs (30 to 60), and the constitution of reserves (10-40 percent). In limited cases, 10 percent of the resources were transferred to the Regional Health Department (see Tables 6.15 and 6.16, forward).

Table 6.14: UF Collected by Region in 2004 Regions Number ofhealth facilities Revenue collected (CFAF franc) Medical Drugs visits Total Bissau 9 14411,968 18,35 1,919 33,163,887 Bafata 12 22,072,788 10,577,610 32,650,398 Biombo 7 2,164,016 6,217,091 8,38 1,107 Bij ag6s 11 5,533,896 3,936,770 9,470,666 Bolama 6 3,195,760 1,702,936 4,898,696 Cacheu 10 3,697,184 4,683,923 8,38 1,107 Gabu 18 18,494,288 9,865,349 28,359,637 Oio 15 12,151,044 9,159,084 2 1,3 10,128 SBo Domingos 7 10,348,664 12,585,677 22,934,341 Quinara 13 3,189,988 3,142,699 6,332,687 Tombali 19 11,127,404 5,249,189 16,376,593 Total 127 106,787,000 85,472,247 192,259,247 Source: Ministry of Health of Guinea-Bissau

Table 6.15: Utilization of UF from Medical Consultations in 2004 (in percentage of total revenue collected) Incentives Drugs, and Transfers to Savings to the other the Regional personnel maintenance Health costs Direction Bissau National Hospital 30 60 10 Laboratory services 20 60 20 Regions Regional Hospitals 30 40 30 Health Centers 20 40 10 30 Basic Health Units 20 30 10 40 Source: Ministry of Health of Guinea-Bissau

Table 6.16: Utilization of UF from Drug Sales in 2004 (in percentage of total revenue collected) Personnel Medical and Savings Incentives other costs maintenance costs

Bissau Central Depot 10 60 30 Health Centers 20 40 40 RBgions Regional Depots 20 40 40 Health Centers 30 40 30 Basic Health Units 30 40 30 Source: Ministry of Health of Guinea-Bissau

- 109- 6.33 It is too soon to determine the impact of the UF initiative on heath outputs and outcomes. In other parts of Africa there still seems to be a debate as to the overall impact of the introduction of UF on the delivery of health services. Findings from a number of countries (Swaziland, Mozambique, Ghana, the Gambia, Zambia) show a decline in utilization rates when user charges were not matched b an improvement in the availability of drugs, access to services and quality of care.'" Evidence so far from Guinea-Bissau as well as the broader region suggests improved levels ofdrug availability following the BI and more recently the introduction ofUF.' l6

6.34 The UF initiative is, however, believed to have improved health service utilization. Since its inception in 2002, the UF system is believed to have substantially improved the availability of drugs and the access to health services at affordable costs. Total UF collected in 2004 reached 5 percent of the health sector budget, ofwhich adult and infantkhildren consultation services and drug sales represented 45 and 55 percent of total, respectively (see Table 6.14 above). However, the cost of other specialized consultation services, such as stomach illnesses and rheumatology, remain relatively high for the average citizen. The majority of patients were treated for pregnancy, respiratory and digestive ailments, and malaria. In the case of malaria for example noticeable progress has been made in the provision of drugs and consultation services, in contrast with the situation prior to the UF system.

6.35 Nevertheless, there is a need for further information on the potential equity and distributional consequences of the UF initiative. Evidence from focus groups conducted as part ofthe 2005 QS suggest that the poorest elements ofthe population may have difficulties covering the costs even of basic adult and infant/children consultation services and drug costs. It may therefore be necessary to consider targeted exemption policies for those unable to pay.

6.36 In order to ensure equitable access to drugs in rural areas, part of the transport costs would need to be subsidized by consumers living in Bissau. Given the emphasis on cost-recovery, moreover, drug prices are often higher in rural areas compared to Bissau due to the cost of moving the goods from Bissau to the regions. As the majority of the poor live in rural areas, this situation may increase the inequality in' the access to drugs in the context of the BI. Not addressing this issue can reduce the potential benefit of the UF for the poor. A possible way to address the issue is to have part of the transport costs of drugs shipped to rural areas subsidized by consumers living in Bissau.

6.37 While there have been clear improvements in the delivery ofhealth services over the last couple of years, the poor are still facing many constraints in accessing the services provided. Beyond the costs of service provision, distance, the non-availability of technical staff (e.g. doctors), and the quality of care are also important factors in excluding people from accessing health services.

115 Shehu (1997); Haddad and Fournier (1995); Abel-Smith and Rawal(l992) 'I6 Soucat, (1995); Soucat et al, (1997)

- 110- 6.38 Health facilities would need to be adequately equipped to meet increasing turn out rates. The frequentation of health facilities is gradually increasing in both central and regional health centers, due to the regular availability of drugs and improved service. However, the majority of the facilities, mainly in rural areas, are ill-prepared to meet this surge in demand. Medical personnel are often insufficient and lack adequate skills and training. Moreover, housing problems and low salaries discourage the staff to take assignments in rural areas."' It is critical that the authorities accelerate the implementation of the ongoing reform rogram in the context of the PNDS, with the support of the Bank and other donors."'In the absence of improvements in these areas, many turn to alternative providers of health services, including traditional healers (mumndade), self-diagnosis/medication, and in urban areas also to private clinics. Recommendations: Train additional medical staff (both technical and administrative). Build housing facilities for medical staff in rural areas.

C. A COMMUNITY-BASEDSERVICE DELIVERY FORREDUCING POVERTY

6.39 There is a growing consensus among development practitioners that participatory mechanisms are necessary for improving service delivery in weak institutional environments. In the case of Guinea-Bissau, given the limited State capacity, the authorities might consider establishing a CSOs/NGOs-managed community development fund in the context ofthe Government's PRSP process (see Box 6.3 for useful guideline for designing a social fund)."g These funds would support micro projects aimed at improving basic education and health service delivery at the community levels, while building on the relative strength ofthe civil society.

6.40 In the case of Guinea-Bissau, using externally funded NGOs collaborating with the Government can improve the delivery of EFA and UF social services. Many national and international NGOs serve an intermediary function through channeling development resources to community-based organizations, providing them with services or technical assistance, or helping to strengthen their capacity. Although NGOs vary in their ability and commitment to work with the poor, in many cases they have advantages as intermediaries in reaching people-including women, ethnic minorities, and the very poor-who are not represented equitably by formal institutions. For example, NGOs may have more field presence in a given area and employ local people familiar with local conditions. They may have better rapport with the poor and a clearer understanding of poor people's survival strategies and perceived needs than other kinds of intermediaries. They may be familiar with low-cost techniques and innovations relevant to poverty alleviation. They also usually have greater flexibility than the staff of official agencies. Many intermediary NGOs have experience in participatory project design, and skills in participatory research, community mobilization, facilitation techniques, and group dynamics.

'17 The Government often rents houses for the medical personnel operating outside Bissau. 118 The reform program includes: (i)the rehabilitation of health centers; (ii) building institutional capacity at the central and regional levels; (iii)improving human resources management; and (iv) improving the M&E system in the health sector by putting in place an integrated Health Information System (SIS). l9 The forthcoming Bank-supported community driven development (CDD) program and similar other externally financed community development projects may be used as platform for this initiative.

- 111 - 6.41 For the next phases of the EFA initiative, the CSOs/NGOs-managed community development find can be used to provide cash transfers to parents who would otherwise let their children work (or get married) at an early age instead ofattending school. CSOs and NGOs can also represent an efficient intermediary for providing other school-related services such as delivering school materials, designing and managing retraining programs for school teachers, and raising awareness of parents about keeping their children (especially girls) in school. With regard to the UF initiative, CSOs and NGOs can be used as efficient tools for designing and managing retraining and capacity building activities for administrative and medical staff in health centers (procurement, financial accounting, and management ofthe facility-based drugstore).

Box 6.3: Designing a Social Fund

The following tips may be helpful for designing a social fund:

Be clear about the objectives ofthe social fund and the centrality ofparticipation in realizing objectives.

Focus on participatory micro projects, and prioritize selection of micro projects according to the intensity of participation. Ask how much local participation there has been in the processes of micro project identification, design, and planning and if the community is involved in implementation.

Encourage or mandate establishment of micro project management committees. It may be necessary to require representation ofwomen and other marginalized groups in the management committees.

Apply participatory data collection methods, such as participatory rural appraisal and beneficiary assessment, to monitor subprojects regularly (assessing 10 percent ofsubprojects a year may be a good target).

Provide for capacity building of intermediary and community organizations as early as possible.

Develop a piloting phase in which participatory mechanisms can be tested, the concept of participation introduced, and actors given time to learn.

Design flexible and transparent procurement and disbursement procedures.

Deliver what has been promised. Avoid delays in processing proposals and delivering services.

Critically monitor the performance ofthe social fund.

Source: Marc and Schmidt (1995).

- 112- D. CONCLUSION AND POLICY CONSIDERATIONS

6.42 As foreshadowed above, while the EFA initiative has contributed to increase enrollments in the primary schools, there remain high risks about the retention of pupils in school. Moreover, there are uncertainties about the sustainability of the EFA because major external supports are ending, while the Government is still not ready to auto finance the program. The adoption of the UF system has brought more autonomy and efficiency to health facilities in the sense that each facility collects and manages its revenues in a relatively transparent manner. The resources are used to renew drug inventories and pay incentives to the personnel. Contrary to the EFA, there are less financial concerns than organizational ones. With the ongoing support of the Bank and other donors, the health sector would have the much needed boost if the reform agenda, agreed upon in the context ofthe PNDS is carried out.

- 113 - REFERENCES

Almeida, Josue Gomes. Technical Coordinator of PRDSP. Personal comment. March 29, 2005.

Ascher, W. and A. Hubbard, Introduction: Central American Recovery and Development” in Central American Recovery and Development: Task Force Report. Durham, NC: Duke University Press, 1989. Pages: 1-18.

Ball, Nicole (1996) “The Challenge of Rebuilding War-Torn Societies,” in Managing Global Chaos: Sources of and Responses to International Conflict, eds. Chester A. Crocker, Fen Osler Hampson, and Pamela Aall. (Virginia: USP Press, 1996), 607.

Baldacci, E., B. Clements, and S. Gupta (2003) “When Reducing Fiscal Deficits Makes Sense for Low-Income Countries” Finance and Development (December 2003), International Monetary Fund. http://www.imf.org/extemal/pubs/ft/fandd/2003/12/pdfhaldacci.pdf

Castro, A. (1980) 0 Sistema Colonial Portugties em Africa, Lisbon: Editorial Caminho.

CILSS/FAO. 2004. Rapport de mission conjointe CILSS/FAO d’Evaluation Preliminaire de la campagne agricole en GuinCe Bissau.

CNJ/SNV. Identificaqzo dos Problemas dos Jovens Guineenses e Revelaqzo das Suas Aspiraqaes.

Chong, A., and M. Rama (2001) “Do Separation Packages Need to be That Generous? Simulations for Government Employees in Guinea-Bissau”, World Bank Economists’ Forum / edited by Shantayanan Devarajan, F. Halsey Rogers, Lyn Squire. Washington, DC : World Bank, 2001-2002.

Coordination RCsidente des ActivitCs OpCrationnelles du Systbme des Nations Unies en GuinCe-Bissau. 2001. Bilan Commun de Pays (CCA) - Guinee-Bissau.

Correia, P. (2005). Personal interview conducted on March 29,2005.

Correia, P. (2005) Electronic Messages on Pension and SSR in the Security Sector

Creppy, E.G (2003) “The Growth Effects of the Socio-Economic Status of Women: A Panel Data Approach” American University.

Cruz, D.V. 2005. ALTERNAG. Personal interview conducted on March 25,2005.

- 114- Datt G., and M. Ravallion (1992) “Growth and Redistribution Components of Changes in Poverty Measures: A Decomposition with Applications to Brazil and India in the 1980s” Living Standards Measurement Study (LSMS) Working Paper No. 83. http://povlibrary.worldbank.org/files/15 180 Growth and redistribution - Ravallion and Datt.pdf

Devarajan S. and D. Go (2002) “A Macroeconomic Framework for Poverty Reduction Strategy Papers with an application to Zambia” World Bank. Washington, D.C. May 2002.

Diop Ndiamey, Paul Brenton, and Yakup Asarkaya (2005) “Trade Costs, Export Development and Poverty in Rwanda” forthcoming.

Dollar, D. and A. Kraay “(2000) Growth Is Good for the Poor” World Bank Working Paper - http://www.worldbank.org/research/~owth/pdfiles/wowth8oodforpoor.pdf

(2001) “Trade, Growth and Poverty), World Bank Working Paper. http://www. worldbank.org/research/growth/pdfiles/Trade5 .pdf

Economist Intelligence Unit (EIU) (2005) Country Report - Guinea-Bissau, April 2005, London: EIU.

Economist Intelligence Unit (2004), Guinea-Bissau Country Report. October. London. http://db.eiu.com/report dl.asp?mode=pdf&eiu issue id=457626645

EU (2001), StratCgie de CoopCration et Programme Indicatif 2001-2007. Brussels. http://europa.eu.int/comm/development/body/csp rsp/print/gw csp fr.pdf#zoom=lOO

Evans, P. (1995) “Embedded Autonomy: State and Industrial Transformation” Princeton: Princeton U.P.

FAO. 2000. Diagnostic de la situation de l’agriculture, Clevage, forets, peches et nutrition - Une contribution au processus CCA.

FAO. 2004. Nouveau partenariat pour le dkveloppement de 1’Afrique (NEPAD) Programme dCtaillC pour le devkloppement de 1’ agriculture Africaine (PDDAA). RCpublique de GuinCe-Bissau TCP/GBS/2907. Programme national d’investissement ii moyen terme (PNIMT). Version provisoire - 26 juillet 2004.

FAO. 2003. Local institutions and livelihoods: Guidelines for analysis.

- 115 - FAO/World Bank. (forthcoming) Diagnosis and emergency restructuring plan to enhance the domestic benefits of the fishing sector and its integration in the world economy. Assistance to Fisheries Sectors in West Afiica. Joint FAO/World Bank Fisheries Sector Memorandum.

Fati, S. 2005. Ex-Major State Chief of Armed Forces. Personal interview conducted on March 29,2005.

FMI. Internal dcument (Draft) Memorando de Politicas Econ6micas e Financeiras para 2005.

Forrest, J. B. (2003) Lineages of State Fragility: Rural civil society in Guinea-Bissau, Oxford: James Currey.

Food and Feed Grains Institute, Kansas State University. 1992. Rice Marketing in Guinea-Bissau: A Study to Develop Information for Policy Makers.

Franqa, F.M.C. 1995. Agro-Neg6cio do Arroz na GuinC-Bissau. Trade and Investment Promotion Project (TIPS) - USAID.

Governo/Comissilo EuroperidINEP. Cooperaqilo Descentralizada: Um Novo Rum0 a Seguir na GuinC-Bissau. Versilo Preliminar.

Governo/INEP/SNV. 2002. A Sucesstio Politica e a Alternhcia do Poder nas Sociedades Tradicionais Guineenses. (FASE 11).

Government of Guinea-Bissau (2004) “Guinea-Bissau Cashew Sector Development Study”, Private Sector Rehabilitation and Development Project of Guinea-Bissau.

Grootaert, C. and Bastelaer T. (eds.) 2002. The Role of Social Capital in Development: An Empirical Assessment. Cambridge, UK, Cambridge University Press.

Grootaert, Christian and Harry Patrinos (1999), eds., Policy Analysis of Child Labor: A Comparative Study. New York, St. Martin’s Press.

GTZ. 2004. Governance Questionnaire - An instrument for analyzing political environments.

INEP/AFRICARE. 1996. Estudo S6cio-Econ6mico da Regiilo de Tombali. Indicadores S6cio-Econ6micos para um Anilise Critica e Longintudinal do Impact0 do PIRST.

INEP. 1996. Subsidio para a Politica de Regionalizaqtio da SNV. AnSlise Regional: Tombali e Quinara.

INEP. 2002a. Estado Modern0 e Poder Traditional (verstio final).

- 116- INEP. 2002b. SucessTio Politica e a Alternancia do Poder nas Sociedades Traditionais Guineenses (FASE 11).

INEP. 2004. Socio-Economia das Tabancas nas Zonas de IntervenqTio do Programa AGIR.

INEP/UNDP. 2005. GovernaqTio de Qualidade para Todos na GuinC-Bissau. Projeto de Apoio a Boa GovernaqTio. Estudos EstratCgicos para uma Boa GovernaqTio.

INEC. 2002. Evaluation de la pauvretk en GuinCe Bissau (2001-2002).

International Monetary Fund (2004) “Benefit of Fiscal Consolidation”. Remarks by Rodrigo de Rato, Managing Director of the IMF to the Real Academia de Doctores, Barcelona, Spain, November 25,2004. http://www.imf.ordextemal/np/speeches/2O04/1 12504b.htm

IMF and World Bank (2000), Guinea-Bissau Interim National Poverty Reduction Strategy and Joint StaffAssessment. Washington DC. http://imaaebank.worldbank.ordservlet/UTDSContentServer/IW3P/IB/2OO- 1/0 1/26/0000 94946 0101 1305321964/Rendered/PDF/multi page.pdf

Jaeger, P., and S. Lynn (2004) “Guinea-Bissau Cashew Sector Development Study” Private Sector Development Project of Guinea-Bissau.

Kakwani, N. (2000) “On Measuring Growth and Inequality Component with Application to Thailand” Journal of Quantitative Economics, Vol. No.

Lin, N. 2001, Social Capital. Cambridge University Press.

Lin, N., Cook, K. and Burt, R. (ed.) 2001. Social Capital: Theory and Research. New York, Aldine de Gruyter.

Louenqo-Lindell, I.(2002) Walking the Tight Rope: Informal Livelihoods and Social Networks in a West African City, Stockholm: Stockholm University.

Lynn, S. Plan0 de Acq6es da EstratCgia de Desenvolvimento do Sector do Caju da GuinC- Bissau (Draft Action Plan as ofNov. 30,2004).

Marc A. and M. Schmidt (1995) “Participation and Social Funds” Participation Series, Environment Department, Paper No. 4, World Bank, Washington, D.C.

Mamdani, M. (1996) Citizen and Subject: Contemporaly Africa and the legacy of late colonialism, Princeton: Princeton University Press.

Manley, Andrew, (1998), Guinea BissadSenegal: War, Civil War and the Casamance Question, mimeo.

- 117- MinistCrio da Agricultura, Florestas, Caqa e Pecuaria. 2003. Programa de Emergencia do Apoio Seguranqa Alimentar (PEASA) (2003-2006).

MinistCrio da Agricultura e do Desenvolvimento Rural. Ante-Proj ecto do Regulamento Geral da Lei da Terra.

MinistCrio da Agricultura, Florestas, Caqa e Pecuhria. Plan0 de EmergCncia. (sem data).

Ministbre de l’agriculture et du dbveloppment ruraVFA0. TCP/GBS/2906(T) Renforcement des capacites organisationnelles et techniques des organisations paysannes. RAPPORT (Draft).

Ministbre de l’agriculture et du dbveloppment ruraVFA0. 2004. TCP/GBS/2906(T) Reforcement des capacities organisationnelles et technique des organization paysannes (OP) Bilan des actions engages dans le domaine de lappui aux OP.

Ministbre de l’agriculture et du dbveloppment ruraVFAO. 2004. TCP/GBS/2906(T) Reforcement des capacities organisationnelles et technique des organization paysannes. Evaluation de l’environnement exterene dans lequel les organizations paysannes (OP) dkveloppent leurs activitks.

MinistCrio do ComCrcio Industria, Turismo e Artesanato. Avaliaqgo da Campanha e Comercializaqgo da Castanha de Caju de 2004 e Recomendaqaes.

MinistCrio da Defesa Nacional. Direcqao-Geral de Produqao e Modernizaqgo das Forqas Armadas. 2004. Memorandum. Versgo Preliminar.

MinistCrio da Defesa Nacional. 2005a. Direcqgo-Geral de Produqiio e Modernizaqiio das Forqas Armadas. Memorandum (Versgo Preliminar).

MinistCrio da Defesa Nacional. 2005b. Direcqgo-Geral de Produqgo e Modernizaqgo das Forqas Armadas. Projetos de Relanqmento Geral de Produqgo das Forqas Armadas.

MinistCrio da Economia e Finanqas. 2003. Anhlise da Fileira do Arroz.

MinistCrio da Energia e dos Recursos Naturais. Comunicaqiio Nacional Initial da GuinC- Bissau sobre as Mudanqas Climhticas. (Relat6rio Draft).

Multiple Indicator Cluster Survey - Guinea-Bissau. December 2000.

OECD. 2001. The Well-being of Nations - The Role of Human and Social Capital. Centre for Education Research and Innovation, OECD.

Ostrom, E. and Ahn, T.K. (ed.) 2003. Foundations of Social Capital. An Elgar Reference Collection. Cheltenham, UK., Northampton, MA, USA.

- 118 - PDRRI. 2001. Programme de Demobilisation, Reinsertion et Reintegration des Ex Combattants. Document de Programme.

PDRRI. 2005. Relat6rio de Implementaqiio do PDRRI.

Pelissier, R. (1997) Hist6ria da Guinb: Portugueses e Africanos na Seneghbia, Lisbon: Editorial Estampa.

Pina, C.L. and CamarQ, T. Janeiro de 2005. Plano de Acqdes para Melhoria da Performance Industrial e Promoqiio do Emprego na GuinC-Bissau.

Programme des Nations Unies pour le DCveloppement (PNUD), Fonds d’Equipement des Nations Unies (FEW) (2003) : Guinea-Bissau, Programme d’ Appui A 1’Emergence et au DCveloppement du secteur de la Microfinance (PAEDMF) (2004 - 2007). http://www.uncdf.or~/francais/microfinance/sectorDev/fenu nbs-prodoc.pdf

Ravallion, M. And S. Chen (2003) “Measuring Pro-poor Growth” Policy Research Working Paper No. 2666 The World Bank, Washington D.C.

Rede Nacional das Associaqdes Juvens (RENAJ) 2003. Visita de Constataqiio Junto As Associaqdes do Interior do Pais. Relat6rio da MissEo.

Relat6rio Nacional de Desenvolvimento Humano. 1997.

Relat6rio Nacional de Desenvolvimento Humano. Guinea-Bissau. 200 1. Seguranqa Humana.

Relat6rio de avaliaqiio da implementaqiio do PRNR apoiado pel0 CRRE. 2004.

Reptiblica da GuinC-Bissau. 2002. Documento de EstratCgia National de Reduqiiao da Pobreza (DENARP).

Reptiblica da GuinC-Bissau. 2002. Programa Nacional de Boa Governaqiio. Volume 2 - Eixos de Intervenqiio.

Reptiblica da Guink-Bissau, (2003) “Plano Nacional de AcqEo Educaciio Para Todo” MinistCrio de Educaqgo Nacional. http ://portal.unesco.ordeducation/en/ev.php- URL ID=18 192&URL DO=DO TOPIC&URL SECTION=201 .html

Reptiblica da GuinC-Bissau, (2002) “ ‘Livro Banco’ da Defesa Nacional da Reptiblica da GuinC-Bissau”, Cooperaqiio Militar Portuguesa, Junho de 2002.

Rudebeck, L. (1990) ‘The Effects of Structural Adjustment in a West-African Village: The Case of Kandjadja in Guinea-Bissau’, Review of African Political Economy, No. 49.

- 119- Rudebeck, L. (1997) “To Seek Happiness’: Development in a West African Village in the Era ofDemocratisation’, Review ofAfrican Political Economy, No. 71.

Rudebeck, L. (2001) ‘On Democracy’s Sustainability: Transition in Guinea-Bissau’, SIDA Studies, No. 4, Gothenburg: SIDA.

Rudebeck, L. (2005) ‘Book Review - Lineages of State Fragility: Rural civil society in Guinea-Bissau’, African Affairs, January 2005, vol. 104, no. 414.

Sanca, 2005. General Secretary ofthe Ministry ofDefense. Personal interview conducted on March 30,2005.

Secretaria de Estado do Plan0 e Orprnento, INEC, INEP Relatbrio do inquerito dos indicadores multifios.

Shehu M. (1997) Overview of the Bamako Initiative programs in Nigeria Progress made and problems encountered. Conference paper on Bamako Initiative. Abuja; Federal Ministry ofHealth.

Sodemann M. and Aaby P. The Bandim Health Project, 2003.

Soucat A. 1995. The Bamako Initiative in Guinea-Bissau. Evaluation Report for UNICEF. New York, UNICEF.

Soucat A. et al. (1997) Health seeking Behavior and house hold health expenditures in Benin and Guinea. The Equity implications ofthe Bamako Initiative. Int J. Health Plan Manage ; 12:S137 - S163.

Technical Sustainable Development, Africa Region (AFTSD) and AFC 14. Fishery Strategy Sector Note. Working Document.

UNESCO (2000) “The Dakar Framework for Action - Education For All: Meeting Our Collective Commitments” the World Education Forum, Dakar, Senegal, 26-28 April 2000. http://unesdoc.unesco.or~irnaaes/OO12/001211/121 147e.pdf

UNESCO (2002) “Education for All - Is the World on Track? Summary Report” UNESCO Publishing, France. http://unesdoc.unesco.ordimaaes/00 12/00 1297/129777e,pdf

UNESCO (2003) “Gender and Education for All, the Leap to Equality: Summary Report”, UNESCO Publishing, France. http://www.unesco.ordeducation/efa report/summary en.pdf

- 120- UNESCO (2004) “Education for All, the Quality Imperative: Summary Report” UNESCO Publishing, France. http://unesdoc.unesco.ordinianes/OO13/00 1373/137334e.pdf

WAF.2002. Plan Cadre des Nations Unies pour l’aide au Dkveloppement de la Guinee-Bissau.

UNDP- Africa Future Program (2003), Crafting Africa’s Futures: National Long-Term Perspective Studies. Abidjan. http ://unpan 1.un.ordintradoc/moups/public/documents/IDEPTCTNPANOO2403 .pdf

WP/GOVERNO/INEP 2001. Estudo Qualitativo Sobre o Perfil da Pobreza na Guink- Bissau: Persepqaes e Soluqaes Locais.

World Bank (2002), Private Sector Rehabilitation and Development Project. Washington DC. http://imaPebank. worldbank.ordservlet/WDSContentSe~er/IW3P/IB/2002/03/22/0000 94946 0203 1304003783/Rendered/PDF/multiOpage.~df

, (2003), National Health Development Project. Project Information Document. Washington DC. http://imagebank.worldbank.orp/servlet/U7DSContentServer/IW3P/IB/2004/07/13/0001 04615 20040714143038/Rendered/PDF/ProiectOInfonn1ntO1OAppraisalOStage.pdf

(2004), Guinea-Bissau Public Expenditure Review. Washington DC. http://ima~ebank.worldbank.or~/servlet/WDSContentServer/IU73P/IB/2004/03/02/000O 12009 20040302125640/Rendered/PDF/271750GW.~df

(200 1) Economic Rehabilitation and Recovery Credit (ERRC) Memorandum of the President, Report No. P7373.

(2004) World Development Indicators (WDI), World Bank.

- 121 - ANNEXES

ANNEX 1. VAR MODEL

Specification of the Model

The fiscal consolidation analysis, carried out in the fourth chapter, is based on a thee- variate VAR model'20. The model seeks to assess the short-term effects of a reduction of wage expenditures on growth and Government finance. The VAR estimation is performed using the following three variables: terms of trade (TOTI), real Government expenditure (GEK), real GDP (GDPK). To better estimate the VAR the first difference of the logarithm ofthe variables are used12'. The proposed VAR does not pretend to capture perfectly all aspects of Guinea-Bissau's economy. However, it allows assessing the effects of shocks to Government expenditures and terms of trade on growth and Government finance.

VAR Functions ( ct,, are the coefficients ofthe VAR, and D denotes the first difference operator):

After estimating the coefficients of the VAR, impulse-response coefficients are derived for ten consecutive periods. Given data limitations, the VAR was run on only' 18 observation points covering the period 1987-2004 (16 observations after first differentiation of the variable). Due to this limitation, the traditional painvise Granger causality tests supporting the proposed Cholesky ordering was not performed. Therefore the analysis mainly relied on the standard methodology proposed by the model (see impulse response fbnctions below). This methodology assumed that TOTI and GEK are the most exogenous variables, while GDPK was found to be the most endogenous variable. In the case ofthe shock on Government expenditure and terms oftrade the order followed was GEK-GDPK-TOT1 and TOTI-GDPK-GEK, respectively. The proposed orderings indicate that in both cases the variable that suffers the shock (GEK and TOTI) causes GDPK.

The impulse-response coefficients are then normalized as indicated in Tables A5 and A6 to derive the impulse response elasticities ofthe cyclical components of GEK and GDPK to a shock to the cyclical component ofGEK and TOTI.

120 The VAR model was initially used as a module for the growth analysis of the 123-PRSP General EquilibriumModel (see Devarajan and Go, 2002). First differentiation is used to address the issue of stationarity.

- 122.- Table A1 1: Impact on GDPK of a shock to the cyclical component of GEK and TOTI (formulation)

1 disturbances to Impulse response coefficients for GDP growth (normalized). Responses to changes in Shock * )wth of:

Shock to GEK (X)

GEK Shock to -Period (P’) TOTI (K)

-1 P’1** s11 -2 P’2 3 P’3 4 P’4

10 * Year (up to the model horizon 2005 1 2006 1 2007 1 2008 1 1 Short-term adjustments to baseline GDP growth caused by the shock(s) T2 T3 T4 S2 1=(XlxP’2)+(KlxP2) ,s4 l=(X4xP’4)+0(4xP4) S22=(X2xP’ 1)+(K2xP 1) S42+(X4xP’3)+0(4xP3) Tl=S2 1+S22 S43=(X4xP’2)+(K4xP2) S44=(X4xP’ 1)+0(4xP 1) S3 l=(X3xP’3)+0(3xP3) T4=S4 l+S42+S43+S44 S32=(X3xP’2)+(K3xP2) S33=(X3xP’ 1)+(K3xP1) T3=S3 1+S32+S33 * ** P and P’ are the normalized impulse-response vectors (see Table A5 and A6 below forward).

- 123 - Table A1 2: Impact on GEK of a shock to the cyclical component of GEK and TOTI

Impulse response -relasticities of the cyclical components Year 2005 2006 2007 2008 of Gov. exp. to a shock to the cyclical Shock periods Period 1 2 3 4

Year (up to the model horizon) 2005 2006 2007 2008

Short-term adjustments to baseline Gov. exp. growth caused by the shock(s) T’2 T’3 T’4

S’21=(X’ lxG’2)+(K’lxG’2) S’4 1=(X’4xGY4)+(K’4xG’4) S’22=(XY2xG’1)+(K’2xGY 1) S ’42+(X’4xG’3)+(Ky4xG’3) T’ 1=S’2 1+S’22 S ’43=(X’4xG’2)+(K’4xG’2) S’44=(XY4xG’l)+(K’4xG’ 1) S’3 1=(XY3xG’3)+(K’3xG’3) T’4=S’4 1+S’42+S’43+SY44 S’~~=(X’~XG’~)+(K’~XG’~) S’33=(X’3xG’l)+(K’3xG’l) T’3=S’3 l+S’32+S’33 * ** G and G’are the normalized impulse-response vectors (see Table A5 and A6 below).

- 124 - Results of the Model Table A1 3: VAR Impulse Response Function for TOT (effect of one standard deviation innovation in TOTI) D(GDPK) D(TOT1) D(GEK) Period (A) (B) (C)

1 0.000000 7.293586 0.000000 2 -3.718151 -0.978541 -2.532837 3 -0,165532 -1.676084 1.973 122 4 4.649 106 1.748743 -0.854378 5 -2.868643 -1.442438 0.028058 6 -0.998386 0.006189 0.466874 7 2.336172 1.1798 11 -0.3 19047 8 -1.062426 -0.822889 -0.180593 9 -0.703927 -0.24493 1 0.397265 10 1.281981 0.726088 -0.188677

Cholesky Ordering: D(TOT1) D(GDPK) D(GEK)

Table A1 4: VAR Impulse-Response Function for GEK (effect of one standard deviation innovation in GEK) D(GDPK) D(GEK) D(TOT1) Period (A’) (B’) (C’)

1 0.000000 27.75677 0.000000 2 - 15.7966 1 -1 1.18754 6.694277 3 -17.29761 2.957747 -0.771224 4 9.925773 -4.365820 5.195 807 5 2.357527 1.3 10706 -2.293782 6 -4.494417 1.795006 -2.928903 7 1.937402 -0.754168 3.037846 8 0.841652 -1.148274 0.043779 9 -2.356979 1.334068 -1.896382 10 1.496444 -0.362855 1.407479

Cholesky Ordering: D(GEK) D(GDPK) D(TOT1)

- 125 - Table A1 5: Normalized Impulse-Response Coefficients for GDP Growth Responses to changes in Responses to changes in growth of TOTI growth of GEK Period (P)=(A)/(B) (P ’)=(A’)/(B ’)

1 0.000 0.000 2 -0.510 -0.569 3 -0.023 -0.623 4 0.637 0.358 5 -0.393 0.085 6 -0.137 -0.162 7 0.320 0.070 8 -0.146 0.030 9 -0.097 -0.085 10 0.176 0.054

Table A1 6: Normalized Impulse-Response Coefficients for Changes in Government Spending Responses to changes in Responses to changes in growth of TOTI growth of GEK Period (G)=(C)/(B) (G’)=(B ’)/(B ’)

1 0.000 1.000 2 0.000 -0.403 3 -0.347 0.107 4 -0.22 1 -0.157 5 0.271 0.047 6 -0.255 0.065 7 -0.117 -0.027 8 -0.188 -0.041 9 0.004 0.048 10 -0.180 -0.013

The normalized impulse-response coefficients in Tables A5 and A6 are derived from the impulse-response coefficients of Tables A13 and A14.

The Underlying assumptions ofthe model are that Government revenue may increase due to improved efficiency in tax collection; however, the authorities would not take any discretionary measures to increase tax rates in the short-term (empirical evidence suggests that fiscal consolidation policies targeting expenditure contraction are more efficient than those aiming at increasing revenues). As Guinea-Bissau’s economy is highly dependant on the export ofraw cashew nuts, the model allows a constant 1 percent annual depreciation ofthe terms oftrade over the simulation horizon

- 126 - 0 8

0 8 1

I ANNEX2. PUBLIC SECTOR DOWNSIZINGAND SEPARATION PACKAGES

The following discussion is derived from Chong and Rama (2001). The paper uses two approaches (direct and indirect) to estimate the total loss. The latter is used to determine the “right” amount ofcompensation for redundant civil servants. The analysis is based on the 1991 ILAP survey data.

The direct approach measures the total loss that the separated public sector worker would experience based on the present values ofthe earnings loss the worker would experience in all the years since displacement until retirement, and the income loss over all the years between retirement and death.’22 The indirect approach, in turn, measures the total loss based on the value of the predicted change in earnings. Put differently, public sector employees would value the benefits associated with their jobs, including old-age income security, at roughly one-fifth of their cash earnings. The direct approach suggests that some government employees would experience a very small loss in the event of separation. According to the indirect approach, however, roughly 85 percent of the employees would face a total loss in excess ofCFAF400,OOO.

Based on these two approaches, four alternative packages identified with the letters A to D have been estimated. Package A is a lump sum. Package B is a multiple of the employee’s salary in government. In package C, the amount of compensation is set in months of salary per year of service. Package D, on the other hand, is based on an analysis of the distribution of predicted losses in benefits in earnings from separation. These losses, unlike in the other packages, are nonlinear because they involve the individual characteristics ofthe employees, as well as the characteristics oftheir jobs.

The performance of the four separation packages can be evaluated on two grounds: cost and fairness. For each government employee, the separation package can be compared to the expected loss in earnings and benefits. If the package exceeds the expected loss, it is assumed that the employee would be satisfied. It follows that each separation package is associated with a satisfaction rate, ranging from 0 to 100 percent. The cost and fairness indicators were calculated for a pre-specified satisfaction rate, for instance, 60 percent. Cost was measured as the average package ofthose employees who would be satisfied by the amount of Compensation received. Fairness was measured as the correlation- coefficient (not shown in Table A21) between the compensation and the loss in earnings and benefits, for those employees who would be satisfied. Implicitly, both indicators used voluntary separations as a benchmark, because the fraction of employees satisfied should be the same as the fraction of employees who would take the package and leave the public sector, if they had a choice.

In practice, every “reasonable” package will overcompensate some employees and undercompensate others. Only exceptionally will compensation match exactly the predicted loss in earnings and benefits. Therefore, the employees who are satisfied with the package are most likely overcompensated. It follows that the average cost of a

122 It is assumed that no loss in old-age income if the worker gets a formal sector job after separation.

- 129 - package is higher than the average loss in earnings and benefits of those employees who are satisfied with it. In terms of the voluntary separation benchmark, mistakes in the direction of overcompensation materialize, whereas mistakes in the direction of undercompensation have no practical implication. Overcompensation also implies that the fairness of a separation package can be low. If the package received by all satisfied employees was equal to their predicted loss (or a multiple of it), the fairness coefficient would be equal to one. In practice, however, some employees get a windfall whereas others are compensated “just right,” so that the fairness coefficient can be close to zero, or even negative.

Table A21 reports the cost and fairness of each ofthe four packages, for nine satisfaction rates ranging from 10 to 90 percent. The zero satisfaction rate is uninteresting, because it would require no compensation. At the other end, the 100 percent satisfaction rate can only be reached at unreasonably high separation costs, because the package has to be generous enough to match the highest loss in earnings and benefits throughout the sample, and this is often an outlier. The only difference between the direct and indirect approaches concerns the loss indicator used. The figures in bold in the table indicate the cheapest package, and the fairest package, for each satisfaction rate. In the case of Guinea-Bissau, packages A and B are quite similar. This is because of the very small variation in earnings across employees in categories “0” to “Z”. If all workers have roughly the same salary in government, a package defined as a multiple of this salary is not too different from a lump-sum payment. Whereas salaries in government are roughly the same for all employees in categories “0” to “z”, losses from separation are not. As the correlation coefficient between a constant and a variable is zero, packages A and B perform very poorly in terms of fairness. However, they are cost-effective at low satisfaction rates. If downsizing can be carried out without hlly compensating more than one-third of government employees, “flat” packages of this sort could keep downsizing expenditures under control. At the other end, “tailored” compensation, as in package D, is the most cost-effective alternative if satisfaction rates need to be high. Packages C and D have a similar performance at low satisfaction rates. The average cost per separated worker is slightly higher for package C, but the fairness coefficient can be higher as well. In order to attain high satisfaction rates, however, the number of months of salary per year of service required by package C becomes too high. At a 70 percent satisfaction rate, package C costs roughly twice as much as package D, but the latter does much better in terms of fairness. The gap between the two packages becomes dramatic at very high- satisfaction rates. In fact, “tailored” compensation dominates all the alternatives, both on grounds ofcost and fairness, if a majority ofthe workers need to be “bought off..”

The compensation formula that the government ofGuinea-Bissau was planning to use for its public sector downsizing operation, which is a special case ofpackage Cyturns out to be quite onerous. This formula would lead to a satisfaction rate of 78 percent, based on the direct approach, or 73 percent, based on the indirect approach. The average cost per satisfied worker would be roughly CFAF 2 million, and the fairness coefficient would be. around 45 percent. Better outcomes could be attained spending less than half. A lump sum of less than CFAF 1 million would achieve the same satisfaction rate of the formula considered in Guinea-Bissau. A similar result, at a similar cost, would be obtained

- 130 - offering less than seven years of salary as compensation. Package D would be even less expensive, as it would cost roughly CFAF 880,000 per satisfied worker. The fairness coefficient would, in this case, be close to 70 percent.

Table A2 1: Average compensation per Separated employee (thousands of CFAF) 11-

Direct Approach Indirect Approach

Percentage of Alternative packages Alternative packages employee satisfied

A B C D A B C D 10 176.0 138.8 227.1 232.7 358.0 36 1.O 441.0 412.7 20 300.5 300.4 451.5 397.8 437.6 437.6 636.0 556.7 30 468.8 464.3 660.6 465.2 513.0 514.0 813.5 620.6 40 632.5 633.9 818.8 567.2 624.2 618.7 989.7 650.4 50 722.0 724.3 964.3 667.0 760.0 762.4 1,188.3 705.5 60 818.0 816.2 1,157.3 736.7 843.0 84 1.6 1,446.7 783.0 70 939.5 944.2 1,464.0 809.1 932.0 930.6 1,862.8 861.0 80 1,007.4 1,014.4 2,143.6 871.8 1,065.5 1,064.9 2,759.8 940.1 90 1,106.3 1,107.2 3,642.0 927.3 1,193.0 1,194.8 4,806.3 1,052.8 m:Chong and Ram (2001), Tables 5 and 6. 1/ Package A is a variable lump sum; package B is a variable multiple of the salary in government; package C is a variable multiple of the salary multiplied by the number of years of service; package D is a variable lump sum, plus half a month of salary per year of service, minus CFAF 50,000 per year of completed education, plus CFAF 60,000 for female employees, plus CFAF 150,000 for employees who work out of Bissau. The variable component of each package is increased until the percentage of satisfied employees indicated in the first column of the table is reached. The figures in bold indicate the cheapest and fairest package for each satisfaction rate.

Boubacar-Sid Barry C:\Documents and Settings\wb224607Wy Documents\GNBWOV ASS\IPSAUlain Report\Gray cover\IPSA Gray Cover Volume 1.doc 08/01/2006 11:36:00 AM

- 131 -