Report No. 16465-GUI Guinea A Socioeconomic Assessment of Well-Being and Poverty March 31,1997 Institutionl al andl Social Policy Unit & Country Department 16 Africa Region Document of the World Bank GUINEA A SOCIOENONOMIC ASSESSMENT OF WELL-BEING AND POVERTY CURRENCY EQUIVALENTS Currency Unit: Guinean Franc (GNF) (Period average, US $1 equivalent) 1988 = GNF 474 1989 = GNF 592 1990 = GNF 660 1991.= GNF 754 1992= GNF 902 1993 = GNF 956 1994 = GNF 977 1995 = GNF 991 WEIGHTS AND MEASURES Metric System FISCAL YEAR January 1 -December 31 Vice President: Jean-Louis Sarbib, AFR Country Director: Mamadou Dia, AFC16 Division Chief: Roger C. Sullivan, AFTI I Staff: Boniface Essama Nssah, AFTI 1 Guinea at a glance Sub- POVERTY and SOCIAL Saharan Low- Guinea Africa Income Development dlamondr Populaton mid-1994 (milWons) 8.4 565 3,176 Life expectancy GNP per capla 1994 (USS) 509 510 390 GNP 1994 (bilons USS) 3.3 288 1,239 Avenge annual growth, 19S0-94 Population (%) 2.9 2.9 1.8 Laborforce (X) 1.8 2.7 118 GNP Groas per primary Moat recent estimate (ast year avaltable since 1989) capita enrollment Poverty: headcount Index (% of population) enrollment Urban population (% of total populaton) 30 28 26 Life expectancy at birth (years) 44 52 65 Inftnt mortalty (per 1,000 li bIrths) 134 94 53 Child malnutrition (% of children under 5) 18 38 Access to safe water Access to safe water (% ofpopulatIon) 60 .. 67 lllteracy (% otpopulafon age 15+) 76 50 41 Gross primary enrollment (% of school-eo population) 46 68 108 -Guinea Male 61 77 112 -Low-income group Female 30 62 100 KEY ECONOMIC RATIOS and LONG-TERM TRENDS 1976 1988 1994 1995 Economic ratloa GDP (billfons USS) .. .. 3.4 3.7 Gross domestic hivestmentVGDP .. .. 13.6 14.8 Openness of economy Exports of goods and non-factor servesGDP .. .. 19.8 21.0 Gros domesti savtngs/GDP .. .. 9.0 10.7 Gros natlonal savIngalGDP .. .. 4.3 5.5 Current account balancetGDP .. .. -9.3 -8.6 Savings Investment Interest paymentS/GDP .. .. 2.2 2.0 Total debtlGDP .. .. 91.6 86.6 Total debt service/exports .. .. 33.4 20.7 Preent vae of debt/GDP .. .. 61.2 Present vakie of debt/exports .. .. 302.3 .. Indebtedness 197644 198545 1994 1995 199644 (average annual growth) Guinea GOP .. 3.7 4.0 4.6 4.9 -LGuinea GNP per capita .. 1.4 t.6 1.8 1.9 Low-income group Exports of goods and nfs .. 3.1 -8.4 7.8 7.2 ---- - STRUCTURE of the ECONOMY 1975 1985 1994 1995 Growth rates of output and Invetment% (X of GOP) I sr Agriculture .. .. 24.1 24.1 / Industry .. .. 30.8 30.5 *o Manufacturing .. .. 4.6 4.7 s Services .. . 45.1 45.4 o Private consumption .. .. 82.3 80.9 -10 General govemment consumption .. .. 8.7 8.5 Imports ofgoods and non-factor services .. .. 24.4 22.3 GDt e-GDP 197544 198545 1994 1995 Growth rates of exports and Imports %) (avenage annual growth) Agriculture 4.2 5.1 5.0 is Industry 25 3.2 3.8 1e Manufacturing Services 44. 3 4.0 4.9 Private consumption . 4 6.5 2.3 4 General govemnment consumption 5.2 -2.2 4.8 V /4 Grou domestic nvestment .. 1.6 -9.0 14.0 .1 import of goods and non-factor servkes .. 2.0 -10.7 0.8 Gross national product .. 4.3 4.4 4.6 -Expods * ImpoZ Note: 1995 data are prellminary estimates. Figures Italics are for years other than those specified. The diamonds show four key indiators in the country (in bold) compared with its icome-group average. If data are missing, the diamond will be Incomplete. Guinea PRICES and GOVERNMENT FINANCE 1975 19S5 1994 1995 Innaon (%) Domsc 30 (% change) Consumer prices 4.1 5.3 20 Impliat GDP deflator 5.0 5.3 GoMrnnnt finance 10 (% of GDP) _ Currentrevenue 10.4 11.0 g0 01 92 3 4 9 Cunent budget balance 4.5 6.1 oo sCP 4 Overall surplus/defcit -3.6 -2.5 -GWd.r - CPI TRADE 1976 1986 1994 1996 Export and Import ovals (nill. USS) Total exporls (fob) 626 725 1 000 Bauxite 272 299 Akumina prb103 110 506 Manufactures . . . II~1l Total Imports (elf) .. 687 690 j5j0mj0o Food .. 74 73 Fuel and energy 18 82 68 82 Cap(tal goods 153 182 Export of goods end non-faector services.. . 672 774 0 - _- Imports of goods andnon8-actorces .828 823 oo 901 9 93 94 9r Res ource balance 18 0Ep -156 -49 Netfactor Income . .. -118 -133 s Net current transfers . .. -41 -35 C-urrnt account balance, _ _ _ _ _ bebfre offdil tranders ,orvs -315 -217 1_ FInancing Items (net) . .. 280 348 Changes In net reserves -3 -2 34 -31 2 Resource balance -156 -49~~~~~~~~~~~~~~~~~~~-1 Memo: ____ Re erves includig gold (mill. USS) .. .. 164 209 Convrsion rate (not)280) 20.3 24.3 976.7 990.8 EXTERNAL DEBT and RESOURCE FLOWS (miNons USS) 1975 1986 1993 1994 Conpositon of totl debt 1994 (mll. US$) Total debt outstanding and disbursed 760 1,455 2,843 3,104 G IBRD 68 55 0 0 F 152 IDA 0 117 681 773 93 B 773 Total debt service 38 72 84 97 IBRD 8 8 0 0 IDA 0 1 6 8 Composition of net resource flows C Officlalgrants 3 45 166 162 71 Offidal creditors 54 48 242 131 Prvbatecretr 4 18 3 -9 E D Foregndiredldnvment 0 1 3 30 1479 Portfollo equty 0 0 0 0 World Bank pmgrem Conmitments 21 33 15 48 A - IBRD E-- BntIal Disbursements 0 22 134 63 B-IDA D-Otlermlblatl F-Pdlvats Principal repayments 2 5 1 2 C- IMF G Short-twm Net fow -2 17 132 61 Intrest payments 6 4 4 6 Nettrandefrs -8 12 128 55 Intemational Economics Deparlment 317196 Table of Contents Foreword ............................................................ 1 Executive Summary. 2 I. INTRODUCTION .16 II. PAST AGGREGATE PERFORMANCE IN LIVING STANDARD .20 CHOICE OF LIVING STANDARD INDICATORS .20 ACHIEVEMENT IN HUMAN CAPITAL .22 IMPACT OF ECONOMIC GROWTH ON INCOME POVERTY .24 WELFARE IMPLICATIONS OF SOME PAST ADJUSTMENT POLICIES .30 III. THE CURRENT DISTRIBUTION OF ECONOMIC WELFARE 35 POLICY TARGETING AND DATA DISAGGREGATION .35 CHOICE OF SOCIAL EVALUATION FUNCTIONS .36 REGIONAL DISPARITIES .38 GENDER DIFFERENTIALS. 5 1 LITERACY, HOUSEHOLD SIZE AND POVERTY .57 IV. STRUCTURAL AND POLICY DETERMINANTS OF WELFARE AND POVERTY .61 LAND TENURE AND WATER RESOURCES .61 LACK OF DIVERSIFICATION OF PRODUCTIVE ACTIVITIES .65 LOW POVERTY RESPONSIVENESS TO ECONOMIC GROWTH .67 MANAGEMENT CAPACITY .68 INFORMAL SECTOR .69 V. HUMAN CAPITAL DEVELOPMENT .72 PAST DEVELOPMENTS IN SOCIAL SECTORS .72 HEALTH SECTOR .75 EDUCATION AND TRAINING SECTOR .81 VI. TOWARD A POVERTY REDUCTION STRATEGY .89 DESIGN ISSUES .89 STRATEGY FOR WELFARE IMPROVEMENT AND POVERTY REDUCTION .91 IMPLEMENTATION ISSUES .99 MONITORING SYSTEM .102 Bibliography .104 iii TABLES Table 1.1 Selected Human Development Indicators (UNDP) ......................................... 16 Table 2.1 Bounds for Some Living Standard Indicators .... ........................................... 22 Table 3.1 Urban vs. Rural Inequality (Atkinson) ....................................... 43 Table 3.2 Regional Primary Enrollment Rates and Home to School Distance in KI .... 50 Table 3.3 Poverty Comparisons by Education of Household Head ................................ 58 Table 3.4 Poverty Comparisons by Household Size ....................................................... 58 Table 3.5 Inequality by Expenditure Components (Income Elasticities) ........................ 59 Table 4.1 Traditional Land Tenure (Fouta Jalon) ............................................................ 63 Table 4.2 Evolution of Growth Elasticities of Poverty (Cornell Poverty Line) .............. 68 Table 4.3 Evolution of Growth Elasticities of Poverty (GOG Poverty Line) ................. 68 Table 4.4 Sectoral Distribution of GDP ............................................................ 70 BOXES Box 1.1 Assessing Living Standards ............................................................ 21 Box 3.1 Atkinson's Social Welfare Function ............................................................ 37 Box 3.2 The Extended Gini Coefficient ............................................................ 41 Box 3.3 Decomposability of Atkinson's Index of Inequality .......................................... 44 Box 3.4 The Structure of Foster Greer Thorbecke Poverty Indices ................................ 47 Box 3.5 The Normative Content of the Foster Greer and Thorbecke Poverty Indices ............................................................ 49 Box 3.6 The Importance of Labor Markets ............................................................ 53 Box 3.7 The Structure of the Urban Labor Market in Guinea ......................................... 54 Box 3.8 The Effects of Mother's Employment on Child Nutrition ................................. 57 Box 5.1 Girls' Education in Guinea ............................................................ 88 Box 6.1 Strategic Choices ............................................................ 90 Box 6.2 The Socioeconomic Role of Government ......................................................... 93 Box 6.3 The Importance of Education ............................................................ 95 Box 6.4 The Concept of Governance ............................................................ 100 FIGURES Figure 2.1 Performance in Longevity and Infant Survival (Kakwani's Improvement Index for Selected Years) ............................................................ 23 Figure 2.2 Changes in Educational Achievement (Kakwani's Improvement Index for Selected Years) ............................................................ 24 Figure 2.3 Evolution of Per Capita Consumption (Base Year 1986) .............................. 25 Figure 2.4 Poverty Incidence (Multiple Poverty Lines) .................................................. 27 Figure 2.5 Estimated Number of Poor (Multiple Poverty Lines) .................................... 28 Figure 2.6 Comparing Observed Growth to that Required for Poverty Maintenance ..... 30 Figure 3.1 Regional Distribution of Land Area, Population and Household Expenditure ............................................................ 39 Figure 3.2 Structure of the Regional Lorenz Curve ........................................................ 40 iv Figure 3.3 Regional Inequality (Extended Gini) ........................................................ 42 Figure 3.4 Regional Inequality (Atkinson's Index) ........................................................ 43 Figure 3.5 Welfare Comparisons (Equally-Distributed Equivalent Expenditure by Region) ........................................................ 45 Figure 3.6 Regional Poverty (FGT Index) ........................................................ 46 Figure 3.7 Incidence of Poverty by Region ................................ ........................ 47 Figure 3.8 Intensity of Poverty by Region ............................. ........................... 48 Figure 3.9 Severity of Poverty by Region ........................... ............................. 50 Figure 3.10 Inequality by Gender of Household Head (Atkinson Index) ............ 51 Figure 3.11 Welfare Comparisons by Gender of Household Head (Equally- Distributed Equivalent Expenditure by Gender ....................................... 52 Figure 3.12 Incidence of Poverty by Gender of Household Head ............ ....................... 53 Figure 4.1 Sectoral Linkages (Rasmussen Indices) ........................................................ 66 Figure 5.1 Health Facility Use for Illness by Quintile .................................................... 78 Figure 5.2 Per Capita Expenditures on Health ...................................................... .. 79 Figure 5.3 Per Capita Public Expenditures on Health, by Region ................. ................. 80 Figure 5.4 Gross Primary Enrollment Rates by Region .................................................. 83 Figure 5.5 Per Capita Distribution of Public Education Expenditure by Region ............ 86 Figure 5.6 Per Student Household Expenditures on Education ....................................... 87 v Foreword 1. In essence, a poverty assessment is an exercise in social evaluation in the sense that it hypothesizes alternative social arrangements. The fitness of a social arrangement may be judged on the basis of the sustainable quality of the lives which it supports. In this regard improvement of their living standards is one of the fundamental expectations of those who are governed. The World Bank has, therefore, declared sustainable poverty reduction its overarching objective and a benchmark measure of its performance as a Development institution. This declaration is equivalent to a choice of social evaluation rule. 2. The living standard enjoyed by an individual may be interpreted as an outcome of the interaction between opportunities offered by society and the readiness and ability of the individual to exploit such opportunities. Thus poverty defined in terms of living standard, has both an individual and a collective dimension; and achievements at the aggregate level, such as longevity and literacy, are key determinants of the possibilities at the individual level. The purpose of this report is to draw strategic implications about such possibilities from available information on the distribution of economic welfare in Guinea. It adds to a framework currently under development by the Government of Guinea to restructure its socioeconomic development policies. 3. Given the necessity of value judgments, sensitivity analysis is applied to the choice of social evaluation functions. Atkinson's Equally Distributed Equivalent measures of welfare are used alongside the standard Foster-Greer-Thorbecke poverty indices. Inequality is measured by means of extended Gini and concentration indices. The report emphasizes the importance of a sound implementation strategy to ensure impact. This critically hinges on governance issues: who should hold institutions accountable, and how, if they are to be effective in reducing poverty? The empowerment of stakeholders to influence institutional behavior is a key to sustained progress on poverty reduction. 4. The report was drafted by B. Essama-Nssah in close collaboration with Mark C. Woodward, Ali Ghobadi and our counterparts in Guinea, namely: Omar Diallo, Mamadou Louda Bah, Sidibe Marlyatou Diallo, and Ibrahima Sory Sangare. Within the Bank, Essimi Menye, Robert Ngong and John Ngwafon provided technical support on statistical matters. Helpful comments were received from Laura Cooley, Lionel Demery, Peter J. Lambert (University of York), Branko Milanovic, Ok Pannenborg, Mustapha Rouis, Ken Sigrist, Dominique van de Walle, and Jack W. van Holst Pellekaan. The report also makes extensive use of work done on Guinea by Cornell University, in particular on labor market issues. Document preparation was handled by Sandra Ghosh, Anne-Sophie Ville', and Nadege Thadey. 1 Executive Summary Context 1. Guinea is considered one of the poorest countries in Africa. Although its per capita GDP, currently estimated at US$ 532, is slightly above the average for Sub- Saharan Africa (US$ 520), the United Nations Report on Human Development has for several years ranked this country last among some 160 countries. Yet the country is richly endowed with agricultural, mineral and energy resources. Depending on the chosen poverty line, on the basis of data from an integrated household survey (IHS), 40 to 57 percent of the entire population may have lived in poverty in 1994. Life expectancy at birth is estimated at 45 years. And roughly 77 percent of the adult population is considered illiterate. This situation is mostly blamed on past policy failures and on the limitations of the current policy stance. 2. The development of the country since independence in 1958 may be broken into two major periods. Prior to 1985, the socioeconomic policy framework aimed at modernizing and industrializing the country. Its implementation was based on a command and control approach and led to very poor economic performance and a dismal social situation. It is estimated that real per capita GDP declined by 1I percent between 1960 and 1965, and then by about 1 percent between 1965 and 19701. From 1960 to 1974, GDP per capita declined 4 percent. Growth was estimated at 3.5 percent between 1975 and 1980. This positive growth was essentially due to the opening of some major new bauxite operations. But, as soon as these reached their full capacity in 1980, overall growth started to decline, following the trend of the non-mining economy. 3. The development strategy driven by the public sector led not only to poor aggregate growth but also to some undesirable structural changes in the economy. Increasingly, private-sector activities moved into the informal sector, conducting their business through parallel channels. While the official sector relied on mining exports for its income and foreign exchange, the informal sector depended more upon clandestine exports and remittances from abroad. The agricultural incentive structure led to a virtual cessation of production for official markets. Thus Guinea went from being a major exporter of agricultural products to a major exporter of minerals and a net food importer. 4. The poor performance of the para-statal sector, combined with the flight of private activities from official channels, led to a deterioration of public finances. The situation was exacerbated by increasing budgetary subsidies to public enterprises. The financial position of the government was further squeezed by external imbalances stemming from poorly designed investment projects that failed to yield the expected returns to meet the Arulpragasam and Sahn (1994:53). 2 World Bank (1990a:1). 2 underlying debt. This made it increasingly difficult for the government to meet its necessary investment in education, health and physical infrastructure. 5. A military coup in 1984 brought a new regime to power and the establishment of the Second Republic. This new regime initiated a public policy reversal both in political and economic activities. In late 1985, the new government launched the Economic and Financial Reform Program (PREF)3. This program was intended to lead to a radical restructuring of the Guinean economy through a significant reduction of the role of the state in the productive sphere of the economy, and an increasing reliance on the private sector to lead future socioeconomic development. The new program was seen as the only solution to the extreme failure of centrally planned economic development in Guinea. 6. The reform program produced encouraging results in its first 5 years (1986-1990). A key component of this reform was a substantial devaluation of the national currency in early 1986. This devaluation, which brought the official exchange rate in line with that of the parallel market, also marked a shift from a fixed to a managed float regime. This was a very important step in view of the fundamental role played by the exchange rate as a key determinant of a country's competitiveness. As a result of this improved incentive structure, export value increased about 16 percent between 1986 and 1989. Most of this increase emerged from the non-mining sector. During the same period, the economy grew about 4.5 percent in real terms and about 2 percent in per capita terms. This growth was more concentrated in small-holder farning and in the service sector, Furthermore, inflation dropped significantly from 72 percent to approximately 28 percent. 7. The second phase of the reform program proved more difficult than the first and coincided with political developments that may have been a complicating factor. Starting in 1991, the implementation pace slowed down significantly. Some positive results were achieved, however. Several non-viable projects were deleted from the Government's public investment portfolio, while the distribution of petroleum, water and electricity was privatized. The inflation rate was reduced from 19.4 percent in 1990 to 7.1 in 1993 and further in 1994 to about 4 percent. Non-mining revenues increased from 4.6 percent of GDP in 1988 to about 7.4 percent in 1994. 8. On the political front the new-found openness of the system allowed events that revealed to some extent the limited governance capacity within the country . Free press, strikes and demonstrations were authorized in 1990. In 1991, a new constitution was adopted and multi-party presidential elections were held in 1993. Within this new context the government has had to deal with increasing social unrest. Public-sector salaries were doubled in 1991 as a result of violent demonstrations by public sector Programne de Reformes Economiques et Financieres. See World Bank (1990a: 10). 4 Govemrance capacity is taken here to mean the ability of the state apparatus to craft and enforce the rules of the game and to arbitrate effectively any conflict that may arise either over the rules themselves or their enforcement (Frischtak 1994). 3 employees. Additional social unrest, characterized by increasing criminality, posed a direct challenge to governmental authority and created a state of uncertainty that led to a significant reduction in private investment. As a consequence, economic growth faltered. The growth rate fell from a yearly average of 4 percent in the second half of the 1980s to 2.4 percent in 1991. The current economic growth rate is estimated at 4 percent. The Distribution of Welfare and Poverty 9. Maintaining and improving the living standard of the population may be considered the ultimate objective of socioeconomic policies and one of the fundamental expectations of those who are governed. A member of a given society is deemed poor if his or her standard of living falls short of a minimum considered acceptable in accordance with prevailing social customs and cultural norms. The living standard enjoyed by an individual has an income and a non-income dimension, and may be viewed as an outcome of interaction between opportunities offered by society, and the willingness and ability of the individual to take advantage of those opportunities. Thus poverty in any country, defined as a low level of achievement in term of living standard, has both an individual and a collective dimension. 10. The enactment of a policy is bound to involve a differential distribution of costs and benefits amnong socioeconomic groups. The desirability of such a policy depends on how the gains of the winners are balanced against the losses of the losers. Its feasibility, however, hinges crucially on the actual weighing of those gains and losses within the prevailing socio-political system. Therefore, the assessment of the distribution of economic welfare within a society is an important step in the analysis of both the desirability andfeasibility of a policy proposal. When the social objective is focused on poverty reduction, then a picture of the distribution of the standard of living at one point in time provides a benchmark against which to assess future progress. It also helps identify target groups or commodities that may need special treatment in the design and implementation of poverty alleviating policies. 11. INCOME POvERTY. The assessment of income poverty in Guinea is based on the 1994 Integrated Household Survey and primarily uses two poverty lines set by the government: an upper poverty line of GNF 293,713 and a lower one of GNF 172,283. The latter is, in fact, the estimated cost of achieving the minimum daily requirement of 2,000 kilo-calories. Based on these poverty lines, three measures of poverty have been computed indicating incidence, intensity and severity respectively. Incidence relates to the proportion of the population living below the minimum standard, whereas intensity reflects the extent to which the well-being of the average poor person or household falls below the poverty line. Lastly, severity relates to the degree of inequality among the poor. 12. It is estimated that at least 40 percent of the total population live below the upper poverty line. A geographical breakdown of this information reveals that, even though urban poverty is significant, poverty in Guinea remains a rural phenomenon. In terms of 4 natural regions, Upper Guinea is the poorest, followed by Middle Guinea and Lower Guinea. Forest Guinea and Conakry represent the two least poor regions of the country. This regional ranking stays the same regardless of the poverty line used or the poverty dimension considered (incidence, intensity or severity). In fact, one gets exactly the same ranking by switching to another type of welfare indicator such as the equally distributed equivalent expenditure (this is the level of per capita expenditure which, if enjoyed equally by every individual, would provide the same level of social well-being as the current distribution). 13. INEQUALITY. In the broadest sense, income poverty in a country is determined by the size and the distribution of the national income. Several indicators of inequality were considered in this assessment. The ordinary Gini coefficient is estimated at 45 percent, while the Atkinson Index of inequality is 24 percent (when the degree of inequality aversion is set equal to one). The Lorenz structure associated with the regional distribution of per capita household expenditure, reveals that Conakry and the Forest region have a share of expenditure higher than their respective share of the total population. In terms of intra-regional inequality, the available data reveals that Conakry is the region with the highest inequality (measured by the ordinary Gini coefficient) in the distribution of per capita expenditure. Middle Guinea, the Forest region and Upper Guinea are ranked second, third and forth respectively. 14. The decomposition of the overall welfare indicator has very useful policy implications. Concentration indices may reveal which commodities are mostly consumed by the poor and which by the non-poor.5 Such information may be used to justify subsidies in well-targeted interventions for poverty alleviation. Furthermore, the ratio of the concentration index to the overall Gini coefficient may be interpreted as an income elasticity. These elasticities may help determine the progressivity of any policy, such as a tax, affecting a particular expenditure component. A policy is considered progressive if its implementation leads to a reduction of the real income Gini index. Thus, a tax on an expenditure item is progressive if the income elasticity of the commodity is greater than one. In such a case, the tax base is increasing with income because a higher burden is imposed on the rich in proportional terms. Hence, inequality in real terms decreases. 15. Income elasticities have been estimated for ten expenditure components: (1) food, (2) clothing, (3) housing, (4) durables, (5) personal care, (6) education, (7) transportation and communication, (8) recreation, (9) ceremonies and other transfers, and (10) other. The results show that expenditures on durables, personal care, transportation and communication, recreation, ceremonies and other (non-classified) do contribute positively to the overall observed inequality. The fact that the income elasticity of these 5 When a specific expenditure component is arranged in ascending order of the total expenditure and one plots the cumulative proportions of the components against the corresponding cumulative proportions of the population, one gets the concentration curve. The concentration index is equal to one minus the area under the concentration curve. This curve may lie above the line of perfect equality (the 45 degree line) in which case the concentration index is negative (Podder 1993:52). 5 expenditures is greater than one suggests that they could be targeted in the search for a base to an indirect tax system. If such a tax system could be implemented, the above results show that such a system would be progressive. Some Structural and Policy Determinants of Welfare and Poverty 16. Within a given society, social welfare, in general, and poverty, in particular, depend essentially on the structure of the economy and the policy stance. These outcomes depend also on the growth and inequality characteristics of the system. The interaction between these two factors may limit or enhance economic growth. It may also increase or reduce inequality. The structure of an economy may be defined in terms of resource endowments and the institutional setting (markets, government, community and family). 17. LAND TENURE. Traditional land-tenure systems in Guinea appear to involve severe gender bias and discrimination against certain castes within the traditional society. This maybe illustrated by the following example from the Fouta Jalon.6 In this area approximately 16 percent of farms or 13 percent of cultivated land have property titles, whereas 69 percent of all farms or 73 percent of cultivated lands are allocated according to custom. Under the latter arrangement, there is a distinction between the ownership of the means of production and claim to what is produced by these resources. The cultivator may claim all that which he or she produces with only a slight remuneration to the landowner. As a result there are no direct means to improve the agricultural system nor are there peasants without land. 18. The two groups which suffer the most from this system of land tenure are the women and the descendants of captives. The women are doubly disadvantaged because they may only borrow land indirectly through a male, and it is a male's responsibility to distribute land between members of his family. Also, even when women succeed in borrowing land they only cultivate less fertile and smaller areas. 19. Land belongs to descendants of the Fulbe who conquered the Fouta Jalon from the original occupants. These lands are passed on to the eldest son at death or divided among living sons during the landowner's life time. Ex-captives now cultivate lands alongside the Fulbe but their use rights to the land are more restricted than those rights retained by the original owners. 20. Because of the small size of the parcels, in the Fouta Jalon most farmers are obliged to borrow land from their neighbors or parents. As a result, a farmer is never certain of where he will cultivate. The limited rights translate into restricted tenure security. Security of holding in turn influences the land user's ability to make long term management decisions and to adopt practices to protect and regenerate natural resources. 6 The Fouta Jalon is a mountainous area covering most of the Middle Guinea. Fisher (1993). 6 21. Former captives' adoption of new land uses or management practices that increase the land user's claim on the land, such as the planting of permanent tree crops (trees remain the property of the planter as distinct from the owner of the land) are likely to be prohibited by the nobles who hold ultimate rights to the land. It would be interesting to investigate the extend to which this example applies to the rest of Guinea. 22. LACK OF DIVERSIFICATION OF PRODUCTIVE ACTIVITIES. The structure of the Guinean economy may be characterized in terms of the sectoral distribution of GDP and on the basis of indices of sectoral linkages. In terms of sectoral shares, agriculture represents 24 percent of GDP.7 Industry accounts for about 31 percent of GDP and the service sector about 45 percent. The sectoral linkages are computed terms of the Rasmussen Indices of backward andforward linkages for 1990. The index of backward linkage is a normalized measure of the total impact on the economy of a one unit increase in the demand for the output of the sector under consideration. A value of this index greater than one indicates that an increase in the demand for the output of this sector will generate a greater than average impact throughout the rest of the economy. Similarly, the forward linkage index is a normalized measure of the impact on a given sector from a unit increase in all final demand. A forward linkage index greater than unity means that the sector under consideration would receive a greater than average impact from an increase in all final demands. In this Rasmussen framework, sectors with linkage indices greater than one are considered key sectors. 23. The results reveal both the lack of a diversification of the productive sphere of the Guinean economy and also the weak sectoral inter-connectedness. Indeed, only the manufacturing sector shows both backward and forward linkages greater than one. However, it would be erroneous to consider this sector a key sector in light of its small share in GDP (about 5%). In the particular case of agriculture, which ought to be a leading sector in the economy in view of its share in total employment, the backward linkage is less than one, which means the stimulation of the demand for agricultural products will not result in any significant impact throughout the economy. This is probably due to the subsistence characteristics of Guinean agriculture and the fact that it does not rely significantly on inputs from the rest of the economy nor does it provide significant input to other processing industries. 24. If the agricultural sector is to play a leading role in promoting growth and reducing poverty, then policy measures must be found to enhance its productivity and its ties with the rest of the economy. In this context, it is to be noted that the current agricultural sector strategy is encoded in the Letter of Agricultural Development Policy (LPDA). Among the stated objectives are to: (1) promote food security; (2) promote agricultural exports; (3) provide essential services to farmers; and (4) ensure a rational management of natural resources. With respect to the achievement offood security, the International Economics Department data. World Bank, 1995. Lettre de Politique de Developpement Agricole, 1991, which is currently under revision. 7 government is bent on using tariffs to curtail the importation of rice. This seems to be the wrong approach to the problem, as the welfare cost to the poor would be higher than any other potential benefit. 25. A DISTORTED LABOR MARKET. The importance of the labor market is underscored by the fact that in a developing country such as Guinea, most households derive their livelihood only from the sale of their labor services. This fact underlies indeed the recommendation of labor-intensive economic growth as a fundamental element of a poverty alleviation strategy (World Development Report 1990). Labor mobility is a crucial determinant of whether certain workers will earn a premium on their labor return in the short run. Government policies will affect the way losses and gains associated with labor market adjustments will be shared. 26. On the basis of a nationally representative sample of 1,728 households living in Conakry in 1990-1991, it was found that, of the 2,565 people surveyed who were engaged in income earnings activities, 24 percent worked in the private wage sector, 24 percent in the public wage sector and the remaining 52 percent were self-employed in household based enterprises. The sample also revealed that non-wage workers earn more on the average than the wage workers. The hourly average earnings were estimated at GNF 347 in the non-wage sector versus GNF 305 for the wage sector.9 Furthermore, the average worker in the wage sector is likely to be a male with a higher level of education than the average worker in the non-wage sector. These descriptive statistics suggest labor market segmentation. 27. The results from further statistical analysis showed that: (1) age and formal education have a positive relationship with the probability of residing in the wage sector; (2) being female has a very strong negative effect on the probability of selecting the wage sector of employment. Overall, the analysis supports the view that returns to individual characteristics are significantly different in the wage and non-wage sectors. It was also found that women face high search costs and other barriers in the wage sector market and will therefore accept more readily available non-wage sector employment at lower expected earnings. 28. Further there is evidence of heterogeneity of the urban labor market in Guinea in the sense that the determinants of earnings are not the same in its different segments. The higher the level of education, the less likely is the person to be self-employed. With respect to the private wage sector, more education increases the probability for a woman to be employed in that sector but it reduces that of a man. In that particular sector, very few women are found in either unskilled jobs or in skilled trades. Furthermore, over half the women are found in professional/managerial or clerical jobs versus 16 percent for men. 9 Mills and Sahn (1996:3). 8 29. An assessment of the impact of gender alone on entry to different segments of the labor market, revealed that women have a much lower chance of entering wage employment.1° The predicted overall probability of working turned out to be lower for women than for men. Even at lower skill levels men have better chances than women. It was estimated that men who have not completed primary education have an 18 percent chance of entering the wage sector of the labor market compared with 0.6 percent for women. 30. These gender gaps in wage employment are due to endogenous factors linked to the operation of the labor market and exogenous ones related to parental investment in boys and girls education. On the demand side of the labor market, there is evidence of barriers to women's employment in the private wage sector. This could be due to the traditional reluctance of many employers to hire women on the basis of the perceived costs associated with maternity and child-rearing. This could also be due to simple biases or stereotyping. It has been established that in Guinea, graduates from female vocational training have difficulties finding jobs in their fields of training. 31. On the supply side, traditional attitudes towards work for women outside the home combined with culturally-conditioned low career aspirations may tend to contribute to the observed outcome. But this needs to be established more firmly. 32. LIMITED MANAGEMENT CAPACITY. Given the low level of human capital within the society at large, it is no surprise that the skill level within the civil service is quite low and unable to meet the requirements of the management of a modern market economy. Furthermore, a Civil Service Reform Management Review performed in 1994 (in collaboration with the World Bank) contains evidence of a civil service characterized by an incoherent structure and an irrational mode of operations. The incoherent structure stems principally from the vagueness of the Civil Service Statutes. This poor specification of the rules of the game creates ample opportunity for those in charge of enforcement to circumvent the code at no penalty. But even if well specified they might not be followed. The report cites instances where productivity allowances have been granted universally regardless of individual performance, even though the statutes require that this be linked to an annual staff performance evaluation. 33. CONSTRAINTS TO INFORMAL SECTOR DEVELOPMENT. The real challenge to the government in Guinea is, of course, to provide employment for the majority of people, in particular the poor, who have too limited skills to operate in the formal wage labor market. In this respect, the informal sector potentially holds a key to the solution of this problem, particularly in this context of public sector retrenchment imposed by the 10 The assessment is based on probabilities of sector participation by sex and education level on the basis of a separate application of a logit model to males and females setting all other explanatory variables to the mean values for the entire sample (i.e. males and females combined). The obtained results thus correspond to sector entry probabilities for a man and a woman with the same characteristics. 9 necessities of policy reforms. In Guinea, it is estimated that at least two-thirds of the economically active population are employed in the informal sector. "I 34. There is evidence that, on average, poor households derive a higher share of their revenue from independent micro-enterprises located in the informal sector, than from other activities. This shows that informal sector development has direct implications for poverty alleviation in Guinea. Most micro-enterprises in Conakry (about 73%) involve commercial activities. Most are headed by women, generally with less education than their male counterparts, who tend to reside in households with the lowest per capita consumption. 35. The next most important class of enterprises in this sector is that of small industry. These tend to be more capital-intensive and headed by men. These observations suggest that women may be facing some discrimination in the credit market. Such discrimination, combined maybe with the low level of education, makes it hardfor them to start small businesses in the non-commercial sector. 36. There are both individual and environmental constraints to informal sector development in Guinea. People operating in the informal sector, regardless of gender, are limited by their lower level of education as compared to the general population, particularly those operating in the formal sector. At the environmental level, it is reported that such development is constrained by poor infrastructure in transportation and communication. This, in turn, restricts the expansion of enterprises that may prove locally profitable. There is also a limited access to formal financial services, and therefore these enterprises cannot borrow for medium to long-term projects. Toward a Strategy for Poverty Alleviation 37. BLUE PRINT. The structure of a desirable strategy for overall social welfare improvement and poverty reduction in particular is well known. Its description may be found in many widely distributed World Bank reports such as the 1990 World Development Report on poverty, the 1993 Poverty Reduction Handbook and the 1995 World Bank Contribution to the Social Summit entitled "Advancing Social Development. " In fact, given the acute limitation of human capacity in Guinea and for ease of reference, the design of a strategy for the country relies more on the blue print presented in the latter publication. 38. According to Becker (1995), a country's standard of living is primarily determined by how well it succeeds in developing and utilizing the skills, knowledge, health, and habits of its population. This view is consistent with the human development approach adopted by the Government of Guinea in the context of the ongoing Program for Human Development spearheaded by UNDP. Furthermore, this view stems from the 1 Walker (1995). 10 fact that there is interaction between investments in human capital and the structure and performance of the economy. 39. Past and more recent performance in living standard reveal that a lot remains to be done in order to improve the quality of life for the majority of Guineans and in particular the most vulnerable. Given this objective for public policy, and the fact that overall welfare is fundamentally determined by efficiency and equity considerations in the allocation of resources, the poverty alleviation strategy proposed below is a two-pronged approach which advocates the formation of social capital along with the promotion of broad-based economic growth in order to create productive opportunities for the built-up stock of capital. 40. As stated above, the government is currently working with UNDP, the World Bank and other donors on the development of a policy framework to foster social development in Guinea. This policy framework will be promulgated in the form of a "Lettre de Politique de Developpement Humain " (LPDH). A preliminary version of this document reveals that the proposed framework is no different from the blue print discussed above. In particular, it recommends that growth enhancing policies be pursued both at the macro and sectoral levels to address some structural imbalances in the economy: (a) weak capacity of the private sector; (b) the persistence of macroeconomic disequilibria; (c) the low productivity of the agricultural sector; and (d) the lack of diversification of the economy. On the equity side, the draft letter recommends that the government deal with both regional disparities and gender differentials. 41. MACROECONOMIC STABILITY. A stable macroeconomic foundation is recognized as a prerequisite for sustainable growth and poverty reduction. Some levels and financing methods of the public deficit may lead to high inflation and the crowding out of private investment. Both of these outcomes would be harmful to the objective of poverty alleviation. Inflation would harm the poor directly while limited private investment would restrict the growth potential and hence the ability to further reduce at least income poverty. 42. In the particular case of Guinea, work done in the context of the Public Expenditure Review reveals that the government has not yet succeeded in developing a viable fiscalframework. This stems from the excessive reliance on mining revenues and foreign aid for the financing of public expenditure. Both these sources of financing are deemed very uncertain. Thus there is a needfor the government to diversify its tax base. 43. This implies necessarily measures to foster the diversification of the non-mining sphere of the economy which, in turn, cannot be accomplished without a proper strategy for private sector development (including policies to promote the formalization of the informal economy). In the search for a new tax base, it is important that the policy change be progressive (as oppose to regressive) to avoid an undue burden on the poor. 11 44. To achieve its equity objectives and promote broad based human capital development, the government is seriously considering the restructuring of public expenditures recommended in the current public expenditure review. 45. RURAL DEVELOPMENT. Given that the majority of the people draw their livelihood from the agricultural sector and that poverty in Guinea remains an overwhelmingly rural phenomenon, it is recommended that emphasis be put on rural development. The government lacks a coherent strategy in that sector. A sound poverty alleviation strategy necessarily must include a major component consisting of policy and program actions designed to achieve lasting improvements in the quality of life of the rural population. Effective rural development will require measures to increase productivity in both agricultural and non-agricultural activities, and efficiency in the storage, processing and marketing of produce. Increased productivity and efficiency will require availability of inputs, good feeder roads, effective transfer of appropriate technology and adequate research-extension services. 46. Given that small farmers dominate agricultural production in Guinea, policies and programs designed to improve their standard of living and that of other poor rural households (through increased productivity and purchasing power) can be expected to have a significant multiplier effect. An increase in small farmer incomes should stimulate the demand for locally-produced goods and services, thus strengthening the linkages between farm and non-farm productive activities. It could also reduce rural migration to urban areas with a beneficial effect on urban poverty. 47. PRIORITIES IN HUMAN DEVELOPMENT. Investments in the social sector should be primarily aimed at: i) expanding access to basic education, with a special focus on women, ii) improving the quality of education, and iii) increasing access (particularly the rural for populations) to essential health services. Increasing Access to Basic Education and Training with Special Attention to Women 48. The high rate of illiteracy in Guinea suggests that available resources should be redeployed in favor of universal primary education. Attention should be given to both formal and non-formnal education. In the formal sector, effective measures are needed to improve pupils' enrollment and retention in the primary schools. Ways must also be found to attract and retrain young drop-outs and adults willing to improve their knowledge and skills through practical or vocational training. 49. Emphasis should be placed on girls education to reduce existing gender disparities with regard to access to basic education, and priority should be given to rural and poor urban areas. 12 Improve the Quality of Education 50. Efforts to enhance the quality of education in Guinea should also include a reevaluation of the content of existing curricula in terms of their relevance to the revised objectives of the various educational programs. Given the above considerations, it is recommended that curriculum reform be aimed at increasing basic competencies in primary and general secondary schools. This may require that the number of subjects be curtailed and adequate time be allocated for teaching fundamental subjects such as language, mathematics and science. It is desirable that any curriculum be designed and implemented with a focus on outcomes relative to these subjects. This particularly requires the establishment ofperformance standards for students and teachers as well as administrators. It also entails the development of assessment systems to monitor what is taught and learned and how resources are managed to meet agreed upon educational objectives. 51. Private schooling in Guinea also has much room for expansion. While the government must retain its role in providing education as a basic service, private schools can accommodate wealthier pupils and capture the costs of some infrastructure and additional salaries, freeing public space for poorer students. Also, private vocational training centers have been expanding rapidly in urban areas, and more than 30 currently offer secretarial and managerial training. The higher education sector still requires large- scale reform. Financial and budgeting reform are needed along with rationalization of the remaining institutions of higher education. Continued improvement requires that the government perpetually reevaluates its programs and adapts to the evolving needs of the country. Implement the Health Sector strategy 52. The largest health policy problem for the Guinean government is not its sectoral strategy, which is basically sound, but rather the implementation of the strategy and the timetable set for its completion. 53. Focusing on sector quality and quantity, the government must ensure the completion and continued support of its essential drugs policy to standardize the use of basic medicines and to make them available to the public, especially the poor, at a reasonable cost. Attention must also be focused such that training and staff development are in line with the sector's human resource needs. This will require the continued refinement of efforts to improve the curriculum of medical and nursing schools and to provide specialized training courses for newly recruited personnel, as well as continued education for veteran personnel. A rational referral program must also be instituted in order to ensure that the health system treats individual cases at the appropriate level, resulting in both greater efficiency and effectiveness. The completion of this stage would be a prerequisite to the future reform of the tertiary care sector and the country's hospitals. 13 54. SOCLAL SAFETY NETS. Safety nets are intended primarily for those among the poor who cannot fully participate in the socioeconomic activity of their community. They constitute mechanisms to prevent the vulnerable groups from falling into absolute poverty. However they may also be designed to help the poor reintegrate into the economy. In the particular case of Guinea, a full picture has not yet emerged as to the identity of appropriate target groups, nor the extent and effectiveness of the traditional safety net. It is, however, commonly held that the handicapped, old people living alone and homeless children (particularly in urban areas), and some castes in the rural areas are among the most vulnerable. But no statistical evidence is available to guide social planning. Some background reports to the Lettre de Politique de Developpement Humain also list among the vulnerable small farmers, rural women, the urban unemployed and some micro-entrepreneurs in the informal sector. 55. At this stage of the Guinean experience, it may be worthwhile investigating the following issues in order to factor the result into the next policy adjustment: (1) What are the target groups and their coping mechanisms? (2) What is the nature and the effectiveness of the current social security system (3) It is widely believed that the traditional system of solidarity is breaking down in the urban areas. What is the extent of this breakdown and is there regional variation involved? (4) What role could be played by intermediate organizations such as NGOs, municipal governments and organized local conmmunities. The institutional capacity of these entities as well as their willingness and ability to work with the central government should be carefully assessed. (5) What are the political factors that would affect the feasibility of such mechanisms? Implementation Issues 56. It is useless to design a strategy if it cannot be implemented successfully on a sustainable basis. Feasibility and sustainability require political and non-political resources, and have design implications. The fact that the outcome hinges on the interaction between the individual and society implies that there ought to be a clarification of roles. This raises the issue of responsibility of the individual (family, community) and of the state. Role clarification may help with the consensus building so essential for the effective implementation of the strategy. The success of such an implementation strategy depends on the soundness of the structure of incentives (social and otherwise), and of the information policy underpinning the whole process. The information policy should ensure wide dissemination of reliable information on the life of society. Fundamentally, the implementation strategy should enable key stakeholders to effectively play their roles. 57. The implementation experience of policy reforms both at the macroeconomic and sectoral levels suggests that the existing stock of governance capacity may not be up to the task at hand. Could this mean that the perceived risk of capture of the state by some powerful social groups is so great that the government has been unwilling or unable to go all the way in making poverty reduction the vision upon which to base its necessary autonomy vis-a-vis such groups'. If this were to be the case, then the donor community 14 would face the challenge of designing, in collaboration with the Guinean authorities, a truly implementable and sustainable strategy for poverty reduction that is commensurate with the existing human and social capacity in the country. This leads us to the conclusion that, focusing on governance is critical to poverty reduction in Guinea. Monitoring 58. As in many other sectors, the legacy of the First Republic in terms of a national statistical system may be considered disastrous. Whatever statistics may have been available were provided exclusively by government departments and tended to be at variance with observed reality. The reliability of most socioeconomic data was undermined by the repression of private initiative. This led to the development of an underground and parallel economy whose activity could not be captured by the official statistics. Evidence of this statistical problem is clearly presented by the fact that all pages in the World Bank World Tables pertaining to the economic situation in Guinea prior to 1986 are basically blank. 59. Limited local capacity makes both the sustainability of poverty monitoring and the effective use of data and analyses uncertain. Yet poverty monitoring remains vital, both to inform public policy and as a means of monitoring performance. Therefore, with respect to sustainability, donor support will continue to be necessary for some time. With respect to policy impact, there is a need to learn from past mistakes and build new processes. Experience in other countries suggests that adopting more participatory approaches may enhance the sense of "ownership" and encourage the inclusion of social issues on the agendas of decision-makers. 60. There are several ways in which this may be done. Quantitative data collection may be complemented with qualitative surveys to provide socio-cultural data and to give voice to the poor and vulnerable. A broad range of stakeholders should be involved in both the design of poverty monitoring tools and in the discussion and analysis of collected data. This should help bring social concerns to the attention of policy-makers and contribute to creating greater demand for better policies. But it is not just data and analyses that need to be broadly debated. The same is true for the public policies which the former seek to inform. Thus it remains critical that the donor community emphasize, in the context of dialogue with government, civil society, and other actors, the importance ofplacing poverty at the center of government's priorities. 15 1. INTRODUCTION 1.1 Guinea is considered one of the poorest countries in Africa, although its per capita GDP, currently estimated at US$ 532, is slightly above the average for Sub-Saharan Africa (US$ 520). As Table 1.1 indicates, the United Nations Report on Human Development has for several years ranked this country last among some 160 countries. Yet the country is richly endowed with agricultural, mineral and energy resources. Depending on the chosen poverty line, on the basis of data from an integrated household survey (IHS), 40 to 57 percent of the entire population may have lived in poverty in 1994. Life expectancy at birth is estimated at 45 years. And roughly 77 percent of the adult population is considered illiterate. This situation is mostly blamed on past policy failures and on the limitations of the current policy stance. Table 1.1 Guinea: Selected Human Development Indicators (UNDP) Year Life Expectancy at Birth Adult Literacy Rate (%) Real GDP per capita HDI Rank (PPP $) (Rank/Total) Guinea Sub-Saharan Guinea Sub-Sahanea Sub-ea Sub-Saharan Guinea Africa Africa Afica 1991 43.5a 51.8 16.80 45 910 1,180 158/160 1992 43.5a 51.8 24a 51 602 1,187 160/160 1993 43.5' 51.8 24" 47 501 1,200 173/173 1994 43.9c 51.1 26.9c 51 500 1,250 173/173 1990, 1985, 1992 Source: UNDP Human Development Report 1991, 1992, 1993, 1994 1.2 The history of the modem state in Guinea starts with independence from France in 1958. The development of the country since that date may be broken into two major periods, depending on the policy stance of the regime in power. Prior to 1985, the socioeconomic policy framework aimed at modernizing and industrializing the country. Its implementation was based on a command and control approach and led to very poor economic performance and a dismal social situation. It is estimated that real per capita GDP declined by 11 percent between 1960 and 1965, and then by about 1 percent between 1965 and 1970.1 From 1960 to 1974, GDP per capita declined 4 percent.2 Growth was estimated at 3.5 percent between 1975 and 1980. This positive growth was essentially due to the opening of some major new bauxite operations. And as soon as Arulpragasam and Sahn (1994:53). 2 World Bank (1990a:1). these reached their full capacity in 1980 overall growth started to decline, following the trend of the non-mining economy. 1.3 The development strategy driven by the public sector led not only to poor aggregate growth but also to some undesirable structural changes in the economy. Increasingly, private-sector activities moved into the informal sector, conducting its business through parallel channels. While the official sector relied on mining exports for its income and foreign exchange, the informal sector depended more upon clandestine exports and remittances from abroad. The agricultural incentive structure led to a virtual cessation of production for official markets. Thus Guinea went from being a major exporter of agricultural products to a major exporter of minerals and a net food importer. 1.4 The poor performance of the para-statal sector, combined with the flight of private activities from official channels, led to a deterioration of public finances. The situation was exacerbated by increasing budgetary subsidies to public enterprises. The financial position of the government was further squeezed by external imbalances stemming from poorly designed investment projects that failed to yield the expected returns to meet the underlying debt. This made it increasingly difficult for the government to meet its necessary investment in education, health and physical infrastructure. 1.5 A military coup in 1984 brought a new regime to power and the establishment of the Second Republic. This new regime initiated a public policy reversal both in political and economic activities. In late 1985, the new government launched the Economic and Financial Reform Program (PREF).3 This program was intended to lead to a radical restructuring of the Guinean economy through a significant reduction of the role of the state in the productive sphere of the economy, and an increasing reliance on the private sector to lead future socio-economic development. The new program was seen as the only solution to the extreme failure of centrally planned economic development in Guinea. The PREF was implemented in two phases. Phase one included: (1) a large- scale devaluation to realign the exchange rate; (2) liberalization of internal and external trade by removing impediments such as price controls and state marketing; (3) restructuring the banking and financial system by closing state banks and promoting commercial banking; (4) improving public sector management through privatization, liquidation, and restructuring of para-statals as well as civil service retrenchment; (5) restructuring of public investmnent to improve productivity; and (6) commercial and institutional reforms that would promote private sector savings and investment.4 1.6 The second phase of the PREF was meant to consolidate phase one reforms. It focused on: (1) strengthening local capacity in economic policy management; (2) removing sector-specific bottlenecks; (3) improving the legal and institutional Programme de Reformes Economiques et Financi&res. See World Bank (1990a: 10). Aruipragasarn and Sahn (1993a:66-67). 17 framework; and (4) developing a social policy protective of those most vulnerable to the effects of policy reform.5 1.7 The reform program produced encouraging results in its first 5 years (1986-1990). A key component of this reform was a substantial devaluation of the national currency in early 1986. This devaluation, which brought the official exchange rate in line with that of the parallel market, also marked a shift from a fixed to a managed float regime. This was a very important step in view of the fundamental role played by the exchange rate as a key determinant of a country's competitiveness. As a result of this improved incentive structure, export value increased about 16 percent between 1986 and 1989. Most of this increase emerged from the non-mining sector. During the same period, the economy grew about 4.5 percent in real terms-about 2 percent in per capita terms. This growth was more concentrated in small-holder farming and in the service sector. Furthermore, inflation dropped significantly from 72 percent to approximately 28 percent. 1.8 The second phase of the reform program proved more difficult than the first and coincided with political developments that may have been a complicating factor. Starting in 1991, the implementation pace slowed down significantly. Some positive results were achieved, however. Several non-viable projects were deleted from the Government's public investment portfolio, while the distribution of petroleum, water and electricity was privatized. The inflation rate was reduced from 19.4 percent in 1990 to 7.1 in 1993 and further in 1994 to about 4 percent. Non-mining revenues increased from 4.6 percent of GDP in 1988 to about 7.4 percent in 1994. 1.9 On the political front the new-found openness of the system allowed events that revealed to some extent the limited governance capacity within the country.6 Free press, strikes and demonstrations were authorized in 1990. In 1991, a new constitution was adopted and multi-party presidential elections were held in 1993. Within this new context the government has had to deal with increasing social unrest. Public-sector salaries were doubled in 1991 as a result of violent demonstrations by public sector employees. Additional social unrest, characterized by increasing criminality, posed a direct challenge to governmental authority and created a state of uncertainty that led to a significant reduction in private investment. As a consequence, economic growth faltered. The growth rate fell from a yearly average of 4 percent in the second half of the 1980s to 2.4 percent in 1991. The current economic growth rate is estimated at 4 percent. 1.10 The above story raises the fundamental issue which concerns this report: given that the country is basically on the right track for economic policy reforms, there is a need to carefully assess the adjustments in the current policy stance in order to achieve a significant improvement in welfare in general and in poverty reduction in particular. To 5 Ibid. 6 Governance capacity is taken here to mean the ability of the state apparatus to craft and enforce the rules of the game and to arbitrate effectively any conflict that may arise either over the rules themselves or their enforcement (Frischtak 1994). 18 do so requires that we trace the social welfare implications of both past and current developments in a manner that helps clarify the desirability, the feasibility and the sustainability of the proposed reforms. 1.11 The analysis in this report is based on the collective choice approach to social evaluation. This framework choice is motivated by the World Bank's declaration that sustainable poverty reduction is its overall objective and a benchmark in its performance as a development institution. This declaration is equivalent to a choice of a social evaluation rule (or social welfare function). A social welfare function is a device that allows translation of individual preferences or conditions into social ones. It is structured by the underlying value judgments. In the context of a poverty assessment, such value judgments may stem from the view that more is better: (1) at the level of the individual this means more of "a good life" is better; (2) at the societal level, more equality is preferred to less. These conflicting demands create policy trade-offs between efficiency and equity. Thus, given two states of social well-being, as represented by different distributions of the living standard, the one with less poverty is preferable, ceteris paribus. Given the uncertainty created by value judgments, sensitivity analysis is extended to the choice of social evaluation functions. In this context, both Atkinson's and Foster-Greer-Thorbecke's frameworks are applied to the same data set. 1.12 Furthermore, the collective choice framework is equally suitable for the discussion of governance issues. In this context, a social state may be characterized by a configuration of diverging private interests. The existing governance capacity may then be interpreted as a social welfare function that aggregates these private interests into a desired public policy outcome. 1.13 The report is structured as follows: the second chapter briefly reviews past performance in the aggregate living standard. Chapter three describes the structure of inequality and welfare in 1994, as well as the distribution of poverty. Chapter four concentrates on the welfare implications of economic management, focusing on macroeconomic issues and rural development. Chapter five examines issues related to human capital development. Chapter six discusses the design of an implementable strategy for sustainable poverty reduction. 19 2. PAST AGGREGATE PERFORMANCE IN LIVING STANDARD CHOICE OF LIVING STANDARD INDICATORS 2.1 Maintaining and improving a population's living standard may be considered the ultimate objective of socioeconomic policy and a fundamental expectation of those who are governed.9 The standard of living is, then, the measure used to assess the performance of a socioeconomic system. An individual's reward for participation in the life of society is reflected in his standard of living. This multidimensional concept, however, is not easily measured in practice. 2.2 In the context of this poverty assessment, we adopt the view that an individual's standard of living hinges crucially on both collective and individual choices. These choices simultaneously determine the opportunities offered by the socioeconomic system as well as the individual's willingness and ability to identify and exploit them. The individual's ability is largely a function of accumulated human capital, while the set of opportunities is dependent on the efficiency and equity characteristics of the socioeconomic system. Household expenditures are our main indicator of the living standard. This measure will be supplemented, however, by such other information (indicators on health, literacy, access to social and economic infrastructure, etc.) as will help ascertain the scope of the capability set. 2.3 The purpose of this chapter is to develop the welfare implications of the socioeconomic developments that were briefly presented in the introduction. The analysis is based on both social and economic indicators of well-being. At the aggregate level (i.e., society as a whole) achievements in such things as longevity and literacy are considered the determinants of capabilities at the household or individual level (see Box 1). 9 Sen, et al. (1987:vii). [Note also that this is in fact a policy implication of Aristotle's conception of a socio-political arrangement (Nussbaum, 1986:3). According to this view, the goodness of a social arrangement may be judged on the basis of the wellness of the lives led by the participants in such an arrangement. Thus every policy-maker should strive to make people capable of choosing to live well.] . . t t E t t00 f;;00;.0. . v. . .;. .. .. ; -0 ;f:iE0;; -Approahes to measuring living stanidrds depend essentially on the i 1mporanceatt to individuals'owvn judgmentsgaboutstheir well-ing, as well as: on therange:of determnin g ftorsincluded ini the living standard indicator. IJnthe contet oftewefist approach, indicatorsteof the standard of livigare based. solely on lindividual preferences, or utility. Given the fa that utility isj unobervable, the jimpleetion. of this apprach must rely uponI proxymeasures. The assumption,0t0hen, that consumers pdfchase the best bundle. of affoirdable goods implies 0that, fior iadgroupof individals having the esame preferences and facing ithe same market constraints,; their ranking: accoing to the level of jexpenditure is equivalentto thleir ranking aordigh to the utility level, Thts Wapproach ten, ind practice, 0to: favr measures based only on thegoods 0and servicestconsumed by households,taaking into considerationmtheir size and demoahic composition . Thus, the, talt0 consumption expenditure on all goodsalnd services suitablynormalize6d fbr differing market conditions .household size tand composition is the prefeired measure of we.belitng ( 0avalion,0 9948). S0tSrictly speaking, this s a- flownmeasure of economic wlaend tendsj toignore such factors:as non-market goods and non-material human condition*swhose value wis nottralatedr into consumption.behaior,andntertempora life-tcle issues. - tThe non-welfstapproach pays fittle attention to individual preferences. oAnAssessment of well-being is instead founded on certain basic achievements. This mayj involve fr instance th ability to affordadequate nutritio'n, clothing and housig. In this contxt as well as in te welfarist approach, the command hofan individual over gods aid eservices remains anj importa insrumntaffecting well-beinig.~ 0Sen'scpab apprfoach;0 ;to tha e assessment of livin conditions distinguishes cn ng fmwcapabilities (1987, 1992) :Functionings are a set of interrelated activities ad states (d;oings and bein,gs)achieved byan imndvidual,ywhereas caabilities a0define th ptunity set,forithe individual to achiev functioig.AsSeexlIn,the cpbilt setisaloust the bdget et inconsumer tery an.elcs.h esn'reo ..to chos from posble tht b dAl9 ti 'VHtugeW:is!a di> i t ~ : t eeshfii toXd se I SI liigs (1992:40). He rpsstha this isageneral aprach to, the evaluation o oca arrangeents. uch arangements wouldb assesse,o the basisv of VIndiviual fuctoins with respectq tothe-extent they enale people to achieve the fbnctionis the deem valuable.'0 If an evaluation of a social arrangement must' be based on the extent towhich hthe indivi y .ahiee functionings -reeal the ability of the arrangementWwtoma peope function tbest,00 IthienT thied ti4State 00should focus 00dits effiorts on theformulation tind impletentation of policies that? enhaneidiidual capability0to0 function well,:0 In this0 contextone ;mayX distinguish -between. iOnt,ernal-and externa :n0calit The ;formaerHrelates to mental and physical abl ities of an individual and are enhanced by health and education.,Exter nal capabiityis deiedJ byIthe; .exerciseof internalcapability.1ASc chcircumstancesmay b viewed in tennsof socioeconomicinstittions(deiningthe rulesoftheeogame )andnthe 'quality f the oecolgical environments. These factors may impede or enhance the achievement of ihtaoning oesndingto Ca tined ternal . capabilty.b J 4 10 In this context, Sen (1992:52) insists that choosing must be seen as part of living and that "doing something" is not equivalent to "choosing to do it and doing it." Consequently, any achievement must be assessed in light of the freedom implied by the capability set. 21 ACHIEVEMENTS IN HUMAN CAPITAL 2.4 The assessment of this topic for Guinea is based on Kakwani's Achievement Index and the associated Improvement Index (1993). The Achievement Index is consistent with the view that the higher the living standard the harder it is to improve upon it. It is a function of the observed value of the living standard indicator (e.g. longevity or literacy), a lower and an upper bound, where the indicator varies from zero at the lower to one at the upper bound. Kakwani proposes the following bounds for life expectancy, infant mortality and literacy rate: Table 2.1 Bounds for Some Living Standard Indicators Lower Bound Upper Bound Life Expectancy 30 80 Infant Mortality 300 5 Literacy Rate 10 100 Source: Kakwani (1993:316) 2.5 The application of the above framework12 to infant mortality data uses the definition of infant survival rate as the number of infants per thousand live births per year who survive at least until the first birthday. This indicator is equal to 1,000 minus the corresponding infant mortality rate. 2.6 Figure 2.1 illustrates the evolution of the improvement index (i.e., change in the achievement index) for life expectancy and infant survival for selected subperiods within the sample's 1982 to 1993 period in Guinea. More specifically, the indices obey the following axioms: (1) the achievement index equals one when the living standard indicator moves from its minimum and approaches its upper bound; (2) the achievement index is a decreasing function of the initial value of the indicator and an increasing function of its end value; (3) the improvement index must be positive, negative or zero according as the standard of living is increasing, decreasing or showing no change; (4) the improvement index in a given period is the sum of individual improvement indices for the sub-period; (5) the index should give more weight to improvement occurring at higher levels of living standard (because this is a more difficult task); and (6) the index should allow meaningful comparisons of performance of any two countries at different levels of standards of living. Furthermore, the index has the structure of a constant elasticity utility function where the elasticity is interpreted as inequality aversion. The values reported here are consistent with an inequality aversion equal to unity. 12 The reader may wonder why bother with this framework instead of using untransformed social indicators. The key motivation here is methodological uniformity. Indeed, what makes Kakwani's framework attractive is the fact that it makes explicit the value judgments underlying the evaluation. Furthermore, it is consistent with Atkinson's approach to to social welfare assessment which we use in the next chapter. 22 Figure 2.1 Guinea: Performance in Longevity and Infant Survival (Kakwani's Improvement Index for Selected Years) 0.012 i 0006 _N_* Longevity c 0.004- __aIfn O Infant t> 0.002- 4Survival 4' -0.002- _ -0.004 _ -0.006- -0.008 _ First Second Second Second Republic Republic Republic Republic (1982- (1985- (1988- (1991- 1984) 1987) 1990) 1993) 2.7 Available data show that life expectancy in Guinea at birth was estimated in 1960 to be 34 years. It reached roughly 40 years in 1982, after which it declined to 39.3 years in 1984. The entire period of 1985 to 1993 shows a sligh7t improvement in life expectancy from 41 to 45 years. This improvement may not be as significant as indicated by the magnitude of the improvement index. 2.8 To understand more fully how negligible this improvement was, it should be noted that both Guinea and Mali had achieved roughly the same level of per capita GDP in 1960 ($410 and $400 respectively) and very nearly the same life expectancy at birth, 34 for Guinea and 35 for Mali. 1 3Yet from 1971 to 198 1, Mali shows an improvement of 2.7 points in life expectancy and 2.2 points in infant mortality (Kak-wani 1993:317). Furthermore, the apparent improvement in longevity in Guinea is in fact accompanied by a deterioration of living standards when measured by child survival. 2.9 The outcome with respect to education is no better (see Figure 2.2). The literacy rate in Guinea inched from 9 percent in 1965 to only 17 percent in 1985. This disappointing figure resulted in spite of the fact that the First Republic had vowed to achieve universal education. Indeed, free education was guaranteed to all. Yet despite dramatic increases in enrollments from 1958 through the early 1980s, Guinea lacked both the human and financial resources to secure adequate and relevant training. 13In terms of purchasing power parity (I1985 PPP US$) (World Bank 1994:225). 23 Figure 2.2 Guinea: Changes in Educational Achievement (Kakwani's Improvement Index for Selected Years) 0.025 =2 0.0 - 0.005 / First Republic Second Republic Second Republic (1970-1985) (1985-1990) (1991-1992) IMPACT OF ECONOMIC GROWTH ON INCOME POVERT' 2.10 This section's assessment is based on per capita private consumption as reported by the national accounts. This is analogous to using per capita expenditures drawn from a household survey as an indicator of well-being. Figure 2.3 shows the evolution of per capita consumption both in nominal and in real terms for the period 1986 to 1996.14 Poverty indices are estimated on the assumption that the Lorenz curve associated with the distribution of the per capita expenditures revealed by the 1994 Integrated Household Survey may be considered as a good approximation of inequality in the distribution of living standards during the decade under consideration. It is also assumed that economic growth has occurred in a distributionally neutral manner. The figure for 1996 is a projection. 24 Figure 2.3 Guinea: Evolution of Per Capita Consumption (Base Year 1986) 600,000 460,000 400,000 360,000 300,000 -9- --R e al- U. 250,000 200,000 -ea 160.000 100,000 60,000 0 * 1986 1987 1988 1989 1990 1991 1992 1993 1994 1986 1996 Year 2.11 To check the sensitivity of the results, three poverty lines have been used. 15 Two of these have been set by the government and a third was established by Cornell University in 1990 within the context of a study on welfare in Conakry.16 2.12 Both the government and Cornell rely in their approach on the concept of a food poverty line. All three lines are anchored to a minimum nutritional requirement of 2,000 kilo-calories per person per day. The cost of food providing this level of energy intake is then calculated and adjusted to account for non-food requirements. According to the food share approach (which both the government and Cornell use), the adjustment is made by dividing the result bt the share of food in the total expenditures of those households deemed to be poor. 2.13 Cornell used the expenditure pattern of the lowest 30 percentile of Conakry's population in 1990. This corresponds to a poverty line of GNF 20,800 per capita per month for that year. For the purpose of the present analysis, this line has been further adjusted to account for inflation and is expressed in yearly increments. Thus, for 1994, this poverty line is roughly equal to GNF 390,876. 2.14 The government of Guinea has instead chosen to base its adjustment of the food line on the expenditure pattern of the poorest 20 percent of the population. For the purpose of comparison, it is important to note that the data have been corrected to account for inflation during the survey period. Furthermore, all regional expenditure 15 Note that a poverty line is in fact a cost-of-living index in the sense that it measures the cost of a reference consumption bundle below which an individual or household is considered poor (Ravallion, 1995, Lecture on Economics of Poverty). See del Ninno (1994) or Anldpragasam (1994). 17 For more on this approach, see Ravallion (1994:33-34). 25 been adjusted to account for regional price differences. In particular, Conakry is used as the base region. Given these considerations, the government proposes an upper poverty line of GNF 293,713 and a lower one of GNF 172,283. The latter is, in fact, the estimated cost of achieving the minimum daily requirement of 2,000 kilo-calories. 2.15 Based on these poverty lines, three measures of poverty have been computed indicating incidence, intensity and severity respectively. Incidence relates to the proportion of the population living below the minimum standard, whereas intensity reflects the extent to which the well-being of the average poor person or household falls below the poverty line. Lastly, severity relates to the degree of inequality among the poor. Figures 2.4 and 2.5 show the results for poverty incidence based on the three poverty lines. 26 Percent I~~~~~~~~~~~~~~~~~~~~~~~~~I A A~~~~~~~~~~~~~~~~ *H S t-i ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ ~~~~~~~~~~~~~~~~~~t Q~~~~~~~~~~~~~~~~~~~~~~ CA~~~~~~~~~~~~~~~~ Figure 2.5 Guinea (1986-1996) Estimated Number of Poor (Multiple Poverty Lines) 4,M,000 4,000,000 - 3,SOO,000 3,000,000 - l Comell 2,500,000 -O-U- GOG A -h-~~~~~~~~~~~~~~~~~~~GOG 8- Niniunu 1,000,000 600,000 0 1986 197 1 1989 1990 1891 1992 1993 1994 1996 1996 Year 28 2.16 Figure 2.4 suggests that the incidence of poverty may have increased from 1986 to 1989, tending to decrease somewhat thereafter. This same pattern was also observed across the three poverty lines for intensity and severity. Given Guinea's history of reform, these results may reflect the up-front costs of adjustment, while their associated benefits are only to be enjoyed down stream, provided a sustained effort. Thus it appears that economic policy reform in Guinea had a positive impact on income poverty. Two issues arise immediately in connection with this conclusion: (1) how significant was this impact; and (2) what were the associated social costs? Answers to these questions are significant to the design of policy reforms. 2.17 To assess the real significance of the impact of growth on poverty (assuming distributional neutrality), we compare the observed growth in real GDP with that required to keep the number of poor at a constant level.18 Figure 2.6 shows these results. The actual growth record for the decade 1986 to 1996 is, overall, below target. The only exception to this observation seems to be associated with the results corresponding to the government's lower poverty line. In this case, observed growth exceeded the target rate for 1988 while matching it in 1993 and 1995. These results suggest that the actual growth record may not be the fundamental cause of the observed declined in poverty. We are left with two possibilities: either the trend is not statistically significant, or the decline in poverty is due to other factors such as trade liberalization that may have rendered certain of the items consumed by the poor less expensive. This remark leads us to the issue of the welfare implications of past adjustment policies. 8 Such estimates rely on the assumption that the share of the national income consumed by households also remains constant. On this see Jayarajah et al. (1996:71). It can be shown that the growth elasticity of the headcount index is equal to the ratio of the growth rate of the number of poor (corrected for the overall population growth) to the grwth rate of per capita income. Other things being equal, the higher this elasticity the lower the rate of economic growth necessary to keep the number of poor constant. 29 Figure 2.6 Guinea: Comparing Observed Growth to that Required for Poverty Maintenance 0.08 0.07 /- - 4Real GDP 0.06. -uPoverty * 0.05 - Maintenance (Cornell) X 0.04 - --Poverty OD 0.03 Maintenance (GOG) 0.02 - Poverty 0.01 , . . . . . . . . . Mainenance (GOG Min.) 0 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Year WELFARE IMPLICATIONS OF SOME PAST ADJUSTMENT POLICIES Trade liberalization and public sector retrenchment are two key components of many structural adjustment packages. These policy packages have fundamentally aimed at enhancing both the coffpetitiveness of the economy and the peiformance of the public sector. Political and social risks are associated with the implementation of these programs, however, and an assessment of these risks must rely essentially on the welfare implications of such policies. Trade Liberalization 19 Nature cf the Policy reform 2.18 The international trade liberalization initiated by the government in early 1986 entailed both a shift in the exchange rate regime from a fixed rate to a more flexible one (similar to a managed peg) and the removal of trade barriers such as tariffs and import licenses. The initial step in this process produced a marked 17-fold devaluation in the local currency. This occurred between the last quarter of 1985 and the first quarter of 1986, roughly aligning the official exchange rate with the parallel rate. The state monopoly on import and export activities was lifted in 1984 while import licensing was abolished in 1986. 19 This section draws on the findings of a study by Jehan Arulpragasain and Carlo del Ninno entitled "Do Cheap Imports Harm the Poor? Rural-Urban Tradeoffs in Guinea (See David E. Sahn, ed. Economic Reform and the Poor in AErica Oxford University Press, 1996). 30 Welfare implications 2.19 This liberalization produced significant effects on the domestic food market in terms of availability and price, thus impacting the welfare of various consumer groups. Arulpragasam and del Ninno report that these reforms were followed by a rapid increase in the volume of food imports, particularly rice, a staple food and major crop in Guinea. At the national level, rice availability augmented from 50 kilograms per capita per year between 1980 and 1982 to 70 kilograms between 1989 and 1990. This increased availability meant that on average since 1986, consumers had access to rice that was approximately 30 percent cheaper than locally grown rice. 2.20 Arulpragasam and del Ninno also report that in Conakry from 1990 to 1992, imported rice accounted on average for 60 percent of all starchy food consumed per capita and 84 percent of the total rice consumption. For poor households it was estimated that imported rice accounted for more than 68 percent of starch consumption. In terms of the combined imported staples such as rice and wheat, it was found that these items accounted for 77.4 percent of the total starchy food consumed on average in Conakry and 84.5 percent of starch consumption by the poor. It was also found that for poor households (the poorest quintile), the expenditure share of imported rice was 12.3 percent versus 2.4 percent for the most affluent quintile. 2.21 In terms of energy intake, imported rice turned out to be the cheapest source of calories in town. This commodity accounted for 32 percent of caloric intake among all households and 42 percent of caloric intake among the poor. 2.22 The large household budget share associated with rice imports implies that we should expect significant income effects associated with changes in their price. This income effect would then tend to dominate the substitution effect, leading to an inverse relationship between the price of imported rice and the consumption of such other items in the consumer budget as different grains, roots and tubers as well as other non-food items (Arulpragasam and del Ninno, 1996). It is also to be expected then, that policies causing an increase in the price of imported rice will depress the demand for many local non-tradable goods due to this dominance of the substitution effect by the income effect. Policy Lessons 2.23 The availability of cheaper imported rice has been considered by policy makers as a threat to national food security. As a result, a 10 percent tariff was placed on rice in 1990. Import tariffs on most other food items were raised from 2 percent in 1986 to 10 percent in 1988 and then 20 percent in 1990. The tariff on rice was increased in 1992 from 10 to 20 percent, while tariffs on other food items were augmented from 20 to 30 percent. There are proposals pending for additional tariff increases. The attraction of this policy instrument is that local production is expected to increase and thereby raise rural real incomes and provide an easy source of revenue for the government. An important issue must then be addressed: How desirable is such a course of action? 31 2.24 Arulpragasam and del Ninno employ a regionally focused multi-market model to demonstrate that "the welfare benefit to urban and rural households associated with cheaper imported rice is greater than the real income loss of a production decrease."20 Thus the gradual reimposition of tariffs after liberalization is harmful to both urban and rural welfare. It was also established that it would be possible to generate an equivalent amount of revenue from a tax on a commodity other than rice (i.e., sugar) at a lesser cost in terms of poverty and calorie intake among the poor. Additionally, any policies tending to lower marketing and transformation costs of agricultural produce will enhance welfare. Public Sector Retrenchment 21 2.25 Civil servants in Guinea constitute, as in many other countries, an important political base for the regime in power. Their alienation could easily cripple the implementation of many policy reforms. This makes public sector retrenchment a particularly tricky operation. Large-scale layoffs may spill over to private labor markets. Furthermore, there may be equity concerns to be addressed in the particular case of women, who may face more discrimination in these private labor markets. Finally, retrenched workers may bear a great cost of the labor market adjustment, when finding alternative sources of employment. 2.26 Throughout the 1970s, public sector employment grew at a rate of about 7 percent per year. By 1979, the state apparatus was employing some 140,830 people out of a population of 4,400,000. This implies that most of the budget was allocated to operating costs. In an attempt to exert fiscal control, nominal wages were frozen between 1965 and 1980, remaining constant in real terms between 1980 and 1984. In 1985, the average civil servant earned an equivalent of US$ 18 per month. The deteriorating employment conditions in the public sector, coupled with lack of mechanisms to ensure accountability, led to rent-seeking behavior, corruption and an increase in the cost of economic transactions. The Nature of the Retrenchment Program 2.27 The entrenchment program sought to: (i) reduce public sector employment by about 25,000 persons; (ii) craft a new pay and benefit system; (iii) sharpen the skills of the remaining employees; and (iv) increase efficiency and accountability through institutional reform. A base survey conducted between December 1985 and April 1986 20 Arulpragasam and del Ninno's (1996) model is calibrated to Lower Guinea because the data for this region are considered the most reliable and because it is in this region that the pervasiveness of imported food is greatest and the issue of imported rice competing with the local rice is the most relevant. 21 This account is based on Mills and Sahn (1996) and Mills and Sahn (1995). 32 revealed the following structure of public employment: 70,989 people employed directly by the civil service; 17,111 assigned to para-statals, state banks or as "detaches" in mining companies; 12,700 in military service; and 2,000 unclassified. 2.28 The program imposed a hiring freeze, mandatory retirement for civil servants aged 55 or those with more than 30 years of service, and ended guaranteed employment for university graduates. Optional early retirement and voluntary departure packages with substantial benefits were offered to induce workers to leave the sector. Compensation packages were also designed to "sweeten" the deal for those forced out of their jobs. Results 2.29 Another census taken in 1989 revealed that 32,639 workers had been removed from the public sector payroll since 1985. Wages were also increased dramatically during the course of the program to boost the morale and the productivity of those remaining. In 1989, measures to link pay to performance translated into a real increase in remuneration of about 23%. In 1991, however, under pressure from a national strike by government employees, the nominal base salary was increased by 145 percent-far exceeding the rise in inflation. Some Welfare Implications 2.30 Mills and Sahn studied the labor market transition paths followed by the retrenched workers in terms of factors such as earnings differentials between sectors, severance payments, gender, age and a dummy variable to account for the 1986 economic reform. They found that females are about twice (1.94) as likely as their male counterparts to leave unemployment in any given period to enter the non-wage sector.22 In fact, females are three times (2.8) more likely than males to enter this sector of employment. Based on this analysis, over 90 percent of females are expected to leave unemployment and enter non-wage employment within three years their departure from the public sector. The figure for men is estimated at 59 percent. This phenomenon could be due to the fact that women face high search costs and other barriers in the wage sector market and will therefore more readily accept available non-wage sector employment at lower expected earnings. 2.31 In a similar study published in 1995 by Mills and Sahn, the per capita consumption of former public sector departee households was compared to that of the general population in 1990. It was observed that those individuals forced out of public sector employment were particularly vulnerable, especially when no alternative employment could be found. Between 35 and 44 percent of such individuals fell, 22 Mills and Sahn (1996:16). 33 depending on whether they are redeployee or retiree, into the lowest 30 percent of the distribution of per capita consumption. 2.32 Indeed, for those who left the public sector before 1990 and succeeded in finding alternative employment, only 32 percent resided in the lowest 30 percent of the distribution, as compared with 38 percent for those who could find such employment. This confirms the difficulties in designing and implementing a retrenchment program where there is no meaningful alternative in the private sector. The situation must be even more dramatic when the departees are endowed with so little human capital to the point that they cannot use their compensation package to create employment opportunities for themselves and others. 34 3. THE CURRENT DISTRIBUTION OF ECONOMIC WELFARE POLICY TARGETING AND DATA DISAGGREGATION 3.1 The enactment of a policy is bound to involve a differential distribution of costs and benefits among socioeconomic groups. The desirability of such a policy depends on how the gains of the winners are balanced against the losses of the losers. Its feasibility however hinges crucially on the actual weighing of those gains and losses within the prevailing socio-political system. Therefore the assessment of the distribution of economic welfare within a society is an important step in the analysis of both the desirability andfeasibility of a policy proposal. When the social objective is focused on poverty reduction, then a picture of the distribution of the standard of living at one point in time provides a benchmark against which to assess future progress. It also helps identify target groups or commodities that may need special treatment in the design and implementation of poverty alleviating policies. 3.2 It is important to note that the level of aggregation of the available information has significant implications in terms of policy design. This is clear in the case where the only information available is in terms of indicators such as GDP per capita, life expectancy, etc. If a country is classified as poor on the basis of such indicators, then poverty alleviation policies would have to shoot in the dark, as it were, with the hope of reaching some poor in the process. With respect to the provision of basic social services, for instance, universal provision will be the only way to proceed. This makes obvious the necessity of a disaggregated analysis that involves the identification of and discrimination among the would-be beneficiaries of socioeconomic policies. The issue then is, as Sen puts it, "how far to push discrimination and where to stop " (1995:15). 3.3 The answer to the above preoccupation lies in the costs and benefits of targeting. In principle the argument for targeting is based on cost-effectiveness of policy implementation. This is to be understood as a matter of maximizing the poverty alleviation benefits associated with an accurately targeted policy from a given cost (Sen, 1995:12). However, there are both direct and indirect costs associated with the design and the implementation of a targeted policy. Such costs stem from the need for identification and policing and from the fact that such a policy may create incentives for beneficiaries and others to adjust their behavior. The more accurate the targeting the costlier. Thus in the search for characteristics by which to classify individuals (or households) it is desirable to consider in priority those that may be difficult to adjust (such as gender characteristics and age) and those that lead to the identification of broader groups (such as the region of residence). 3.4 For the purpose of this analysis, we focus on geographical, gender and literacy dimensions of the underlying population. Some consideration will also be given to household size. CHOICE OF SOCIAL EVALUATION FUNCTIONS 3.5 A fundamental feature of the social choice approach underlying the analysis in this report is the aggregation of individual conditions into a social state. In the particular context of welfare analysis, this becomes a matter of aggregating individual living conditions into an indicator of social welfare. Such an aggregation is performed on the basis of a social evaluation rule or a social welfare function. The structure of such a function is determined by the value judgments underlying the evaluation exercise. Given that we have chosen the distribution of some indicator of the living standard (per capita expenditure) as a social state indicator, many of the value judgments used here imply that different weights are to be assigned to different segments of the distribution of the living standard. Two classes of social evaluation functions are used in this analysis. The first class bases welfare assessment on the entire distribution of per capita expenditure. This includes both Atkinson frameworks (1970, 1983). The second class relies on a truncated distribution and includes the FGT family of poverty indices which are technically analogous to a social welfare function (Kanbur et al. 1995:92).23 3.6 All these approaches involve the use of an inequality aversion parameter which encapsulates some or all of the underlying value judgments. In the case of the first group of approaches, the higher the value of the aversion parameter the more weight the evaluation rule assigns to the poorest individual (or group). In the particular case of the FGT poverty indices, value judgments are imbedded in both the poverty line and the parameter of inequality aversion among the poor. Furthermore, the approaches chosen here all lead to indices that may be interpreted in terms of abbreviated social welfare functions. An abbreviated social welfare function is basically written as a function of two basic characteristics of a distribution of the living standard: the mean and the inequality. The structure of the selected welfare criteria is explained in Boxes 3.1 through 3.4. 23 Kanbur et al. (1995). In the same context it is worth noticing the following statement by Yitzhaki (1996:10), "The Foster, Greer and Thorbecke (1984) index of poverty is related to Atkinson's (1970) index of inequality in the same way that Sen's index is related to the Gini coefficient." 36 ..r .... ..ino .. Social flfrFuction Basic'.ally a vsoivalatiAon rule is anwexpression oftahe value judgm ents u ng a c pison of isocial states.0 This :box0 discusses iin alittle more detailfo the intereste readerthe stucur of Atison'srue '' 'Th,e'Atkinson approachto socIal evauation is based on the following value judgments. -Assumniing {that income (ortexpenditure) is: used as an: indicator of livingsadr then we take6it to be thle ase that: (1 households do not difsignificly in their non-incomecacteristicsd (homogeneity);(2)the indithevidual i th best judge of heror his well-being(individuaism)it) (3) iscomelp'rovides'td utilitfy toAincome units; more, income is preferred to less; Y(5) it does not matter socially who receives which income f(symimry), (6) overall well-ng is equal t eo te avege income utility -acrossX individuas (additivity and separabili; (7) rich-to-por income trasfirsare a 63v gothing in the sensethat t0ey increase overall welf e (iquity 4aversion).24 It 0-; 0000Cis interesting to note th0at0t the requirements that more income be preferred to lessiand tht rich-to-poor trnsfers ipov a wel frllows:the isocialwelfare finction tto hAdle the conflict ,teen social preferene for equality ain 9th for efficiency.25 Te be drk c 'oncept inis framework is tht Mof eql diribueqivaet income.; This; rsepres ients EstheIeIf per capita icme which,0 if Tenjoye equallyby ever individual, would&provideIthe 0same Vlevel of social welfare as ftie current distribdtion for some choice: of utility function.- Inthis treort,;4we *will calli ally distribue i emuexpndftur. Te numeical valueof thisinditcatoSrturns out& tobe tequal to the mean inmultiplied byi one mninus the Atkinson index ofiuit nequali It ismworth noticinghtat the mean of the income dstribetion is the maximal valueI ofthe Meql distribued equivalentAinCOme. Furthermore, for a given levelof finequalty aversion thet diffitereneietween thiemean income and the equally distributed equivalent one1is interpreted as th -per capIta cost (interms* of welfre Aloss ofinequality. i Dividing th'is dffrenceby th ye mean of the distribution, we get the At on index of relativeinequali. I Nt thatthe t rionWof gudging t:he fsocioeconomiC pem o a rcowt on the IbasisofpercapitaCDP iscorn istentw ithAtkinsonh6: approach with no aversion to inequalt. 24 See Lamnbert (1993a:207-213). 25 Jenkins (1991:24). In this context, social welfare functions that are consistent with the above value judgments are known as individualistic additively separable symmetric and inequality-averse social evaluation functions (Larnbert, 1993a). 26 This measure is analogous to a measure of risk aversion. In that context the measure would stand for what a risk averse player would be willing to pay for a sure thing in exchange for a risky promise with equivalent expected pay-off, Thus we take this to represent the maximum value (in terms of the chosen living standard indicator) that an inequality averse society is willing to sacrifice in exchange for a more equal distribution but socially equivalent to the original one (Jenkins 1991:25). 37 REGIONAL DISPARITIES 3.7 General Description: Guinea is typically divided into four natural regions (see distribution in Figure 3.1):27 (i) Lower Guinea (or coastal Guinea) covers about 18 percent of the national territory, is mostly inhabited by the Soussou and accounts for about 38 percent of the 1994 total population. This region includes Conakly, the capital city, which has a population of about one million inhabitants. (ii) Middle Guinea, also known as the Fouta Djallon, represents 22 percent of national territory, and is home to roughly 22 percent of the total population, which is predominately Peuhl (Fulani). (iii) Upper Guinea is a savanna area covering 40 percent of the total land area but accounting only for about 21 percent of the population and is predominately Malinke. (iv) The Forest region covers about 20 percent of the national territory and accounts for about 19 percent of the population. This population is made up of several small ethnic groups such as the Toma, Manon, Kissi and Guerze, collectively known as the Forestiers. 27 This account relies on information from both the 1994 Integrated Household Survey and the 1991 Priority Survey or Enquetes sur les Informations Prioritaires (ESIP), Ministare du Plan et des Finances. 38 Figure 3.1 Guinea (1994): Regional Distribution of Land Area, Population and Household Expenditure .. . ..... 40- _ 36- /-1 = 30- z , 0 Land Area e 2S- U- 25Population IL 20- j| C0 Expenditure 0 Lower Guinea' Middle Guinea Upper Guinea Forest Guinea Region * Includes Conakry 3.8 Regional Inequality: In the context of the 1994 Integrated Household Survey, Conakry has been pulled out of Lower Guinea to form a separate region. Figure 3.2 reveals the Lorenz structure of the regional distribution of per capita household expenditure. According to this figure, only Conakry and the Forest region have a share of expenditure higher than their respective share of the total population. This suggests that there may be significant regional inequality in Guinea with a preliminary indication that Conakry and the Forest region might be the most favored, and Upper Guinea the least favored of all regions. 39 Figure 3.2 Guinea (1994): Structure of the Regional Lorenz Curve 351 30- 25 2_. _ __*Share of Total c 20- Expenditure trmr U Shar of Total i. 15 i i *i Population 10 I I I 5 III. | 0 4 . Conakry Lower Middle Upper Forest Guinea Guinea Guinea Guinea Region 3.9 Given that a Lorenz curve is merely a device describing relative inequality in a distribution and that the mean and the population size of such a distribution cannot be retrieved from the information contained in the Lorenz curve, it follows that any ranking of these regions in terms of relative inequality will depend on the underlying value judgments, and any welfare ranking will need to take into consideration both the size and the inequality dimensions of the underlying distribution. 40 Box 3.2 The00 Exended (in! Coeffcient :JbThe icoefficient is the most commonly used indef relative income inequaity2t Itsi1relationfto the standard Lorenz curve stems from its definition as one minus twice the area un-derthiefiLenz curve or twice the area between the Lorenzcurve and the line ofXpeflect inequality. However, its expression does not explicitly reveal ithe value judmets underlyig the inex.0 Asit turns out, the ordinary Gini is a memberm fof a faily: of indices thatt depend on:a subjective value judgments parameter (Yitzhaki 1983). This fily of indices is representielb thexteided Gini c.oefficient.2 The normativeXcontent of the Gini faily tof indices is derived by factoringQ inthe utility function such feelings as&deprivation or ism (Lambert :1a 9b) - The computation of this coefficient requires tfe ispecification of distibutional. Value judgmentsi thgh a parameter which may be interpeted, as-an equality . pref or inequality aversion parameter. Any analysis based on this fatil of indices isMnecessarily influenced by the hoice ofth einequality aversion parameter. Operationallyt an estiate of anet ended &Gini ma betcalculated as follows (Lerman and Yitzhaki 1984): (1) raink all observations-winan increasing oder :0o income (r the selected living standard indicatr); t(2) divide the rank of each observationt b the total number of observations; (3) compute one minus this relative rank and raise-the-result to,: a power which equals the aversion parameter minus one ;O (4)co:mte the fcovariance blietwreen 0fthe income variable and theO-transformedrank obtained jin(3) fi ally (5) multiply-the covariance by the negative of the ratio of the aversion parameter to the mean Diffrent Cvalues of the aversion parameter imply that differentrelative weis ar assigned to variousgquintiles of the distribution. MI general, the&iank of the quinle receivingthe m.aximum lweight is a decreasing function of the aversi parameter.In th e fc hdase oftheordinary Gini th c,aversion parameter is equal to two; the;maximum weigt is assigned- to themedian and w,eigts are-symetrically distributed around this quintile. Quintiles below the median tend to receivef higbertweights the higher the value of the aversion parameter. Thee reverse is tuet for the quintiles above theAmedian. 3.10 Figures 3.3 and 3.4 show the ranking of natural regions in Guinea on the basis of intraregional inequality measured both by the extended Gini coefficient and Atkinson's index of inequality. Box 3.2 explains further the nature and structure of the extended Gini coefficient. 28 Relative indices of inequality are generally symmetric, responsive to transfers and scale-invariant. Indices of relative inequality are reduced if the Lorenz curve shifts inward, and unaffected if the Lorenz curve does not change. 29 The extended Gini is akin to the Atkinson Index based on the concept of equally-distributed equivalent income. This is a level of per capita income which depends on the inequality aversion parameter and is such that, if given to every individual, the resulting distribution would be socially indifferent to the initial one. 30 Lambert (1993b). 31 Lerman and Yitzhaki (1984). 41 Figure 3.3 Guinea (1994): Regional Inequality ( Extended Gini) 5 4 * -X-- Conkary --W Lower Guinea 3 - iK A~~~~~~~~~~~~~~~~~~~~~-- Middle Guinea AI + -Upper Guinea a! Forest Guinea 0 0 1.01 2 3 4 5 6 7 8 9 10 Degree of Inequality Aversion 3.11 On the basis of the ordinary Gini (i.e. when the aversion parameter is equal to 2), Figure 3.3 reveals that Conakry is the region with the highest inequality in the distribution of per capita expenditure. Lower Guinea has the least inequality. Middle Guinea, the Forest and Upper Guinea are ranked second, third and forth respectively. It is to be noted that this ranking remains constant on a wide range of choice of the aversion parameter. Indeed there is a reversal at the top only after the value of six and at the bottom only after the aversion parameter has reached 8. 3.12 Figure 3.4 presents a ranking of the regions on the basis of intraregional inequality as expressed by the Atkinson Index. This figure shows more frequent reversal in regional ranking according to the choice of the degree of aversion to inequality. Atkinson (1970:260) has found that when the aversion parameter in his framework is equal to one then, the ranking of distributions by his index is closer to the one achieved by the ordinary Gini coefficient. This result is confirmed by Figure 3.4 which shows that, when the aversion parameter is equal to one, the regional ranking is almost the same as the one obtained on the basis of the extended Gini for a degree of aversion equal to two. 42 Figure 3.4 Guinea (1994): Regional Inequality (Atkinson's Index) 5 4. ---Conkary -a-- Lower Guna t 3 1/ f \ / \ / - / / < ^ Mddle Guinea E4-=Upper Gulnea 2 m~~~~~~~~~~~~~~~~~~~~~~-- Forest Guinea 0 1 2 3 4 5 6 7 8 9 10 Degree of Inequality Aversion 3.13 Table 3.1 presents results about urban and rural inequality on the basis of Atkinson's index. The results reveal that the global index stands at 24 percent when aversion is equal to one. This means that for the chosen value judgments, only 76 percent of total expenditure needs to be distributed equally to achieve the same level of welfare implied by the current distribution. In other terms, current inequality implies a 24 percent loss in social welfare. Note that the ordinary Gini coefficient is equal to 45 percent. This indicates a significant amount of inequality in the distribution. Table 3.1 Guinea (1994): Urban vs. Rural Inequality (Atkinson) Inequality Aversion Global Index 1 2 3 4 5 6 National 0.241 0.388 0.488 0.559 0.610 0.649 Urban 0.226 0.373 0.483 0.572 0.643 0.697 Rurl 0.173 0.295 0.387 0.457 0.511 0.553 Inter Group Inequality 0.041 0.053 0.052 0.044 0.033 0.023 Intra Group Inequalty 0.200 0.334 0.436 0.515 0.577 0.625 Soure: 1994 1IS 3.14 The table also shows that, regardless of the degree of inequality aversion, urban inequality is higher than rural inequality. Furthermore, the contribution of the intergroup and intragroup components to total inequality also depends on the aversion parameter. Intergroup inequality accounts for about 17 percent of total inequality when the aversion 43 parameter is equal to one. This declines to about 3 percent as the focus on the poorest group increases. At the same time intragroup inequality increases from 83 to about 97 percent. This suggests that, with respect to the urban/rural split, intragroup inequality is more important than intergroup inequality in accounting for total inequality. Box 3.3 Decomposability ofAtkinson isIndex ofInequality In a special sense, the Atkinson index is additively decomposable between intergroup and intragroup inequality (Donaldson 1995). Jenkins (1991:32) explains that an inequality index is additively decomposable by population subgroups if it is such that: (I) having sorted the population exhaustively into mutually exclusive groups, the overall inequality may be computed from only the number of members, the mean, and the inequality in each and every subgroup; (2) an increase in inequality in one subgroup leads to an increase in the overall inequality; and (3) the overall inequality can be written as the weighted sum of intragroup inequality plus a term representing a between group inequality calculated on the basis of mean incomes and group size. The Atkinson index has this property when the mean is replaced by the equally distributed equivalent share. This idea may explained as follows. Consider the following three distributions: (a) the actual distribution of expenditure; (b) the distribution where each individual receives the equally distributed equivalent expenditure of his subgroup (there is no intragroup inequality in this case); and (c) the distribution where everybody receives the overall equally-distributed expenditure. Note that all these three distributions yield the same level of social welfare. In other terms, going from (a) to (b) removes intragroup inequality welfare neutrally from the original distribution. From (b) to (c) intergroup inequality is removed welfare neutrally. Also, moving from (a) to (c) removes the whole inequality at once (welfare neutrally). Welfare Comparisons 3.15 Inequality comparison in the context of welfare assessment provides only a partial picture of the situation since total welfare is determined by both the size and the inequality of the living standard distribution. Figure 3.5 presents the ranking of regions in terms of the equally equivalent distributed expenditure (EDE). On the basis of this figure, it is clear that, regardless the chosen value judgments in terms of inequality aversion, Conakry and the Forest region are the best off while Middle and Upper Guinea are the worst off 44 Figure 3.5 Guinea (1994): Welfare Companisons (Equally-Distributed Equivalent Expenditure by Region) 5- * * . S * S * S * 0 4 - E + Conkary - a-Lower Guinea 3*-* Middie Guinea +---Upper Guinea = 2 - A A A A * * * *- **-Forest Guinea t- * . s * * * * * . 0 1 2 3 4 5 6 7 8 9 10 Degree of Inequality Aversion Regional Poverty 3.16 Focusing now on poverty, it is interesting to notice that, on the basis of figure 3.6, the ranking of the regions for all three poverty lines and levels of aversion to inequality among the poor remains the same as the one based on the EDE.32 In this context it is quite revealing to note the correspondence between the choice of a poverty line and that of the aversion parameter within the Atkinson 'sframework. 32 The levels of aversion associated with Figure 3.6 have an ordinal interpretation. The headcount index corresponds to 1, the poverty gap index to 2, and the square poverty gap index to 3. 45 Figure 3.6 Regional Poverty (FGT Index) a 4- - *Conkary 0 3 - *-*-* -Lower Guinea -A--Middle Guinea 2 *_ - -Upper Guinea * Forest Guinea 1 2 3 Degree of Aversion 3.17 Other things being equal, a poverty line may be regarded as an equally distributed equivalent living standard. This stems from the fact that the poverty line is a commodity bundle that, if given to everybody in the society, there will be no more poverty for that period of time. Under this interpretation, we note that the Government 's poverty line of GNF 293, 713 corresponds to a value of about 2 for the Atkinson parameter of aversion to inequality.33 For this level of aversion, the welfare loss of inequality in the distribution climbs to about 39 percent. 3.18 Figures 3.7 through 3.9 give a more detailed picture of the regional distribution of poverty according the three selected poverty lines and considering incidence, intensity and severity. The exact value is 1.94425. The Govermnent lower poverty line of GNF 172,283 corresponds to a level of aversion of 5.6572 while the Comell line of GNF 390,876 is associated with a degree of aversion equal to 0.67124. 46 Figure 3.7 Guinea (1994): Incidence of Poverty by Region * Conakry 60./ tr Lowr Guinea o40 OMiddie Guinea a0. 0 Upper Guinea * Forest Guinea 20. 6~~~~~~~~~~~~~~~~~~~~~~ National GNF 390,876 GNF 293,713 GNF 172,283 Poverty Line I3ox 34 TheStrctre o Fose Greer Th orbecke Povertyn Idices Within the FGT framework, individual poverty is measured as a functon of the iniidual poverty gap. This involves aparameter sumtmarizing our valuej]udgmnents interms ~of inequ;alityl aversion among the poor. On the basis of the chosen poverty inhe (a choice that also involves value judgments), negative poverty gaps are set equal to zero. Aggregate poverty at the society's level is equal to the following weighted average of individual gaps. This defines a class of poverty indices that: (1) are,adie decomposable along a population profile; (2) satisfy the monotonicity axiom (i.e. the gap increas as incomne, dereass); and (3) indicate an increase in poverty if ceteris pan bus, the income of a poor household is reduced or there is a pure transfer from a poor household to a richer one.34 34 Additivity ensures sub-group consistency. That means if poverty increases in any sub-grup of the population without decreasing elsewhere, then aggregate poverty also increases. In this context, aggregate poverty may be written as a weighted average of subgroup poverty with the weights equal to the population share of each subgroup. 47 Figure 3.8 Guinea (1994): Intensity of Poverty by Region 35-/ 30-/ 25./ _ Conakry _ 20- / _ . .* Lower Guinea X .____________________________________ _ ] . O M iddle Guinea IL 15. 0 Upper Guinea * Forest Guinea $11 11 N National GNF 390,876 GNF 293,713 GNF 172,283 Poverty Line tox 3.5- The Normative Content of The Foster Greer and horbecke :Poerty Indicesn Poveiiy ordeins accing tor Mg an, idex have both a descripive and a native ontet. We Fnow, constder the normative content of poverty comparisons made on the basis' of te dFGT lcass of indices. We giXwill fino-e that these indices' are structurally euivalentq to oial sc wefare functionsin thle fnrdanmental sensel at Ithe social condition is in ferredfrom a weighted&average of individual poverty gaps. TMe dixtribution of these Poverty gaps may be. taken as an indicator of a social ftt FUrtbermer4 the value judments underlyingdthe evaluation process are embedded' in the ;poverty: l:ine tand he aversion. Povertycompari:sons made on the basisi of the: headcount index d0pded entirely on the chosen p.verty 0;lne,>-. but 00griven th2*e00; uncertaintiestsurrounding fde0eithe estimlation of I g sadds and povertylniesit; woJld be useifl to have a: feel:of the robustness tof povrty ranings prescribed by a Ipcular measure. Sutpposehn we would like to compare lwo different distibutionsi of ingstadars Aand Bwith respect to poverty incidenc eas measured by the headcount ratio. Ploting, for Peach disibution, all possible valuesWof 'te poo.vertyl.inealog; -theI horzontal axis and the associated values of the head-ofunt index:on thever tia laxis -yields th,e vertyjiincidence cwrve,(Ravallion, 1993). This curve is in fact the cumulative disirtibton function. wEach vpoint onthe curve reveals the percentage of the population with a standard of iving below the level measured on the horizonttalkaxis. itf in: the relevant rangqeof the'welfare indicat the ditibt ue. pertaining to:A lies nowhere iaboe t associated, withB,then regardless. of theipoverty libe chosen we: must conclude that overty is tlowerhintate AthanS in stateB.hi: other terms B idoninMates A cordig to thfirst oderdmin, ance test. What are the implications of ffirstoedomninanceI: for. povety o s made0on the basis of fhe povrt gap4findex? If the first ordertdominance condition is met, can we:still conclude tath poverty in state A islessJ thanJinW state B according tothe poy gap index? The answer is yes tbeause the poety gap measures' the area under the,povertincidence: curve, divided by the povety line (AVallion, 1993:126-129). Therefore, if thepoverty: incidence Curve A lies nowhere above curve B, then the area under A: will: beno greaterthanthe area6 under B. Likewise, fori:valuesiof the povertyaversion parameter at least equal to two;.itt :can be shon that te povet measuretis a fnction:of th poverty seve curve defined as the area underthe povertydeficit curve.~ it can be shown that the nested structure of 'the above indices and .associated curvesOipies atthe fitst ord dominance condition reveals poverty rankings not only for the head-count index,;but forigWer order pover measures as well (Ravalion1993). Wet therefore restate it as follows: lXIf' the distibution curve in wstateA is nowhere 'abvethat instate'B, 1thenpoverty in A can be no greater than in B 0regardless of the poqvtyq line:or.measure (at least within the FGT class). What happens if distributi rves do intersect? Then, no unambigWus ordering jof p tmay; bet Sclaimed: ADon. thre :basis of the head-count index. The outcome now depends on the chosen: poverty line. r'ever;if the taSea under At isstill lfessthan:the ara undero B, then measuring povertyMwithte poety gap Vindext ,and tsiquare pover:: gapt will lead t th conclusion that tiere is less poverty in A than itB.i Thisis t.he Asen-ordne0r domsnance VAtes In principle,if this test is tfinconclusive, oemay fufth strit the ange f admtissible measures of poveryand proceedtanalogously : Fostean ;Shorocks have show that (1988a:181, 1988b): (1) First-ordertochasticoinane is equivalent 0-to;: ranting by 0;social welfare ;functions th6at 4jare:: symmetric and increasin in the level 0.of livn standard; (2) orderings by s-00lthle poerN$ty gap finzdex agree with those madeon asis otonicnd Cinequality ;daverse::social welfr fuicttions t[In accordance: to Shorrocks' theorem this isI equivalentAhtothe generalized Lodomi curve dominance which reduces to the simple WLrenz dominance when the two distribttions havethe hsame mean (1983)]; oand: (3) 0povet orerings by the squaredpovert Vgap indictoare backed up by valu judgments reflecedt in:symmetic, monotoniic, inqual1ity Kaveseanditrfsensitve scialwelfafcions.; 49 Figure 3.9 Guinea (1994): Severity of Poverty by Region 20- 16 14 Conakry 12. U Lower Guinea p 10 f0 Middle Guinea 8 _ Upper Guinea -; w atU_ I Forest Guinea 0 National 2- 0- GNF 390,876 GNF 293,713 GNF 172,283 Poverty Line 3.19 With respect to the urban-rural breakdown the evidence shows that even though urban poverty is significant, poverty in Guinea remains a rural phenomenon. On the basis of the Cornell poverty line for instance, it is estimated that the rural sector contributes for about 83 percent of national poverty versus 17 percent for the urban sector. 3.20 Finally it is important to note that social indicators also confirm the regional disparities revealed by the data on per capita expenditure. The following table gives gross regional enrollment rates at the primary level and average distance from home to school Table 3.2 Guinea (1991): RegionalPrimary EnrollmentRates and Home to School Distance in Knm Region Gross Enrollment Distance Conakry 79.4 Forest Guinea 35.0 7.6 Lower Guinea 28.4 12.3 Middle Guinea 23.3 10.5 Upper Guinea 20.9 15.8 Source: Government of Guinea 50 GENDER DIFERENTIALS 3.21 Increasingly, there is evidence that human capital investment in women is socially more productive than investment in men, furthermore gender attributes are not easily adjustable. For these reasons it is important to investigate the gender dimension of welfare in Guinea.3 3.22 It is estimated that women represent 51 percent of the total population in Guinea. The overall gender ratio is about 96 men for 100 women. This ratio is below 100 for all regions except Conakry and Upper Guinea. In Conakry for instance, the ratio is estimated at 106. The observed reversal from these two regions is mainly a result of migratory movements (by men in search of opportunities) to the capital city and to gold and diamond mines in Upper Guinea. On the basis of age groups, there tend to be more men than women up to the age of 19. Women tend to dominate the age group 20 to 54. And above the age of 55, there tend to be more men than women. Figure 3.10 Guinea (1994): Inequality by Gender of Household Head (Atkinson Index) 4 3 -h-Urban Males o /-U-=Urban Females J 2 - // i- -fi Rural Males -0- Rural Fenmas 0 0 1 2 4 6 Degree of Inequality Aversion 35 See Schultz (1994). The author reviews evidence that not only such investment brings high individual market remns, it is also associated with positive externalities such as reduced child mortality and fertility, involves intergenerational redistribution in the form of better heath and education of children and slower population growth. 51 Figure 3.11 Guinea (1994): Welfare Comparisons by Gender of Household Head (Equally- Distributed Equivalent Expenditure by Gender) 3 ;_ I -a-- Urban Males 2_-- Urban Females _ ,4- ___ Rural Males -0--Rural Females 0 I I 1 2 4 6 Degree of Inequality Aversion 3.23 Figures 3.10 through 3.12 show respectively gender comparisons of inequality, welfare and poverty. The distinction is based on data pertaining to the gender of the head of household. The data show that inequality tends to be higher among female-headed households than male-headed households. Furthermore, taking into consideration both the average expenditure and the within group inequality, female headed households tend to be better off and less poor than male-headed households. 3.24 It is to be noted that with respect to poverty, the pattern remains the same for all three dimensions: incidence, intensity and severity. Furthermore, the result that female- headed households tend to be better off than male-headed ones is consistent with what is revealed by the analysis of data from Cote d'lvoire (Kakwani 1993a:51).36 Given that the sample included only 10 of the population in female-headed households and not all these persons are females, the true picture of gender inequality may have to be sought elsewhere. For this we turn to the labor market where most Guineans earn their living. 36 In this context, Kakwani cites an explanation from Glewwe(1988). The explanation given in this case is that female-headed households are disproportionately located in urban areas which are richer than the nrual areas. However, in the case of Guinea, the data indicate that 10 percent of the total population live in female-headed households: 4 percent in urban areas and 6 percent in rural areas, so this explanation does not apply. 52 Figure 3.12 Guinea (1994): Incidence of Poverty by Gender of Household Head 70- 60 / | - ||a. * Urban Male so - 2~~~~~~~~~~~~~~~~~~~ Urban Female 40. ~~~~~~~~~~~~~~~~~~~~~0 Rural Male 30. 0~~~~~~~~~~~~~~~~~~~~ Rural Female 20- N~~~~~~~~~~~~~~~~~~~~~ Population GNF 390,876 GNF 293,713 GNF 172,283 Poverty Line Bo 3.:6The Impeortance of Labor Markets A policy reform or a shock dwill affect employment and eamings throug ths intrec6nnectedness of productlabor and capital markets. Varous adjustments that occur in the piOduct maket will affec th derived demand for labor depending on lat l ensity of d;iet sectors of activities. Changs in tedeed demand will ultely act household livelihoods. Labor mobility is a crucial determinant of whether certain workers will earn a pemimonth0eirlabor return in the short run. Government policies will afct the way losses and gainsassociatd with these adjustmes will be shared. Ptoduct market adjustments do affect labor supply as well. An incrase in the prices of Ceinai essential commodities may Read to an increase in labor force participation by other members of the household besides the main income eamer (e.g. school-age children). The sme changes may lead to people working longer hours or holding miultiple jobs. The importance ofthe labor market is further underscored by the fat tha in a developing country such as Guinea, most households derive their livelihood only frm the se of their labor services. This fat underiies indeed the trecommendation of labor-intensive economic growth asa fimndariental element of a poverty alleviation strategy (World Development Rert 1990). 53 Box 347 The Sr rofshe Urban Lbor Market in G-inea This bo summarizes thefin ings of several studies: by Corell Univest based on :a household survey conducted in 1990 Teo focus is on segmenation1 participation and: eaings. A labor market is any pess that brings together the supply of and demiand -for a ceran ype of labor Labor may be differentiated by such characteristics as skills, otio, gender Of the supplie, etc. Labor.market seentadion exists when similar worker charcteristicsreceive significantly different rates of returznt according to the subsector of employent. Tmhis hetrogeneity is a structural rigiditywith potential implications for the transition:costs zing laid-offworkers. This is a fundamel facor to be considered in the desig .and implemetawtion of public sector reform ptograms. For the urban labor market in Guinea, Mills and Sahn (1996) present evidence of segmentation between theVwage and non-wage sector. On the, basis of a nationally representative sample of 1,728 households living in Conakry in. 1990-1991 it was found thaft of the 2,565 people surveyed who were engaged in income eanings activities, 24 percent worked inthe private wage sector, 24 percent in the publicwae- sector and the remaining 52 percent were self-employed in household based:e.ntrprises,. The sample also revealed that non-wage workers earn more on the average than the wage workers. The hourly average earings were estimated at GNF 347 in the non-wage sector versus, GNF 305 for thfe wage sector? Furthermore, the average worker in the wage sector is likely to be .a muale with a higher level of education than the average worker in, the non-wage sectr. These descriptive statistics suggest labor market segnentation. The results from fiurther statistical analysis showed that: (1) age. and formal education. have a positive relationship with the probability of residing in the w:ag sector;. (2) beinig female has -a vey strong negative effect on the. p.robability of selecting the wage sector of employment4 Overall, the analysis supports the view that returns to individual characteristics:are significantly differt in the wage and non-wage sectors. It was also found that women face high.search cost and other barriers in the wage sector market -and will therefore accept maor readily available non-wage sector employment at lower expected eamings. 3.25 Glick and Sahn (1995a ) provide an in-depth analysis of labor market issues in Guinea using the same Conakry data set. They consider four possible employment outcomes for both male and female participants: (1) nonparticipation; (2) self- employment; (3) private sector wage employment; and (4) public sector wage employment. With respect to market segmentation, they investigate whether these labor market segments differ significantly with respect to the determinants of entry and earnings with particular emphasis on gender differentials. Their analysis is based on a multinomial logit model.33 The next two sections report on their findings. 37 Mills and Sahn (1996:3). 38 See Annex. 54 Participation 3.26 Labor force participation involves all individuals who are currently employed as well as those who are looking for work. On the basis of a sample of 1,725 households comprising 3,566 men and 3,306 women 15 years of age and older, Glick and Sahn (1995) found that the participation rate for men between 30 and 50 was close to 100 percent. The rate for women in the 15 to 65 age bracket was estimated at 40 percent. Men are relatively equally spread across the three sectors of employment: (1) private wage; (2) public wage; and (3) self employment. Women tend to be concentrated in self- employment and those who are self-employed tend to come from the lower deciles of per capita household income distribution. 3.27 With respect to educational attainment, it was found that both men and women employed in the public sector tend to be well educated. Men have an average of 10.2 years of schooling compared to 11.9 years for women. Average years of schooling are very low in the self-employment sector: about 3 years for men and 2 for women. In the private wage sector, women average over 9 years of schooling and men only 5. 3.28 Further analysis confirmed the heterogeneity of the urban labor market in Guinea in the sense that the determinants into its different segments are not the same. The higher the level of education, the less likely is the person to be self-employed. With respect to the private wage sector, more education increases the probability for a woman to be employed in that sector but it reduces that of a man. In that particular sector, very few women are found in either unskilled jobs or in skilled trades. Furthermore, over half the women are found in professional/managerial or clerical jobs versus 16 percent for men. 3.29 In an assessment by the authors of the impact of gender alone on entry to different segments of the labor market, it was found that women have a much lower chance of entering wage employment.39 The predicted overall probability of working turned out to be lower for women than for men. Even at lower skill levels men have better chances than women. It was estimated that men who have not completed primary education have an 18 percent chance of entering the wage sector of the labor market compared with 0.6 percent for women. 3.30 The study explains that these gender gaps in wage employment are due to endogenous factors linked to the operation of the labor market and exogenous ones related to parental investment in boys and girls education. On the demand side of the labor market, there is evidence of barriers to women's employment in the private wage sector. This could be due to the traditional reluctance of many employers to hire women on the basis of the perceived costs associated with maternity and child-rearing. This 39 The assessment is based on probabilities of sector participation by sex and education level on the basis of a separate application of a logit model to males and females setting all other explanatory variables to the mean values for the entire sample (i.e. males and females combined). The obtained results thus correspond to sector entry probabilities for a man and a woman with the same characteristics. 55 could also be due to simple biases or stereotyping. It has been established that in Guinea, graduates from female vocational training have difficulties finding jobs in those fields. 3.31 On the supply side, traditional attitudes towards work for women outside the home combined with culturally-conditioned low career aspirations may tend to contribute to the observed outcome. But this needs to be established more firmly. Earnings 3.32 Glick and Sahn (1995a: 13-20) found evidence that the determinants of earnings do vary from sector to sector. This is further evidence of heterogeneity of the urban labor market in Guinea. For the particular case of education it was found that the returns to schooling do not vary significantly across segments of the labor market. Their findings may be summarized as follows: * Male Education: (i) in self-employment men who have completed secondary education earn twice as much as those who have not completed primary school; those who have attended university earned three times as much as the reference group; (ii) in the public sector returns to primary education were slightly higher than in the private wage sector, but the situation reverses itself at high level of education: post primary education pays more in the private sector than in the public sector. - Female Education: (i) in self-employment women who completed primary school earn 30 percent more (hourly) than women who did not. Unlike the case of men, post primary education has no significant impact on what women earn through small enterprises in self-employment. This suggests that post-primary education may not be useful for the types of activities in which self-employed women are involved. (ii) in the public sector, the returns to female education are significant at all levels. No significant results could be established by the authors in the case of private sector employment. * Experience: It was found that (i) in private wage employment, age is associated with higher earnings for men; (ii) for both men and women in self- employment, occupational experience had a similar effect; and (iii) in the public sector, occupational experience has a small but significant impact on women's wages (but not for men).40 * Enterprise Capital: measured by a proxy: the value of capital purchases made for the enterprise over the past year. It was observed that businesses run by men employ more capital than those run by women. This may be a reflection of the constraints that women face on the credit market. 40 This was captured by the number of years of occupational experience and dummies for the age. 56 Gender Dimension 3.33 Average hourly profits for men exceed those for women by 120% in self- employment (Glick and Sahn, 1995a:20-25). In the public sector average male earnings exceed female earnings by about 20%. However, in the private wage sector female wages are slightly higher than males'. This could be due to superior education of the female employees in the private sector. These observed differences raise the issue of whether they are due to discrimination in pay or to such characteristics as education and experience. Box1381 The Efet f ohr 1 Epoment on Child Nutto Based o n tsur vey data n fo Con co akty and using the anthrpometric foutcomeof child Wheight Stdardized eand sex as an indicator of nutrition sas heightor-geGlik and Sahni (199*16 foud that.: (1) thereiJs~ a~ tradeoff between the negatve effects,on childcare assoad with the mothers labor karket partipatio and the psitive effets of the derived -nome; e,j(2) the 0overall ;effect ils 3however fnegative; i(3) ladditions tto ;materal ltabor ;income y5ield; Jan improvemetinf ithEe titioal j statmstI offthe chld t;at is great than that associated with eqivalent additions: to. non-ma er houeholdhincome. Th is sgesthmat mother cares e est.n:: other *erms,)" household itncome may nothbecomlet ely pooled withinthe ho Ihold and tht preferenes dffrt among famiymembers over sthe allocation of resource sto chldeare, versus teruse.s' tPocy ImPlimtons Inetg iwomen s education or rducing Igenderbiasin credit. kmarkets Xxwill im;lxprove ;both dwomen's earingstandwill be more efetive- iimpvg cil nutritio'n thanmeasuresttthatindiscr-iinaly inctease:i overallthousehold'income. H OWeverthere is; a- ntitional risk associatedwith tis poliS y Xthat may be dealt withbyhthe 6provision dof aeuate childIare alternative for mothes woenter thie labor force. hiis afford sanopporunity for privat e seor deve,lopment LITERACY, HOUSEHOLD SIZE AND POVERTY 3.34 With respect to literacy, the results from the integrated survey show that 81 percent of the total population reside in households whose heads are illiterate. Among households with literate heads, 3 percent of the population reside in households whose head has finished primary school, 2 percent in households headed by graduates from first level of secondary school and 1 percent in households headed by second level of secondary school graduates, 3 percent in households headed by graduates from a technical school, 3 percent in households headed by university graduates, and 8 percent in households whose heads have some sort of education. These results are presented in Table 3.3. 57 Table 3.3 Guinea (1994): Poverty Comparisons by Education of Household Head Number of Education Level of Households in Population Per Capita Incidence Intensity Severity Head of Household Sample Share Expenditure Illiterate 3,109 80.5 399,827 62 25 13 Primary Graduate 188 3.1 596,010 42 13 6 Secondary I Grad. 124 1.8 746,561 19 5 1 Secondary 2 Grad. 100 1.3 755,408 17 4 1 Technical Graduate 247 3.2 919,039 16 4 2 University Graduate 222 2.5 1,676,167 5 1 0 Some Education 426 7.6 519,828 53 16 7 All Literate 1,307 19.5 781,325 34 10 4 National 4,416 100 474,332 56 22 11 Source: 1994 IHS Household Size 3.35 Household size ranges from 1 to 39 individuals. The median household contains 6 persons. For the purpose of this analysis we consider three classes of households. Households containing 1 to 4 individuals will be considered small. They account for about 15 percent of the total population. Medium-size households have 5 to 6 members and account for 21 percent of the total population. Finally, households of seven or more members are considered large. They account for about 64 percent of the entire population. This classification follows the one use by Kakwani in 1993 in a study on C6te d'Ivoire.41 Table 3.4 Guinea (1994): Poverty Comparisons by Household Size Household size Number of Population Per Capita Incidence Intensity Severity in the Sample Share Expenditure I to 4 persons 1,571 14.5 715,017 34 10 4 5 to 6 persons 996 20.9 486,467 53 21 10 27 persons 1,849 64.6 416,446 62 25 13 Total 4,416 100 474,332 56 22 11 Data Source: 1994 IHS 3.36 The above results confirm those from several other empirical studies showing that larger households are more likely to live in poverty. 41 Kakwani (1993a:53). 58 Expenditure Components 3.37 The decomposition of the overall welfare indicator has very useful policy implications. Concentration indices may reveal which commodities are mostly consumed 42 by the poor and which by the nonpoor. Such information may be used to justify subsidies in well-targeted interventions for poverty alleviation. Furthermore, the ratio of the concentration index to the overall Gini coefficient may be interpreted as an income elasticity (Yitzhaki 1994). These elasticities may help determine the progressivity of any policy, such as a tax, affecting a particular expenditure component. It is well known that a policy is considered progressive if its implementation leads to a reduction of the real income Gini index. Thus, a tax on an expenditure item is progressive if the income elasticity of the commodity is greater than one. In such a case, the tax base is increasing with income because a higher burden is imposed on the rich in proportional terms. Hence, inequality in real terms decreases. 3.38 Table 3.5 presents income elasticities estimated for ten expenditure components on the basis of extended concentration coefficients. Table 3.5 Guinea (1994): Inequality by Expenditure Components (Income Elasticities) Level of Inequality Aversion Component Group 1.01 2 3 4 5 6 7 8 9 10 Food 0.529 0.779 0.822 0.845 0.861 0.873 0.882 0.889 0.895 0.900 Clothing 0.824 1.016 1.037 1.044 1.047 1.048 1.048 1.049 1.048 1.048 Housing 1.294 0.872 0.844 0.834 0.830 0.827 0.826 0.825 0.825 0.825 Durables 2.294 1.285 1.244 1.271 1.197 1.183 1.171 1.162 1.154 1.147 Personal Care 0.706 1.105 1.124 1.128 1.126 1.124 1.121 1.118 1.115 1.113 Education 0.412 0.812 0.910 0.966 1.001 1.023 1.038 1.048 1.055 1.060 Transportation & 2.529 1.597 1.464 1.398 1.357 1.329 1.309 1.292 1.279 1.268 Communication Recreation 1.118 1.510 1.407 1.354 1.321 1.298 1.280 1.267 1.256 1.247 Ceremonies and 0.824 1.178 1.158 1.141 1.128 1.117 1.108 1.101 1.094 1.089 Other Transfers Other 0.941 1.336 1.289 1.263 1.246 1.235 1.226 1.219 1.213 1.208 Gini 0.017 0.453 0.573 0.633 0.669 0.694 0.713 0.727 0.739 0.748 Source: 1994 IHS 3.39 The structure of these elasticities is such that we can also tell whether a particular component has a negative or a positive effect on total inequality. As Podder (1993:53) explains it, if a component is proportional to total expenditure it will have no particular effect on total inequality. Its concentration index will be equal to the overall Gini. Thus, 42 When a specific expenditure component is arranged in ascending order of the total expenditure and one plots the cumulative proportions of the components against the corresponding cumulative proportions of the population, one gets the concentration curve. The concentration index is equal to one minus the area under the concentration curve. This curve may lie above the line of perfect equality (the 45 degree line) in which case the concentration index is negative (Podder 1993:52). 59 income elasticity will be equal to one. If the component rises more than proportionately with total expenditures, however, the associated income elasticity will be greater than one and we may conclude that the component has a positive effect on overall inequality. Similarly, an income elasticity less than one will lead us to believe that the component has a negative effect on total inequality. 3.40 Based on the above interpretation, it is clear that expenditures on durables, personal care, transportation and communication, recreation, ceremonies and other (non- classified) do contribute positively to the overall observed inequality. The fact that the income elasticity of these expenditures is greater than one suggests that they could be targeted in the search for a base to an indirect tax system. If such a tax system could be implemented, the above results show that such a system would be progressive. This conclusion is invariant to the underlying value judgments as the elasticities remain substantially above one for a wide range of values in the domain of the inequality aversion parameter. 60 4. STRUCTURAL AND POLICY DETERMINANTS OF WELFARE AND POVERTY 4.1 Social welfare in general, and poverty in particular, within a given society depend essentially on the structure of the economy and the policy stance. The outcome depends also on the growth and inequality characteristics of the system. This chapter focuses on the implications of economic management both at the macro and sector levels, subject to the structural features of the economy. The emphasis is placed on the factors that either limit economic growth or increase inequality. LAND TENURE AND WATER RESOURCES 4.2 One important characteristic of the poor in general is that they tend to be located in areas where investment opportunities are limited and rates of return low. Two types of assets are to be considered: material assets such as land and credit, and human capital. In many developing countries landlessness is a prime cause of rural poverty. Thus an important means of alleviating rural poverty may involve some sort of land reform. However, there are severe constraints to the implementation of land reform. The reform tends to be administratively too complex and politically too difficult and may take a long time. Furthermore, land reform may disrupt production at a critical time. This would make governments quite reluctant to meddle with land tenure systems if doing so entails a short-run loss of output.45 4.3 There is evidence that traditional land-tenure systems in Guinea may involve a severe gender bias and discrimination of certain castes within the traditional society, as illustrates the following example from Fouta Jalon.46 In the province of Fouta Jalon, the US Agency for International Development (USAID) conducted a study of natural resource management and land tenure which was released in April 1995. Approximately 16 percent of farms or 13 percent of cultivated land have property titles, whereas 69 percent of all farms or 73 percent of cultivated lands are allocated according to custom. Under the latter arrangement, there is a distinction between the ownership of the means of production and claim to what is produced by these resources. The cultivator may claim all that which he or she produces with only a slight remuneration to the landowner. This S Demery and Addison (1987). 46 Fisher (1993). is particularly true in Middle Guinea. As a result there are no direct means to improve the agricultural system nor are there plantations or peasants without land. Gender Differences 4.4 The two groups which suffer the most from this system of land tenure are the women and the descendants of captives. The women are doubly disadvantaged because they may only borrow land indirectly through a male, and it is a male's responsibility to distribute land between members of his family. Also, even when women succeed in borrowing land they only cultivate less fertile and smaller areas. 4.5 Women's access to the land and right to the produce of the land are determined according to one of the following criteria: (1) frequently the titular owner of the land receives a payment after every harvest (between 1/10 to 1/3 of the crop); (2) the woman must contribute labor to certain projects undertaken on the parcel lender's own property; or (3) the women's children need to be available for Kile in the fields of the landowner. Also women were traditionally restricted to the cultivation of "suntuure" or inner fields, not the outer fields that are the exclusive domain of men. 4.6 The intensive cultivation practiced in the suntuure requires that soil fertility be maintained through consistent manuring and mulching. Maintenance of this fertility is the sole responsibility of a woman, and her investment in soil building will vary according to her work load and physical strength. Outer fields are traditionally the men's domain. People prefer cultivating valley-bottom fields but Forestry Agents have increasingly prohibited farming in these areas 4.7 With the decline of rainfall since the 1970s, yields have reduced in both inner and outer fields. In the latter case, the villagers have abandoned long-season, slow-growing varieties Due to population pressures and the governmental restrictions on land use fallow periods have shortened and today average between three to ten years. On river bank and valley bottom fields the fallow rarely exceeds seven years. Fears of reprisals and heavy fines have led to the abandonment of these fields for longer periods. 4.8 Many women now cultivate parcels in outer fields on a portion of their husband's fields. This is because, despite the fact that the fields were traditionally the domain of men, the male exodus for wage labor has left women to fill the gap. Caste or Class Differences 4.9 Land belongs to descendants of the original occupants or conquerors, the Fulbe. These lands are passed on to the eldest son at death or divided among living sons during the landowner's life time. Ex-captives now cultivate lands alongside the Fulbe but their use rights to the land are more restricted than those rights retained by the original owners. 62 4.10 Because of the small size of the parcels, in Fouta Jalon most farmers are obliged to borrow land from their neighbors or parents. As a result, a farmer is never certain of where he will cultivate. The limited rights translate into restricted tenure security. Security of holding in turn influences the land user's ability to make long termn management decisions and to adopt practices to protect and regenerate natural resources. 4.11 Former captives' adoption of new land uses or management practices that increase the land user's claim on the land, such as the planting of permanent tree crops (trees remain the property of the planter as distinct from the owner of the land) are likely to be prohibited by the nobles who hold ultimate rights to the land. The Land Code 4.12 In March 1992 a new land tenure code was promulgated based on the right to private property. This was designed to favor private investment and to support efforts for farmers to obtain credit. In practice, however, this system is subject to numerous constraints above and beyond those associated with costs and technical implementation. 4.13 There is no private appropriation of land possible on certain bas fonds where the relationships between existing titular owners of rights and the residents are inextricable and every act of authority runs the risk of reducing the access of the poorest to irrigated land. 4.14 Because of the degree of indebtedness of rural populations, particularly the smaller farms, the use of the land as security could rapidly lead to loss of the land and its acquisition by larger farming operations. Also, there are no established systems for agricultural worker compensation outside of farnily relationships. The following table describes various categories of land rights. Table 4.1 Guinea: Traditional Land Tenure (Fouta Jalon) Right of use Description Limited Rights to grow annual crops and to collect frewood and fruits Full Rights to grow annual and perennial crops, to make certain permanent improvements to collect firewood and fruit among others Transfer rlghis Rights to lend, rent, mortgage, give, bequeath and sell Limited Land user has no permanent transfer or alienation rights but may have some temporary privileges to lend or rent the land Preferential Limited to transfer within family or lineage Complete Total right to alienate land to those outside of the village 63 Water Resources 4.15 Rainfall has declined since the 1970s with negative implications for water resources. This has affected crop choices and cultivation location and has caused herd animals to be kept further afield. 4.16 The gender and class differences described in the study underscore how difficult it is to actually convert a land tenure system and transform an agricultural community with diminishing resources and competing claims to water, fields for cultivation and grazing. The allocation of resources and property rights continue to follow patterns established prior to the First Republic and even prior to the colonial period. Now because of government environmental management concerns, there is the additional constraint of government inspectors preventing cultivation of some of the most fertile land. 4.17 Rights and security in land in the watershed vary according to social rank gender, age and resident status. The complex array of land rights and the variability that exists in and between villages will affect what project interventions should be selected and for whom. There is a widespread fear among villagers that they will lose their land to the state or project. 4.18 Grazing resources are considered common-pool assets, though the land on which the forage is found is owned privately. The inhabitants of the watershed have very few rules managing and regulating herd activities. Village territorial limits have little significance in relation to pastoral activities Herds crisscross village territories, especially during the dry season, facing virtually no restrictions. The villagers are not at all satisfied with the state's "early burning" policy for the bowal. They argue that application of the policy causes more harm than good, and they resent their loss of autonomy in this resource-management activity. 4.19 Water resources are recognized as common-property resources and are managed communally by villagers. The management rules have fallen into disuse with increased dependence on household wells. Many water sources are considered to be filled with spirits and supernatural powers, which sometimes affects the local inhabitants' use of the resource. More than one village can use a single water source, though all natural springs are located within village territories and are common property of that village. In the past there has been open access to fish in seasonal and perennial ponds. Project investment in fishing resources may increase the interest in and competition for access to these resources. Some Policy Implications 4.20 Devolve power to the local level. This may involve the creation of Local-level Natural Resource Management ( NRM) institutions should be created. However, adoption of the terroir villageois approach should remain supple in its operation, with constant awareness of the specificities of the place and social groups to whom it is being applied. Efforts need to be made to insure that the villagers acquire the necessary skills, 64 information, and knowledge of their local NR base to make local-level management feasible and sustainable. 4.21 A longer-term resource management approach is needed. And given that villagers are fearful of losing their land to the state or to a project, development practitioners need to be extremely sensitive to this and diminish their concerns. In this context, it should be noted that villagers often confuse an NRM "project" with the "state" and its desire to "take land from the them to develop and use." 4.22 Rights to trees and rights to land where trees are to be planted need to be discussed. Tree ownership is linked to land-ownership in Fouta Jalon. Tree planting is therefore strictly controlled to prohibit non-owners of land from planting trees. Ex- captives and women are the most affected, as neither group is able to freely plant trees. 4.23 A grazing land policy is needed. Grazing resources are considered common-pool assets, though the land on which the forage is found is owned privately. The inhabitants of the watershed have very few rules managing and regulating herding activities. 4.24 Project managers should choose water sources in consultation with local inhabitants, because traditional belief systems dictate which water source is used. Water resources are considered common property. Management rules have fallen into disuse with increased dependence on household wells. Furthermore, open access to fishing resources needs to be re-investigated as this was common in the past. LACK OF DIVERSIFICATION OF PRODUCTIVE ACTIVITIES 4.25 The structure of the Guinean economy may be characterized in terms of the sectoral distribution of GDP and on the basis of indices of sectoral linkages. In terms of sectoral shares, agriculture represents 24 percent of GDP.47 Industry accounts for about 31 percent of GDP and the service sector about 45 percent. The sectoral linkages are shown in Figure 4.1 in terms of the Rasmussen Indices of backward andforward linkages for 1990. The index of backward linkage is a normalized measure of the total impact on the economy of a one unit increase in the demand for the output of the sector under consideration. A value of this index greater than one indicates that an increase in the demand for the output of this sector will generate a greater than average impact throughout the rest of the economy. Similarly, the forward linkage index is a normalized measure of the impact on a given sector from a unit increase in all final demand. A forward linkage index greater than unity means that the sector under consideration would receive a greater than average impact from an increase in all final demands. In this Rasmussen framework, sectors with linkage indices greater than one are considered key sectors. 7 International Economics Department data. World Bank, 1995. 65 Figure 4.1 Guinea (1990): Sectoral Linkages (Rasmussen Indices) I- -* Agriculture 1.4- / O Mining o Manufacturing 0 O Public Works and Construction la 0.6- 8 Private Services U) 0.4- 0 Public Services Backward Forward Linkage 4.26 These results reveal both the lack of a diversification of the productive sphere of the Guinean economy and also the weak sectoral inter-connectedness. Indeed, only the manufacturing sector shows both backward and forward linkages, greater than one. However, it would be erroneous to consider this sector a key sector in light of its small share in GDP (about 5%). In the particular case of agriculture which ought to be a leading sector in the economy in view of its share in total employment, the backward linkage is less than one, which means the stimulation of the demand for agricultural products will not result in any significant impact throughout the economy. This is probably due to the subsistence characteristics of Guinean agriculture and the fact that it does not rely significantly on inputs from the rest of the economy nor does it provide significant input to other processing industries. 4.27 If the agricultural sector is to play a leading role in promoting growth and reducing poverty, then policy measures must be found to enhance its productivity and its ties with the rest of the economy. In this context, it is to be noted that the current agricultural sector strategy is encoded in the so-called Letter of Agricultural Development Policy (LPDA).48 Among the stated objectives are to (1) promote food security; (2) promote agricultural exports; (3) provide essential services to farmers; and (4) ensure a rational management of natural resources. With respect to the achievement offood security, the government is bent on using tariffs to curtail the import of rice. As stated earlier, this seems to be the wrong approach to the problem, as the welfare cost to the poor seems to be higher than any other potential benefit. Lettre de Politique de Developpement Agricole, 1991, which is currently under revision. 66 Low POVERTY RESPONSIVENESS TO ECONOMIC GROWTH 4.28 The two fundamental determinants of income poverty are the size and the distribution of national income. Holding inequality constant, the potential of a country for poverty reduction depends on the responsiveness of poverty to economic growth. Such responsiveness may be measured by the growth elasticities of various poverty measures.49 Tables 4.1 and 4.2 present poverty elasticities for the headcount index based on both Cornell's and the government's upper poverty lines. These elasticities are quite low, indicating thus low potential for the Guinean economy to alleviate poverty through growth. Indeed the low level of these elasticities explains the result presented in Chapter two where it was observed for the sample period 1986-1996 that the economy was unable to reach a growth rate that would keep the number of poor constant. 4.29 These elasticities are also low by international standards. Jayarajah et al. (1996) have computed such growth elasticities for a sample of 46 countries. They found that these elasticities tend to be higher (between two and four) in East and South Asia, compared to Sub-Saharan Africa. For the latter, the growth elasticity was consistently less than one. The results from Guinea presented in the tables below do not seem to deviate significantly from these observations. This situation may be linked to the structural characteristics of the economy. As the results on inter-sectoral linkages demonstrate, the Guinea economy has a dual nature, where most of the poor operate either in the subsistence agricultural sector (in the rural areas) or in an informal sector (in the urban areas) with weak links to a very small industrial sector. Under such circumstances, one would not expect growth in the industrial sector to have a significant impact on poverty in the subsistence or the informal sector. 4.30 To promote high poverty responsiveness to growth, it is important to increase the productivity in the subsistence or the informal sector. This may require investment in human as well as in physical capital and restructuring of incentives. A smooth operation of the labor market is important in this regard. Increased productivity in this sector should lead eventually to both increased inter-connectedness of the productive sector and factor mobility within the economy in a response to changes in the structure of incentives. In any event, the necessary changes in the structure of incentives must be guided by the regional as well as the socio-economic distribution ofpoverty within the country. See Jayarajah et al. (1996:70-73). Such indicators measure the percentage change in observed poverty associated with a given change in mean income. In the particular case of the poverty incidence, it can be shown that the poverty elasticity is the negative of the elasticity of the cumulative function at the poverty line. Thus the size of the corresponding elasticity is determined by the size of the headcount and the steepness of the cumulative distribution at the poverty line. As it turns out, the higher the headcount the lower the growth elasticity the steeper the distribution at the poverty line the higher the elasticity. 67 Table 4.2 Guinea: Evolution of Growth Elasticities of Poverty (Cornell Poverty Line) Year Incidence Intensity Severity 1986 -0.76 -1.29 -1.67 1987 -0.68 -1.22 -1.59 1988 -0.75 -1.29 -1.66 1989 -0.60 -1.14 -1.51 1990 -0.67 -1.21 -1.58 1991 -0.69 -1.23 -1.60 1992 -0.76 -1.30 -1.68 1993 -0.81 -1.35 -1.73 1994 -1.01 -1.58 -1.99 1995 -0.88 -1.42 -1.82 1996 -0.86 -1.41 -1.80 Table 4.3 Guinea: Evolution of Growth Elasticities of Poverty (GOG Poverty Line) Year Incidence Intensity Severity 1986 -1.10 -1.67 -2.11 1987 -1.01 -1.57 -1.99 1988 -1.09 -1.66 -2.09 1989 -0.93 -1.48 -1.89 1990 -1.00 -1.56 -1.98 1991 -1.02 -1.59 -2.01 1992 -1.10 -1.68 -2.11 1993 -1.16 -1.74 -2.19 1994 -1.23 -1.83 -2.29 1995 -1.24 -1.84 -2.30 1996 -1.22 -1.82 -2.27 MLANAGEMENT CAPACITY 4.31 As stated earlier, the economic and financial reform program (PREF) launched in 1985 was seen as the only way out of the legacy of the centrally planned economic development in Guinea. The content of the program was very sound and indeed led to initial positive results. However, the limited nature of these results may be linked to the severe weakness of management capacity available in the country. Of course, this limited capacity itself is a reflection of the extremely low human capacity resulting from the policy failures. 4.32 Given the low level of human capital within the society at large, it is no surprise that the skill level within the civil service is quite low and unable to meet the requirements of the management of a market economy. Furthermore, a Civil Service Reform Management Review performed in 1994 (in collaboration with the World Bank) contains evidence of a civil service characterized by an incoherent structure and an irrational mode of operations. The incoherent structure stems principally from the 68 vagueness of the Civil Service Statutes. This poor specification of the rules of the game creates ample opportunity for those in charge of enforcement to circumvent the code at no penalty. The report cites instances where productivity allowances have been granted universally regardless of individual performance, even though the statutes require that this be linked to an annual staff performance evaluation. 4.33 As for the erratic behavior of the system, it is to be noted that career advancements, according to the statutes, must meet strict educational criteria and conform to specified hiring (competency exams) and promotional (job productivity and risk) procedures. In particular, it is required that both the Minister of Civil Service and the Minister of Finance jointly appoint staff within the civil service who meet these criteria. Yet, there is evidence that many of these rules are not followed. For instance, it appears that most of the employees hired during 1993 were never subjected to competitive examinations. Furthermore, the system is characterized by a lack of staffing charts and job descriptions. This situation can hardly promote a meaningful system of personnel management. 4.34 Given the scope and complexity of the PREF and the weak management capacity, the World Bank, in collaboration with other major Donors, has attempted to address this issue through a series of technical assistance projects for economic management. Two technical assistance projects were approved in 1985 and 1988 respectively along with two structural adjustment credits. The two projects aimed at facilitating the implementation of the government's reform program in the short-run and at fostering long-term institutional development and capacity building in the domain of economic management. The audit of both projects rated the overall outcome as marginally satisfactory and considered the institutional impact as modest, judging as uncertain the sustainability of the effort. The audit cited the principal reasons as: (1) over-ambitious objectives; (2) complex design; (3) lack of Donor coordination; and (4) the inability to implement the strengthening of local capacity components (due to technical assistant resources fully taken by the day-to-day demands of the reform program's implementation). 4.35 Based on the experience of this project, it was recommended that an integrated approach be adopted for the formulation of an institutional development strategy in Guinea Such an approach should deal with systemic issues such as the incentive structure in the civil service. It should also be based on effective donor coordination and mechanisms for the transfer of skills. In this particular case, it is recommended that emphasis be placed on in-service training of local staff and that formal training be more specifically targeted. Furthermore, such a strategy should be based on a merit system, efficient resource allocation and the fostering of a culture that promotes and rewards professionalism, integrity and self-improvement. INFORMAL SECTOR 4.36 The real challenge to the government in Guinea is, of course, to provide employment for the majority of people, in particular the poor, who have very limited 69 skills to operate in the formal wage labor market. In this respect, the informal sector potentially holds a key to the solution of this problem, particularly in this context of public sector retrenchment imposed by the necessities of policy reforms. In Guinea, it is estimated that at least two-thirds of the economically active population are employed in the informal sector.50 4.37 Mills and Sahn present evidence that, on average, poor households derive a higher share of their revenue from independent micro-enterprises, located in the informal sector than from other activities. This shows that informal sector development has direct implications for poverty alleviation in Guinea. Most micro-enterprises in Conakry (about 73%) involve commercial activities. Most are headed by women, generally with less education than their male counterparts, who tend to reside in households with the lowest per capita consumption. 4.38 The next most important class of enterprises in this sector is that of small industry. These tend to be more capital-intensive and headed by men. These observations suggest that women may be facing some discrimination in the credit market. Such discrimination, combined maybe with the low level of education, makes it hardfor them to start small business in the non-commercial sector. Table 4.4 Guinea: Sectoral Distribution of GDP Sector Informal Sector Formal Sectors Total Primary 46 4 29 Secondary 12 70 34 Tertiary 42 26 37 Total 100 100 100 Source: Walker (1995:70) 4.39 There are both individual and environmental constraints to informal sector development in Guinea. People operating in the informal sector, regardless of gender, are limited by their lower level of education as compared to the general population, particularly those operating in the formal sector. At the environmental level, Walker reports that such development is constrained by poor infrastructure in transportation and communication. This, in turn, restricts the expansion of enterprises that may prove locally profitable. She also notes the limited access to formal financial services, and therefore these enterprises cannot borrow for medium to long-term projects. 4.40 Given the important role that the informal sector plays in providing livelihoods for the poorer segments of the population, it is crucial that any poverty alleviation strategy address these constraints to the sector's development. In particular, consideration must be given to the establishment of a proper regulatory framework that encourages the development of this sector. Furthermore, with respect to purely financial 50 Walker (1995). 70 constraints ways must be found to provide access to adequate financial services, for informal sector operators on a sustainable basis. The design and the implementation of such a policy must be especially targeted at women, ensuring that they not only increase the productivity of their current enterprises but also to empower them to expand into the traditional activities dominated by men. 4.41 The fact that most micro-enterprises in the informal sector tend to be headed by women who also form the poorest expenditure quintile would suggest that programs designed to promote informal sector activities should be targeted at female-headed enterprises. 71 5. HUMAN CAPITAL DEVELOPMENT 5.1 It is now commonly accepted that for economic growth to be sustainable and equitable, it must involve the poor in a significant way. That is, they must have access to the necessary productive assets such as land, water, credit and markets, productive employment opportunities, and above all they must have access to essential social services that enhance their human capital such as health and education.43 5.2 The issue therefore arises as to what extent current social policies (existing institutions, their mode of operation and the funding of such service) do enhance human capital formation by the population at large and the poor in particular? To what extent do individual willingness and ability (private costs of existing public services) determine the use of available services? 5.3 For Guinea's citizens in general and the poor in particular, the social sector in Guinea can be a key contributor to Guineans' overall living standard. As one of the poorest nations in the world, however, living standards are reflected in extremely low health and education indicators, and the pervasive shortcomings of available social services. The requirement of maintaining a sound macroeconomic framework in the face of fluctuating world prices for Guinea's main export (bauxite) and a weak non-mining revenue base imposes the need for a careful, efficient and effective allocation of available funds for public investment to both social and economic infrastructure. In recent years, expenditures on social sectors has declined to a figure that is low even by regional standards, which are already lower than the average for developing countries. PAST DEVELOPMENTS IN SOCIAL SECTORS 5.4 The new government has committed itself to overhauling the health system and gave priority in the health sector to developing curative and preventive services at the primary level, with a particular focus on previously neglected rural areas. Despite the government's stated goals, better allocation of existing resources has been slow to materialize because of weak capacity in the government to implement policy directives and an apparent lack of commitment toward improving and protecting basic health services. The system remains weakest at the prefectural and sub-prefectural level, with 43 This chapter is based on work performed by the Health and Education Units of AF5PH, and contributions from AF5CO, within the context of the ongoing Public Expenditure Review in Guinea. drugs being scarce in many rural areas and facilities being insufficiently staffed and equipped. While efforts have been initiated to reverse this trend, the failure of the system at peripheral levels has led to inefficient use of higher level facilities, with hospital resources being used to treat many simple cases. 5.5 Over the past several years, the Ministry of Health (MoH) has responded by developing a set of action programs and operational targets that give priority to strengthening primary level services and seeking to redistribute the benefits of public spending for health in favor of the most disadvantaged and vulnerable social groups (women and children in rural areas). The implementation of each program is under the direct control of specific MoH units, and a central coordinating, monitoring, and evaluation unit has been created to guide the programs' implementation. Nevertheless, resources are highly strained, and the percentage of the government's recurrent budget allocated to the health sector is too low to allow for even basic services at the current population level, even if resources were efficiently allocated and the budget was fully implemented. 5.6 In the education sector as well, the government launched in 1984 a program of reform and rehabilitation to improve the operation of the educational system and to make it more responsive to the needs of the population. In the educational sector, the reform included placing educational administration under a single ministry, reorganizing regional educational management, rationalizing higher education, extending secondary education one year, and providing in-service training for teachers. Although some progress was made, the reform program did not address many of the sector's fundamental constraints, and resource allocation was slow to follow the declared policies. 5.7 In response to the disappointing outcome of the previous reforms and in recognition that the education sector was under-funded, the Guinean government prepared a program to address the major issues in the sector as an intrinsic part of the country's macroeconomic adjustment program. This new reform program, which was adopted at the end of 1989, attempted to give highest priority to improving basic education services, particularly in rural areas, while acknowledging the constraints imposed by the government's plan of overall fiscal austerity. Specifically, priority was to be assigned to key cost-effective pedagogical inputs and primary school construction with community participation in rural areas. Expansion of higher education was to be delayed and no major investments were to be made in the sector for research facilities or associated institutes, although qualitative and efficiency improvements would be sought. At the management and operations level, priority was to go to training, technical assistance and essential operating funds to promote efficiency gains, implement revised norms and standards, and to build sector management capacity. The present state of the sectors in this context will be examined in the following sections. 5.8 While extensive reform has taken place during the Second Republic, more efficient allocation and use of public resources is needed within key social sectors. In addition, questions remain about the affordability of current sectoral strategies and 73 inefficient allocations may crowd out funding for priority expenditures. In Guinea, expenditures on the social sectors declined by 25 percent between 1991 and 1994, although their share in total expenditures remained constant during this time (near 19 percent). These figures are low even by regional standards: expenditures on the social sectors in Sub-Saharan Africa averaged 26 percent between 1985 and 1989, and the average in all developing countries is 38 percent (1990). Further examination by sector reveals that while Guinea's spending on education has now reached SSAs average of the previous five years, health expenditures still remain almost 30 percent less than the SSA average between 1985 and 1990. 5.9 Relative to the high incidence of preventable diseases in Guinea, insufficient resources are allocated to primary and preventative health programs, and much of the expenditure disproportionately benefits wealthier citizens. In fact, public health care is relatively more expensive for the poor than private health care is for the wealthy. Furthermore, the health care that is available is generally concentrated in Conakry and in a few other geographical areas, limiting access to and the quality of services in many other, particularly poorer and rural, districts. Integrating the disparate elements of Guinea's health care system while ensuring that the poor increase their access to primary and preventive health care in the context of continued fiscal austerity is the difficult task that the Guinean government now faces in the health sector. 5.10 In the education sector, the government's reforms that were meant to deal with the legacy of a poorly functioning educational system have had some positive results: an increase in gross enrollment rates in primary education, notably among girls; improved budgetary management; and decentralization at several government ministry levels. As with other sectors, however, the government still faces a number of serious problems. Although enrollment rates have greatly increased, they are still very low, as are completion rates. Since the poor are less likely to enroll in schools, they receive a smaller proportion of public expenditures on education; poor children (especially those from rural areas and girls in particular) have far lower enrollment rates than wealthy, and the situation is not likely to improve as long as public expenditures on education continue to decline (1991 to 1993). At higher levels of education, wealthy students benefit even more, creating inequalities in per capita recurrent costs by educational level. Finally, Guinea's higher education system has not seen any comparable reform and remains grossly inefficient, although it continues to absorb more than one third of public education expenses. With private sector education among the least developed in the region, the government must ensure continued reform and renewed commitment if its current strategy is to have any hope of long term success. 74 HEALTH SECTOR Current Institutional Setting 5.11 The Ministry of Health (MoH) is responsible for the formulation and delivery of national population, health, and nutrition policy in Guinea, including the overall strategic planning, management, administration, and coordination of the country's health system. The health care system is based upon 33 prefectures in the four major geographical regions which contain 376 health centers (centres de sante or CS) and 347 village health posts (postes de sante or PS). In addition, there are 33 prefectural hospitals and four regional national hospitals, two of which are in Conakry. Management and coordination of health activities in the regions are the responsibility of Regional Health Inspectors, and a Prefectural Director of Health is responsible for overseeing technical activities, staff performance, drug distribution, and contacts with local administrative authorities in each prefecture. The activities of the CS are primary care, infant health, normal maternal labor (in rural areas), family planning, simple diagnostics, sale of essential drugs, and information. The activities of the PS are promotion of health education, growth surveillance of infants and pregnant women, primary care, sale of essential drugs, and referral to the CS, since the health system is based on patient referral from lower levels to more specialized services at higher levels. More than a dozen maternal and child care centers, located mostly in urban areas, are also part of the public health care system. Pharmaguinee, a former public enterprise now under the MoH, is responsible for the procurement and distribution of drugs to public health facilities, but the supply and distribution of essential drugs is highly dependent on external donors and NGOs who heavily supplement the provision of health services by financing infrastructure investments, basic operating costs, drugs, and some salaries. 5.12 Formally, entry into the public health system is at the sub-prefectural health center level (CS or PS), but problems that cannot be dealt with at the health center level are in principle referred to the prefectural hospital. At the highest levels, the regional hospitals and two university hospitals provide tertiary care. Many health centers have now adopted cost-recovery schemes, whereby fees are collected for the different health services provided, in particular the prescription and direct provision of drugs to patients. However, these fees do not fully cover the excessive costs for personnel and other specific services; hence, most health centers remain dependent on centralized financial support. 5.13 In addition to the public health system, the government has made continuous efforts to promote private sector participation, including non-profit organizations, in the health sector. While the number of private practitioners and clinics remains small, more than 100 private pharmacies have been established. Of course, several thousand traditional healers continue to provide health services and several traditional birth attendants are available for deliveries in most villages. 75 5.14 Available resources are often concentrated at the hospital level, encouraging those people with access to hospitals to use that level for routine services. Health centers and posts are often isolated geographically, resulting in a weak system of logistical support and supervision. Although the distribution of physical facilities throughout the country is reasonably equitable, rural facilities are often disadvantaged in terms of staffing, equipment, and operating funds. Government Sector Strategy and Spending Pattern 5.15 The government's overall objective for health care system is to improve the quality and expand the coverage of health services, particularly those oriented towards women and children. In order to do this, the government is to concentrate resources on expanding and improving primary health services, particularly in rural areas, and on developing programs to combat the most prevalent diseases and reach the most vulnerable groups. As budget resources are obviously constrained in this regard, greater efficiency in the planning, management, and organization of the health care system is the key to its success. The government's strategy is to make more effective use of existing resources by: a) strengthening the MoH's organizational structure with a view towards consolidation and decentralization of responsibilities; b) strengthening the managerial capacity of the MoH to (i) allocate the sector's scarce resources, both human and financial, along the lines of established policy and priorities, (ii) implement the sector's action programs, (iii) collect, analyze and utilize epidemiological, management and financial data; (iv) prepare and monitor future programs in the sector; and (v) decentralize decision making in the sector; c) improving the quality and frequency of in-service training programs and routine supervision and inspection at all levels in order to improve the quality of the MoH's staff resources; d) ensuring the rational procurement and efficient distribution of low-cost essential drugs to all levels; and e) encouraging the provision of family planning, nutrition, and public health education to women and communities. 5.16 Continued mobilization of resources for the health sector is to be aided by: a) increasing the relative level of the government's budgetary contribution to the health sector; b) introducing cost recovery measures at all levels of the system to mobilize additional resources for sector operations; c) promoting private initiatives in health care including physicians, clinics, pharmacies and imports through wholesale pharmaceutical companies; and d) obtaining support from donors to expand the health delivery system. The government has developed a set of action programs to operationalize the above policy measures, and while the programs are far reaching and ambitious, progress in initiating and sustaining some measures has been slow and will require stronger government commitment, particularly for politically sensitive areas such as personnel redeployment, allocation of recurrent expenditures, and cost recovery. 5.17 Although public health expenditures as a share of total expenditures have been fairly constant, they have been declining at an alarming rate in real terms over the past several years. During the 1988-1994 period, they fell from US$ 7.1 to US$ 1.8 per 76 capita. The decline in public financing in real terms and the imbalances due to an insufficient and declining current non-salary budget limit the effectiveness and efficiency of the sector. In response, the government in 1995 increased the current budget for health from 3.9 percent of the total budget to 5.4 percent. But despite the increase, the current budget for health remains inadequate because of the large proportion of the funds used for salaries (94 percent in 1994) and incomplete budget implementation. 5.18 Local governments also allocate part of their budgets to the health sector to finance salaries of certain personnel and to defray other operating costs. These expenditures are usually financed through local tax revenues and by some direct contributions from the central government. Although local government, such as those at the prefecture level, in principle may allocate up to 30 percent of their tax revenues to health and education, the total sum is quite small in comparison to the expenditures budgeted by the central government. 5.19 During the 199 1-1994 period, expenditures for primary care, secondary care, tertiary care and administration represented an average of 24 percent, 22 percent, 22 percent and 18 percent of the total expenditure on health, respectively. This allocation is inconsistent with the government's official sector policy stated above, and there have been no recent increases in the proportion of funds allocated to primary health care. However, as noted earlier, many of the non-primary level health establishments (such as hospitals) often diagnose and treat primary ailments. Of greater concern is the inadequacy of current non-wage expenditures for primary health care facilities. While the bulk of these expenditures go toward purchasing drugs, the amount remains too small; in 1994, public expenditure for medicine was increased to 64 percent of current non-wage expenditure, but this totals less than US$ 0.25 per capita, far less than the regional expenditure of US$ 2 per capita. In addition, the share remaining for other non-wage expenditures, such as infrastructure and vehicle maintenance, is grossly insufficient, creating difficulties in supervision, patient referral, and vehicle operation. Finally, government policy to redeploy health personnel to the primary care positions is not visible in expenditure figures. In 1991, salaries of personnel in primary health care represented 29 percent of total expenditure for wages, and secondary, tertiary, and administration personnel comprised 36 percent, 16.5 percent, and 18 percent of the total respectively. These figures have generally remained the same over the past several years, casting doubt on the government's capability of implementing its stated policy. Use and Poverty Effects of Health Services 5.20 In Guinea, the health care system presently benefits the wealthy far more than the poor, even though access to health care centers has improved considerably over the past five years. This is due to many factors, including geographic, cultural, and monetary considerations. In 1995, about 56 percent of the population outside Conakry had access to CS or PS level health facilities within 5 km and 81 percent had such access within 10 km. The latter figure varies by region between 74 and 85 percent. Utilization rates of health facilities also vary across the regions: for primary care, 31 percent of the 77 population outside Conakry utilized CS or PS (37 percent for inhabitants within 5 km and 22 percent for inhabitants within 10 km); for prenatal consultations, a comparable estimate is 68 percent overall. However, obstetrical care utilization is only 15 percent overall. In comparison, the utilization rates in Conakry for primary care, prenatal consultation, and obstetrical care are 23, 63, and 23 percent respectively, reflecting the availability of hospital care. Referral rates from the CS level to regional hospitals are quite low, estimated at about 1 percent. Figure 5.1 Guinea (1994): Health Facility Use for Illness by Quintile 0.6-/r 0.5- , U . I _ _ _ _ _ _ Q No Treatment C 0.4-O N Private m 0.3-o 18 _ _0 Public Hospital m 0.2-o 0 Public Health Center 0.1 .7I0iIL 0 - Poorest 2 3 4 Richest Quintile Source: 1994 Public Expenditure Review based on IHS data 5.21 As Figure 4.1 indicates wealthier groups access the health care system far more than the poor. Based on the 1994 IHS, of individuals who reported an illness in the previous month, 60 percent of the poorest quintile did not seek any treatment, compared with about 30 percent for the wealthiest quintile. The poor that did seek treatment overwhelmingly relied on private traditional medicine; only about 12 percent used the public health system. In contrast, more than 50 percent of the wealthiest quintile used public hospitals, which are far more expensive than other health care facilities. The result is that wealthier Guineans receive much larger per capita public expenditures on health than poorer Guineans; the per capita figure is GNF 4,062 for the wealthiest quintile and only GNF 346 for the poorest quintile (see Figure 4.2). This gap is primarily attributable to the unequal distribution of hospital expenditures, as the wealthiest quintiles absorb almost all of the hospital budget. Given that only the wealthier groups use hospital services, the government's policies and actions are at cross 78 purposes when the MoH spends twice as much of its current budget on hospital care, rather than on primary care. Figure 5.2 Guinea (1994): Per Capita Public Expenditures on Health C!, z 3000 . . . as 25000- | * Health Centers ioooJ, 1000 '* Hospitals 000 a. 0 0 D of Quintiles Source: 1994 Public Expenditure Review based on IHS data 5.22 At the primary health care level, the differences in health care facility utilization are not as dramatic. The poorest quintile receive GNF 265 per capita, while the wealthiest receive GNF 975 per capita. Expenditures are also higher in urban areas and especially the capital, Conakry, where they are twice as high as in other regions (see Figure 4.3). 7hus it is clear that the government's policy to redeploy resources to primary care levels has yet to be effectively implemented. Across the board, the expenditure figures favor the wealthy; in hospital expenditures, the poorest quintile receives about 2 percent of public expenditures while the wealthiest quintile receives about 55 percent. This distribution is among the most unequal in Africa, and is generalizable to all public health expenditures: the poor receive a smaller share of public health expenditures than their share of the population. Even when considering public expenditures on primary care health centers, the poorest quintile receives 10 percent while the wealthiest receives 36 percent. 79 Figure 5.3 Guinea (1994): Per Capita Public Expenditures on Health, by Region 2500- Z 2000m C, _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ U~~~~~~ Health 1501500 Centers U Hospitals * 1000- Cu 500 -ii 0 I G) 0 C0 0 O .,,,0 _ Region Source: 1994 Public Expenditure Review based on IHS data 5.23 The government's health care system, on average, contributes 55 percent of total per capita expenditures on public health, and households contribute the remaining 45 percent themselves. Public health expenditures are more important for the poorer groups, contributing a larger share of the total spent on public health; total spending on health increases dramatically by expenditure quintile, and per capita spending on public health is 13 times larger on the wealthiest quintile than for the poorest. Relative to household expenditures, the burden of health care spending on the poor is greater than on the non-poor. In fact, public health care is more expensive for the poor than private health care is for the wealthy. Also, the poor spend almost twice as much of their nonfood expenditures on health as the wealthy do. Finally, poorer groups spend more on private health care (primarily traditional health services) than public health care. 5.24 One can speculate as to why the poor choose not to use public health services any more than they do, and the answer is, of course, due to more than just cost recovery fees. While they may pay more for fees in the private sector, the consultation fee for a visit to a public sector facility is added to travel time, perceived level of benefit, and opportunity costs for that time. All in all, public sector services are much more expensive for the poor than private sector services. For example, a visit to a public health center takes an average of 9 hours, whereas a visit to a traditional caregiver takes less than 1 hour. Thus, 80 it is likely that costs associated with travel and waiting contribute to making public health care too expensive for the poor. Undoubtedly, the social determinants of health service utilization are also complex and deserve extensive study to determine their contribution. EDUCATION AND TRAINING SECTOR Current Institutional Environment and Outcome 5.25 The education sector in Guinea is similar to other francophone countries in Africa; it consists of six years of primary level, seven years of secondary level split into a four- year cycle for lower secondary and vocational training and a three-year upper secondary cycle, and five years of tertiary level education. In addition, certain vocational and technical education programs of various length are available at the secondary and tertiary levels. In the 1993-1994 school year, the system accommodated 598,290 students, with 79 percent in primary school, 18 percent in secondary school, 1.4 percent in technical schools and 1.4 percent in universities. In addition, the education system hosted 1,049 participants in a teacher training program. Responsibility for the educational sector is divided along the lines of pre-university and higher education. All pre-university structures are directed by the Ministry of Pre-University Education and Professional Training (MEPUFP) through eight regional academic inspectorates, in turn administering 38 prefect-level directorates of education and 210 sub-prefect level pedagogic delegates. Higher education in Guinea is managed jointly by the Ministry of Higher Education, Scientific Research, and Culture (MESRC), three sectoral ministries, and some of the institutions of higher learning themselves. 5.26 Over the last decade, the Guinean government, in recognition of the legacy of the poorly designed and ineffective education sector of the First Republic, has implemented a series of reforms that sought to increase enrollment in primary education, improve access to education for the most disadvantaged groups, and increase public spending on education, particularly at the primary level. The reforms have in many ways been successful: gross enrollment rates have increased notably; budget management has been improved; and the MEPUFP has decentralized and increased the efficiency of many of its functions. Nevertheless, the educational system still requires much improvement, and current goals require significant reallocation of resources if they are to be accomplished. 5.27 Guinea's pre-university educational system has improved considerably since the first phase of the latest reform program (Programme d'ajustement sectoriel de l'ducation, PASE, 1990-1994) was begun. The program has increased the gross enrollment rate (GER) in primary schools from one of the world's lowest, 29 percent in 1989, to 40 percent in 1994. Currently, the government appears to be ahead of schedule on increasing gross enrollment to 53 percent by the year 2000. The success of recent programs can be attributed to the government's development of a ten-year strategic plan covering the whole education sector and incorporating realistic assumptions about the country's macroeconomic context. The elements of the program included: an ambitious school construction program supported by community groups, NGOs and donors, and 81 accompanied by implementation of low-cost construction norms and standards; an extensive redeployment program shifting over 2,200 teachers from administrative posts and secondary school positions to primary school classrooms; and reallocating budget expenditures toward the education sector, particularly primary education. The record so far demonstrates the importance of a long-term framework for education system support, and the need to integrate program elements into a coherent sectoral strategy. 5.28 Vocational training in Guinea is the domain of seventeen lower secondary Vocational Education and Training (VET) institutions located in the various regions, along with a small number of post-secondary programs. The programs offered in VET schools have remained the same for many years; the schools continue to operate separate from and without regard for Guinea's labor market needs. Although the government is committed to moving to a demand-driven approach toward vocational training, it must still develop the sector's management capacity and establish basic quality assurance, maintenance, and upkeep functions. Higher education institutions continue to absorb a large share of the sectoral budget while producing relatively low numbers of graduates who are of limited employability outside the public sector. The large policy shifts of the past two decades have served as a negative impact on performance, and indicators continue to deteriorate. Women, particularly, have lost their share of access, with numbers declining from 19 percent attending in 1984 to 6 percent presently. Higher education institutions are overcrowded and under-funded, with a large build-up of internal arrears for food, equipment, and supply purchases. Monthly stipends to students alone represent 48.7 percent of the institutions' budgets and 21.3 percent of the entire MESRC budget. 5.29 Notwithstanding the relative improvement throughout the pre-university sector, only 12.8 percent of first-year enrollees went on to complete the primary level, and thus only 3.7 percent of the age cohort actually completed primary school; of this small group, only 5.7 percent were from rural areas and only 1.1 percent were girls. Large disparities in enrollment by region (see Figure 4.4), gender, and urban/rural location are still prevalent throughout Guinea. By region, Conakry has by far the highest enrollment rates, reflecting the fact that it is the main urban center, as well as the recipient of the largest amount of funding. Along gender lines, while primary enrollment rates for girls have increased from 19 percent to 23 percent between 1989 and 1993, the level of enrollment for all levels has remained at 32 percent for the past ten years. And the percentage of female teachers (less than 23 percent) is even lower than the percentage of female students. Finally, the differences between urban and rural enrollment is also great. The GER among rural students during 1993-94 was 27 percent for rural students and 78 percent for urban students. 82 Figure 5.4 Guinea (1992-1993): Gross Primary Enrollment Rates by Region 90- 80- 70- 60- a Boys| 0. 40 | | | | | Girls | 20I- - 0 0 Jj COTE I KILOMETERS / Yo ) r,ii c ~~~~~~~D'IVOIRE f Uc) LiBERIA \ i l> The bound-ries>, co-, denoninai- anrd ny ather oforn,atioo L-I B E R I A t - - C s~~~~~~~UEIA ~ > h-r -n tIns -ep da nat ir,ply, o- the port of The Wo-ld a-k G-op, L I,. RI0 r any judgment on the legal status of any terrtory, or any endorsement - -- N oor aceptonce of such boundaries 108 _______________~~~~~~~~~'' C IMAGING Report No.: 16465 GUI Type: SR