Report No. 21506-MOR Kingdom of Morocco Poverty Update (In Two Volumes) Volume 1: Main Report March 30, 2001 Middle East and North Africa Human Development Sector (MNSHD) u Document of the MrId Bank CURRENCY EQUIVALENTS Unit of Currency = Moroccan Dirhams (DM) Period Average Exchange Rates (DM per US dollar) 1996 1997 1998 1999 2000 2001* 8.72 9.52 9.60 9.80 10.63 10.73 * Rate for 2001 is average January 1- April 20 FISCAL YEAR January 1 - December 31 ACRONYMS AND ABBREVIATIONS AMB Associations Musulmanes de Bienfaisance (Muslim Charity Associations) BAJ Barnamaj Aoulaouiyat Ijtimaiya (Social Priorities Program) BTP Bdtiments Travaux Publics CET Centres d'Education et du Travail (Education and Work Centers) CC Chantiers Collectiviti CFP Centres de Formation Professionnelle (Vocational Training Centers) CIMR Caisse Interprofessionnelle Marocaine de Retraite (Moroccan Interprofessional Pension Fund) CMR Caisse Marocaine de Retraite (Moroccan Pension Fund) CNOPS Caisse Nationale des Organismes de Prgvoyance Sociale (Social Security Provisions Office) CNSS Caisse Nationale de Sicuritg Sociale (Social Security Office) COSEF Royal Commission on Education Reform CPI Consumer Price Index CSE Centres Socio-iducatifs (Socio-educational Centers) EN Entraide Nationale (Welfare Program) FNBT Farine de Bld Tendre (Low grade flour) GDP Gross Domestic Product LFPR Labor Force Participation Rates LMI Low Middle Income LSMS Living Standards Measurement Survey MEN Ministry of National Education MENA Middle East and North Africa MEST Ministry of Higher Education (complete name) MLD Mean Logarithmic Deviation MOH Ministry of Health NGO Non Governmental Organization OCP Office Ch9rifien des Phosphates (Sherif Phosphates Office) ODEP Office d'Exploitation des Ports (Port Operations Office) OFPPT Office de la Formation Professionnelle et de la Promotion du Travail (Vocational Training Office) ONCF Office National des Chemins de Fer (National Railways Office) ONE Office National de l'Electricitg (National Electric Office) PAGER Rural Portable Water Project WFP World Food Program (Programme Alimentaire Mondial - PAM). PN Promotion Nationale (Public Works) PPP Purchasing Power Parity RCAR Rigime Collectifd'Allocations de Retraite (Retirement Benefits Collective Plan) REER Real Effective Exchange Rate SMAG Agricultural Minimum Wage SMIG Industrial Minimum Wage ULC Unit Labor Cost UNDP United Nations Development Program Vice President: Jean-Louis Sarbib Country Director: Christian Delvoie Sector Director: Jacques Baudouy Task Team Leader: Setareh Razmara ACKNOWLEDGEMENTS This study has been prepared by a team of several people and is based on the findings of a mission in February 2000. It was written by Setareh Razmara (Task team leader and principal author) and Giovanni Vecchi (Consultant), under the supervision of Zafiris Tzannatos (SP sector manager). Inputs were provided by Daniel Dulitzky (pension system), Karim El Aynaoui (agriculture policy and macro background), Guillermo Hakim and Sophal Ear (labor markets) and Furio Rosati (child labor and transfers). Dominique van de Walle and Martin Ravallion were the principal advisors for the poverty analysis. Bahjat Achikbache (MNSHD) supervised the preparation and the carrying out of the 1998/99 LSMS survey and facilitated the collaboration with the Statistical Office. The poverty lines update and the calculation of poverty incidence were prepared by the Moroccan team (Mr. Douidich, Living Conditions Observatory) and technicians from the statistical Office (Mr. Douidich and Mr. Bennani) visited headquarters in October 1999 to discuss a preliminary poverty profile with the Bank (Olivier Dupriez and Stefano Paternostro (AFTM3). Valuable comments and suggestions were received from Willem van Eeghen, Paolo Zacchia, Linda Likar, Maryse Pierre-Louis, Nicole Klingen, Regina Bendokat, Eluned Roberts-Schweitzer, Mark Thomas, Jeffrey Waite, Sonia Hamman, Pedro Alba, Douglas Lister. Peer reviewers were Peter Lanjouw (DECRG) and Tamar Manuelyan Atinc (EASPR). Ms. Emma Etori was responsible for formatting the paper. Special thanks to Moroccan authorities for their support and active collaboration. Particularly Mr. Cherkaoui (Director of the Statistical Office) supported the preparation of the poverty analysis and Mr. Abzad (in charge of Household Surveys) and Mr. Douidich (Living Conditions Observatory) and their team kindly provided assistance and information from the LSMS data. KINGDOM OF MOROCCO POVERTY UPDATE TABLE OF CONTENTS EXECUTIVE SUMMARY AND POLICY RECOMMENDATIONS.....................................................................i CHAPTER I: POVERTY IN MOROCCO: A COMPREHENSIVE OVERVIEW..............................................1 A. Recent developments on poverty reduction and social policies ...................................................................... B. Poverty and its many dimensions ....................................................................................................................2 C. Opportunity, security and empowerment in M orocco..........................................................................o-........ 3 CHAPTER II: WHAT HAPPENED TO POVERTY DURING 1990-98 AND WHO ARE THE POOR?..........5 A. Changes in poverty from 1990 to 1998 ........................................................................................................... 5 B . Ineq u ality ......................................................................................................................................................... 7 C. W ho are the poor in 1998? .............................................................................................................................. 8 D. Effect of transfers on poverty.......................................................................................................................13 CHAPTER III: WHAT ARE THE MAIN FACTORS THAT EXPLAIN THE OBSERVED INCREASE IN POVERTY? ................................................................................................................................................................15 A. M acroeconomic Evolution in 1990's ............................................................................................................ 15 B. Contribution of Slow Economic Growth To Poverty Increase......................................................................17 C. Rural sector is in an impasse......................................................................................................................... 18 D. Export Growth, Employment and W ages.....................................................................................................19 E. Role of W orker Remittances......................................................................................................................... 22 F. Public Finance in the Social Sectors ............................................................................................................. 23 G. Policy Implication ......................................................................................................................................... 25 CHAPTER IV: HOW ARE THE POOR SUPPORTED BY PUBLIC SPENDING IN SOCIAL SECTORS?.28 A. Access to social services during 1990-98......................................................................................................28 B. Education and Literacy.................................................................................................................................. 29 C. Health care system ......................................................................................................................................... 34 D. Social insurance system ................................................................................................................................ 39 E. Social assistance mechanisms ....................................................................................................................... 42 Consumer Food Subsidies............................................................................................................................. 42 Public W orks Program (PN)......................................................................................................................... 45 W elfare Program - Entraide Nationale (EN) .................................................................................................47 ANNEXES A. Poverty Measurement and Analysis B. Child Labor C. Effect of Transfers on Poverty D. Incidence of Public Expenditures in Education and Health E. Pension System F. Social Assistance Programs Statistical Annex EXECUTIVE SUMMARY AND POLICY RECOMMENDATIONS During the last decade, social indicators improved substantially in Morocco. This was due to an increase in public social expenditures and the Government's increasing focus on rural areas (e.g., rural electrification, rural roads, potable water and BAJ). Yet, both absolute poverty and economic vulnerability increased. Sluggish GDP growth driven mainly by the drop of agriculture value-added, a collapse in employment creation, and growing inequality in rural areas are the key forces behind this poverty increase. Public expenditures in social sectors, although increasing in recent years and shifting towards rural areas, were also insufficient for lifting people out of poverty and reducing inequalities. To reduce poverty in a sustained manner in the future, Morocco will need a combined policy for economic growth at the macro level and efficient spending on social policies at the sectoral level. It may also need to spend more, and in a more targeted way, on social sectors: despite the recent efforts, its level of public expenditures remains low in relation to comparator countries. This increase is possible. It can come from an efficient use of budgetary resources, especially from its expected privatization revenues, reallocation of public spending and a redirection of the savings obtained from the rationalization and selective abolition of consumer food subsidies to strengthen its safety nets. Based on a partial analysis of a Living Standards Measurement Survey (LSMS) conducted in 1998-99 (the survey data have not been completely processed), this report assesses the extent and trend of poverty over the last decade and reviews the role of key social policies (such as education, health, social insurance and social assistance). The aim is to provide directions for policies which can foster growth and improve the living standards of the poor. A. Trends in Poverty and Social Conditions The recent LSMS survey, undertaken by the Statistical Office, shows that the total number of poor increased from 13.1% of the population in 1990/91 to 19% in 1998/99 (or, from 3.4 to 5.3 million). The number of "economically vulnerable", i.e. those who are at or below 50% above the poverty line, increased from 35 to 44% (or, from 9 to 12 million). Poverty remains largely a rural phenomenon: almost one Moroccan out of four is poor in rural areas compared to one out of ten in urban areas. Although the rural population represents 46% of the total population, 66% of the poor live in rural areas. But poverty in urban areas is rising also. The share of urban poor increased from 27 to 34% of total poor during this period, almost four times faster than rural/ urban population shift. The increase in poverty did not coincide with a deterioration of the main social indicators. For example, adult illiteracy was reduced (from 55 to 48%), and the net primary enrollment rate has substantially improved from 58% in 1990 to 70% in 1998, particularly among rural girls (from 28 to 47%). Child labor was also reduced from 18 to 14%. The crude death rate also decreased (from 7.7 per 1000 population to 6.3) and life expectancy increased from 67 to 69 years. Access to basic infrastructure services improved both in urban and rural areas. Despite these improvements, however, Morocco's social indicators continue to lag behind comparable regionallincome level countries and significant disparities remain between urban and rural areas. The primary enrollment rate remains 20% below the average enrollment rate in low middle-income countries (LMI); infant mortality and maternal mortality are high; and access to potable water and electricity in rural areas remains one of the lowest in the region. This calls for steady and increasing efforts by the Government over the coming years, as the legacy of the past remains huge. ii B. Why Poverty Increased in the 1990s? Poverty increased mainly because of slow growth which led to a decline in per capita consumption of the households. The decline in the GDP growth rate during the 1990s was driven mainly by the drop of the agriculture growth which was not offset by the small growth of non-agriculture GDP. Overall, the economy has grown annually by only 1.9% during 1991-98, compared to 4.1% during 1986-91, and non-agriculture GDP growth dropped from 4.2 to 2.8% p.a., particularly due to the slow growth in manufacturing. Given a population growth rate of 1.9% p.a., virtually the per capita income stagnated and real private consumption per capita, based on the national accounts, decreased by 1.4% p.a., compared to an increase of 1 9% over 1986-91. The LSMS data also shows similar decline in per capita consumption during the 1990s (1.6% p.a.). Repeated droughts in the 1990s account partly for this poor performance, and have i -creased vulnerability among the poor. Slow economic growth accounted for roughly 84% of the increase in poverty at the national level in the 1990s. The rest was due to a more unequal income distribution. Although at the national level relative inequality hardly changed during the 1990s (Gini coefficient stagnated around 39%), this appears to mask different patterns of inequality across urban and rural areas. Based on other inequality measures which are more sensitive to changes occurring in the left tail of the expenditure distribution, the impact of sluggish growth and deterioration of equality on poverty was different according to regions: in urban areas, improvement in redistribution among the poor dampened the negative impact on poverty attributable to lack of growth: the lowest decile received 22% of total per capita expenditure in 1998/99 (compared to 20.9% in 1990/91); while in rural areas both lack of growth and worsening of inequality played equal roles in the increase of rural poverty: the lowest decile received 25.4% of total per capita expenditure in 1998/99 (compared to 26.6% in 1990/91). More specifically, the following changes are associated with the increase in poverty: * Decline in manufacturing exports. The appreciation of the real effective exchange rate and the increase of unit labor costs (3% compared to -7% in 1986-91) in manufacturing in the 1990s led to a reversal of the gains achieved during 1986-91. Export growth in manufacturing sector declined substantially (from 20 to 6% p.a.) and job opportunities in the manufacturing exportable 3ector - mainly clothing and textiles - declined significantly. This was a labor intensive sector employing mainly the low-skilled in urban areas. * Slowdown in employment creation led to an increase in urban unemployment. Increasing urbanization has been coupled with declining growth in labor demand (3% during 1991-98 compared to 4.6% in 1986-91) and an increase in the urban unemployment rate (from 15% in 1991 te 22% in 1999). The unemployment rate among the urban poor also increased from 30 to 32% in the 1990s, and the share of poor among the unemployed increased (from 21% in 1991 to 28% in 1998). Although poor households are larger in size, the number of unemployed members in poor households has been also increasing during the last decade (from 0.7 to 0.8). * Drought, agricultural policies and price changes have all contributed to an increase in rural poverty. Unfavorable climatic conditions reduced agricultural output by 1.2% p.a. in the 1990s.1 High fluctuation in agricultural GDP is mainly the result of Government policies during the 1980s, which has promoted the expansion of cereal production on rainfed, less productive lands, worsening the impact of droughts. Since most of the rural poor earn their living from small agriculture and off- farm activities, they have been negatively affected. Moreover, the decline in rural income was accentuated by the adverse price evolution of cereals and cattle products during the 1990s. The two critical survey years for the present analysis, that is. 1990191 and 1998/99, correspond to similar climatic conditions and they were both good agriculture years iii * Decline in international worker remittances. In the 1990s, worker remittances from abroad declined (0.9% p.a.) and dropped, according to national accounts, from about 11% of total private consumption in 1991 to 8.4 % in 1998. Although worker remittances benefit more the middle-income groups in urban areas, their decline during the 1990s may have contributed to some extent to the rise in poverty incidence. Based on 1998/99 LSMS data, on average they represented about 10% of total household expenditures per capita in urban areas and about 7% in rural areas. * Increased public expenditures in consumer food subsidies diverted public resources which could have been spent on the poor. As a result of higher international prices for food subsidies, especially for wheat in the mid-1990s, the budgetary cost of consumer food subsidies increased by over 20% (from 1.3% of GDP in 1990 to 1.6% in 1998). This compares to 0.2% of GDP spent on public works (Promotion Nationale) which creates about 40,000 person-year jobs benefiting the rural unemployed, including the poor. While subsidies are poorly targeted as only 25% reach the poor. * Public spending in social sectors increased during the 1990s but remains insufficient and inefficient to lift people out of poverty. Poverty and social indicators might have deteriorated further if it had not been for an increase in public social spending during the 1990s. Social spending rose from 35 to 43% of total public expenditures, and from 9.1 to 12.5% as percentage of GDP between 1991 and 1998.2 Yet this increased level pales in comparison to other countries in the region and at the same level of income, who spend on average close to 20% of GDP on social programs. Two notable changes are worth noting: the decline in child labor and an increase in school enrollment, particularly in rural areas and for girls: in 1998, 66% of the children aged 7 to 15 were attending school, compared to 58% in 1991; in parallel, the percentage of working children was reduced to 14% from 18%. The reduction in child labor came mainly from non-poor households although school attendance increased for both poor and non-poor households. Nevertheless, the returns on social programs are poor (in particular for education); most programs are poorly targeted (i.e., consumer food subsidies); a large share of all social programs (education, health, social insurance, etc.) is mainly urban biased and designed to benefit the higher-middle income and the rich; and there is no comprehensive safety net in place for protecting the poor (i.e., health insurance, safety net for poor elderly and disabled, family benefits). C. Who are the Poor, their Sources of Income and why are they Poor? Despite the increase in poverty, the profile of the poor remained relatively constant. Both in rural and urban areas the poor generally lack human capital and live in large households, with many children and few working members. Incidence of poverty among children (less than 15 years of age) is about 1.5 times higher than among the adults: about 44% of the poor are children under 15 years of age and about 25% of the total children are poor compared to 16% of the total adults. About 64% of the poor are illiterate (compared to 52% for the non poor), and most of the poor households are headed by men. Most of the poor live in the Center and Northwest regions, and the incidence of poverty in BAJ provinces (27%) is higher than in the non-BAJ provinces (16%). Urban poor household heads are most likely to be working in services and construction as private sector employees or self-employed. Their rural counterparts are mainly self-employed and work in agriculture. Unemployment among the urban poor (31.5%) is higher than the national level (25.7%) and has increased during the last decade. The incidence of poverty has increased only slightly among the elderly (from 15 to 17.1%) while it has dramatically increased for both the disabled (from 13.5 to 22.6%), and the pensioners (retraitis) (from 2.5 to 7%). Thus, poverty during the 1990s appear to have hit the disabled most severely 2 Social expenditures include: education, health care, basic infrastructure, social security, employment programs, vocational training, consumer food subsidies, Promotion Nationale, Entraide Nationale, rural development programs and alphabetization. iv and it appears that even the pension recipients among the non-employed are also facing the poverty risk and they are not relatively well hedged against economic vulnerability. Labor is the main income-generating asset of the poor, particularly in urban areas. But rural households derive their income from multiple sources within the rural economy (agricultural own-production, production for market, sale of labor in rural markets, rental of productive assets such as land or capital goods, sale of crafts and other small-scale manufactures) and from the urban economy as well. Once the information on various sources of income and assets would be available, the discrepancy between the labor income of the poor and the non-poor as well as their assets, in both rural and urban areas would need to be investigated. At national level, public and private transfers have helped to reduce th2 poverty incidence by 24% (19 instead of 24.7%) and the poverty gap by 50% (4.4 instead of 10.8%). Private transfers have a larger impact on the incidence of poverty than public transfers (which are mainly formal pensions), particularly in rural areas, while public transfers are more effective in improving the condition of the very poor particularly in the urban areas. Because of the dynamics of urbanization in Morocco, the structural causes of urban and rural poverty are closely related. Poor rural households engaged in productive activities typically have access to agriculture land but their landholdings are small, rarely irrigated, and often exhibit low productivity, especially in rainfed areas; and due to lack of land titles and registration they cannot obtain formal credit and invest in their property. On the other hand the urban poor are also facing multiple deprivations such ais lack of employment, inadequacy of access to land, housing and basic services. D. Policy Directions for Poverty Reduction Promoting Sustained Economic Growth The key to reduce poverty will be to revitalize economic growth, as the elasticity of poverty to growth in Morocco is high (2.7 during 1991-98). The impact of growth on poverty reduction is expec ted to be higher in urban areas than in rural ones where a decline in inequality can also lead to a significant poverty reduction. If the macro policies of the 1990s continue, the GDP will grow at about 3% per annum (corresponding to about 1% annual per capita private consumption growth over 2000-2005) and poverty incidence would reach about 4.5 million in 2005 (about 15% of the population). But with a more aggressive and comprehensive growth oriented strategy, assuming that growth will be about 6% per annum, (this corresponds to about 2.5% annual per capita private consumption growth over 2000-2005) and poverty would decline to about 3.6 million in 2005 (about 12% of the population). Since most of the poor live in rural areas, pro-rural growth and particularly pro-agricultural sector growth would have a higher impact on poverty reduction. However, given the high variability of agriculture output to climate conditions, particularly the cereal production, the prospects for the agricultural sector are uncertain. Sustainable growth is more likely to originate from the development of the industrial sector especially in labor intensive sub-sectors where there are specific comparative advantages such as manufacturing, tourism and fisheries. To this effect, an export-oriented approach needs to be adopted through acceleration of trade liberalization and a more flexible exchange rate policy. In addition to tight fiscal and active monetary policies, an export-oriented growth strategy would also require simplifying procedures for setting up new small and medium enterprises and increasing access to land and credits for the SME. Improving Efficiency and Ouality of Public Expenditures in Social Sectors Restoration of economic growth is a pre-requisite for reducing poverty but will not be sufficient. Policies that reduce inequality should also provide additional ammunition for poverty and vulnerability reduction 3 See "Morocco: Sources of Growth", World Bank, April 2000. V in Morocco. Such policies can be distinguished between those which contribute to longer term development (such as investments in education, health, infrastructure, better urban planning, action to facilitate empowerment of the poor); those which enable better employment outcomes (especially among the less skilled and poorer workers) and provide sustainable, yet efficient labor insurance (such as pensions), and those which directly affect the poor (such as social assistance). Since late 1990s, there has been a new emphasis on social and rural programs to reduce regional gaps and improve social conditions of the poor. Among these recent measures which have been introduced or are under development are (i) acceleration of rural infrastructure programs (electricity, potable water and roads) with the aim of reaching 60-70% of the rural population by 2004; (ii) consensus building on education reform; (iii) creation of a Social Fund which will support community-based development; (iv) restructuring of the Entraide Nationale; (v) Government endorsement of the Gender Action Plan prepared by NGOs and women representatives; (vi) partnerships between Government and civil society for education and literacy programs; (vii) drought emergency programs targeted to poor rural areas; (viii) participatory rural development programs; and (ix) extension of medical insurance to the poor. In addition, since the mid 1990s, NGOs are increasingly becoming active in social programs; they are providing new avenues for public/private partnerships and new approaches for poverty reduction. Inevitably, these programs will help the poor/vulnerable, but these efforts will need to be substantially increased to have a rapid impact on poverty reduction. The available data and analysis in this report do not address all these issues. The following selective list of policy directions is driven by the need for (i) improving the efficiency and quality of established social development mechanisms; (ii) introducing various social safety net measures for preventing the vulnerable from falling into poverty and helping those who are temporarily or permanently unable to take advantage of income earning opportunities; and (iii) increasing budgetary resources towards social sectors. The increase in budgetary resources for social sectors could be obtained through reallocating budgetary spending, redirecting savings obtained from the consumer food subsidies reform (about 0.8% of GDP), additional general revenues notably from privatization proceeds, and redirecting budgetary wage expenses. It is roughly estimated that the Government could obtain a saving of about 2% of GDP that could be allocated to social sectors. Education and Literacy. Literacy and quality education are integral parts of a multi-dimensional concept of poverty. Based on the 1998/99 LSMS findings, actions to build up human assets of the poor need to focus on: * Universal Lower Basic Education. To increase lower basic schooling and reduce gender gaps, factors that are affecting low school attendance of the poor, such as indirect costs and access, need to be addressed. Although reaching under-represented groups may be relatively costly, it would require (i) reallocating the education budget towards basic schooling (with special priority on serving rural girls); (ii) continuing building school facilities in remote rural areas and designing the school facilities to attract rural girls; (iii) providing extra services and additional incentives to attract rural children, particularly girls (i.e. school meals, free books, cash transfers for attending school, etc.); (iv) expanding BAJ experiences in other poor rural provinces; (v) fostering school autonomy in order to give accountability and ownership at local level; and (vi) reinforcing NGOs and non-formal education programs at the lower-basic level while evaluating and monitoring their outcomes. * Increase Access and Improve the Ouality of Upper Basic and Secondary Education. As shown by the 1998/99 LSMS the main reasons for high dropouts at 12 years of age, is lack of access to upper basic education (mainly in rural areas) and lack of incentive and value-added in staying in school. Improving the quality and increasing access to upper basic as well as secondary education would help to keep children out of labor markets, provide them opportunities for continuing schooling, and increase their skills as future labor force. The cost of increasing access to upper basic and secondary vi education may be substantial, but it could be financed through introduction of selective cost recovery mechanisms at the secondary and higher education levels. * Adult illiteracy reduction. Given the high correlation between poverty and illiteracy, and the positive links between maternal literacy and health conditions, the Government goal is to reduce adult illiteracy to around 20% over a period of 14 years. This means the illiteracy program will need to reach about 500,000 people a year, compared to the current level of about 100,000. Thus tt e current functional literacy programs undertaken by the Ministry of Social Development, which contract out provision of services to other public/private agencies and local NGOs, after careful review of their effectiveness, need to be expanded and adequately funded. Based on preliminary estimates, to achieve its goal, the adult literacy budget will need to increase by threefold (from DH 30 million to .bout DH 90 million per year or about 0.4% of education budget). Health Care. Despite important achievements in health outcomes over the last decade, policies still favor the hospital subsector in the major urban areas, thereby de-emphasizing, de facto, priority to primary health services, particularly in rural areas. The 1998/99 LSMS data indicate that (i) the urban-rural disparity in access to services is still far from being reduced; (ii) the rural population and urban poor are deprived of adequate provision of health services; and (iii) the low level of utilization of public services indicates the need to increase both access and quality of public services. To achieve these goals, based on the 1998/99 LSMS and other sector work on health, the following priority areas would ne--d to be addressed/strengthened: * Reducing infant and maternal mortality rates. This would require strengthening maternal and child health programs, especially in rural areas and increasing responsibilities of local authori ties. * Improving equity and expanding health coverage. This would require redirection of he Ath care resources towards primary health care, more emphasis on cost-effective interventions, and strengthening partnerships with the private sector, non-governmental organizations and local communities and authorities. This would improve quality and access to health services and contribute to reducing rural-urban and inter-regional inequities. * Ensuring long-term sector financial sustainability. To improve financial sustainability through health financing reforms that will enhance health sector revenues, and financial access to cost- effective health services, to move towards universal health insurance coverage and to implement hospital reforms to improve the efficiency of the system and promote cost recovery. The first step in this process could be to introduce a compulsory health insurance scheme for the formal sector and expand access to health care services by putting in place a national health assistance fund for indigents. * Increasing financial resources for the health sector. To improve quality and increase access to health services, there is a need for an increase of the efficiency of public and private spending in order to get more value for the monies spent. In addition, the Government needs to consider increasing resources for the health sector, as public health spending is below the level of other comparable income countries. * Improving overall sector management. This could be achieved through the implementation of effective decentralization policies and the strengthening of core functions at the central level. Social Insurance System. The system coverage is limited to public and formal private sector and targets mainly the middle-income urban wage earners. Coverage is low mainly due to evasion among salaried workers as well as the large proportion of self-employed who are not affiliated. Overall the poor do not benefit from family allowances and pension benefits: in 1998/99 only about 5% of poor elderly received a formal pension. However, as shown by the 1998/99 LSMS, public transfers, which are mainly composed of formal pensions, have a non-negligible impact on poverty and income inequality reduction: Assuming there are no household behavioral responses and no substitution effect between public and private vii transfers, eliminating public transfers, mainly formal pension benefits, would increase the incidence of poverty from 19 to 21%, and particularly the poor would become poorer because the mean income gap would almost double, from 4.4 to 8.7% and the distribution among the poor would worsen. Moreover with an aging population, addressing the financial sustainability of the formal pension system is relevant from a poverty perspective because it could impose additional burden on the budget as well as jeopardizing the old age protection for the formal urban wage earners, including for the poorest among the poor. From a poverty perspective the following areas could be reformed: * Introducinp safety net schemes for poor households with children. Except for family allowances, there is no other cash transfer scheme in Morocco for providing income support for families to raise their children. As part of the safety net program, options for selective cash transfers to the poorest groups could be considered. Such a transfer can be scholarships to poor children, financed through general revenues. The level of benefits, targeted groups and cost would need to be evaluated. * Introducina social safety nets for elderly poor. Since there are no alternative safety nets for the old age in Morocco, introduction of targeted old-age cash transfer programs which would provide a basic pension as a safety net to every poor individual who reaches 60 years of age could be considered. This scheme could be means tested and financed by the general revenues. Although this type of old age cash transfer system can be expensive, in places with a tight family structure, such as Morocco, it will have the advantage of providing a safety net not only to the elderly poor but to their families. * Addressing Financial sustainability of the pension system. Creating links between benefits and contributions would improve the financial sustainability of the scheme while reducing evasion and distortions on the functioning of the labor market. The Government should explore the feasibility of implementing parametric or structural reforms. Among the parametric reforms, it is recommended to review benefit formula and eligibility rules, in particular the years of covered work requirement. Among the structural reforms, the government could consider the creation of a second pillar to channel part of the retirement savings. Based on the results of the ongoing actuarial projections a financially sustainable pension system could be designed. Social Assistance Programs. Overall social assistance programs are marginally targeted to the poor, they face efficiency problems and are limited in scope and effectiveness, and their coverage is modest compared to the existing number of poor and needy population. * Reforming food subsidies for better targeting of the funds to the poor. Eighty percent of the assistance programs in Morocco is devoted to universal consumer food subsidies: they are mainly transfers to processors and producers, and only 25% of funds reach the poor. Although, the poor benefit four times more than the rich in relative terms from food subsidies, Government resources could be more efficiently targeted to the poor through a parallel reduction of tariff protection and subsidies on food products and introduction of complementary targeted assistance to low-income groups. This would allow the reallocation of budgetary funds freed up by reduced subsidy expenditures, estimated at DH 2.7 billion (or 0.8% GDP) savings, to programs benefiting the poor (e.g., literacy, basic services in rural areas, public works programs and assistance programs targeted to the poor). Reduction of tariff protection would lead to domestic price decline which in turn would benefit consumers (domestic prices of vegetable oils, flour, would be reduced by about 10% and 14% respectively, and it would remain stable for sugar). In addition, in the medium term it would allow to strengthen agricultural policy by focusing on crops where Morocco has a comparative advantage in order to ensure adequate income for the rural population, including the poor. However, in the short- term about 2 million farmers and agriculture workers would be affected by the tariff reduction and would need compensatory measures (i.e., programs to help small land holders to convert to other and better suited alternatives; expansion of the existing labor intensive public works projects (PN); basic infrastructure development in affected rural areas; and community development programs in remote viii rural areas). To start dismantling the border protection system, since November 2000, the Government has begun liberalizing oil prices, and has reduced the import tariffs on oilseeds. As a result, the consumer prices on oil have dropped, and the savings are expected to be used for decentralized rural programs to compensate farmers. Nevertheless, these reforms need to be extended to the other two commodities (sugar and wheat) which are politically more sensitive. * Expanding existing Public Works (PN) to labor intensive activities. Currently publi c works, managed by PN, are largely targeted to rural areas and about 40% of the funds are reaching the poorest among the unemployed and those who are affected by transitional shocks, such as drought. Given the recent increase in poverty, the PN, without any institutional changes, could: (i) implement large scale employment programs in rural areas while ensuring that labor intensive methods are applied to implement projects targeted to the most disadvantaged regions; (ii) help other Government agencies in selecting public works' projects that are executed using labor intensive methods, and (iii) contract out certain urban projects to the private sector and NGOs while focusing its resources on labor intensive rural works located in the most disadvantaged regions. * Restructuring the in-kind Welfare Program (EN). To improve the efficiency of EN, since 1999, the Government has launched a two-year restructuring effort piloted in four provinces to introduce changes in service delivery. The objective of this effort was to test the impact of the changes and examine restructuring alternatives. Based on lessons learned from the restructuring in piloted provinces, an expansion of the restructuring of EN needs to be launched in order to (i) better identify target groups and their needs; (ii) better focus EN activities, improve cost-effectivenes3 of the programs and increase beneficiary coverage; and (iii) strengthen its institutional capacity to monitor the impact of its activities. E. Future Agenda The report addressed the trend of poverty over the last decade and the role of key social policies based on a partial analysis of the 1998/99 LSMS due to data limitations. But additional work would be re.,uired to address the following issues: * Evaluation of regional poverty and better understanding of the determinants of poverty in both rural and urban areas, through the econometric modeling/analysis of the risk of being poor. * Refinement and update of the analysis (poverty profile and social public expenditures), once the data from the 2001 Household Expenditure survey is available. * Assessment of the sensitivity of the poverty profile to economies of scale. * Evaluation of sources of income for the poor in both urban and rural areas (including agricuhure and off-farm activities, physical assets, migration, land, etc.) * Impact of rural development and ongoing agriculture policies on the poor. * Urban development and livelihoods of urban poor (access to land, housing, basic services, etc.). * Fine tuning the micro and macro analysis and clarifying the labor market issues. * Impact of devaluation and export oriented growth on the poor. * Links between employment expansion, including informal sector, and poverty. * Incidence analysis of consumer food subsidies based on the 1998/99 LSMS data and clarification on who are the net producers and consumers. * Construction of a poverty map to improve targeting of public policies. * Participatory poverty assessment to review how the poor people view poverty and well-being? What are their problems and priorities? What is their experience with the institutions of the state, markets and civil society? and How are gender relations faring within households and communities? Morocco: Strategy for Poverty Reduction Findings Objectives Policy recommendations Poverty Increase: * In 1998/98 about 5.3 million of Morocco's population were * Reduce poverty. * Reach roughly 6% annual GDP growth through promoting considered poor (19%), compared to 3.4 million in 1990/91 export-oriented growth strategy through: accelerating trade (13.1%); and the number of "economically vulnerable", i.e., those liberalization and realignment of the exchange rates; who are at or below 50% above the poverty line, increased from 35 promoting growth in industrial sector, especially in labor to 44% (or, from 9 to 12 million). intensive sub-sectors such as manufacturing, tourism and * 84% of the poverty increase is attributed to slow economic growth fisheries; tight fiscal and an active monetary policies; and the rest to the deterioration of income distribution, simplifying procedures for setting up new small and medium * During the 1990s inequality at the national level stagnated (Gini * Reduce inequality and enterprises; and increasing access of small businesses to land coefficient around 39%), but inequality in rural areas increased, disparity among regions. and credits. while it decreased in urban areas. * Poverty continues to be mostly a rural phenomenon but urban * Reallocate public expenditures towards programs benefiting poverty has also increased from 27 to 34%. disadvantages groups and remote areas. * Incidence of poverty is higher in BAJ provinces (27%) than the non- BAJ (16%) provinces and about 36% of the poor (about 2 million) live in BAJ provinces. * Prepare a poverty map which would allow identifying the poor * With an elasticity of poverty to growth of 2.7, poverty incidence provinces and expanding BAJ programs. will be reduced by 8% p.a. if the GDP growth reaches roughly 6% p.a. (corresponding to 2.5% per capita household expenditure growth); a pro-agricultural growth would have a higher impact on poverty reduction (-13% p.a.) while an industry biased growth would have the lowest impact on poverty (-2% p.a.). Public Expenditures in Social Sectors: * Public expenditures in social sectors increased from 9.1% of GDP in * Provide additional * Increase budgetary resources towards social sectors through 1991 to 12.5% in 1998 and social indicators improved but they ammunition for poverty reallocating public spending, redirecting savings obtained from continue to lag substantially behind those of comparator countries and reduction. the consumer food subsidies reforms and increasing general there are substantial dispanities mn access to social services between revenues. mteallurban areas. Education (costing 5.9% of GDP): * Enrollment has increased at all levels and it has helped to reduce * Reach universal lower * Reallocate education budget towards basic schooling; expand child labor over the last decade. basic schooling and BAJ experiences; combine school construction in remote rural * Main impediments to enrollment for the poor are: low income reduce gender gaps and areas with quality improvement measures and design the because they cannot afford to send their children to school; lack of regional disparities. school facilities to attract rural girls; provide extra services to access to school, particularly in rural areas; and cultural attitude, attract rural children, particularly girls (i.e. school meals, free particularly for rural girls. * Increase access and books, cash transfers for attending school, etc.); give * Poor, particularly in rural areas, allocate a larger share of their improve quality of upper accountability and ownership at local level; reinforce NGOs expenditures (1.5%) to education than the better off (0.5%). basic and secondary and non-formal education programs at the lower-basic level, * Lower-basic education is pro-poor and subsidies accruing to poorest education. to keep after monitoring their outcomes. (13.9% of HH expenditures) are higher than for the rich (0.6% of children out of labor * Introduce selective cost recovery at the secondary and higher X Findings Objectives Policy recommendations HH expenditures). market, provide them education levels. * Upper-basic education is uniformly benefiting all expenditure opportunities for * Expand functional literacy programs undertaken by the groups, continuing schooling, Government with other public/private agencies and local * Secondary and higher education benefit mainly the non-poor, and increase their skills. NGOs, after careful review of their effectiveness. * Reduce adult illiteracy to 20% by 2010. Health (costing 1.1% of GDP): * Infant and maternal mortality rates are above regional averages. * Reduce infant and * Introduce maternal and child health programs, especially in * Almost 80% of recurrent public budget on health is allocated to maternal mortality rates. rural areas and increase responsibilities of local authorities. public hospitals in urban areas mainly used by the better-off. * Improve equity and expand health coverage through * Less that 15% of the population has access to formal health redirecting health care resources to primary health care, insurance but almost 99% of them are non-poor. * Improve quality and emphasizing on cost effective interventions services, * Travel time and costs of seeking medical care are senous access access to health services increasing the efficiency of public and pnvate spending, and obstacles for all households in rural areas, including poor. and reduce rural-urban strengthening partnerships with private sector, non- * Households behavior confirms inefficiency and low quality of inter-regional inequities. governmental organizations and local communities. health services in the public sector: consultations with private doctors is predominant for all expenditure groups, including poor, * Introduce health financing and hospital reforms to enhance and exceed visits to public hospitals. health sector revenues, improve financial access to cost- * Out of pocket expenditures are high, particularly for the poor: urban effective health services and move towards universal health households spend almost twice on health care that the rural one; and insurance coverage, e.g. introduce compulsory health the poor spend 2.4% of there budget on health care, compared to insurance scheme for the formal sector and expanding access 4.3% for the better-off. to health care services for indigents. * Regardless of health facilities, the better-off use more often medical consultations. * Ensure long-term * Increase financial resources for the health sector and improve * The better-off consult more often the private sector while the poor financial sustainability. management of the system. resort to public health facilities (dispensaries, health centers, public hospitals). * Public expenditures on health (1.2% of GDP) are low compared to other comparable income countries (2.8%). Social Insurance System (costing 2.5% of GDP): * Only 2.3 million wage earners in the formal urban sector contribute to the social insurance system, corresponding to about 28% of the total labor force, and have pension coverage; and about 13% of the wage earners receive family allowances (family benefit). * The system is urban biased and does not provide any benefits for vulnerable groups who have no permanent source of income. * Family allowances, similar to wage supplements to families with * Since there is no other * Evaluate options for selective cash transfers to the poorest children, (i) increase the labor cost because they are financed by safety net mechanism, groups. Such transfers can be scholarships to poor children payroll taxation (8.9% wage bill); (ii) indirectly finance the pension provide income support and fininced thrnugh gcneral r-nu-c benefits; (iii) are inequitable because their coverage is limited to for low income families public and formal private sector. with children. * Although only about 4% of the elderly poor receive formal pension * Improve the safety net * Exanune means tested old age cash transfer program which benefits in the absence of these benefits (main component of public for the old age of the would provide a basic pension as a safety net to poor xi Findings Objectives Policy recommendations transfers to households) poverty would increase, the poor would poor. individuals who reach 60 years of age and evaluate its cost to become poorer and inequality among the poorest would worsen. be financed by the general revenues. * Minimum pensions are modest and well below minimum * Ensure long term * Based on the ongoing actuarial projections, explore feasibility subsistence level: for CNSS they correspond to 30% of SMIG and financial sustainability of of implementing parametric or structural reforms to increase for CMR to 40% of the SMIG. the pension system. coverage and sustainability of the pension system. Social Assistance Programs (costing 2% of GDP): Consumer food subsidies (1.6% of GDP): * Consumer food subsidies (cooking oil, sugar, low-grade flour) are * Increase impact of public * Reduce/eliminate tariff protection and subsidies on food universally available and poorly targeted. spending on the poor and products and use savings (about DH 2.7 billion) for targeted * Only 25% of the subsidies reach the poor, but they account for a reduce distortion in assistance to low-income consumers and farmers and programs higher proportion of poor's expenditures and represent about 40% of agriculture. benefiting the poor (e.g., literacy, basic services in rural areas, their calone intake. * Strengthen agricultural public works programs and assistance programs targeted to the * Substantial share of subsidies is a transfer to producers and policy to focus on crops poor). processors (50% for flour, 25% for sugar, and 21% for cookng oil) where Morocco has a e Since November 2000, GOM has begun liberalizing oil prices, and generates significant economic distortions in agriculture sector. comparative advantage. and has reduced the import tariffs on oilseeds but these reforms Public Works (PN) (0.2% of GDP): need to be extended to the other two commodities (sugar and * Provide employment in rural areas through productive investments * Expand large scale wheat). utilizing to some extent labor intensive methods: creates about 10.4 employment programs in million person day of employment per year at the average cost of rural areas and enforce * Reorient resources to labor intensive activities in rural areas. DH 43/day; and quality basic infrastructure construction in low- the self-targeting * Help other Government agencies in selecting labor intensive income commuities. mechanisms. methods for implementing their projects. * PN has the capacity of scaling up labor intensive rural activities, * To focus resources on labor intensive activities in particularly during droughts. disadvantaged regions, contract-out some projects to the * Recently resources have shifted towards urban areas and about 50% private sector/NGOS. of its budget is supporting local government and only 40% of jobs created are labor intensive. * Except in BA provinces, there is no explicit miomtoring of outcome. Welfare Programs (EN) (0.1% of GDP): * EN is the only public agency that has the mandate to help the poor * Better identify target and works in partnership with other agencies (incl. NGO). groups; improve cost- * Its services (literacy to poor mothers, basic training for illiterate effectiveness of the girls, shelters for school-aged children and orphans, pre-schooling) programs; increase * Evaluate results (in terms of service delivery and coverage) of are self-targeted because they do not attract the non-poor beneficiary coverage; EN in piloted provinces. * Programs cover only a tiny proportion of the poor (about 80,000 or and strengthen * Based on the lessons learned, consider restructuring expansion. 1.6% of the poor) and are inefficiently run. institutional capacity to * Administration is highly centralized, overstaffed and there is no monitor the impact of momtonng. activities. * Since 2/3 of the staff are temporary and from low-income groups it can be assumed that between 70-80 of the EN's budget reach the poor. I_I_I 1 CHAPTER I POVERTY IN MOROCCO: A COMPREHENSIVE OVERVIEW According to the 1998-99 Living Standards Measurement Survey conducted in Morocco, poverty showed a disturbing increase during the 1990s, regardless of how poverty is measured. This report updates the poverty profile for Morocco following the comprehensive approach suggested by the latest World Development Report. Therefore, it presents a detailed analysis of (i) what happened to poverty over the 1990s (Chapter II), (ii) which factors are key in explaining the observed increase in poverty (Chapter III), and (iii) the support given to the poor by the Government intervention (Chapter IV). An overview of the main initiatives taken by the authorities as well as by NGOs and private businesses in the late 1990s is presented in this chapter. Their effects, though not captured by the 1998/99 data, are likely to have a significant impact on poverty reduction in the coining years. A. RECENT DEVELOPMENTS ON POVERTY REDUCTION AND SOCIAL POLICIES 1.1 The main objective of this report is to provide a comprehensive poverty update for Morocco, on the basis of a Living Standards Measurement Survey conducted in 1998-99. The report recognizes that poverty is, by its very nature, a multifaceted phenomenon affected by the interaction of economic, political and social processes. Accordingly, the report adopts a comprehensive approach to poverty, and focuses on both income and non-income measures of poverty, along the lines suggested by the latest World Development Report.4 Because of the time span covered by the data, however, the results reported in the following chapters make no allowance for possible reduction in poverty attributable to the numerous initiatives taken by the Government since the late 1990s. In fact, social and human development policies have increasingly gained momentum since 1997, mainly because of the actions of the central Government, but also thanks to the initiatives taken by NGOs, the Palace, and private businesses. In all likelihood, these recent initiatives will have positive outcomes in terms of promoting opportunity, enhancing the security of vulnerable groups, as well as facilitating the empowerment of the poor. 1.2. Among the Government measures which have been introduced over the past four years or are under preparation are: (i) acceleration of rural infrastructure programs (electricity, potable water and roads) with the aim of reaching 60-70% of the rural population by 2004; (ii) a Royal Commission has reached a consensus on a Charter for Education that provides for the first time a strong basis for successful education reform; (iii) Law to establish an Agence de D6veloppement Social (ADS) with Tripartite Board including for the first time NGOs. By supporting community-based development, this Social Fund has a strong potential to efficiently reduce the risks faced by the poor; (iv) restructuring of the Entraide Nationale; (v) Government endorsement of the Gender Action Plan prepared by a group of NGOs and women's representatives; (vi) partnerships between Government and civil society for education and literacy programs; (vii) drought emergency programs targeted to poor rural areas; (viii) participatory rural development programs; (ix) extension of medical insurance to the poor; 4 See The World Bank (2001), World Development Report 2000/2001, Attacking Poverty, Oxford University Press. 2 (x) the King's establishment of Foundations/Funds to address poverty and promote social development - Mohamed V Foundation for Solidarity and Poverty Reduction, and F assan II Fund for Social and Economic Development; (xi) Judicial Reform and fight against corruption. 1.3 A detailed evaluation of most of these recent initiatives is being prepared in a separate World Bank document "Social Protection Strategy Note". This poverty update evaluates a selection of Government social programs (Chapter IV) which accounts for about 90% of total social expenditures. B. POVERTY AND ITS MANY DIMENSIONS 1.4 According to the LSMS data, the incidence of poverty in Morocco has increased from 13.1% in 1990/91 to 19% in 1998/99. The increasing trend in poverty emerges also if other standard measures of poverty are considered: during the 1990s poverty increased both in depth (the poor got poorer) and severity (the distribution of welfare among the poor became more unequal). However, poverty is a complex phenomenon, intrinsically multidimensional. Being aware of this leads to a more comprehensive approach, where income-based estimates of poverty are flanked by other indicators, ranging from 'social indicators' (e.g., progress in survival, access to basic facilities, education, health, etc.) to measures which take into account other spheres of deprivation such as opportunity, security, and empowerment. 1.5 However, the increase in expenditure-based poverty measures in the 1990s was coupled with a substantial improvement in most social indicators (see Chapter IV). Indeed, as a result of an increasing share of the budget devoted by the Government to social expenditures (from 8,4 to 12.5% as a percentage of GDP between 1990 and 1998), a large set of social indicators showed an unambiguously positive trend during the 1990s: adult illiteracy, primary enrollment rates, death rates, life expectancy, access to electricity and piped water, have all recorded significant improvements. Despite the progress reccrded in aggregate social indicators, two main issues remain, which should be of great concern to the Government. First, aggregate figures mask wide imbalances between urban and rural areas, as well as gender disparities.s Secondly, slow economic growth during the 1990s leads to a reappraisal of Morocco's performance when put in an international perspective: in fact, although the percentage discrepancy between some social indicators in Morocco and other comparable countries has decreased, it is still large and the pace of improvement should be accelerated further. 1.6 A more comprehensive assessment of Morocco's performance during the 1990s based on a set of composite indices confirms the results of this report (see Table 1).6 The main results can be summarized as follows. * The Human Development Index (HDI) in Morocco increased from 0.54 in 1990 to 0.58 in 1997; this level remains however the lowest among other countries in the region. * The Gender-related Development Index (GDI) confirms that Morocco continues to lag considerably behind all other comparable countries: the improvements in female literacy/enrollment rates documented in Chapter IV of this report are to be firmly corroborated in the coming years. s In Chapter IV, Table 14 documents the extent of such improvements. 6 The Human Development Index (HDI) measures the overall achievements in three dimensions of development: longevity, educational attainment and income; the Gender-related Development Index (GDI) measures the same basic capabilities as the HDI, but improves upon it by taking into account the inequality in achievement between women and men: the gr-ater the gender disparity, the lower the GDI compared to the HDI; the Gender Empowerment Measure (GEM) captures the extent to which women and men are able to actively participate in economic and political life and take part in decision-making; the Human Poverty Index (HPI) provides an aggregate measure of non-income poverty: the HPI takes into account longevity, knowledge and overall economic provisioning (as measured by the percentage of people with access to health services, safe water and the percentage of malnourished children under five). 3 * The Gender Empowerment Measure (GEM), which is a good measure of women's empowerment, shows that Morocco fares well in this respect. * The Human Poverty Index (HPI) shows that in 1997 Morocco had a very high HPI - about 55% higher than the average recorded in medium human development countries. This confirms the fact that poverty over the 1990s increased not only in its income dimension, but also - and possibly more disturbingly - in most other dimensions of deprivation. Table 1: Comparative composite indices (1997) GNP HDI GDI EM HPI Country (per capita, US$ Human Gender-related Gender Human 1997) Development Development Index Empowerment Poverty Index Measure Index Jordan 1520 0.715 - 0.220 9.8 Iran 1780 0.715 0.696 0.264 20.4 Tunisia 2110 0.695 0.681 0.353 23.1 Algeria 1500 0.664 0.642 0.245 28.8 Egypt 1200 0.616 0.603 0.275 33 Medium Human 1280 0.662 0.658 - 25.3 DevelopmentI Note: Techneal detais about the idces i columns 3-6 arc n the Human Development Report (1999) Source: UNDP Human Development Report (1999), Social Indicators of Development. C. OPPORTUNrY, SECURITY AND EMPOWERMENT IN MOROCCO 1.7 Although macroeconomic management has successfully ensured stability - a prerequisite to higher economic growth - most of the poverty increase recorded in Morocco during the 1990s was due to the poor performance of the economy. Over the years 1991-98 real GDP grew only by 1.9% per year, as a result of a pronounced drop in the agriculture value added (-1.9% per year) and the fact that non- agriculture GDP did not expand fast enough to absorb the effects of the declining growth of the traditional sector. Real per capita private consumption mirrored such an impasse and declined at 1.4% a year. Nor did the trend in inequality help: although in urban areas the increase in poverty was dampened by a decline in inequality, in rural areas - where the majority of the poor are concentrated - rising inequality significantly fostered the increase in poverty. 1.8 The strong causal links between poor economic growth and the worsening of both income and non-income dimensions of poverty made no exception for Morocco. The poor performance of the Moroccan economy during the 1990s greatly inhibited the promotion of material opportunities for the poor. This report emphasizes the importance of promoting opportunity and explicitly addresses the question of how to achieve pro-poor growth (chapter III). Moreover, an assessment of the extent to which the poor were helped by public intervention in building up their human and physical assets, is provided in Chapter IV. The analysis takes into account the many facets of opportunity, namely the process of job creation, the accumulation of social overhead capital (roads, electricity, water, and sanitation), the provision of financial services, as well as access to and fruition of schools and health services. 4 1.9 A further dimension of poverty considered by this report is the vulnerability of the poor to a wide array of shocks, ranging from country-level adverse shocks to micro-idiosyncratic shocks affecting individuals and/or households. The estimates reported in Chapter II show that more than 40% of the Moroccan population is clustered just above the poverty line, thereby bringing to the Government's attention the need to set up a comprehensive strategy to enhance security for the poor. This report attaches great importance to the reduction of economic vulnerability, and devotes most of chapter IV to the investigation of its many facets. The upcoming "Social Protection Strategy Note" complements the analysis of this report by (i) evaluating the existing social protection programs; (ii) acknowledging the recent efforts made by the Government in helping people manage risk; and (iii) reviewing the role of the NGOs in social sectors. 5 CHAPTER II WHAT HAPPENED TO POVERTY DURING 1990-98 AND WHO ARE THE POOR? While the 1980s were marked by a substantial decline in poverty, the 1990s saw a substantial increase in both the incidence of poverty and the economic vulnerability. In 1998/99 about 19% (5.3 million) of Morocco's population are considered poor compared to 13,1% (3.4 million) in 1991 and 21,1% (4.6 million) in 1985. Poverty continues to be mostly a rural phenomenon: almost one Moroccan out of four is poor in rural areas compared to one out of ten in urban areas. Both depth and severity of poverty have increased during the 1990s particularly in rural areas. Incidence of poverty has increased in all regions, with the notable exception of the oriental region. The poor continue to be characterized by large household size, high dependency ratio, and low education level. Poor households are more likely to be either wage earners or self-employed and most of the poor work in the agriculture, services and construction sectors. In urban areas, the increase in poverty seems to be correlated with unemployment increase. Among the inactive, most of the poor are to be found among households where heads are disabled and elderly. Both private and public transfers have substantial poverty and redistributional impact. A. CHANGES IN POVERTY FROM 1990 To 1998 2.1 People are classified as poor if their consumption expenditure falls below a poverty line. The poverty line is defined as the value of a basket of basic foodstuffs and non-food goods.7 For 1998/99 the poverty lines are DH 3.922 per person and per year for the urban level, DH 3.037 per person and per year for the rural level, and DH 3.337 per person and per year at the national level (compared respectively to DH 2.674, DH 2.384 and DH 2.495 in 1990/91).8 (see Table 2). Table 2: Poverty lines, incidence, depth and severity of poverty (dh/person/year) for 1990-91 and 1998-99 I9199998-99 Changes duigthe 909 Poverty lie Urban Rural National Urban Rural National Urban Rural National Poverty line (DH/perlyear) . 2674 1 2384 2495 3922 13037 3337 1 47% 27% 34% Incidence of poverty (%) 7.6 J18. 131 12.0 127.2 19.0 158% 51% 1 45% Number of poor (in 91 2.448 3 360 I.1 3.496 5 307 98% 43% 58% thousands .......... .. . . . . . .. ... .. Number of econ. vulnerable 2.312 6.640 8.952 5.034 7.122 12 156 118% 7,3% 35% (50%>PL) (thousands) I ::al$opulation 12w005_13_603 2 0 15 27 1 5_ _ _5% POeYCv Ga mdwex (%I 1S 3.8 2.7 2.5 6.7 4.4 67% 76% 63% Severin*inde.t() 0.4 2 0.8 7 8 2.5 200 208% I 2o0 i no is eLnrnued b. % pping th. i lng inpincnt t to rithe ioo p- erl line and b, re-esurnniing [he ro,d demand mi .kI cto calcul. the allowance for non-food goods Source : Statistical Office and World Bank estimates based on 1990/91 & 1998199 LSMS 7 The methodology used to measure the poverty lines for 1998/99 is based on the one used for the Poverty Report on Morocco of 1994. For the details see Annex I describing the methodology for constructing the poverty lines for 1998/99. It should be noted that throughout this report it has been assumed that prices do not vary systematically within Morocco. This is consistent with both the evidence provided by the 1994 report on poverty in Morocco (see Kingdom of Morocco: Poverty, Adjustment & Growth, Report no. 11918-MOR), and the view that internal market integration has been increasing considerably over the 1990s. In terms of 1993 PPP US$/capita/day these poverty lines correspond to: US$2.3 in rural, US$2.9 in urban and US$2 5 national. 6 2.2 The 1998/99 survey confirms the increase of poverty incidence during the last decade and the rural characteristics of poverty already noticed in the 1990/91 and 1984/85 surveys.9 In 1998/99 the poverty incidence at the national level is estimated to be 19,0% (about one out of five Moroccans lives under the poverty line). There are 5,3 million poor persons compared to 3,4 million in 1990/91 (13,1% of the population) and 5,7 million (21,1%) in 1984/85 (see Table 2). 2.3 Poverty in Morocco continues to be mostly a rural phenomenon: in 1998/99 more than oae fourth of the population living in rural areas are poor, as compared to one-tenth in urban areas. Although the rural population represents 46% of the total population, 66% of the poor live in rural areas (compared to 73% in 1990/91), a percentage which rises to 77% for the very poor and over 90% for the extremely poor. 2.4 According to the 1998/99 data, a large proportion of the population of Morocco can be defined as "economically vulnerable". During 1990-99, the number of "economically vulnerable", i.e., tLose with consumption per capita at or below 1.5 times the poverty line, has also increased: in 1998/99, the total number of poor would increase to 12 million (44% of the population), compared to 9 million 35%) in 1990/91. Therefore, in addition to the absolute poor, about 7 million people in 1998/99 (compared to 5.6 million in 1990/91) can be considered "economically vulnerable" to natural, economic and social shocks (i.e., droughts, job loss, illness, old age, etc.) and fall into poverty (Table 2). Figure 1: Poverty Incidence Curves for 1990-91 and 1998-99 507100 15 1007000 25 20%7 0 90%_ 60% 0 aSW 40%- E 20% 0% 50 75 100 125 150 175 200 225 250 275 300 Percentage of the poverty line rural 1998/99 -- urban 1998/99 0 rural 1990/91 - urban 1990/91 Source: Statistical Office and World Bank estimates based on 1990/91 & 1998/99 LSMS * According to estimates based on the LSMS, in 1990/91 mean private consumption per person was DH 6.780, which compares to DH 6.384 resulting from the National Accounts (NA) (see World Bank, "Kingdom of Morocco: Poverty, Adjustment and Growth"; Report No. 11918-MOR; January 1994). By adjusting the survey data to make them consistent with the NA estimates, the incidence of poverty for 1990/91 rises to 14.6%, thereby resulting 1.5% higher than the unadjusted estimates of poverty used throughout this report (13.1%). Although the main conclusions reached in this report about the trend of poverty over the 1990s are not affected by the above adjustment, the increase in poverty over the 1990s appears less severe than in the absence of the NA-adjustment (see Direction de la Statistique, Analyse du profil et de la dynamique de la pauvretd: un fondement de l'att6nuation des d6nuements - forthcoming). 7 2.5 According to the poverty incidence curves for urban and rural areas in 1990/91 and 1998/99, it does not really matter where the poverty line is drawn. For any given poverty line: (i) poverty has increased over the 1990-1998 years; (ii) poverty in Morocco is by far higher in rural than in urban areas; and (iii) urban poverty is increasing faster than rural poverty (Figure 1). 2.6 Both the depth and the severity of poverty have unambiguously increased over the 1990-98 years, particularly in rural areas (see Table 2). In 1998-99 the estimated cost of raising the consumption level of every poor person to the (upper) poverty line was 4.4% of the cost of giving every person, regardless of how poor, a transfer equal to the poverty line (compared to 2.7% in 1990/91).'o In other words, there would be sizable gains from a perfect targeting. As far as the severity of poverty is concerned, the estimates point to a worsening in the distribution of welfare among the poor: between 1990/91 and 1998/99 the index has almost doubled, both at the national level and within urban and rural areas, meaning that among the poor the distribution of income has become more unequal. B. INEQUALITY 2.7 Between 1990/91 and 1998/99, the same pattern of inequality is found in both urban and rural areas, although the former is slightly more unequal than the latter. In both years, the richest 50% of the population makes almost 75% of the total expenditure, while the top 10% accounts for more than 30% of the total expenditure on consumption. The data on the distribution of expenditure shares across population deciles are consistent with the estimates of the Gini coefficient: during the 1990s, the Gini coefficient stagnated at around 39%. Less than 20% of the aggregate inequality is due to differences between the mean expenditures between urban and rural areas: 80% of the observed inequality at the national level would disappear if - ceteris paribus - the disparities within each area would be eliminated. Table 3: Cumulative distribution of expenditures in urban and rural areas by decile (in %) Decile Urban Rural National Urban Rural National 1 2.6 3.6 2.7 2.9 3.3 2.6 2 6.6 8.5 6.6 7.0 8.1 6.5 3 11.7 14.5 11.4 12.1 14.0 11.3 4 18.2 21.2 17.0 18.0 20.7 17.1 5 25.1 28.9 23.8 24.9 28.5 23.9 6 33.8 37.5 32.3 33.0 37.5 31.9 7 43.3 47.7 41.7 42.6 47.9 41.4 8 55.1 59.7 53.8 54.5 60.3 53.2 9 70.8 74.9 69.2 70.2 75.7 68.8 10 100.0 100.0 100.0 100.0 100.0 100.0 Nominal Mean expenditure 9,224 4,623 6,780 10,154 5,085 7,823 (DH/person/year) 1991 Mean expenditure 9.224 4,623 6,780 7,543 3,942 5,890 (DH/person/year) 37.7 31.2 39.3 37.7 31.6 39.5 (1.04) (0.66) (0.72) (0.58) (0.57) (0.47) Notes: Deciles refer to households' per capita expenditure Bootstrap standard errors in parentheses Source: Statistical Office, 1990/91 & 1998/99 LSMS. Poverty gap provides also important information to policymakers because it indicates the potential gains by targeting poverty alleviation programs vis-a-vis sharing out transfers to every poor by an amount equal to the poverty line. 8 2.8 To what extent may the increase in poverty in Morocco during the 1990-98 period be attributed to increased inequality? Unambiguously, relative inequality hardly changed during the period 1990-98, and in fact, neither the levels of inequality at the national level, nor the inequality within each area -hanged (see Table 3)". These findings suggest that Morocco during the 1990s appears as a highly static society, where lack of mobility adds to unequal living conditions (as measured by the per capita expenc-iture on consumption). This flat distributive trend in the 1990s rules out that inequality can explain the observed increase in poverty, and it ultimately suggests that during the 1990-98, the trend in per capita expenditure (income) is most likely to be responsible for the increase in poverty (see Chapter III). C. WHO ARE THE POOR IN 1998? 2.9 Regional Poverty Profile.12 Morocco has become a predominantly urban society: in 1999 about 54% of the population lived in urban areas, compared to 49% in 1991. Although about two-thirds of the poor live in rural areas, there are marked disparities in poverty among the regions. Along with increasing urbanization the share of urban poor increased from 27% in 1990/91 to 34% in 1990/91. However increase in urban poverty is almost four times faster than rural/urban population shift. 2.10 Based on the new poverty lines for 1998/99, the incidence of poverty is highest in the ceni-er south (Meknds Tafilalet) (29%), center-north (particularly, in Fes-Boulmane, Taza-Al and Hoceima-T iounate) (28%), and northwest (Doukala-Abda, Marrakech-Tensift-Alhaouz) (24%): these regions represer t almost 40% of the total share of the poor population. The incidence of poverty is also higher in BAJ provinces (27%) than the non-BAJ (16%) ones: about 26% of the population and 36% (about 2 million) of the poor live in BAJ provinces." The rural areas most affected by poverty are the regions of Meknes-Tafilalet (36%); F6s-Boulemane, Taza-Al Hoceima-Taounate (31%) and Tanger-T6touan (31%). As for the urban poverty, it is relatively widespread in the regions of F6s-Boulemne, Taza-Al Hoceima-Taounate (24 %), Meknes-Tafilalet (23%) and the Orient (16%). i urban areas, the incidence of poverty is lowest in Grand- Casablanca (5%), Chaouia-Ouardigha (4%), and Rgions Sud (6%). The major contribution to a ggregate poverty comes from the region in Tensift-Al Haouz with 15% of the total poor living there. Elimination of rural poverty within the three poorest regions (Tensift, Taza-Hal Hoceima, and Souss-Massa-Daraa) would reduce the total number of the poor by 42%. 2.11 Overall between 1990/91 and 1998/99 the incidence of poverty increased in all regions, with the notable exception of the oriental region. Because of the change in the definition of regions in 1994, It is safer not to push the comparisons too far. In terms of the incidence of poverty, however, it is sufficiently clear from the data that some regions did better than others. As far as rural poverty is concerned, for example, the incidence of poverty increased dramatically in the center and northwest, whereas it did not change in the oriental region. As for urban poverty, the most dramatic increase has occurred in northwest again, whereas the Tensift region saw a relatively less severe increase. 2.12 Demographic and Socioeconomic Characteristics of the Poor. The 1998/99 LSMS data suggest almost the same profile as in 1990/91. First, in Morocco larger household size is associated with greater incidence of poverty in both urban and rural areas: at the national level, the incidence of poverty rises from 1.8% for households sized 1-2, to 30.9% for households sized 9 or more (poor households size is 7.7 in 1 See Annex A for the illustration of the results of the test carried out to check whether the observed differences i a the Gini indices are statistically significant. These results confirm that inequality did not change over the 1990s. 12 The 1998/99 household survey was not designed for a regional and provincial breakdown of poverty indices. Theiefore it is important that the results discussed in this paragraph be treated cautiously. 1 The World Bank has introduced in FY96 a social priorities program called BAJ in fourteen poorest provinces. It comprises three complementary projects designed to improve the quality and quantity of basic health and education, enhance rural infrastructure using labor intensive techniques, and improve the Government's capacity to monitor and analyze the socio- economic status of the population. 9 98/99 compared to the national average 5.9); Second, both the child/adult ratio and the dependency ratio are higher in poor households (the former equals 1.08 for the poor, compared to 0.62 for the better off. Third, households headed by young breadwinners (aged between 15 and 24) as well as by heads aged 60 years or more have a lower risk of poverty than the average. Fourth, incidence of poverty among children is about 1.5 times higher than among the adults: about 44% of the poor are children under 15 years of age and about 25% of the total children are poor compared to 16% of the total adults. Finally, the gender of the head of the household is not a significant determinant of poverty: 90.9% of poor households are headed by males and 9.1% by females, to be compared with 87.8 and 12.1%, respectively, in the population. Thus, there is no evidence supporting the "feminization of poverty" hypothesis, although it should be noted that this finding requires cautious interpretation. 2.13 Education and Poverty. Education has a strong correlation to poverty risk and usually the education of household heads plays a key role in determining whether a household is poor or not. In Morocco levels of education are inversely correlated with the incidence of poverty: at the national level the incidence of poverty increases from 3% for households headed by top-educated breadwinners (sup&ieur) to 23% for households headed by persons with no schooling. An estimated 64.3% of the poor have no education, and 17.5% have only primary education, compared to the proportions among non-poor households estimated, respectively, at 52 and 24%. Such pattern is similar both in rural and urban areas albeit stronger in rural ones. Comparing 1990/91 and 1998/99 we find a similar pattern in the relative risks of poverty by level of education: according to 1990/91 LSMS data, the incidence of poverty increases from 2% for households headed by breadwinners with a secondary diploma or higher to 16% for those headed by non-educated breadwinners. In 1990/91, approximately 72% of the poor have no education, and 12% have only primary education. Table 4: Child Work and Enrollment for Poor and Non-Poor (in %) 1998 1991 Urban Rural Urban Rural Poor Non- Poor Non- Poor Non- Poor Non- poor Poor poor Poor School only 70.0 87.2 37.9 51.9 70.7 84.1 34.3 43.2 Work only 6.4 4.3 22.9 22.8 8.0 4.2 22.4 26.5 Neither school nor work 23.4 8.4 37.9 23.9 21.3 11.6 43.2 29.9 Source: Statistical Office and World Bank estimates based on 1990/91 & 1998/99 LSMS 2.14 At the national level, 27% of all children in the 10-15 age-group are illiterate, while in the poorest regions that percentage may rise to 40% or more (for instance, 42.3% in Taza-Hal Hoceima, 37.8% in Tensift). These children deprived of even a basic education in childhood have very poor labor market prospects for the future. There is a very clear link between the poverty of a family and the poor achievements of its children. Most of the adolescents who leave school to seek employment before or just after completing the primary level, are from poor families. About 24% of all children between 10 and 15 did not complete even primary education and are no longer in school. 2.15 Although poverty has increased during the last decade, child labor has been declining mainly as a result of the increasing school attendance, particularly in rural areas and for girls. However, the reduction 1 Allowing for economies of scale would not alter the trend in poverty during the 1990s. Nevertheless further research is required to refine the poverty profile adjusted by economies of scale and thereby helping to improve targeting of pro-poor policies. For instance (i) since the approach used for calculating poverty lines does not take into account economy of scale for larger households, the result may be spurious, in the sense that households headed by females tend to be smaller in size than average; (ii) the result does not rule out that women may be trapped by poverty in other respects; and (iii) women may be under-represented as head of poor households because respondents in the Surveys often cite a male relative as head of household. 10 in share of working children is mainly concentrated in non-poor households although school at :endance increased for both poor and non-poor households (See Box 1). Box 1: Working Children Although poverty has been increasing during the last decade, child labor has been declining mainly as a result of the increasing school attendance. According to the LSMS , in 1998 about 64% of the children aged 7 to 15 were attending school, compared to about 58% in 1991. In parallel, in 1998 about 16% of the children were only working and about 18 % were reported as doing nothing, compared to 18 and 25% respectively in 1991. The increase in male school attendance (70% in 1998 compared to 68% in 1991) was associated mainly with a reduction in the number of working children (16% in 1998 compared to 18% in 1991), while for females, the enrollment increase (58i in 1998 compared to 48% in 1991) was coupled mainly by a reduction in the number of "idle" children (25% in 1998 compare J to 36% in 1991). The increase in total school enrollment during 1991-1998 has been largely due to the increase in enrollmer: of rural females (38% in 1998 compared to 24% in 1991), which was associated with a reduction of the number children performing household chores (36% in 1998 compared to 50% in 1991). Dropouts and child labor are highly correlated. Particularl ' in rural areas we observe that by the age of 15, less than 10% of the rural girls are in school (compared to about 24% for rural 1oys) and almost 50% perform household chores (compared to 8% for rural boys). Most of the children classified either as 'working only' or 'combining school with work' do perform their activities within the family farm or business (see Table 4 1. As a consequence, they are likely to work in an environment that is more friendly and less harmful than that experienced by children working for non family members. However, many of these children, especially females, do not attend school and this i! likely to strongly limit their possibility of development and social promotion. The reduction in share of working children is mainly concentrated in non-poor households although school attendance ncreased in both poor and non-poor households: in 1998, 69% of poor urban children and 39% of poor rural children were enrolled compared to 71% and 34% respectively in 1991. Based on LSMS data, the increase in school attendance was high. r for the "middle class" (third and fourth income quintiles) while the reduction in child labor was the smallest among the poor. Factors that are relevant for the child labor in Morocco are: (i) probability of working increases with the age of .ne child, particularly for boys in both urban and rural areas, while probability of attending school decreases. But for girls only the probability of doing households chores increases with age; (ii) level of income does not influence the decision of sending a child to work but has a significant influence on the decision to be "idle"; (iii) family structure influences the decision of the household differently depending on the region of residence: effect of the household structure on boys and girls refle :s, most likely, the gender specialization within the household; (iv) in rural areas, where most people are farmers, the effects of the household structure are more linked to the return to work in the agricultural sector; (v) in both rural and urban an as, large numbers of children in the household increase the probability of sending some of them to work and the number of siblings tends to reduce the probability of attending school for boys and girls; (vi) education level of the parents, particularly of t!e father, reduce the probability of sending children to work; and (vii) school availability does not influence household choices in urban areas but in rural communities the availability of schools appears particularly relevant: probability of both working and being idle decreases if a school is present and the more grades the school offers. Effective policies for reducing child labor in Morocco are: increasing access to schools in rural areas (i.e., if children in every village had access to a primar: school, enrollment would rise by 6% in rural areas); reducing fertility; promoting mother education for reducing gender gap; and reducing household chores particularly for rural girls (availability of water). (See Annex B). 2.16 Employment Characteristics of the Poor. The LSMS data show that both labor market status (the occupational status of the head of the household such as employment, unemployment or inactive status) and the sector of employment correlate with poverty. At the national level amongst the poor 80% are employed, 6% are unemployed, and 14% are inactive (i.e., housewife, elderly, handicapped, etc). This pattern does not change in urban and rural areas: in rural areas, 87.5% of the poor are active, 3.7% are unemployed, 8.8% are inactive; and in urban areas, 64.8% of the poor are active, 10.3% are unemployed, 24.9% are inactive. (Table 5). 2.17 At the national level, the number of income earners per household is Icw in Morocco: in the average household sized 5.93, there are only 2.04 monetarily gainful employed members. Reliance on a sole earner for each household means high exposure to risks of illness or loss of employment: there are approximately 1.7 million single-earner households (corresponding to almost 9 million individuals), while 0.§ million households do not have any member classified as actively employed. Slightly more than 1 million individuals live within households where all members are classified as inactive. The latter estimates point to the existence of a sizable proportion of the population that is likely to be chronically poor, i.e. likely to be trapped below the poverty line for long periods of time, possibly forever. 11 Table 5: Poverty profile by employment status of the household head for 1998/99 Employment status % Population % Poor population of household head Urban Rural National Urban Rural National Employed: -Wage earner 42.2 21.0 32.4 34.4 27.2 29.9 -Self-employed 21.6 60 2 39.4 29.3 59.7 49.9 -Employer 3.8 1.2 2.6 0.8 0.0 0.3 -Others 1.4 1.6 1.5 0.2 0.6 0.5 Not Employed: -Unemployed 4.4 2.9 3.7 10.3 3.7 5.9 -Disabled & Ill 4.1 3.2 3.7 7.9 2.5 4.3 -Pensioners 8.6 1.6 5.4 4.7 0.6 2.0 -Others 13.9 8.2 11.3 12.3 5.7 8.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source . Statistical Office and World Bank estimates based on the 1998/99 LSMS 2.18 According to both 1990/91 and 1998/99 surveys, urban poor household heads are most likely to be either wage earners in the private sector or self-employed, as opposed to the rural poor ones who are mainly self-employed: according to the 1998/99 LSMS, the urban poor are living either in households headed by wage-earners (53.2%) or by self-employed (45.3%), while in rural areas, most of the poor are headed by self-employed (68.2%) followed by wage earners (31.1%). Among the poor households headed by wage earners, more than 90% are in the private sector, irrespective of the area of residence. The distribution of the poor by sector of activity of the head of the household, at the national level, shows that most of the poor work in the agriculture sector (56.6%) followed by services (25.5%) and construction (BTP) (12.6%). More specifically in urban areas 58% of the poor have their breadwinners working in services (59.7% self- employed and 40.3% wage-earners), 21% in BTP (92.6% wage-earners and only 7.4% self-employed) and 12.9% in industry. While in rural areas 75.3% of the poor have their breadwinners in the agricultural sector (80.9% self-employed and 18.4% wage-earners), 13.0% in services (51.2% self-employed and 47.7% wage- earners) and 9.4% in BTP. 2.19 Unemployment and the Poor. Most of the poor in rural areas cannot afford to be unemployed, but in urban areas unemployment has increased for all expenditure groups. Based on LSMS data, to some extent, the increase in poverty is correlated with unemployment increase in urban areas. While unemployment is mainly an urban phenomenon (urban unemployment is 25.7% compared to 7.2% in rural areas) underemployment is likely to be more relevant for the rural areas.15 According to the 1998/99 LSMS, in urban areas only about 10.3% of the poor live in households where the head is unemployed and 28% of the unemployed are poor (compared to 12.9 and 20.6% in 1990/91). However, the unemployment among the urban poor (31.5%) is higher than in aggregate (25.7%) and it has increased during the last decade (compare to 30 and 20.6% in 1990/91, respectively). Taking into account unemployed members inside the household, the number of unemployed is also higher among the poor than the non-poor: in 1998/99, in urban areas the average number of unemployed per household equals 0.83 for the poor and 0.58 for the non poor (compared to 0.75 and 0.44 in 1990/91). The average number of unemployed per household also shows a similar pattern for both 1990/91 and 1998/99, i.e. it tends to drop as expenditure levels increase. In 1998/99, in the lowest quintile the number of household members who are unemployed is higher with respect to the highest quintile (0.09 compared to 0.07). Similarly, in 1990/91 the number of unemployed inside the household decreases from 0.07 in the lowest quintile to 0.05 in the highest. 15 In respect to underemployment, the 1998/99 survey does not provide sufficient information to address this issue. 12 Table 6: Summary characteristics of the poor for 1990/91 and 1998/99 1990/91 1998/99 NLAMER OF POOR - P--r 3 4 million people P0or 3 iji lbon pcunle Rural areas: 73% of all poor Rural areas: 66% of all poor Urban areas: 27% of all poor Urban areas: 34% of all poor SOCIO-DEMIOGRAPHIC CII.RACTERISTIC, Lare hiousOhold 17 5 indi% idial.1 iarge hous:hold,i - rdlJU1h High child/adult ratios (1.31 per household) High childladult ratios (1.091 per household) Lack of schoolina (73% of the poor are without schooling) Lack of schooling (64% of the poor are without scholing) OCCLTPAtIQN)QtSTATU Employed Employed Urban: 49% self-employed, 38% wage-earners. Urban: 53% wage-earners, 45% self-employed Rural: 58% self-employed, 30% wage-earners. Rural: 68% self-employed, 31% wage-carners. Unemployed Unemployed Urban: 13.0% Urban: 10.3% Rural: 6.4% Rural: 3.7% Inactive Inactive Urban: n.a Urban: 32% disabled, 19% pensioners, Rural: n.a. 18% elderly. Rural: 41% elderly, 28% disabled, 21% housevife. Sector of activity Sector of activity Urban: 65% services, 23% industry, 10% BTP Urban: 58% services, 21% BTP, 13% industry Rural: 83% agriculture, 7% services, 5% BTP Rural: 75% agriculture, 13% services, 9% BTP Note: n.a (not available) when the number of observations does not allow for reliable estimates Source: Statistical Office and World Bank estimates based on 1990/91 & 1998/99 LSMS. 2.20 Inactivity and the Poor. Amongst the inactive, poverty is most prevalent for those who are incapable of working and who do not possess any potential for employment: at the national leve-l, within the inactive group, most of the poor are to be found among households where the heads are disEbled and elderly (30.5% disabled, 27.5% elderly, 16.4% pensioners and "rentiers"). In urban areas 31.5% are disabled, 19% are pensioners, and 18.1% elderly; while in rural areas 41.2% are elderly and 23.9% are disabled. The incidence of poverty has increased only slightly among the elderly (from 15 to 17.1%) while it has dramatically increased for both the disabled (from 13.5 to 22.6%), and the pensioners (retraihs) (from 2.5 to 7 %). However, the most spectacular change concerns the mix of elderly and disabled among the poor; in 1990/91: out of 100 inactive-headed households below the poverty line, 47 were run by elderly and 18 by disabled breadwinners, to be compared with 27 and 30 respectively, in 1998/99. Thus, among the inactive, poverty during the 1990s appears to have hit disabled most severely. Although this inlormation should be treated with caution, it seems that even the pension recipients among the non-employe(J are also facing the poverty risk and there are not relatively well hedged against economic vulnerability. 2.21 Income earnings opportunities of the poor. Information on income sources based on the 1998/99 LSMS have not yet been processed at the time of the preparation of this report. Thus at this point little is known about how many income sources are pursued by households, or their relative imiortance. Particularly information on agricultural revenues as well as off farm sources of income are not;.,vailable. However the most important source of income for the typical Moroccan household is labor. Usually the poor depend on labor income even to a greater extent than the non-poor and labor is the main income- generating asset of the poor. Once the information on various sources of income will be available the discrepancy between the labor income of the poor and non-poor as well as the differences in pay between poor and non-poor would be investigated in more detail. It is also important to evaluate to what extent the 13 difference in labor income for the poor and the non-poor can be explained by differences in the endowments, and especially by differences in human capital. 2.22 In addition, diversification of income sources is an efficient way to cope with poverty risk. Typically rural households derive their income from multiple sources within the rural economy (agricultural own-production, production for market, sale of labor in rural markets, rental of productive assets such as land or capital goods, sale of crafts and other small-scale manufactures) and from the urban economy as well. D. EFFECT OF TRANSFERS ON POVERTY 2.23 Transfers represent an important component of all households income in Morocco, and particularly the family system does perform a set of functions similar to those performed by a public social security system (see Annex C). According to the 1998/99 LSMS, about 48% of households receive private transfers16 and about 41% of them give transfers. Transfers from abroad are received by a substantially smaller number of households (about 13%), and public transfers17 are received by about 15% of households and they are mainly targeted to urban areas. The proportion of households receiving private transfers is roughly the same in both rural (about 46% of households) and urban areas (50% of households). In urban areas, however, households are more likely to receive transfers from abroad (about 15 versus 10% in rural areas) or public transfers (20 versus 8% in rural areas). 2.24 The probability of receiving private transfers is higher for households belonging to the lowest and to the highest income quintile while public transfers are more likely to be received by households belonging to higher income quintiles. Table 7: Effects of public and private transfers on poverty by area of residence in 1998/99" Poverty Index Total Net of 2.vate transfers Net of jic transfers Net of al transfers Nat ional Head count 18.9 22.3 21.4 24.7 Poverty gap 4.4 6.5 87 10. Urban Head count 12 15.2 15 18.7 Poverty gap 2.5 4.6 9.8 12.1 Rural Head Count 27.2 30.5 28.4 31.7 ,poverty gap 6.7 8.6 7.4 9.3 Note : 1/ Net impact of transfers should be treated with caution: it is calculated under the assumption that there are no household behavioral responses to receiving transfers and there is no substitution between public and private transfers. Source: Statistical Office and World Bank estimates based on the 1998/99 LSMS 2.25 Both private and public transfers are progressive, meaning their share declines with income and lack of these transfers would increase inequality, particularly in urban areas. However private transfers are better targeted towards poorer households and have a larger redistributive impact (reducing inequality) than public transfers. For private transfers, poorer households, first quintile, receive almost twice as much as the higher expenditure groups. Assuming that there are no household behavioral 16 The definition of private transfers (given and received) excludes those originating from Fettra or Zackat. 17 Public transfers include formal pension, family allowances, special aids, wage bonuses, insurance for work related injuries. The definition of public transfers includes all transfers from administration, parastatal, public or semi-public enterprises. 14 responses to receiving these transfers and there is no substitution between public and private transfers, in the absence of private transfers the inequality of income distribution (as measured by the Atkinson index) would increase by 14%. In urban areas, private transfers reduce income inequality by about 30%, while in rural areas the impact is smaller, about 10%. Public transfers also are progressive and tend to reduce income inequality, but to a much smaller extent than private transfers. Public transfers, which are mainly pensions (about 70%), are concentrated in urban areas, where they reduce income concentration by about 12%, but there impact in rural areas is negligible (see Table 7). 2.26 Public and private transfers have also a substantial effect on both the incidence and depth of poverty (see Table 7). At the national level, lack of transfers would increase the incidence of poverty from 19 to 24.7% and would more than double the poverty gap (from 4.4 to 10.9%). Therefore the poor would become poorer. Private transfers have a larger impact on the incidence of poverty than public transfers, particularly in rural areas. But public transfers have a larger impact on the poverty gap than private transfers. Although public transfers have a smaller redistributive impact than private ransfers, they are more effective in alleviating the condition of the very poor in the urban areas. At the national level, the severity index for expenditures including public transfers (1%) is lower than for expe-nditures net of public transfers (1.6%), showing that among the poor, public transfers are benefiting the relatively less poor. However the net impact of transfers on poverty and income distribution should be treated with caution because there are behavioral responses from the households depending on whether they get transfers or not. 15 CHAPTER III WHAT ARE THE MAIN FACTORS THAT EXPLAIN THE OBSERVED INCREASE IN POVERTY? During the 1990s, there was a close relationship between the poverty increase and the performance of all sectors where the poor were mainly employed. The main causes of the poverty increase were the decline in agriculture growth and the insufficient growth in non-agriculture sector, particularly services and manufacturing exportable sectors. As a result, the increase in employment was insufficient to absorb the labor force and unemployment, particularly in urban areas, increased. During the 1990s, roughly 84% of the increase in poverty has been attributed to lack of economic growth and the rest to deterioration of income distribution. Increase in unemployment in urban areas seems to be an important factor in rising urban poverty. Decline in worker remittances during the 1990s, although benefiting the middle-income groups, may have also played a significant role in rising poverty. Morocco faces the challenge of generating sustained broad based growth and ensuring that the benefits of growth are distributed across all income groups. Moroccan authorities need to promote economic growth as well as foster progressive policies to reduce inequality. A. MACROECONOMIC EVOLUTION IN THE 1990s 3.1 Overall, macroeconomic performance in the 1990s has been weak as compared to the progress realized in the second half of the previous decade (see Table 8). In the early 1990s, Morocco was coming out of a nearly decade long period of strong economic stabilization and structural adjustment reforms (1983- 1991). These efforts saved the country from falling into an economic tailspin with potentially devastating consequences for the most vulnerable groups. This difficult period of adjustment coincided with a favorable evolution in terms of trade and with propitious weather that raised agricultural output. These mutually supporting circumstances enabled Morocco to achieve admirable stabilization results with growth, income expansion, and a positive impact on poverty reduction (incidence of poverty fell from 21,1% of the total population in 1984 to 13,1% by 1990/91). Table 8: Selected economic indicators for 1986-1998 1986-91 1991-98 Real GDP growth 4.1 1.9 Agriculture 4.0 -1.8 Industry, of which: 3.6 3.1 Manufacturing 4.5 2.6 Construction 3.8 1.7 Services 4.5 2.6 Real GDP non-agriculture growth 4.2 2.8 Real per capita private consumption growth 1.9 -1.4 Investment/GDP 22.4 21.5 Real growth in exports 14.3 4.9 Real growth in manufacturing exports 20.2 6.1 Contribution of Export to Non-Agr GDP growth 1.4 0.5 Urban labor force growth 5.0 3.3 Urban Employment growth 4.6 3.0 Unemployment 15.3 17.9 Note: Growth rates are annual. Source: "Morocco: Sources of Growth", World Bank, April 2000. 16 3.2 During the 1990s Morocco saw a reversal of fortune from the 1980s which led to poverty increase. The economy's growth rate fell by around 2 percentage points in the 1990s and non agricultural GDP growth rate dropped from an annual average of 4.2% in 1986-91 to 2.8% in 1991-98. The decline in GDP growth during the 1990s was driven mainly by the drop of agriculture value added and the fact that non- agriculture GDP was not expanding fast enough to absorb the effects of the declining growth pace of this traditional sector. Also, Morocco experienced an unpredictable growth pattern between 1991 and 1998 which has negatively affected poverty: four years of negative GDP growth rates, of which two episodes of sharp fall (4% or more). Moreover, slow expansion rates in services, industry and manufacturing sectors, which were driving forces of growth in the 1980s, have been the major factors contributing to the weak performance of non-agricultural GDP growth during 1991-98." In addition to agriculture and services, where the poor are mainly employed, the construction sector has also been growing much slower during the 1990s. 3.3 In fact, the impact of the agriculture sector on economic growth is increasingly st ong and negative. It seems that the performance of the agriculture sector has an impact on non agricullure GDP growth. The contribution of agriculture to GDP growth went from 0.8 percentage points in 1986-91 to minus 0.3 percentage points in 1991-98. The service sector also reduced its contribution to the overall GDP growth rate of almost one percentage point, while the industrial sector maintained its contribution unchanged (see Figure 2). This was mainly due to lower growth in manufacturing. With an: average growth rate of 4.5% per year, the manufacturing sector had stimulated industrial growth in the sezond half of the 1980s, while its annual average growth has decreased to 2.6% in 1991-98. This decline prevented sustained growth in the industrial sector, and accounted for a third of the decline of non-agricultural GDP growth between the two periods. 3.4 The decline of the service sector can also be partly attributed to its strong dependence, through demand effects, on economic developments in the rural sector, although some other important sub- sectors, such as tourism, which are insulated from agricultural shocks, also performed relatively poorly during most of the 1990s compared to the late 1980s. Figure 2: Morocco: sectoral contribution to GDP 5.0% $ 4.0% - 1Agric 4 3.0% - 1.0% - 0.0%0 Source: "Morocco Sources of Growth", World Bank, April 2000 1 See "Morocco: Sources of Growth", World Bank, April 2000. 17 3.5 Sluggish economic growth, F4"3: ItvateQmas#ionPerO#tafbrl 98 particularly in the agricultural sector, has = led to a decline in living standards of the 3100 households and to an increase in poverty. 300 According to the national accounts, real 20 private consumptionl9 growth rate fell by c U 2800 more than 3.7 percentage points in the 1990s, from an annual average of 4.1 percent over 1986-91 to 0.3 percent between 1991 and 1998, and real private 2W0 consumption per capita dropped from DH 240 3160 in 1991 to around DH 2860 in 1998, a si as84a's?'A789991 92 2 94 9 9 at 1980 prices. This decline is equivalent to an annual average decrease of -1.4 percent, + -4 " 4ogamW(5 compared to an increase of 1.9 percent over the 1986-91 period. Source: "Morocco: Sources of Growth", World Bank, April 2000 B. CONTRIBUION OF SLOW ECONOMIC GRowrH To POVERTY INCREASE 3.6 Two factors can explain poverty increases over time: average incomes decline and/or worsening of income distribution (increase in inequality). At the national level relative inequality hardly changed from 1990 to 1998: this result applies to all per capita expenditure deciles (see Table 3 in chapter II). However, the estimates of inequality measures which are more sensitive to changes occurring in the left tail of the expenditure distribution point to the presence of non negligible redistributive effects among the poor.20 In fact, the overall inequality trend appears to be masking different patterns of inequality across urban and rural areas, particularly in the lowest deciles (deciles 1 to 3). 2' 3.7 According to LSMS data, over the 1990s, inequality in urban areas decreased: in 1990/91 the lowest 3 deciles received 20.9% of total per capita expenditure, while in 1998/99 the same deciles accounted for 22% of total expenditure. This compares to rural areas, where inequality increased, the above percentages being 26.6% in 1990/91, and 25.4% in 1998/99. 3.8 What are the relative contributions of lack of growth and change in inequality to poverty increase during 1990/91-1998/99? 22 At the national level, 84% of the change in aggregate poverty incidence is attributable to the lack of growth in mean per capita expenditure, and the change in inequality accounts for only 16% of the poverty increase. The same result is obtained by decomposing the poverty gap index. On the other hand, the decomposition of the severity index shows that although the lack of growth is still responsible for the largest part of the increase in poverty (72%), the worsening of inequality nationally did play a significant role (28%). In urban areas, the reduction in inequality was an important force 19 Expressed in local currency and deflated by the CPI - 1980 prices. 20 At the national level, during 1990 and 1998, the Gini coefficient increased by 0.5%, whereas the Atkinson measure, which is more sensitive to changes in inequality among the poor, increased by 0.9%, 2.4% and 5.4% for the coefficient of inequality aversion set equal to 1, 2, and 4 respectively. Thus, if more weight is given to the poor, the increase in aggregate inequality during the 1990s confirms worsening of inequality among the poor. Changes in Atkinson index between 1990 and 1998 are statistically significant particularly when the inequality aversion parameter is set higher than 2. For more details see Annex A. 21 See Datt and Ravallion (1991). 22 This is based on the methodology introduced by Ravallion and Huppi (1989) which allows to identify two causes of a change in poverty: a change in the mean expenditure level at given inequality, and a change in the inequality of expenditure at given mean. The former can be interpreted as the "growth effect on poverty" while the latter is the "distributional effect". 18 offsetting the increase in poverty caused by the lack of expenditure growth. In fact in urban areas, the redistribution that occurred amongst the poor dampened the negative impact on poverty attributable to lack of growth in poors' income. In the absence of redistribution which favored the poorest, the increase in aggregate urban poverty would have been much more dramatic than it actually was. On the other hand, in rural areas the decomposition of the poverty change shows that both the growth and the inequality component contributed to the increase in poverty: the effect on rural poverty due to the decrease in the mean expenditure was significantly fostered by the changes in the distribution. Table 9: Decomposition of poverty change for 1998/99 Poverty increase due to: Total change in Growth Redistribution Poverty measure Areaport %()(% poverty (%) (%) (%) National 45.0 84 16 Headcount Index Rural 57.9 48 52 Urban 51.1 123 -23 National 63.7 84 16 Poverty Gap Rural 75.8 45 55 Urban 69.4 156 -56 National 92.7 72 28 Severity Index Rural 118.3 3 62 Urban 79.5 173 -73 Note: The residuals of the decompositions have been averaged out (see Ravallion, 1992). Source : Statistical Office and World Bank estimates based on the 1998/99 LSMS C. RURAL SECTOR IS IN AN IMPASSE 3.9 In Morocco, about 47% of the population and over 60% of the poor are located in rural areas and earn their living mainly from small agriculture and off-farm activities. Given that the two years of the surveys, 1990/91 and 1998/99, were good agriculture years and climatic conditions were comparable, the poverty incidence in rural areas may be even underestimated. If 1998/99 was a drought year, the incidence of poverty in 1998 would have increased from 19 to 21 %.23 The poor performance in the agriculture sector during the 1990s is therefore the key determinant of the living standards of rural households and is potentially the most important factor explaining the recent increase in poverty. Empirical evidences are showing that in countries where most of the poor live in rural areas, agricultural growth is significant for poverty reduction2. During the 1990s, both Government agriculture policies and price evolution of cereals and cattle products have not favored small and vulnerable rural producers. Once the agriculture information of the 1998/99 LSMS will be available, the impact of these policies on the poor should be evaluated in more detail. 3.10 Agriculture continues to account for almost 20% of GDP and employs about 40% of the labor force. Even though a real potential in other activities, such as tourism, exists, the economy still lacks diversification away from agriculture. The continuous vulnerability of agriculture to recurrent droughts and its strong impact on GDP makes this lack of diversification even more problematic in the context of a rapidly growing labor force and increasing poverty. Furthermore, the climatic fluctuations have exposed large shares of the rural population to extreme vulnerability. 3.11 Despite high border protection, the price policies and the State interventions on cereal markets, have neither ensured domestic cereal supplies nor favored the small and vulnerable rural producers. In the mid-1980s, the Government, with the objective of "self-sufficiency" in food production and through cereal support prices, has promoted the expansion of cereal production on less productive lands in areas 23 Simulated head-count index are calculated after adjusting the expenditures of the agricultural households. 24 See WDR 2000 on poverty. 19 with lower and high rainfall variability. As a result, despite inappropriate rainfall climate in Morocco, agriculture sector experienced phenomenal growth mainly due to the expansion of cereal cultivation, and in particular soft wheat. Higher level of profitability allowed Moroccan cereal producers to increase the quantity and improve the quality of inputs, in particular seeds and fertilizers. However, since that time, the tendency has been rather stagnant, and agricultural value added proven to be highly vulnerable to drought-induced fluctuations. Given that cereal production is the major driver of agricultural GDP, its high variability to climate conditions has led to high fluctuation in agricultural GDP. Within this incentive structure, agricultural resources are pulled into cereals production instead of export crops on which Morocco has a comparative advantage. Therefore, the current agriculture policy, which is encouraging low productivity and value-added growth, creates obstacle to employment creation in dynamic and labor intensive exportable productions. Since most vulnerable small cereal producers are located in marginal zones and most of the rural poor are landless labor, they are negatively affected by high variability of cereal production, while large farmers and traders are the main beneficiaries of the support price policy (see Chapter IV). 3.12 Over the 1990s, price evolution of cereals and cattle products - livestock is an important source of revenue and plays a saving role for the small farmers - have also negatively affected the income level of rural households. Since 1992, real prices of the main cereals - soft wheat, hard wheat and barley - and animal products have been oriented downwards. D. EXPORT GROWTH, EMPLOYMENT AND WAGES 3.13 Export and Employment Growth. During 1986-91, the main factor explaining poverty alleviation was the fast expansion of job opportunities in the manufacturing exportable sector, mainly clothing and textiles. During this period, growth in manufacturing exports was mainly explained by strong gains in manufacturing competitiveness, namely due to falling unit labor costs and a depreciating real effective exchange rate, in addition there was preferential access to European markets for Moroccan products. However, during the 1990s these trends were reversed. Weak export performance has been the leading cause of declining output growth in manufacturing and slow employment generation. The substantial slowdown of export growth in the 1990s (in dollar terms, overall exports drop from 14% annual average growth in 1986-91 to 5% in 1991-98) was the result of the deterioration in competitiveness particularly due to the appreciation of the real effective exchange rate and the increase of unit labor costs in manufacturing. 3.14 The expansion in employment during 1986-91 was concentrated in low-skill, low-pay jobs, mainly under the form of temporary jobs. The expansion of low-pay jobs had a significantly positive impact on poverty alleviation. Although workers were being paid lower wages, labor force participation rates increased, with a net effect of increasing average household income. But the halt in manufacturing export growth since the early 1990s, also meant a halt in job expansion, particularly for the low-skilled labor, and an increase in poverty incidence. 3.15 Increase in Urban Labor Force and Worsening of Urban unemployment. The urban labor force has grown at a faster pace compared to the rate of job creation: labor force grew during 1991-98 at around 3.3% per year (compared to 5% per year during 1986-91) while employment grew by 3% (compared to 4.6% in 1986-91). However, the decrease in labor force growth in the 1990s is mainly explained by the decline in participation rates for women who are getting discouraged from looking for jobs in the formal sector. Main supply-side factors explaining rapid labor force growth are population growth and steady inflow of rural immigrants. Expansion of urban labor force was coupled with a decline in the demand for workers because of a lower GDP growth and contraction in export growth, and as a result urban unemployment increased. 20 Table 10: Competitiveness and Employment in Manufacturing (per annum growth) 1986/91 1991/96 1986/96 Wages 0.7% -0.1% 0.6% Labor Productivity 6.0% 1.1% 2.9% ULC -6.6% 3.1% 0.1% REER -1.8% 2.3% 0.3% Employment Manuf. 8 0% 1.5% 4.2% Notes: ULC is the wage rate /productivity; REER is measured by reference to a weighted average exchange rate of the currencies of the Morocco's trading partners Source: UNIDO & World Bank staff estimates using least square growth rate. 3.16 Although population growth has declined (1.9% by the end of the 1990s), rural migration continues to put pressure on urban areas. Approximately 220,000 people leave rural areas for th e cities every year - an increase of 30% over the previous decade. Although this can explain to some extent the faster increase in the absolute number of poor in the cities, urban poverty is increasing almost four times faster than the rurallurban population shift: most of the poor still live in rural areas but between 1990/91 and 1998/99 the number of poor in urban areas doubled (from 0.9 to 1.8 million), compared to an increase of 43% in rural areas. Moreover, urbanization is likely to continue in the next decade: the proportion of the urban population in Morocco (54% in 1998) is still low compared to the average urbanization in other lower middle-income countries (63%). 3.17 Official estimates show that not only has urban unemployment grown, but total employment has declined in absolute terms since 1997 due to job destruction in recent years. As a result of slow economic growth and sluggish job creation, the total number of unemployed (urban and rural) reached 1.5 million people in 1999, and urban unemployment rate in 1999 reached 21.8% compared to 15% in 199 1. 3.18 What is the impact of unemployment increase on poverty? As shown by the 1998/99 LSMS data, the increase in unemployment is correlated with the increase in poverty among the unemployed. In 1998/99, the unemployment rate among the urban poor reached 31.5%, compared to 30% in 1990191 and the number of unemployed members inside the households is higher among the poor than the non-poor and has been increasing during the last decade. About 25% of the urban poor in the 15-24 age group and about 16% of those in the 25-44 age group are unemployed. While at the national level about 14 and 8% are unemployed among the two age groups, respectively. 3.19 The composition of the unemployed has also changed in the last few years. Most of the unemployed continue to be young and have no previous work experience, and unemployment among household heads at the national level is stable (around 4%) over 1991-98. However for th. poor, unemployment among household heads, although much higher than the national level, has declined over the period: in 1998/99 among the urban poor household heads, 10.3% are unemployed (3.7% in rural areas) compared to 13% in 1990/91 (6.4% in rural areas). Unemployment among higher education graduates has jumped from about 20% in the mid-1990s to above 30% in 1999, and the duration of unemployment spells among the better educated is also getting longer. For the poor, the more educated also show higher rates of unemployment: in 1998/99 those with no schooling had the lowest rate of unemployment (1.9% compared to 2% in 1990/91) but those with primary schooling report a 11% rate (10% in 1990/91) and those with secondary school reached 25% (14% in 1990/91). 25 See the annual Urban Labor Force Survey (ENPAU). 21 3.20 Evolution of Wages. Average real wages in the manufacturing sector have almost stagnated during the 1990s and labor productivity growth, measured as real value added per worker, has slowed considerably from 6.0% p.a. during 1986/91 to 1.1% p.a. in Figure 4: Nominal and Real SMIG (Deflated by CPI) for 1980-98 1991/96 (Table 10). During 1986- 800 2000 91 unit labor cost in -Z 6 manufacturing decreased 600-- 0 0 E substantially (-6.6% per year) f 1200 boosting competitiveness in 400 . specific sectors like textiles and 800 clothing, but between 1991 and 200- 400 E 00 1996, the unit labor cost increased by 3.1% per year, reversing the 0 1 10 (M < 1* 100 denote pro-poor public spending; <100 denote spending in favor of the better off. Source : Statistical Office and World Bank estimates based on the 1998/99 LSMS. Reform Options 4.14 Investment in education is a crucial factor to increase the level of education of the poor and build up their human assets. This is particularly important in order to break inter-generational cycles of poverty and give poor children today the opportunities that their parents never had. The Government is giving high priority to reforming the education system and in 1999 a Royal Commission on Education Reform (COSEF) was established. The reform measures proposed by COSEF include: (i) launching the process of decentralization by gradually delegating management to regions and promoting local responsibilities; (ii) introducing cost recovery at the secondary and higher levels; (iii) recognizing language of instruction issues; (iv) increasing partnership with other actors (private sector, NGOs, communities, parent -teacher associations); (v) increasing efficiency through more multi-grade and double shift classes in primary schools, and higher student/teacher ratios; (vi) redesigning all education cycles to increase the economic relevance of the curricula and improve the quality; (vii) rationalizing staffing across and within primary and secondary levels; (viii) improving human resources conditions; and (ix) strengthening teacher training programs and upgrading teaching methods, both in basic and secondary education as well as in post- secondary education institutions. However, this large reform agenda does not clearly set the priori ties and address the institutional capacities of implementing agencies. Based on findings from the 1998/99 LSMS data, increasing access and improving quality of the basic and secondary education, and reducirg adult illiteracy are among priorities to build up human assets of the poor. * Universal Lower Basic Education: As shown by the 1998/99 LSMS, over the last decade, despite raising poverty, increase in school enrollment has helped to reduce child labor. To increase lower basic schooling and reduce gender gaps, factors that are affecting low school attendance of the poor, such as indirect costs and access, need to be addressed. Although reaching under-represented groups may be relatively costly, it would require (i) reallocating the education budget toward basic schooling (with special priority on serving rural girls); (ii) designing the school facilities and providing extra services and additional incentives to attract rural children, particularly girls (i.e. school meals, free books, cash transfers for attending school, etc.); (iii) continuing building school facilities in remote rural areas while combining it with quality improvement measures; (iv) hiring local teachers (particularly women); (v) expanding BAJ experiences in other poor rural provinces; (vi) adapting to the extent possible school schedules to local habits and culture and expanding more flexible school timetables; (vii) fostering school autonomy in order to give accountability and ownership at local level; and (ix) continuing to work with NGOs and non-formal education programs while evaluating and monitoring their outcomes. To reach universal lower basic enrollment, detailed costs need to be identified. 34 * Increase Access and Improve the Quality of Higher Basic and Secondary Education: As shown by the 1998/99 LSMS the main reasons for high dropouts at 12 years of age, is lack of access to upper basic education (mainly in rural areas) and lack of incentive and value-added in staying in school. Improving the quality and increasing access to higher basic as well as secondary education would help: (i) keeping children out of labor market; (ii) giving the opportunity of continuing school for those who finish the lower basic education; and (iii) increasing the skills of the future labor force. Once access to lower basic education is increased, it is critical to address access to secondary education. The existing distributional gap between urbanized and rural Morocco, and between men and women, will not be eradicated unless there is truly equal educational opportunity for all Moroccan children. The cost of increasing access to upper basic and secondary, which may be substantially high, needs to be estimated and financing sources need to be identified. Among various financing mechanisms, introduction of selective cost recovery at the secondary and higher education levels could be considered. * Adult illiteracy reduction: The LSMS data confirm the strong correlation between poverty and illiteracy. The Government plans to reduce overall adult illiteracy to around 20 percent over a period of fourteen years. This would require expansion of the adult literacy programs initiated by the Ministry of Social Development to reach about 500,000 people a year, compared to the current level of about 100,000. To raise the literacy rate the Government has begun to support through public funds the use of flexible mechanisms by emphasizing partnerships with NGOs, the private sector, and civil society as well as other government agencies (e.g., MEN) and the performance targets are monitored (such as numbers of adult illiterates reached). However the quality and effectiveness of these new mechanisms need to be evaluated before being expanded. Roughly the cost of expanding the literacy program to half a million people a year would be about DH 90 million, which is about three times its current budget (about DH 30 million), and the Government needs to identify the necessary financing sources.41 However the cost of expanding adult literacy is only about 10% of the cost of new civil servant hiring each year (roughly DH 900 million). C. HEALTH CARE SYSTEM 4.15 The Government - largely through the Ministry of Health (MOH) - is responsible for basic public health activities, management and regulation of the sector, and is the major provider of services at all levels (basic and hospital care -- including general and specialty hospitals) as well as the social safety net for the poor.42 In principle, all Moroccans are eligible to receive health care in MOH facilities, (health centers, dispensaries, diagnostic centers, and public hospitals), either free of charge, if indigent, or by paying subsidized fees for the better-off. 4.16 Morocco does not have a compulsory health insurance system, but it has a voluntary system for the formal sector including mutual insurers for civil servants, certain categories of professionals (largely in banking), and public enterprises; mutual and private health insurance for private sector enterprises and individuals; and a compulsory social security system (CNSS - Caisse Nationale de Sgcuritj Sociale) which does not provide medical insurance but does provide some limited health benefits for children through its family allowance system. Formal health insurance coverage is quite limited, covering about 15% of the population, three-fourths of whom are civil servants. Among civil servants, about 80% are voluntarily enrolled in CNOPS (Caisse Nationale des Organismes de Privoyance Sociale) but since November 1999 the system has become compulsory. According to the 1998/99 LSMS, almost 99% of CNOPS beneficiaries are in urban areas and non-poor. 41 Cost of adult literacy is estimated at about DH 180 per person. This includes (i) the cost of literacy course which is about DH 130 for a 200-hour course; and (ii) the books, estimated at DH 100 but is expected to be reduced to DH 40. 42 See "Morocco Health Financing Brief', WB, December 1999. 35 4.17 Access. The 1998/99 LSMS data indicate that access to health services varies by household living standards (see Table 16). Most of the poor use the services of public health facilities while the better-off opt mainly for the private sector. However, the data reveal that there is strong evidence of self- selection. The non-poor clearly opt out of the public sector: in urban areas, 75% of individuals in the richest expenditure group chose private health care (compared to 23% for the poorest), while in rural areas the percentage is 61% (compared to 28% for the poorest). Table 16: Access to health services by region and household consumption expenditure quintile for 1998/99" uFbwo Riir4l..tin~ Quintile I Quintile V Quintile I Quintile V Quintile I Qu,ntile V % who reported being ill & consulted health facilities2' 595 80.0 40.0 73.9 45.1 77.2 Sector: - (%) Public 76.1 20 5 66.8 38.5 70.6 29.3 - (%) Semi-public 0.5 4.7 5.0 0.2 1.6 3.1 - (%) Private 23.4 74.8 28 2 61.3 27.8 67.6 Total 100.0 100.0 100.0 100.0 100.0 100.0 Person consulted: - (%) Doctor 89.1 89.9 68.4 85.6 73 3 89.3 - (%) Pharmacist/Dentist 2.4 6.8 5 4 6.7 2.9 64 - (%)Nurse 4.7 0.2 19.1 39 16.7 0.8 - (%) Other 3.3 1.8 7.2 3.6 63 2.2 Total 100.0 100.0 100.0 100.0 100.0 100.0 Distance (kin) 7,9 13.7 17.7 34.4 14.8 14.9 Health insurance coverage (%) 5.6 43.2 1.8 7.9 2.4 35.0 Notes: 1/ Quintiles refer to total individuals classified by household consumption expenditure per capita 2/ Refers to percentage of people reporting an illness in the preceding month of the survey Sourre: Source : Statistical Office and World Bank estimates based on the 1998/99 LSMS. 4.18 Morocco' s health delivery system has improved substantially over the past 30 years, but this progress has been uneven across the territory. According to 1998/99 LSMS data, the travel time and the costs in seeking medical care represent serious access obstacles for households in many of Morocco's rural areas. On average, rural residents are 21 km from a public health facility (31 km from a private one), while urban residents are 5 km from public (11 km from private facilities). On average, the rural s' travel time to reach a public health facility is almost three times as much as for individuals living in urban areas. About 50% of urban respondents were able to reach the health facilities on foot, compared to 14% for the rurals (65% of individuals living in rural areas needed a vehicle) (see Table 16). 4.19 Utilization of public/private health facilities and the type of treatment sought by individuals differ also within expenditure quintiles (Figure 9). For instance, of all those reporting ill in the poorest quintile, about one-third took treatment from primary basic dispensaries (either public or private), about 20% from public hospitals and private doctors, 17% from primary health centers. The LSMS data also show that (i) health care provided by private doctors is predominant across all expenditure quintiles except the poorest; and (ii) visits to private sector exceeded those to public hospitals for all quintiles. 4.20 High utilization of private health facilities confirms to some extent that both the inefficiencies and low quality of the public health system are perceived by the households. According to 1998/99 LSMS data, in urban areas, about 60% of the poor sought treatment either in public hospitals (32.4%) or in dispensaries (27.4%), to be compared to 33% for the non-poor. In rural areas, dispensaries are the most common choice by the poor (35.8%); however, a relatively high percentage of the poor (20.895) were treated by private doctors, compared to 15.3% by public hospitals. 36 Figure 9: Utilization of Health Facilities by Expenditure Quintiles for 1998/99 600- 500- 40 0- 10DispenaryExp.425.7t18. 14.6 7.7 4.21 Bud ealtcetr Acorin7 t Natinal 14alt Accunt fo 39798. iity6fHat 2%,voluntary m ras adri20.4 for2abo46 intrntia arai a .9 3ce t a ri. 6.4 IS Other 5.8 3.8 2.6 3.5 6.2 Expenditure quintile Source : Statistical Office and World Bank estiates based on the 1998/98/99 LSMS. 4.21 Budgetary costs. According to National Health Accounts for 1997/98, Ministry of Health expenditures represent 22% of total health expenditures, other Ministries health expenditures represent 2%, voluntary mutuals and private health insurance account for about 16%, direct out of pocket payments from household account for 54%, private enterprises and public organization represent 4% and the international cooperation and local collectivities account for the remaining 2%. Public expenditures on health in Morocco, estimated at around 1.1% of GDP in 1998/99, compared to 0.9% in 1990, are well below those found in other comparable income countries (2.6% in other MENA countries). As a share of total government expenditures, health expenditures have increased slightly, from 3.4% in 1991 to 3.7% in 1998, and in real terms they have increased by 2.9% p.a. over 1991-98. Since about 86% of the spending is for operating expenses, this increase was largely driven by the increase in salaries (+3.7% per year over the same period). However, expenditures for equipment decreased substantially over the last decade (-2.4% per year).o Moreover a disproportionate share of the MOH recurrent expenditures is for hospitals (78.8%) and the remaining is allocated to preventive care. Overall, 47% of the MOH expenditures are affected to public hospitals (including teaching hospitals). Among public hospitals, the two teaching hospitals, which are located in Rabat and Casablanca, are absorbing about 30% of recurrent expenditures of the MOH. 4.22 Health costs for the Household. The 1998/99 LSMS data show that the expenditures for health vary with household standard of living, ranging from about 2% of household budget share for the poorest households to 6% for the richest. Households' expenditures for health increase constantly with total expenditure, both in urban and in rural areas. On average, the yearly per capita expenditure on health for urban households (DH 511) is more than two times as much as for rural households (DH 189). Overall, 43 This pattern is consistent with the presence of quality problems concerning the health facilities, particularly the lack of critical equipment and supplies Most equipment is over 15 years old: 40% of operating rooms, 22% of reanimation; 39% of lab, and 32% of radiological equipment. 37 in both urban and rural areas, households in the poorest deciles allocate a lower share (2.2%) of their expenditure to health than those in the richest deciles (5.8%). 4.23 Although health care is in principle free for the indigent, the 1998/99 LSMS shows that out-of pocket payments for health services are far from being negligible. At the national level, the household budget share on health for the poor is 2.4%, compared to 4.3% for the non-poor. In urban areas, the budget shares on health for the poor and the better-off are 3.1 and 4.7%, respectively, compared to 2.1 and 3.7% in rural areas. The same results are also obtained using the quintiles: at the national level, the budget share for health services in the lowest quintile is 2.3% (2.9% in urban and 2.1% in rural), which compares to 5.5% in the highest quintile (5.6% in urban and 4.9% in rural areas). Table 17: Distribution of people utilizing health facilities by expenditure quintile for 1998/99 (%) Urban areas Dispensary" 23.9 18 2 24.9 25.5 7.6 100. Health center 2 19.8 18.7 34 1 17.7 9.8 100 C Public hospital 16 1 19.5 17.6 257 21.1 100 C Private consultation 43 10.0 16.1 24.9 448 100 C % of sick people with 59.5 67.5 69.7 72.9 80.0 71'7 consultation Rural areas Dispensary 11.8 21.1 248 20.0 22.5 100.0 Health center 14.7 13.9 21 2 30.2 20.0 100.0 Public hospital 95 11.9 17.6 21.3 39.7 100.0 Private consultation 4.1 10.0 18.2 24.6 42.9 100.0 % of sick people with 40.0 40.0 54.1 59.5 73.9 56.3 consultation National Dispensary 17.5 22.3 20.1 22.5 17.7 100.0 Health center 13.9 21.0 250 26.9 13.2 100.0 Public hospital 9 0 15 0 21.0 22.7 32.3 100.0 Private consultation 3.8 9 5 13 8 25.1 47.9 100.0 % of sick people with 45.1 543 64.9 71.1 77.2 660 consultation Notes: (i) This table shows the percentage of people reporting access to health facilities in the preceding month of the s Irvey. (ii) Dispensary is the lowest level of health care without any doctor; Health centers have a doctor and provide the basic health care services; (iii) Quintiles refer to total individuals classified by household consumption expenditure per capita. Source* Statistical Office and World Bank estimates based on the 1998/99 LSMS. 4.24 Distribution of Health Subsidies. Based on preliminary 1998/99 LSMS data, the pattern of utilization of public health facilities provide the following observations (see Table 17): * The better-off obtain more often health consultation, no matter what health facility is chosen, particularly in the urban area. At the national level, 45% of the poor have access to medical consultations compared with 77% of the rich. * The better-off often use private consultations no matter their area of residence, while the poor most often resort to public health facilities (dispensary, health center and public hospital). 38 * Public hospitals. At the national level as well as in the rural area, the richest 20% of the population benefits four times more than the 20 percent poorest individuals from public hospitals. But in the urban area, the poor as well as the rich benefit from public hospitals. * Health centers. In rural areas, community health centers are used by all expenditure groups, while in urban areas health centers are mostly used by low expenditure groups: the poorest 40% of urban population benefits two times more from health centers than the 20% richest urban individuals. * Dispensaries. In urban areas, dispensaries are largely targeted to low and middle expenditure groups: the poorest quintile is three times more likely to use dispensaries than the richest quintile. Although in rural areas all expenditure groups use public dispensaries, the richest quintile benefits two times more than the poorest quintile from public dispensaries. 4.25 Options for Reforms. The 1998/99 LSMS data indicate that limited resources combined with poorly targeted public health programs and inefficiently run public hospitals have resulted in (i) inequitable distribution of health services and therefore the urban-rural disparity is still far from being reduced; (ii) both access and utilization levels are strongly biased toward the urban households; (iii) the rural population and urban poor are deprived from adequate provision of health services, particularly less- off women and children in rural areas; and (iv) due to low quality of public services, utilization of private services by all expenditure groups is higher than utilization of public hospital. The low level of utilization of public services indicate the need to increase both the access and the quality of services. To upgrade health service coverage, and improve access and quality of the health care for all income groups, based on other sector work" the following measures should be considered/strengthened: * Reducing infant and maternal mortality rates. This would require strengthening maternal and child health programs, especially in rural areas. These issues will also need to be linked to improvements in health education, female literacy, gender issues regarding health personnel, and education particularly for women in rural areas. * Improving equity and expanding health coverage. This could be achieved through redirecting health care resources towards primary health care, emphasizing on cost-effective interventions and strengthening partnerships with private sector, non-governmental organizations and local communities and authorities. The strategic objective would be to (i) improve quality and access to health services for the population in both urban and rural areas, and (ii) reduce rural-urban and inter- regional inequities in access to health care services. * Ensuring long-term sector financial sustainability. Based on the 1998/99 LSMS, the use of private services by all income groups has underlined the poor quality of public health facilities but also to some extent has shown the willingness to pay, by all expenditure groups, for quality services. Currently the public cost recovery efforts are weak, low compared to the real costs (less than 10% of costs), inefficiently enforced, and resulted in large hospital subsidies to the better-off. Moreover, the existing insurance system (CNOPS) has led to discrimination in favor of urban areas. Improving financial sustainability of the health system will require (i) health financing reforms that will enhance health sector revenues, improve financial access to cost-effective health services and move towards universal health insurance coverage; and (ii) hospital reforms which would improve efficiency of the system and promote cost recovery. The first step in this process could be to introduce a compulsory health insurance scheme for the formal sector and expand access to health care services by putting in place a national health assistance fund for indigents. To address cost recovery, a large scale See "Morocco Health Financing Brief', World Bank, December 1999. 39 evaluation of the impact of fee increases requires knowledge of price elasticities across different expenditure groups and real willingness to pay for measures, but such estimates are not iurrently available. * Increasing the public financial resources for the health sector. Investments through BAJ projects as well as projects financed by other donors have helped to increase access to health care services in poor areas. However, they are still insufficient given the rural disparities. To improve quality and increase access to health services there is a need for an increase in the efficiency of public and private spending in order to get more value for the monies spent. In addition, the Government ieeds to consider increasing resources for the health sector, as public health spending is below the level in other comparable income countries. * Improving the management of the system. This could be achieved through implementation of effective decentralization policies and the strengthening of core functions at the central level. D. SOCIAL INSURANCE SYSTEM 4.26 The social insurance system protects those in the formal economy that are wage earners and their dependents. Benefits include family allowances; limited health insurance (see Health Section); old age, and disability and survivors insurance. With about 2,300,000 contributors, the system provides (i) pension coverage for only 28% of the labor force; (ii) access to formal health insurance for less than 1591 of total population; and (iii) family allowances for about 470,000 wage earners (about 6% of the labor force or 13% of the population). Table 18: Social Insurance Contribution Rates Private sector (CNSS & Civil Service (CMR & CIMR) RCAR) Old age and survivors - 9.12% (6.08% employer 14% (7% employer & 7% insurance & 3.04% employee) employee) Disability insurance - 0.66% (0.44% employer & 0.22% employee) Health insurance Mutuals or Priv. Insurance CNOPS/Mutuals (voluntary) 7% (3.5% employer & 6% (3.5% employer & 3.5% employee) 2.5% employee) Family allowances 8.87% employer (for Budget financing CNSS) Total 25.65% (18.89% employer 20% (10.5% employer & & 6.76% employee) 9.5% employee) Source: Official data. 4.27 CNSS covers about 1 million workers in the private sector and CMR covers about 800,000 government employees. Mandatory contributions, which are not high by international standards, are for the old age benefits and family allowances: for the private sector the mandatory payroll taxes paid to the CNSS represent 19% of the gross wage bill compared to 14% to CMR for the civil servants (see Table 18). Total contributions including the voluntary health insurance contributions to mutuals and private insurance reach 26 and 20% for the private and public sector respectively. However, the weak links between contributions paid and benefits received tend to give the perception that contributions (for s Total 2,300,000 contributors, represent around 45% of the urban labor force, high for regional standards. Taking into account about 15 million people who are between 15 and 59 years of age, and the labor participation rate of about 30%, total labor force (including rural workers) would be about 8 million. Therefore total covered population by the insurance system is about 28%, placing Morocco on the low side of coverage by regional standards. 40 pension and family allowances) are a tax rather than a deferred wage and therefore directly affect labor supply. 4.28 Overall the poor do not benefit from family allowances and formal pension benefits. The system fails to cover those who need it most, such as those who do not have a permanent source of income, who do not have a secure job or who lose it, who are vulnerable to external shocks and sickness, death of the household head in the informal sector, and divorce/repudiation. These vulnerable groups in Morocco generally seek help from intra-household and informal support networks. 4.29 Family Allowances. Similar to other countries in the region (Tunisia), family allowances are intended to provide wage supplements to families with children, and they are financed through social insurance system by payroll taxation paid entirely by the employer (8.9% of gross wage bill for private sector and Central Government budget financing for civil servants). This is equivalent to a disguised tax which raises the labor costs in the private sector and creates distortion in the labor market. Since family benefits contributions in the private sector exceed outlays, surpluses are used indirectly (through cross- subsidization) to finance the pension benefits in the private sector. For the Government employees, contributions are implicitly netted out of their salaries. Family allowances are also unfair because, being paid by social insurance system, there coverage is limited to public and formal private sector and targeted to middle-income urban wage earners. According to the 1998/99 LSMS, the lowest quintile do not receive these family allowances. In fact the design does not differentiate between risk and poverty objectives: the benefits are intended to provide financial support to families for the cost of raising children but they are administered by the social insurance system as though they were a form of insurance. As a result the system is inequitable because there is no similar benefits provided to the poor and vulnerable with children (i.e., unemployed, elderly, disabled, widows, etc.). 4.30 Pension System. The pension system is fragmented, leading to high administrative costs and hindering labor mobility across sectors.4 It is dominated by four funds (CMR, CNSS, RCAR, and CIMR) which cover about 97% of contributors. The consolidated analysis reveals a system's dependency ratio of 0.27, which is relatively low considering that it includes disabled and survivors; annual surplus of the system is about 0.5% of GDP and total reserve fund represents about 9% of GDP. However closer inspection reveals a number of problems: the high fragmentation into different schemes affect labor mobility and increase administrative costs; there are perverse incentives embedded in some of these schemes rules; some of the schemes have annual deficits or a very small reserve fund; and the benefits are generous in principle given the available resources and the benefit formula, and yet they provide inadequate benefits in practice for most of the population. 4.31 In 1998/99 only about 4% of the elderly poor received formal pension: out of 2 million elderly, 10% have formal pensions (17% in urban areas and only 2% in rural areas); about 13% of the elderly are poor and about 4% of the elderly poor receive pension benefits. However, as shown by the 1998/99 LSMS, public transfers, which are mainly composed of formal pension, have non-negligible impact on poverty and income inequality reduction. Assuming there are no household behavioral responses and no substitution effect between public and private transfers, eliminating public transfers, mainly formal pension benefit, would increase the incidence of poverty from 19 to 21%, and particularly the poor would become poorer because the mean income gap would almost double, from 4.4 to 8.7% and the distribution among the poor would worsen. Increase in poverty gap confirms that public transfers, particularly 46 There are four main schemes: the CNSS --Caisse Nationale de Sgcuriti Sociale-- for private sector employees, the CMR -- Caisse Marocaine de Retraite-- for civil servants, the CIMR --Caisse Interprofessionnelle Marocaine de Retraites-- which is voluntary and a supplement to the benefits provided by CNSS, and the RCAR --Rggime Collectif d'Allocation de Retraite-- for temporary workers in the public sector. Additionally, six public enterprises offer their own pension plans: Bank Al- Magrib, Office Chdrifien des Phosphates (OCP), Office d'Exploitation des Ports (ODEP), Oice National des Chemins de Fer (ONCF), Office National d'Electricit6 (ONE), and Rggie des Tabacs. 41 pension benefits, affect mainly those who are already poor and in absence of public transfers the distribution of income of the poor would shift backward (the poor get poorer), and would sp:ead out (inequality increases because the poorest among the poor get poorer). Therefore based on these findings and with aging population, addressing the financial sustainability of the formal pension system is relevant from a poverty perspective because it could impose additional burdens on the budget as well as jeopardize the old age protection for the formal urban wage earners, including for the poorest among the pool. 4.32 The main shortcomings of the pension system are: (i) although the demographic structure is still favorable, the proportion of the elderly in the population is expected to grow fast, leading to a high dependency rate and imposing a potential burden on pay-as-you-go plans finances; (ii) eligibility conditions are generous and negatively affect the financial sustainability of te system (i.e.. under- reporting, low retirement age, high replacement rate, etc.); (iii) given the existing benefit formula, pension contributions in the private sector (9.12% of payroll tax for CNSS) are insufficient for covering the benefits; (iv) real financial situation of the funds is hidden due to cross subsidization among benefits: for example CNSS is already experiencing annual deficits but the family allowances, which have accumulated significant reserves, are used to cover their deficit; (v) amount of the minimum pension is not sufficient for protecting the elderly from dropping into poverty: the minimum pension paid by the CNSS is about 30% of the industrial minimum wage (SMIG) (or about 20% of the average salary in manufacturing) and the one paid by the CMR is about 45% of SMIG (or about 25% of the average salary for civil servants); and (vi) the coverage is urban biased. For more details see Annex E. 4.33 Reform Options for Social Insurance System. In addition to introducing compulsory health insurance and free health care scheme for the poor and vulnerable, the financial sustainability if other components of the social insurance system needs also to be improved. In addition given the needs of the poor, the introduction of selective social safety net scheme for protecting the poor elderly and the poor households with children could be considered. From the point of view of the poverty strategy, the following reform options could be considered: * Introducing safety net schemes for poor households with children. As shown by the 1998/9) LSMS data, family allowances are mainly benefiting the middle income groups and the well-off in the urban areas. Since there is no cash transfer scheme in Morocco for providing income support to poor families with children, as part of the safety net program, options for selective cash transfers to the poorest groups (e.g. poor unemployed, widows, elderly, disabled heads of household with children) could be considered. One such transfer can be scholarships to poor children which could be financed through general revenues. The level of benefits, targeted groups and cost would need to be evaluated for these selective reforms. * Introducing social safety nets for elderly poor. Based on the LSMS data, in 1998/99, among the inactive at the national level, 27.5% of the poor lived in households where the head is elderly. Since there is no alternative social safety for the elderly poor, introduction of targeted old-age cash transfer program which would provide a basic pension as a safety net to every poor individual who reaches 60 years of age could be considered. This scheme could be financed by general revenues and can be tested. Further evaluations are required to estimate the cost of this program. Although this type of old age cash transfer system can be expensive, in places with a tight family structure, such as Morocco, it will have the advantage of providing a safety net not only to the elderly poor but to their families. * Addressing Financial Sustainability of the Pension System. The pension system is in no immediate danger of collapsing, but some of the schemes like CNSS are already experiencing annual deficits and the civil servant pension plan (CMR) represents roughly US$2 billion contingent liability from the 42 Government which keeps accumulating in arrears. In fact inadequate pension parameters would imply a large burden on the government which is the de-facto guarantor of the pension plans for at least CNSS and CMR. This in turn could impose financial burdens on Government resources and jeopardize public expenditures in social sectors that benefit the poor. To avoid the financial insolvency and ensure that a large fraction of the population would not retire poor in the next few decades, remedial measures need to be implemented. Since there is no unique solution to this problem, the government should explore the feasibility of implementing parametric or structural reforms. Among the first, it is recommended to review benefit formula and eligibility rules. Among the second type, is the preparation of a comprehensive social security system reform that takes into account the socio-economic characteristics of Morocco and adequately hedges against the risks faced by its population as they grow older. Based on the ongoing actuarial projections for each of the pension schemes, with assistance from the World Bank, the cost of implementing different types of reforms will be estimated. E. SOCIAL ASSISTANCE MECHANISMS 4.34 Social assistance programs in Morocco, costing over 2% of GDP, consist of three programs: (i) universal consumer food subsidies; (ii) public works employment programs administered by the Promotion Nationale (PN); and (iii) in-kind welfare program for the needy administered by the Entraide Nationale (EN). Despite the increase in poverty, the budgetary cost of these programs has remained almost constant in the last decade. Four-fifths of government expenditures on social assistance programs are devoted to consumer food subsidies, and only about one-tenth go to public works. Overall they are marginally targeted to the poor and the vulnerable, they face efficiency problems and are limited in scope and effectiveness: they mostly benefit the non-poor, leakages are high, and their coverage is modest compared to the existing number of poor and needy population. This section reviews the existing assistance programs and proposes policy recommendations for improving their targeting. Consumer Food Subsidies 4.35 Consumer food subsidies were introduced in Morocco to stabilize prices of strategic goods with no explicit focus on the poor.47 Although the poor benefit from consumer food subsidies, they are a very inefficient and expensive way to help reduce poverty and have unfortunate economic side effects as well. In fact food subsidies in Morocco tend to compensate consumers for the incidence of high border tariffs, as well as to protect crop producers and industrial processors. Except for low-grade flour (FNBT) 48 which is available at an annual aggregate limit of 10 million quintals, cooking oil and sugar are universally available at subsidized prices in unlimited quantities to anyone who chose to buy them. Cost of food subsidies, which is the largest social assistance program, has reached DH 5.3 billion in 1999 (about 1.6% of GDP compared to 1.3% in 1990). Since July 1996, import quotas and licensing controls have been eliminated and the system is financed by (i) part of the custom revenues (equivalent tariffs) collected on soft wheat, sugar, oilseeds and their main derivatives; and (ii) direct budget contributions. 4.36 Distributional Incidence. Overall, the system of universal and flat food subsidies in Morocco is disproportionately favoring the rich in absolute terms, who consume relatively more of the subsidized goods. Therefore they are not well-targeted to poor consumers and cannot be justified as a mean to redistribute income to the poor: only about 25% of the food subsidies reach the poor. However their welfare effect is stronger for the poor because subsidized products account for a higher proportion of the poor's expenditure and 40% of their caloric intake. 47 See "Consumer Food Subsidies in MENA", World Bank, December 1999. 48 High-extraction rate flour is known as farine nationale de bi tendre (FNBT). 43 Table 19: Impact of Consumer Food Subsidies, by Quintile, 1995 % Absolute Avg. Subsidies per capita 1 2 3 4 5 Total FNBT 23% 24% 22% 18% 13% 100% Sugar 15% 19% 20% 21% 25% 103% Cooking Oils 11% 13% 20% 24% 32% 103% Total Absolute incidence 15% 19% 20% 21% 25% 10)% Relative Incidence : Avg. per cap. Subsidies as % of Total Avg. per Cap. Exp. FNBT 1.2% 0.8% 0.5% 0.3% 0.1% 0.3% Sugar 2.1% 1.6% 1.2% 0.9% 0.5% 0.')% Cooking Oils 0.8% 0.7% 0.6% 0.5% 0.3% 0.5% Total Relative incidence 4.0% 3.0% 2.4% 1.7% 0.9% 1.1% Nutritional Importance FNBT 4.5% 3.4% 2.7% 1.9% 1.2% 2.4% Sugar 5.3% 4.7% 4.7% 4.5% 4.3% 4.6% Cooking Oils 3.1% 2.8% 3.5% 3.6% 4.2% 3.5% Total Subsidized Calories as % of Total Calories Acquired 12.9% 10.9% 10.8% 100% 9.7% 106% Note: Detailed expenditures on food products based on 1998/99 LSMS data are not yet available. Source: Lindert and Glewwe, World Bank, 1995, based on 1990/91 LSMS data. 4.37 In absolute terms, a large portion of subsidy expenditures never reaches the consumer, and of the portion that does, only a small fraction reaches the poor. Incidence of subsidies falls disproportionately on the top quintile of the population, which receives 25 percent of total subsidy spending compared to only 15 percent for the bottom quintile.49 (see Annex F and Table 19). 4.38 Despite high degree of leakages, the subsidies in Morocco constitute an important transfer to the poor. Overall, the poor benefit four times more than the rich in relative terms from food sL.bsidies. Similar to all food related transfers, all the three food subsidies in Morocco are progressive and as share of total per capita expenditures of the household they drop whenever income increases. Moreover subsidized goods are more important to the diet of the poor than the non-poor, and the calories from subsidized goods represent a larger portion of total caloric intake by the poor compared to the rich (see Table 19). 4.39 Link between Consumer Food Subsidies and Agriculture Sector Policy. In addition to distorting agricultural production patterns, in practice, consumer food subsidies serve to counterac import tariffs and to make up for the high border protection on food items. Therefore, they mainly con stitute a transfer to producers rather than consumers: with respect to flour, for example, on average about 50% of the subsidy goes to producers; for sugar and cooking oil, producer transfers amounted to at least 25 and 21% of total subsidy outlay on each, respectively. 4.40 At current international prices, nominal protection rates for wheat, sugar and oilseeds, and their major derivatives range from 84 to 157%. This substantial public transfer to the processing units creates a distorted price and incentive structure. Consequently, competition is weak, inefficient processors and weak performers are maintained, and significant incentives for fraudulent behavior are promoted particularly for wheat subsidies (FNBT). 4.41 On the agricultural side, persistently high farm-gate prices for soft wheat, sugar crops and oilseeds generate significant economic distortions. Particularly in the case of soft wheat, it encourages substitution effects leading to (i) an undesirable extension of land sowed in soft wheat, contributing to soil erosion; (ii) miss-allocation of productive resources and increasing the variability of yields with no real 49 Incidence analysis is based on the 1990/91 LSMS data Once the 1998/99 LSMS data on detailed food expenditures products is available, this incidence analysis will need to be updated. 44 positive effect on productivity; (iii) discouraging farmers from planting crops which are better suited to the climate. For oilseeds and sugar, the protection policy has been mainly driven by industrial transformation interests rather than by producer interests. The sugar processing sector consists of a number of uncompetitive refineries that survive behind protective barriers. The oilseed refining industry is even more concentrated, and the primary refiner is well placed politically to maintain its favored position regardless of whether it procures its inputs from domestic producers or from imports. 4.42 Reforming food subsidies for better targeting the funds to the poor. The Moroccan Government is well aware that the food subsidy program coupled with the highly protectionist agricultural policy is costly and inefficient. Therefore various reform options are being assessed. Elimination of consumer food subsidies without addressing the agriculture policy and reducing tariff protection would reduce budgetary costs but it would also lead to an increase in all three consumer food prices and would result in an increase of poverty, particularly in rural areas. Therefore the best reform options in terms of better targeting the funds are (i) parallel reduction of tariff protection and subsidies on food products; and (ii) introduction of targeted assistance to low-income groups negatively affected by the reduction of domestic protection and consumer food subsidies. This would allow to reallocate budgetary funds freed up by reduced subsidy expenditures to programs benefiting the poor (e.g., alphabetization, basic services in rural areas, public works programs and assistance programs targeted to the poor). In addition in the medium term it would allow to strengthen the agricultural policy through focusing on crops where Morocco has a competitive advantage in order to ensure adequate income for rural population, including the poor. Overall, eliminating the subsidy program and cutting the custom tariffs would generate about DH 2.7 billion (or 0.8% GDP) savings that could finance the economic and social costs of the transition period (nearly three times the investment budget of the Ministry of Health for fiscal year 2000) through financing targeted assistance for those who will be affected by the reform. The dismantling of the border protection system has already started. Since November 2000, the Government has begun liberalizing oil prices, and has reduced the import tariffs on oilseeds. As a result the consumer prices have dropped and the savings are expected to be used for decentralized rural programs to compensate farmers. Nevertheless, these reforms need to be extended to the other two commodities (sugar and wheat) which are politically more sensitive. * Producers, farmers and processors. Adjustment process will affect everyone to some extent, but mostly 3540,000 sunflower producers, 60,000 farmers in regions with no comparative advantage for sugar crops, and nearly all cereal producers (about 1.5 million farmers). In total about 2 million farmers and agriculture workers would be affected by the tariff reduction. Compensatory measures could include (i) programs to help small land holders to convert to other and better suited alternatives; (ii) expansion of the existing labor intensive public works projects (PN) to assist low income households; (iii) basic infrastructure development in affected rural areas to increase access of the vulnerable population to basic social services (i.e., potable water, electricity, rural roads); and (iv) community development programs in remote rural areas. For oilseed, sunflower producers could receive a direct temporary payment to protect their revenue which will cost about DH 120 million. In the milling activity and the sugar, increasing the competitive pressure would lead to a desirable restructuring process. Within this context, the four public producers of sugar should be immediately privatized when possible, and liquidated for some of them. Here, the State could finance the social costs following a liquidation. * Consumers. Reduction of tariff protection would lead to domestic price decline which in turn would benefit consumers (domestic prices of vegetable oil, flour, would be reduced respectively by about 10% and 14%, and it would remain stable for sugar). Nevertheless the timing for the phasing out subsidies and choice of products to be eliminated at a more rapid pace should be guided by the impact of the removal of the subsidies on the poor and the share of existing subsidies accruing to the poor. Subsidies on cooking oil appear to be suitable for more aggressive elimination because of its lower 45 contribution to total per capita expenditures nutritional intake of the poor, but for sugar arid FNBT subsidies should be removed more gradually in light of their relatively larger contribution to the total expenditures and caloric acquisitions of the poor. Based on the 1998/99 LSMS the incidence of eliminating subsidies for the poor needs to be calculated once the disaggregated food information is available and the introduction of targeted food programs, particularly for poor children, should be considered and their costs estimated. Public Works Program (PN) 4.43 The Promotion Nationale (PN), an autonomous directorate under the Ministry of Interior, is responsible for the implementation of labor-intensive public works to fight rural under/unemployment and to improve rural infrastructure. Overall PN, which uses force account procedures, has been an important source of temporary employment in rural areas, although more recently it has also shifted some resources towards jobs in urban areas. Although the 1998/99 LSMS data does not allow to assess the relevance of PN on poverty, given the low level of wages and the labor intensity of the activities, there is no doubt that this program helps to provide income support to the rural population, including the poor particularly during droughts. Strengths of PN * Provide employment in rural areas through productive investments utilizing labor intensive construction methods in the provision of works and services. * Create high quality basic infrastructure work in low-income communities. * Establish participatory development process to involve communities in the planning and execution of development projects of national and/or local interest. * Provide appropriate and efficient mechanisms to create temporary employment for under/unemployed, and particularly provide income-earning opportunities to the rural poor, although the urban programs are targeted to first job seekers who are not necessarily poor. * During 1990-1999 the PN has created about 104 million person-days of employment (on average about 10.4 million person-days of employment per year or 40,000 person-year) for a total ccst of DH 4.5 billion.5) Average cost of job created per day is low, about DH 43 (or $4/day), though there are significant annual and regional variations in the number of jobs created and funds allocated by various programs. * Vast majority of the beneficiaries are unskilled workers, and are paid the agricultural minimum wage, SMAG (currently about 4$/day). * Ability to respond rapidly to regional and national crises when funds are available through creating temporary employment: PN has the administrative capacity of scaling up, particularly during droughts and inactive agricultural period, the labor intensive rural activities to generate temporary in come for the underemployed and the poor. * Low project cost compared to other similar activities: Unit cost of PN projects (particularly for social infrastructure) is comparable with similar activities undertaken by other agencies while for rtral roads PN's project unit cost is about half the cost of other agencies. * Share of administrating the program is very low by all standards: less that 6% of the investment cost compared to other countries' social funds administrative cost which is about 10% of the total 50 During 1997-99, BAJ I - the Public Works project has created about 1.8 million person-days of employment. The cost of person-day is estimated at DH 85 ($8.5). This cost is based on the assumption that 90% of the projects are rural roads with a coefficient of labor share of about 55% in average and the remaining 10% of the activities are buildings and rural water supply projects with an average labor share of 35%. 46 investment. Moreover it has an appropriate administration structure and is present in almost every region.51 Weaknesses in targeting and cost effectiveness. * Jobs created are not always based on labor intensive mechanisms and the PN's original mandate has been diluted significantly during the last decade: almost half of the resources (particularly in Chantiers Collectivit6s) are spent on supporting local government wage payments to administrative staff and other minor jobs, rather than directly helping low-income communities through labor intensive activities. Originally, this wage support to local governments, benefiting mainly the low- income unemployed, was expected to be temporary and over time to be transferred to local governments' budget. But, as a result of insufficient local budget, these activities are still financed by the PN. * During 1990-99, about 40% of the total employment was created in traditional chantiers of PN (excluding the CCs with 50% share of labor in total costs). The labor intensity of traditional chantiers varies by type of infrastructure and by region.52 For rural roads the share of labor in total project cost is between 17 and 86%; for water supply projects it is between 31 and 47%; and for social infrastructure the share of labor is around 30 and 40%. * Projects are not financially sustainable. Although once projects are completed, the local communities are involved in their maintenance through the local communities' presidents, the unavailability of resources prevents them from maintaining the projects. In fact there are no practical arrangements put in place to ensure that maintenance will take place and beneficiaries are responsible for it. * Except in BAJ provinces, there is no explicit monitoring of the outcomes. 4.44 Options for Reforms. Despite their weaknesses, PN's programs are currently the most efficient employment scheme targeted to the rural population, including the poor. Without requiring any institutional changes, PN could continue to play an important role in providing temporary income schemes for the rural population. To increase the effectiveness of public works, the PN could: * Implement large scale employment programs: PN should continue to play its role as an emergency agency providing large-scale employment in areas hit by droughts and other natural disasters. This would require phasing out the Chantiers Collectivit6s to free funds as well as ensuring that labor intensive methods are applied to implement projects in the most disadvantaged regions of the country. With the existing annual investment budget allocated to PN, amounting to DH 500 million, assuming daily agriculture minimum wage (SMAG) at DH 45, labor share in project at 50% and assuming each person will work for 20 person-days53 the program could reach 280,000 workers. In case the investment budget of the PN is doubled (DH 1 billion), the program could reach nearly 75% of the 5.3 million poor. * PN could act as an advisor to other Government agencies in selecting projects: projects should use labor intensive methods as well as being productive. But this would require effective coordination within the Government, to a certain extent already done at provincial level. * Involve private sector in PNprograms: PN could also initiate to contract-out certain projects to the private sector and NGOs. At the moment, contractors or suppliers are only used to provide materials on PN construction sites. Private contractors could be instructed by contract to use labor intensive methods and local workers, similar to the method used by PN. This would allow the PN to focus its 51 Total staff of PN is around 1,200 of which around 476 people are high-level staff and about 70 people are military staff 52 See audit report prepared for the BAJ I, March 2000. 53 In principle PN procedure mandates that each worker must cede to others in line after working for 10 days, but we are assuming that the normal turn over of workers is set at 20 days per year because the workers should be able to return if they are in need of income support. 47 resources on rural works located in the most disadvantaged regions rather than scattering its resources throughout the country as it does today. Welfare Program - Entraide Nationale (EN) 4.45 Entraide Nationale offers many different but small-scale services - literacy, basic training, shelters, pre-schooling etc. - to disadvantaged women, children and youths (See box 2). However its performance is uneven, and its coverage is inadequate: out of an estimated 5 million poor, of wl.ich some will need EN assistance, only about 80,000 people are reached by EN programs (or 1.6% of the poor). The Administration is highly centralized and organizational capacities are weak; aind both cost/performance monitoring and follow-up of outcomes are absent. 4.46 EN programs are for the most part pro-poor, but the existing programs are loosely focused and try to achieve too many objectives.54 Although the budget for EN activities is modest (about DH 200 million or 0.1 percent of GDP in 1998), most of it is allocated to recurrent costs, mainly the salaries of about 6,300 persons of which about 1,900 are permanent staff and 4,400 are temporary (all the temporary staff are women and earn about DH 1,400/month). The cost of programs by beneficiaries differs from one program to another, but in average it reaches DH 2,500 per beneficiary per year, compared to DH 3,300/year/student in basic education and DH 8,000-12,000 per student per year in professional vocational training. Box 2: Program of Entraide Nationale * Basic literacy training to mothers and basic nutritional advice for their children (Centres Soci( - Educatifs - CSE); * Training centers for girls who are illiterate and/or dropouts from schools (Centres d'Education at du Travail - CET); * Shelter for school age children and orphans (mainly boys) by subsidizing and providing staffing to local charities (Associations Musulmanes de Bienfaisance - AMB); * Pre-school centers for children between ages 3 to 6 (garderies); * Technical training (2 years) to needy children/basic education dropouts (above ages 10-12 year,) and orphans of the AMB (Centres de Formation Professionnelle - CFP); * Subsidy/food transfers to handicapped associations and training for handicapped children. Strengths of EN * EN has an important mandate to assist the poorest of the poor. In fact, it is the only public agency in the country which tries to reach the poor through in-kind services. * Although EN is reaching a small portion of the poor, it has centers located all over the country (and particularly in the South), potentially enabling it to reach many remote areas. * To some extent the self-targeting schemes are already in place due to the location of the centers and the type of services provided which are not attractive to the non-poor: although there is no information on the socio-economic characteristics of the beneficiaries, and the LSMS has no information on the beneficiaries of EN, on the basis of the location of the centers and the kind and quality of services provided, it can be assumed that most services are self-targeted. It is estimated that about 70 to 80% of EN's budget reaches poor or low-income groups. In addition to the direct beneficiaries, a large part of the EN's non-permanent staff belong to low income groups. * EN is a public agency and because of its legal status it has more flexibility in managing and implementing services than other line ministries. 5 See " Morocco: Social Protection Reform - Entraide Nationale Component"; WB 1998. 48 * EN has a tradition of private support and has a sound experience of partnership arrangements with other organizations (particularly in the case of the AMB centers). Weaknesses of EN * The administration is highly centralized, allowing little chance of local decision-making initiatives' and targeting of services. The D616gu6 (regional coordinator at the provincial level for EN programs) has no power to make decisions about personnel, nor activities. * Given the lack of a comprehensive social strategy, EN's activities have become scattered, with little focus, and some activities are less well targeted to the poor (e.g., "garderies" and CFP). * There is no system for monitoring the cost of each program, the performance of the centers (e.g., AMB, CET) and there is no follow-up on the outcomes (i.e., insertion rates of the CFP, number of girls taught to read and write). * Some staff and infrastructure are not really used or are under-used (particularly in the case of CSE) 55and because of its small investment budget, EN's infrastructure is dilapidated (i.e., AMB, CSE, CET). * Staffing levels are considerably higher than needed and staff are not flexible and difficult to be re-trained (24% of EN staff are illiterate women): The EN's recurrent expenditures are high compared to its number of beneficiaries (about 80% of its budget is spent on staff salaries), and staffing levels have changed very little over the last 8 years, and yet the number of beneficiaries has dropped by about 90%-- mainly due to the discontinuation of food aid (i.e., before 1992, when food-aid distribution in the CSE existed, the clientele of EN reached about 750,000 beneficiaries, compared to 80,000 today). 4.47 Reform Options. There is a strong and demonstrated need for an efficient assistance system and given that there are no other public institutions in Morocco that have direct responsibility for helping the poor, a major restructuring of EN is necessary. To improve the efficiency of EN, since 1999, the Government has launched a two-year restructuring effort piloted in four provinces to introduce fundamental changes in service delivery. The objective of this initiative was to test the impact of the changes and examine restructuring alternatives. Several concrete measures have been introduced, including: (i) selection of the pilot provinces on the basis of well-defined criteria; (ii) assessment of the effectiveness of EN's Vocational Training Centers (CFPs) and transfer of eligible students to the vocational training offices (OFPPT's); and (iii) increasing partnerships with other agencies and local NGOs for implemention of the activities: in 1999, Foundation Mohamed V, has invested in the AMB to improve efficiency and quality of services. 4.48 However the results of this restructuring in pilot provinces are not yet available. Moreover, an extensive restructuring to reach a larger share of the poor and deliver fewer, more focused programs more efficiently has not yet been implemented. Although the budget allocation to EN is clearly too small for the State to carry out minimum social protection objectives, it is not advisable to increase EN resources before evaluating the results of the ongoing restructuring in pilot provinces. 4.49 The restructuring should enable EN to:" (i) better identify target groups and their needs; (ii) better focus its activities, make them more cost-effective, and increase their beneficiary coverage; and (iii) strengthen its institutional capacity to monitor the impact of its activities. To achieve these objectives, EN needs to: (i) become more accountable to the beneficiaries and strengthen its institutional capacity to monitor the impact of its programs; (ii) continue working in partnership with other agencies (public/private) and local NGOs in order to improve the efficiency of its services; (iii) introduce major reforms in management and operation, including decentralization of responsibilities, reduction of over-staff, redesigning management responsibilities, establishing performance targets and a monitoring system, reviewing partnership arrangements, re-training of staff at all levels, etc. 5s Capacity utilization rates are especially low at CSE (1.5%) and CET (31%) compared to AMB (71%) and CFP (80%). 56 See "Morocco: Social Protection Reform - Entraide Nationale Component"; World Bank 1998. 49 BIBLIOGRAPHICAL REFERENCES Ravallion Martin, Poverty Comparisons, Chur: Harvood Academic Publishers, 1994. Ravallion Martin and Datt Gaurav, "Growth and Redistribution Components of Changes in Poverty Measures: A Decomposition with Applications to Brazil and India in the 1980s", Journal of Development Economics, 38, 275-95, 1992. Ravallion Martin and Huppi Monica, "Measuring Changes in Poverty: A Methodological Case Study of Indonesia During an Adjustment Period", World Bank Economic Review, 5, 57-84, 1991. Moroccan Studies Direction de la Statistique, Analyse du profil et de la dynamique de la pauvrete: unfondement de l'attgnuation des denuements, forthcoming. Direction de la Statistique, Observatoire de Niveau de Vie, "Role des envois de fonds des travailleurs a l'etranger", internal document, February 2000. Direction de la Statistique, Observatoire de Niveau de Vie, "Approche pratique de la pauvrete", internal document, November 1999. World Bank Reports World Bank , "Kingdom of Morocco: Poverty , Adjustment and Growth"; Report No. 11918-MOR; January 1994. World Bank, "Kingdom of Morocco: Social Protection Reform - Entraide Nationale Component", informal report, 1998. World Bank, "Kingdom of Morocco: Sources of Growth", yellow cover, March 2000. World Bank, "Kingdom of Morocco: Health Financing Brief', informal report, January 2000. World Bank, "Consumer Food Subsidy Programs in MENA"; Report No. 19561-MNA, December 1999. World Bank, "Kingdom of Morocco: Audit Report BAJI - Promotion Nationale"; informal report, March 2000. World Bank, "World Development Report - Social Indicators of Development", 1999.