Ghana - Poverty past, present and future

Report No. 14504-GH Ghana Poverty Past, Present and Futu re June 29, 1995 Population andl Hunian Resources Division West Central Africa Department Africa Region Document of the World Bank  CURRENCY EQUIVALENTS Currency Unit Cedi US$1 = Cedi 1160.6 Cedi I = US$ 0.00086 (as of June 27, 1995) ABBREVIATIONS AND ACRONYMS ARI - Acute Respiratory Infections CPI - Consumer Price Index COCOBOD - Ghana Cocoa Board EPS - Extended Poverty Study ERP - Economic Recovery Program FAO - United Nations Food and Agriculture Organization GDP - Gross Domestic Product GLSS - Ghana Living Standards Survey GNP - Gross National Product GSS - Ghana Statistical Service JSS - Junior Secondary School MoH - Ministry of Health NGOs - Non-government Organizations PPA - Participatory Poverty Assessment PTA - Parent-Teacher Association SSS - Senior Secondary School FISCAL YEAR January 1 - December 31  TABLE OF CONTENTS FOREWORD EXECUTIVE SUMMARY . .................................................... i . INTRODUCTION .1 2. POVERTY PRESENT: A PROFILE OF POVERTY IN GHANA. 3 A QUANTITATIVE POVERTY ASSESSMENT, 1992. 3 Measuring Poverty .3 The 'upper'poverty line. 4 A 'lower'poverty line. 6 The poverty index. 6 Poverty in 1992 . 8 Demographic characteristics of the poor 9 Economic characteristics of the poor .13 A PARTICIPATORY POVERTY ASSESSMENT, 1994 .14 The nature of poverty 14 Results of the priority rankings .15 Strategies for moving out of poverty .17 Vulnerability, seasonality, and poverty 19 Comparison with the quantitative assessment 21 TARGETING PUBLIC SOCIAL SPENDING TO THE POOR. 22 Social expenditure incidence 22 Attitudes to social services provision in Ghana 29 Perceptions of healthcare services 30 Local perceptions of water and sanitation services 32 Perceptions of education services .32 3. POVERTY PAST: ECONOMIC RECOVERY AND THE POOR IN THE 1980S 36 Policy reforms and the economic recovery .36 Preparing the evidence .41 Has poverty fallen? .42 Corroborating evidence .47 Poverty past -- summary .48  4. POVERTY FUTURE: AN AGENDA FOR FURTHER POVERTY REDUCTION ............................................... 51 Continued economic growth is critical ................................... 52 Improve rural livelihoods and agricultural proJuctivity .......... 53 Deepen human capital .............................................. 56 Pursue reforms of the health sector ....................................... 57 Increase government investment in health .............................. 57 Re-orientate government health expenditures to rural communities ............................................... 58 Encourage the use of peripheral health facilities by the poor .. 58 Reveiw and investigate health charges in the field ........... ....... 59 Persevere with education reforms ......................................... 59 Strengthen direct targeted poverty interventions ........... ......... 60 Continue and strengthen poverty monitoring ......................... 60 Establish policy research priorities ......................................... 61 REFERENCES ...................................................... 63 MAP  TABLES Table 2.1: Estimated Daily Calorie Consumption of Individuals in Vicinity of Poverty Line, 1992 ............................................................. 5 Table 2.2: Headcount and Poverty Gap Indices by Area and Socioeconomic Group, Upper Poverty Line, 1992 ............................................................. 9 Table 2.3: Poverty Indices by Gender of Household Head, 1992 ................................... 12 Table 2.4: Unit Subsidies in the Education and Health Sectors, 1992 . ........................... 24 Table 2.5: Incidence of Public Spending on Education and Health by Quintile, 1992 ..... 25 Table 2.6: Household and Government Spending on Publicly Provided Education and Health, 1992 ........................ ... 28 Table 2.7: The Incidence of Public Sector Social Spending in Selected Countries .......... 29 Table 3.1 Main Macroeconomic Indicators, 1984-94 ......... 39 Table 3.2: Headcount and Poverty-gap Indices by Area, 1988-92 .................................. 44 Table 3.3: Headcount and Poverty-gap Indices by Socioeconomic Group, 1988-92 .................................................. 45 Table 3.4: Comparison of Headcount Indices With and Without Seasonal Adjustment, 1988-92 ................................................. 45 Table 3.5: Sources of Household Income by Area, 1988-92 .......................................... 47 Table 3.6: Selected Social Indicators, 1988 and 1993 ................................................. 48 FIGURES Figure 2.1: Reported Illness and Response by Quintile, 1992 11....................................... 1 Figure 2.2: Income Sources of Poor and Non-poor Households, 1992 ......................... 13 Figure 2.3: Distribution of Subsidies for Health and Education .................................... 26 Figure 2.4: Mean Education Subsidies by Level, Quintile and Gender, 1992 27 Figure 2.5: Mean Health Subsidies by Facility, Quintile and Gender, 1992 27 Figure 3.1: Macroeconomic Trends, 1984-94 .......................................3.......7.......... . 37 Figure 3.2: Production of Food Crops in Calories, 1980-94 .. .................................. 40 Figure 4.1: Government Spending on Health and Education 1987-92 ........................... 57 BOXES Box 2.1: Interpreting Poverty Indices .7 Box 2.2: Survival and Coping Strategies .18 Box 2.3: Expenditure Incidence Analysis .................................. ................. 23   FOREWORD After years of desperate economic decline, the Government of Ghana embarked on an Economic Recovery Program in 1983. Since then, there has been a recovery in output, and a marked improvement in economic performance. The World Bank report Ghana 2000 and Beyond: Setting the Stage for Accelerated Growth, took stock of these achievements and reviewed the prospects for improved economic performance in the future. It highlighted the need for accelerated growth in the fight against poverty, and at the same time emphasized the importance of improving Ghana's human resources as a basis for such growth. More recent household data have become available for policy analysis. This has given a fresh impetus to poverty work in Ghana, and provided the opportunity for the Extended Poverty Study. This report synthesizes the main findings of the Ghana Extended Poverty Study. It was prepared by Lionel Demery (PSP, Task Manager) and draws heavily on background papers prepared by Rend Bernier (consultant), Ellen Bortei-Dorku (consultant), Shiyan Chao (AF4PH), Harold Coulombe (consultant), Tony Dogbe (consultant), Christine Jones (PRDMG), David Korboe (consultant), Andrew McKay (consultant), Kalpana Mehra (PSP), Andy Norton (AFTHR), Uma Subramanian (ASTEN) and Ye Xiao (PRDMG). It also draws on work by the Ghana Statistical Service and the Ministry of Health. The Peer Reviewers were Martin Ravallion (PRDPH), Jack van Hoist Pellekaan (AFTHR) and Paul Glewwe (PRDPH). It has benefited from the encouragement, advice and comments of many, including (in Ghana) Kwaku Twum-baah, Peter Digby, Anarfi Asamoah-Baah, Ken Williams, and (at the World Bank) Jose Sokol, Ian Porter, Gotz Schreiber, Kazi Matin, Nicholas Bennett, Sudharshan Canagarajah, Asad Alamn, David Peters. The funding assistance of UNICEF, ODA (UK) and CIDA (Canada) is also gratefully acknowledged. The report has been prepared in close collaboration with the Government. Key components of the Extended Poverty Study were produced by the Ghana Statistical Service and the Ministry of Health. Government participation was also arranged through a Counterpart Team, led by Mr. K.B. Amissah-Arthur, the Deputy Minister of Finance, which also reviewed the Country Economic Memorandum, Ghana: Growth, Private Sector, and Poverty Reduction. Major reviews were held at the National Development Planning Commission. The encouragement and advice from these consultations is gratefully acknowledged.   EXECUTIVE SUMMARY 1. This report synthesizes the results of the Ghana Extended Poverty Study (EPS). This collaborative effort of the World Bank and the Government of Ghana takes as its starting point the conclusions of the report Ghana 2000 and Beyond: Setting the Stage for Accelerated Growth, which reviewed the prospects for accelerated broad-based growth in the country. The study concluded that given the right conditions, economic growth in Ghana would be effective in reducing poverty. Yet, to achieve accelerated growth, direct actions are needed by the government to improve the income earning opportunities of the poor. Ghana 2000 and Beyond emphasized the need for a well targeted and concerted effort by the Government to improve the access of poorer Ghanaians to the social and economic infrastructure. Without such an effort, economic growth is unlikely to accelerate or to benefit most of the population. Improvements in human welfare constitute both the ends and the means of an accelerated growth strategy. 2. Unlike most countries of the region, Ghana initiated its Economic Recovery Program (ERP) promptly-as early as 1983. This brought about a recovery in output and incomes from the desperately low levels of the early years of the decade; and this growth by all accounts has been broad-based, benefiting most Ghanaians. Although there remain problems not adequately addressed by the ERP, compared with most countries of the region, the results have been encouraging. First, government persistence with the reforms has paid off. Economic growth rates of around 5 percent per annum have been sustained for more than a decade. Export and import volumes have grown, despite the continued deterioration in the country's terms of trade, and the macroeconomic balances have improved. But is the past a reliable guide to the future? If Ghana can accelerate its growth, are the poor likely to benefit? The EPS provides an opportunity to answer these and other questions. It has established three principal objectives: * To identify the main characteristics of poverty in Ghana, and to document the views of the poor about their difficulties and potential solutions to them. In short, understanding poverty present. * To investigate whether recent economic growth, attributed in part to Ghana's ERP, has benefited the poor-understanding poverty past. * To assess whether current conditions are conducive to broad-based growth, emphasizing the access of the poor to human capital and infrastructure-anticipating poverty future. 3. Ghana 2000 and Beyond was constrained by available household data. Since then (and as part of the EPS initiative), more recent data have become available for policy analysis. This has given a fresh impetus to poverty work in Ghana. POVERTY PRESENT 4. The EPS has approached poverty as a multidimensional problem. In seeking to understand poverty present in Ghana, it has adopted two broad approaches. First, a quantitative approach, emphasizing the economic dimensions of poverty, utilizes the results of the 1992 round of the Ghana Living Standards Survey (GLSS) conducted by the Ghana Statistical Service. Second, a  participatory assessment of poverty was conducted in 1993-94, in which the views of the poor were obtained using qualitative research methods. The quantitative and qualitative dimensions of poverty 5. With no generally agreed poverty line in Ghana, much of what follows is concerned with the measurement of poverty. Two poverty lines are defined-as two-thirds and one-half of mean expenditures in 1988. Unlike previous studies, these were subjected to further review. Households in the vicinity of the upper poverty line were found to derive just over 2,100 calories from food expenditures, which provides support for this benchmark in the quantitative analysis. Just over 31 percent of the Ghanaian population reported per capita expenditure levels below the upper poverty line in 1992 (and just under 15 percent below the lower poverty line). Taking the upper poverty line, poverty is found to be greater in the Rural Savannah and Rural Forest areas, which account for 60 percent of total poverty in the country. Both the incidence and depth of poverty are found to be greater in Rural Savannah than any other area. Living standards are lower in general in the Rural Savannah, even for the non-poor. Welfare indicators such as school enrollments or availability of clean water supply, are worse for the non-poor in the Rural Savannah than they are for the poor in Accra. The participatory assessment also emphasized the fact that whole communities in the northern areas of Ghana were poor (which was not always the case elsewhere). 6. The incidence of poverty is lowest in the capital Accra (at 23 percent of its population). Compared with earlier profiles, these results point to a much smaller gap between urban and rural poverty. If household expenditures are expressed in adult equivalents (rather than in per capita terms) the urban-rural differences in poverty incidence all but disappear. For example, under one set of adult equivalent weights, poverty in Rural Coastal is about the same as in Accra. The quantitative poverty assessment therefore gives greater weight than previous studies to urban poverty. This is also bome out in the participatory assessment. 7. Gender is an important dimension of poverty, especially in northern Ghana. Women play significant roles in rural economic activities-in Rural Coastal and Rural Forest they are responsible for 40 percent of all household agricultural activities, and they completely dominate agricultural processing activities. The participatory assessment confirms the sharp distinction between the income-eaming roles of men and women in Ghana. Women therefore bear a disproportionate share of the burden of being poor-they are obliged to spent a great deal of time not only at working in family enterprises, but in the nurture and rearing of children, and in important household tasks, such as cooking, and fetching water and fire wood. 8. Many of the findings of the participatory assessment (PPA) were broadly consistent with the quantitative poverty analysis. Both agree that poverty is greatest in the rural north. While not surprising, this message rings out clearly from both approaches. The inadequacy of public services in the rural areas is a major conclusion drawn from the quantitative and qualitative assessments. The participatory assessment highlighted the concern of the poor for improved water supply. This priority (especially in the north of Ghana) is amply demonstrated in the findings of the GLSS. On average, members of rural households spend 37 minutes per day fetching water (in 1992). In Rural Savannah, they are obliged to devote 48 minutes each day to this activity. The GLSS data also confirm that this task is borne mainly by girls and women. Females in Rural Savannah spend on average 70 minutes per day collecting water. 11  9. The participatory assessment obtained information which complemented the quantitative data. It reviewed the attitudes of the poor towards their poverty, and their views about how best to resolve their problems. l The qualitative assessment contrasted the conception of poverty in the Rural North, where the problem was thought to affect whole communities and to threaten food security (in the 'hungry' season), with that in urban areas, where poverty is viewed more as an individual condition, and where well-being is defined in terms of access to stable employment rather than food security. 10. The views of the pcor about their priority needs also varied by area. Those in the rural north emphasized access to improved water supply, health care and transport infrastructure. There were more diverse views expressed in the rural south. In most, though not all communities, food security did not seem to be a preoccupation; most placed emphasis on the provision of health care and schooling. Many of the rural poor expressed concern over environmental changes-the declining water table, decreasing stream flows, soil infertility, etc. In urban communities far greater emphasis was placed on better employment opportunities and access to credit. Even in urban areas, concern was also expressed over the availability of potable water and improved sanitation. The views of the poor about effective strategies for moving out of poverty reflected the above priorities. The rural poor emphasized community involvement in an anti-poverty strategy, including improved transportation, water supply, and basic health and education services. The urban poor highlighted improved emplovment opportunities and better credit provision for small- scale enterprises. 11. Much of the quantitative data used in assessing poverty 'present' relates to 1992, and is therefore somewhat outdated. There is evidence that living standards have deteriorated since then, due in part to accelerating inflation. This is discussed below. The targeting of social spending 12. The GLSS also offered an opportunity to assess the extent to which government social spending (on health and education) reaches the poor. Using the expenditure incidence approach, it is possible to trace how spending in these sectors is distributed to the poor in Ghana. This involves combining information on the unit costs of education and health services (in terms of costs per enrollment or per health facility visit) with the use of such services by households-the former derived from public expenditure data and the latter obtained from the GLSS. Table I reports the results of the exercise. 13. On average, each Ghanaian gained Cedis 10,644 from government spending on education and Cedis 3,959 from spending on health in 1992. Education resources were more evenly distributed and better targeted to the poor. The bottom quintile gained 16 percent of education spending but only 12 percent from government spending on health. And the top quintile's share of education spending (21 percent of the total subsidy) contrasts with that of health (33 percent). On average, urban residents gain more than those in rural areas. An urban resident received on average Cedis 14,047 from public education spending and Cedis 5,808 from health spending, compared with just Cedis 9,218 and Cedis 3,309 respectively in rural areas. The relative disadvantage of rural residents was greater in health. There are reasons to believe that the data underestimate the share of subsidies going to urban areas. The inequality in the distribution of social spending is likely to be greater than indicated in Table 1. But, overall public-sector education and health spending are distributed more equally than income/expenditures. If iii  households were given income transfers in place of these subsidized services, income/expenditure distributions would become more equal. I Table 1: Incidence of Public Spending on Education and Health by Quintile, 1992 Education Health Per capita Share of Per capita Share of subsidy total subsidy total subsidy subsidy (cedis) (%) (cedis) (%) Quintile: 1 8731 16.4 2296 11.6 2 11021 20.7 3065 15.5 3 11196 21.0 3692 18.7 4 11207 21.1 4228 21.4 5 11067 20.8 6515 32.9 All Ghana: 10644 100 3959 100 Urban 14047 42.1 5808 48.7 Rural 9218 57.9 3039 51.3 Source: Demery et al (1995) 14. There is evidence of gender differences in the distribution of social spending in Ghana, especially on education. Taking the country as a whole, girls received 45 percent of primary-level subsidies and just 40 percent of secondary subsidies. This gender bias was more marked among poorer households, but was nevertheless in evidence among all quimtiles. In health, women gained more of the public subsidy than men, in part because of differing health needs. 15. The PPA investigated the attitudes of the poor towards public social services. The public medical system is widely regarded as expensive, in part because of the high opportunity costs of care, but also because of extra charges (for example in many outreach services). Many reported uncertainty on the part of service providers about the correct scale of charges in government hospitals. Even the non-poor in the north of the country found it difficult to raise the cash nee7ded for treatment. Schemes to allow exemption from user charges for the poor were not considered to be effective. 16. The cost factor was also prominent in the views expressed about public education. In addition to school expenses (books, school supplies, desks, school building maintenance, etc.), the poor were particularly concerned about the opportunity cost of schooling, in terms of the time lost to potentially income-earning activities. The poor quality of education was the single most important concern of the participants. With parents often responsible for school building maintenance and furnishings, poor communities are not able to maintain a satisfactory standard of education facility. Teacher quality and supervision were viewed by many respondents to be in urgent need of improvement. Participants consistently stated that their children could not read or write upon completion of primary schooling. This is confirmed in an analysis of the second round of the GLSS. iv  17. The quantitative and qualitative approaches to poverty assessments have complemented each other in the analysis of social services in Ghana. Incidence analysis quantified the distribution of government spending, on health and education, whereas the PPA provided valuable insights into the factors underlying the use (or non-use) of government services by the poor. Both point to a critical problem of access to health care services for the rural poor, and to the problem of quality in the provision of education services. POVERTY PAST 18. Living standards have fluctuated significantly in Ghana over the past two decades or so. GNP per capita fell sharply before 1983, and has since recovered. But Alderman (1994) estimates that even with a GNP rate of growth of 5.8 percent per annum, it would take until the end of the century for Ghanaian living standards to return to their 1965 level-and this after some ten years of growth already. The interest of this section lies not so much with this longer perspective as with the shorter-run changes during the 1980s and early 1990s-a period of economic adjustment and reform. 19. Ghana initiated an effective adjustment as early as 1983, and avoided many of the problems which come from unwarranted delays. This relatively agile response to the economic crisis meant that a recovery was sustained for much of the decade. The center-piece of the ERP was exchange rate policy. The real exchange rate depreciated sharply between 1984 and 1987. Inflows of external transfers led to a lull in this depreciation despite continued terms of trade deterioration between 1987 and 1991-significantly the period covered by the GLSS data. Trade reforms were also pursued vigorously. The result was improved macroeconomic balances, export growth, and a recovery in GDP growth. 20. However, three broad weaknesses characterize the ERP. First, the slow pace of some reforms, such as export-crop marketing and privatization. These have given mixed signals to the private sector. Second, since 1992, an election-induced fiscal shock has led to fiscal imbalances and to uneven economic growth because of slippages in economic performance. Third, there has been a disappointing private-sector response to the reforms, particularly in terms of investment. However, there is some uncertainty about the role played by agriculture in the recovery. The national accounts indicate slow growth in agriculture during the period covered by the GLSS data. However, these accounts are currently being revised. Additionally, government crop statistics (reported by the Ministry of Food and Agriculture) show a marked increase in agricultural output. 21. Did poor households benefit, and does the evidence suggest that poverty fell as a result of the ERP-induced recovery? To address this question this section relies mainly on evidence from GLSS data for 1988, 1989, and 1992. The period covered by these data follows the more dramatic recovery in output during the early years of the adjustment (1983-87). Taking the real exchange rate as the key proxy variable for the adjustment effort, it is clear that by the time the surveys take up the story, the major shift to sustainable macroeconomic management was largely achieved. GDP growth had subsided somewhat. The growth in real private consumption decelerated slightly after 1987. In this sense the GLSS data do not reflect the full impact of the ERP on household welfare. v  22. A number of adjustments were required to ensure that the GLSS data are comparable over time. In addition to the usual constructs needed to take account of price variations (both geographic and over tipne), adjustments were needed to correct for potentially higher recall errors in the early years of the GLSS. The suitably modified data indicate a significant fall in poverty-the headcount for the country as a whole falling from 37 percent in 1988 and 42 percent in 1989 to just 31 percent in 1992, with most of this poverty reduction occurring in rural areas. Poverty in Rural Coastal fell from 38 percent in 1988 to just 29 percent in 1992 (with similar orders of magnitude in other rural areas). Table 2: Poverty Headcount Indices by Area, 1988-92 (Percent) 1988 1989 1992 Accra 8.5 21.9 23.0 Other Urban 33.4 35.1 27.7 Rural Coastal 37.7 44.6 28.6 Rural Forest 38.1 41.9 33.0 Rural Savannah 49.4 54.8 38.3 5 Ghana 3:6::.i : JUt7 0 :9 41.8 31.4 Source: GSS (1995a) 23. But poverty has clearly persisted as a major problem, especially in the capital city. The increase in poverty incidence in Accra over just one year from only 9 percent in 1988 to 22 percent in 1989 is hardly credible. It is likely that the GLSS I survey under-estimated poverty in Accra, and that the period has witnessed only a gradual increase in poverty in the capital city (by just one percentage point between 1989 and 1992). Against a general setting of economic recovery and growth, this gradual decline in living standards in Accra is a source of major concern. The decline affected mainly middle income groups, especially wage earners. Real expenditures did not decline as much in households engaged in non-farm self employment and those which relied on remittance income-households which are more likely to be poor. The frequently-expressed view that poverty in Ghana has increased as a result of the ERP is probably a reflection of what has happened in Accra. First, the evidence strongly indicates a major improvement in economic welfare elsewhere in Ghana. Accra residents may well be rnistakenly generalizing from their own experience. Second, even within Accra, it is the better-off, who are invariably the most vocal, who have lost out in recent years. 24. Since 1992, there is evidence of a slow down in the rate of economic growth and of a deterioration in living standards. In part, this was due to deteriorating macroeconomic balances, brought about by fiscal expansion. Inflation has accelerated. These recent events may have reversed some of the gains achieved up to 1992. vi  POVERTY FUTURE 25. The encouraging messages about poverty past should not lead to any complacency about poverty in the future. First, poverty remains a serious problem in Ghana. Both the quantitative and qualitative assessments underscore the difficulties facing large sections of the population currently living in poverty. In some cases, food security and nutritional status are threatened, especially in the north of the country. Second, there are pockets of poverty which have been resistant to the economic recovery, especially in urban areas. Third, the past is not necessarily a reliable guide to the future. It remains to be seen whether future growth will be as broad-based in its effects as the past recovery. 26. Mindful of this, and in part a response to the findings of the EPS, the Government has recently taken action to develop a meaningful and effective poverty-reducing strategy. It has set up a poverty coordinating committee, which will set the policy agenda on poverty for the coming years. Among other objectives, it will review the EPS, and other similar work, establishing what the findings mean for Government policy. It will also set priorities for further work on poverty in Ghana. While not seeking to anticipate the deliberations of the committee, this section reviews the key messages emerging from the first phase of the Extended Poverty Study for government policy in Ghana. 27. Accelerated economic growth is essential: With annual GDP per capita growth of 3 percent, it will take 12 years to reduce the incidence of poverty to 10 percent. If growth can be accelerated to 5 percent, it will take only 7 years. The recent deceleration in growth (primarily caused by increasing macroeconomic imbalances) is therefore a cause for concem. 28. Improve rural livelihoods and agricultural productivity: Growth will only benefit the poor if it is sufficiently broad based. Given the continuing importance of agriculture for the poor, this means that agricultural productivity growth must be raised. The government must avoid over- taxing agriculture, either indirectly through real exchange rate overvaluation, or directly through cocoa pricing policies. But the problems facing agriculture cannot be resolved simply through changes in the incentive structure, important though they be. There is a critical need for a major improvement in rural infrastructure-improved water supply, better roads and communications, and more effective research and extension activities. This is a long-term policy agenda, but a central one if growth is to benefit the poor. 29. Deepen human capital: Investing in human resources confers direct benefits to the poor in raising the quality of life (better health and education, for example). It also enhances long-run economic growth. Ghana still has some way to go to achieve the depth of human capital which proved so important to the East Asian countries in accelerating economic growth. Although there have been significant improvements in the provision of health, water, and education services since the early 1980s, much remains to be done. Government social expenditures remain largely untargeted to the poor, especially on health. First, policy reforms in the health sector must be pursued. Ghana does not spend enough on the public provision of health, and existing spending is too focused on urban-based services. Recent reforms on financing rural health services must be monitored carefully. Changes in health user charges should also be considered, encouraging self- selection in service use, and diverting the better-off away from, and the poor towards primary-level vii  services. The government should also persevere with its education sector reforms. Here the main need is to enhance the quality of schooling. The government must further irnprove the management of the education secto4, and raise standards of teaching, mainly through better teacher discipline. It must also increase its allocations to non-salary costs, providing more blackboards, books, stationery, desks and chairs, and better building maintenance. 30. Strengthen direct targeted poverty interventions: The persistence of poverty in some parts of the rural sector and in urban areas, especially Accra, points to the need for other targeted interventions to assist the poor to participate in the growth process. A case can be made for establishing a social fund which would target the needs of poorer groups, including those in the urban informal sector. Although the nature of the support needed is not yet clearly established, the focus would likely be on employment creation and the delivery of basic social services, with an emphasis on strengthening NGOs, local governments, communities and small-scale private initiatives. 31. From the EPS results, and especially the PPA findings, a social fund is likely to be more effective in improving the welfare of the poorest households and communities if its resources can be effectively targeted to the most needy. This means that its resources should be targeted geographically. For the poorest communities in the north, an essential objective would be the creation of employment and income-earning opportunities during the lean season. If the activities so funded improve the rural infrastructure, the social fund would have a double edged effect on poverty-on the one hand raising current incomes and improving food security today, and on the other, creating an enabling environment for future growth. 32. Continue and strengthen poverty monitoring: The GSS has proven itself able to field and process a survey as complex as the GLSS. Yet, it has taken some time to process and clean the data, so that the GLSS3 report was published some three years after the fieldwork was completed. There is therefore a clear need to further strengthen poverty monitoring in Ghana, and to enhance the capacity of the GSS to undertake this responsibility. Further consideration needs to be given to obtaining a consensus on the nature of poverty and on the poverty line itself. This involves building on both the quantitative and qualitative work of the first phase of the EPS. 33. Establish policy research priorities: The work summarized in this report is only a beginning. It is a first phase of a program of work that will continue. The recent establishment of a government poverty coordinating committee provides an opportunity for the government to establish its own policy priorities and their implied research agenda. Topics receiving attention in a second phase of work include labor markets, migration and poverty, agriculture, the environment and the poor, gender aspects of poverty, and the access of the poor to social services. viii  I 1. INTRODUCTION 1.1 This report synthesizes the results of the first phase of work of the Ghana Extended Poverty Study (EPS). This collaborative effort of the World Bank and the Government of Ghana takes as its starting point the conclusions of the study Ghana 2000 and Beyond: Setting the Stage for Accelerated Growth, which sets out the prospects for accelerated broad-based growth in the country. This study concluded that given the right conditions, economic growth in Ghana would be effective in reducing poverty. Yet, to achieve accelerated growth, direct actions are needed by the government to improve the income earning opportunities of the poor. Ghana 2000 and Beyond emphasized the need for a well targeted and concerted effort by the government to improve the access of poorer Ghanaians to the social and economic infrastructure of the country. Without such an effort, economic growth is unlikely to accelerate or to benefit the majority of the population. Improvements in human welfare constitute therefore both the ends and the means of an accelerated growth strategy. 1.2 Unlike most countries of the region, Ghana's Economic Recovery Program (ERP) was initiated without undue delay-as early as 1983. This brought about a recovery in output and incomes from the desperately low levels of the early years of the decade. And this growth by all accounts has been broad based, benefiting most Ghanaians. In these respects Ghana's recent experience has been exceptional in the regional context. Although there remain problems not adequately addressed by the ERP, compared with most countries of the region, the results have been encouraging. First, the government has persisted with the reforms. And this has paid off. Economic growth rates of around 5 percent per annum have been sustained for more than a decade. Export and import volumes have grown, despite the continued deterioration in the country's terms of trade, and macroeconomic balances have improved. But is the past a reliable guide to the future? If Ghana can achieve accelerated growth, are the poor likely to benefit as much from such growth? The first phase of work of the EPS provides an opportunity to answer these and other questions. It has set itself three main objectives: * To identify the main characteristics of poverty in Ghana, and to document the views of the poor about their difficulties and potential solutions to them. In short, understandingpovertypresent. * To investigate whether recent economic growth, attributed in part to Ghana's ERP, has benefited the poor-understanding poverty past. * To assess whether current conditions are conducive to broad-based growth, emphasizing the access of the poor to human capital and infrastructure- anticipating povertyfuture.  2 1.3 Ghana 2000 and Beyond was constrained by available household data, lacking the third year (1991/2) of the Ghana Living Standards Survey (GLSS). Since then (and as part of the EPS initiative) the GLSS3 data have been processed and have become available for policy analysis. This has given a fresh impetus to poverty work in Ghana. The EPS treats poverty as a multidimensional problem. And it has therefore approached it from several angles. Its agenda can be divided into two phases. The first comprises four major pieces of work: a poverty profile of Ghana; a Participatory Poverty Assessment (PPA); analyses of poverty change in Ghana during 1988-92; and an assessment of the incidence of public social spending. These first-phase studies (listed in the bibliography) are reviewed in this report. Combined, these contributions have improved our understanding of poverty-present, past, and future. The second phase seeks to develop a poverty- reduction strategy in Ghana, and to advance understanding of a selected issues, both initiatives emerging from the first-phase work. 1.4 The following chapter deals with poverty present. It combines the results of the poverty profile (based on the GLSS data-see Ghana Statistical Service, 1995) and the PPA (Norton, et al, 1995, Korboe, 1995). It also reviews how government spending on the social sectors benefits the poor (Demery et al, 1995). This is followed by an analysis of changes in poverty over the recent past. This chapter assesses whether the economic recovery induced by the ERP resulted in a sustained increase in the standard of living of households in general, and poor households in particular. Finally, chapter 4 addresses the question of poverty future. The focus is on whether conditions are currently in place for the benefits of growth to be shared broadly among the Ghanaian population. The chapter identifies the critical policy implications of the EPS to ensure that future economic growth will be effective in reducing poverty.  2. POVERTY PRESENT: A PROFILE OF POVERTY IN GHANA 2.1 This chapter reviews what is presently understood of poverty in the Ghanaian context. With no generally agreed poverty line in Ghana, much of what follows is concerned with the measurement of poverty. But the section also deals with the characteristics of the poor, and how they view their own circumstances. 2.2 Poverty is a complex and multidimensional phenomenon. As a consequence, there are several approaches which can be taken in its analysis, each emphasizing different aspects. In seeking to understand poverty present in Ghana, the EPS has adopted two broad approaches. First, a quantitative approach-emphasizing the economic dimensions of poverty-utilizes the results of the GLSS conducted by the Ghana Statistical Service (GSS). These household surveys were structured and are nationally representative. The results of this analysis are summarized in the next section This is followed by a review of the results from the PPA. The final section of the chapter provides an analysis of the targeting of social spending to the poor in 1992, together with a discussion (drawn from the PPA) of the attitudes of the poor to public services. A QUANTITATIVE POVERTY ASSESSMENT, 1992 2.3 Our concern in this section is to provide a profile of poverty in Ghana using the GLSS survey conducted in 1991/92. l The section draws from three EPS papers-GSS (1995a), Coulombe and McKay (1995) and Jones and Xiao (1995), and brings up to date the previous poverty profiles prepared for Ghana (see Boateng et al, 1990, and Glewwe and Twum-Baah, 1990). Measuring Poverty 2.4 To identify which households and individuals in Ghana may be considered as poor a measure of the economic welfare or standard of living must be selected. The preferred measure in the EPS is real household total expenditure. There are strong theoretical reasons for this choice (Deaton and Muellbauer, 1980). And measures of household expenditure are generally more accurate and reliable than household income. For the purposes of the poverty profile, household expenditures are expressed in per capita terms and measured in constant Accra May 1992 prices. Geographical cost-of-living differences are therefore accounted for (see GSS, 1995a for details). 1 The GLSS surveys have been fielded in three years. The first (GLSS1) covered the period September 1987-August 1988, the second (GLSS2), October 1988-September 1989, and the third (GLSS3), October 1991-September 1992. For simplicity, these are termed respectively 1988, 1989 and 1992 throughout this report.  4 2.5 The next step is to agree upon a poverty line. This has to be expressed in terms of the welfare measure selected. Deciding on a poverty line is invariably a matter of judgment-an art as much as a science. It is important that the benchmark be generally accepted as a standard for policy analysis and prescription. Unfortunately, in Ghana there is as yet no consensus about a poverty line, and there was no 'ready-made' formula available to the EPS. Previous work (such as Boateng et al, 1990) had prescribed an explicitly arbitrary line, without any attempt to relate it to an absolute standard. Essentially, this involved determining poverty as a ratio of mean expenditures in the base year, placing emphasis not so much on the absolute numbers of people below such a line as on the patterns of poverty that are thus generated. 2.6 Given the continuing absence of any generally accepted standard, the EPS has adopted a similar approach, but with some additional analysis. The starting point was to define poverty in terms of mean expenditures in the base year-1988. This line was then fixed in real terms for meaningful, over time comparisons. Details of the specific procedures used to compute the poverty lines in the EPS work can be found in Ghana Statistical Service (1995) and Coulombe and McKay (1995a). Two such poverty lines were selected, an upper line of two-thirds of 1988 mean per capita household expenditures and a lower line fixed at one-half of mean expenditures. 2.7 Unlike previous studies, however, the EPS subjected the selected benchmarks to further analysis. This gives more meaning to the selected poverty lines, and confirms their appropriateness for policy analysis. It is hoped that this work will take the discussion of poverty further in Ghana itself, and serve to encourage a growing consensus about the poverty standard in the country. The 'upper' poverty line 2.8 After making a number of important adjustments to the expenditure data (discussed below),2 a poverty line of two-thirds of mean household total expenditure per capita in 1988 (and expressed in constant May 1992 Accra prices) is given as Cedis 132,230 per person per year. Is this a reasonable basis for poverty analysis? What level of deprivation is implied for individuals unable to achieve such expenditure levels? What levels of caloric consumption are permitted at expenditure levels within the proximity of the higher poverty line? To make this assessment, households were selected whose per capita expenditures (in 1992-i.e., using the GLSS3 data) were in the range of + 5 percent of the poverty line, and their food consumption patterns analyzed. Expenditures on 49 individual food items were translated into physical quantities (generally kilograms of the food item) using GLSS3 price data. Caloric conversion factors (based on Platt, 1962) were then applied to these quantities to obtain estimates of the calories derived from such consumption. The results are reported in Table 2.1. 2 The most important adjustment was to correct for recall error in the expenditure data for 1988 and 1989. Since this affected mean expenditures in 1988, the adjustment also influenced the poverty line.  5 Table 2.1: Estimated Daily Calorie Consumption of Individuals in Vicinity of Poverty Line, 1992' Food Item Quantity consumed Share in food Estimated daily (Kg. per person per year) expenditure calorie intake (%/) (per person) Maize 17.3 2.6 172 Maize flour/products 16.0 3.2 159 Rice 8.4 3.3 81 Other cereals/products 8.0 3.8 67 Cassava 160.3 10.3 672 Cassava products 13.9 2.9 130 Plantain 34.5 5.5 121 Yam 33.5 5.3 96 Cocoyam 25.0 4.0 77 Fish 12.0 13.5 87 Meat 3.0 4.2 22 Poultry 1.4 1.8 6 Milk/eggs 6.8 0.8 29 Fats/oils 5.8 2.0 142 Beans/nuts 19.2 4.2 126 Fruits/vegetables 26.7 9.7 23 Sugar 2.2 1.1 25 Other foods/beverages 12.8 3.3 106 Total - 81.5 2,141 a/ The sample for this computation comprised households i 5 percent of the poverty line of Cedis 132,230.28. There were 232 households and 1370 individuals selected for the exercise. Source: GLSS3 2.9 Given the price and caloric conversion data available, it was possible to cover just over 81 percent of total food expenditure of the household selected. From such consumption, households in the proximity of the poverty line gained over 2,100 calories per person per day.3 Although high by African standards, this consumption of calories provides adequate nutrition to satisfy FAO standards. These findings therefore can be taken as giving some support the two-thirds of mean formula for the poverty line. 3 It is uncertain what the effect of the remaining expenditure items not covered would be on total caloric intake.  6 A 'lower'poverty fize 2.10 Other approaches to assessing the meaningfulness of the two-thirds formula were reported by GSS (1995a). First, the line was compared with the World Bank's poverty line of PPP$1 per person per day, which translates into Cedis 107,188 per person per annum (in Accra May 1992 prices). This is just 77 percent of the poverty line as given by the two-thirds formula. GSS (1995a) also compared the two-thirds formula with the general minimum wage in Ghana (set at Cedis 800 per day in 1994).4 Assuming an individual works on average 5.5 days per week, and that an average of 49 percent of household members are economically active, this suggests a poverty line of around Cedis 82,500 (again in May 1992 Accra prices)-or just 60 percent of the two-thirds formula. 2.11 Although these comparisons are somewhat crude, and the methods used to translate them into an equivalent poverty line in May 1992 Accra prices are only approximations (see GSS, 1995a), they do suggest that there would be advantage in using an alternative, lower poverty line in the EPS. 5 Such a line was fixed at one-halfof mean household expenditures per capita in 1988, or Cedis 99,173 per person per annum (in Accra May 1992 prices). This more closely corresponds to the World Bank standard, and to the official minimum wage benchmark. 2.12 Repeating the same analysis of the caloric intake in the vicinity for the lower poverty line indicates that members of households (whose total expenditures are