Ghana Growth, Private Sector, and Poverty Reduction A Country Economic Memorandum

Transcription

1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Report No GH Ghana Growth, Private Sector, and Poverty Reduction A Country Economic Memorandum May 15, 1995 Country Operations Division West Central Africa Department U *Document of te World-Bank.

2 CURRENCY EQUIVALENTS Currency Unit Cedi US$1 = Cedi 1130 Cedi I - US$ (as of end-april, 1995) ABBREVIATIONS AND ACRONYMS AGE - Ashanti Goldfields Corporation BOG - Bank of Ghana CEM - Country Economic Memorandum CMC - Cocoa Marketing Company COCOBOD - Ghana Cocoa Board CSPIP - Civil Service Performance Improvement Progran EBRD - European Bank for Reconstruction and Development ECG - Electricity Corporation of Ghana E&D - Exploration and Development ERP - Economic Recovery Program ESAF - Enhanced Structural Adjustment Facility GDP - Gross Domestic Product GNP - Gross National Product GIPC - Ghana Investment Promotion Council GLSS - Ghana Living Standards Survey GNPC - Ghana National Petroleum Corporation GOG - Government of Ghana GOIL - Ghana Oil Corporation ICOR - Incremental Capital-Output Ratio IMF - International Monetary Fund IPP - Independent Power Producer LBC - Local Buying Company MDA - Ministries and Departmental Agencies MIGA - Multilateral Investment Guarantee Agency NBFI - Non-Bank Financial Institution NIRP - National Institutional Renewal Program O&M - Operations and Maintenance PBC - Produce Buying Company PFMRP - Public Financial Management Reformn Program SOE - State-Owned Enterprise SSNIT - Social Security and National Insurance Trust TOR - Tema Oil Refinery VAT - Value-Added Tax VRA - Volta River Authority FISCAL YEAR January I - December 31

3 TABLE OF CONTENTS FO R EW O R D... i EXECUTIVE SUMMARY... ii CHAPTER 1 RECENT ECONOMIC DEVELOPMENTS AND IMPLICATIONS... 1 Recent Policy Reforms... 1 Recent Fiscal Performance... 3 The Fiscal Turnaround of 1993 and Recent Economic Performance... 7 Conclusions CHAPTER 2 POLICY REFORMS AND THE PATTERN OF GROWTH.14 Policy Reforms and Performance.14 Estimating Private Sector Value-Added.19 Private Sector's Share of Output.21 Is Growth Public Sector-Led?.23 Conclusions.24 CHAPTER 3 CHAPTER 4 CHAPTER 5 GROWTH AND POVERTY REDUCTION.25 The Definition of Poverty.25 Aggregate Poverty.26 Urban-Rural Poverty.27 Gender and Poverty.29 What Explains the Reduction in Poverty?.30 Looking to the Future.32 Conclusions.33 PUBLIC SOCIAL EXPENDITURES AND POVERTY.35 The Level of Human Capital.35 The Incidence of Public Social Spending.37 Conclusions.46 PROMOTING PRIVATE SECTOR GROWTH.48 Sustaining Macroeconomic Stability.48 Investing in Infrastructure.50 Strengthening the Financial Sector.57 Liberalizing the Petroleum Sector.59 Removing the Distortions in Cocoa Trading.61 CHAPTER 6 MEDIUM-TERM PROSPECTS.66 Macroeconomic Prospects.66 External Financing:

4 External Debt Conclusions CHAPTER 7 CONCLUSIONS Agenda for Further Work APPENDIX TABLES MAP LIST OF TEXT TABLES Table 1.1 Selected Economic Indicators. 4 Table 1.2 Financing of the Budget. 9 Table 2.1 Selected Policy and Performance Indicators.16 Table 2.2 Non-Traditional Exports by Product, Table 2.3 Private and Public Sector Contribution to GDP in Table 2.4 Private Sector Share of Value-Added, Table 3.1 Sources of Household Income, by Poor and Non-Poor.27 Table 3.2 Poverty Profile, by Urban and Rural Areas.27 Table 3.3 Sources of Household Income, by Urban and Rural Areas.28 Table 3.4 Poverty Profile, by Rural Economic Group.28 Table 3.5 Sources of Household Income, by Rural Economic Group.29 Table 3.6 Poverty Profile, by Gender of Economic Head of Household.29 Table 3.7 Sources of Household Income, by Gender of Economic Head of Household.30 Table 3.8 Sectoral Pattem of Rural Self-Employment, by Individual.32 Table 4.1 Human Capital Indicators.36 Table 4.2 Distribution of Public Recurrent Expenditures on Health, by Urban- Rural Areas.39 Table 4.3 Distribution of Public Recurrent Expenditures on Health, by Gender 40 Table 4.4 Distribution of Mean Public Expenditures on Education, by Urban- Rural Areas.43 Table 4.5 Distribution of Public Expenditures on Education, by Gender.44 Table 5.1 Ownership Options for Infrastructure Provision: Ghana and Best 51 Practice Compared. Table 5.2 Barriers to Private Sector Investment in infrastructure in Ghana 52 Table 5.3 World Cocoa Production and Market Share.62 Table 5.4 Distribution of Cocoa Proceeds.64 Table 6.1 Medium-Term Macroeconomnic Prospects, Table 6.2 Ghana: Aid Disbursements, Table 6.3 Extemal Financing Requirements and Sources, Table 6.4 Extenal Debt Service,

5 LIST OF TEXT FIGURES Figure 1.1 Investment in Ghana, Figure 1.2 Savings in Ghana, Figure 1.3 Money Supply and Prices, Figure 1.4 Nominal and Real Interest Rates, Figure 1.5 Budgetary and Current Account Deficits, Figure 2.1 World Cocoa Price, Terms of Trade, and Real Exchange Rate 15 Indices, Figure 2.2 Cocoa Terms of Trade and Cocoa Output, Figure 2.3 Sectoral Growth Rates Figure 2.4 Sectoral Shares of GDP Figure 4.1 The Incidence of Public Social Spending , Figure 4.2 Distribution of Public Recurrent Spending on Health by Expenditure Quintile, Figure 4.3 Distribution of Public Spending on Education by Expenditure Quintile, LIST OF TEXT BOXES Box 3.1 Box 3.2 Box 4.1 Box 4.2 Box 4.3 Box 6.1 Decomposition of Poverty Changes: Some International Comparisons How long for Poverty Eradication for the Average Poor? What is 'Expenditure Incidence' Analysis? Cost Recovery in Health: Economic Theory and the Lessons of International Experience.42 Cost Recovery in Education: Economic Theory and the Lessons of International Experience Ghana--Vision 2020: An Overview LIST OF APPENDIX TABLES Table Al National Accounts -- Current Price Data, Table A2 National Accounts -- Shares of GDP, Table A3 National Accounts -- Constant Price Data, Table A4 National Accounts -- Annual Growth Rates, Table A5 Merchandise Exports and Imports, Table A6 Balance of Payments, Table A7 Public Finance, Table A8 Key External Debt Indicators, Table A9 Major Prices,

6

7 FOREWORD After a decade of stabilization and adjustment in Ghana -- intended to build an incentive structure that is market-friendly and leads to sustainable growth with poverty reduction -- poverty continues to be a pressing concern. There are also concerns in Ghana about the potential for adverse effects on poverty of a private sector-led growth strategy. It is, therefore, important to develop the linkage between growth and poverty reduction. Given the economic developments of 1992 and 1993, especially the drop in private investment, this is an appropriate time to examine this relationship, identify the macro- and micro-level constraints to private sector growth, and suggest the policy changes that are required to stimulate it. This Country Economic Memorandum (CEM) on Ghana is, therefore, thematic, focusing on the theme of "Growth, Private Sector, and Poverty Reduction." However, the report does not undertake an investigation of the reasons behind differences in sectoral growth rates, nor of the constraints on agriculture, which will be the subject of future economic and sectoral work. The poverty results in the report draw upon the ongoing work of the Extended Poverty Study, currently being prepared by the World Bank. They are based on analyses of the Ghana Living Standards Surveys of , , and Since the Extended Poverty Study is still in progress, the results presented here are subject to further refinements. The CEM was prepared by a team consisting of Asad Alam (Task Manager), Hans-Martin Boehmer, Sibabrata Das (AF4CO), Paul Glewwe (PRDPH), Christine W. Jones, Xiao Ye (PRDMG), Philip Keefer (PRDFP), Sudharshan Canagarajah, Saji Thomas (AF4PH), Liaqat Ali and Anthony Tsekpo (Consultants), under the overall supervision of Kazi M. Matin (AF4CO). The Lead Adviser was Ishrat Husain (Chief Economist, EAPVP) and the Peer Reviewers were Lionel Demery (ESP) and John Nellis (PSD). The report also benefited from advice and comments from Bension Varon (Economic Adviser, AFRCE), Jose Sokol (Lead Economist, AF4DR), and Ngozi Okonjo-Iweala (Chief, AF4CO). The report was prepared with substantial participation from the Government at key stages -- the conceptualization of the thematic focus, the drafting of the issues paper, the discussion of key background papers, and the more recent green cover discussions. Government participation was through a Counterpart Team led by Mr. K.B. Amissah-Arthur, the Deputy Minister of Finance. The final report was also discussed with Dr. K. Botchwey, the Minister of Finance. The report also benefited from a workshop on key background papers with the National Development Planning Commission. The support and cooperation of the Ghana Statistical Service in providing the Ghana Living Standards Survey data and the latest National Accounts statistics is gratefully acknowledged. The comments of the donor community in Accra on key background papers and on the Executive Summary are also gratefully acknowledged.

8 GROWTH, PRIVATE SECTOR, AND POVERTY REDUCTION EXECUTIVE SUMMARY 1. There is now clear evidence that policies that promoted growth in Ghana during the 1980s also helped to reduce poverty-and substantially so. Between and , Ghana's poor fell from 37 percent to 32 percent of the population. This reduction was widespread and broadbased, benefiting mainly rural areas and the most vulnerable group-female-headed households. Almost all of the reduction can be traced to economic growth in general and growth in non-farm self-employment in particular. Growth in private-sector services, especially wholesale and retail trade, has been the driving force behind non-farm self-employment. Moreover, growth has been distributionally-neutral between and No country has achieved sustained poverty reduction without economic growth. The private sector has been the engine of rapid growth in fast-growing countries. Growth reduces poverty by providing economic opportunities for the poor to utilize their labor and generate higher personal incomes; at the same time, growth forms the basis for larger tax revenues from which to finance the public expenditures on the social sectors to enable the poor to take advantage of the economic opportunities. In particular, growth that exploits the comparative advantage of a country, using its more abundant factor of production (i.e., labor), enables the income gains generated by it to be pro-poor. A growth strategy that promotes macroeconomic stability and a liberal and neutral incentive regime fits the bill. Ghana has been pursuing such a strategy since its Economic Recovery Program (ERP) was launched in Ideally, an analysis of Ghana's growth and poverty reduction should start with the level and pattem of poverty in 1983, when the Government began a decade of stabilization and adjustment to speed growth and reduce poverty. Since such data were not available, the starting point becomes the Ghana Living Standards Surveys of and These indeed show that Ghana's economic strategy has benefited the poor. True, these comparisons cover a much later stage in the recovery program. But that means that the actual reduction in poverty since the 1983 policy reforms has been greater. 4. What were the policy changes that spurred growth? And how did the pattern of growth help the poor? The Policy Reforms and Growth 5. The policy reforms since 1983 have reduced the fiscal deficit and inflation, helped improve infrastructure services, and shifted relative prices and incentives towards the tradables sector, in general, and towards exports, in particular. The key element of fiscal consolidation up to 1991 has been the growth in government revenues, whose share of GDP rose from 6 percent in 1983 to 13 percent in 1986 and to 16 percent in Higher revenues made it possible to reduce the fiscal deficit and, at the same time, increase public investment in infrastructure which had virtually collapsed prior to Prudent monetary management also led to inflation falling from 123 percent in 1983 to 40 percent in 1986 and 18 percent in The resulting improvements in ii

9 macroeconomic stability made it possible for farms and firms to respond to the shift in production incentives induced by the policy reforms. 6. As a result of these reforms, the economy turned around. Although economic activity witnessed its biggest surge during the early years of the ERP (5.3 percent annually during ), aggregate growth has averaged 4.7 percent per annum since The private sector has made a significant contribution to growth. However, this growth performance has not been uniform across sectors. Agriculture recorded an annual growth rate of only 1.9 percent since 1987 while services have grown at an average annual rate of 7.4 percent over the same period. Merchandise exports and imports have grown faster than GDP and with it, complementary wholesale and retail trade activity which have dominated GDP growth under the ERP. The share of external trade in GDP increased from about 5 percent in 1983 to 32 percent in 1986, 35 percent in 1991, and 55 percent in During the initial years , merchandise imports grew at a rate lower than that of merchandise exports, but thereafter, merchandise imports grew faster due to the growth in non-oil imports. The export that showed the strongest growth was gold, benefiting no doubt from the liberalization of the mining sector, which has now become the largest export earner overtaking cocoa. Non-traditional exports also grew faster since 1986 reflecting not only the growth in export demand but also favorable policy changes at home. 7. The differential pattern of sectoral performance has significantly altered the relative importance of the sectors in the Ghanaian economy. While in , agriculture accounted for almost 50 percent of GDP, its share fell to just over 40 percent in At the same time, the share of services increased from about 38 percent to about 45 percent of GDP, making it the largest sector in the economy. The share of industry barely changed-from 13 to 14 percent. Poverty Reduction 8. Poverty is predominantly a rural phenomenon in Ghana. So, too, is poverty reduction. Rural areas are home to two-thirds of the population but account for 75 percent of all poverty. While aggregate poverty declined from 37 in to 32 percent in , rural poverty fell from 42 percent to 34 percent (see Table 1). In fact, the rural poverty reduction accounts for most of the fall in aggregate poverty. The overall depth of poverty also declined. The mean expenditures of the poor were 12 percent below the poverty line in but by , it had improved to 8 percent. Table 1: Poverty Profile, by Urban and Rural Areas Area Share of Population Incidence of Poverty Depth of Poverty Contribution to (Po) (Pi) National Poverty (%) Urban Accra OtherUrban Rural ALL Source: GLSSI & GLSS3; Bank Staff calculations..

10 9. The poverty differential between rural and urban areas appears to have narrowed from 15 percentage points in to 7 in Urban poverty fell only marginally during the period, with improvements in other cities and towns offsetting a dramatic rise in Accra-from 9 percent to 23 percent. While further analysis is required to explain the rise in poverty in Accra, poverty there remains well below the national average. But the large fall in rural poverty strongly suggests that economic growth has effectively 'pulled up' the rural poor who have gained proportionately more. 10. Moreover, there was a dramatic reduction in poverty among households headed by women-from 39 percent to 29 percent. At the same time, poverty among male-headed households fell from 36 percent to 33 percent. This difference has two implications. First, most of the aggregate reduction comes from the fall in poverty of households headed by women. Second, the differential in the incidence of poverty between the sexes has been reversed. Explaining the Poverty Reduction 11. The reduction in poverty is due largely to economic growth. Ghana's growth in averaged almost 5 percent per annum, which translated into 2 percent growth in per capita GDP. The impact on poverty reduction, of course, depends on the nature and pattern of the growth. In Ghana, economic growth accounts for most of the poverty reduction. The poorest groups also benefited from growth. This is significant. In many other countries, the reduction in poverty is typically accompanied by a worsening in income distribution. 12. For the poor and the non-poor alike, farm income is the dominant but declining source of income. Non-farm self-employment is the second most important (and growing) source, and wage employment the third. The decline in rural poverty appears largely due to growth in non-farm selfemployment. Non-farm self-employment cuts across all real sectors but is essentially service-based (see Table 2). In , close to 54 percent of rural non-farm self-employment came from services; by , that share had risen to 62 percent. The share of trading activity in non-farm self-employment rose from 50 percent in to 55 percent in At the same time, the share of agriculture fell significantly. Growth in domestic trading appears to be the driving force behind growth in incomes and expenditures for rural groups. This growth is no doubt due to the surge in both imports and exports during the period of the ERP. 13. Despite the gains that Ghana has made, poverty remains a serious and extensive problem. For 32 percent of the population, or about 5 million people, expenditures per capita in 1992 were less than US$25 a month. However, poverty reduction can be expedited through proper policies. This suggests three things. First, since strong economic growth, that is at least distributionally neutral, has been shown to be a fundamental requirement for effective poverty reduction in Ghana (as elsewhere), the pursuit of proper macroeconomic policies to promote private sector growth is a must. Second, since poverty is largely a rural phenomenon, agriculture sector policies are crucial in any strategy of poverty reduction. And third, as is argued below, investments in human capital and better targeting of public social expenditures are required if growth is to be sustained and poverty is to be reduced. iv

11 Table 2: Sectoral Pattern of Rural Self-Employment, by Individual Sector Farm Self-Employment Non-Farm Self-Employment Pattern of Rural Non-Farm Self-Employment Non-Farm Agriculture Industry Mining Manufacturing Electricity & Construction Services Wholesale & Retail Trade Transportation & Communication Community, Social, & Other Services Source: GLSSI & GLSS3; Bank staff calculations a/ Covers forestry, logging, and fishing. The sumn of Farm and Non-Farm Agriculture comprises the agriculture sector as defined in the national income accounts. Investing in People 14. Even with the best economic policies, rapid growth with poverty reduction will elude Ghana unless more is invested (and more effectively) in health and education. Though spending on the social sectors has increased significantly from the low levels of 1983, current levels remain low relative to comparable countries. In the five years to 1994, public expenditures on health averaged 1.2 percent of GDP, while that on education averaged 3.7 percent of GDP. Despite steps (under the ERP) to reorganize health care delivery and to promote a more effective and equitable health care system, the quality of health care services is inadequate. In education, wages, salaries, and other staff costs absorb about 95 percent of the recurrent public spending in the sector. This leaves little for non-wage operations and maintenance, both essential to improve the quality of education. 15. The aggregate levels of public spending on health and education are less regressive than household income and expenditure distribution; the bottom fifth's share of health and education spending far exceeds its meager 4 percent share of all household spending (see Figure 1). But these public expenditures are generally regressive in absolute terms-that is, less goes to the poor, more to the rich. Between 1989 and 1992, health spending on the poorest fifth of the population fell from 12 to 11 percent of total health spending; the share for the richest fifth rose from 31 to 34 percent. Education spending was only marginally regressive in absolute terms in 1989 and 1992, being distributed almost uniformly across income groups. In 1988, the poorest fifth had 17 percent of total education spending, compared with 24 percent for the wealthiest fifth; in 1992, the poorest got 16 percent and the richest 21 percent. v

12 Figure 1 The Incidence of Public Social Spending , 1992 *-Lorenz Curve e- Education *-e - -Education Health Health ; a50 " 30 (20~ J-~ Cumulative Distribution of Population (%) Source: Demery, et al, Deficient Public Health Targeting. There has been little change in the regressivity of health expenditures in the four years, In 1992, the poorest's share of public spending on health centers and clinics was only 10 percent. Their share of out-patient services was 13 percent and of inpatient services 11 percent. For the richest fifth of the population the comparable figures are 31 percent for health centers and clinics, 35 percent for outpatient services and 32 percent for inpatient services. 17. There is also a huge urban bias in the regional distribution of health spending. In 1989, urban Ghana (a third of the population) took 42 percent of the total health budget and over 50 percent of outpatient spending; by 1992, these figures had increased to 49 percent and 55 percent. There are also gender differences in public health spending. Girls and women gained a larger and increasing share of expenditures-from 55 percent of total recurrent health spending in 1989 to 56 percent in But that change disguises a bigger and increasing share going to the wealthiest. In 1992, the women in the poorest quintile received 9 percent of total health spending, the richest 38 percent. 18. Clearly there is a strong case for increased targeting towards the poorest quintiles, rural areas, and poor women. Targeting rural areas and women should be relatively easy. But targeting the poorest in general is more difficult. Those who can pay should, with the money used to help the poor who cannot pay and to make improvements in service quality. Cost recovery, through price signals, can direct users to the appropriate type of medical facility leading to a more efficient use of these services. In fact, both economic theory and international experience suggest that cost recovery in high cost curative health care can lead to reduced fiscal pressures, and if used properly, to improvements in quality and in access to health facilities. vi

13 19. Recently, Ghana increased charges for drugs (and other curative services) at govenument health clinics and hospitals under the "cash and carry" scheme (a revolving fund for drugs), which in fact involves drug prices somewhat higher than their full cost. There were, of course, exemptions but few went to the poor. Both economic theory and experience in other countries suggest that cost recovery should not be attempted for preventive care while low-cost curative health services should be priced to cover some, if not all, costs. 20. Learning the Hard Way. In education, too, the story is largely the same-more to the rich,jless to the poor-except in primary education where the poorest fifth received 22 percent of public spending. For the richest, it was 14 percent. In secondary education, the poorest received 15 percent and the richest 20 percent. In higher education, spending is very regressive, with the share for the lowest quintile only 6 percent and that for the richest 45 percent. 21. Rural-urban inequalities are also evident, with urban areas receiving about 50 percent more than the rural areas in per capita terms. While per capita expenditures for primary education are higher in the rural areas, the rural-urban inequalities are strong and worsening in secondary and higher education. There are also gender inequalities across all education levels but it is worst in secondary education. In general, boys receive almost 50 percent more than girls. 22. Although net primary enrollment in Ghana increased from 68 percent in to 74 percent in , it is still not universal. Enrollments actually fell for the poorest quintile. At the secondary level, the enrollment rate increased from 33 percent in to 35 percent in , with only a minor fall for the poorest quintile from 29 percent to 28 percent. Economic theory and the lessons from other countries suggest that cost recovery should not be pursued at the primary level unless net enrollment rates are almost universal (95 percent or higher). 23. In 1987, under the first education reform program, small user fees were charged at the primary and junior secondary school levels, textbook fees were introduced for grades 3-10, and full cost recovery was introduced for meals at secondary schools. Thereafter, official fees were increased to keep pace with inflation, but user fees, as a percentage of Ministry of Education costs recovered, declined. This was because teacher and school administrator salaries were increased in real terms. Within two years, however, unofficial fees charged by schools increased dramatically, resulting in increased real costs to parents. In recent years, the Government has tried to rein in these unofficial fees. 24. In higher education, cost recovery should be as complete as possible. Before 1989, university education in Ghana was free of charge, and housing and meals were also provided at no cost to students. In 1989, a move toward cost recovery was made by charging students for meals (although not for tuition), a small step toward cost recovery. Recent Developments 25. Having established a good macroeconomic framework in 1991, Ghana was poised for rapid economic growth. Inflation was down to 18 percent. The budget surplus was equivalent to 1.5 percent of GDP. The growth in money supply was still high but falling, and real GDP growth was 5.3 percent. Private investment had risen continuously to a ten-year high of 8.1 percent of GDP. Sustained macroeconomic stability, rapid privatization, and further liberalization of the agriculture and financial sectors could have accelerated both investment and growth. vii

14 26. Instead, the 1992 fiscal shock, triggered by across-the-board wage increases of 80 percent for civil servants and by a fall in tax revenues undermined macroeconomic stability. This was further accentuated by a shortfall in foreign program grants. Private investnent fell precipitously-to 4.3 percent of GDP in The narrow fiscal accounts slumped from a surplus of 1.5 percent of GDP in 1991 to a deficit of 4.8 percent in 1992, a reversal of 6.3 percent of GDP. The corresponding change for the enlarged fiscal accounts was from a deficit of 4.0 percent to a deficit of 10.7 percent. 27. In 1993 and 1994, the Government was able to engineer a substantial fiscal turnaround of around 4 percent of GDP, although lower than originally planned. In 1994, divestiture receipts, helped to generate a fiscal surplus of 2.1 percent of GDP. But central bank borrowing by the Ghana National Petroleum Corporation led to a 46 percent rise in money supply and higher inflation. 28. The Government implemented new policy initiatives during to promote an enabling environment for private-sector development and reduce the role of government in the economy. There has been accelerated privatization of state-owned enterprises, including the Ashanti Goldfields Corporation, rapid growth of the stock exchange, improved dialogue with the private sector, and further liberalization of the regulatory framework. Despite these reforms, Ghana's efforts to promote private investment have been hampered by macroeconomic instability. The fiscal shock has not only highlighted the underlying fragility of the fiscal situation (i.e. the structural problem of an oversized public service), but has also generated new imbalances (interest bill) and new uncertainties (possibility of another large wage shock). 29. So what lessons can be drawn from Ghana's economic experience, particularly in view of the forthcoming general elections in 1996? There are three. First, avoid another large wage shock. Even a 30 percent across-the-board wage increase in 1996 would move the projected fiscal surplus of 1.2 percent of GDP into a deficit of 1.4 percent. At 16 percent of GDP, tax revenues are already high and there is limited scope for increasing them further to cover expenditure increases without hurting private-sector growth. Moreover, balance of payments needs are already substantial and mainly reflect overall budget deficits. 30. Second, reduce public expenditures as a percentage of GDP. Spending on wages and salaries, subsidies and transfers, and interest payments are high as a percentage of GDP. They need to be reduced to ease financing requirements, promote macroeconomic stability, and provide greater resources for non-wage operation and maintenance expenditures. This will mean reducing the size of an overstaffed public sector and improving public financial management systems. 31. Third, maintain tight monetary control. This is needed so that money supply can provide the nominal anchor to the financial system, and monetary policy can be used to control inflation. Access to central bank credit, an aggravating factor for monetary expansion, should be restricted and the excess liquidity in the system mopped up through the sale of official debt. The Government needs to adopt measures (from aligning the term structure of interest rates to developing new financial instruments) to enhance the effectiveness of instruments of indirect control and to encourage the development of a secondary interbank market in domestic debt instruments. viii

15 What Else Needs To Be Done? 32. Ghana's past success in reducing poverty clearly shows the crucial role of broad-based economic growth, most of which came from the private sector. Future growth, too, must come from the private sector. Rapid agricultural growth will be the key to continued reductions in poverty. Sustaining macroeconomic stability and increasing the quantity and quality of investments in human resources will be essential to promote rapid private sector growth. But there are other actions, too, that the Government can take to promote private-sector growth-encourage private investment in infrastructure, strengthen the financial sector, remove distortions in cocoa trading, and liberalize the petroleum sector. 33. Mobilizing Private Investment in Infrastructure. The Ghana 2000 study identified poor infrastructure as a major obstacle to growth and welfare improvements. Ghana's provision of communication and transportation services is no worse than that of its neighbors but is lower than the poorer countries of Latin America. For instance, Ghana has only 2.8 telephone main lines per 1000 inhabitants compared with 21 in Guatemala, 17 in Honduras, and 26 in Peru. Because of public sector resource constraints, private investment will be needed. 34. Private investment-both domestic and foreign-needs to be harnessed to fill the resource gap and expand infrastructure investment. The Government has already indicated its interest in taking this route. In doing so, Ghana is following a strategy pursued by many countries in Latin America and elsewhere that have encouraged private-sector provision of telecommunications services, power generation and distribution, water and sanitation facilities, as well as airport services, airlines, and railroads. 35. International best practice has demonstrated that private ownership and competition, combined with an appropriate regulatory framework, can trigger better productivity and service expansion in most infrastructure sectors. But in Ghana, current ownership of infrastructure is a far cry from international best practice. There are numerous barriers to private sector involvement, most of which are within the power of the Government to correct. These include government restrictions on entry, unpredictable criteria for tariff-setting, poor cost recovery, various risks (country, commercial, and currency), and poor regulation. These can be obviated by * Establishing and following transparent and predictable regulatory procedures; * Removing pricing and entry regulations; * Introducing detailed regulations for power and telecommunications governing prices, conditions of interconnection, service obligations, concession areas, and so on; * Drawing up a timetable for auctioning concessions to provide local power distribution, waste collection, urban transport, and water system billing and maintenance; and * Where private sector investment is possible (water, sanitation, urban bus transport, telecom), increasing prices to cover all costs of service provision and enforcing payment for services. 36. Strengthening the Financial Sector. Weaknesses in the financial sector that restrict financing opportunities for productive private investment are a serious impedimento business growth and expansion. Six years after the beginning of financial sector reforms in 1988, Ghana's financial system remains shallow. Broad money holdings (M2) are equivalent to 17 percent of GDP, as compared with 79 percent in Malaysia, 46 percent in Indonesia, 37 percent in Kenya, and ix

16 30 percent in Zimbabwe. The low level of intermediation is due mainly to lack of confidence in the financial system, insufficient competition, and weak financial infrastructure. 37. Confidence in the financial system suffered because of several policy interventions by the Government. In 1979, it undertook a currency conversion. In 1982, it demonetized the 50 cedi note, froze bank deposits in excess of 50,000 cedis pending investigation for tax liability, restricted bank loans for financing trade inventories, and required that business deals of more than 1,000 cedis be conducted by check. Whatever the intention, these restrictions undermined private sector confidence in the banking system and prompted substantial financial disintermediation. Confidence has been slow to recover. 38. Also, until recently, real interest rates on deposits have been negative, giving individuals little incentive to hold savings in bank accounts. People prefer to hold cash, so that monetary holdings are large when compared with monetary aggregates. Meanwhile, a proliferation of highretum, low-risk government paper has led to high interest rate spreads and has been crowding out bank lending to the private sector. There is also a low rate of financial innovation and poor quality of service throughouthe financial sector. 39. Ghana's formal banking system is dominated by state-owned banks that have a monopoly in banking services. Rural areas are particularly underserved by financial institutions. Ghana's informal financial sector is large, mobilizing an estimated 45 percent of all private sector financial savings, but its capacity to intermediate between savers and investors is limited. This is due to people's savings behavior and the absence of strong links with the formal sector. There may be considerable scope for profitable contacts between the informal and the formal sectors such as bringing the savings generated by susu collectors into the formal system and allowing traders to draw on formal credit by simplifying the use of collateral. Uncertainty about the Government's view towards the desirability of unmonitored informal financial intermediaries has interfered with these linkages. 40. The main challenge is to bring more savings into the financial system to finance productive private investment. This means * Reviving confidence in the banking system, including divesting state-owned banks; * Promoting competition in the banking and the non-bank financial sector, where the Social Security and National Insurance Trust and the State Insurance Corporation have a virtual monopoly; * Supporting informal financial intermediaries; * Investing in legal and financial infrastructure; and * Enhancing intermediation by rural financial institutions, particularly rural banks. 41. Liberalizing the Cocoa Sector. Ghana was the world's leading cocoa producer for most of the 20th century. Production peaked at 550,000 tons in percent of the world market. By 1983, however, production had fallen to 159,000 tons and world market share to 9 percent as a result of the high taxation of cocoa. There has been a modest recovery since, with production rising to 245,000 tons in 1993 (10 percent of the world market). 42. There are two main policy distortions in the cocoa sector. First, the Government maintains a monopoly in the export of cocoa and a near-monopsony in the domestic market. Second, the x

17 cocoa sector is taxed heavily, both explicitly and implicitly, with the result that the cocoa farmer receives only a small share of the producer price, reducing incentives for investment and production. 43. These distortions have been primarily responsible for the lack of investments and the decline in output. Some policy measures have sought to address these concerns, including raising nominal producer prices and liberalizing intemal marketing by granting licenses to private licensed buying companies. But these companies had difficulty raising finance and the marketing margin allow,ed them is insufficient to cover costs. Effective competition has been elusive, however, and will remain so unless cocoa pricing is liberalized and private traders have direct access to export markets. Liberalizing cocoa marketing is critical to restoring the health and vitality of the sector and to increasing incomes of the poor. These reforms include * Liberalizing extemal marketing, which would boost producer prices for cocoa farmers, eliminate the inefficiency costs of the monopoly, and create incentives for rapid growth in the sector; * Reducing the taxation of the cocoa sector, regardless of the marketing arrangements in place, and making taxation tranisparent and predictable; and * Removing infrastructural impediments to the cocoa trade -absence of feeder roads, poor maintenance, the high cost of transportation, and so on. 44. Liberalizing Petroleum. Petroleum is the single largest import, accounting for more than 10 percent of merchandise imports in 1994 and equivalent to about 15 percent of merchandise exports. Petroleum is also a major source of revenue. Taxes on petroleum accounted for a quarter of tax revenues in 1994, equivalent to almost 4 percent of GDP. 45. The Ghana National Petroleum Corporation is responsible for the exploration and procurement of crude oil. Refining is handled by the Tema Oil Refinery, and the Ghana Oil Corporation distributes about a third of all petroleum products. All these companies are in state hands. 46. The sector suffers big economic inefficiencies arising from the public monopoly in refining. Moreover, administrative regulation of retail prices made it more difficult to pass on to the consumer the higher import costs due to the depreciation of the cedi. Private sector growth would be promoted by * Removing the monopoly of the Ghana National Petroleum Corporation over the procurement of crude petroleum; * Commercializing the Tema Oil Refinery in the short-term, * Liberalizing petroleum prices with negotiated mark-up margins for oil distributors and an ad valorem tax rate for excise taxes, and with frequent review of marketers' margins; and * Privatizing the Ghana National Petroleum Corporation, the Tema Oil Refinery, and the Ghana Oil Corporation. And Elsewhere in the Private Sector 47. The private sector is the main contributor to GDP (two-thirds) and to GDP growth. Despite ten years of reform, however, and a running divestiture program, the private sector has xi

18 been unable to expand its share of GDP to any significant degree. Accelerated growth requires a major stimulus from the private sector. Moreover, just as low agricultural growth has limited aggregate GDP growth, so has it put the brakes on private sector growth. Faster growth can only be achieved by boosting agricultural productivity. 48. At the same time, the private industrial sector, in particular manufacturing, has not vet fully contributed to economic growth. Private investment in manufacturing needs to be encouraged, through the maintenance of an adequate enabling environment, to generate emnployrnent and a solid private industrial base. Following the liberalization of the trade regime, the manufacturing sector was pressed to improve efficiency and reduce production costs. This environment, however, has now persisted for a number of years and the initial adjustments are likely to have occurred. Higher investment and growth in private manufacturing may remain elusive unless it is supported by a viable and stable macroeconomic environment and the availability of adequate financing. Economic Prospects 49. Provided that the Government is able to stabilize the economy and proceed in a significant manner on the policy reforms identified above to promote private sector growth, then the private sector response over is projected to be more favorable. GDP growth rate is projected to increase from 3.8 percent in 1994 to 6.3 percent in 2004.' Agriculture's growth rate is projected to increase from 1.0 percent in 1994 to 3.3 percent in 2004, so that the per capita growth rate in agriculture will become positive. The increase in agricultural growth rate is projected to come from the continued liberalization of the sector, particularly in cocoa export marketing, and from investments in infrastructure. Mining, especially gold, is projected to continue to grow rapidly over the medium-term. Mining output growth rate is projected to increase from 5.1 percent in 1994 to 8.0 percent in This is based on the output effects of recent investments in largescale gold mining. Manufacturing sector growth should also improve from 1.5 percent in 1994 to 6.1 percent in 2004 due to projected increases in investment and efficiency. The growth in tradables -- in agriculture, mining, and manufacturing -- should provide support to service sector growth which is projected to continue having strong growth, improving from 6.3 percent in 1994 to 7.5 percent in Prospects for the export of both traditional and non-traditional products are also good, reflecting expected positive developments in both domestic and foreign markets. Investment and savings are projected to recover and grow well beyond their 1991 levels. By 2004, gross investment is projected to reach 21.2 percent of GDP while national savings is projected to reach 18.1 percent of GDP, substantially narrowing the investment-savings gap. 50. In order to sustain the above growth projections, concessional external assistance, even if declining, will continue to play a significant role in Ghana's economic development over Commitments of US$ 800 million per year are sought from donors for this period. Program aid is expected to decline and a bigger share of project aid is expected to go towards human resource development and rural infrastructure. These are World Bank projections based on the expected pace of reforms based on the Government's commitments. If, however, reforms proceed faster than expected, higher growth rates could be achieved; conversely, if the reforms proceed at a slower pace, lower growth rates may result. xi.

19 Agenda for Further Work 51. This report has raised several important questions. These are not only in connection with the analyses of the poverty changes but also with respect to Ghana's growth prospects. The ongoing Extended Poverty Study will seek to provide answers to the differential pattem of poverty reduction over regions and over socioeconomic groups. With respect to Ghana's growth prospects and its potential for reducing poverty, there are concems over whether future growth will be as broad-based as past experience so that the poverty reduction is as sizable. This concem arises from, the low human capital base of the poor, the lack of adequate targeting of public social expenditures, and the generally poor quality of health and education services. In addition, the human capital constraints and the low rates of agricultural growth also raise concem over the sustainability of long-term growth in Ghana. 52. The above concems require an investigation of the following questions: * What are the reasons behind agriculture's slow growth under the ERP? * Is the growth in services that has been witnessed under the ERP sustainable? To what extent has it been driven by foreign aid? * How critical are human capital constraints to Ghana's growth prospects? * What can be done to ensure that Ghana's future growth will be as broad-based as past experience over ? These questions are the agenda for future economic work which will seek to inform the debate on Ghana's path of accelerated growth with poverty reduction. xiii

20

21 Chapter 1 RECENT ECONOMIC DEVELOPMENTS AND IMPLICATIONS In the last two years, Ghana crossed a new threshold in signaling the Government's commitment to the private sector through significant policy reforms and through major fiscal turnarounds in 1993 and Unfortunately, the fiscal shock of 1992 has undermined a private investment response to those reforms. Notwithstanding the turnaround in fiscal deficit equivalent to 4 percent of GDP, the legacy of the fiscal shock continues to haunt macroeconomic performance. The fiscal shock reversed the progress that had been made in controlling inflation and in achieving internal and external balance. Ghana's ability to accelerate growth depends on controlling inflation quickly and sustaining macroeconomic stability in the medium-term. 1.1 Since 1983, the Government of Ghana has implemented a gradualist but sustained adjustment strategy under the rubric of the Economic Recovery Program (ERP). The reforms under the ERP have successfully turned the economy around. In the decade prior to the introduction of the ERP, the growth rate of real GDP averaged -2 percent per annum; in the decade since then, it has averaged around 5 percent. Average per capita income growth rate has increased during the same period from a negative 5 percent per annum to a positive 2 percent per annum. Inflation was reduced from an average of 56 percent per annum -- and a peak of 123 percent in to an average of 26 percent per annum, with the lowest rate of 10 percent in The year 1991 was the high point in Ghana's economic recovery since 1983 with inflation down to i8 percent, the budget in a surplus equivalento 1.5 percent of GDP, the growth of the money supply, still high at 27 percent, but falling, and real GDP growth rate at 5.3 percent. Private investment had risen continuously to a decade high of 8.1 percent of GDP. But the 1992 fiscal shock marked a dramatic break in the otherwise sustained progress in Ghana in terms of inflation, private investment, fiscal balance, and the current account balance. Subsequento the jump in wage expenditures, money supply increased by more than 50 percent, the current account deficit widened to almost 9 percent of GDP, the cedi depreciated sharply, and inflation accelerated, leading to the collapse of private investmento 4.3 percent in This chapter examines Ghana's recent economic performance and its implications for the future. For this purpose, the three years, , are contrasted with the three years, , against the backdrop of the recent policy reforms and the sharp recession of which led to the adoption of the ERP. Recent Policy Reforms 1.4 Over , the Government undertook several new policy initiatives to foster a sea change in the enabling environment for increased private sector activity and the role of the Government in the economy. These key policy changes related to privatization of state-owned enterprises (SOEs), development of the Stock Exchange, dialogue with the private sector, and changes in the regulatory framework. 1.5 Privatization. The pace of divestiture of SOEs has picked up steam since March The sale of the Government's majority shares in the Ashanti Goldfields Corporation (AGE) and minority holdings in seven of the most profitable enterprises marked a significanthreshold in private sector development. The sale of the Governnent's holdings of AGE shares -- reducing its shareholdings from 55 percent to 30 percent - to farmers, workers, and civil servants in Ghana has generated a broader constituency for

22 2 accelerated privatization. Similarly, the sale of its minority holdings in seven companies listed on the Ghana Stock Exchange to international fund managers has improved Ghana's image among foreign investors. Of the 260 majority-owned SOEs, owned both directly and indirectly by the Government, only 64 SOEs were divested between 1988 and Another 10 small and medium-sized majority-owned SOEs were sold to the private sector in 1994 under the Government's new accelerated divestiture program. 1.6 Two aspects of the Government's proposed accelerated privatization program are worthy of mention. First, as part of this program, the Government proposes to prepare for privatization four large and strategic enterprises -- Ghana Telecom, State Insurance Corporation, State Housing Corporation, and the Ghana Oil Company -- which have significant linkages with the rest of the economy. Second, in order to accelerate the pace of privatization, the Government aims to enhance substantially the transparency of divestiture procedures, use various methods for divesting the state-owned enterprises, and contract private firms to implement the divestiture procedures. In order to facilitate the privatization program, the Government passed the Statutory Corporations Act in 1993 which converted a number of public statutory corporations into limited liability companies with salable shares. 1.7 Development of the Stock Exchange. The Ghana Stock Exchange (GSE) commenced operations in November 1990 for trading in corporate equities and bonds, and government securities. However, the subsequent offloadings of the Government's minority holdings in seven companies and the listing of AGE shares boosted market capitalization to Cedis 1,972 billion (US$2.1 billion) by the end of May 1994, equivalento 34 percent of GDP. Shares of the AGE alone, however, accounted for over 85 percent of this capitalization. The proposed accelerated privatization program will further boost total market capitalization which already exceeds that of most Sub-Saharan African countries including Kenya (US$1.8 billion) and Zimbabwe (US$1.8 billion) although still very small compared to those in the fastgrowing East Asian economies. 1.8 Dialogue with the Private Sector. In 1991, the Government initiated a dialogue with selected leaders of the private sector under the Private Sector Advisory Group. This was subsequently broadened and strengthened through the Private Sector Roundtable which is composed of a representative spectrum of Ghanaian entrepreneurs and senior government officials. The Private Sector Roundtable meets regularly to discuss relevant policy issues. In 1994, the Roundtable prepared reports on privatization, technology development, financial sector, and legal and commercial infrastructure which have since been adopted by the Government. The dialogue with the private sector has helped to cultivate improved private sector perceptions of the Government's attitudes towards it. 1.9 Changes in the Regulatory Framework. The Government has signaled strong support for the private sector through actual and planned changes in the regulatory framework. First, the Government has approved a more liberal Investment Act to improve the climate for private investment.' A key feature of the Act is the elimination of the need for a formal government approval before investors can invest; as long as the investment satisfies the conditions in the Act, the investor can register with the Registrar and proceed to invest with a report to the Ghana Investment Promotion Center (GIPC). In addition, the Act reduces the minimum foreign capital requirements to encourage joint ventures between foreign investors and domestic small-scale enterprises. Also, discretionary and non-uniform fiscal incentives have been drastically reduced relative to the 1985 Act; a low corporate tax rate and duty exemptions on capital equipment provide uniform incentives for all investors. Second, the Government accepted the obligations under Article VIII of I The Ghana Investments Promotion Center Act, 1994.

23 3 the IMF's Articles of Agreement which stipulates that the Government will seek to achieve its balance of payments objectives through appropriate financial and exchange rate policies rather than through quantitative controls on the current account. This will have the effect of further boosting Ghana's image with international investors. Third, the Government has submitted a bill to Parliament for facilitating private participation in the communicationsector. Fourth, the Government has announced its intentions to seek private investment in power generation The Government has continued to deepen reforms in other areas, too, especially the tax system. The dispersion in import tax rates has been reduced. The 4-tiered import tariff structure was simplified to a 3-tiered structure of 25, 10, and 0 percent in The Value-Added Tax (VAT) was implemented in March 1995 covering the retail and service sectors. 2 Furthermore, the liberalization of domestic cocoa marketing was initiated in Recent Fiscal Performance 1.11 Notwithstanding the recent policy reforms, Ghana's efforts to promote private investment have been hampered by renewed macroeconomic instability in 1993 and 1994 due to the large fiscal shock in The shock not only highlighted the underlying fragility of the fiscal situation (i.e., the structural problem of a very large public service) but also generated new imbalances (i.e., interest bill) and new uncertainties (i.e., possibility of another large wage shock) Fiscal reform has been at the core of the ERP. Its primary objective was to restore fiscal discipline and to increase public investment in physical and social infrastructure. By 1991, the ERP appeared to have reigned in the narrow fiscal accounts 3 which improved from an average deficit of 3.5 percent of GDP during the economic trough of to an average surplus of 0.8 percent in (see Table 1.1). In 1991, the narrow fiscal surplus had reached an ERP-period high of 1.5 percent of GDP. The fiscal consolidation was particularly impressive as it occurred along with substantial increases in tax revenue, which increased from an average of 5.3 percent of GDP in to 11.8 percent in , and with sizable increases in capital expenditure from own revenue, which increased from 1.0 percent of GDP to 2.6 percent during the same period. When aid-financed expenditures are included, total capital expenditure increased from an average of 1.2 percent of GDP to 5.0 percent during the same period. This is in contrast to the experience of most of the adjusting countries in Sub-Saharan Africa where the reductions in budget 4 deficits had to be accompanied by lower capital expenditures because of only small increases in revenues. 2 The 17.5 percent VAT, which replaced the IS percent sales tax, applies to a broad range of services such as electricity supply, telecommunications, and professional services, that were not taxed previously. 3 The narrow fiscal accounts refer to the budget exclusive of capital expenditures financed through foreign project aid (both loans and grants) but includes program grants. This is more commonly used because by excluding foreign project aid, it insulates the budget from delays or disruptions in the foreign aid funding of projects and also provides a better indication of the stance of fiscal policy. 4 See, for instance, The World Bank, 1993, "Adjustment in Africa," pages 4748.

24 4 Table 1.1: SELECTED ECONOMIC INDICATORS % NATIONAL ACCOUNTS (million cedis) Nominal GDP (ad market prices) B * 2; 5 404E...,,4 Real GDP (contant 1975 prices) d,...T.... ' ;.. (annual percentage chan 5 e) Rea GDP (conrant 1975 prices) , 5. Agriculture ' 5 Industry Services I. Gross Investment (% of GDP) , Public Investmntd Private Investment National Savings :0 Public Savinp O,,..4 *5 PrivateSavings MONEY & PRICES (annual percenrage change) Broad Money Inflation TreauryBill Rate (91-day) ,.: 5.36 (indices) TermsofTrade(1984=100) t.5 Real Exchange Rate (1984=100) i i.- Exchange Rate (cedis/uss) i i.250 6:3- NARROW FISCAL ACCOUNTS A (% of GDP) TotalRevenueaandGrants TaxRevenue Non-TaxRevenue ' 5. Grants Divestiture Ok TotalExpenditure&NetL.nding CurrentExpenditures , Ii s; Wages&Salaries Other Goods&Services , InterestPayments s Subsidies & Transfers OtwrExpenditures ss 0 1' Special Efficiency Capital Expenditures&Net s: Lending Capital Expenditure ' &. Net Lending ' 5 Current Account Deficit x Narrow Budget Deficit ('mf) lcludinggrants : Excluding Grants Enlarged Budget Deficit s Enlarged Capit Exp S. U.. s BALANCE OF PAYMENTS (millions of US dollars) Merchandise Expot (fob) ::sai: >ss5.::::::: w mports (cif) s': sfs, 1133.? '. S Merchandise Trade Balance , Resource Balance CurrentAccountBalance , 5.59'.. 4 i. Capital Account (net) S 3hI S31 Overall Balance : Memno Iteim.: GDP per Capita (USS) $4; 40f$ Current Account Deficit ( / of GDP) , R.4. a/ The narrow fiscal accounts excludes foreign project aid (both loans and gramns) but includes program grants. Source: Government of Ghana; World Bank saff esimates

25 The Anatomy of the 1992 Fiscal Shock. The fiscal shock of 1992 was triggered by a large across-the-board increase in wages for civil servants 5 and by a fall in tax revenue. This was accentuated by a shortfall in foreign program grants. Total expenditures and net lending increased by 3.3 percentage points of GDP in 1992 (and by another 3.8 percentage points in 1993) while tax revenue fell by 2.4 percentage points of GDP in Grants fell by 0.3 percentage points of GDP. These combined to produce a sizable fiscal shock that reversed more than 7 years of sustained improvement in fiscal control and discipline. The narrow fiscal accounts slumped from a surplus of 1.5 percent of GDP in 1991 to a deficit of 4.8 percent in Expenditure Side. The policy decisions that led to the fiscal shock took place in late Partly as a result of that, and partly because progressively higher government borrowings to meet the budgetary deficits led to increasing interest payments, the expenditure increases occurred in both 1992 and On the expenditure side, total expenditures and net lending rose by 25 percent in 1992 and by a further 22 percent in 1993 (from 13.7 percent of GDP in 1991 to 17.0 percent in 1992 and 20.8 percent in 1993). As a result, average total expenditures and net lending increased from 13.7 percent of GDP in to 20.0 percent in Most of the increase in total expenditures reflected increases in current expenditures which increased from 10.2 percent of GDP in 1991 to 12.4 percent in 1992 and 16.3 percent in Compared with levels of current expenditures in of 10.2 percent of GDP, current expenditures were more than 50 percent higher during at 15.6 percent of GDP There were three sources for the increase in current expenditures. First, as mentioned above, the Govermment provided an across-the-board wage increase for civil servants averaging 80 percent. This was meant to reduce the disparities in wages between the civil servants and employees in the rest of the public sector. This increase was made effective from July 1, 1992, and led to an increase in wages and salaries from 4.2 percent of GDP in to 5.7 in Second, the Government provided an increase in the wages of employees in subvented organizations and in pensions. This was reflected in an increase in net subsidies and transfers from 2.2 percent of GDP in to 3.1 percent in Third, interest payments increased from 1.4 percent of GDP in to 3.3 percent in reflecting the increase in public domestic debt and revised higher estimates of Ghana's external debt In addition, total expenditures also rose from higher outlays on "special efficiency" and on capital expenditures. The "special efficiency" account reflects public sector retirement and end-of-service benefits. Towards the end of 1992, the Government agreed to a programn to encash accumulated retirement benefits or end-of-service benefits as a palliative for workers to move to a new pension scheme. These payments were to be phased out over four years, with the first payment of 20 percent of the total for stateowned enterprises and 25 percent of the total for subvented organizations in As a result of these payments, expenditures under "special efficiency" increased from 0.5 percent of GDP in to 1.1 percent in At the same time, capital expenditures rose from 2.6 percent of GDP in to 3.1 percent in Most of this increase was actually due to increased outlays in 1992 on development projects in the rural areas prior to the elections. 5 In 1992, the Goverment faced a series of strike activities as it found itself under increasing pressure from various labor unions for pay increases. The doctors went on strike in May, the nurses in June, the workers of the Ghana Cocoa Board in July, and the railway employees and the civil servants in September. 6 The corresponding change for the enlarged fiscal accounts, which includes capital expenditures financed through foreign project aid, was from a deficit of 4.0 percent to a deficit of 10.7 percent.

26 Revenue Side. Between and , total revenues and grants increased from 14.5 percent of GDP to 18.3 percent reflecting increases in both tax and non-tax revenue. During the same period, tax revenues increased from 11.8 percent of GDP to 12.9 percent while non-tax revenues increased from 1.2 percent of GDP to 4.2 percent. Grants fell marginally from 1.4 percent of GDP to 1.2 percent However, the period did not see a uniform trend in revenue increases. In 1992, the budgetary problem of a 3.3 percentage point increase in total expenditures was compounded by a shortfall on the revenue side. Total revenue and grants fell from 15.2 percent of GDP in 1991 to 12.2 percent in 1992, or by 20 percent. This fall was caused primarily by three sources. First, taxes on domestic goods and services fell from 5.4 percent of GDP in 1991 to 4.3 percent largely on account of diminished petroleum tax receipts. The latter occurred because the Government (a) failed to pass through the increases in the domestic price of petroleum resulting from the depreciation of the cedi; and (b) simultaneously wrote off cedis 14 billion that were due from the Ghana National Petroleum Corporation (the state petroleum import agency) as revenue from the petroleum tax Second, taxes on international trade and transactions fell from 4.6 percent of GDP in 1991 to 3.7 percent in This was due to (a) diminished revenues from cocoa export taxes because of a fall in world cocoa prices and higher operating costs of the Cocoa Board; and (b) fall in import duty revenues because of the reduction in the level of the special tax on imports, the abolition of the super sales tax, the elimination of the import duties and sales taxes on all building materials, and a general increase in the inefficiencies in tax collection. And third, foreign program grants fell from 1.4 percent of the GDP in 1991 to 1.1 percent in 1992 due to delays in disbursements on account of slippages in the Government's program implementation Implications of the Fiscal Shock. The fiscal shock has three major implications for restoring macroeconomic stability. First, levels of current expenditures on wage-related items 7 and interest payments, were 56 percent higher in than in Wage-related expenditures rose from an average of about 6.4 percent of GDP in to 8.8 percent of GDP in ; in 1994 it stood at 9.0 percent of GDP. It thus accounts for more than one-half of tax revenue and exactly one-half of current expenditures. Even if such expenditures grow at only the rate of inflation, their share will not fall below 7.8 percent of GDP by If, on the other hand, there is a real increase in wages of around 5 percent per year, there will be no change in the share of wage-related expenditures in GDP Interest paymnents have risen even faster from 1.4 percent of GDP to 3.3 percent over the same period. Virtually all of this increase relates to domestic debt and has occurred in 1993 and 1994 due to the large open-market operations carried out by the Bank of Ghana (BOG) to mop-up the excess liquidity. In view of the large monetary expansion in 1994, the interest bill is likely to peak next year at around 4.8 percent befbre beginning its gradual decline Second, the above implies that non-wage operations and maintenance expenditures (O&M), i.e., expenditures on goods and services necessary to secure satisfactory returns on capital investments, have been squeezed. Though their levels have increased between the two periods from 2.4 percent of GDP to 2.8 percent, this increase is not commensurate with the expansion in the Government's capital stock during the same period. 7 The sum of Wages & Salaries and Subsidies & Transfers.

27 Third, though tax revenues have been raised from 12.4 percent of GDP in 1991 to 15.9 percent in 1994, the rise has not even kept pace with the rise in current expenditures. Clearly, there are limits to raising taxes for meeting current expenditures under a private sector development strategy. Also, divestiture receipts are likely to taper off. In any case, divestiture receipts are best used for the reduction of domestic or external debt, or the acquisition of foreign assets, and not for budgetary purposes. The Fiscal Turnaround of 1993 and In 1993, the Government was able to engineer a major turnaround in the fiscal situation even if the turnaround was lower than originally planned. The narrow fiscal deficit was cut to 2.5 percent of GDP from 4.8 percent of GDP in 1992, implying a turnaround of 48 percent. This turnaround was largely due to a substantial improvement in total revenues which increased by 5.6 percentage points of GDP. The turnaround in the narrow deficit was less than the Government's ambitious target of a surplus of 0.3 percent due to a shortfall in petroleum tax revenue because the second round of increase in excise duty was not implemented on the advice of the Parliament. Thus money supply grew by almost 30 percent In 1994, too, the Government brought about a fiscal turnaround of 44 percent (excluding excess divestiture receipts), albeit less than planned. The Government's target was to achieve a fiscal surplus of 0.9 percent of GDP but, in reality, the Government was only able to reduce the deficit to 1. I percent of GDP. However, the use of the excess divestiture receipts averted any monetary expansion and yielded a fiscal surplus of 2.2 percent of GDP. Other items of the fiscal account actually fell short of targets by 2.0 percent of GDP. These were primarily due to two sources: (i) a shortfall in import tax revenues due to the fall in import volume, a shift in the composition of imports away from consumer goods, the failure in charging the 10 percent duty on exempted goods, as planned, and the inefficiencies in customs collection; and (ii) an overrun in current expenditures on personal emoluments and domestic interest payments. Recent Economic Performance 1.27 The fiscal shock was particularly damaging to macroeconomic stability and growth that had been nurtured through sustained and gradual policy reforms. Its ramifications for the economy, as we shall see below, have been severe and have proven difficult to reverse Investment. Gross investment exhibited an upward trend up to 1991, rising from 4.7 percent of GDP in to 14.5 percent in But gross investment in averaged the same due wholly to its sharp fall in from 15.8 percent in 1991 to 12.8 percent in 1992 (see Figure 1.1) The fall in private investment was due to political uncertainty and macroeconomic instability. The latter arose from the fiscal shock of The slack in private investment was offset partly by an increase in public investment from 7.7 percent of GDP in 1991 to 8.5 percent in Since 1992, private investment has not been able to recover and hovers around 4.5 percent of GDP while public investment has grown to 11.5 percent of GDP in As a result, public investmnent has increased its share of total investrnent from 49 percent of gross investment in 1991 to 72 percent in The cost to Ghana in terms of the lost private investment due to the fiscal shock has been high. If private investment had remained at 1991 levels during (and assuming unchanged public investment), gross investment over would have been higher by almnost $950 million. If, on the

28 8 other hand, private investment had maintained its pre-1992 trend during (see Figure 1.1), private investment would have been much higher, reaching 15 percent of GDP by Figure 1.1 Investment in Ghana, (% of GDP) 1F_ -Gross Investment -- Public Investnent - - Private Investmentt 16.00% % 12.00% % % _-{ oc % " 4.00% O % t_---- a 0.00% I Figure 1.2 Savings in Ghana, (% of GDP) National Savings Public Savings Private Savings 10.00% 8.00%. 6.00% % % A 0 00% 1I I I I I % i % 1.31 Savings. As with gross investment, gross national savings also maintained an upward trend up to 1991, doubling from a low of 4.0 percent of GDP in to 8.0 percent in During this same period, private savings increased from 4.6 percent of GDP to 5.6 percent while public savings increased from -0.6 percent of GDP to 2.4 percent. But the fiscal shock sent gross national savings into a tailspin, averaging only 3.8 percent of GDP in The decline in national savings took place in 1992 and 1993, concomitantly with the fiscal shock. From 9.3 percent of GDP in 1991, gross national savings fell to 4.2 percent in 1992 and to 1.4 percent in 1993 (see Figure 1.2). More than 50 percent of this decline was accounted for by the fall in public savings (the fiscal current account) which fell from a surplus of 3.2 percent of the GDP in 1991 to a deficit of World Bank estimates of gross national savings are lower than the RAF's because they do not include net official transfers from abroad.

29 9 percent in Private savings also fell during the same period from 6.1 percent of GDP in 1991 to 2.3 percent of GDP in In 1994, national savings saw some recovery, rising to 7.3 percent of GDP. This increase was fueled from public sources, primarily the receipts from the divestiture of SOEs. Private savings rose only marginally to 2.6 percent of GDP which suggests that even the recent increase in gross national savings, though still very small, may not be sustainable As with private investment, the cost to Ghana of the decline in private savings over has been high. Assuming that private savings had stayed at their 1991 levels, then the cumulative additional flows of private savings over would have amounted to $620 million in current values. If private savings had maintained its pre-1992 trend (see Figure 1.2), the gains would have been much greater and private savings in 1994 would have been almost 10 percent of GDP Monetary Policy and Prices. The budgetary deficit of 1992 was met by increased borrowings from the banking system. Money supply (broad money - M2) increased by more than 50 percent that year. Although the growth in money supply has been reduced since then, it averaged about 41 percent per annum in against 23 percent in The year 1992 also marked a reversal in the pattern of Government's financing of the budget deficit. Over most of the earlier ERP years, the Government was making net repayments to the banking system. In 1991, net repayments amounted to 2.0 percent of the GDP; in 1992, however, the Government had net borrowings from the banking system equivalent to 3.4 percent of GDP or three-quarters of the narrow fiscal deficit (see Table 1.2). As a result, excessive liquidity was introduced into the economy, fueling inflation. Table 1.2: Financing of the Budget (percent of GDP) Total Financing Net Foreign Financing Borrowing Amortization Net Domestic Financing Banking System Nonbanks Source: IMF 1.37 In 1993, money supply growth was brought down to 27 percent. However, in 1994, it rose to 34 percent despite a significant budgetary surplus. This will exceed the program target of 5 percent by a substantial margin. Unlike earlier years, the 1994 money supply growth is not predicated by higher fiscal borrowings but by higher borrowings by parastatals and by lower repayments to the banking system by the Government. The latter was largely due to the significant shortfalls in foreign program loans which were $120 m less than projected With the money supply expanding fast, inflation has not been far behind (see Figure 1.3). Sustained reductions in money supply growth over had succeeded in bringing inflation down from 40.0 percent in 1987 to 18.0 percent in This continued for the first half of But money supply expanded rapidly thereafter. Since this was a late development, its inflationary impact was not felt that year and inflation averaged 10.1 percent in But in 1993, inflation shot up to 25 percent and in 1994, it averaged 24.8 percent instead of the programmed 15 percent.

30 10 Figure 1.3 Money Supply and Prices, Broad Money (M2) Consumer Price Index - CI - Food Price Index 160 ~ i o100 ' 80 \ E W While changes in aggregate demand have been largely responsible for the price increases, changes in the weather, and thereby in food outputs, have also contributed to inflation in the short term. Food prices, which account for almost half of the consumer price index, closely follow domestic food production. For instance, 1990 was a year of drought and food scarcity pushed food inflation to more than 40 percent. This fueled the increase in the consumer price index that year despite the decline in the rate of money growth. The following year, 1991, was a particularly good year for rains. As a result, agricultural output was high and the rate of increase of food prices was low at 9.0 percent. In 1992, although the climatic conditions were not very favorable, inflationary pressures were contained through greater availability of food imports and greater releases from public stocks. Inflation was kept down even as money supply growth accelerated. Food inflation continued to fuel consumer price increases in 1993 and Interest Rates. The improved financial situation in the early part of 1992 had permitted nominal interest rates to fall to 19 percent from their high 1991 level of 29 percent (see Figure 1.4). But the need to contain the rapid monetary expansion in late 1992 and 1993 that followed the fiscal shock led to an increase in nominal rates by more than 50 percent in In addition to their increased volatility, the rise in nominal interest rates fed directly into the fiscal shock, increasing the domestic interest burden for the Government Movements in the real interest rate were also affected by the fiscal shock. Real interest rates fell in 1992 to 6.1 percent from the ERP period peak of 18.9 percent in They fell again in 1993 to 3.3 percent as the rise in the nominal interest rate was insufficiento offset the 150 percent increase in inflation that year. The increase in inflation in 1994 has further eroded the real interest rate to 1.0 percent.

31 11 Figure 1.4 Nominal and Real Interest Rates, Nominal Interest Rate---Real Inerest Rate A ~~~~~-0 / -5 0 W% 0% 0% _ - _ Note: ' The nominal interest rate used here is the end-period 91 -day Treasury Bill rate which is deflated by the end-period inflation rate to get the real interest rate Balance of Payments. In 1992, Ghana's overall balance of payments deteriorated from a surplus of $171 million in 1991 to a deficit of $124 million in This was due to the sharp deterioration in both the current account and the capital account. The budgetary explosion of 1992 is reflected in the increase in the current account deficit from 6.7 percent of GDP in 1991 to 8.6 percent of GDP in 1992 (see Figure 1.5). At the same time, the capital account worsened because of (i) delays in the disbursement of long-term loans from multilateral and bilateral sources due to slippages in the Government's implementation of policies, and (ii) substantial private short-term capital outflows due to the looser domestic money market conditions. The response of short-term capital flows was dramatic. A net inflow of $16 million in 1991 was reversed into a net outflow of $119 million Comparing the periods and , the current account balance has worsened, increasing from a deficit of 6.7 percent of GDP to 10.2 percent. In fact, the recent current account deficits have been the largest since the start of the ERP and, as Figure 1.5 shows, they closely mirror not only the trends but also the levels of the overall budgetary deficits The above has had major implications for Ghana: (i) it suggests that during , Ghana was heavily dependent on capital inflows, and more so than during the earlier years of the ERP; (ii) the financing of the higher current account deficits must imply that external debt has risen; (iii) it also suggests that the private sector accounts -- private savings and private investments, are largely in balance; and (iv) therefore, the capital inflows primarily finance the overall budget deficit These implications were indeed realized. Net flows on the capital account averaged $651 million in compared with $539 million during and $141 million during Net foreign borrowings, both public and private, have increased from $291 million in to $458 million in Almost half of this increase was accounted for by higher private borrowings. External debt to GDP

32 12 ratio increased from 61 percent in to 78 percent in although debt service as a percentage of the exports of goods and services improved from 38 percent in to 29 percent in because of the decline in interest and principal payments. The private investment-savings gap has moved within a narrow band of -2 to 3 percent of GDP over the ERP period, being in almost close balance in 1985 and in Figure 1. 5 Budgetary and Current Account Deficits, Overall Budget Deficit (enlargedef.) - Current Account Deficit (% of GDP) rrlij a. o Note: The Overall Budget Deficit (enlarged definition) includes capital expenditures financed through foreign project aid (grants and loans) Output Growth. As if the fiscal shock and its concomitant effects on the economy were not bad enough, real GDP growth during averaged 4.2 percent, slightly lower than the average of 4.6 percent for (but still much higher than the negative 3.8 percent that prevailed during which marked the height of the economic crisis in Ghana). But the decline in the real GDP growth rate in had more to do with exogenous shocks in agriculture due to poor weather conditions than to the fiscal shock. Agricultural growth rates fell from 2.3 percent in to 1.1 percent in The output decline would have been greater if industry and services had not grown to mitigate the negative effects of the decline in the agricultural growth rate. The rate of growth of industry and services stayed constant between the two periods at about 4.4 percent and 6.8 percent of GDP respectively After falling from 5.3 percent in 1991 to 3.9 percent in 1992, real GDP growth recovered in 1993 to the ERP average of 5.0 percent. But it again fell short in 1994 for which it averaged 3.8 percent (see Table 1.1). The lower growth rate in 1994 was caused by the untimely rains in the Akosombo Dam cachment area, which reduced power generation and affected industrial output, and by the ethnic conflict in the north which disrupted agricultural production. Conclusions 1.48 At the beginning of this decade, Ghana stood poised for accelerated growth. The Economic Recovery Program had turned the economy around in a significant fashion. But the fiscal shock of 1992 severely undermined macroeconomic stability, reversed many of the macroeconomic gains that had been made, and seriously dampened the environment for private sector development. Ghana is still trying to

33 13 recover to growth and private investment rates of Ghana's ability to do so, and to accelerate growth, depends in large measure on the sustainability of macroeconomic stability, which is currently precarious. The macroeconomic implications of the fiscal shock clearly argue for (i) a surplus in the narrow fiscal accounts driven by expenditure reductions, (ii) the adoption of sound monetary policies not only for controlling inflation which deters savings and investment, but also for preventing short-term capital outflows. The latter also argues for a cessation of the Government's access to central bank credit. These measures are essential to providing a viable and stable macroeconomic environment for accelerated growth.

34 Chapter 2 POLICY REFORMS AND THE PATTERN OF GROWTH Policy reforms since 1983 have shifted relative prices and incentives towards the tradables sector and have turned the economy around. Although economic activity witnessed its biggest surge during the early years of the ERP (i.e ), aggregate growth performance has averaged 4.7 percent since However, the performance has not been uniform across sectors. Agriculture recorded an average annual growth rate of only 1.9 percent since 1987 and saw its share of GDP fall significantly while services have grown at an average annual rate of 7.4 percent over the same period and account now for 45 percent of GDP, making it the largest sector in the economy. Merchandise exports and imports have grown faster than GDP and with it, wholesale and retail trade have dominated GDP growth under the ERP. The performance of the private sector has similarly been mixed. Private sector share of GDP was around 66 percent in Mining, financial services, and wholesale & retail trade benefited from the liberalized economic environment, experienced strong growth and increased their share of GDP. 2.1 A key objective of Ghana's Economic Recovery Program was to achieve accelerated private sector growth to facilitate a sustained reduction in poverty. This chapter attempts to give an overview of policy reforms and the pattem of growth in Ghana under the ERP, especially since 1987, by following the development in the real sectors of the economy and deternining the contributions made by the private and public sectors. Particular attention is paid to the evolution of private sector growth in Ghana as such a time-series analysis has not been previously done for Ghana. Policy Reforms and Performance 2.2 Ghana's policy refonns since 1983 have reduced the fiscal deficit and inflation, improved infastructure, and moved relative prices and incentives in favor of tradables, in general, and exports, in particular. The key element of fiscal consolidation up to 1991 has been the growth in govenment revenues whose share of GDP rose from 6 percent in 1983 to 13 percent in 1986 and to 16 percent in Higher revenues made it possible to reduce fiscal deficit and at the samne time increase public investment infrastructure, which had collapsed before Inflation fell from 123 percent in 1983 to 40 percent in 1986 and 18 percent in The resutirg inprovements in macroeconomic stability and in infrastructure made it possible for farms and firms to respond to the shift in production incentives induced by policy reforrn. 2.3 Pre-1987 Reforms and Performance. Fiscal consolidation and murnerous nominal devaluations of the Cedi (from Cedis 2.75 per dollar in 1983 to Cedis 90 per dollar in 1986) resulted in substantial real depreciation and significant inprovements in the relative incentives for the production of tradables (see Figure 2.1). Within tradables, relative prices shifted in favor of exportables like cocoa, timber and minerals and against irnportsubstitutes like manufactures as a result of the elimination of import licensing, the liberalization of export restrictions on timber and minerals and administered increases in cocoa producer prices. The real producer price of cocoa more than doubled between 1983 and 1986 (see Figure 2.2). These shifts in relative prices improved the rural-urban terms of trade, given the dominance of primary exports.

35 15 Figure 2.1 World Cocoa Price, Terms of Trade, and Real Exchange Rate Indices, TC-Terms of Trade - World Cocoa Price -X- Real Exchange Rate %0 20 O-. Ir l0 l l0 I' I I l oo % % 0 0% 0% 20% " Source: Bank Staff calculations Note: (i) The real exchange rate has been inverted here so that a decrease in the real exchange rate index reflects depreciation. (ii) The World Cocoa Price is the London price for Ghanaian cocoa. Figure 2.2 Cocoa Terms of Trade and Cocoa Output, { - Real Producer Price (1977 base) Cocoa Output 300D I. 30D *~ O' I I I I I I 0 I AI-I I ISO 2w~ i Source: Ghana Cocoa Board and the Quarterly Digest of Statistics Note: The Real Producer Price for cocoa is obtained by deflating the Nominal Producer Price by the Rural Consumer Price Index (base 1977).

36 As a result, the economy turned around, with exports of cocoa, logs and gold growing very rapidly (see Table 2.1). Overall real GDP growth increased from minus 1 percent per annum before the ERP to around 5.3 percent per annum between 1983 and Total merchandisexports grew at around 22 percent during that penod, with logs and cocoa growing the fastest. The share in GDP of exports rose from about 2 percent in 1983 to 16 percent in 1986 and that of total trade from about 5 percento 32 percent during the same period. Imports also grew rapidly but at a rate lower than export growth with non-oil inports growing at about half the rate of total merchandisexports. Particularly strong during this period was service sector growth which exceeded that of the overall economy (see Figure 2.3). Though government services grew during this period, growth in wholesale and retail trade and in transport services was significant. These services grew because of their complemientarity with exports and imports, which grew rapidly during this period. The growth in such services was no doubt facilitated by increased public investments in roads, rail and ports. Mining was another sector which grew rapidly with the liberalization of the sector, albeit from a very low base. Manuficturingrowth, after an initial spurt due to freer internediate imports, fell off because of substantially lower protection. Table 2.1: Selected Policy and Performance Indicators ( period averages) Relative Price Indicators Real Exchange Rate a (1984=100) Black Market Premia b(%) c Domestic CPI Inflation Rate (CPI) Nominal Interest Rate (end-period 91 -day T-Bill rate) Real Interest Rate d Urban-Rural Terms of Trade '(1977=100) n.a. Performance Indicators ( ) (annual percentage change in volume) Merchandise Exports Cocoa Logs Gold Merchandise Imports Oil Non-Oil Source: Bank staff calculations; Quarterly Digest of Statistics; IMF a/ The real exchange rate is the nominal exchange rate deflated by the ratio of the domestic consumner price index to the international inflation index. An increase in the real exchange rate here reflects depreciation and vice versa b/ Kapur I., M.T. Hadjinichael, P. Hilbers, J. Schiff, and P. Szymczak, 1991, "Ghana: Adjustment and Growth, ," IMF Occasional Paper No. 86, September, Bank staff estimates. c/ Average for ; 1994 is not available. d/ End-of-period nominal interest rate less end-of-period inflation. e/ Urban-Rural terms of trade are calculated here as the ratio of the wholesale price index for manufactured goods relative to the wholesale price index for agriculture, forestry and fishing. The price indices were obtained from the Ghana Quarterly Digest of Statistics.

37 17 Figure 2.3 Sectoral Growth Rates 12.0% U10.0% % x -Eg t0 t 3 L * % 0~ 0~~~~~ (u~ 6.0%' 4.0% 2.0% 0.0% -4.0% 2.5 Refonrns and Performance during The ternms of trade deteriorated considerably after 1987 with the decline in the world price of Ghanaian cocoa. The dollar price of Ghanaian cocoa fell at an annual rate of over 7 percent over and started to recover only in 1993 (see Figure 2.2). Consequently, the cocoa sector has not benefited consistently from the nominal increases in producer prices for cocoa. While real producer prices did rise during , they fell consistently over But despite the deterioration in the terms of trade, there was little real depreciation (see Figure 2.1). Growth in the inflow of net foreign aid during this period prevented this and dampened the relative incentives for exports, even as tariffs were reduced. 2.6 This period also saw the liberalization of investment regulations and the rationalization and reduction of direct and indirectaxes. The reforms also saw the price of labor fall relative to capital, as interest rates rose with the reform of the financial sector. The period average real interest rates, while still slightly negative during this period, were substantially less so than during the earlier period. In fact, real interest rates had become positive by 1991, boosting incentives for savings and investment. However, the faihire to liberalize the marketing of agricultural inputs and of food crop outputs continued to undermine incentives for investment and production in the agricultural sector (except in cocoa). Similarly, the continue dominance of the manufacturing sector by inefficient and protection-dependent public enterprises undermined growth in the sector. Thus only the liberalized mining sector and the service sector benefited from the reforms. 2.7 During this period, average growth was around 4.8 percent, somewhat lower than during the previous period, but still quite robust and substantially higher than the growth rate in the pre-erp period. The service sector maintained its growth, with transport services taking the lead followed by wholesale and retail trade services. Manufcturing growth fell off sharply, though there is evidence of growth in micro and small-scale enterprises in the ural sector which is not captured by the national accounts. Agriculture coninued to grow at an average of around 2 percent per annum. The mining sector grew at nearly twice the rate of the previous period. Merchandisexports continued to grow faster than GDP -- responding to the sustained favorable policy enviromunt for exports - though at a rate of 9 percent, which was lower than the spurt between 1983 and Exports of gold grew very fast during this period. Non-traditional exports also grew much faster, almost

38 18 doubling between 1988 and 1989 (see Table 2.2). The growth in non-traditional exports reflected growth in export demand, improvements in the duty drawback system, and increased foreign exchange retention rights. Non-oil imports continued to grow at over 10 percent annually, slightly more than the rate of growth of merchandisexports. By 1991, the share of trade in GDP increased further to 35 percent. Table 2.2: Non-Traditional Exports By Product, (in thousand USS) (Jan.-Sep.) Agricultural Products 21,169 28,781 33,040 22,059 26,134 31,500 Horticulture Products 2,527 4,936 8,386 6,769 7,703 6,291 Fish& Seafood Products 13,236 21,592 20,047 11,091 14,755 8,855 Other Products 5,406 2,253 4,607 4,199 3,676 15,354 Processed & Semi-Processed Products 13,354 33,106 28,632 44,888 42,989 48,442 Wood Products 3,764 5,546 6,237 15,512 18,807 20,512 Aluminum Products 2,536 9,793 5,475 4,940 3,607 3,374 Other Products 7,054 17,767 16,920 24,436 20,575 24,556 Handicraft Items ,471 2,576 1,786 Total Non-Traditional Exports 34,719 62,342 62,551 68,418 71,699 81,728 Source: Ghana Export Promotion Council 2.8 Post-1991 Reforms and Performance. More recently, there have been both setbacks and progress in furthering policy reforms. The setback was the reversal on the fiscal front in 1992 (see Chapter 1). But significant progress was made in many other areas. Exchange rate liberalization went further apace, leading to the elimination of the very large black market exchange premium that had preceded the ERP and had persisted substantially through most of the eighties. Financial sector reforms continued to keep real interest rates positive. Domestic marketing in cocoa was liberalized in 1992 with the entry of private traders and substantially higher nominal producer price rises were given in 1994 so as to pass on a larger share of the cocoa export price to the farmers and increase incentives for investment in, and production of, cocoa. Cocoa output rose to an ERP-period peak of almost 320,000 metric tons in 1992, but the weakness in international cocoa prices, the removal of subsidies on farm inputs, and a higher incidence of crop disease have contributed to fluctuations in cocoa output. 2.9 During this period, average GDP growth slowed down slightly to 4.2 percent with services still showing robust and sustained growth (see Figure 2.3). Among the service subsectors, wholesale and retail trade grew strongest, at more than double the national growth rate. Growth in mining also continued to be strong but agricultural growth showed no recovery, slipping further to average only 1. I percent during this period. The decline in agricultural growth was largely due to poor weather conditions in 1992 due to which agricultural output actually dropped by 0.6 percent. And in 1994, the ethnic conflict in the north has also disrupted agricultural production. While both merchandise exports and imports have also slowed down, their growth rates are still above the GDP growth rate. As a result, the share of total merchandise trade in GDP has climbed further to 54.5 percent in The Changing Sectoral Pattern of Growth. This differential pattern of sectoral performance has significantly altered the relative importance of the sectors in the Ghanaian economy (see Figure 2.4). While in , agriculture accounted for almost 50 percent of GDP, its share fell to just over 40 percent in

39 At the same time, the share of services increased from about 38 percent to about 45 percent of GDP. The share of industry increased only nominally from 13 to 14 percent. Figure % 50% l. 40% 30% Sectoral Shares of GDP (1975 prices) M U~~~~~~~~~~~~~~~~~~~~~~. U~~~~~~~~~~~~~~~~~~~~~~~l e ~. :- 11cei 20% c -- a 0% 2.11 Since most of the growth has occurred from the service sector which is predominantly in the private sector, this raises the interesting question as to the contribution of the private sector in promoting growth in Ghana. The following sections, therefore, attempt to estimate the share of the private sector in domestic value-added and trace its evolution during the later years of the ERP. Estimating Private Sector Value-Added 2.12 In the absence of official data on private and public sector contributions to GDP, one has to estimate private and public sector value-added using available economic indicators. Paul (1990) made an earlier effort to decompose Ghana's GDP into public and private sectors contributions for the period using several data sources-the 1984 Population Census, the 1987 Industrial Survey, reports from various ministries and associations, as well as "informed judgments about the economy."9 The results suggested that the private sector contributed about 64 percent of total GDP in 1986 (see Table 2.3) Employing similar methods, the private sector share of GDP has been estimated to be around 65 percent in Cote d'ivoire, 73 percent in Madagascar, 75 percent in Kenya, and 87 percent in Senegal.' 0 More recently, the European Bank for Reconstruction and Development (EBRD) estimated that the private 9 Samuel Paul, 'Assessment of the Private Sector," World Bank Discussion Papers, No 93, June Figures are taken Private Sector Assessments for the various countries and reported in C6te d'ivoire: Private Sector Assessment, World Bank Report No IVC. The reported nunbers refer to the time period

40 20 sector contributes to no more than half of the GDP in all but a handful of formerly centrally planned economies." The private sector was estimated to be the strongest in the Czech Republic, accounting for about 65 percent of GDP in Table 2.3: Private and Public Sector Contribution to GDP in 1986 (in percent of GDP- percent of sub-sector in parentbeses) Total Private Sector Public Sector Total GDP Agriculture Cocoa (60) 3.1 (40) Forestry (75) 1.4 (25) Agr., Livestock, Fishing (85) 5.4 (15) Industry Mining (30) 0.7 (70) Manufacturing (75) 2.2 (25) Utilities (1) 0.9 (99) Construction (50) 1.1 (50) Services Transport (60) 1.7 (40) Wholesale & Retail (95) 0.5 (5) Financial Services (20) 6.8 (80) Govermment Serices (0) 12.2 (100) Other Services (100) 0.0 (0) Source: Derived from Paul, S. (1990) ibid., World Bank data, and Bank staff estimates 2.14 The methodological problems and data deficiencies in doing such an exercise are always substantial, as the EBRD report also acknowledges. Given these caveats, however, this chapter attempts to do so using various economic indicators for different sectors, such as private sector share of total employment, total assets (e.g. in the financial sector), or gross output (e.g. in mining). The results provided here are derived by first estimating the share of private sector value-added within each sub-sector. Applying the breakdown on a sub-sectoral basis to the data provided in the national accounts then yields the aggregate contribution of the private sector. Changes in total private sector contribution to GDP can thus vary either as a result of higher private sector shares of output within sub-sectors, or as a result of faster growth in sectors that are dominated by the private sector. Both factors play an important role in understanding the contributions that the private sector makes to economic growth The estimates derived below are likely to understate the actual growth in private sector activity for two reasons. First, the growth of the informal sector is not adequately captured in the national income accounts and since the informal sector activity is wholly in the private domain, private sector shares based on the national income accounts will be under-estimated.' 2 Second, the time series tracking the development of the private sector is based on the assumption that the share of the private sector value added has remained by and large unchanged within sub-sectors, except where meaningful indicators 11 European Bank for Reconstruction and Development, Transition Report, October Ghana's national accounts data in themselves may not reflect the true value-added by individual sectors of the economy. For example, the value-added by manufacturing is based on a sample of firms taken in the mid-70s which concentrated on fimns with 10 or more employees. While exiting firms have been replaced, the strong growth of microenterprises that has occurred since the beginning of the ERP is not likely to be reflected in official data.

41 21 suggest otherwise.' 3 Thus, when the growth of the private sector was not reflected in available indicators, such as private sector share of total assets, the conservative assumption was made that the private sector share did not grow within the sector. Private Sector's Share of Output 2.16 Several sectors have experienced a shift towards greater private sector participation since the introduction of the ERP. While the Government's divestiture program, which began in 1987, has led to a reduction in the number of state-owned enterprises in several sub-sectors, three sub-sectors may have benefited the most. The introduction of the 1986 Mining Law resulted in a sharp increase in private investment, both domestic and foreign, in the mining sector. Agricultural sector reforms have reduced state involvement in agricultural production through the closing and divestiture of cocoa plantations. The financial sector has benefited from sectoral reforms that began in 1987 and have led to the opening of several private bank and nonbank financial institutions Because of the dominance of the private sector in agriculture, which accounts for roughly half of GDP, total production has always been dominated by the private sector. At the beginning of the ERP period, the private sector accounted for about 64 percent of total value-added. This share did not rise significantly until the late 1980s when a more aggressive divestiture program and the implementation of sectoral reform programs supported growth in the private sector. By 1994, private sector's share of GDP had increased by a modest 2 percentage points over the level of 1987 (see Table 2.4). Table 2.4: Private Sector Share of Value-Added, (in percent of total GDP) Total Agriculture Cocoa Forestry Agr., Livestock, Fishing Industry Mining Manufacturing Utilities Construction Services Transport Wholesale & Retail Financial Services Government Services Other Services Source: Bank Staff estimates 13 Indicators of annual change in private sector share of value-added were available for cocoa, mining, and financial services.

42 Agriculture. The private sector's share in agricultural production has increased only modestly since the implementation of sectoral reforms in While private activities have been dominant in all sub-sectors, government control of cocoa marketing-which is included under agriculture in the national accounts statistics-has limited the private sector's share of output in the this particular sub-sector.' 4 Private sector share of output in the other agricultural sub-sectors has remained constant and the participation of the public sector limited The importance of agriculture for assessing the activities of the private sector cannot be overlooked. Private sector agricultural production accounted for over 40 percent of total GDP at the beginning of the ERP period, but average growth rates in the sector have been weak. While total GDP has grown at an average annual rate of almost 5 percent, agricultural output rose at an average rate of less than 2 percent. This means that private agricultural production has been a drag on overall private sector growth. As a result, its share of total GDP has fallen to less than 33 percent in Within this sector, growth of food crops, livestock, and fishing output has been particularly slow, averaging only 1.6 percent per year, while accounting for the largest share of agricultural output. Private cocoa value-added has benefited from the divestiture of state-owned plantations, as well as from a shift away from state marketing Industry. T'he fact that private agricultural output has declined significantly relative to total GDP during the ERP period meant that private sector share in the other sectors had to rise just to maintain the private sector's overall position in the economy. However, this was not the case in the industrial sector, where the private sector's share of output has remained virtually unchanged throughout the ERP period. Despite faster growth immediately following the implementation of the ERP, private industry's share of total output has stagnated since 1987 at just over 8 percent (see Table 2.4) Manufacturing, the largest industrial sub-sector, is dominated by the private sector, but poor growth performance led to a fall in the private sector share of total industrial production from a peak of 62 percent in 1986 to 60 percent in Private manufacturing's share of total output was just over 6 percent in The private sector has made substantial progress in the mining sector following the liberalization and the influx of foreign investment that began with the implementation of the 1986 Mining Law. However, even though the share of private sector output within the mining industry is estimated to have grown from 32 percent in 1987 to 65 percent in 1994, the important of the sector to total value-added is 14 The private sector share of value-added in the cocoa sub-sector has been determined as a residual after subtracting government marketing and production from the total. Government marketing as percentage of sectoral value added has been derived from the accounts of the COCOBOD. Private Sector Share of Value Added in Cocoa Production (in percent) Source: COCOBOD and Bank staff estimates

43 23 not big enough to give private industrial production a boost. 5 Private mining's share of total output was still less than I percent in Services. The private sector has gained the most in the services sector, where its share of total services output grew from 40 percent at the beginning of the Economic Recovery Programn to 46 percent in 1987 to over 51 percent in Most significantly was the strong growth in the wholesale and retail sector. This sub-sector is dominated by small private enterprises, limiting the public sector to only about 5 percent of sectoral output. While the private sector's share of output in this sub-sector has probably remained constant over the past several years, an average annual growth rate of over 10 percent between 1987 and 1993 has made this the most significant service sub-sector. Private sector growth was also partially due to the reforms in the financial sector that were initiated with the Financial Sector Adjustment Program in 1987 and allowed private bank and nonbank financial institutions to expand their business.' As a result of strong growth in wholesale and retail services, private services now account for almost one-fourth of GDP (see Table 2.4). This particular sector appears to have benefited the most from the liberalized economic environment. It has to be noted, however, that the service industry in Ghana consists predominantly of small and micro enterprises that operate in the informal and semi-informal economy. While this sector generates a large number of jobs, investment in physical capital is likely to be low. Is Growth Public Sector-Led? 2.25 The above analysis suggest that not only the lion's share of the value-added but also most of GDP growth comes from the private sector. The common perception that growth in Ghana under the ERP was public sector-led is explicable more in terms of investment shares than GDP shares. Public investment grew from a low of about 4 percent of GDP during to 7.6 percent in and to 10.0 percent in During the same period, private investment increased from a period average of 4.0 percent of GDP in to 5.4 percent in before slipping to 4.5 percent in is The share of private sector output is estimated using available information on mining output by frms and share of public ownership for each fir. Private Sector Share of Value-Added in Mining (in percent) Source: Bank staff estimates. 16 One indicator used in deriving the relative share of the private sector was its share of total assets in the banking system. However, this was supplemented by information about the share of public sector assets in the insurance industry. Private Sector Share of Value-Added in Financial Services (in percent) InBanking In all Financial Services Source: Bank Staff estimates.

44 It is also true that private investment in itself is too low to fully explain the growth in private sector production. Excluding any complementarity between public investment and private growth, this analysis would suggest an unusually low ICOR of 1.3 since Clearly, private investment cannot be the only driving force behind private sector growth. Substantial complementarities between public investment and private sector growth explain part of private sector growth. Much of the increase in public investment has been in restoring basic infrastructure which is critical for private sector growth. It is very likely that without substantial public investment, private sector growth would not have been nearly as strong. But it is also important to recognize that private investment where it occurred, as in agriculture and in building construction, are under-estimated in the national accounts data. For example, the planting of cocoa trees by small- and medium-sized cocoa farmers and construction activities by small firms and individuals, are not captured. Much of private sector growth has also come from sectors that require relatively low investments as is obvious from the rapid growth in retail and wholesale services It is then possible to modify the assertion that growth in Ghana is public sector-led by saying that investment has, by and large, been public sector-led. Although growth in total investment has come from the private sector prior to 1992, one has to bear in mind that this reflected a recovery from extremely low levels during the mid-80s (see Chapter 1, Table 1.1). Since 1992, public sector investment has significantly exceeded private investment. While the private sector accounts for two-thirds of production, it accounts now for less than a third of total investment. V. Conclusions 2.28 Policy reforms undertaken by Ghana under the Economic Recovery Program have shifted relative prices towards the tradables sector, inducing increased export growth and international trade. Complementary services, particularly wholesale and retail, have benefited from this. The private sector is the main contributor to GDP and the main factor behind GDP growth under the ERP. However, the share of the private sector in GDP has not changed significantly since Low rate of agricultural growth has limited aggregate GDP growth as well as private sector growth in Ghana. The private industrial sector, in particular manufacturing, has not yet fully contributed to economic growth. Private investment in manufacturing needs to be encouraged, through an adequate enabling environment, to generate employment and a solid private industrial base. Following the liberalization of the trade regime, the manufacturing sector was pressed to improve efficiency and reduce production costs. This environment, however, has now persisted for a number of years and the initial adjustments are likely to have occurred. Higher investment and growth in private manufacturing, however, may remain elusive unless it is supported by a stable macroeconomic environment and the availability of adequate financing. A stronger signal of private sector support needs to be sent through an accelerated privatization of strategic state-owned enterprises and state-owned banks.

45 Chapter 3 GROWTH AND POVERTY REDUCTION Ghana made sizable gains in reducing poverty from 37 percent in to 32 percent in Growth, which has been distributionally-neutral, has driven the reduction in aggregate poverty. The reduction in poverty has been widespread and broad-based, benefiting mainly the rural areas, and the more vulnerable segments of the population --female-headed households. Most of the reduction in rural poverty is due to growth in non-farm self employment, particularly trading services which are related to the surge in exports and imports. But despite the gains in poverty reduction, at current growth rates, the average poor can expect to escape poverty only in ; and the average extreme poor only in This needs to be expedited by macroeconomic and sectoral policies that promote faster economic growth, especially agricultural growth,. 3.1 International experience clearly suggests that growth is a necessary, though not sufficient, condition for poverty reduction. 17 No country has achieved sustained poverty reduction without economic growth. Growth attacks poverty by providing economic opportunities for the poor to utilize their labor and generate higher personal incomes; at the same time, growth forms the basis for larger tax revenues from which to finance the public expenditures on the social sectors to enable the poor to take advantage of the economic opportunities. In particular, growth that exploits the comparative advantage of a country, using the more abundant factor of production (i.e. labor), enables the income gains from growth to be 'pro-poor'. A growth strategy that fits this criterion is private sector growth under a liberal and neutral incentive regime. Ghana embarked upon such a growth strategy under the Economic Recovery Program (see Chapter 2). 3.2 This chapter provides an analysis of the levels and patterns of poverty in Ghana and investigates the association between growth and poverty reduction in rural Ghana. 1 It shows that Ghana has indeed had a significant reduction in poverty between and , which has been broad-based and has benefited more the vulnerable segments of the economy and the population. The Definition of Poverty 3.3 The treatment of poverty in this chapter relies solely on monetary measures of living standards. The welfare of the population is measured by total household expenditures per capita. This is not to imply that other dimensions of poverty such as nutrition, literacy, and health outcomes are not important.19 The 17 See, for instance, The World Bank, 1990, "The World Development Report 1990," Oxford University Press. Is The analysis in this chapter is based on the Ghana Living Standard Surveys (GLSS) of and which were conducted by the Ghana Statistical Service. These provide, for the first time, the means by which an investigation of the changes in poverty can be carried out over a period of time. Although there was an intermediate survey in , the analysis is restricted to the initial and terminal years so as to have a broader range of years over which to draw conclusions. The actual periods of the three surveys were: September 1987-August 1988 (GLSSI); October September 1989 (GLSS2); and October 1991-September 1992 (GLSS3). 19 These are discussed in the Extended Poverty Study. See, for instance, Ghana Statistical Service, 1995, "The Pattern of Poverty in Ghana, ," mimeo, January.

46 26 poverty line used below is defined as two-thirds of the mean per capita expenditure in the first year of the survey ( ), expressed in constant May 1992 Accra prices. This translates into a poverty line of Cedis 132,230 per capita per year.20 The discussion below refers only to two commonly-used poverty indices -- the headcount index and the depth of poverty index. The former estimates the proportion of the population below the poverty line while the latter the proportion by which the mean expenditures of the poor fall below the poverty line. For the purpose of comparability over time, the expenditure data has been corrected in the base year for the change in recall period between the surveys. Aggregate Poverty 3.4 The poverty profile reveals that the incidence of poverty fell between and from 36.9 percent to 31.5 percent. The depth of poverty also declined -- the mean expenditures of the poor were 11.9 percent below the poverty line in and 8.1 percent below the poverty line in These indicate a major success in efforts to combat poverty and, as we shall see later, reflect the impact of growth in pulling up the poor An examination of the sources of income for the poor and the non-poor reveals that farm income is the dominant but declining source of income for both groups, non-farm self-employment is the second most important and growing source of income for both, and wage employment is the third most important source of income (see Table 3.1). 24 But there are interesting differences between the income sources of the poor and non-poor. (i) Farm income provides a smaller share of total income for the non-poor relative to the poor. (ii) Non-farm self-employment contributes a larger share to the income of the non-poor than that of the poor. And (iii) wage employment income contributes a higher but declining share of total income for the non-poor compared with the poor. 3.6 It is important to note that the aggregate reduction in poverty and the changes in the sources of income are largely reflective of the changes across most disaggregated groups of the population. This suggests that the reduction in poverty has indeed been broad-based, touching the vast majority of the population, as is indeed evidenced by the sizable reduction in rural poverty. 20 Coulombe, H. and Andrew McKay, 1995, "Pattern and Changes in Poverty in Ghana, 1987/ /92: An Assessment of Robustness," mimeo, Department of Economics, University of Warwick, January GLSSI has a recall period for food expenditures of 14 days whereas GLSS3 has a much shorter 2-3 day recall period. Since longer recall periods are associated with recall error, the GLSSI estimates of food expenditures were adjusted upwards by 2.9 percent for every additional day of longer recall period, up to the seventh day, compared with GLSS3 (see Scott and Amenuvegbe, 1990, "Effect of Recall Duration on Reporting of Household Expenditures - An Experimental Study in Ghana", Social Dimensions of Adjustment in Sub-Saharan Africa Working Paper No. 6, The World Bank, Washington, D.C. and Coulombe, H. and Andrew McKay, 1995, ibid.) 22 The estimates for the annual expenditures on subsistence consumption used for the poverty profile assume that the surveys were conducted uniformly throughouthe year in all localities. 23 The analysis of poverty over and is over a later time slice of the ERP period and, therefore, underestimates the gains in poverty reduction that Ghana has made during the ERP. 24 Measures of income levels in household surveys are generally unreliable. This is one reason why the welfare measure is based on real expenditures which are generally more accurately measured. But income shares are likely to be less subject to measurement errors than the absolute levels of income. Moreover, to date, there is no evidence to suggest that there are biases in income shares. It is, therefore, reasonable to use income shares to classify households according to their main source of income and to establish whether there have been changes over time.

47 27 Table 3.1: Sources of Household Income, by Poor and Non-Poor Income from Poor Non-Poor Wage Employment Farm Self-Employment' Non-farm Self-Employment Actual and Imputed Rent Remittances Other Source: GLSSI and GLSS3; Bank staff calculations a/ Farm Self-Employment covers the cultivation of crops, the processing and transformation of crops, and the herding and production of livestock-related products. Urban-Rural Poverty 3.7 The reduction in poverty during and is a rural phenomenon. Rural poverty diminished substantially ftom 41.9 percent of the rural population in to 33.9 percent in (see Table 3.2). This suggests that the poverty reduction has been widespread, given that the rural areas account for about two-thirds of the total population. However, poverty increased dramatically in Accra from 8.5 percent in to 23.9 percent in But poverty levels in Accra remain well below the national average. In terms of overall urban poverty, the increase in poverty in Accra was compensated by the reduction in poverty in the other urban areas. Consequently, urban poverty showed only a marginal fall during the period. Table 3.2: Poverty Profile by Urban and Rural Areas Area Share of Population Incidence of Poverty Depth of Poverty Contribution to National (Po) (Pi) Poverty (%) Urban Accra Other Urban Rural ALL Source: GLSSI & GLSS3; Bank Staff calculations 3.8 Poverty itself is also predominantly a rural phenomenon. About 75 percent of the national poverty can be attributed to the poverty in the rural areas which is greater than the share of the rural population. Rural poverty is also much higher than the national average. However, the poverty differential between the rural and urban areas appears to have narrowed from 14.5 percentage points in to 7.4 in The large fall in rural poverty strongly suggests that economic growth has effectively 'pulled up' the rural poor who have gained proportionately more during the period. 25 Further analysis is ongoing to explain the dramatic increase in poverty in Accra. It is possible that this was due to the higher inflation in Accra relative to the rest of the country and to the consequent erosion of real wages in fonnal employment. The greater poverty does not, however, appear to be explained by the income levels of the in-migrants to Accra which are actually higher than those of the resident population.

48 What characterizes the successful poverty outcomes in the rural areas? A look at the sources of income of rural households is revealing and supports the results from the analysis of the income sources for the poor and non-poor (see Table 3.3). The decline in rural poverty appears primarily due to the large 50 percent increase in the share of non-farm self-employment income. As for remittances, they clearly have an urban bias with their contribution to urban households' income 2-3 times that to rural households. 26 Wage employment provides about 8 percent of rural households' income while showing diminished importance in the urban areas. Table 3.3: Sources of Household Income, by Urban and Rural Areas Income from Accra Other-Urban Rural Wage Employment Farm Self-Employment Non-farm Self Employment Actual and Imputed Rent Remittances Other Source: GLSSI and GLSS3; Bank staff calculations 3.10 Poverty across Rural Economic Groups. While the major rural economic groups benefited from the reduction in poverty, the change in poverty was not uniform (see Table 3.4). The largest reductions in poverty were for the non-farming households -- households which derived 50 percent or less of their total income from agriculture -- for whom poverty fell by more than 25 percent. Farmers, who comprise more than 50 percent of the total population and percent of the rural population, saw a 15 percent decrease in their poverty during the period. Among the farmers, poverty decreased substantially for both the cocoa and non-cocoa farrners. But despite the reduction in poverty incidence, the farmers as a whole remain the poorest group with the incidence of poverty among them about 15 percent greater than the national average. Table 3.4: Poverty Profile by Rural Economic Group Share of Population Incidence of Poverty Depth of Poverty Contribution to (Po) P] 1X) National Pov % Farm Households' Cocoa Farmersb Non-Cocoa Farmers Non-Farm Households Source: GLSSI & GLSS3; Bank Staff calculations a/ Farm households are defined as those households who derived more than 50 percent of their total income from farming activities. See Note to Table 3.1 for definition of Fanning. hl Cocoa farm households are defned as those who derived more than 50 percent of crop revenue was from cocoa The sources of income for the two farm groups appear quite similar to those above (see Table 3.5). Farm income is the predominant source of income for all. All groups also have a growing, albeit small, share of income coming from non-farm self employment. There also appears to be reduced reliance on wage employment as a source of income. 26 Remittances are generally higher for non-farm households, including those in rural areas for whom the share of income from remittances is comparable to those of urban households. Farm households have much lower share of income from remittances and, as they comprise more than 50 percent of the total population and percent of the rural population, they bring down the share of remittances in income for the whole rural population.

49 29 Table 3.5: Sources of Household Income, by Rural Economic Group Income from Cocoa Farmers Non-Cocoa Fanners Non-Farm Households Wage Employment Farm Self-Employment Non-farm Self Employment Actual and Imputed Rent Remittances Other Source: GLSSI and GLSS3; Bank staff calculations Gender and Poverty 3.12 The inclusive nature of the poverty reduction in Ghana is also borne out by the dramatic reduction in poverty among female-headed households (see Table 3.6).7 The incidence of poverty among them decreased substantially during the period from 39 percent to 29 percent, a fall of almost 25 percent. At the same time, the incidence of poverty among male-headed households fell from 36 percent to 33 percent, or by about 8 percent. This has two implications: first, most of the aggregate poverty reduction comes from the fall in poverty of female-headed households; second, the poverty incidence differential between the two sexes has been reversed. In , female headed households had a poverty incidence 2 percentage points higher than their male counterparts; in , their poverty incidence was 5 percentage points lower. Table 3.6: Poverty Profile by Gender of Economic Head of Household Gender Population Share Incidence of Poverty Depth of Poverty Contribution to (Po) (PO) National Poverty _ Male Female Source: GLSSI & GLSS3; Bank Staff calculations 3.13 The explanation for the dramatic improvement in the standard of living of female-headed households -- both in absolute terms and relative to the male-headed households -- probably lies in 3 major differences in their pattern of sources of income (see Table 3.7). (i) The dependence of female headed households on farm income is lower than that for their male counterparts, and declining. The share of farm 27 Refers to the economic head of household. Two clarificatory points need to be made here. First, the 'economic head' of the household is defined as the highest wage earner, if wages are equal, then the person who worked the most hours is taken as the economic head. Ideally, the economic head should have been defined on the basis of contribution to total income but this is not possible since many incomes, particularly, agricultural income is lumped together as household's income. The GLSS also provides a second definition of head of household 'as identified by the household'. The overlap between the two definitions is percent and the use of the GLSS, in fact, leads to a bigger reduction in poverty among the female-headed households (from 41 percent in to 29 percent in ). The use of the 'economic head' definition, therefore, provides a more conservativestimate. Second, there are changes in the questionnaires in the definition of who is a household member. GLSSI/2 consider a person as part of the household if the person has been away from home for no more than 9 months; in GLSS3, the allowable period of absence is 3 months. It may appear that this will lead to an upward bias in the proportion of female-headed households in GLSS3 if one assumes male members as more likely to be away from home but in reality this does not appear to make any difference for two reasons: (i) the period of absence criteria in the GLSS applies to household membership but not to the household head as 'identified by the household', and (ii) the proportion of female-headed households has remained fairly constant between the two periods, regardless of which definition of head of household one uses.

50 30 income in the total household income for them fell from 42 percent to 29 percent between and compared with a much smaller fall for the male-headed households from 58 percent to 55 percent. (ii) Female-headed households rely predominantly on non-farm self employment as their major source of income. In , the share of non-farm self employment income in total household income was 40 percent for the female-headed households and 21 percent for the male-headed households. This was higher than the shares in by 42 percent and 35 percent, respectively. And (iii) female-headed households have a higher and increasing reliance on income from remittances compared to the male-headed households. Whereas for the latter, the share of remittances in total income stayed at around 3 percent, it increased for the female-headed households from 13 percent to 14 percent between and Table 3.7: Sources of Household Income, by Gender of Economic Head of Household Income from Male Femsle Wage Employment Farm Self-Employment Non-farm Self-Employment Actual and Inputed Rent Remittances Other Source: GLSSI and GLSS3; Bank staff calculations What Explains the Reduction in Poverty? 3.14 The reduction in aggregate poverty is due largely to economic growth. Between and , GDP growth averaged close to 5 percent per annum. In per capita terms, this translated into 2 percent growth per annum. The size of the impact on poverty reduction is, of course, contingent on the nature and pattern of the growth itself. In particular, growth induces a change in the income distribution; improvements in income distribution reduce the number of the poor below the poverty line and vice versa. This is evident from a decomposition of the change in poverty into its growth and redistribution components. 2 The decomposition indicates that the contribution of the growth in mean expenditures to poverty reduction was more than 85 percent (see Box 3.1). What is also interesting is that growth has been distributionally-neutral. As a result, the redistribution component has not been antithetical to growth in reducing poverty. This is very significant as in many other countries, the reduction in poverty is typically accompanied by a worsening in the income distribution. 28 See Datt, G. and M. Ravallion, 1992, "Growth and Redistribution Components of Changes in Poverty Measures", Journal of Development Economics, 38, pg

51 31 Box 3.1: Decomposition of Poverty Changes: Some International Comparisons How much of any observed change in poverty can be attributed to a growth in living standards as distinct from a change in the distribution of living standards? The seminal paper by Datt and Ravallion (1992) has shown that any change in poverty can be decomposed as the sum of a growth component (the change in poverty from a change in mean income/expenditures, keeping the Lorenz curve, i.e. income distribution, constant), a redistribution component (the change in poverty from a change in income distribution, keeping mean income/expenditures constant), and a residual (the interaction between the growth and redistribution effects). The table below presents the decompositions for Ghana and for a few other countries and demonstrates a variety of country experiences. Three key messages come out - (i) Nowhere has poverty reduction taken place without growth in mean incomes and expenditures. In CMte d'ivoire, where economic decline was acute, poverty increased dramatically despite partial mitigation by an improvement in income distribution. (ii) In Ghana, most of the reduction in poverty caine from growth, as also borne out by the experiences of India, Brazil, Indonesia, and Kenya. This is contrary to the experience of Colombia where most of the poverty reduction came from improvements in income distribution. And (iii) in Ghana, growth was distributionally-neutral, the redistribution component being close to zero. This is contrary to the experience of, say, Brazil and Kenya, where the redistribution component is large and positive, implying a worsening of income inequality concomitantly with growth, which partly dampened the growth-induced reduction in poverty. Country Period Average Real GDP Change in Poverty Growth Redistribution Residual Growth Rate (%) (Headcount Index)c Component Component Ghana India (rural) Brazila Colombia (urban)b C6te d'ivoireb Kenyab Indonesia (rural)b a/ from Datt and Ravallion (1992), ibid. bl from Demery, L., B. Sen and T. Vishwanath (1994), "Poverty, Inequality, and Growth," mimeo, November; see the original sources therein. c/ various poverty lines; hence not strictly comparable The marked reduction in rural poverty appears to be due to growth in non-farm self-employment which has been shown to be a source of growing income for rural communities. This is also supported by the sectoral pattern of rural self-employment. Between and , employment in farming for all individuals fell from 89 percent to 87 percent; non-farm self-employment showed a corresponding small 29Fu eoe esetrlpte but significant increase from 11 percent to 13 percent (see Table 3.8). Furthermore, the sectoral pattern of growth in the National Accounts is confirmed by GLSS data on income and on individual employment. While non-farm self-employment activities cut across all real sectors, they are essentially service-based. Thus, in , almost 54 percent of all working rural self-employed reported working in the service sector, with 50 percent in wholesale and retail trade; by , this had increased to 62 percent and 55 percent respectively. The share of manufacturing employment -- of the micro- and small-scale type -- declined from 37 percent to 31 percent. Non-farm agricultural employment -- comprising forestry, logging, and fishing -- also fell from 6.6 percent to 3.9 percent. This fall in non-farm agriculture coupled with the declining share of farming activity suggests contraction in the share of agricultural employment, a finding which is consistent with the declining share of agriculture in the national income accounts (see 29 The sector of employment is identified in the surveys by the respondent and provided with an International Standard Industrial Classification (ISIC) code. Subsectors are grouped together in Table 3.8 to provide for an adequate number of individuals in each group.

52 32 Chapter 2). The importance of non-farm activity is also reflected in the sizable contribution to the reduction in rural poverty by the non-farm rural households which are engaged primarily in non-farm activities The growth in domestic trading then appears to be driving the growth in employment, incomes and expenditures for the rural groups. During , National Account estimates indicate that services grew by almost 8 percent, and that wholesale and retail trade grew by more than 10 percent. The growth in domestic trading is, in part, due to the surge in both imports and exports during the ERP period and, in part, a correction for the excessively depressed levels of trading activity before the ERP. This was facilitated by the enormous improvement in transport infrastructure during the same period. Table 3.8 : Sectoral Pattern of Rural Self-Employment, b Individual Sector Farm Self-Employment Non-Farm Self-Employment Pattern of Rural Non-Farm Self-Employment Non-Farm Agriculture a Industry Mining Manufacturing Electricity & Construction Services Wholesale & Retail Trade Transportation & Communication Cormnunity, Social, & Other Services Source: GLSSI & GLSS3; Bank staff calculations a/ Covers forestry, logging, and fishing. The sumn of Farm and Non-Farm Agriculture comprises the agriculture sector as defined in the National Accounts. Looking to the Future 3.17 Despite the gains that Ghana has made in reducing poverty, poverty remains a serious and extensive problem. With 31 percent of the population with expenditures of less than US$25 per month, that translates into almost 5 million poor people. When can the average Ghanaian poor expect to emerge out of poverty? Box 3.2 suggests that the year may prove to be the cross-over year for the average poor if current growth rates are maintained and if no other economic conditions become a binding constraint. But for the extreme poor -- those with expenditures one-half of the poverty line -- poverty alleviation is still a long way off in ! Ghana can, of course, expedite its riddance of poverty by accelerating growth and by targeting better its public social spending which, as Chapter 4 shows, is currently largely regressive in character. At the same time, the regressivity of public social expenditures and the low levels of human capital attainment suggest that human capital constraints can well become binding to further growth (if not already so). This means that the prospect for poverty reduction to continue apace at past trends may not materialize unless human capital improvements are made.

53 33 Box 3.2: How long for Poverty Eradication for the Average Poor? Despite the growth-induced reduction in poverty, poverty still remains severe with 31.5 percent of the households in under the poverty line. This raises two interesting questions: (i) How long will it take for the average poor to climb over the poverty line at current rates of growth of per capita income and can the human capital base of Ghana sustain such projections? and (ii) Do the actual poverty numbers for compare with such a projection for it based on numbers? First, the arithmetic of the projection. Let x be the income of the average poor, and z the poverty line. If per capita incomes grow at rate g per annum, then the average poor will cross over the poverty line in T years, where T is given by: x (I+g)`- z. Therefore, T = {In(zlxy{ln(l+g)}, where In represents the natural logarithm. Implicit here is the assumption that growth is distribution-neutral - i.e., changes in the mean income do not affect the Lorenz curve. The poverty profile uses a poverty line of Cedis 132,230 per capita per annum in May 1992 Accra prices. This is two-thirds of the mean per capita expenditures in By this criterion, 31.5 percent of the population were poor in with a mean per capita consumption of Cedis 98,301 which is 74.3 percent of the poverty line. With a population growth rate of 3 percent per annum and national income growth rate of 5 per cent per annum, income per capita will grow at 2 percent per annum. If the average poor's income also grows at this rate, then upon applying the above formula, it will take the average poor approximately 15 years (from ) to cross the poverty line, which gives year as the cross-over year. However, for the average extreme poor - with expenditures below one-half of the poverty line -- it will take another 46 years (from ) to cross the poverty line. This gives the year as the cross-over year. (This is consistent with the forecast of the Ghana 2000 report which used the same assumptions and concluded that from the base year of 1987, it would take the average poor 20 years to move out of poverty and the extreme poor 53 years.) But does such an extrapolation make sense? Do poverty numbers correspond with a similar projection based on figures? Yes, they do. In , the incidence of poverty, using the same poverty line, was 36.9 percent with the average poor's mean consumption equaling Cedis 89,794 which is 67.9 percent of the poverty line. Assuming the same per capita income growth rate, this would have implied that the mean per capita consumption of the average poor in (T=4) would have been Cedis 97,196 which closely approximates the actual outcome in of Cedis 98,301. Thus, the reality does correspond to a forecast based on the above formula and assumptions. The reason that it does so is because income distribution has not changed much (slight improvement only) and, as Box 3.3 has shown, almost all of the reduction in poverty has come from an increase in mean expenditures. While that is so, the actual per capita consumption expenditures in suggest an implicit growth of per capita incomes of the poor of 2.3 percent - higher than the national average. This means that the poor participated more than proportionately in the growth process. It is remarkable that this should have occurred despite the regressiveness of public social spending, as the next chapter demonstrates. This then further suggests that the benefits to the poor would have been further enhanced and their exit from poverty expedited if there had been better targeting of public social spending. At the same time, the regressivity of public social expenditures and the low levels of human capital attainment (see Chapter 4) suggest that human capital constraints can well become binding to further growth, if they are not already. This means that the prospect for poverty reduction to continue apace at past trends may not materialize unless human capital improvements are made. Conclusions 3.18 Ghana has made substantial gains in reducing poverty. Poverty was reduced from 37 percent in to 32 percent in The reduction in poverty has been associated with sustained and distributionally-neutral growth. The reduction in poverty has also been broad-based and has benefited the more vulnerable groups in Ghana. The incidence of poverty among female-headed households has also fallen dramatically reversing the earlier disadvantage that they had relative to male-headed households; more than 60 percent of the reduction in aggregate poverty comes from the reduction in poverty of femaleheaded. While poverty continues to be a predominantly rural phenomenon, most of the poverty reduction has also come from the rural areas. Growth and increasing dependence on non-farm self-employment income, particularly from trading and manufacturing, appear to drive the reduction in poverty.

54 But poverty in Ghana still remains a serious and extensive problem. Almost 5 million people remain poor in Ghana. At current rates of growth, the average poor can expect to escape poverty only in , and the extreme poor in , provided that no other economic condition becomes binding. But poverty alleviation can be expedited through proper policies. This suggests four things. First, since economic growth has been shown to be a fundamental requirement for effective poverty reduction in Ghana, the pursuit of proper macroeconomic policies to promote growth is a must. Second, since poverty is largely a rural phenomenon, appropriate agriculture sector policies are crucial in any strategy of poverty reduction. Third, trade policies which remove the price distortions in agriculture will promote the growth of labor-intensive agricultural production and result in poverty reduction. And fourth, as the following chapter brings out, investments in human capital and a better targeting of public social expenditures is required if growth is to be sustained and poverty is to be reduced.

55 Chapter 4 PUBLIC SOCIAL EXPENDITURES AND POVERTY Ghana's poor have participated in the growth process despite generally regressive public social expenditures during Public recurrent expenditures on health suggest a generally regressive pattern while expenditures on education were less regressive, due mainly to the progressive pattern of spending on primary education. There is substantial scope for improving the poverty-targeting of social expenditures which can also be assisted by increased cost recovery, particularly in tertiary education and high-cost curative care. Ghana's growth and poverty reduction effort would have been greater if the poor had benefited more from the social expenditures; at the same time, current low levels and regressive patterns of targeting will make it more difficult to accelerate or even sustain broad-based growth and poverty reduction in the future. 4.1 The preceding chapter has shown that economic growth over was associated with a strong and broad based reduction in rural poverty. The pervasiveness of growth and poverty reduction efforts is in large part determined by the human capital stock of the poor as it is their access to such social services as health and education that determine their participation in the growth process. Therefore, expanding access to these social services is a key component of the long-term strategies for sustained economic growth and poverty reduction. This chapter attempts to determine to what extent the public funding of such services has been targeted to the poor in Ghana. This is used to draw inferences about the role that public social spending can play in promoting accelerated growth and poverty reduction. The Level of Human Capital 4.2 The Ghana 2000 study emphasized that it is in the development of human resources -- health and education - that Ghana has to make the biggest strides if it is to break into the league of fast growing economies. Even with the best of economic policies, rapid economic growth will not be feasible unless Ghana invests more, and more effectively, in human capital. 4.3 Level of Human Resource Development. UNDP's Human Development Report classifies Ghana as a country with low human development. In 1993, the report gave Ghana a ranking of 131 out of the 173 countries ranked; a year later, Ghana's ranking had slipped to Table 4.1 provides some indicators of human capital in Ghana. Infant and under-five mortality rates have decreased in recent years, but there are large regional disparities.31 For instance, under-five mortality is higher in rural areas (149 per 1000 live births) than in urban areas (90 per 1000 live births). Rates of immunization are particularly low at only 39 percent. Primary school net enrollments are far short of universal levels. Female disadvantage is quite apparent with their substantially lower net enrollment rates at both the primary and secondary levels. While literacy rates are higher than those for Sub-Saharan African countries, they are only at the average for the low-income countries. 30 The ordinal ranking is on the basis of the Human Development Index which is an unweighted average of the relative distances measured in terms of life expectancy, literacy, mean years of schooling, and income per capita; the measure of distance is the difference between the value of the variable in a country and a minimum, divided by the range of the variable. 31 See Ghana Statistical Service (GSS) and Macro International Inc. (MI), 1994, Ghana Demographic and Health Survey 1993, GSS and Ml.

56 36 Table 4.1: Human Capital Indicators (atest year) Indicator Unit of Measure Ghana Sub-Saharan Low-Income Africa Countries Health Infant Mortality per '000 live births Under-5 Mortality " Maternal Mortality per 100,000 live births Life Expectancy years at birth Immunization % of age group Education Net Primary Enrollment Ratios % of age group Female " Net Secondary Enrollment Ratios " Female " Literacy % of pop., age Source: Ghana Statistical Service (GSS) and Macro International Inc. (Ml), 1994, Ghana Demographic and Health Survey 1993, GSS and MI; World Bank, 1994, Social Indicators of Development 1993, Johns Hopkins University Press, Baltimore and London 4.4 Level of Investment in the Social Sectors. Though spending on the social sectors has increased significantly from the low levels in 1983, current levels of public expenditures in health and education remain low relative to comparable countries. During the last five years, , health expenditures averaged 1.2 percent of GDP while education expenditures averaged 3.7 percent of GDP. 4.5 Compared with the levels of public health expenditures in other countries, Ghana's levels remain low despite a health reform program under the ERP which was designed to reorganize health care delivery and to promote a more effective and equitable health care system.32 Major policy strategies include decentralizing health management to lower levels, intensifying health campaigns to reduce morbidity and mortality, and improving the quality of services. 4.6 Public expenditures on education have also been low despite the adoption of the comprehensive education reform program in 1986 under the ERP. In addition, personal emoluments continue to account for about 95 percent of the recurrent public spending in the sector, leaving very little public allocation for non-wage operations and maintenance expenditures which have obvious adverse effects on the quality of education services. 4.7 Given the already poor human capital base, this argues for a reallocation of public expenditures towards the social sectors. In addition, as this chapter will show, social expenditures need to be oriented more towards basic health and basic education from which the economic gains are the highest and which are also more accessible and relevant for the poor. 4.8 While there is some scope for reallocating expenditures towards these sectors, any large increase in public expenditures on them is infeasible given the fiscal situation and the concerns over macroeconomic stability. This suggests that additional resources will have to be sought either through cost recovery or through direct private sector participation, for which also cost recovery is required. Hence, the promotion 32 The World Development Report 1993 estimated that in Sub-Saharan Africa, Asia, and Latin America, public expenditures on health averaged about 2 percent of GDP (WDR 1993, pp. 52).

57 37 of private participation in the delivery of health and education services has to be a necessary component of a program for enhancing the human capital stock of Ghana. The Incidence of Public Social Spending Overview. Who benefits from government spending on the social sectors in Ghana? The following sections investigate this adopting the approach known as 'expenditure incidence' analysis (see Box 4.1). Public expenditures in health are generally regressive in absolute terms (see Figure 4.1). Aggregate health expenditure data for also do not support the increasing utilization of health services by the poor. In 1989, the poorest quintile appropriated only 12 percent of total health spending whereas the top quintile appropriated 31 percent of the total health budget; by 1992, the share of the poorest quintile fell to 11 percent while that of the richest quintile increased to 34 percent. It is not possible to say whether the regressivity of health spending has fallen over the period since the distribution of incidence curves for 1989 and 1992 cross each other. In terms of actual cedis, the top quintile is estimated to have gained over Cedis 6,670 per annum from its use of subsidized health services in 1992 whereas the lowest quintile received just Cedis 2,278. Box 4.1: What is 'Expenditure Incidence' Analysis? Who benefits from government spending on the social sectors in Ghana? This can be answered using expenditure incidence analysis, also referred to as 'benefit incidence' analysis. This approach has become popular since the pathbreaking work on Malaysia by Meerman (1979) and on Colombia by Selowsky (1979). There has been a recent resurgence of interest in the approach, reviewed in Van de Walle and Nead (1995). Few applications have been attempted in Africa-the exceptions being World Bank (1995) and Demery et al (1995). Expenditure incidence analysis seeks to measure how government subsidies on services such as health and education are distributed across groups in society. The distribution of these subsidies is determined by two broad factors. First, it depends on government spending itself, and how it is allocated within the sector. The lower the spending, and the greater the effective cost recovery, the lower will be the subsidy embodied in the service provided. Second, the distribution will depend on household behavior-on who uses the service that the government provides. It is only by using the service (by, for instance, sending a child to a priimary school, or visiting the outpatient department at a hospital) that individuals and households can lay a claim to the in-kind transfer that is implicit in the subsidy. Expenditure incidence analysis therefore brings together two sources of information: data on the government subsidy (estimated as the unit cost of providing the service less any cost recovery back to the government) allocated to the different categories of service (primary schooling, in-patient hospital care, etc.); and information on the use of these services by individuals and households, which is usually obtained from household surveys. In general, government expenditures will be more equally distributed when the spending is concentrated on services which are used widely by the population, and used especially by poorer groups. If public expenditures are concentrated on primary education, or on primary-health facilities such as clinics or health centers-which are widely used by the poor-public expenditures will tend to be more equally distributed. However, if governments spend more on high-cost services which are not generally used by poorer groups (such as university education or in-patient hospital care), the incidence of spending is likely to be more unequal. In sum, the expenditure incidence of public spending depends both on the allocation of public expenditures within the sector, and on the behavior of households. References: Demery, L., R. Bernier, S. Chao, K. Mehra, 1995, "The Incidence of Social Spending in Ghana, ," mimeo, February. Meerman, J., 1979, Public Expenditures in Malaysia: nho Benefits and "hy?, New York: Oxford University Press. Selowsky, M., 1979, Who benefits from Government Expenditure?, New York: Oxford University Press. World Bank, 1995, Poverty Assessment - Kenya, Report No This section draws on Demery, L., R. Bernier, S. Chao, K. Mehra, 1995, "The Incidence of Social Spending in Ghana, ," mimeo, February.

58 Education spending was less regressive than health spending in absolute terms in 1989 and 1992, being distributed almost uniformly over the expenditure quintiles (see Figure 4.1). In 1988, the poorest quintile received 17 percent of the total subsidy compared with 24 percent for the top quintile; in 1992, the shares were similar with the poorest quintile getting 16 percent and the richest quintile getting a smaller 21 percent. The middle quintiles have gained a larger share of the education spending which is reflected in the distribution curve for 1992 (for the 2nd to the 5th quintile) being slightly above that for Whereas the top quintile is estimated to have gained over Cedis 11,279 per annum from its use of subsidized education services in 1992, the lowest quintile received just Cedis 8, While the aggregate levels of health and education are generally regressive in absolute tenns, they are progressive relative to the distribution of income (see Figure 4.1). The Lorenz curve for 1992 lies well below the distribution of incidence curves for health and education for the respective years.34 For instance, the bottom quintile's receipt of health and education spending far exceeded its 4 percent share of total household expenditures. This pattem is strikingly uniform across all population quintiles. Figure 4.1 The Incidence of Public Social Spending , 1992 Lorenz Curve Education U - Education i- Health Healthl ~ ~ ~ ; ~~~~~~~~~~ Cumulative Distribution of Population (%) Source: Demery eta, The Incidence of Public Spending on Health. The pattemn of aggregate health expenditures above closely reflects the pattern of expenditures over different types of health facilities and over expenditure quintiles (sec Figure 4.2). Since the distribution remains almost constant over the period, only the 1992 distribution is shown. The poorest (bottom) quintile does not appear to gain from the health services, provided throug any type of facility, in proportion to their share in population. In 1992, their share of public expenditures on health centers and clinics amnounts to just 10 percent. Their share in the use of out-patient hospital services was 13 percent while their share of the in-patient subsidy was 11 percent. In general, the contrast with the richest quintile is dramatic. In 1992, of the total public spending on health centers and clinics, the share of the richest quintile was 31 percent in 1992 while of the spending on 34 The Lorenz curve for 1989 almost overlaps with that for 1992 and is therefore not presented here in the interest of clarity.

59 39 outpatient care, they appropriated 35 percent in Of the in-patient public spending, their share was 32 percent in The inequality of public expenditures in health is also reflected in its regional distribution (see Table 4.2). In 1989, urban Ghana, which represents under a third of the total population, obtained 42 percent of the total health budget, and over 50 percent of out-patient spending; in 1992, this had increased to 49 percent of the health budget and 55 percent of out-patient spending. By contrast, the rural population in 1989 received 58 percent of the health spending and 50 percent of the out-patient spending; in 1992, these had fallen to 51 percent and 45 percent respectively. On average, urban residents received a subsidy over the year of Cedis 2,223 in 1989, compared with just Cedis 1,459 for rural residents - a differential of 52 percent; in 1992, the comparable figures were Cedis 5,808 for the urban and Cedis 3,039 for the rural population, and the differential had widened to 91 percent. The pro-urban bias has, therefore, become noticeably more marked. 35 Figure Distribution of Public Recurrent Spending on Health by Expenditure Quintile, u las i Outpatient Inpatia Clinics, ec. Source: Demery et al, 1995 Expenditure Quintiles Table 4.2: Distribution of Public Recurrent Expenditures on Health by Urban-Rural Areas (perc nt) Urban Rural All (Mean subsidy, in cedis) (2,223) (5,808) (1,459) (3,039) Outpatient Inpatient Health Centers & Clinics Source: Extracted from Demery et al, The analysis underestmtes the allocation of health spending in Accra since the GLSS had very few respondents seeking in-patient care in Accra. Thus, the analysis actually underestimates the true degree of regressivity in the health system of the country.

60 There is also evidence of gender differences36 in public health spending incidence - females gained a larger and increasing share of the expenditures - from 55 percent of total health recurrent spending in 1989 to 56 percent in 1992 (see Table 4.3). Spending per capita was 13 percent higher for females relative to males in 1989, increasing to 21 percent in But this bigger and increasing share is true primarily for the richest quintile. In 1992, the poorest quintile among females obtained only 9 percent of total health spending whereas the richest quintile received 38 percent compared with 1989 shares of 10 and 34 percent, respectively, suggestive of increasing inequalities among females. Table 4.3: Distribution of Public Recurrent Expenditures on Health, by Gender (%) Expenditure Quintile Male Female ALL (Mean subsidy, in cedis) (1,598) (3,576) (1,811) (4,321) First Second Third Fourth Fifth Source: Extracted from Demery et al, Cost Recovery in Health in Ghana? The above analysis clearly lays out the case for increased targeting towards the poorer quintiles, rural areas, and women. Targeting rural areas and women is easier as readily available specific criteria can be used for identifying these groups, for instance, though geographical distribution, sex, and type of health need. But targeting the poorer quintiles in general is more difficult. One way that this can be facilitated is through increased cost recovery. Cost recovery can be used to charge those who can pay for the use of service and to channel these revenues to expanding services for those who cannot. At the same time, cost recovery introduces price signals which direct users to the appropriate level or type of medical facility leading to a more efficient use of these services. In fact, both economic theory and international experience suggest that cost recovery in health can lead to reduced fiscal pressures, increased resources, and if used properly, to improvements in quality and to increased access to health facilities (see Box 4.2) In Ghana, charges for publicly provided health services were introduced in 1971 with the Hospital Fee Act, but the fees charged were very low. In 1985, a much larger move toward cost recovery began. Receipts from user fees increased from 5 percent of the recurrent budget of the Ministry of Health in 1985 to 12 percent in However, after 1987 the contribution of user charges to operating costs declined because the fees were not adjusted to keep pace with inflation A more recent move toward cost recovery has been the implementation of increased charges for drugs (and other curative services) at government health clinics and hospitals under the "cash and carry" scheme (a revolving fund for drugs), which in fact involves drug prices somewhat higher than their full cost. This change was implemented in Greater Accra in 1990 and then in most other regions in Exemptions from payment are, however, made for Ministry of Health staff, indigent individuals, persons requiring drugs or treatment for particular diseases (leprosy, tuberculosis and psychiatric disorders), and 36 The ternn "differences" is used here instead of "bias" since (unlike with education) there are significant differences in the health needs of males and females. It is not possible, therefore, to draw conclusions about bias in health spending without quantiffing differences in needs.

61 41 for services for children (immunizations, antenatal and postnatal services, and "treatment at child welfare clinics"). It appears that the vast majority of exemptions are given to Ministry of Health staff and their family members, and very few are given to poor persons. But exemptions granted to children for preventative care are being honored. In Ghana as a whole the percentage of children between the ages of 13 and 24 months that were fully immunized rose from 46.7 percent in 1988 to 54.8 percent in 1993, an improvement which took place in both urban and rural areas One potential risk of instituting cost recovery for drugs is that it may decrease use of health care services by the population, and the greatest drops may occur among the poor. But evidence from the Ghana Living Standards Survey suggest that for Greater Accra between 1988 and 1991 (i) there is no evidence of declining use of health services; (ii) there is no evidence of a shift from the public to the private sector - in both years a slightly higher percentage visited public sector facilities than visited private facilities; and (iii) there is no evidence that differences in the use of health services between the poorer and wealthier quintiles have widened - in fact, the opposite seems to have happened Both economic theory and the lessons from other countries suggest that (a) cost recovery should not be attempted for preventative care; (b) low-cost curative health services should be priced to cover most, if not all, of their costs; and (c) high cost curative care should be priced for civil servants and formal sector workers to the extent that insurance schemes can be developed. In Ghana's context, this suggests that a self-financing insurance program for civil service and formal sector workers can be introduced to recover costs of high-cost curative care. This can, perhaps, be done prospectively with progressively higher rates of cost recovery for future entrants into formal sector employment. 37 Ghana Statistical Service and the Institute for Resource Development/Macro Systems, 1989, "Ghana Demographic and Household Survey 1988," Columbia, Md.; Ghana Statistical Service and the Institute for Resource Development/Macro Systems, 1994, "Preliminary Report on the Ghana Denographic and Household Survey 1993", Columbia, Md.

62 42 Box 4.2: Cost Recovery in Health: Economic Theory and the Lessons of International Experience It is useful for analytical purposes to divide health care services into preventative care and curative care. The former services are intended to prevent possible future health and nutrition problems, and the latter services are treatments of existing health problems. Curative care can be further divided into relatively inexpensive curative care (such as outpatient purchases of medicines for common health problems) and relatively expensive care (such as treatments that would require hospitalization). Both economic theory and the lessons from other countries suggest that (a) cost recovery should not be attempted for preventative care; (b) low-cost curative health services should be priced to cover most, if not all, of their costs; and (c) high cost curative care should be priced for civil servants and formal sector workers for whom insurance schemes can be developed. There are several factors which suggest that preventative health care should be subsidized. (i) There are large positive externalities involved in providing it, for instance, prevention of contagious diseases; (ii) Distributional objectives can be served since most forms of preventative health care can be dispensed at local health clinics and the service is not confined to major urban areas. Thus, it is likely to be used more by the poorer rural population. (iii) The cost of providing preventative services is usually not large, so the scope for cost recovery is not large. And (iv) it is likely that many households would not be aware of the value of preventative care, even if public awareness campaigns are mounted. For curative care there is more scope for cost recovery. This is because there is little problem either with respect to inadequate infornation or externalities. For relatively inexpensive care there are also few problems with respect to market failure and redistributional objectives - since the costs are not large there is little need for credit and insurance and the amounts to be redistributed are also small. However, for relatively expensive care there are market failure problems (lack of insurance) and the potentially regressive nature of any cost recovery measures. Both can, however, be overcome through an insurance system with wide coverage that pays for itself - i.e., insurance preminums cover the cost of treatment of the insured groups. While this would be hard to do for all, it should be possible for formal sector employees. Since households with formal sector employees usually have higher incomes than the others, the distributional implications of charging full cost insurance premiums are "pro-poor." Several developing countries have attempted to apply cost recovery for low-cost curative health services. A particularly informative experimental program was tried in the Adamaowa province of Cameroon (Litvack and Bodart, 1993). Fees in rural health centers for curative drugs and medical supplies (bandages, etc.) were raised from zero to more than their full cost (to help pay for overhead and to cross-subsidize health outreach services) but the money raised was used to ensure that these drugs and supplies were indeed available at the clinic. This increase in quality (certainty of getting the drugs and supplies) led to substantial increases in use of health centers. Most interesting was the finding that the increase was highest for the poorest groups. This study provides one concrete example of how increased cost-recovery can increase the delivery of curative health services if thefunds raised are used to increase the quality of the service. Indeed, in countries that raised fees without raising quality (Lesotho, Mozambique and Swaziland), use of facilities declined substantially (World Bank, 1993). Collection also appears to be encouraged if fees are retained at the clinics where they are collected and not sent to the capital city (World Bank, 1993). Concern over the adverse effects on the poor of the high costs of some types of curative care, e.g. hospital care, has led some countries to institute cost recovery but with fee waivers or reductions for poorer households. But, in practice, this is difficult to do because hospital administrators generally cannot determine the incomes of patients. A study of targeting in Latin American examined two such countries - Belize and the Dominican Republic (Grosh, 1994). The findings support the view that waivers are difficult to implement; it was not even clear that the waiver recipients were, in fact, the desired target group. As for instituting formal insurance schemes which reach the poor, this is probably not feasible for most low-income countries - the one case of a successful nationwide insurance program aimed at the poor, that in Costa Rica (Grosh 1994), depended on the high rate of literacy of the population and the high incidence of formal (documentable) economic transactions. But insurance schemes covering formal sector and/or civil servants have been adopted in some African countries, e.g., Mali, Senegal, Sudan and Zaire. Private insurance for wealthy individuals who are neither civil servants nor formal sector employees has also developed in Cbte dlvoire, Ghana, Senegal and Zimbabwe. Sources. Grosh, M., 1994, "Administering Targeted Social Programs in Latin America: From Platitudes to Practice", World Bank Regional and Sector Studies, Washington, D.C. Litvack, Jennie, and Claude Bodart, 1993, "User Fees plus Quality Equals Improved Access to Health Care: Results from a Field Experiment in Cameroon," Social Science and Medicine, Vol.37, No.3, pp World Bank, 1993, "Better Health in Africa, " Africa Technical Department Working Paper No. 7, Washington, D.C.

63 The Incidence of Public Spending on Education. The aggregate pattems of public spending in education hides large variations over the different levels of education (see Figure 4.3). The pattems of the primary education subsidy across the quintiles have remained remarkably stable over the period and so only the pattem for 1992 is shown. The share of primnary spending received by the poorest quintile was 22 percent while that for the top quintile was 14 percent. A similar pattem is observed for secondary education with the poorest quintile receiving 15 percent and richest quintile receiving 20 percent. But the pattem of tertiary spending is very regressive with the share of the lowest quintile only 6 percent while that for the richest quintile 45 percent. Figure Distribution of Public Spending on Education by Expenditure Quintile, 1992 Ž L 20, llll U, Prinary Secondary Tertiary Expenditure Quintiles Source: Deinery, et al, Rural-urban inequalities in education spending are also evident with urban areas receiving about 50 percent more than the rural areas in per capita terms (see Table 4.4). While per capita expenditures for primary education are higher in the rural areas, the rural-urban inequalities are strong in secondary and tertiary education. In secondary education, the urban areas received 11 percent more than the rural areas in 1989; by 1992, this differential had widened to 47 percent. In tertiary education, urban areas received about 600 percent more in 1989; but by 1992, this differential had narrowed slightly to about 500 percent. Table 4.4: Distribution of Mean Public Expenditures on Educaton, by Urban-Rural Areas (cedis) Level of Education Urban Rural All 5,310 13,580 3,414 9,218 (% of total expenditures) (41.3) (42.1) (58.7) (57.9) Primaiy 1,883 3,987 2,140 4,628 Secondary 1,077 5, ,981 Tertiary 2,104 3, Source: Extracted from Demery et al, 1995 a/ Annual, per capita expenditures.

64 There are also gender inequalities across all education levels but the most acute for secondary education (see Table 4.5). Males, in general, received almost 50 percent more of the spending relative to females. In secondary education, females received only 37 percent of the spending in 1989 which increased a little to 40 percent by The gender inequalities have worsened slightly for primary education spending but improved for the other levels of education. Table 4.5: Distribution of Mean Public Expenditures on Education, by Gender (%) Level of Education Male Female All (Mean subsidy, in cedis) (4,782) (12,803) (3,157) (8,614) PrimnarY Secondary Tertiary Source: Extracted from Demery et al, Cost Recovery in Education in Ghana? Economic theory and the lessons from other countries suggest that (a) cost recovery should be done cautiously at the primary level, and only when enrollment rates are almost universal (95 percent or higher); (b) there is more room for cost-recovery at the secondary level, but not full cost recovery; and (c) cost recovery at the tertiary level should be as complete as possible (see Box 4.3) Though there was very little cost recovery in education before 1987 in Ghana, secondary schools with boarding facilities were often closed for weeks due to lack of food for boarders, and at all levels of education, textbooks were often completely unavailable. The only substantial fees were for meals eaten by secondary school students who lived in boarding facilities (no fees were paid for rooms), and even these fees covered only about one third of the costs In 1987, the first education reform program began. Small user fees were charged at the primary and JSS (junior secondary school) levels, textbook fees were introduced for grades 3-10, and full cost recovery was introduced for meals at secondary schools. Thereafter, official fees increased so as to keep pace with inflation, but cost recovery in terms of the percentage of Ministry of Education costs recovered from user fees declined because teacher and school administrator salaries were increased in real terms. However, within two years of the introduction of decentralization in education (at the primary and JSS levels) in 1988, "unofficial" fees charged by schools increased dramatically, resulting in increased real costs to parents. In recent years, the Government has tried to reign in "unofficial" fees, which were primarily used to "top off' salaries and raise money for local governments, not to improve school quality Although net primary enrollment in Ghana has increased from 68 percent in to 74 percent in , it is still not universal, and Ghana is, therefore, still far from the point where some form of cost recovery should take place. Net primary enrollment rates have increased for all expenditure quintiles, including the poorest. At the secondary level, the enrollment rate has increased from 33 percent in to 38 percent , with all quintiles again participating in the increase. Overall, the results suggest that the very modest moves toward cost recovery at the primary and secondary level do not appear to have reduced school enrollment.

65 45 Box 4.3: Cost Recovery in Education: Economic Theory and the Lessons of International Experience Economic theory and the lessons from other countries suggest that (a) cost recovery should be done cautiously at the primary level, and only when enrollment rates are almost universal (95 percent or higher); (b) there is more room for cost-recovery at the secondary level, but not full cost recovery, and (c) cost recovery at the tertiary level should be as complete as possible, with student loans used to overcome problems of incomplete credit markets. There is sound economic reasons why primary education should be universal, since both the private and social rates of return from it are very high. However, cost recovery cannot work because of several reasons: (i) The per student costs are very small and, therefore, the scope for cost recovery limited; (ii) Many families may not send their children to primary school and it may be difficult for governments to correct for this as parents may be entrenched in their views because of their own lack of formal education. (iii) Even if families want to send their children to primary school, they may be constrained by lack of own finds or of access to credit. Here, too, governments may be unable to provide a remedy through, for instance, a public loan program because (a) the transactions costs involved in processing the small loans that will be required will be high, and (b) the infeasibility of collecting loan payments through direct paycheck deductions if, as may well be the case, most of the children end up outside of formal employment. With respect to secondary schooling, it is not clear whether universality is the socially optimal outcome. There also appears to be some more room for cost recovery since (i) the costs per student are much higher (due to more skilled teachers and special coursework); and (ii) parents will be better inforned about the value of education. But the problem of parents not having access to credit remains and, in fact, becomes more acute because of the larger amounts involved. At the tertiary level, cost recovery should be made as complete as possible for three reasons: (i) There is little reason to think that positive externalities result in a sub-optimal number of people entering institutions of higher learning; (ii) There is little reason to believe that households in which someone has successfully completed secondary education do not realize the value of tertiary education; and (iii) The income distribution effects of charging user fees are almost certainly pro-poor since individuals in tertiary education come from households with much higher than average incomes. The main problem, however, is the lack of access to credit again but it is quite possible to put mechanisms into place which will enable students to take loans which they can later repay. There have been very few studies of the feasibility of cost recovery at the primary and secondary levels of schooling. Two studies at the primary level in Mali and Malawi suggest that it may well be possible to raise user fees while maintaining enrollment by using the revenues from the fees to increase school quality (Birdsall 1987 and Tan et al 1984). But these two studies were based on econometric estimates and not on randomized experiments, so one needs to proceed cautiously given the lack of hard evidence. There are a few remedies for reducing the burden on poorer households and disadvantaged groups. In Bangladesh, girls pay lower tuition rates than boys at the secondary level. A similar policy has been instituted in northern Nigeria, where tuition fees are waved for all girls. There is also scope for waiving or reducing fees on the basis of a geographical targeting of the poor. Individual household targeting, while fine in principle, is impracticable as with health fees. Student loans for secondary schooling remains a possibility but, as yet, an untried one. In virtually every country in the world students enrolled at the tertiary (university) level are from better off families, and this is particularly true of developing countries. Thus there is no reason to subsidize tertiary education from the perspective of redistributing from rich to poor. The only reason to provide financial assistance to attend universities is the problem of incomplete credit markets. Student loan programs are the obvious mechanism to correct for this, and indeed many developing countries, particularly in Latin America, have instituted such programs. But the programs in Latin America recover at most 25 percent of costs from student fees. Still this amount is quite high compared to most other developing countries. Of particular interest is the case of Chile, which moved from a heavily subsidized program in the 1960's and early 1970's to substantial cost recovery in the late 1970's and 1980's. The distribution of students across income groups was no more skewed (and indeed was probably less skewed) toward wealthier groups after the move to substantial cost recovery was made. Note that the interest rates on these loans are almost always lower than market rates. (A general rule of thumb is that a loan should be repaid over a period of ten years, and the payments should not exceed 10 percent of the university graduate's income.) In terms of the political feasibility of introducing fees at the tertiary level, given the possibility of student unrest, is to introduce the fees gradually, and only to new students (Carlson 1992). This has been done successfully in Sweden (in the 1960's) and, more recently, in Great Britain. Sources: Birdsall, N. 1987, "Demand for Primary Schooling in Rural Mali: Should User Fees Be Increased?" Technical Note 874, Population, Health and Nutrition Department, The World Bank, Washington, D.C. Carlson, S., 1992, "Private Financing of Higher Education in Latin America and the Caribbean." Latin American and the Caribbean Technical Department, Regional Studies Program Report No. 18, The World Bank. Washington, D.C. Tan, Jee-Peng, Kiong Hock Lee and Alain Mingat, 1984, "User Charges for Education: The Ability to Pay in Malawi," World Bank Staff Working Paper No. 661, The World Bank, Washington, D.C.

66 Before 1989, university education in Ghana was free of charge and housing and meals were also provided at no cost to students. In 1989, a small move towards cost recovery was made by charging students for meals, although there were still no tuition charges. 38 To help students finance meal costs, a student loan scheme was instituted allowing students to borrow up to $200. Repayment begins once the graduate finds employment in the civil service or the formal private sector -- a 5 percent payroll deduction (plus a 12.5 percent contribution from the employer) that would normally go to the Social Security Administration is instead used to repay the student loan. This implies lower contributions to the social security system over the lifetime of the individual. However, since most workers accumulate maximum retirement benefits long before reaching the age of retirement, this program effectively costs the students nothing. All it does is transfer money from the Social Security system to the Ministry of Education. In addition, the interest rates on the loans are only 3 percent in nominal terms, which given an annual inflation rate of percent implies significantly negative real rates of interest. Briefly, this attempt at cost recovery is a very small one. Conclusions 4.28 The education and health sectors in Ghana are not functioning effectively in the sense that the majority of the population do not have access to good quality education and health services. Health sector targeting is poor, and exhibits a much more regressive pattem than education in absolute terms. The problem does not appear to be in the allocations across the types of health facilities, as the poor seem to use hospitals as much as clinics, but in the lack of availability and poor quality of services. Education targeting overall is also slightly regressive in absolute terms with primary education somewhat progressive and tertiary education highly regressive. The regional distribution of social spending across rural and urban areas also reveal regressive patterns with the rural areas receiving less than their share of the population as well as less in per capita terms. Significant gender differences also exist, particularly in secondary and tertiary education While the patterns of spending are regressive in absolute terms, they are progressive relative to the distribution of income. Even tertiary education, which is the most regressive of all the various levels of social spending, is also progressive relative to the income distribution. Be that as it may, the overall pattern of public social spending suggests that future growth may well be constrained by human capital and may not be as broad-based as the past experience. This makes investment in human capital all the more critical if Ghana is to achieve accelerated growth with poverty reduction Measures to improve the targeting of social expenditures can be assisted by selective measures of cost recovery. So far, cost recovery in the health and education sectors is still quite modest in Ghana and there has been very little increase in cost recovery during the ERP. Moreover, the increase in cost recovery between and does not appear to have affected the poor's use of social services. In this context, three lessons emerge for Ghana. First, in the interest of equity and greater access to the services by the poor, cost recovery should be avoided for primary education and preventative care. Second, there is some scope for recovering costs for secondary schooling and for low-cost curative care by targeting the non-poor, perhaps on a geographical basis. This would imply free access for those in the poor areas and limited cost recovery for the rest. Third, tertiary education and high-cost curative care are more amenable to cost recovery. In education, this can be done through progressively increasing targets for cost recovery 38 Initially, students were also charged for housing, but student unrest resulted in eliminating housing fees. Also, a small matriculation (registration) fee was added, but the level of that fee was very low, amounting to only US$5-7.

67 47 in tertiary education with the aim of aclieving full cost recovery in the medium- to long-run. In healt, cost recovery can be started through the introduction of a self-financing insurance program for formal sector workers to recover costs of high-cost curative care. This can, perhaps, be done prospectively with progressively higher rates of cost recovery for future entrants into formal sector employment.

68 Chapter 5 PROMOTING PRIVATE SECTOR GROWTH In addition to investing in human capital, there are at leastfive areas of deficiency which have inhibited private sector growth in Ghana. Policy responses to these relate to sustaining macroeconomic stability, investing in infrastructure, strengthening the financial sector, liberalizing the petroleum sector, and removing the distortions in cocoa trading. The adoption of policies towards these ends is essential if Ghana is to move towards accelerated growth. 5.1 The experience of Ghana with poverty reduction has clearly brought out the crucial role of broadbased economic growth, most of which was generated in economic activities undertaken by the private sector. But growth in Ghana continues to be constrained by deficiencies in at least six areas -- human capital, macroeconomic stability, infrastructure, financial sector, the petroleum sector, and the cocoa sector. The previous chapter has underlined the need to improve the stock of human capital and stressed the importance of orienting social spending towards the poor so that they are better placed to take advantage of the economic opportunities as and when they arise. This chapter examines how best the policy environment in the remaining five areas can be improved so as to stimulate private sector growth. Sustaining Macroeconomic Stability 5.2 It has been shown in Chapter 1 that macroeconomic instability arising from the fiscal shock of 1992, and excessive borrowing by parastatals, has characterized Ghana's recent economic enviromnent. The fiscal and the monetary excesses have accentuated the need for fiscal discipline and tighter monetary control. In the light of Ghana's economic experience over the last few years, what lessons are there for the future, particularly in view of the forthcomin general elections in 1996? Three key lessons can be clearly drawn: 5.3 Avoid Another Wage Shock. The avoidance of another wage shock in the prelude to the 1996 elections is an absolute necessity. There are three reasons why this is paramount. First, any recourse to such a measure would further undermine the macroeconomic situation and destroy the credibility of government policy. Second, there is hardly any scope of increasing revenues further to cover expenditure increases. Tax revenues are already more than 15 percent of GDP and the scope for increasing them any further is limited in a private sector-based development strategy. Moreover, the Government has been using divestiture receipts to offset budgetary expenditures. Not only are such uses not advisable but even the recourse to them in the future is limited until the very large state-owned enterprises are sold. Third, the balance of payments needs of the country are already substantial and derive in large measure from the overall budget deficits. Any measure to increase these deficits is only going to further worsen the current account deficits which would push Ghana into major extemal financing problems. 5.4 Reduce Expenditures on Wage-Related Items and Interest Payments. Public expenditures on wages and salaries, subsidies and transfers, and interest payments are already very high as a proportion of GDP and need to be reduced in order to ease the financing requirements, promote macroeconomic stability, and provide greater resources for non-wage O&M expenditures. 5.5 Any reduction in these expenditures require action on two fronts. First, it requires a restructuring and the downsizing of the public sector. This is being done through the divestiture of many public sector undertakings and through civil service reform. It is particularly important to rationalize the size,

69 49 composition, grading structure, and remuneration of the civil service. This will enable the wage structure of a rationalized civil service to counter both the erosion of real wages as well as the widening of wage differentials between the public and the private sector so that adequate incentives for civil service employment and performance can be provided. To this end, the Government needs to (i) implement fully an integrated personnel and payroll system; (ii) set manpower ceilings for all government agencies, including subvented agencies; (iii) set hiring policy guidelines; (iv) introduce a differential pay grade system to stretch pay relativities in favor of middle and upper management staff; and (v) expand the retrenchment program. 5.6 Recent steps announced by the Government have been in this direction. It recently launched the National Institutional Renewal Program (NIRP) the purpose of which is to review the functioning of the whole public sector, assess its relationship with the private sector, and make recommendations for reform. Under the aegis of NIRP, a Civil Service Performance Improvement Program (CSPIP) was also initiated in February, 1995, with the declared goal of improving the efficiency of the civil service. 5.7 Second, public financial management systems need to be improved if the Government is to enhance the control, efficiency, and effectiveness of the use of public resources. The Government's Public Expenditure Review for 1993 highlighted several problems in public financial management including (i) weak budget preparation; (ii) weak expenditure monitoring and control mechanisms; (iii) lack of a proper accounting system; (iv) lack of ownership and accountability on part of the various Ministries and Departmental Agencies (MDAs); and (v) problems in controlling and monitoring the multi-year contracting system. In order to address these problems, the Government developed in late 1994 a medium-term Public Financial Management Reform Program (PFMRP). New systems have been introduced on a pilot basis in three key ministries - Health, Education, and Information, and will gradually be extended to all spending units of the Government. The proper development and implementation of the program will be critical to the Government's ability to monitor and control public consumption expenditures. 5.8 Maintain a Tight Control over Monetary Policy. Monetary policy in Ghana needs to be tight so that money supply can provide the nominal anchor to the financial system, and monetary policy can be directed to the task of controlling inflation. The control of inflation requires that monetary growth be consistent with the inflation objectives. Mopping up of the excess liquidity in the system needs to be done through the sale of official debt. But Ghana's open market operations are not fully effective. The Government needs to adopt measures to enhance the effectiveness of instruments of indirect control and to encourage the development of a secondary interbank market in domestic debt instruments. Measures to achieve this should include (i) refraining from signaling indicative reserve prices for Treasury Bills and Bank of Ghana Bills; (ii) monitoring the term structure of interest rates to reflect inflation targets; (iii) reducing the range of maturities, particularly the 30-day bills, to foster the development of a secondary market in bills; (iv). developing new financial instruments, including bankers acceptance and other commercial paper, to be traded in the secondary market by trading houses; and (v) keeping liquidity ratios under review along with monetary policy developments with a view to reducing them to levels needed for prudential purposes once the system of indirect monetary control is sufficiently strengthened. Access to central bank credit, which has led to inflationary monetary expansion in the past, should also be terminated.

70 50 Investing in Infrastructure The Ghana 2000 study identified infrastructural inadequacies as a major obstacle to growth and welfare improvements in the country. It is still the case, however, that Ghana's provision of communication and transportation services, although similar to levels in neighboring countries (Benin and Togo, for example), is lower than the poorer countries of Latin America. For instance, Ghana has only 2.8 telephone main lines per 1000 inhabitants compared to 21 in Guatemala, 17 in Honduras, and 26 in Peru. These differences are high even taking into account, for example, the greater urbanization of Peru. Evidence from around the world suggests that steps can be taken to close these gaps. Chile enhanced the role of the private sector in its telecommunications and electricity sectors in the mid 1980s. From 1980 to 1990, the number of telephone connections in Chile increased 137 percent, and electricity production rose 59.4 percent. The UK divested its telephone system in 1983; starting from a high number of connections in 1980, unsurprising for a high-income country, it nevertheless saw a 43 percent increase in the number of telephone connections over the decade. In Ghana, on the other hand, the number of telephone connections increased from 1980 to 1990 by only 20 percent, even though Ghana started from a much lower base of connections. Electricity production in Ghana during the same period increased by a meager 2.4 percent. While these comparisons are only indicative, they nevertheless make a powerful case for considering new ways of approaching infrastructure development Increases in investment in infrastructure can come from both the public and the private sectors. Ghana has done a remarkable job of expanding and rehabilitating roads, ports, as well as power generation and transmission over the last decade. However, rapid increases in the public contribution to infrastructure will be hampered by constraints on fiscal revenues, and by problems of implementation, especially in its infrastructure sector. Even if implementation improved, public sector resource constraints would limit the extent to which the Government can, by itself, relieve these infrastructure bottlenecks Private investment -- both domestic and foreign -- needs to be harnessed to fill the resource gap and to expand infrastructure investment. The Government has already indicated its interest in taking this path. In so doing, Ghana is following a strategy pursued by many countries in Latin America and elsewhere that have encouraged private sector provision of telecommunications services, power generation and distribution, water and sanitation facilities, as well as airport services, airlines and railroads The following sections (i) distill the lessons of international experience on private participation in infrastructure, (ii) examine the current policy arrangements in place in Ghana to assess the extent to which these fall within the policy frontier as defined by international best practices, and (iii) propose recommendations International Best Practice and Ghana. International best practice regarding the appropriate market structure and level of private sector involvement varies across infrastructure sectors. Table 5.1 summarizes international best practice with respect to the ownership structure of infrastructure and compares it to current Ghanaian practice. International best practice has demonstrated that private ownership and competition, combined with an appropriate regulatory framework, can trigger productivity improvements and service expansion in most infrastructure sectors. The exceptions to this are in sectors such as urban roads, where the expense of metering usage and collecting payment make private ownership 39 This section relies on a background paper by Philip Keefer, 1995, "Private Participation in Infrastructure: Ghana and International Experience," mimeo.

71 51 difficult. However, in Ghana, government ownership and management of infrastructure is still pervasive in all of these sectors, suggesting that the Government has to undertake more policy reforms if it is to approximate more closely intemational best practice. Table 5.1: Ownership Options for Infrastructure Provision: Ghana and Best Practice Compared Government Public Private Private Private Ownership Best Sector Department Enterprise Lease Concession More Less Practice Regulated Regulated Countries (less (more _ompetitive) competitive) Telecom Basic services Ghana Best US, Britain, Long distance Ghana Best Argentina, Value added Ghana Ghana'/Best Peru Power Thermal generation Ghana Best Chile, Britain, Transmission Ghana Best Norway, Distribution Ghana Best US Transport Urban bus Ghana Sri Lanka Primary and secondary roads Ghana 2 /Best Most countries Bridges and Highways Ghana 2 /Best Best Porl, Airport and Rail Services Ghana Ghana Best Chile, US, ~ ~~~~~~~~Britain Water Urban piped network Ghana Best Best Mexico, Nonpiped systems Ghana 3 /Best Argentina Sanitation On-site sewerage Ghana Best Ghana Most indust. Piped sewerage and treatment Ghana Best Best Chile,. ~ ~~~~~Agenfina Waste Municipalities in the US, Collection Ghana Best Ghana Europe, Sanitary disposal Ghana Best Ghana Latin Am. Notes: 1/ Only for a few value-added services does Ghana allow private entry with little or no price regulation. 2/ Although best (or at least mogt common) practice often involves control by government department, Ghana's allocation and expenditure of road fands remains non-transparent, employing criteria that often seem arbitrary. 3/ The problem in Ghana is, of course, that there is great reliance on unpiped water Competition catalyzes the greatest benefits from private participation in infrastructure. Not all infrastructure sectors, however, have the same competitive potential. There are three different levels of competitiveness into which different infrastructure sectors can be grouped. The first category contains sectors that lend themselves naturally to competitive market structures without price regulation. Examples are value-added services in telecommunications (cellular phone service, telephones, directories and directory assistance), nonpiped sanitation systems, and some airport and rail services (such as catering). All that is required for these markets to function competitively is the clear allocation of property rights (for example, rights of cellular operators to scarce radio spectrum), and sanctions on cartel behavior by firms. While Ghana has made significant progress in allowing cellular operators to enter, much needs to be done in freeing other markets for private investment.

72 The second category of infrastructure sectors contains those that are competitive but require more government oversight and regulation. This group includes long distance and basic telecommunications services and power generation. These sectors exhibit considerably higher capital requirements, and multiple providers do not often emerge immediately. Some form of interim price regulation is therefore often needed. In addition, even where there are multiple providers, the competitiveness of these sectors is more reliant on regulation. The Government of Ghana has slated greater private sector participation in power and telecom. For instance, in its telecommunications plan the Government contemplates ambitious levels of private sector participation, including nearly unregulated access to local telecom markets, following the examples set by Chile, New Zealand and the UK In the third category of infrastructure sectors, industry characteristics such as economies of scale or "contiguity", or network economies make it inefficient to have multiple firms operating in the same market. This is true for urban piped water systems, waste collection, urban bus systems and power transmission. It is also true for catering services inside any one public hospital, or port and airport management for any single port or airport, although across all hospitals, ports and airports, there are ample opportunities for different private providers. Best practice in these areas is to create concession areas, identify service responsibilities, and then auction off rights to the concession for a certain number of years. In the case of different hospitals, ports or airports, the performance of different providers can be compared, creating more competitive conditions in these markets. The policy of assigning concessions is done irregularly in Ghana in waste collection, and, despite the private provision that characterizes the sector, not at all in urban bus transport Barriers to Private Sector Investment in Infrastructure in Ghana. Several obstacles confront expanded competition and private sector involvement in Ghanaian infrastructure. These barriers relate to government restrictions on entry, insufficient regulation of the rules of the game, unpredictable tariff-setting criteria, lack of cost coverage by tariffs, and financing risks (see Table 5.2). Many of these are within the power of the Governmento remove Government Restrictions on Entry. Competitive private participation in markets is hindered in Ghana by the absence of regulations that allow private sector participation. With few exceptions, Ghanaian law granting monopoly rights to public entities for the provision of most infrastructure services must be changed before private investment in infrastructure will be possible. Legislation permitting private entry into the telecommunications sector is closest to being passed.40 Until this legislation is passed, and regulations necessary for its implementation are developed, only entry into cellular communications can be considered relatively unrestricted. In other sectors (for example, power and water), government policy with respect to private involvement is still evolving, and no legislation is yet in place. Current power legislation allows Independent Power Producers (IPPs) to enter but without further legislation, IPPs will only be able to sell to the Volta River Authority (VRA), delaying the onset of competition in the market. 40 Afer the legislation was approved by the Parliament, the President has returned it with some observations.

73 53 Table 5.2: Barriers to Private Sector Investment in Infrastructure in Ghana Insuflldent Covernment Regulation of Unpredictable Lack of Cost Restrictions on Sector Ent* Rules of the Game 2 Tariff-Setting Criteria 5 Coverage by Tarifs 4 Ftnancinz Risk 5 Telecom Basic services High High High High High Long distance High High High Medium Medium Value added Medium Medium Medium Medium Medium Power Thermal generation Medium High Medium Low-Medium High Transmission High High Medium Low-Medium High Distribution High High Medium Medium High Transport Urban bus Medium Medium Medium Medium Medium Primary and secondary roads High High N/A N/A High Bridges and Highways High High Medium N/A High Port, Airport and Rail Services High-Medium Medium Medium Medium Medium-Low Water Urban piped network High High High High High Nonpiped systems Low Low Low Low Low Sanitation On-site sewerage Medium-Low Medium Low Low Low Piped sewerage and treatment High High High High High Waste Collection Medium Medium Medium High Low Sanitary disposal High Medium Medium Medium Low Notes: 1/ Government restrictions on entry: this is high when government policy does not permit private entry into the sector. 2/ Insusfllcent regulation of rules of the game: this is high when rules related to private sector ownership rights, interconnection charges, market definition, etc., are undefined. 3/ Unpredictable tariff-setting criteria: This barrier is highest when non-technical, undefined criteria substantially affect final tariff levels, and when there are no criteria to define allowable costs. 4/ Lack of cost coverage by tariffs: This barrier is high when current pricing does not cover (approximately) long run marginal costs. 5/ Financing risks: Given the levels of risk in Ghana, financing risks are high when sunk costs, and thereforexposure to risk, in the sector are high Insufficient Regulation of (Non-price) Rules of the Game. The removal of government restrictions on entry is a necessary but not a sufficient condition for competition and private participation in infrastructure. Regulations will be needed to establish key rules of the game. These relate to (i) interconnection charges, (ii) concession terms, and (iii) the enforceability of payments Interconnection relates primarily to telecom and power. The profitability of investments in power generation, in power distribution, and in basic and long distance telecom services depends on the difference between tariffs and the costs of interconnection to the power transmission grid or between local telephone exchanges. Interconnection rules are not yet established in Ghana, although the Government is drafting plans for the telecom and power industries. In developing these rules for telecom, the Government can consider one of several alternative interconnection rules, although experience with these is too recent to determine which is superior. In Britain, intermediate users of the Oftel network are charged no more than retail users of that network. Australia, on the other hand, seeks to reimburse telecom network providers only for the long run incremental costs incurred by network use. In making its choice, Ghana should ask

74 54 whether the rule will attract investment, whether it is clear and reduces the potential for dispute, and whether it places low demands on government expertise and resources Concession terrns, such as the length of the concession, the geographic market area, the permissible customers (to which high-voltage customers, for example, can IPPs sell directly), and the output prices, are an additional prerequisite of private sector entry into infrastructure provision. The length of concessions should be sufficient to give the private operator an incentive to invest in efficient technology. The geographic market area should be such that economies of contiguity and scale are optimized. One water company for most of Ghana clearly implies an excessively large market, a point recognized by the Government, while the absence of any market definition at all for urban transport and waste collection is clearly insufficient. Permissible customers and output prices are bound together by the issue of crosssubsidies. IPPs will have the greatest impact if they are allowed to sell to any customer. However, some potential customers of IPPs (for example, the high voltage customers of ECG) effectively cross-subsidize other users (for example, the low voltage customers of ECG). This market cannot be opened to competition without a simultaneous decision to forego the cross-subsidies Perhaps the most important rule of the game that governments need to enforce is the principle that users must pay for their consumption of infrastructure services. A substantial fraction of system losses in Ghana's water sector are likely due to unauthorized diversion. Cost recovery is likely to be difficult in waste disposal, as well, because Ghana has few mechanisms to compel individuals to utilize the waste disposal services that the government or private franchisers offer. The risk of sanction for illegal dumping is low, so it is difficult for private waste collectors or owners of sanitary disposal sites to recover their costs. If private investors have doubts about their ability to collect payment from users, they will hesitate to enter Unpredictable Tariff-Setting Criteria. Current government plans, particularly in telecommunications, call for substantial reliance on competition for the determination of tariffs. However, the experience of other countries indicates strongly that initial entrants may require a period of exclusivity during which price regulation will be required Current tariff-setting systems, however, suffer from two significant drawbacks. The first is the absence of price incentives for efficient production and the second is the politicization of the tariff-setting process. With the exception of the power sector (where a long-run marginal cost benchmark has been established), tariff changes are based simply on the difference between projected revenues and costs, largely extrapolated from previous year performance, providing little incentive for efficient management. Either a cost-based regulatory system (following the Chilean model), or a price-cap system (following the UK model), would be an improvement over this In choosing from among various regulatory options, Ghana will need to balance the demands of efficiency, the desired distribution of surplus between firms and consumers, and the Govemment's ability to undertake complex regulatory tasks. Price cap regulation is potentially easier to implemnent, but may allow for greater transfers from consumers to producers than is desirable. Rate-of-retum tariff-setting may reduce consumer to producer transfers, but is likely to place greater burdens on regulatory authorities and discourage more efficient production The second drawback of current tariff regulation is the politicization of the process. While parliamentary approval does infuse necessary transparency into the rate-setting process and allows users to influence rate outcomes, independent rate-setting agencies can accomplish the same end with greater predictability and reliance on economic criteria. Countries such as Hong Kong, Britain, France, USA, and

75 55 Singapore, all rely on highly qualified regulators, following well-established procedures for regulation and dispute resolution, with notice given when major policy changes are designed, and with opportunities provided to comment on those changes Lack of Cost Coverage by Tariffs. Procedures for setting prices that are transparent and predictable must also offer investors the assurance that regulated prices will cover their costs. Where the pricing regime fails to compensate investors for the costs of their investment plus a risk-adjusted rate of return commensurate with what they can eam elsewhere, private sector participation is unlikely to be forthcoming. The same is true if investors have to depend on government subsidies to compensate for low prices, or if they are concemed that the prices that provide them with an adequate return are not politically sustainable. The best indicator for investors of future govemrnment attitudes towards infrastructure tariffs is the current level of tariffs. In the power sector, Ghana has set prices so that they approximately cover the long-run marginal costs of producing and distributing electricity. However, prices currently charged by other govenmment-owned infrastructure providers fail to cover costs. In water, waste collection, telecommunications and urban bus transport, prices do not allow for capital replacement Financing Risks. All countries expose infrastructure investors to certain levels of sector policy, country, commercial, and currency risk. Where the amount of capital needed for a project is large, as is the case for many infrastructure sectors in Ghana, the impact of these risks is proportionately greater. In Ghana, most of the deteminants of sector policy risk are described in the discussion above on the other barriers to private sector entry. In addition, however, investors face country, commercial, and currency risks Commercial and currency risks are self-explanatory; a stable, balanced macroeconomic environment is a necessary condition for keeping them low. Country risk is high in countries where investors have doubts about the security of their investments from expropriation, or about the reliability of dispute resolution mechanisms that are available, particularly to settle disagreements with the state. It is unlikely that investors would perceive country risk in Ghana to be low relative to other countries that they.night be considering. Consequently, early investors into Ghanaian infrastructure are likely to require guarantees that minimize these risks. These might include guarantees with respect to demand for services, greater specificity about exact tariffs, and exclusive concessions rather than competitive markets The Govenmment can mitigate concems about country risk in several ways: first, through greater transparency and predictability in policy making and greater, more visible attempts to reassure investors regarding future governent policies; and second, by allowing firms to seek extemal guarantees, such as those provided by the Multilateral Investment Guarantee Agency (MIGA). Ghana already has some investments that are guaranteed by the MIGA facility Although commercial risks are ideally bome by the private investors, who are best placed to know the costs of production and to project demand, in practice efforts (such as government purchase guarantees) to offset sector and country risk are likely also to reduce the commercial risk bome by private investors. Although this outcome may be unavoidable, the terms of these concessions should be transparent, particularly to reduce the uncertainties of subsequent entrants (e.g., independent power producers) 41 Procedural guarantees and transparency, however, must be balanced against the need to make regulatory decisions in a timely manner. The regulatory process in the US tends to be the most time-consuming because it offers the most exhaustive opportunities to different interests to question regulatory proceedings and to extend them. Other countries give different interests the opportunity to be heard, but restrict challenges to regulatory decisions.

76 56 regarding the regulatory advantages that incumbent power producers might enjoy. After seeing the favorable experience of earlier investors, later entrants will be willing to accept less generous terms Because they are more comfortable in Ghana's political and economic environment, local and expatriate investors will require fewer guarantees, and so should be encouraged to participate in infrastructure development. This can be done, first, by listing some infrastructure projects on the Stock Exchange in Ghana; second, by considering the Jamaican example of an Infrastructure Investment Fund; and third, by strengthening local collateral laws and their enforcement. Expansion and improvement of urban bus transport is handicapped by the fact that banks do not accept buses as collateral, fearing correctly that they will be spirited across one of Ghana's international borders. The problem of moveable security is a difficult one to resolve; however, strengthening the ability of borrowers to use immovable security (land and buildings) as collateral would also be an important step towards increasing local investment Conclusions. Ghana faces major infrastructural inadequacies which constrain private sector growth. In the face of problems with the implementation of public investment projects and binding resource constraints of the Government, infrastructural investments will have to come increasingly from the private sector. International best practice is testimony to how private ownership and competition, combined with an appropriate regulatory framework, can lead to service expansion and productivity improvements in most infrastructure sectors. But, in Ghana, the current state of government ownership and management of infrastructure is a far cry from international best practice and numerous barriers to private sector investmnent exist most of which are within the power of the Governmento correct In the light of the above discussion, the following are the policy recommendations to promote greater private investment in the infrastructure sector: * Establish and follow transparent and predictable regulatory procedures and criteria; create regular, orderly channels of communication between regulatory decision makers and those affected by the decisions. * Accelerate the drafting of detailed regulations governing prices, conditions of interconnection, service obligations, concession areas, etc. for the power and telecom sectors. * Establish a timetable for creating and auctioning off concessions to provide local power distribution services in selected areas. * Remove pricing and entry regulations in competitive infrastructure sectors such as construction and maintenance of on-site sanitation systems and value-added services in telecommunications; but accelerate regulations for spectrum allocation for cellular services. * Increase prices to cover all costs of service provision (including the opportunity cost of capital) in service categories where private sector investment is possible (water, sanitation, urban bus transport, telecom). * Eliminate differences in tipping fees and other conditions of doing business that exist between government and private providers of on-site sanitation services. * Evaluate the possibility of establishing and auctioning off concessions to waste collection, urban transport routes, and water system billing and maintenance. * Dedicate resources to the enforcement of payment obligations of infrastructure beneficiaries, as in water.

77 57 Strengthening the Financial Sector Private sector growth in Ghana has been severely constrained by the low level of financial intermediation which restricts financing opportunities for productive private investment. Six years after the beginning of financial sector reforms in 1988, Ghana's financial system has remained shallow. Broad money holdings (M2) are equivalento 17 percent of GDP, as compared with 79 percent in Malaysia, 46 percent in Indonesia, 37 percent in Kenya, and 30 percent in Zimbabwe. The low level of financial intermediation derives from the low level of confidence in the financial system, insufficient competition, and weak financial infrastructure Confidence in the Financial System. Confidence in the financial system has suffered because of several policy interventions by the Government. In 1979, the Government undertook a currency conversion. In 1982, it demonetized the 50 note, froze bank deposits in excess of 50,000 pending investigation for tax liability, restricted bank loans for the financing of trade inventories, and required that business transactions in excess of 1,000 be conducted by check. Whatever the Govermment's intentions had been, these restrictions greatly undermined private sector confidence in the banking system and prompted substantial financial disintermediation. Although all financial restrictions that had been imposed prior to the ERP were subsequently lifted, public confidence in the financial system was slow to recover Also, until recently, real interest rates on deposits have been negative, giving individuals little incentive to hold savings in bank accounts. People prefer to keep their money in cash, so that currency holdings are proportionately large when compared with monetary aggregates. Meanwhile, a proliferation of high-return, low-risk government paper has been crowding out bank lending to the private sector. This has lead to high interest-rate spreads, a low rate of financial innovation, and poor-quality service throughouthe financial sector Competition in the Financial Sector. Ghana's formal banking sector is comprised of the central bank -- the Bank of Ghana --9 commercial banks, 3 merchant banks, and over 100 rural unit banks. The sector is dominated by state-owned institutions and shows few signs of competition. Non-bank financial institutions (NBFIs) are comprised of a stock exchange, 21 insurance companies, the Social Security and National Insurance Trust, two discount houses, the Home Finance Company, numerous building societies, a venture capital company, a unit trust, and a leasing company. Most of these are also in the public sector Financial Infrastructure. The formal banking institutions have been unable to provide much genuine intermediation between savers and investors. Demand deposits are the main source of funds for banks, and loans and advances account for a relatively small portion of their assets. High reserve requirements have prevented banks from building up their loan books, while hefty returns paid on government securities have afforded banks an easy life. Privately managed banks have been more adept and efficient at financial intermediation The NBFIs have yet to emerge as significant players in the financial system, in large part because of macroeconomic uncertainty, pervasive state ownership, and, until recently, the absence of defining legislation and adequate supervision arrangements. The (mostly state-owned) contractual savings 42 This section draws on the recent report - The World Bank, 1994, "Ghana - Financial Sector Review: Bringing Savers and Investors Together," Report No GH.

78 58 institutions have undertaken reverse term-transformation by investing in short-term government securities. Other than for real estate, NBFIs have provided very little medium- and long-term finance to the economy Rural financial markets in Ghana consist of formal institutions (such as branches of commercial banks and rural banks) and informal institutions, such as susu collectors. Both can play an important role in intermediating between savers and investors, but thus far Ghana's rural financial markets have not been able to provide people in the countryside with adequate savings opportunities or access to credit -- particularly in the agricultural sector. The development of rural financial markets has been limited by factors that are typical for Sub-Saharan Africa -- asymmetric information, absence of credit histories, high transaction costs involved in lending to smallholders and rural microenterprises, and lack of acceptable collateral. In addition, an unstable macroeconomic environment, along with relatively high inflation compared to low past levels, and lack of confidence in formal financial institutions, has led people and banks alike to shun formal financial arrangements in rural markets, and thereby to lose the benefits of intermediation However, some recent developments have been encouraging. Among the NBFIs, leasing companies, building societies, and savings and loan associations have been innovative in serving savers and borrowers. Also, even though the stock exchange has only 17 listed companies and low trading volumes, recent success in placing Ashanti Goldfields Company shares has sparked new interest in it. With the emergence of the Home Finance Company, housing finance also shows potential for rapid growth, and prospects for establishing primary and secondary mortgage markets are promising Although Ghana's informal financial sector is large, with an estimated 45 percent of all private sector financial savings mobilized initially through informal channels, its capacity to intermediate between savers and investors is limited, in part by people's savings behavior, and in part by the absence of strong links with the formal sector. There may be considerable opportunity for profitable contacts between the informal and the formal sectors such as bringing the savings generated by susu collectors into the formal system, and allowing traders to draw on formal credit by simplifying the use of collateral. However, uncertainty about the Government's view towards the desirability of unmonitored informal financial intermediaries has complicated these linkages Conclusions. Today, the main challenge for the financial sector is to bring more savings into the financial system in order to finance more productive private investments. The resultant financial savings and lower projected public sector borrowing requirements would allow private investment finance to increase sharply. The recently concluded Financial Sector Study has recommended the following measures to improve the sector's ability to provide financing for the private sector investment: * Revive confidence in the banking system and send a strong signal favoring private participation in the financial sector by proceeding quickly with the divestiture of state-owned banks. * Promote competition in the non-bank financial sector by remr;oving the virtual monopoly positions of the Social Security and National Insurance Trust and the State Insurance Corporation and by creating a level playing field where the private sector can compete effectively. * Support informal financial intermediaries by clearly stating that the Government encourages informal operators to join the formal system through an expansion of their business activity in order to qualify for registration under the Law. * Increase access to formal credit by strengthening the legal infrastructure for debt recovery and by expanding the range of usable collateral.

79 59 Enhance intermediation by rural financial institutions and, in particular, rural banks by enhancing Bank ofghana (BOG) supervisory capacities and by making changes in the institutional design of rural banks The expected gains from improved financial intermediation are large. An increase of private sector credit equivalent to three percent of GDP, in addition to the increase resulting from a reduction in public sector borrowing, is feasible by bringing more of already existing informal savings into the formal sector. While this would still leave Ghana's ratio of private sector credit to GDP well short of ratios recorded in many other countries, it would have an immediate impact on growth through improved efficiency of investment, and could raise long-term growth rates by increasing the depth of the financial system. These are essential to the promotion of private sector growth in Ghana. Liberalizing the Petroleum Sector 5.46 Petroleum is a key commodity in the economic development of Ghana. It represents the most important source of commercial energy. It is also the single largest import item -- and, thereby, the largest item of foreign exchange outflows -- comprising more than 10 percent of merchandise imports in 1994 and equivalent to about 15 percent of merchandise exports. Petroleum also represents a key source of Government revenues. Taxes on petroleum accounted for a quarter of tax revenues in 1994, equivalent to almost 4 percent of GDP The chain of operations in the petroleum sector -- exploration, procurement, refining, and distribution -- were entirely in the hands of the private sector until the early 1970s when the operations were turned over to the public sector. Currently, the Ghana National Petroleum Corporation (GNPC) is responsible for the exploration and procurement of crude petroleum, the Tema Oil Refinery (TOR) handles the refining, and the Ghana Oil Corporation (GOIL) distributes about one-third of the petroleum products. 43 In addition, the pricing of petroleum products is administratively done on a cost-plus basis rather than through an automatic formula which would serve to insulate the pricing mechanism from political pressures Exploration and Development. With respect to the risk of E&D, Ghana is not considered to be a geologic region of high prospectivity which means the cost of development is high whereas the potential rate of return may be low. This implies that the downside risk is considerable and the opportunity cost for public sector involvement in this area indefensibly high. For this reason, Ghana should encourage private risk capital into E&D, as many other countries have done. In Egypt, for instance, the public sector holds the exploration rights which are contracted out to the private sector on a production sharing basis. The private sector puts up all the risk capital while the public sector shares in the profit Procurement. Public procurement in Ghana is relatively efficient with the actual cost of petroleum imports quite close to import parity prices (derived from prices on international markets plus transport). This is quite in contrast to most other West African countries where prices (inclusive of carriage, insurance, and freight) are well above import parity levels and procurement is the most promising 43 See The World Bank, 1993, "Ghana - Energy Sector ReviewV', Report No GH. The remaining two thirds is distributed by private companies.

80 60 area of savings.44 But the cost to the refinery of the imported petroleum is higher than the procurement price since the GNPC adds on a purchasing fee and the cost of commercial credit. Hence, by inducing competition in the procurement of the crude petroleum, or by even just permitting the TOR to import crude petroleum directly, the costs to the refinery can be reduced significantly Refming. Petroleum refining offers scope for significant efficiency gains. Although the Tema Oil Refinery is competitive, there are gains to be made from reducing volume losses due to wastage, cutting administrative costs, and eliminating trading losses on the sale of fuel. 45 The higher refining costs of the TOR emerge from its sheltered existence whereby the price for the refined product that it receives is on a cost-plus tolling basis. But exposing the TOR to market competition would limit its margin to a net-of-cost basis from market-determined ex-refinery prices and thereby induce increased efficiency in its operations Distribution. With respect to petroleum product distribution, there is substantial private sector activity. But GOIL still commands a large share of the market. There is no rationale for the continued involvement of a public sector entity in petroleum mnarketing which can be efficiently performed by the private sector Retail Pricing System. Any petroleum pricing system should reflect true economic costs, be automatic, and provide the Government with a reliable stream of tax revenues. But in Ghana, none of these desirable attributes have been satisfied fully. The current retail pricing system for petroleum products is supposed to be on a cost-plus basis, taking into account the cedi cost for imported crude, and processing and other costs. But the administrative regulation of retail prices has made it difficult for the Government to pass on filly to the consumers the higher import cost due to the cedi depreciation. As a result, the Government has had to bear the cost of large foregone fiscal revenues. However, if the pricing system was made automatic, then this would preclude such problems Conclusions. The petroleum sector in Ghana suffers from significant economic inefficiencies arising from the public monopoly in the refining of petroleum. Private sector growth in the country can be promoted by opening up procuremento the private sector and by shrinking the size of public involvement in the sector. In addition, the administrative setting of prices has delayed adjustments and resulted in large foregone fiscal revenues for the Government. In view of the inefficiencies which exist in the petroleum sector, the following recommendations can be made: * Remove the monopoly of GNPC over the procurement of crude petroleum. * Commercialize the operations of the Tema Oil Refinery in the short-term. * Liberalize petroleum prices with negotiated mark-up margins for oil distributors and an ad valorem tax rate for excise taxes, and with frequent review of marketer's margins. * Privatize GNPC, the Tema Oil Refinery, and GOIL. 44 See Cuneo e Associati, 1992, "Petroleum Products Supply and Distribution in Sub-Saharan Africa," miimeo and Miguel Schloss, 1993, "Does Petroleum Procurement and Trade Matter? The Case of Sub-Saharan Africa," Finance & Development, March. The c.i.f. dollar prices for 1990 were found to be well above the import parity prices, especially in Cote d'lvoire (63 percent), Burkina Faso (44 percent), Guinea-Bissau (51 percent), and Senegal (42 percent). Ghana was the only country in the eight country study to have c.i.f. prices lower than the import parity price, albeit by only a small percentage (5 percent). 45 The Cuneo study (ibid.) estimated that the potential savings in refming for Ghana were $4 million for 1990.

81 The liberalization of the petroleum sector will, of course, need to take place under a legal framework which ensures fair trading, adequate insurance against liability, and the proper application of health, safety, and environmental standards. It will also need to be phased to allow the TOR to adjust to the changing competitiv environment. Removing the Distortions in Cocoa Trading Historical Background. Ghana was the world's leading cocoa producer during most of the 20th century. Ghanaian production reached a peak of 550,000 tons in 1964 with a world market share of 33 percent. But it began to falter thereafter. By 1975, production had fallen to 416,000 tons and world market share to 28 percent; by 1983, the comparable figures were 159,000 tons and 9 percent. Since then, however, production has shown a modest recovery, rising by 1993 to 250,000 tons and 10 percent of the world market share Of the cocoa producing countries of Africa, cocoa's decline in Ghana was mirrored partially by Nigeria (see Table 5.3). It is striking to note that the decline in Ghana and Nigeria took place even as world cocoa prices were high and increasing. C6te d'lvoire took advantage of the favorable international environmento increase its production from 180,000 tons in 1970 to 840,000 tons by In the African context, this compensated in absolute terms for the decline in Ghana's and Nigeria's production and African cocoa production increased from 1,098,000 tons in 1970 to 1,360,000 tons by 1993, or by over 35 percent. But relative to the world market, Africa's share still declined from 71 percent in 1970 to 55 percent in 1993, largely because of inroads made into the cocoa market by Indonesia and Malaysia The Ghanaian cocoa sector's decline is also reflected in the lower share of cocoa value-added in GDP and in the lower share of cocoa export revenue in total export. Cocoa's share in GDP fell from 14 percent in 1970 to 5.5 percent in 1980, and has since risen to 9 percent in It's share of export revenues has fallen from more than 75 percent in 1970 to 24 percent in Moreover, the cost to Ghana, in terms of lost export revenues, from its declining market share has been enormous. Had Ghana maintained its 1970 market share (28 percent) in 1993, then its total cocoa earnings at the 1993 price of US$1,232 per ton would have been US$850 million instead of the US$309 million received -- that is, US$540 million more in just Even if Ghana had only maintained its production level to that of 1970 (434,000 metric tons), then its cocoa export revenues would have been US$125 million more. These numbers are suggestive of the high cost the country has paid for the decline of the cocoa sector While the problems in the cocoa sector reflect both infrastructural inadequacies -- inadequate feeder roads and transport facilities, the incidence of the swollen shoot disease, and the lack of chemicals -- and policy distortions, the former are not new and Ghana was able to produce substantially larger amounts of crop in the past despite the prevalence of these problems. Over the years, however, the policy environment became very rigid and, as we shall see below, created distortions which do not provide adequate incentives for increasing productivity in the sector Distortions in Cocoa Trading. There are two main policy distortions in the cocoa sector, all of which relate to cocoa trading. First, the Government maintains a monopoly in the extemal marketing of cocoa and a near-monopsony in the domestic market for cocoa. And second, the cocoa sector is taxed 46 This section relies on a background paper by Liaqat Ali, 1994, "Distortions in Ghana's Cocoa Marketing," mimeo.

82 62 heavily, both explicitly and implicitly, with the result that the cocoa farmer receives only a small share of the producer price. Table 5.3: World Cocoa Production and Market Share (in parentheses) ('O0 MT) Country Crop Year ' Cameroon (7.2) (7.9) (7.3) (6.1) (4.6) (4.1) Cote dlvoire (11.6) (16.3) (22.3) (29.4) (32.1) (34.1) Ghana (28.1) (27.9) (16.8) (11.1) (11.7) (10.0) Nigeria (20.0) (14.4) (10.2) (7.6) (7.0) (5.5) Total Africa 1,098 1,010 1,023 1,106 1,418 1,360 (71.0) (67.8) (60.0) (56.0) (63.8) (54.9) Brazil (11.7) (17.3) (20.8) (18.5) (14.7) (12.0) Indonesia (0.1) (0.2) (0.4) (2.0) (6.0) (10.5) Malaysia (0.3) (0.9) (2.1) (6.6) (8.8) (8.9) World total 1,554 1,489 1,695 1,974 2,507 2,474 * Estimate. Source: International Cocoa Organization; Ghana Cocoa Board; and Gill and Dufus, conunodity brokers; World Bank staff estimates 5.60 Public Control on Cocoa Marketing. The Government maintains a complete monopoly over the external marketing of cocoa through the external arm of the Ghana Cocoa Board (COCOBOD), the Cocoa Marketing Company (CMC). The CMC buys the cocoa from the Produce Buying Company (PBC), a wholly-owned subsidiary of COCOBOD, and the private Licensed Buying Companies (LBCs). As we shall see below, the COCOBOD is a high-cost government monopoly and the inefficiencies inherent in the monopoly marketing of cocoa translate into lower potential producer prices for cocoa and thereby in reduced incentives for investments by farmers In the internal marketing, the COCOBOD has had a complete monopsony. But in 1992, as part of its efforts to liberalize the sector, the Government invited applications from the private sector for participation in the internal marketing of cocoa. Subsequently, licenses were granted and five LBCs participated for the first time in the 1993/94 crop season, in addition to the PBC. That year, the PBC purchased nearly 80 percent of the domestic production and the remaining 20 percent was bought by the LBCs The LBCs low share of the domestic market reflect problems of transition as well as of the inherent weaknesses of the current arrangements. First, the 1993/94 crop year being the first year of operations for the private LBCs, they had to overcome financing difficulties, poor transportation facilities in the rural areas, lack of infrastructure (such as weighing scales, tarpaulins, and storage space), and lengthy delays for cocoa quality control, in unloading cocoa at the harbor due to shortage of storage capacity, and in receiving payments from the CMC Second, the marketing margin provided to the LBCs has been very small. As long as the price that the LBCs have to pay the farmers and the price that they receive from the COCOBOD is fixed, the

83 63 incentives for profit maximization are very limited. In the 1993/94 crop year, the price for the producers was fixed at Cedis 308,000 per ton while the purchase price from the LBCs was fixed at Cedis 362,000 per ton. The difference between the producer price and the purchase price from the LBCs was thus Cedis 54,000 per ton. While the purchase price from the LBCs took into account the costs of labor, material and equipment, transport and quality control, the interest paid on borrowed funds, and a profit margin, some of these costs appear to have been underestimated and the marketing margin, therefore, inadequate. For instance, the financing cost was calculated at the rate of interest of 25 percent whereas all the LBCs paid more than 31 percent. Nor does the marketing margin allow for the costs incurred because of delays on part of the CMC. The narrowness of the margin allowed to the LBCs does not permit themn to undertake the large upfront fixed investments in storage facilities, hauling equipment and ancillary services which are required to start an operation and act as a barrier to entry. For the 1994/95 crop year, the Producer Price Review Committee has determined a producer price of Cedis 700,000 per ton and a purchase price from the LBCs of Cedis 117,000 per ton. The higher marketing margin of Cedis 117,000 per ton should make LBC activity more profitable Third, the LBCs faced financing difficulties. These arose on 3 counts. First, under the existing banking regulations, the commercial banks are not allowed to lend more than 25 percent of their net assets and not more than 10 percent to one individual client (if unsecured). 4 ' In the 1993/94 crop year, the Agricultural Development Bank (ADB) exceeded regulations on lending operations, and they were questioned by the Bank of Ghana. Second, commercial banks suffer from liquidity shortage and are therefore unable to meet the LBCs' demand on their own. In the 1993/94 crop year, the financial requirement of the LBCs was Cedis 18,100 million 48 ; in the 1994/95 crop year, this is likely to be at least Cedis 40,850 million. During 1993/94, the COCOBOD used its market power to borrow at cheaper rates in the international financial world and then lend it to commercial banks at around 25 percent for onlending to the LBCs which paid 30.5 percent. 49 Third, LBCs do not have enough collateral guarantees and, therefore, the commercial banks are reluctant to lend to them. Current arrangements do not allow LBCs to pledge their cocoa as collateral once it has been delivered to the CMC depot Excessive Taxation of the Cocoa Sector. Table 5.4 demonstrates the distribution of cocoa earnings among producers, the COCOBOD and the Government. Producers' share in the crop year 1993/94 averaged around 40 percent although a part of this compensation is still to be paid by COCOBOD. However, in 1994/95 crop year, with a 127 percent increase in the nominal price for cocoa, the producers' share is expected to rise to 51 percent, the highest share ever. In contrast to the Ghanaian situation, producers in Indonesia and Malaysia earn around 90 percent of the FOB price. Except for Ghana, cocoa producers everywhere routinely receive at least 65 percent of the free-on-board price One reason for the low share of the producer price in the free-on-board price is the inefficiencies in COCOBOD operations which imnply an implicit taxation of the sector. For the crop year 1994/95, the COCOBOD's budget has been determined at Cedis 78.9 billion or about 17 percent of the expected total 47 Single client exposure for secured loans is 25 percent. 4 For 1993/94, this is calculated on the basis of the producer price (including the internal marketing margin) of e 362,000 per ton, and on the assumption that LBCs bought 50,000 tons (approximately 20 percent of the crop). For 1994/95, the crop quantity purchased is kept the same, but the producer price assumed (including the internal marketing margin is C 817, The commercial banks charge 4 percent on top of 25 percent as commitment, monitoring, and guarantee fees, plus an additional 1.5 percent interest for printing and payment of checks.

84 64 earnings (Tables 5.4). In the crop year 1993/94, this was 19 percent of total earniings. While COCOBOD costs contain elements of non-marketing activities some of which, such as extension services and research, can be rationalized on the grounds that the market will underprovide these services -- as the positive social benefits from them will not be captured -- total costs are still high and impose an implicit tax on the sector. Table 5.4: Distribution of Cocoa Proceeds Calculation of Proceeds 1993/94 (actual)_ 1994/95 (estimate) Crop size (metric tons) 251, ,000 FOB price per MT (cedis) 1,206,754 1,564,030 Earnings (millions of cedis) 303, ,569 Distribution of Proceeds million cedis % of Total million cedis % of Total Total 303, , Producers' Share 123,126' , COCOBOD Budgetb 58, , Surplus (Govt's share) 122, , Source: Ministry of Finance (April 28, 1995) a/ Includes compensation (approx. one-quarter of the total) which is still to be paid by the COCOBOD. b/ COCOBOD's budget includes private buyer's costs 5.67 A second reason for the low share of the producer price in the free-on-board price is the excessive explicit taxation of the cocoa sector despite the cocoa farmers being among the average poor in Ghana (see Chapter 3). The explicit cocoa tax is obtained as a residual -- the entire surplus of the COCOBOD after deducting costs from total earnings. It, therefore, varies from year to year and in 1993 was 44 percent of the free-on-board price. In contrast to Ghana, no tax is imposed on cocoa in Indonesia and Malaysia, where a standard personal or company income tax applies to farm income generally. Nigeria has also abolished discriminatory taxation of cocoa by the federal government although there remains a separate levy disguised as inspection fees collected by the major cocoa-producing states. In Brazil and Cote d'lvoire, moderate taxation of cocoa earnings, amounting to about 15 percent, is imposed This is not to suggest that Ghana should abolish all taxation of cocoa. Ghana does indeed command about 10 percent of the world cocoa market and the theory of commercial policy is very clear in its prescriptions that where monopoly power in exports exists, national welfare is maximized by restricting exports through an optimal export tax. Several studies have sought to calculate the optimal cocoa export tax for Ghana. 50 Their estimates vary between 4 percent and 21 percent depending upon the assumptions used, particularly the calculated elasticities for export demand. But they clearly state the case for a reduction in the level of taxes -- both explicit and implicit -- imposed on cocoa Conclusions. The cocoa sector continues to suffer from two major policy distortions. These relate to the visible hand of the government in cocoa marketing and the excessive taxation of the sector. These distortions have been primarily responsible for the lack of investments and the general decline in output. Some policy measures, such as the provision of higher nominal producer prices, the liberalization of internal marketing, etc., have sought to address these concerns. However, the experience of the LBCs 50 See for instance hnran, M. and R. Duncan, 1988, "Optimal; Export Taxes for Exporters of Perennial Crops," Working Paper Series No. 10, The World Bank; Akiyama, T. and Donald F. Larson, 1994, "Adding-Up Problem - Strategies for Primaiy Commodity Exports in Sub-Saharan Africa," Working Paper Series No. 1245, The World Bank-; and Panagariya, A. and M. Schiff, 1990 "Commodity Exports and Real Income in Africa," Working Paper Series No. 537, The World Bank.

85 65 with intemal marketing reflect the inherent weakness of the current arrangements. Efifctive competition has been elusive and will remain so unless the price of cocoa is liberalized and private traders have direct access to export markets. Continued policy reforms in cocoa marketing are critical to restoring the health and vitality of the cocoa sector and in assisting Ghana in achieving accelerated growth The above analysis suggests the following recommendations: * Liberalize cocoa external marketing. The flow of benefits from such a measure would include higher producer prices for the cocoa farmers, the elimination of the inefficiency costs inherent in the monopoly external marketing of cocoa, and the creation of the incentives to promote rapid growth in the cocoa sector. * Reduce the taxation of the cocoa sector. Explicit taxation should be brought down to levels consistent with optimal cocoa export taxes for Ghana while implicit taxation should be completely eliminated. The latter would be an automatic corollary of a complete liberalization of the internal and external marketing arrangements. * Make the taxation of cocoa transparent and predictable so that prospective investors and the Government do not face uncertainty with respect to what their receipts are going to be. * Make the removal of infiastructural impediments to the cocoa trade - absence of feeder roads, poor maintenance, the high cost of transportation -- a high priority for the Government in its development agenda.

86 Chapter 6 MEDIUM-TERM PROSPECTS Ghana is expected to gradually improve its GDP growth rate over the mediumterm to reach 5.5 percent by 1997 and 6.3 percent by Growth is expected to be broad-based, deriving from enhancements in agricultural productivity, growth in manufacturing, and continued growth in gold mining and services. Export prospects are also good Investment and savings are also projected to rise, reaching 21.2 percent and 18.1 percent of GDP, respectively, by 2004 substantially narrowing the investment-savings gap. However, concessional external assistance, even if declining, will continue to play a significant role in Ghana 's economic development. Commitments of US$800 million per year are sought from donors for the period Program aid is expected to decline and a bigger share of project aid is expected to go towards human resource development and rural infrastructure. 6.1 Ghana's medium-term objective is to further deepen and extend the policy reforms under the Economic Recovery Program and promote private sector growth. These objectives and the Government's projections for the medium- and long-term are detailed in the Ghana -- Vision 2020 report (see Box 6.1). This chapter provides a brief overview of the macroeconomic prospects for the medium-term, , and the financing requirements over the three years, These projections are based on the expected pace of reforms. If, however, reforms proceed faster than expected, higher growth rates can be achieved. Macroeconomic Prospects 6.2 The medium-term macroeconomic forecast for Ghana assumes successful stabilization and strong policy performance (Table 6.1). GDP growth is projected to increase ftom 3.8 percent in 1994 to 5.5 percent in 1997 and 6.3 percent in The growth in GDP is projected to come from improved agricultural performance and growth in manufacturing. Faster growth can be achieved if policy reforms and public investment projects are implemented more vigorously and macroeconomic stability is restored more quickly and sustained thereafter. 6.3 Sectoral Growth. Improved productivity in agriculture is projected to raise agricultural growth rates to 3.1 percent by 1997, and 3.3 percent in 2004, so that per capita agricultural growth rates become positive. The recent liberalization of the agriculture sector (and the impending private participation in cocoa exports) as well as infrastructure investments to date are projected to generate productivity growth in both traditional and non-traditional products. Foreign investments in the processing of vegetables and fish, and domestic investment in horticulture, provide the first signs of a supply response in non-traditional agricultural exports. Nevertheless, the agriculture growth is projected conservatively, given uncertainties regarding the rate of adoption of new technology -- including better seeds, more fertilizer, and improved implements. Mining, especially gold, is projected to continue to grow rapidly over the medium-term. This is based on the output effects of recent investments in large-scale mining, especially gold. Manufacturing sector growth rates should also improve from 1.5 percent in 1994 to 5.8 percent in 1997 and 6.1 percent in 2004 on the basis of projected increases in private investment and in efficiency. These are predicated on the restoration and maintenance of macroeconomic stability, the acceleration of privatization (including the

87 67 divestiture of commercial banks), the approval of the new liberal Investment Code, and improvements in the dialogue between the Government and the private sector. The increase in tradable goods activity should support the growth of the service sector which is projected to increase from 6.3 percent in 1994 to 6.8 percent in 1997 and to 7.5 percent in Box 6.1: Ghana - Vision 2020: An Overview Ghana - Vision 2020 (The First Step: ) is a coordinated program of economic and social development policies for the five years, , that the President of Ghana presented to Parliament on January 6, This was in fulfillment of his obligation under Article 36 of the Constitution of the Republic of Ghana (1992). and is the first in a series of five-year National Development Planning Frameworks upon which Ghana's long-term (25 year) vision - to become a middle-income country - will be built. The report lays down the underlying goal of development as improving the quality of life of all Ghanaians by reducing poverty and raising living standards through a sustained increase in national wealth and a more equitable distribution of the benefits from it. The long-term objectives set out in the report focus on five themes: human development, economic growth, rural development, urban development, and an enabling environment. The basic goals of human development are to improve health and life expectancy, to expand and improve education and training, to have access to clean water, adequate sanitation, and affordable housing, and to reduce poverty and income disparities. Economic growth is seen to come from an open and liberal market economy that optimizes the rate of economic development and ensures the maximum welfare and material well-being of all Ghanaians. This is to be fostered through a strengthening of the enabling environment to encourage private investment through improvements in human development, the administrative and legal system, and the economic infrastructure. The report envisions a long-term growth rate of 8 percent per annum. Rural development is aimed at reducing disparities between the incomes and standards of living of the rural and urban populations. Urban development is aimed at ensuring that small and medium-sized towns and cities adequately fulfill their role as service centers of the rural hinterland and that the process of urbanization contributes positively to development. The medium-term objectives in the report provide guidelines for the formulation of strategies to achieve the long-term objectives and thereby consolidate the foundations for accelerated economic and social development. This is seen to be facilitated by the political reforms undertaken over the past five years -- decentralization, democratization, and the return to constitutional government. But a number of factors are identified as constraints to development. These include a high rate of population growth, inadequate child care, low levels of domestic savings and financial intermediation, low level of productivity, constraints on energy supply, low status of awareness of science, technology and the environment, and a biased international economic order. The guidelines are aimed at overcoming these constraints to development. The framework laid out in the report will be the basis for the preparation of medium-term five-year plans by ministries, sectoral agencies, and district assemblies. These sectoral and district plans will then be subjected to coordination, rationalization, and harmonization by the National Development Planning Commission in order to ensure mutual compatibility and consistency with the national development objectives. Out of this process will emerge the national Five- Year Rolling Plan for the period Source: Government of Ghana, 1995, Ghana--Vision 2020 (The First Step: ): Presidential Report on Coordinated Program of Economic and Social Development Policies, January. 6.4 Investment and Savings. Though higher foreign direct investment would be desirable, actual increases in such investment over the medium-term are unlikely to be a large source of investment financing. The increases in private savings and investment levels are projected conservatively and are feasible, given the levels attained in Private investment is projected to recover gradually, approaching its 1991 share in GDP of 8.1 percent by Thereafter, it is projected to increase further to 12.0 percent by This may be exceeded if private investment in the telecommunications and the power sector grows more rapidly than projected. Increases in private investment are expected to be financed mainly through increases in domestic savings. Both public and private savings are forecasted to recover so that gross national savings are expected to be 11.6 percent of GDP by 1997 and 18.1 percent by 2004 reflecting restrained growth in wage-related expenditures and a fall in interest payments. Public

88 68 savings are expected to increase gradually to 4.1 percent of GDP by 1997, reflecting the implementation of the Value-Added Tax, and to 6.7 percent of GDP by The improved macroeconomic environment, increased competition in the financial sector, and additional investments in financial infrastructure should stimulate private savings which are projected to increase to 7.6 percent of GDP by 1997 and to 11.4 percent by Table 6.1: Medium-Term Macroeconomic Prospects, Estimated Projected GDP Growth (% p.a.) Agriculture Industry Manufacturing Mining Services Gross Investment (% of GDP in current prices) Public Private National Savings (% of GDP in current prices) Public Private Narrow Fiscal Budget (% of GDP; Exports (% of GDP in current prices) Current Account Balance (% of GDP)b GDP Growth per Capita (% p.a.) Inflation (% annual average change in CPI) Source: Ghana Statistical Service; World Bank staff estimates and projections a/ Excludes capital expenditure financed through external project aid and the correspondin grants and loans. bt Excludes grants. 6.5 Exports. Ghana's export prospects are good. Future prices on the international cocoa market indicate a continued finning of demand conditions, while domestic production is responding favorably to the large real increase in producer prices which took place in Shipments of gold, which is now Ghana's major export, are also projected to grow healthily through the medium-term, albeit at more modest rates than were recorded over the past two years. In addition, electricity exports should rebound now that water levels on the Volta Lake have returned to normal. Nevertheless, export volume is projected to grow by around 6 percent per year during , a rate similar to that achieved in the past. This, coupled with upward movements in gold and cocoa prices, should permit growth in import volumes and improvement in the trade balance simultaneously. Non-traditional exports are projected to grow at a substantially greater rate than total exports. External Financing: Ghana will seek commitments of concessional financing of around US$800 million per year for the period. This will consist of US$600 million of project aid (including technical assistance) and US$200 million of program aid (including food aid). With the current pipeline of undisbursed balance, this translates into annual average disbursements of around US$400 million of project aid and around US$230

89 69 million of program assistance over the period. Annual disbursements in 1995 will be higher than the average reflecting the carryover of program aid from The aid pipeline at the beginning of 1995 is estimated at US$2.0 billion. Of this US$693 million is expected to be disbursed in 1995, US$608 million in 1996, and US$599 million in An annual average of US$800 million in new commitment is recommende during to meet the projected financing requirements in those years and to maintain a sufficiently large project aid pipeline. Most of the program loans for 1995 are already committed. 6.8 Aid Trends. The projected level of annual commitments for period is less than that for the period when commitments of concessional assistance amounted to US$900 million per year. Slightly more than 40 percent of project aid commitment has been from IDA. Disbursements of total concessional assistance is expected to be similar to that in , i.e. around US$650 million (see Table 6.2). Table 6.2: Ghana: Aid Disbursements, (in millions of US $) Official Development Assistance ' 1996P 1997P Gross Disbursements * As % of lnports of Goods and Nonfactor Services Source: World Bank a/ Grants plus concessionaloans. e/ Estimate. p/ Projected. 6.9 Concessional external assistance will continue to play a significant role in Ghana's economic development, even under fairly optimistic assumptions about exports and national savings. During , Ghana's current account deficit, excluding interest payments and official grants, is projected at US$1.3 billion (see Table 6.3); in addition, debt service obligations, including the reduction of arrears, are expected to total US$1.0 billion. The program also incorporates a targeted increase in gross international reserves of US$0.4 billion, equivalento 5.3 months of imports, by the end of On this basis, Ghana's total external financial requirement of is projected at US$2.6 billion The financing needs will be covered primarily by official grants and long-term concessionaloan inflows, including disbursements under a three-year ESAF arrangement with the IMF. It has been assumed that proceeds from the Government's divestiture program will amount to over US$110 million and direct foreign investment will amount to US$150 million during the three-year period Investment-Savings Gap. The external resource implications of the medium-term economic outlook can also be calculated as the gap between investment and savings. Ghana is expected to finance its rising investment increasingly from its own savings. The investment-savings gap is expected to fall from its peak in Private investment and private savings will also rise, given successful stabilization in 1995, from 4.4 and 2.5 percent of GDP, respectively, in 1994 to 8.1 percent and 7.6 percent, respectively, in Also, increased public saving is expected to help lower the investment-savings gap over the medium-term.

90 Program Aid. Program aid commitments (including food) of around US$200 million a year is recommended over this period. More than half is expected from non-ida sources. It is recommended that this program assistance be allocated through the interbank market and the transactions in that market be monitored on a regular basis. Such an approach will facilitate disbursements and private sector access, and provide for a more efficient allocation of resources. Given the dependence of domestic food production on the weather, food assistance can make a useful contribution to meeting financing requirements as it would substitute for commercial imports when food imports are needed Project Aid. Project aid commitments of about US$600 million a year are recommended for the period. The amount is similar to past levels and reflects the important role of public investment in infrastructure and human resource development. Such assistance could fill, in part, the financing gaps in road maintenance and telecommunications. Financing gaps also exist in the urban sector and the social sectors. It is recommended that donors finance a high proportion of the initial costs and, where possible, also finance recurrent and maintenance requirements, especially in the social sectors. Table 6.3: External Financing Requirements and Sources, (in millions of US S) Actual Estimated Projected L REQUIREMIENTS CurrentAccountDeficit Official MLTAmortization IMF Repurchases Reserve Requirements IL SOURCES Grants (Gross) Concessional IDA Project Program Other Multilateral & Bilateral IMf Purchases Private Capital Direct Foreign Investment Others (Divestiture, Suppliers' Credit etc.) Non-Concessional Short-term Capital Other flows (incl. Errors and Omissionsya Source: World Bank a/ includes net arrears, bilateral balances, other MLT liabilities and erors and reduction.

91 Role of Donors. Donors can contribute to the Government's efforts to promote private domestic and foreign investment in several ways. The continued availability of external support should raise the confidence of investors in the sustainability of the Economic Recovery Program. Implementation of welltargeted public investment projects that improve the provision of essential services, and thus reduce the costs of doing business in Ghana, would promote higher levels of both private domestic and foreign investment. Donors can also provide technical assistance to identify microeconomic distortions and to facilitate the private sector's access to factor and financial markets, improve entrepreneurial ability, and provide for a legal and regulatory framework which is conducive for business. Financing can be provided for support services for businesses, particularly export promotion. Bilateral donors, in particular, could make a contribution by helping to strengthen private business associations through technical assistance and training programs. Donors can ensure that projects with parastatals or the central government do not stifle competition, but promote private participation by avoiding areas that would crowd out private investors Composition of Aid. These projections incorporate a falling share of program and adjustment loans relative to the earlier period. This is, in part, because the World Bank is reducing the share of adjustment lending in its total comnmitments. Therefore, around 20 percent of the total disbursements of concessional financing for the period will consist of program loans (including food aid). Two factors influence the speed with which program or adjustment lending can be reduced by the Bank. First, the level of resource inflow has to be maintained. Sharp and sudden cuts in adjustment lending from the Bank would reduce net resource transfers to Ghana given the current rate of implementation of public investment projects. Hence, concerted efforts are being made to increase disbursements on investment lending. Second, it is expected that other donors will maintain their levels of program aid during the transition period. External Debt 6.16 Ghana has continued to service its debt without rescheduling though some arrears have emerged last year. At the end of 1994, external payment arrears amounted to USS93 million. The Government attaches the highest priority to the speedy regularization of relations with its external creditors in order to ensure that unimpeded access to new credits is maintained. These arrears will be cleared during the Policy Framework period through cash payments and rescheduling on a bilateral basis, and no new external arrears will be incurred. Ghana will not seek debt relief under the auspices of the Paris Club Ghana's external public debt outstanding and disbursed (excluding IMF obligations), at the end of 1994, stood at US$4.8 billion, nearly 80 percent of GDP. Bilateral creditors account for nearly 23 percent and multilateral creditors for about 70 percent (see Table 6.4). The rest is owed to private creditors Although, the stock of debt is growing each year, the outlook for the debt service burden is better. The debt service burden (excluding IMF repurchases) falls to 15 percent in Also, the stock of debt continues to be on concessional terms.

92 72 Table 6.4: External Debt Service, (in millions of US $) Actual Estimated Projected Concessional & Non-Concessional Debt Service Amortization Concessional Non-Concessional Interest IMF Debt Service Repurchases Charges Total Debt Service Debt Service Ratios(%) Excluding IMF IMF Memorandum Item Exports ofgoods and Services' Source: World Bank; IMF Note: Figures are rounded. a] Sum of Exports of Goods and Non-factor services, Factor Receipts and Workers' Remittances Conclusions 6.19 Ghana is expected to gradually improve its GDP growth rate over the medium-term to reach 5.5 percent by 1997 and 6.3 percent by Growth is expected to be broad-based, deriving from enhancements in agricultural productivity, growth in manufacturing, and continued growth in gold mining and services. However, concessional external assistance, even if declining, will continue to play a significant role in Ghana's economic development. Commitments of US$800 million per year are sought from donors for the period. Program aid is expected to decline and a bigger share of project aid is expected to go towards human resource development and rural infrastructure.

93 Chapter 7 CONCLUSIONS Policy reforms since 1983 have shifted relative prices and incentives towards the tradables sector and have promoted distributionally-neutral growth. Growth has led to a sizable reduction in poverty over The reduction in poverty has been broadbased, benefiting most the rural areas, and has been driven by growth in private-sector services, especially wholesale and retail trade. This has occurred despite low levels of and investment in, human capital and despite generally poor targeting of the social expenditures; further growth with poverty reduction is contingent upon human capital not becoming a binding constraint to growth. In addition to investments in human capital, growth needs to be promoted by investments in infrastructure, the attainment of macroeconomic stability, reforms in the financial sector, the liberalization of the petroleum sector, and the removal of distortions in the cocoa sector. With these measures, Ghana will be able not only to sustain, but also accelerate growth with poverty reduction. For this, concessional externalfinancing will continue to be needed, even if at declining levels. 7.1 Policy reforms undertaken under the Economic Recovery Progran have reduced the fiscal deficit and inflation, improved infrastructure, and shifted relative prices and incentives towards the tradables sector, in general, and towards exports, in particular. The resulting improvements in macroeconomic stability have made it possible for farms and finns to respond to the shift in production incentives induced by the policy reforms and the economy has turned around. Although economic activity witnessed its biggest surge during the early years of the ERP (5.3 percent annually during ), aggregate growth has averaged 4.7 percent since Even this modest rate of growth has led to a sizable reduction in poverty in Ghana. Between and , poverty fell from 36.9 percent of the population to 31.5 percent. What is striking about this is the widespread nature of this poverty reduction. Rural poverty fell from 41.9 percent to 33.9 percent, accounting for most of the poverty reduction. As a result, the poverty differential between the rural and urban areas has improved from 14.5 percentage points in to 7.4 in The large fall in rural poverty strongly suggests that economic growth has effectively 'pulled up' the rural poor who have gained proportionately. Vulnerable economic groups have also benefited from the poverty reduction. Poverty among female-headed households declined dramatically from 38.5 percent to 28.5 percent, reversing the poverty differential between the sexes. 7.3 While a definitive analysis of the reasons for the change in overall poverty and in poverty over different economic and social groups remains to be completed, ongoing work confirms that economic growth accounts for the bulk of the reduction in poverty. It is also clear that this growth has been distributionally-neutral in the aggregate, so that the positive impact on poverty of an increase in mean expenditures has not been offset by adverse changes in income distribution. This is significant as in many other countries, the reduction in poverty is typically accompanied by a worsening in income distribution. 7.4 The decline in rural poverty appears largely due to growth in non-farm self employment activities, much of which is service-based -- wholesale and retail trade, in particular. The growth in trading activity reflects the strong growth in both exports and imports during the period. This is also an area of activity which is almost wholly dominated by the private sector.

94 But can future growth be as broad-based as past experience? Can the poverty reduction then be as sizable? Concern over this arises from the low human capital base of the poor. Levels of human capital in Ghana are very low, on average. For the poor, then, human capital attainments must be even worse. This is not being helped by the aggregate levels of spending on health and education which appear to be generally poorly targeted. While the poor may have benefited from the growth process in the past, human capital constraints may well become binding in the future and undermine the potential gains in poverty reduction. 7.6 But regardless of the broad-based character of future growth, there is also the more fundamental question of whether future growth itself can be sustained. Concern over this arises from three sources. First, past growth has been shown to be driven by growth in services. But this growth in services may not be sustainable. If the high rate of growth in services is, in part, a recovery from the compressed levels prior to the ERP, and, in part, fueled by aid, then the prospects for sustainability are dim. On the other hand, if service growth is complementary with growth in tradables, then rapid growth in exports will continue to spur service growth. Growth in tradables is predicated upon growth in the agriculture and the manufacturing sectors. However, a firm answer to the sustainability of service sector growth must await a careful and detailed investigation. 7.7 Second, rapid and sustained future growth is likely to remain elusive unless agriculture grows rapidly, too. The potential for agricultural growth through productivity increases is considerable. This can be expected from both existing crops, where the current yields are exceedingly low, and from new highvalue crops (e.g. in horticulture), where yields are high. Agricultural exports can play a major role in that growth. While the exploitation of such potential was hampered in the past by delayed policy reforms in agriculture and by slow improvements in marketing and transport infrastructure, recent reforms and investments in infrastructure place agriculture in a better position for faster growth. In any case, further economic and sector work is planned to analyze the reasons for poor productivity growth, in general, and for inadequate adoption of technology (e.g. use of high-yielding seeds, fertilizer, and better implements), in particular. 7.8 Third, human capital constraints may not only limit the benefits of growth to the poorer segments of the population but also impinge on the growth process itself. The diffusion of technological improvements in agriculture, and the growth of a manufacturing base, among other things, are all critically dependent on the availability of educated and skilled manpower. Bigger and better investments in basic health and education are, therefore, essential to obviate this ostensible constraint to economic growth. 7.9 But even if human capital constraints are lessened, future growth will need to be stimulated by other short- and medium-term policy actions. Short-term policy relates to the attainment of macroeconomic stability through the avoidance of any wage shock and the prudent control of monetary growth. Mediumterm policy measures require mobilizing private participation in infrastructure, strengthening the financial sector, and liberalizing the cocoa and petroleum sectors. These measures will go a long way towards promoting private sector growth which, given the dominant contribution of the private sector to GDP, is essential if economic growth is to be substantially accelerated and sustained The Government is committed to moving forward on the above fronts to promote private sector activity. Several policy measures over have also been supportive of greater private sector involvement. These have included an accelerated privatization program, improved dialogue with the private sector, further liberalization of the regulatory framework, and growth of the stock exchange.