A Note on Public Expenditure and Poverty Reduction in Ghana

By Abena D. Oduro

Centre for Policy Analysis

Accra.

Revised Version of the Paper Presented at a Workshop on MACROECONOMIC STABILITY, GROWTH AND POVERTY REDUCTION IN GHANA

Organised by ISSER and CEPA, in Collaboration with Cornell University

May 2001.

A Note on Public Expenditure and Poverty Reduction

Introduction

Empirical investigations into the determinants of poverty in Ghana find that it is correlated with an inadequate stock of assets (physical and non-physical), limited access to infrastructure and basic utilities and services (Asenso-Okyere et al, 1997; Canagarajah et al., 1998; Boateng et al, 2000). The primary school net enrolment rate in 1998/99 is estimated to be 83%. Amongst the poor the net enrolment rate is much lower at 73%. In the rural areas 35.4% of the population still obtain water from natural sources, i.e. rivers, lakes and ponds. Some districts in Ghana do not have a hospital. In rural Ghana about 37% of the population is estimated to be living within 30 minutes of a health facility. The poor therefore face severe constraints in their ability to undertake income generating activities that will yield income levels that allow them attain the minimum desired standard of living. They are handicapped in taking advantage of perceived opportunities. In addition, they have a limited capacity to protect themselves from risk and to reduce the impact of shocks that they may experience. In Ghana poverty is concentrated largely in rural communities. Rural poverty is estimated to contribute approximately 90% to national poverty. A feature of rural poverty is the extent of subsistence production. This raises two issues. The first is the existence of functioning rural markets. The second is improving the rural poor’s access to markets.

Ghana’s poverty reduction strategy must therefore aim at building up on the poor’s stock of assets, improving their access to infrastructure, basic utilities and services and the creation of well functioning markets. These actions will facilitate an improvement in the productivity of the poor and increase the opportunity for income generating activities. Public sector expenditure has an important role to play in shaping and influencing some of the determinants of poverty. The tax system, i.e. taxes and tax reliefs can be used as instruments to address some poverty issues. However the impact of tax instruments is limited by the fact that a large section of the population operates within the informal sector and many fall outside of the tax net. Second the tax system is not an effective instrument to deal directly with some of the causes of poverty, i.e. low primary school enrolment rates, inadequate stock of well-equipped health centres, clean water and access roads. Public expenditure actions will be more effective in addressing these problems. The tax system is important to provide funding for the necessary public expenditures.

How Public Expenditure impacts on poverty reduction

Public expenditure can have a mitigating effect on poverty through:

The provision of infrastructure and services to the poor

Creating the conditions that will enhance the ability of the poor to accumulate assets

Facilitating the creation of institutions that will reduce the incidence of risk facing the poor

Reduce the impact of negative shocks through the provision of safety nets.

In countries such as Ghana, public sector spending is an important component of the poverty reduction strategy. At the same time the economic reform programmes require restraint in the growth of public expenditures. Public expenditure programmes for poverty reduction must include a strategy on how finances will be generated to fund the programme. This is to prevent the emergence of large budget deficits that will create economic instability and dampen economic growth. In designing public expenditure programmes for poverty reduction the first step should be to ensure that the total expenditure is consistent with macroeconomic targets. The second step is to identify the sources of funding for the expenditures. The third step is to allocate public resources so as to have the maximum impact on poverty reduction given the budget constraint.

From a poverty perspective however, the concern should not only be with the quantum of public expenditure but also with the components or structure of such spending. Even though rising public expenditure may have a negative effect on economic growth it is recognised that it may also have a "crowding-in" effect that encourages and indeed facilitates the expansion of private sector investment and therefore growth.

For public expenditure to be poverty reducing the following is required:

Provision of adequate resources for activities that will be poverty reducing. It is the consensus that spending on basic education, primary health care, feeder roads and rural water and sanitation are necessary types of expenditure for poverty reduction. In assessing whether public expenditure is pro-poor an examination of trends in the allocation of public resources to these sub-sectors is usually undertaken.

Education’s share of the central government budget is estimated to have declined from 23.4% in 1992 to 16.5% in 1999. This decline in spending on education has occurred at the same time that the share of interest payments has risen from about 12% to 22.6%. Cursory examination of trends in central government spending suggest that there has been a crowding out of poverty related spending by the rising interest payment obligations.

The share of basic education in central government’s education budget has declined in the 1990s. This has translated into a decline in the level of the unit subsidy for basic education between 1994 and 1998 (Canagarajah and Ye, 2001). Central government spending on health has not suffered as severe a decline in its share as has education. It’s share of spending declined from 7.6% in 1992 to 6.7% in 1999 (ISSER, 1999).

Provision of quality services. A sub-standard quality of service provided will reduce the demand for the service and translate into a waste of resources. Evidence abounds in Ghana and other developing countries that the utilisation of subsidised sub-standard services provided by the public sector can be lower than the utilisation of more expensive but better quality services provided by the private sector.

Efficiency in the use of resources. Improved efficiency implies a reduction in waste and therefore being able to obtain more output from the available inputs. This requires improved timing in the supply of inputs and ensuring that resources do reach the intended destination. The quality of the service provided is directly linked to the issue of efficiency in resource use.

Target the poor. Evidence suggests that public spending in Ghana is not adequately pro-poor. It is because the subsidies going to sectors that will benefit the poor are not sufficiently taken up by the poor. Canagarajah and Ye (2001) using benefit incidence analysis find that between 1992 and 1998 the share of the secondary and tertiary education subsidy going to the poorest quintile has improved. However the system still favours the richest quintile. There was a small decline in the share of the primary education subsidy going to the lowest quintile, whilst the share going to the two highest quintiles increased between 1992 and 1998. It is obvious that there is a need for better targeting of pro-poor expenditures. The spending programmes need to be adequately designed to ensure that the poor are encouraged and able to make use of the subsidised service or facility that is being provided.

The Determinants of an Effective Poverty Reducing Public Expenditure Programme.

The public expenditure system needs to be revamped to ensure that public expenditures will indeed achieve the objective of reducing poverty. Several mechanisms need to be put in place and actions taken to ensure that the required resources are made available, there is efficiency, improved quality and targeting of the poor. These mechanisms and actions will be discussed below.

Information. More data needs to be made available to guide the design of a public expenditure programme for poverty reduction. The data is needed to inform policy makers during the prioritisation process as well as to assist in the design of projects that target the poor. It is important that realistic targets are set and that programmes and projects are chosen that will have the maximum impact on poverty reduction.

Data must be differentiated on the basis of location, gender and income. Although the poor are to be found in all regions of Ghana, there is a high concentration of poverty in the three northern regions. Estimates of poverty rates using data from the fourth household survey conducted in 1998/99 suggest a headcount index of 69%, 88% and 84% in the Northern, Upper East and Upper West regions of Ghana. The national average is estimated at 39% (Statistical Service, 2000). An examination of data on infrastructure and services indicates that there is what may be correctly described as a historical bias in the allocation of public resources away from these regions. The low stock of infrastructure and supply of basic utilities can partly explain the persistence of poverty in these regions and the difficulty in registering significant declines in the incidence of poverty despite the years of continuous positive national economic growth. To illustrate the point further Table 1 below provides data on two disadvantaged districts in Ghana, one in the south and the other in the northern part of the country. There is a clear difference in the match between the population and the resources that have been provided by the public sector in the two. The Bongo District with twice the population of the Juaboso-Bia district does not have a hospital and has only one health centre. The number of public primary schools is less than 20% of what has been provided in Juaboso-Bia. The size of the district and the high cost of getting resources to a dispersed population cannot explain this disparity. The evidence seems to suggest the population is less dispersed in the Bongo district than in Juaboso-Bia. Thus it would be less expensive to provide resources in that district.

Table 1. Comparing Two Districts in Ghana

Notes: MOH Ministry of Health

JSS Junior Secondary School

SSS Senior Secondary School

Source: District Profiles prepared for the National Poverty Reduction Programme, Ghana.

For these disparities to be removed there is a need for region and district level data to be used to inform the allocation of central government expenditures. The formula or criteria used to allocate spending on a region and district level must be made publicly available and be clearly articulated. Regional differentiated data on central government expenditure must be made readily available at each of the ministries, departments and agencies. Gender differentiated data must also inform the allocation of resources and the design of projects. The Statistical Service and the data collecting departments of the various ministries and departments must be mandated to collect and present the gender, location and income differentiated data that is required on a regular basis. District Assemblies must also be encouraged to provide on a regular basis, district level data on variables that will facilitate poverty monitoring. The information should be kept both at the district level and at the Statistical Service.

Prioritisation Prioritisation is important within the context of limited resources. It is crucial that at the budget formulation stage priorities are set. For the priority setting process to be effective the available funding is made known. The prioritisation process will be undermined if during the implementation process, priorities are not adhered to due to unplanned cuts in fiscal spending and/or political interference. The Medium Term Expenditure Framework provides a useful framework from which to implement a poverty reducing public expenditure programme. It has been adopted by Ghana since 1999. Unfortunately the MTEF has not been successfully implemented in Ghana. The process has been hampered by several constraints, i.e. poor co-ordination between the Ministry of Finance and the National Development Planning Commission, problems in assessing available revenues and donor funds and resource constraints related to costing the activities.

Participation It is the consensus that if projects and programmes are designed without taking into account the views, needs and constraints facing the intended beneficiaries the expected results may not materialise. The necessary institutional structures must be in place to facilitate the contribution of the intended beneficiaries in the design of the projects and programmes.

In Ghana the decentralisation process is the likely institution where this participation could occur. The District Assembly has the potential to be the appropriate framework to ensure participation of the poor. However a number of risks exist that can undermine its effectiveness. The local elite whose interests may not necessarily coincide with that of the poor could capture the district assembly. The second is the risk of not having an effective monitoring and evaluation mechanisms at the district level. Related to this is the risk that the poor may not be able to sanction members of the district assembly. The third risk is that of inadequate financial and human resources at the district level. If the responsibilities of the district far exceed the resources allocated to it from central government the ability to effectively tackle poverty reduction will be severely constrained.

Several of the District Assemblies are prone to these risks. Reports made in the newspapers suggest that even at the district level spending is not adequately targeted at the poor. The weak or non-existent accounting and monitoring systems result in the misappropriation of funds. The identified weaknesses of the District Assemblies need to be addressed if they are to be the institutions whereby increased participation of the poor is to occur.

Earmarking Revenues. This is a method that is used to protect certain types of expenditure. Thus the concern about adequate funding for poverty reduction may be addressed by earmarking a certain proportion of the central government and district level revenues for that purpose. Earmarking revenues has been adopted in Ghana. The District Assemblies Common Fund is one of such funds. It supports spending at the district level. The District Assemblies Common Fund (DACF) came into existence after the enactment of Act 455 in 1993. The Fund consists of monies allocated by Parliament under section 2 of the Act and any interests and dividends accruing from investments of monies from the Common Fund. Under Section 2 of the Act, Parliament is required to allocate not less than 5% of the tax revenues of Ghana to the Fund.

The success of earmarking revenues will depend on a number of factors. The first is the extent to which central government is committed to ensuring that the earmarked funds are released. If there are no mechanisms to ensure that the allocation of the stipulated amount is automatic and/or if there are no sanctions facing a government that does not release the funds then it is likely that there will be slippage when it suits central government as was the case in Ghana in 2000 when funding was only released to the Districts for the first quarter and half of the second. Article 252 of the Constitution of Ghana states:

There shall be a fund to be known as the District Assemblies Common Fund.

Subject to the provisions of this Constitution, Parliament shall annually make provision for the allocation of not less than five percent of the total revenues of Government of Ghana to the District Assemblies for development; and the amount paid into the District Assemblies Common Fund in quarterly instalments.

It would appear therefore that in 2000 there was a contravention of the Constitution of Ghana when there was a failure to make the quarterly instalments. Neither the District Assemblies nor any other member took the Government up on the matter or presented it to the Supreme Court for a pronouncement.

The second factor is the need to ensure that the earmarked funds do not outlive their usefulness. Earmarking or special funds limit the amount of discretionary funds in the budget. Over time the special funds can contribute to a misallocation of resources if they are not adjusted or removed altogether in the event of changing circumstances.

In the case of Ghana, it is critical that policy makers at the centre are able to remain in tune with the evolving needs at the district. Earmarking of district level spending by the centre may not always coincide with district level needs. The district assemblies are provided with guidelines on the utilisation of the District Assemblies Common Fund. These guidelines are provided by the Minister of Local Government in consultation with the Minister of Finance. The guidelines issued in 1999 earmarked approximately 50% of the DACF monies. At the district level 20% of the common fund is earmarked for "poverty reduction schemes". This requirement was introduced in 1997 because it was noted that the projects being funded by the DACF were not those that directly promoted an expansion in production and employment. Thus in order to " …remove the obstacle of lack of access to credit for self-employed micro, small and medium scale entrepreneurs and to promote the development of micro, small and medium scale enterprises…constrained by lack of access to formal finance….District Assemblies may allocate 20% of their respective yearly shares of the District Assemblies Common Fund to create a line of credit in each District". The poverty alleviation fund is therefore a credit line. However, in 1999, very few of the districts met the requirement. At the regional level the proportion of the DACF that was allocated to poverty alleviation was less than the stipulated 20%. In many instances this was by a wide margin. It is important to understand why most districts were not able to meet the requirement. The failure to meet the 20% target may be because there is no demand for credit on the terms provided by the DACF, applicants cannot meet the requirements to access the credit, or else credit is not the binding constraint of employment and production. Although no answers have been provided here what the evidence suggests is that prior to earmarking revenues it must be clear that there is a demand or need for the service the funds are being earmarked to finance. Again the opportunity for a misallocation of resources occurs if the earmarked funds do not serve the needs of the intended beneficiaries.

Table 2. Disbursement of DACF by Region

Source: District Assemblies Common Fund.

Monitoring and Evaluation There must be in place an effective monitoring and evaluation system. This is needed to ensure that the priorities are being met and that spending is not biased away from agreed upon poverty reducing activities. Input flows must be monitored to ensure that supplies are made at the right time and in the appropriate quantities and prices. Public expenditure tracking surveys are needed to monitor actual delivery. Such surveys contribute to ensuring efficiency in the use of resources as well as ensure that the resources are reaching the intended beneficiaries. Third, outputs must be monitored. This is necessary to inform the setting of priorities and the identification of potential beneficiaries. Many projects have suffered because of weak monitoring and evaluation. In the agriculture sector for example it was found that the monitoring and evaluation offices of most projects are poorly staffed. They need the support of the implementing agencies to collect and process the information that is gathered. Unfortunately the implementing agencies are also handicapped in terms of skills and motivation to do this effectively.

Apart from the formal monitoring and evaluation arrangements that must be made during project design, an informal monitoring and evaluation mechanism can be created. This can be done by providing information to the general public, especially the intended beneficiaries on the objectives of the project or programme, the expected outcomes and the implementation procedures. When the public is armed with information, it is easier to assess actions and at the level of the District Assembly ask the necessary questions. Transparency is an important component of any monitoring and evaluation system.

Co-ordination of Projects Poverty reducing public expenditures must be an integral part of the public expenditure programme. This is necessary to reduce the incidence of conflicting policies and duplication of effort. Thus poverty reduction programmes must not be separate from the general business and budgets of ministries, departments and agencies.

Efficiency Measures It is important to have the right mix of the different types of expenditure (i.e. personal emoluments, other recurrent expenses and investment spending) to ensure efficiency in the use of resources and quality of the service provided. At the project design stage all the inputs for the entire life of the project should be costed. Thus for example, if there should be a school building project it is necessary that the project should not only focus on the cost of the building but also provide costing for the teachers, books and learning materials and maintenance expenses. A school building without teachers and/or without necessary teaching and learning aids is tantamount to being a white elephant and a waste of resources. In the search for resources to finance the project the focus should not be on the component of interest, for example the school building but on the totality of inputs needed for the project to operate successfully. This type of budgeting is a feature of the MTEF process. Unfortunately resource constraints in the way of equipment, trained personnel and other resources have made costing a difficult and lengthy process.

Will a Pro-Poor Public Expenditure Programme be at the Expense of Economic Growth?

Achieving an increase in economic growth is a core component of any poverty reduction package. An increase in growth rates for a given set of tax rates should be accompanied by an increase in revenues and therefore an increase in resources available for spending on poverty reduction programmes. If the economy is not growing the ability to increase public spending for poverty reduction purposes will be severely constrained. There are various options open to government to finance public sector spending in a context of slow or no growth. However most of these options have costs associated with them that will undermine the poverty reduction programme. Borrowing from the banks to finance the excess spending will be inflationary and is likely to hit the poor the hardest. Borrowing from the non-bank public sector puts upward pressure on interest rates. This will discourage investment and can have an adverse impact on the growth of employment opportunities. Recourse to borrowing from abroad could increase the indebtedness of the county unless the funds are borrowed on highly concessional terms. Finally the government could consider increasing the tax rate or the tax base. Increasing the tax rate could erode support for the poverty reduction programme and increasing the tax rate could draw the poor into the tax net. Thus the pro-poor programme must be designed to ensure that maximum growth targets given the targets for poverty reduction are attained. It is for these reasons that an increase in economic growth rates should form a core part of the poverty reduction strategy. Economic growth reduces the costs associated with a poverty reduction programme that requires an increase in public sector spending.

It is quite possible that if poverty reduction is not an objective the patterns of public expenditure will be different and the maximum growth rate could be higher in the short run. If society values poverty reduction it may be willing to trade-off some amount of growth to ensure a reduction in poverty. Pro-poor expenditure policies need not be associated with zero or declining growth. Indeed in countries like Ghana many of the poverty reducing spending programmes may actually be growth promoting so that there is no sharp trade off or conflict between poverty reduction and growth. For example spending programmes that are geared at improving the productivity of farmers and increasing the marketable surplus will be both poverty reducing and growth promoting.

A Proposal for a Pro-Poor Expenditure Programme for Ghana

Ghana’s public spending programme for poverty reduction should have a rural and agricultural bias. This is because the poor are to be found largely in rural communities and they tend to be food crop farmers. Public spending should be geared towards reducing agriculture production and marketing constraints and improving productivity. It should also aim at increasing the opportunities for non-farm employment and income generation. The spending programme must have a component that aims at improving the education and health status of the poor and aim at closing the gap that exists between the three northern regions and the rest of the country in terms of the provision of clean water, education and health services and economic infrastructure.

In order to obtain support for the poverty reduction programme and bearing in mind that there are pockets of severe urban poverty it is necessary that there are poverty reduction projects that benefit the urban areas. This is because the data suggests that a significant amount of revenue is generated outside of the agriculture sector. If tax revenues are to be raised in order to finance the increase in expenditure it is likely that a large proportion of the increase will be borne by the formal sector that tends to be located primarily in the urban areas. The urban population is also the most vocal and easy to organise.

Targeting is essential because of resource constraints. However for effective targeting it is necessary to have more data than is available from the household surveys. Within regions poverty is not evenly spread and it is necessary to have district level and sub-district level information to identify the intra-regional variations in the incidence of poverty, and on the distribution of economic and social infrastructure.

There is very little in the way of safety nets provided by the public sector. It is proposed here that a pension scheme should be developed for persons aged 60 and above irrespective of whether they have worked in the formal sector. It could be designed to provide a pension to individuals who do not receive a state pension. The pension could be set at a percentage of the equivalent of the monthly minimum wage. Potential beneficiaries could provide basic information on their age, gender, occupation etc to district or sub-district offices. At the same time information on recipients of state pensions could also be provided to the districts to allow them cross check on information to prevent recipients of a state pension receiving it twice. Since many people aged 65 and above in rural communities do not live alone this may be a means of providing a safety net to rural households in particular. For some rural households the old age pension might be the only significant source of income during the hunger season.

References:

Asenso-Okyere, W.K., N.N.N. Nsowah-Nuamah and P. Alberson (1997). "Characterising the Poor in Ghana: A Logit Approach", in Sustainable Food Security in West Africa, ed. by W.K. Asenso-Okyere, G. Benneh and W. Tims, Kluwer Academic Publishers, Dordrecht.

Canagarajah, S., D. Mazumdar and X. Ye (1998) The Structure and Determinants of Inequality and Poverty Reduction in Ghana, 1988-92. Policy Research Working Paper, The World Bank, Washington D.C.

Boateng, K., L. Boakye-Yiadom and A.D. Oduro (2001) Poverty in Ghana A Research Report Presented to the African Economic Research Consortium, Nairobi.