Using specific purpose grants to achieve developmental goals: some practical considerations

ABSTRACT This practice note reviews the theory of conditional grants and their use in five countries (Rwanda, Kenya, South Africa, Uganda, and India). The note makes the case for conditional specific purpose grants based on theory and then puts forward five practical suggestions that can guide developing countries in designing specific purpose grants. These include setting clear conditions, establishing robust monitoring systems, developing sub-national capacity, being cognisant of sub-national autonomy, and simplifying the transfer system as best possible.


Introduction
Decentralisation has been trending in developing countries since the turn of the millennia, and alongside this, intergovernmental transfers from central to sub-national governments. It is estimated that intergovernmental transfers account for about 65% of total local government revenues in developing countries, compared to 15% in developed countries (Hart and Welham 2016). Given this trend, questions around the relative share of specific purpose and/or conditional grants versus unconditional general grants become important.
The increasing popularity of intergovernmental transfers has been met with some challenges. There is growing evidence that without adequately taking account of the incentives/disincentives created, large increases in grants can cause unforeseen problems with local government performance and longer-term sustainability (Hart and Welham 2016). For example, transfers might not be spent on national priorities and/or crowd out local contributions. Given this, many have advocated for specific purpose grants or conditional grants as they can enhance fiscal equalisation at the sub-national level (Kalirajan and Otsuka 2012;Martinez-Vazquez and Sepulveda 2018), improve economic efficiency (Martinez-Vazquez and Sepulveda 2018), and promote greater accountability and service delivery (Faguet 2014;Hart and Welham 2016).
A pertinent question for policy makers which follows is: how should such specific purpose grants be designed? This is precisely the type of question that concerns developing countries. I was involved in investigating potential answers to this question as part of a technical assistance programme to the Government of Ethiopia. As part of the assignment, a comprehensive review of the theory of conditional grants and their use in five countries (Rwanda, Kenya, South Africa, Uganda, and India) was undertaken. The present practice note makes the case for conditional specific purpose grants based on theory, and then puts forward five practical suggestions that can guide developing countries in designing specific purpose grants.
What are specific purpose grants?
Grants can be earmarked for a specific purpose vs non-earmarked; conditional (where a set of criteria must be met for disbursement) vs non-conditional. They can also be mandatory (set by law or regulation) or discretionary.
Non-earmarked transfers provide additional revenue to carry out fiscal programmes. They ease the budget constraint of the sub-national government and increase the budget available for all expenditure (Figure 1(a))what economists term a pure income effect (Stiglitz and Rosengard 2015). On the other hand, earmarked or specific purpose transfers increase the fiscal capacity to carry out a specific programme areafor example, designated spending on health. Not only do earmarked transfers provide income, but they also change the relative price of the targeted programme, making it "cheaper" for the region/locality to implement relative to other types of spending. This pivots the budget line implying "more" health can be purchased relative to other programmes (Figure 1(b)).
Earmarked grants can also have conditions attached to them and be linked to performance-based grant systems, where transfers from central to the sub-national government are conditioned upon performance of the latter. Typical examples of performance measures include revenue collection, budget execution, mechanisms of accountability, financial management, and good governance (World Bank 2017, 8). Another common condition employs a matching component, where the sub-national government contributes funds alongside central government.

The case for specific purpose intergovernmental transfers
The case for earmarked transfers can be separated into three arguments around equity, economic efficiency, and accountability ( Figure 2). Each is discussed below.

Equity
Many countries opt for a decentralised system as it assigns greater responsibilities to sub-national levels of government. This has the potential to increase responsiveness to local needs and participation in decision making (Faguet 2014). On the other hand, decentralisation can also lead to inequities in services received by citizens of the same country. Both non-earmarked and earmarked grants can be used to address these inequities. For the present discussion, emphasis will be placed on earmarked transfers. This is because earmarked grants can be used to target places where expenditure Fiscal disparities may arise from varying revenue raising abilities as well as differences in cost structures across regions/localities. Transfers from richer regions to poorer ones via central government can promote horizontal equity (Martinez-Vazquez and Sepulveda 2018). Furthermore, intergovernmental transfers can be used to reduce inequalities around access to services. For example, earmarked transfers from central to sub-national governments can enable jurisdictions with lower fiscal capacity to provide new opportunities for residents. One can imagine an earmarked transfer to a jurisdiction with relatively low water access, where the region has to use the transfer to improve access to water (specific purpose), and may also be required to allocate counterpart funding if an expenditure matching grant is used (conditional). This type of transfer can lead to equity in access to opportunities across regions (Faguet 2014), but not necessarily equity in outcomes. The latter, equity in outcomes, is contingent on access to services; but also utilisation and uptake by citizens, as well as state/region-specific characteristics (Kalirajan and Otsuka 2012).

Economic efficiency
Another argument for conditional specific purpose transfers, and particularly expenditure matching grants, is to correct for the under provision of some services at the regional/local level that generate positive fiscal externalities to other regions/localities and central government. A fiscal externality of this nature is one in which spending by a jurisdiction accrues benefits to the fiscal system as a whole (Martinez-Vazquez and Sepulveda 2018). However, each jurisdiction does not consider third-party benefits when allocating resources, and as a result, under-provides some services. For example, Figure 2. Benefits of specific purpose transfers regional/local spending on primary and secondary education contributes to developing human capital and a capable workforce. These workers become employed and pay income tax to the central government, generating wider (third-party) fiscal and social benefits, which the region/ locality does not consider in choosing how much to invest in education.
From an efficiency perspective, the central government may wish to incentivise more spending on education as a means of maximising national welfare and long-run tax revenue. One way of doing this would be via an expenditure matching grant. A non-matching specific purpose grant that does not require a matched contribution but stipulates spending on a given programme can also be used. However, issues may arise because of fungibility of income and the non-matching earmarked transfer may crowd out or reduce regional allocations to the given programme (Hart and Welham 2016). An expenditure matching grant can address the externality, while simultaneously overcoming the "crowding out" effect. This, in effect, incentivises regions to internalise the positive fiscal externality.

Government accountability
Earmarked transfers can also help to improve both federal and local government accountability by creating stronger incentives for public service delivery than what would ordinarily exist (Faguet 2014;Hart and Welham 2016). For instance, decentralisation can promote political competition among jurisdictions where politicians compete to offer "better public services at a lower corruption-price" (Faguet 2014, 4). With respect to national accountability, earmarked grants can be used to help achieve national development goals (Hart and Welham 2016). By targeting specific spending programmes at the sub-national level which fit into wider national goals, the central government can be more accountable to the citizenry. Earmarked grants can also be used by the central government to ensure regional service delivery is improved, which also strengthens accountability to voters.
The case studies I was tasked with reviewing the use of specific purpose grants in a selection of countries to provide advice to the Government of Ethiopia who are considering re-designing their specific purpose grant system. Five countries were selected using a purposive sampling strategy. These include Kenya, South Africa, India, Rwanda, and Uganda. These countries were selected as they have comparatively different systems of intergovernmental transfers, a range of relationships between central and sub-national governments, and are at different stages in the decentralisation process. Various articles and reports which document the experiences of these countries were consulted, and from this review, practical suggestions were extracted. These are presented in the next section.
Before proceeding, Table 1 summarises the five cases, including the type of decentralisation and the system of intergovernmental transfers. The four African countries are all unitary states, while India is a federal state. All countries have some type of conditional transfer, alongside some unconditional grants. In all cases, conditional grants are used to target spending in certain sectors such as health, education, or agriculture, for example.

Practical design considerationslessons from the case studies
Five practical lessons have been extracted from the five case studies. These relate to setting clear conditions, establishing robust monitoring systems, developing sub-national capacity, being cognisant of sub-national autonomy, and simplifying the transfer system as best possible. Each is discussed in turn.

Clear conditions must be set for conditional transfers
There should be clear agreement between central and sub-national governments on the conditions of the grant. In Kenya, this includes input-based conditions such as a specified expenditure type, output-based conditions such as achieving a set level of service delivery under a performancebased system, and/or a grant-matching requirement where the sub-national government agrees to contribute an agreed percentage (Boex and Smoke 2020). In Rwanda, specific purpose grants target set sub-accounts, and performance contracts or imihigo are agreed with district-level mayors (Chemouni 2016). India requires states to commit a matching grant, which varies based on the level of development in the state (Rao 2017). Uganda uses a performance-based financing approach (World Bank 2017). South Africa legislates conditional grants under the annually updated Division of Revenue Act (Savage 2020). The Act requires all grant programmes to publish a clear framework setting out rules for allocation and use, measurable objectives, disbursement schedules, reporting requirements, and records of past performance.
Conditional transfers can vary in how specific they are. For example, a block conditional grant may require the sub-national government to spend on health broadly, giving the region/locality some autonomy on which parts of the health system to tackle. A more specific condition would indicate that spending should be on vaccinations, for example. The latter may work better when the central government has incomplete information on costs and expenditure needs as it may improve accountability and be less prone to problems with soft budget constraints. On the other hand, such specific conditions are often correlated with high monitoring costs. In all case study countries, conditions have been credited with improving services delivery.
Additionally, conditions can (and should) be set so that there is alignment with national development plans, and the specific sector-level master plan if the grant targets a particular sector. In so doing, the grant is used to both improve service delivery and meet national objectives. Government structure System of transfers

India
Federal system. India has a three-tier federal structure with governments at the union, state, and local levels. The former two are primarily responsible for service delivery (Rao 2017).
There are two main types of transfers: general block grants and specific purpose transfers. In 2016, the ratio between general and specific purpose was 70:30 (Rao 2017

Robust monitoring systems need to be established
Given that clear conditions are agreed, a robust monitoring system is required where results can be documented and tracked. Although Kenya has a strong framework for setting conditions, documenting results is one of the weaknesses of the system (Boex and Smoke 2020). This means that service delivery gains and progress toward development outcomes could be further enhanced. In contrast, Rwanda has drawn on the traditional imihigo system of performance contracts to enhance accountability (Chemouni 2016). Imigiho targets are linked to performance metrics and scores. Monitoring systems have improved significantly in South Africa. Here, monthly and quarterly sub-national government reports on revenue, expenditures, and non-financial performance are submitted to relevant sector departments and the National Treasury, who in turn submit quarterly reports to Parliament (Savage 2020). A strong monitoring system improves intergovernmental accountability and service delivery at the sub-national level. It can also work to improve efficiency and equity across a locality as robust monitoring can be used to inform annual value for money assessment.

Sub-national capacity should be considered
An assumption underlying decentralisation is that sub-national governments can deliver services, given access to finances. Evidence from India and South Africa suggest that regional-level implementation capacity can often constrain service delivery (Roa 2017; Savage 2020). This implies that effort must be placed on strengthening local capacity when intergovernmental transfer systems are being established/reformed. This may be done through a conditional grant similar to the Capacity-Building Grant in South Africa, which aims to develop management, planning, technical, and budgeting skills in municipalities (Savage 2020).
A balance with regional autonomy is needed Some have argued that earmarked transfers, if excessively used, can limit regional spending autonomy. For instance, 80% of transfers to districts in Rwanda are earmarked for precise activities, with only 20% left for discretionary spending (Chemouni 2016). Conditional transfers should therefore be cognisant of the larger decentralisation agenda, which often aims to give more power to sub-national governments. It should also aim to limit any potential "crowding out" of regional spending. For example, if regions are responsible for education, an earmarked grant can be complementary by focusing on the poorest schools or poorest children. In so doing, the fiscal autonomy of the regions is preserved, but pro-poor spending is enhanced.

Keep it simple (where possible)
Specific purpose grants have been particularly equalising in the case of India, with noteworthy examples like the National Health Mission, the Universal Elementary Education Programme, and the Mahatma Gandhi National Rural Employment Guarantee (Rao 2017: 15-24). However, the government has adopted some 28 initiatives under the Centrally Sponsored Scheme, and another 45 central sector schemes (Rao 2017). Such a vast number of schemes compete for limited resources and results in a thin spread of funds, which may affect service delivery. This case may be specific to large federal systems; though, as a general guide to avoiding these issues, governments should simplify/streamline the transfer system and limit the number of interventions within a particular grant to a manageable number that is easy to monitor and cognisant of regional capacity.