Skip to navigation – Site map

Results-Based Aid: Potential and Limits of an Innovative Modality in Development Cooperation

L’aide axée sur les résultats : atouts et limites d’une modalité innovante dans le cadre de la coopération au développement
La ayuda basada en los resultados: posibilidades y límites de una innovadora modalidad de cooperación para el desarrollo
Stephan Klingebiel and Heiner Janus


Results-based aid aims to identify outputs or outcomes that can be measured and quantified – that is, results that can be directly linked to development activities. A contract between the donor and partner country stipulates that for every incremental success a set amount (’reward’) will be paid. So far, experience of such aid programmes in practice remains limited. Sometimes they are a refinement of performance-based budget supports, but in other cases it has been suggested to incentivise single indicators (i.e. pay per pupil completing school). While this concept can be promising in theory, its practical implementation might turn out to be difficult. For instance, there is a risk of creating adverse incentives because all available resources might be focused on achieving just the one indicator to the detriment of other development objectives.

Top of page

Full text

1. Introduction

1In principle, all development cooperation approaches are aimed at achieving ‘results’. The international debate on results-based approaches differs from previous debates in that development cooperation in practice often focuses on inputs and processes. In many cases, aid resources are intended for investment (in the construction of schools, for example) or advisory services (for the education sector, for instance), but it is rarely possible to show ‘results’ in terms of measurable improvements in the well-being of people (i.e. in the sense of outputs and especially outcomes). Instead, aid success is recorded through input and process indicators, showing, for example, whether expenditure on a country’s education sector is increasing or whether agreed reform documents (such as a sectoral strategy for education) have been approved.

2Although documenting reforms and expenditures indicates how development activities in a partner country can be assessed, the information content is limited for two reasons. First, it is often unclear whether the intended results have actually been achieved. For example, are more children being educated and what is the quality of education that graduates have received? Second, what did development aid contribute exactly? If results have been achieved, are they causally linked to the development cooperation activity?

3Results-based approaches seek to identify quantifiable and measurable results that can be attributed as directly as possible to the effects of development cooperation.  This makes them a promising approach for implementing post-2015 development goals, where the link between aid intervention and development results needs to be documented. Results-based approaches aim to establish a close link between aid disbursements and incentives to encourage development results. These results-based approaches are envisioned as overcoming several drawbacks of input-oriented aid, such as ‘no clear result evidence line’, heavy transaction costs, and the bypassing of partner countries’ national systems. Knowledge on and understanding of conceptual underpinnings and actual experiences in implementing results-based approaches is, however, still limited (Klingebiel, 2012; Perrin, 2013; Janus, 2014). This paper therefore adopts an exploratory approach in order to better understand the concept of results-based approaches and document initial experiences.

Overview: A Brief Literature Review

4Research on results-based approaches relates to several literature strands on foreign aid, including literature on the impact of foreign aid, aid allocation and conditionality, and aid modalities. The empirical literature on the impact of foreign aid often focuses on either demonstrating a positive (Clemens et al.; 2004, Minoiu and Reddy, 2010; Addison et al., 2013) or negative effect of aid on economic growth (Easterly et al., 2003; Rajan and Subramanian, 2008). However, evidence of the overall impact of aid on economic growth remains mixed.  In addition, there is more nuanced analysis on whether aid promotes human development, especially in the areas of education and health. Here, substantive proof exists that aid can reach intended beneficiaries and can provide them with key services, which increase human welfare. (Fielding et al., 2006; Morrissey, 2010; Kenny, 2011). Depending on the size and level of certain results-based approaches, research into them can contribute to the debate regarding the macro-effects of aid, including effects on growth and human development.

5Literature on the impact of foreign aid also increasingly focuses on micro-level interventions that measure the effects of individual aid interventions within a confined environment. Especially, rigorous microeconomic impact evaluations, often including the use of randomized control trials, have demonstrated the potential for well-designed project interventions to generate development results (Duflo and Banerjee, 2011; Karlan and Appel, 2011). Small-scale, results-based approaches therefore strongly relate to research that focuses on demonstrating and comparing the measurable effects of individual aid programmes.

6Another strand of relevant literature analyses aid allocation and conditionality. Donors have a history of making aid disbursements conditional on policy reforms in recipient countries. There are broadly two generations of conditionality in the literature. First generation conditionality was aimed at reform of economic policy, mainly market liberalisation as promoted by the World Bank and International Monetary Fund through structural adjustment programmes (Stokke, 2004). Second generation conditionality in contrast is more closely related to political conditionality, often linked to accountability, transparency, and recognition of human rights (Van de Walle, 2005; Collier, 2007). As part of second generation conditionality scholars have argued that aid is more effective in the right policy environment (Burnside and Dollar, 2000) and proposed ‘performance-based allocation’ of aid (also called aid selectivity). Based on those findings the USA introduced the Millennium Challenge Corporation (MCC) in 2004; research in this regard indicates that aid allocation can support the reform-mindedness of recipient countries and that an "MCC effect" exists (e.g. Öhler et al., 2010; Parks and Rice, 2013).

7Yet, there are also indications that aid conditionality possesses only limited effectiveness (Collier and Dollar, 2002; Killick, 2004; Temple, 2010), not least due to the inherent motivations of donors to disburse aid budgets, thereby undermining the effectiveness of their conditionality. In response to this challenge donors attempted to move towards ex-post conditionality (Svensson, 2003; Booth, 2011), meaning that donors should switch from disbursing against pre-specified menus of policy changes to the allocation of aid on the basis of periodic overall assessments of government results. This modification to second generation conditionality might avoid issues such as moral hazard (e.g. interruption of reforms after the aid transfer has taken place) and follows a disbursement logic, where the partner country is expected to deliver results first before resources are transferred. Research on the contract character of aid is devoted to those aspects (e.g. Azam and Laffont, 2003). Such an approach also promises to provide fiduciary assurances to donors but also requires a better understanding of the links between specific policy actions and expected medium-term outcomes (Koeberle et al., 2005). While the idea of ex-post conditionality has been debated in the academic literature for some time, it is only recently that aid programmes have turned this thinking into practice in the form of results-based aid (also RBA).

8One important distinction between second generation conditionality and conditionality in the context of results-based aid is the nature of conditions. Second-generation conditionality focuses on political conditionality aspects, meaning conditions that are not directly related to the specific aid intervention. Aid is used to incentivise reforms of the broader political system beyond the scope of donor-funded interventions in specific sectors. Conditionality in the context of several results-based aid approaches, however, explicitly addresses technical triggers closely related to the outcome of a specific aid intervention. For example, disbursements for education and health programmes would be based on the independent assessment of ‘school completers’ or people vaccinated. The idea behind this more technical type of conditionality is to avoid politicised donor–recipient discussions, and focus on implementation challenges. Whether this theoretical distinction between political conditionality and technical triggers is feasible also in practice will be revisited in the final chapter of this paper.

9Lastly, results-based approaches relate to literature and policy debates on the effectiveness of different aid modalities (Tilley and Tavakoli, 2012). The debates on aid effectiveness as presented in events in Paris, Accra (Wood et al, 2011) and Busan have a set of standards and principles in order to make aid more results-oriented. Results-based aid is one major attempt to apply these aid effectiveness standards in a new modality for organising aid relationships between development partners and partner countries. Yet, it is disputed whether results-based aid is able to overcome traditional weaknesses of other aid modalities and whether results-based aid creates additional challenges or not (see, for example, de Renzio and Woods, 2008; Pereira and Villota, 2012).

10Overall, there are open questions regarding whether results-based approaches can improve the impact of aid, including on the macro- and micro-level, whether results-based approaches can set effective conditions for partner countries to reform, and whether results-based approaches represent a modality that conforms to principles of effective aid. These questions are pertinent to the overall future aid discussion in the context of a post-2015 development agenda. Achieving the post-2015 sustainable development agenda will require spending significant aid resources, especially in the poorest countries. The amount of resources that will be channelled through results-based aid and its developmental impact will critically depend on the potential and limits of this approach.

11In order to assess the potential and limits of results-based aid this paper proceeds in four steps. First the concept of results-based aid is defined. Second, potential advantages and disadvantages are introduced. Third, the paper presents initial practical experiences, including the case of results-based aid for fiscal decentralization in Ghana and Tanzania. Finally, the paper draws conclusions.

2. The Concept


12Definitions are a challenge in the debate on results-based approaches since there is no common terminology. The key feature of these approaches is that payments are only made once a pre-defined result has been achieved. In this regard results-based approaches differ from other aid approaches where funds are used to finance specific inputs for achieving results (e.g. schools to improve education, medical equipment to improve the health situation of the population, etc.)

13Several competing terms and concepts are used in the discussion on results-based approaches (output-based aid, performance-based aid, results-based funding, etc.) Yet, there is a growing international consensus on definitions (see, for example, Pearson, 2011; de Hennin and Rozema, 2011; Birdsall and Savedoff, 2011). This paper focuses on results-based aid (also RBA) and provides the following definition: in general terms, results-based aid is a partnership between a development partner (donor) and a partner government (recipient). A main feature of results-based aid is based on the introduction of an ex-post conditionality concept: a contract between both partners that defines incentives to produce measurable results. Aid disbursements or non-disbursements are directly linked to these independently verified measures of results. If these results have been achieved, the aid disbursement will be released; and if they have not, the aid disbursement will not take place. It is necessary to agree upon a ‘unit price’ in advance (e.g. how much aid is provided per student passing the final exam). Donors are not involved in the implementation process (‘hands-off’).

Key Characteristic

14The main characteristic of results-based aid is the link between aid intervention and strong incentives to encourage results. The main underlying assumption is that, in the past, aid approaches focussed mainly on inputs or processes, and only in some cases on outputs. Examples of aid interventions that are directly oriented towards results are rather rare. Typical aid interventions, for example in the education sector, focus on providing the necessary inputs for achieving a desired result. Inputs in this regard might include providing advice to the ministry of education in order to develop a new educational concept or a strategy for increasing school enrolment rates.

15On the investment side, an input-based intervention might consist of funding new primary school buildings or establishing a specific education budget target as a minimum share of the total national budget (e.g. in the context of sector budget support). However, providing inputs does not always lead to the desired results. For example, even with a lot of consulting and investment in school buildings, school enrolment rates and the individual educational achievements of children might not increase. Reasons for this possibly include incentives on a household level to keep children at home or a ministry in charge of education that has had no real political will to implement an effective sector policy that ensures that schools are staffed with adequately trained teachers and equipped with teaching materials. And even if the enrolment rate did improve after the donor intervention, it might not be possible to determine whether this success can be attributed to the donor intervention or whether it would have occurred as well without the donor’s support.

16Results-based aid aims to address this challenge. At least on the conceptual level, the link between the donor intervention and the objective aspired to in terms of measurable results is a close one, since the donor intervention provides strong incentives for results.

17In the present paper, results are defined as the direct and indirect effects of inputs and activities. We can distinguish between different levels of results. Outputs are normally technical results (for example, a newly constructed school). An output might lead to the next level: outcomes (for example, increase in enrolment rates because new school facilities are available). The most ambitious level of results is impact. Impacts are defined as wider developmental effects (for example, poverty reduction due to improved educational outcomes).

18Still, it might be difficult to always make a clear distinction between different categories along an impact chain. For instance, depending on the point of view adopted, an increase of a sector budget share might be defined as an input (e.g. to give more priority to a sector) or as a result if a larger budget is seen as an end in itself (Pearson, 2011, 4).

Figure 1 – Results-Based Aid Impact Chain

Figure 1 – Results-Based Aid — Impact Chain

Source: Author’s own representation

19Results orientation and results management in the framework of aid have many different forms. On a technical level, aid agencies have developed a number of tools for focussing on results (see, e.g. World Bank, 2011a, 35; CoP-MfDR, 2011; OECD and World Bank, 2008). Examples include results-focus in strategies, results-oriented planning and operations tools, and monitoring and evaluation systems focussing on results. However, there is a distinction between results-based management that is a cross-cutting tool for aid agencies to organize their work and results-based aid that refers to a specific financing modality.

Phasing of Results-Based Aid

20Results-based aid is organised in three phases with a clear division of labour between the aid provider and the aid recipient. The first phase of results-based aid is the preparation and finalisation of a contract between a development partner and the partner government. This step is crucial in several respects. Both contract partners have to identify an area, sector, or a specific objective that is important for the development process of the country and suitable for results-based aid at the same time. The selection process presents several challenges. For example, social sectors such as education and health quite often receive more support from donors than do other sectors. If a high level of support in a given sector is available, it might be difficult to identify a relevant ‘result’ since a number of other interventions are being funded already. In addition, an results-based aid programme could compete against other interventions in certain cases. For instance, a partner country that does not achieve the results agreed with one donor could try to substitute this financing gap with funding from another donor. Here the incentives set by results-based aid would be diluted. Other important considerations in the preparation phase are related to the duration of the contract and the sustainability prospects of the supported area (e.g. the question of how the partner country is going to deal with the situation if and when donor support comes to an end).  Also, contract partners have to agree on a measurable result and an appropriate indicator or set of indicators. A key task is to ensure that baseline data is available or can be collected. Further, the process for data collection and data analysis has to be discussed.  Finally, a ‘price per unit of progress’ has to be identified. The contract partners have to discuss and agree upon a performance level for results that is appropriate (At what level do we reward, for example, ‘additional’ students passing the final exam? Is the level achieved last year an appropriate starting point, or should we use an average of the last few years? etc.)

21The second phase is characterised by the implementation of the activities necessary to achieve the results. The nature of the activities could vary. One major bottleneck might be inadequate funding for a task, and the government might now be willing to provide more resources. Perhaps insufficient capacity is a major obstacle and the government would now agree to take specific remedial action (additional staff, training for staff, implementation of a retention strategy, etc.). Other possible obstacles for achieving results might be related to an overall power game within the government. However, as aid disbursement now depends on results, this could impact on internal decision-making processes. Most importantly, the partner country is in charge of the whole implementation process.

22The third phase is based on an assessment of the progress made. This should normally be carried out by a third party in order to ensure high-quality and incontestable data. The data will serve as the basis for the calculation (price per unit of progress) of the aid disbursement, since incremental progress is to be rewarded. The progress assessment is to be carried out on a regular basis (e.g. annually).

3. Potential advantages and challenges

23Practical experience of results-based aid programmes is limited (see Klingebiel, 2012; Perrin, 2013; Janus, 2014). In the area of budget support, the European Union has been using an results-based aid-type instrument in the form of variable tranches, which are based on performance agreements. The result indicators used in this context relate in particular to the health, education, and water sectors. Experience with performance-based tranches in major budget support approaches has been positive, since an incremental payment procedure makes development aid inflows predictable to some extent and acts as a performance incentive. The USA’s MCC programme similarly adopts a resulted-based approach in that aid allocations for different countries are based on meeting the requirements set out in a list of indicators. The United Kingdom’s Department for International Development is currently implementing a results-based aid pilot in Ethiopia in the education sector. In 2012, the World Bank added the Program for Results, a results-based approach, to its set of instruments.

24Against the background of ongoing aid debates, several general considerations can be discussed concerning results-based aid some of which might be relevant to one specific results-based aid type but not to another.

Potential advantages of results-based aid

25Action is directly aimed at providing incentives for results: the behaviour of all actors (development partners as well as partner governments) is significantly influenced by the results. There are direct links between the aid interventions and incentives (which might lead to results); the benefits might be more immediate and quantifiable.

26Incentives for performance: the input of aid creates incentives to perform for the partner country. This performance orientation can have a spillover effect into other sectors of the partner country.

  • 1 In general terms, accountability is the obligation of a person, group, or institution to justify de (...)

27Strengthened ownership on the part of the partners / partner government responsible for implementation: the task of achieving the goals lies with the partner government. The donors have no responsibility for implementation. This strengthens the partners’ political and administrative systems. At the same time, the approach might be supportive of more mutual accountability.1

28Better verification of the results of aid: closing the ‘attribution gap’ (proving a direct causality of aid measures and incentives that lead to results) might be more successful in specific cases. This can help donor countries to demonstrate the concrete benefits of aid (visibility of the development partner). However, there might be a risk of simplistic causal ‘attributions’, especially in those cases where results were achieved; results-based aid does not lead to an easy and automatic attribution of results to aid interventions.

Possible disadvantages and limits of results-based aid

29Responsiveness of the partner’s political system to incentives: the results-based aid concept assumes that the partners are open for incentives to perform better. This applies to those partner countries that show a strong performance orientation (‘good performers’), or at least where there are areas of access, such as in specific, viable institutions (‘pockets of effectiveness’, see Roll, 2011). However, the literature provides examples especially regarding low-income and highly aid-dependent countries where those favourable conditions are non-existent or only partly assured. Particularly, research on political systems in a number of sub-Saharan African countries shows evidence for systemic non-performance in core areas of service delivery (see, e.g. van de Walle, 2005; Chabal and Daloz, 1999).

  • 2 Whether incentives might do good or harm depends on the specific setting in a country. For example, (...)

30Adverse incentives, unintended consequences, and non-systemic strategies: generally speaking, there is a danger of setting adverse incentives; a strong focus on a specific outcome might tend to result in non-systemic analysis and strategy. The pressure to achieve certain goals can thus lead to the neglect of other priorities in the same sector. Indicators that might not be entirely suitable to this approach jeopardise the implementation of policies that are too heavily focussed on quantitative goals.2

31Capacity: the approach implies that the partner countries have the capacity to achieve results. If their capacities and their public financial management system are deficient, this does not seem realistic.

32Sectors and data: results-based aid cannot be implemented equally well in all sectors. Social sectors, such as education and health, as well as sectors dealing with infrastructure services that can more easily be measured (transport, public water supply, etc.) are well suited to results-based aid. In other sectors it may be harder to measure these results or to come to an agreement on them with the partner countries (such as complicated agreements on good governance), and the direct effects cannot always be clearly shown as wider outcomes. This applies, for example, to various areas of public financial management.

33Losing entry points for policy dialogue / de-linking (some) results-based aid approaches and the political context: where results-based aid involves an automatic mechanism for payment following the achievement of certain goals, difficulties might arise if a development partner were forced to pay out, even if faced with an unfavourable political environment, including massive governance problems (such as serious human rights abuses). Results-based aid is not an instrument for expanding opportunities for policy dialogue.

  • 3   In this case, it could be possible to develop schemes for start-up financing.

34Insufficient pre-financing capacities: in the context of this approach, pre-financing by the partner country is intended, or even necessary. Because of tight budgets in a number of low-income countries, this could be a major obstacle.3 There is the further risk that other aid resources in a country might be redirected to this end.

35Fiduciary risks: Results-based aid risks might be similar to programme-based approaches (such as budget support or pooled funding mechanisms) in terms of fiduciary challenges (Leiderer, 2012). Since aid disbursements are not tied to specific activities or procurement procedures, fiduciary risks might be a relevant challenge.

36Time horizon: results-based aid can create a short-term perspective, because it might cause a focus on results that can be achieved quickly. Results that can only be achieved in the medium or long term might clash with short-term political rationales (desire for re-election, etc.)

37Factors out of the government’s control: in those cases where results agreed upon are not reached because of factors not under the control of the government, the effect of incentives for performance vanishes because disbursement of results-based aid funds does not take place, regardless of the development-orientation of the government. For example, if a partner government faces the strong need to cut a budget because of the overall economic situation (due to an international crisis or unfavourable terms of trade), there might also be the need to reduce the budget line for achieving results.

38Danger of unambitious results: whether results are realistic or unrealistic is often vague. Since partner countries and donor agencies have an interest in disbursing the rewards, there might be an implicit tendency to identify less ambitious results.

39Some of the limits and technical challenges mentioned above might be tackled. For example, in those cases where insufficient pre-funding capacity would not allow the use of results-based aid, an adjustment of the approach is reasonable. In such a case several options might be considered. (i) The donor could set up a system where (partial) pre-funding is provided, for example for the first expected cycle of results. However, this course of action might contribute to a significant reduction in the intended incentive and pose a challenge if results are not achieved (How to organise reimbursement in case of non-performance? Is this procedure really enforceable? etc.) (ii) Another option might be the reimbursement of the interest paid in those cases in which the recipient borrows from the capital market in order to be able to make the necessary investments. However, even if a reimbursement was agreed upon in advance, there might be other effects on the recipient’s budget (e.g. due to borrowing limits set by the International Monetary Fund).

40Further provisions might be included in order to specifically support the capacity aspects and the reliability of data; regulations in this regard are included, for example, in the new World Bank approach. An upfront investment earmarked for building the capacities of the institutions concerned could be integrated into results-based aid. This amount could be provided in advance and spent in line with an agreed approach (such as tendering of capacity-building activities in the specific area). A similar approach might be used in order to improve the quality of data needed; for example requiring an agreed amount or share of aid to be used to ensure the regular provision of reliable data.

  • 4   However, one could argue that the predictability risks can mainly be managed by the recipient gov (...)

41Regarding results-based aid, there seems to be a rather high probability of a low disbursement level due to poor or non-performance (at least in those cases where ambitious targets are set). On the one hand, the possibility of partial or non-disbursement is an important feature of results-based aid. It is the intention of the approach to establish a strong link between performance and the provision of aid. If the aid amount is finally made available to the partner country in some way, even in a case of non-performance, there might be an adverse impact on the incentive structure from the beginning. On the other hand, partial or non-disbursement is a crucial challenge for the partner country in terms of the predictability of aid.4 In addition, such a situation might also lead to challenges for the donor (aid management) and to questions about the impact of the country strategy. For example, if the government of the partner country is not able to perform (i.e. to provide results in accordance with the results-based aid agreement) solely due to exceptionally unfavourable international conditions (reduced budget because of adverse international market prices for relevant commodities), the donor might be under pressure to disburse despite non-achievement of results.

42Typically, results-based aid is associated with interventions in social sectors (often education and health) and sometimes with interventions concerning other direct service delivery activities (such as access to sanitation). These intervention areas have some specific features:

  • It is easy to identify development results. In many areas, reference can be made to international and national objectives (e.g. Millennium Development Goals or national poverty reduction strategies) and / or international standards.

  • These results are measurable.

  • Data and baseline information is often available, or easy to collect. High additional transaction costs can normally be avoided.

  • Intense disputes between the different parties around the definition of results, the indicators, the applied methods, and data are not expected.

43The application of results-based aid in other areas seems to be more challenging.  For instance, applying results-based aid to governance presents the challenge of finding indicators that capture quantifiable results on the outcome or impact level. In addition, there are political sensitivities when donors and partner countries have to find mutually agreed results related to politicised issues such as human rights or democracy promotion (Klingebiel, 2012). Thus, within the governance sector, an area such as fiscal decentralisation where measurements are technically straight-forward is more suitable for results-based aid. The sector of fiscal decentralisation is also characterised by good availability of baseline data for measuring progress in improving financial management performance. In fact, donors already have some experience with implementing performance-based grant systems in the area of fiscal decentralisation (Steffensen, 2010) and the World Bank has initiated several performance-based grant systems under the Program for Results lending instrument. Therefore, experience with performance-based grant systems can offer valuable insights into current practices of donors in implementing results-based aid.

4. Initial Experiences with Results-Based Aid in Ghana and Tanzania

44There is only limited evidence of initial implementation experiences as most results-based aid pilots are still at the conceptual stage or in their early years of implementation. Also, there are only a small number of pilot programmes that conform with the ‘pure’ definition of results-based aid provided above. Despite these constraints on case selection, the two specific country cases, Ghana and Tanzania, illustrate some of the potentials and limits of results-based aid in practice (Janus, 2014). Both countries are so called ‘donor darling’ countries that receive significant amounts aid and in which donors are keen to invest in innovative aid modalities. The findings are tentative, and only indicate specific initial experiences in implementing results-based aid in the sector of fiscal decentralisation. While the results cannot be generalised beyond the context studied they do offer important insights into potential challenges in implementing results-based aid.

45The District Development Facility (DDF) in Ghana supported by Canada, Denmark, France, and Germany (MLGRD, 2012) and the Urban Local Government Strengthening Program (ULGSP) in Tanzania (World Bank, 2012) one of the World Bank’s first Programme for Results projects are results-based aid pilots in the area of fiscal decentralization. The DDF in Ghana has been running since 2008 and the ULGSP just started in 2013. Both programmes aim to improve the performance of local governments, provide decentralised infrastructure, and enhance public service delivery at the local level. They rely on annual independent assessments of local administrations against key public financial management indicators that form the basis for disbursements. Results of the annual assessments and the underlying contracts defining indicators and disbursements are public, and local administrations can use the funds in a discretionary manner.

46Both programmes therefore largely comply with the criteria for defining results-based aid, with one exception. Results-based aid indicators in decentralisation mostly target the input or output level, and not the level of outcomes or impact as envisioned by the results-based aid model. Decentralisation programmes focus on resource management and adopt indicators that measure processes (reflecting planning, budgeting, financial management, etc.) These indicators are viewed as intermediate-results or proxy-results indicators, which are important prerequisites for achieving service delivery outcomes. There is a tendency to predominantly rely on these process indicators, as local administrations are mostly responsible for intermediate activities that can then lead to beneficial development outcomes.

47A number of challenges related to the results-based aid modality can be observed. First, in both countries the adoption of the results-based aid modality was mostly driven by donors and the design programme was largely developed by donors. This is understandable, given that both programmes are pilots, but partner countries being in charge of choosing and designing results-based aid programmes is clearly important. Related to this question of ownership both programmes also show mixed records regarding adherence to other principles of aid effectiveness.  For instance, there are signs of greater harmonisation with other donors and using the country’s national systems, but both programmes also require additional administrative structures that need to be created, thereby duplicating structures of other donors or the partner country. This undermines efforts towards alignment of aid interventions with a partner country’s own systems.

48A second challenge relates to the role of conditionality and setting coherent incentives in results-based aid programmes. While disbursements are based on technical triggers in both countries, there is also a political debate on how to adjust conditions for disbursement. In Ghana, the public financial management of districts improved significantly in the first years of the programme but then a performance plateau was reached. In 2011, most districts qualified for grants and scored highly in the assessments (MLGRD, 2012).  While this is largely a positive development it also means that programme incentives and indicators lose relevance once a critical number of local governments meet all targets. Here, indicators need to be adjusted in order to make incentives more relevant again.

49Yet, these adjustments can be influenced by converging donor and partner country interests. The donor often has an interest in disbursing because aid agencies typically reward the actual disbursement of available aid commitments. The partner country, too, wants to mobilise the maximum of available resources. Local politicians might further exert pressure on the ministry in charge to disburse funds, even if results have not been met by local governments. In order to avoid settling for unambitious results, partners and donors have both to ensure that indicators remain relevant and resist disbursement pressure. They need to work jointly to develop a coherent incentive mechanism that is able to enforce its own conditions. This is a technical as well as a political challenge. Therefore, partners and donors should work together towards building a results culture in which results-based aid can be an element of a broader toolbox of aid modalities. At the same time this means that the aid allocation within one country should only be partly and not entirely results-based.

50Finally, in both countries the results-based aid programmes only represent a small part of a larger set of decentralisation activities. Thus, both cases clearly underscore the importance of understanding the local political context. There are no signs of adverse incentives set by the results-based aid programme but both cases demonstrate the need to carefully adapt programme incentives to the local political economy of decentralization. For example, in Ghana the continuous creation of new districts at the local level, partly financed through donor activities, presents the challenge of “recentralisation” (Grossman and Lewis, 2013). In Tanzania, donors in the decentralisation sector also implicitly supported a system that seeks executive expansion and centralized government control without the decentralization of resources that would give localities real power (Tripp, 2012). These broader structural features of the decentralisation process need to be understood thoroughly in order to design meaningful incentive systems for improved performance of local governments.

51In sum, the experiences from Ghana and Tanzania underscore the need for designing results-based aid modalities in adherence to aid effectiveness principles (such as ownership, harmonisation, and alignment), the difficulty of creating coherent incentive mechanisms, and the challenge of adapting results-based aid programmes to the local political context. Despite these challenges both cases also clearly demonstrate the potential of results-based aid. Both programmes are thoroughly designed towards encouraging increased levels of performance, and both programmes set effective incentives for achieving results. In Ghana, a number of potential benefits such as improved financial management and greater interaction between local governments and local citizens are already observable. District officials there report that competition between districts has put local government ‘on their toes’ to perform, and interaction between district administrations and local citizens has increased, thereby also strengthening local accountability structures. In terms of conditionality, experiences in both countries show that the dialogue between donors and recipients is mainly focused on a more technical discussion of implementation challenges related to the results-based aid programme.

5. Conclusions

52The potential benefits and limits of results-based aid strongly depend on the specific approach and design. The international debate is dealing with a variety of different approaches in the field of development aid. The spectrum ranges from a more results-oriented approach in the area of budget support (macro and sectoral levels) to new types of projects that are intended to push one specific result (sub-sectoral level or ‘one result’-specific results-based aid). All approaches have one common feature: they create incentives for achieving results. The analysis shows that if results-based aid programmes are properly designed, they can become an important aid modality, especially in the context of a post-2015 development agenda. While results-based aid can become an important tool for developing countries overall, there might be limited scope for applying demanding results-based aid programmes in fragile situations or countries with weak institutions.

53The discussion on the potential advantages and disadvantages of results-based aid reveals a number of key considerations that should be included in the design and implementation of results-based aid. For instance, there can be a significant risk of adverse incentives and non-systematic strategies if aid focuses too heavily on one measurable result and results-based aid is only donor-driven. In addition, there is a bias of results-based aid in favour of performing countries. The likelihood of ‘good performance’ (achieving results) is more pronounced in cases in which countries have good leadership structures, planning and implementation capacity, and a functioning public financial management system.

54Opportunities for identifying measurable and quantifiable results are straight-forward in social sectors and several (basic) infrastructure-related areas. The governance sector is, in general terms, less favourable in this regard. ‘Political governance’ issues such as political freedom or human rights do not seem to be suitable for results-based aid approaches. However, some other governance areas have the potential to be included; this applies especially to public financial management and fiscal decentralisation. The cases of Ghana and Tanzania demonstrate how results-based aid can work in the area of fiscal decentralisation.

55Further, results-based aid might be regarded as a potential alternative aid modality to budget support, especially in times of waning political support for the latter approach. Overall, donors face the challenge of designing results-based aid programmes that introduce strong incentives for achieving results while retaining flexibility to operate in different policy environments. Currently, there are still knowledge gaps regarding the conceptual basis and implementation experience of results-based aid and more research is needed. But against the background of the analysis provided above, the following conclusions can be drawn:

56First, results-based aid standalone approaches from one donor might be inappropriate, especially in challenging country settings. Frequent and close collaboration with other development partners might be essential. There is a strong need to work closely with other donors in order to have an influential and constructive dialogue with partner governments. Any attempt to focus on just a few specific issues might prove difficult in a complex situation; harmonised and sector-oriented strategies are also important in the context of results-based aid.

57Second, some countries might display a low level of responsiveness to results-based aid and limited opportunities for ‘pockets of effectiveness’. Results-based aid strongly relies on ‘driving forces’ for reforms in the government structure, including at the highest level. If government structures are not receptive to performance orientation, the likelihood of failure (non-performance) is high. In principle, one could think about some niches of government where this approach might work and could have spillover effects, at best. However, since the governance situation dominates all public structures, scope might be limited.

58Third, the process of selecting and designing results-based aid programmes that are adapted to local contexts, critically determines the future success of results-based aid. Therefore, open engagement between donors and partner countries is needed and partner countries have to be in charge of the overall process. Also, results-based aid programmes are technically demanding and sufficient investment in the design of such programmes is likely to outweigh adjustment costs that otherwise might occur as results-based aid programmes are being implemented.

59In conceptual terms, results-based aid links to several academic debates on aid and can clearly contribute to discussions on  aid impact, allocation and conditionality, and aid modalities. In particular, results-based aid has a strong potential to contribute to the debate on aid conditionality. Based on the conceptual design of results-based aid interventions and on initial country experiences it has become clear that results-based aid is neither a pure case for ‘political conditionality’ nor for ‘technical triggers’ in the context of aid. Depending on the concrete arrangement, results-based aid can be used as an instrument to push a donor agenda linked to political conditionality. This is true, for example, with regard to the MCC. However, other results-based aid pilots that are focused on a specific indicator in a given sector concentrate on technical triggers rewarding independently verified results. Thus, results-based aid does not provide a clear-cut example for a second generation of political conditionality. Results-based aid rather represents a case of conditionality that relies more strongly on technical triggers than traditional aid modalities. This specific form of a more technical type of conditionality is the distinctive feature of results-based aid as an innovative aid modality. The examples of Ghana and Tanzania show how results-based aid there is focused on technical triggers. Still, experience in both countries also indicates that even technical triggers might easily become politicised in practice. Thus, a purely technical focus of results-based aid might be difficult to uphold once political interests come into play.

Top of page


Addison, T., S. Singhal and F. Tarp (2013) Aid to Africa: The changing context, WIDER Working Paper No. 2013/144 (Helsinki: UNU-WIDER).

Azam J.-P. and J. Laffont (2003) 'Contracting for aid', Journal of Development Economics, 70(1),  pp. 25-58 (34).

Birdsall, N. and W. D. Savedoff (2011) Cash on Delivery: a new approach to foreign aid, 2nd edition (Washington, DC: Center for Global Development).

Birdsall, N., W. D. Savedoff and A. Mahgoub (2009) Cash on Delivery: exploration of feasibility in Malawi (Washington, DC: Center for Global Development).

Burnside, C. and D. Dollar (2000) 'Aid, Policies, and Growth', The American Economic Review, 90: pp. 847-868.

Booth, D. (2011) Aid effectiveness: bringing country ownership (and politics) back in, ODI Working Paper 336 (London: Overseas Development Institute).

CGD (Center for Global Development) (2006) When will we ever learn? Improving lives through impact evaluation, Report of the Evaluation Gap Working Group (Washington, DC: CGD).

Chabal, P. andJ.-P. Daloz (1999) Africa works: disorder as political instrument (Oxford: Oxford University Press).

Clemens, M., S. Radelet and R. Bhavnani (2004) Aid and Growth: The Current Debate and Some New Evidence (Washington, DC: CGD).

Collier, P., and D. Dollar (2002) 'Aid allocation and poverty reduction', European Economic Review, 46, pp. 1475-1500.

CoP-MfDR (Community of Practice on Managing for Development Results) (2011) Framework for results-based public sector management and country cases (Manila: Asian Development Bank).

De Hennin, C. andH. Rozema (2011) Study on results-based programming and financing, in support of sharing the multi-annual financial framework after 2013 (No place).

Duflo, E. and A. Banerjee (2011) Poor economics: a radical rethinking of the way to fight global poverty (New York: Perseus Books).

De Renzio, P. andN. Woods (2008) The trouble with Cash on Delivery aid: a note on its potential effects on recipient country institutions (GEG. Oxford: University of Oxford), Mimeo.

Easterly, W., R. Levine and D. Roodman (2003) New data, new doubts: A comment on Burnside and Dollar's "aid, policies, and growth" (2000) (Cambridge: National Bureau of Economic Research).

Fielding, D., M. McGillivray and S. Torres (2006) A wider approach to aid effectiveness: Correlated impacts on health, wealth, fertility and education, UNO-WIDER Research Paper No. 2006/23 (Helsinki: UNU-WIDER).

Grossman, G. and J. I. Lewis (2013) 'Administrative Unit Proliferation', American Political Science Review.

Hansen, H. and F. Tarp (2000) Policy Arena Aid Effectiveness Disputed, Journal of International development, 12, pp. 375-398.

Janus, H. (2014) Real innovation or second-best solution? First experiences from results-based aid for fiscal decentralisation in Ghana and Tanzania, Discussion Paper (Bonn: Deutsches Institut für Entwicklungspolitik/ German Development Institute).

Karlan, D. and J. Appel (2011) More than good intentions: How a new economics is helping to solve global poverty (Boston: Dutton Juvenile).

Kenny, C. (2011) Getting better: why global development is succeeding and how we can improve the world even more, (New Xork: Basic Books (AZ)).

Killick, T. (2004) 'Politics, Evidence and the New Aid Agenda', Development Policy Review, 22: pp. 5-29.

King, K. (ed.) (2012) Value for money in international education. A new world of results, impacts and outcomes? NORRAG News 47 (Pencaitland: Network for Policy Research, Review and Advice on Education and Training).

Klingebiel, S. (2011a) Results-based financing – the potential for a greater results focus, Discussion Paper (Bonn: Deutsches Institut für Entwicklungspolitik/ German Development Institute).

Klingebiel, S. (2011b) Results-based aid: limitations of new approaches, Briefing Paper 17/2011 (Bonn: Deutsches Institut für Entwicklungspolitik/ German Development Institute).

Klingebiel, S. (2012) Results-based aid (RBA): new aid approaches, limitations and the application to promote good governance, Discussion Paper 14/2012 (Bonn:  Deutsches Institut für Entwicklungspolitik/ German Development Institute).

Koeberle, S. H. Bedoya, P. Silarszky and G. Verheyen (eds.) (2005) Conditionality revisited: concepts, experiences, and lessons (Washington, DC: World Bank).

Leiderer, S. (2012) Fungibility and the choice of aid modalities: the red herring revisited
UNU-WIDER Working Paper 2012/68 (Helsinki: UNU-WIDER).

Lopez, R. G. and M. G. Moreno (2011) Managing for development results, progress and challenges in Latin America and the Caribbean (Washington, DC: Inter-American Development Bank).

Minoiu, C. and S. G. Reddy (2010) 'Development aid and economic growth: A positive long-run relation', The Quarterly Review of Economics and Finance, 50, pp. 27-39.

MLGRD (Ministry of Local Government and Rural Development of Ghana) (2012) Operational manual for the implementation of the district development facility (Accra) – (2013) Publication of FOAT I-V (2006–2011) results and allocation (DDF & UDG).

Morrissey, O. (2010) 'Aid and the financing of public social sector spending' in Hujo, K. and S. Mcclanahan (eds.) Financing Social Policy: Mobilizing Resources for Social Development (Basingstoke: Palgrave/UNRISD).OECD (Organisation for Economic Co-operation and Development) (2008) Managing for Development Results: Information Sheet, (accessed 14 March 2014).

OECD-DAC and World Bank (2008) Sourcebook: Emerging Good Practices in Managing for Development Results, 3rd Edition, (accessed 15 April 2014).

Öhler, H., P. Nunnenkamp And A. Dreher (2010) Does Conditionality Work? A Test for an Innovative US Aid Scheme, Kiel Working Paper No. 1630 (Kiel: Kiel Institute for the World Economy).

Parks, B. C. and Z. J. Rice (2013) Measuring the Policy Influence of the Millennium Challenge Corporation: A Survey-Based Approach (Williamsburg: The Colleage of William and Mary).

Pereira, J. and C. C. Villota (2012) Hitting the target? Evaluating the effectiveness of results-based approaches to aid (Brussels: Eurodad).

Perrin, B. (2013) Evaluation of Payment by Results (PBR): Current Approaches, Future Needs, Working Paper 39 (London: Department for International Development (DfID)).

Pearson, M. (2011) Results-based aid and results based financing: What are they? Have they delivered results? (London: HLSP Institute).

Rajan, R. G. and A. Subramanian (2008) 'Aid and growth: What does the cross-country evidence really show?', The Review of economics and Statistics, 90, pp. 643-665.

Rodrik, D. and M. R. Rosenzweig (eds.) (2010) Handbook of development economics, vol. 5. (Oxford / Amsterdam: Elsevier BV).

Rogerson, A. (2011) What if development aid really rewarded results? Revisiting the Cash on Delivery (CoD) aid model, Development Brief 1-2011 (Paris: OECD).

Roll, M. (2011) The state that works: pockets of effectiveness as a perspective on stateness in developing countries, Working Paper No 128 (Mainz: University of Mainz).

Savedoff, W. (2011a) Incentive proliferation? Making sense of a new wave of development programs (Washington, DC: Center for Global Development).

Savedoff, W. (2011b) Governance in the health sector: a strategy for measuring determinants and performance (Washington, DC: Center for Global Development).

Steffensen, J.(2010) Performance-based grant systems: concept and international experience (New York: United Nations Capital Development Fund).

Stokke, O. (ed.) (2004) Aid and political conditionality, EADI Series No. 16 (London: Routledge).

Svensson, J. (2003) 'Why conditional aid does not work and what can be done about it? ', Journal of development economics, 70(2), pp. 381-402.

Temple, J. R. W. (2010) 'Aid and conditionality' in D. Rodrik and M. R. Rosenzweig (eds.) Handbook of development economics, vol. 5 (Oxford / Amsterdam: Elsevier BV), pp. 4415–4523.

Tilley, H. and H. Tavakoli (2012) Better aid modalities: are we risking real results? (London: ODI).

Tripp, A. M. (2012) Donor assistance and political reform in Tanzania (Helsinki: United Nations University World Institute for Development Economics Research).

Van de Walle, N. (2005) Overcoming stagnation in aid-dependent countries (Washington, DC: Center for Global Development).

Wood, B., J. Betts, F. Etta, J. Gayfer, D. Kabell, N. Ngwira, F. Sagasti and M. Samaranayake (2011) The evaluation of the Paris Declaration, Phase 2, Final Report (Copenhagen: Danish Institute for International Studies).

World Bank (2011a) A new instrument to advance development effectiveness: Program-for-Results financing (Washington, DC: World Bank).

World Bank (2011b) World Bank for results 2011 (Washington, DC: World Bank).

World Bank (2012) Program appraisal document for urban local government strengthening program (program-for-results operation) (Washington, DC: World Bank).

Top of page


1 In general terms, accountability is the obligation of a person, group, or institution to justify decisions or actions taken. It is associated with sanctions in the event of compliance / non-compliance and is therefore based on incentives. Accountability is relevant in three respects:

- accountability on the part of the donor;
- accountability in partner countries (domestic accountability);
- mutual accountability between partner and donor.

The different levels of accountability do not automatically complement each other. In some cases, they may even compete. Mutual accountability might be costly, entail numerous compromises, and suffer from shortcomings. This is true, for example, for coordinated national development strategies and joint monitoring approaches and policy analyses. Results-based aid can, in principle, strengthen mutual and national (in the partner country) accountability, since both forms are based on the partners’ implementation of policies and their activities.

2 Whether incentives might do good or harm depends on the specific setting in a country. For example, normally we assume that an increase in resources available at the local level can contribute significantly to development. Money can be spent in accordance with local needs and priorities. Decision-making processes developed in close cooperation with the people might improve participation. However, if an increase in resource transfers to the local level is considered by local or national elites as available ‘rents’, it might just create conflicts over ‘access to rents’. If local structures are not prepared for and experienced in managing an increase in resource transfers, this approach might also fail because of limited capacities on the ground. If aid provides strong incentives for an increase in own resources (local taxes, etc.) this might also contribute to problems (for example, for small businesses) because inexperienced staff on the local level push for local revenues that, for example, affect farmers who sell their products.

3   In this case, it could be possible to develop schemes for start-up financing.

4   However, one could argue that the predictability risks can mainly be managed by the recipient government, since it is in charge of implementing those policies necessary for achieving the expected results.

Top of page

List of illustrations

Title Figure 1 – Results-Based Aid — Impact Chain
Credits Source: Author’s own representation
File image/png, 42k
Top of page


Electronic reference

Stephan Klingebiel and Heiner Janus, « Results-Based Aid: Potential and Limits of an Innovative Modality in Development Cooperation », International Development Policy | Revue internationale de politique de développement [Online], 5.2 | 2014, Online since 20 May 2014, connection on 20 September 2017. URL : ; DOI : 10.4000/poldev.1746

Top of page

About the authors

Stephan Klingebiel

Dr Stephan Klingebiel is Department Head (Bilateral and Multilateral Development Cooperation) of the German Development Institute (Deutsches Institut für Entwicklungspolitik (DIE)). Before joining DIE in 1993 he was a researcher at the University of Duisburg (Institute for Development and Peace). From 2007 to 2011 he was Director of KfW Development Bank’s office in Kigali, Rwanda, dealing with development cooperation issues. His research and university teaching focuses on the political economy of aid, aid and development effectiveness, political economy and governance issues in sub-Saharan Africa, and crisis prevention and conflict management. He is co-lead author of the European Report on Development 2013 on the post-2015 global development agenda. Stephan Klingebiel is a regular Visiting Professor at Stanford University.

Heiner Janus

Heiner Janus is a researcher at DIE’s Bi- and Multilateral Development Cooperation department. His research focuses on aid and development effectiveness, results-based approaches, and the role of emerging economies in development cooperation.

Top of page


Creative Commons License
International Development Policy is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Top of page
  • Logo The Graduate Institute of International and Development Studies
  • Logo DOAJ – Directory of Open Access Journals
  • Logo ERIH PLUS | NSD
  • Les cahiers de