10. In focus: Aid effectiveness in Afghanistan, Mali and South Sudan

Christoph Zürcher
University of Ottawa
  • Evaluations over a decade of development interventions in highly fragile situations find that interventions predominantly fail to deliver on their objectives.

  • Development co-operation providers may need to rethink the principle of allocating according to the greatest need in these contexts and instead focus on interventions with the highest probability of success in the contexts in which aid is deployed.

To what extent can aid be effective in highly fragile situations? A recent systematic review of 315 individual evaluations of aid to Afghanistan, Mali and South Sudan provides robust, clear and sobering conclusions: Development aid to the most fragile contexts has not sufficiently met its objectives, whether these are stabilisation (for example, through basic service provision, mediation and conflict resolution capacity building, or the provision of economic opportunities); improvement of state capacities and good governance; or increasing women’s empowerment (2020[1]; Zürcher, 2020[2]). The most important – and most disappointing – finding is that development aid is not a suitable tool for addressing the core problems of such contexts. Aid does not improve governmental capacity, does not lead to better governance and does not provide more stability. Even where assistance was effective to some degree – for instance, in education, health and rural development interventions – results may not be sustainable. These findings should compel donors to fundamentally reconsider basic assumptions regarding their engagement in fragile contexts.

Systematic reviews are exercises in learning. They aim to identify and summarise all existing evidence on a given topic. This particular systematic review covered evaluations published in English or French between 2008 and 2021. In total, 315 evaluations that met a predefined methodological quality threshold were included: 142 on interventions on Afghanistan, 104 on Mali and 69 on South Sudan. All included evaluations were then distributed along ten predefined aid sectors: women’s rights, health, rural development and climate change, the rule of law, stabilisation, education, sustainable economic development, nutrition, humanitarian assistance, and good governance. In the final step, the included evaluation reports were mined for evidence on effectiveness.

The review found that interventions in education, health and rural development were reasonably effective, although results were most likely not sustainable without external support. Interventions contributed to improved livelihoods and strengthened coping mechanisms but did not lead to sustainable economic growth, jobs or income opportunities. Programmes supporting macroeconomic development, macroeconomic policies, financial management and support for the private sector were mostly ineffective. Likewise, interventions for promoting gender equality had little impact. Finally, and most importantly, interventions in stabilisation, good governance and capacity building for the central government were clearly ineffective in all three countries.

Interventions in education, health and rural development were reasonably effective, although results were most likely not sustainable without external support.  

Stabilisation interventions came in different shapes and forms. Stabilisation projects aimed to quickly restore basic services to the population in the hope that this would encourage the population to work with the government or aimed to provide economic opportunities that work as a peace dividend. Other projects aimed to directly strengthen the mediation and conflict resolution capacity of communities and political actors; another type supported political institutions and processes; and some regarded support for civil society actors as a possible contribution to stabilisation. The sample analysed in the review contained 28 evaluations of stabilisation interventions. None of these interventions proved to be effective. This is in line with the wider academic literature on stabilisation, which suggests that aid is seldom a viable tool for reducing violence (Sexton, 2016[3]).1

Interventions to promote “good governance” were also found to be largely ineffective. Good governance is a very broad sector and includes public sector and regulatory policy reform, promoting democracy, election support, anti-corruption programmes, and the rule of law, among other things. Together with stabilisation and capacity building for the government, it is at the core of what donors are trying to achieve in fragile contexts. Again, the evidence is clear. Thirty of the evaluations that covered good governance initiatives suggest that effectiveness was very low in these extremely fragile contexts. Factors that explain the lack of impacts included entrenched, patronage-based practices within the government; a lack of buy-in from the government; donor-driven, top-down project design with little regard for the core institutional requirements and demands of the partner institutions; and lack of political will from the government.

Likewise, capacity-building measures for governments in extremely fragile contexts largely missed their mark. The review found that donors overestimated both existing state capacity as well as the recipient government’s political will for reforms. As a result, programming for capacity building was overambitious, unsustainable and ineffective. The hope that the government would become a partner in service delivery never materialised. The least effective were capacity-building measures in so-called politicised fields, which offer opportunities for lucrative corruption and are therefore vital for maintaining the rentier and patronage state.

The meta-evaluation found that three main factors explain why development co-operation often fails to achieve its aims in extremely fragile contexts. First, in highly fragile settings, political power does not reside with formal political institutions but rather in patronage networks. Patronage depends on rent-seeking, predation and widespread corruption. Elites in such a system are thus not interested in political reforms that would endanger this particular way of exercising power. Second, donors consistently overestimate existing state capacity and design programmes based on largely imagined capacity. Third, the lack of basic security, typical for fragile situations, renders many aid programmes ineffective.

In sum, the systematic review demonstrates that interventions in stabilisation, good governance and capacity building for the government were not effective. This is a very disappointing finding and one that urges donors to come to terms with the fact that while development aid can help improve basic livelihoods and service provision – albeit to a limited extent and not sustainably – it has little political transformative capacity in these very fragile contexts. Aid will not make a fragile state more stable, more capable or better governed.

Every aid dollar that is spent on an ineffective intervention is one that cannot be spent on an intervention that might be effective. Continuing to allocate resources to sectors where the probability of success is low is not only ineffective, it is also ethically wrong because it binds resources that otherwise could be used to improve people’s lives; for example, to provide shelter, increase food security, improve access to health, or teach children how to read and write.

Perhaps the most important implication of this systematic review is that aid managers should always consider opportunity costs when taking allocation decisions. It may be very tempting to allocate aid resources to where the greatest needs seem to be – a lack of security, a weak government, appalling gender inequalities. But once the evidence shows that the probability of success in these sectors is extremely low, then aid resources must be allocated to sectors with a reasonable probability of success, even if the needs are perceived as smaller in these sectors or when these sectors are not aligned with the ideological priorities of Western donors.

Of course, an aid strategy that would prioritise aid sectors with a higher probability of success over aid sectors with a low probability of success has its own challenges. For example, if aid is predominantly provided to sectors such as health, education or livelihoods, there is the risk that aid dependencies will be created, without an easy exit option. Furthermore, focusing aid on such sectors will do little to transform the predatory state structures that are at the core of fragility.

This systematic review highlighted such conundrums. It provided no silver bullets, but faced with the evidence from the review, the aid community can no longer ignore the problem. What is needed now is an honest discussion about a new aid strategy in fragile contexts, one that starts with the acknowledgement of the limits of what aid can accomplish in these settings. This finding is not entirely new. There is a solid academic literature that argues that external actors rarely succeed in strengthening institutions in fragile states and territories (Bliesemann de Guevara, 2010[4]; Ottaway, 2002[5]; Chowdhury, 2017[6]; de Waal, 2015[7]); that fragile states are often trapped in fragility (Pritchett, Woolcock and Andrews, 2013[8]; Collier, 2007[9]; Carment and Samy, 2019[10]); and that aid faces many challenges in fragile states (Gisselquist, 2014[11]). Yet, there is still the prevailing sense among many scholars and practitioners that aid effectiveness might somehow be increased by fine-tuning aid modalities, by better adapting aid to local contexts, by increasing aid flows or by staying engaged for a longer time. This systematic review strongly suggests that this is not the case.

The evidence also shows that aid only has a fair chance of being effective in highly fragile contexts when programmes are modest in their ambitions and relatively small, do not assume unrealistic partner capacities, are aware of the context, and do not spend aid money too fast or in highly insecure regions controlled by insurgents. The most important lesson is that aid should avoid sectors where it is most probably not effective and prioritise sectors where aid has a fair chance of being effective.

What is needed now is an honest discussion about a new aid strategy in fragile contexts, one that starts with the acknowledgement of the limits of what aid can accomplish in these settings.   

Such principles amount to a new paradigm – out with grand visions and in with local, tangible and small gains. Abandoning grand visions in favour of tangible local gains may be a bitter pill to swallow for many aid agencies and aid practitioners, who are often driven by a genuine desire to use aid as a lever to transform societies so that structural impediments to development are removed. Unfortunately, the evidence shows that aid has little transformative power in the most fragile contexts. This is a very uncomfortable truth, but one that can no longer be ignored.

Defining a new approach for supporting the most fragile places will be a long process. One thing that donors can do now is further engage in collecting evidence. More systematic reviews of aid to highly fragile contexts are needed. Systematic reviews of aid to less fragile states are also needed to determine whether aid effectiveness increases once fragility becomes less severe. Systematically varying the fragility score can indicate whether and how aid effectiveness across different sectors is correlated with initial fragility.


[4] Bliesemann de Guevara, B. (2010), “Introduction: The limits of statebuilding and the analysis of state-formation”, Journal of Intervention and Statebuilding, Vol. 4/2, pp. 111-128, https://doi.org/10.1080/17502970903533652.

[10] Carment, D. and Y. Samy (2019), Exiting the Fragility Trap: Rethinking Our Approach to the World’s Most Fragile States, Ohio University Press, Athens, OH.

[6] Chowdhury, A. (2017), The Myth of International Order: Why Weak States Persist and Alternatives to the State Fade Away, Oxford University Press, Oxford.

[9] Collier, P. (2007), The Bottom Billion: Why the Poorest Countries Are Failing and What Can Be Done About It, Oxford University Press, Oxford.

[7] de Waal, A. (2015), The Real Politics of the Horn of Africa: Money, War and the Business of Power, Polity Press, Cambridge.

[11] Gisselquist, R. (2014), “Aid and institution-building in fragile states”, The Annals of the American Academy of Political and Social Science, Vol. 656/1, pp. 6-21, https://www.jstor.org/stable/24541760.

[5] Ottaway, M. (2002), “Rebuilding state institutions in collapsed states”, Development and Change, Vol. 33/5, pp. 1001-1023, https://doi.org/10.1111/1467-7660.t01-1-00258.

[8] Pritchett, L., M. Woolcock and M. Andrews (2013), “Looking like a state: Techniques of persistent failure in state capability for implementation”, Journal of Development Studies, Vol. 49/1, pp. 1-18, https://doi.org/10.1080/00220388.2012.709614.

[3] Sexton, R. (2016), “Aid as a tool against insurgency: Evidence from contested and controlled territory in Afghanistan”, American Political Science Review, Vol. 110/4, pp. 731-749, https://doi.org/10.1017/S0003055416000356.

[2] Zürcher, C. (2020), Meta-review of Evaluations of Development Assistance to Afghanistan, 2008-2018, Chapeau Paper, Federal Ministry for Economic Cooperation and Development, Berlin, https://www.sicherheitneudenken.de/media/download/variant/198198.

[1] Zürcher, C. (2020), “The impact of development aid on organised violence: A systematic assessment”, 3ie Working Paper, No. 37, International Initiative for Impact Evaluation, New Delhi, https://www.3ieimpact.org/sites/default/files/2020-08/WP37-Systematic-Review-Aid-Violence.pdf.


← 1. For a systematic review of the effects of aid on violence, see Zürcher (2020[1]).

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2023

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.