Good Aid in Hard Places: Learning from ‘Successful’ Interventions in Fragile Situations

Abstract

Development assistance to fragile states and conflict-affected areas can be a core component of peacebuilding, providing support for the restoration of government functions, delivery of basic services, the rule of law and economic revitalization. Despite a wealth of research, however, significant gaps remain in our knowledge about what has worked, why and the transferability and scalability of findings. The project presented in this special issue offers new research on ‘successful' projects and programmes in diverse domains and contexts in an effort to address these gaps. This article introduces the special issue and eight case studies included in this volume. Three broad factors are highlighted as significant in understanding why some interventions work better than others: the area of intervention and the related degree of engagement with state institutions; local contextual factors such as capacity and the existence of local supporters; and programme design and management. The article concludes with discussion of how these case studies and similar research can speak to broader debates in the literature on how peace is built, and in particular on the interaction between external actors and endogenous processes.

Development assistance to fragile states and situations is a core component of efforts to assist conflict-ridden countries in creating the conditions for sustainable peace.¹ Indeed, aid is intended to support activities at the heart of peacebuilding, including the restoration of core government functions, delivery of basic services, the rule of law and access to justice, and economic revitalization. Recent statistics suggest that Official Development Assistance (ODA) represents the largest financial inflow to fragile states – about $53 billion in 2011 or 38 per cent of ODA.² Nevertheless, fragile states have clearly lagged behind other countries in achieving the Millennium Development Goals. Fragile states and conflict-affected areas thus will remain a focus for donors in the post-2015 development agenda.³

Among the core and continuing challenges for both scholars and practitioners in this area is the question of how change occurs, peace is built and fragile states become more ‘resilient’.⁴ In particular, how do external interventions interact with local institutions and endogenous processes to influence outcomes?⁵ In the field of international development, this is manifested clearly in the need for better answers to questions about what precisely has worked to support peacebuilding in fragile situations, why and whether experience in one context can be expected to transfer elsewhere or to be scaled up successfully. Despite large and rich related literatures, there are major gaps in existing work in relation to these key questions.

This special issue emerges from research conducted under the Governance and Fragility theme of UNU-WIDER's Research and Communication on Foreign Aid (ReCom) programme (2011–13), and specifically the project on ‘Good Aid in Hard Places: Learning from What Works in Fragile Contexts'.

The ‘Good Aid' project centred around in-depth case studies of ‘successful' aid-supported projects and programmes, eight of which are included in this special issue. This article introduces the project and situates it within the literature on state fragility, impact evaluation and aid successes, to show how these analyses can contribute to both better policy making and more rigorous theory-building. Drawing on the literature and the case studies presented in this collection, this article identifies and discusses three sets of factors as important in understanding why some interventions work better than others: (1) the area of the intervention and the related degree of engagement with domestic state institutions; (2) local contextual factors, in particular windows of opportunity, capacity and the existence of local supporters; and (3) programme design and management. While the third set of factors is largely transferrable and scalable, the first two are less so and should be considered carefully in assessment of the feasibility of extending project and programme models to new contexts and in considering adjustments to improve their viability. It concludes with discussion of broader implications and avenues for future research. This article reflects the views of the author and is not intended as a summary of collective conclusions reached by contributors.

What Can New Analysis of ‘Successful’ Aid Interventions Add?

Development assistance to fragile states and conflict-affected areas is a core component of peacebuilding, with the potential to support each of the five Peacebuilding and Statebuilding Goals (PSGs) advanced by the International Dialogue on Peacebuilding and Statebuilding and set out in the New Deal for Engagement in Fragile States announced at Busan in 2011: legitimate politics, security, justice, economic foundations, and revenues and services. In terms of ‘how to' best address these goals, policy discussions highlight the ten ‘Fragile States Principles' first adopted by OECD ministers in 2007:

1. Take context as the starting point.

2. Ensure all activities do no harm.

3. Focus on statebuilding as the central objective.

4. Prioritize prevention.

5. Recognize the links between political, security and development objectives.

6. Promote non-discrimination as a basis for inclusive and stable societies.

7. Align with local priorities in different ways and in different contexts.

8. Agree on practical co-ordination mechanisms between international actors.

9. Act fast … but stay engaged long enough to give success a chance.

10. Avoid pockets of exclusion (‘aid orphans’).⁶

Yet, despite broad agreement (at least among practitioners) and a large body of related research, such principles provide only very loose guidance as to what precisely development actors should be doing in fragile situations. How precisely should they take context into account, design and implement activities that do no harm and so on? How should they balance the tensions and challenges inherent in the application of such broad principles? Such guidelines also provide little insight into which projects and programmes donors should support, that is, based on past experience, which have worked, why and why should (or shouldn't) we expect the same outcomes elsewhere?

This article and special issue focus on this last set of questions, building on a large research and policy literature on aid, peacebuilding and development in conflict and post-conflict states. Three areas of research are especially relevant, offering important – but also incomplete – insights.

The first focuses on why and how states become fragile and fall into conflict, including considerable discussion of how external interventions have contributed to, or failed to stop, these processes.⁷ This work often serves as the basis for policy guidance on peacebuilding, with the assumption that ‘successful' peacebuilding will result from taking steps to address the factors that lead to failure.⁸ Likewise, this work provides empirical support to many of the Fragile States Principles. Analyses highlighting the diverse nature of fragile situations, for instance, underscore the value of taking local context into account; research into the role of aid in contributing to ‘structural violence' highlights the value of ‘doing no harm’;⁹ and work on horizontal inequality and conflict is consistent with the principle of promoting non-discrimination and inclusive societies.¹⁰

With respect to our central questions, however, this work has clear shortcomings. Analysis tends to be done at the country level, offering relatively little traction on the specifics of projects and programmes. In addition, it has paid more attention to failure and failed interventions than to success and successful interventions. This lopsided focus can be misleading: processes that lead to peace are not necessarily the converse of those leading to conflict.¹¹ Conflict itself may have momentum such that addressing its initial causes, once started, is not enough to trigger and sustain peaceful change. Moreover, research that focuses solely on situations that share the same outcome (here, failure) suffers from a basic methodological flaw; in selecting on the dependent variable,  it cannot test causal hypotheses about the influences of particular factors on variation in the dependent variable.¹²

A second major body of work relevant to our central questions is the literature on impact evaluation and in particular the ‘gold standard' of experimental research using randomized controlled trials (RCTs). Project and programme evaluations of various types have long been carried out, but this literature has increased dramatically since 2008.¹³ Systematic reviews of impact evaluations also help to consolidate findings from multiple studies. RCTs in particular have the potential to isolate the impact of development interventions from other potential influences, thus providing strong leverage on the questions of ‘what has worked' and ‘what could work' in development programming. Drawing on findings from RCTs, Karlan and Appel, for instance, identify ‘seven ideas that work': microsavings; reminders to save; prepaid fertilizer sales; deworming; remedial education in small groups; chlorine dispensers for clean water; and commitment devices.¹⁴ Likewise, Banerjee and Duflo draw on experimental studies to identify a host of promising interventions in areas ranging from health and education to policing.¹⁵

This body of work also has several clear shortcomings. For one, in terms of peacebuilding, the existing evidence base from impact evaluations is underdeveloped in key areas. A recent ‘evidence gap map' review of impact evaluations relevant to 25 categories across the five PSGs, for instance, found that there is little to no evidence on most categories and identifies only two (community-driven reconstruction and psycho-social programmes for victims) as having a large enough number of studies to be promising for evidence synthesis.¹⁶

While some such knowledge gaps might be remedied by prioritizing new research in understudied areas, the nature of experiments also imposes limits on what they can study.¹⁷ Randomized study of some interventions may be impractical or unethical.¹⁸ RCTs also need a large number of observations in order to gain precise estimates, meaning that they are best employed to study interventions that can be randomized at a relatively low level of aggregation such as the village, in a manner that cannot be done for some national reforms. Some interventions can be expected to show effects only over years, meaning that impact is not captured within the time-frame of most impact evaluations. Further, RCTs are difficult to implement well in any situation – and situations of conflict pose additional logistical and practical challenges.

Additionally, while RCTs can have strong internal validity, critics contend they may have weak external validity.¹⁹ In other words, they can provide strong evidence that a given intervention had a specific impact in a given context, but tend to offer much weaker leverage on assessing whether factors specific to that context were necessary for this measured impact. Thus, they are limited in their ability to speak to broader ‘why' questions and in particular to assess whether and why the same intervention applied in a different context should have the same impact – that is, to issues of transferability and scalability.

A third body of work relevant to our central questions consists of aid success stories, many of which have been elaborated by bilateral and multilateral development organizations. Just a handful of examples include the US Agency for International Development's ‘Success Stories'; the UK Department for International Development's ‘Development Tracker'; the World Bank's Aid That Works (2007) and Yes Africa Can (2011); and the Overseas Development Institute's ‘Development Progress' studies funded by the Gates Foundation.²⁰

This body of work has played, for one, a useful role for donors in providing justification for development assistance at a time of contracting resources and strong criticism of aid.²¹ While many of these success stories fall more into the realm of advocacy than scientific research – and thus call for further examination – an emerging number suggest a movement towards more rigorous and independent discussion of success and best practices, including some specific attention to aid to fragile situations.

The project presented in this special issue extends this third approach to focus on the lessons that can be learned from relatively detailed case studies of a set of generally well-regarded aid-supported initiatives in fragile situations. The approach is inspired in particular by Judith Tendler's Good Government in the Tropics which also explores what can be learned from a set of case studies of ‘successful' reform efforts – in Tendler's case of four government sectors in the state of Ceará, Brazil.²² As Tendler sought to address gaps in a literature on development focused on poor government and failed public sector reform by analysing ‘good government', this project seeks to speak to gaps in a literature on development assistance for peacebuilding drawn disproportionately from analysis of the causes of fragility and failed interventions, by analysing ‘good aid'. In so doing, it provides new traction on some of the questions left unanswered in the rich bodies of work summarized above.

This Collection

The ‘Good Aid' project began with a review of the literature, on the basis of which well-regarded aid-supported projects and programmes in fragile states were identified. As case studies were intended to contribute collectively to broader theory-building, cases were then chosen to include both a diversity of peacebuilding areas and geographic regions. Attention was also paid to research gaps in the sense that (with a few exceptions) priority was given to cases for which relatively little in-depth peer-reviewed analysis had already been published. Authors for each case study were then sought and invited on the basis of their knowledge of selected initiatives and ability to provide independent analysis complemented by their scholarly or policy expertise.²³ A variety of disciplinary approaches, including political science, economics, law and education, are thus reflected in the case studies. A contributors' workshop provided an opportunity for discussion of the case studies and reflection on collective findings.²⁴

In order to facilitate comparison across cases, contributors were asked to address several common issues in their analyses, in particular to evaluate outcomes critically and to analyse causal processes and drivers of success, including – but not necessarily limited to – the role of local context, aid modalities, local ownership, project/programme design and timing. Beyond these guidelines, contributors were free to develop their case studies as they saw fit. Some of the articles in this special issue thus focus more on impact, while others more on explanation. Most early versions of articles discussed explicitly each of the causal factors noted above, but by the final versions included here, discussion of some factors judged less significant was dropped. Case studies also draw on a range of methodological tools and data sources.

The first two case studies in this special issue focus on community-driven development (CDD), that is, programmes that involve ‘community control and management of funds and emphasize community inputs into the planning and decision-making process’.²⁵ Social funds may also be included under this broad category. CDD has received considerable attention by donors. As of 2012, the World Bank for instance supported approximately 400 CDD projects in 94 countries.²⁶ The first study by Andrew Beath, Fotini Christia and Ruben Enikolopov considers the National Solidarity Programme (NSP), Afghanistan's largest development programme. The NSP is among the best known CDD projects and has been widely cited as a positive example.²⁷ Beath et al.'s analysis draws on the NSP Impact Evaluation (NSP-IE), which they led, an RCT including 500 villages across ten districts. The article presents detailed discussion of the NSP's impact – showing in particular positive effects on access to drinking water and electricity, acceptance of democratic processes, perceptions of economic well-being and participation of and attitudes towards women, and more limited effects on perceptions of local and national government performance and economic outcomes. It builds on highly detailed analysis of impact to consider causal processes and wider implications for aid.

The second case study, by Lamis Al-Iryani, Alain de Janvry and Elisabeth Sadoulet, focuses on the Yemen Social Fund for Development (SFD). Like the NSP, the SFD has been widely cited as a positive example of CDD in a fragile environment.²⁸ A national programme, it operates through nine regional offices and by 2012 reached almost a quarter of Yemen's villages for instance. Drawing on review of multiple external assessments, internal documents and first-hand experience with the SFD over more than 16 years, the authors argue that four factors are particularly important to the SFD's achievements: stakeholder ownership and a demand-driven approach; trust stemming from political neutrality in the allocation of resources; flexibility; and the relevance of interventions for beneficiaries.

The next two case studies address initiatives to support the restoration of core government functions. As the Peacebuilding and Statebuilding Goals underscore, strengthening public financial management (PFM) is a key component of building more capable states. PFM has thus become a major area for donor engagement in fragile states, although recent analyses stress that more research is needed to assess the degree to which donor support to PFM systems in fact ‘works’.²⁹ This case study builds on an initiative by the World Bank to address the knowledge gap on PFM in post-conflict states, drawing on eight country case studies.³⁰ Of these, Afghanistan, Kosovo, Sierra Leone and West Bank and Gaza were rated as having made ‘substantial progress'. Heidi Tavakoli, Ismaila Ceesay and Winston Cole build on and extend this analysis in examining the Sierra Leone case. Their article highlights four factors in the country's relatively successful reforms: local ownership and support of reforms; a cadre of motivated and well-qualified local technical advisers; a relatively strong starting point in terms of PFM performance after the cease-fire; and considerable international assistance coordinated through budget support operations.

Support for core government functions, along with legitimate politics, is also the focus of Rauli S. Lepistö, Rachel M. Gisselquist and Jussi Ojala's analysis of Finn Church Aid's 2012 initiative to provide assistance in the form of a seconded staff member to the process of setting up the Federal Government of Somalia. A small-scale initiative, the FCA secondment is considered in the context of other aid-funded programmes to support external advisers within developing country government ministries. A number of such initiatives exist in post-conflict countries – the best known of which is probably the Scott Family Liberia Fellows Programme – and they are generally well regarded. However, very little research has been done to assess critically whether, why and how they ultimately support peacebuilding and statebuilding. The article argues that the secondment contributed in significant ways to the Somali transitional government's capacity to implement the transition roadmap, despite some major limitations. While the initiative benefited overall from a high degree of local ownership, the case study also illustrates some of the tensions faced by external actors in supporting such truly locally owned initiatives.

The subsequent two articles address interventions in the areas of rule of law, security and justice. Laura Bacon considers Liberia's post-conflict gender-sensitive police reforms. Liberia's policing and security sector reforms have been described as arguably among the more successful in sub-Saharan Africa, and Bacon's analysis complements a considerable literature by providing new discussion of issues relevant to women, peace and security.³¹ The article discusses the reform as a qualified success in terms of the Liberia National Police's two key goals, representation and responsiveness. It illustrates the positive impacts of timing, context, local ownership and foreign assistance, as well as the challenges due to low technical capacity and weak existing rule of law.

In the next study, a team of researchers led by Martin Gramatikov considers the Judicial Facilitators Programme in Nicaragua, which was first implemented in selected rural communities in 1998 as a pilot project of the Organization of American States. As Nicaragua is not currently a fragile state, it is worth remembering that it was then less than a decade out of its civil war, and the programme operated first in areas in which the judicial reach of the state was weakest. The programme has since been scaled up within the country and used as a model for pilots in other countries in the region, testament to its positive reputation.³² More broadly, community-based paralegal programmes are widely considered to be a promising way of supporting improved access to justice in developing countries. The ‘Timap' programme in post-conflict Sierra Leone in particular is the subject of several analyses.³³ Other examples can be found in Bangladesh, India, Malawi, the Philippines, South Africa and Zimbabwe. Focusing on the scaling up of the Judicial Facilitators Programme from rural to urban areas within Nicaragua, the article approaches evaluation of the community-based paralegal model from a new perspective, building on previous evaluations and drawing on original survey data and in-depth interviews. The data suggest that the expanded programme had some positive impacts on access to justice and levels of conflict, but that these were more modest than expected based on previous experience with the programme in rural areas. Discussion suggests for one that the different characteristics of urban versus rural communities may have influenced outcomes as the success of the programme has been due heavily, for instance, to communities' trust in facilitators.

The final two articles in this special issue focus on the delivery of basic services in the areas of education and health. Returning to the quintessential failed state of Somalia,³⁴ James H. Williams and William C. Cummings consider UNICEF's generally well-regarded Education Programme, which operated from 1996 to 2010.³⁵ Although other international agencies also supported education in Somalia during this period, UNICEF took a leading role. As the article highlights, enrolment rates remain low, but they have improved from an even lower base. It finds that the programme's achievements can be attributed to a variety of factors, including attention to local context, community ownership and particular programme design elements. The article thus provides valuable discussion on how aid supported education, and through it peacebuilding, even in the most fragile of contexts.

The final article in this special issue, by Andrew Rosser and Sharna Bremner addresses the World Bank's health projects in Timor Leste – specifically, the Health Sector Rehabilitation and Development Project (HSRDP I) and the Second Health Sector Rehabilitation and Development Project (HSRDP II). The study builds on and extends Rosser's previous analysis of these projects for the World Bank's Aid That Works volume. ³⁶ By considering developments since this analysis was completed, it offers new assessment of why the projects had the outcomes they did and the relative importance of various factors. The article underscores in particular the role of political economy factors in shaping the relative success of these projects. More broadly, the authors argue, the case illustrates that aid effectiveness is ‘a function not just of the technical quality of project design and the administrative competence of project managers but also the extent to which there is congruence between donor and local elites' agendas'.³⁷

What Works, Why and Implications for Transferability and Scalability

Evaluating Outcomes Using Diverse Methods

The studies in this special issue illustrate the qualified and contingent nature of ‘success' even in some of the most well-regarded initiatives. Some aspects of each initiative are clearly less successful than others, and some have positive results but arguably could have been more successful under other circumstances.

Precise indicators of ‘success' differ across the cases, but in general the focus is on broader ‘outcomes' relevant to peacebuilding objectives, rather than on ‘outputs'. Beath et al.'s analysis of the NSP in Afghanistan, for instance, does not focus on how many grants were distributed but on a host of economic, political and social indicators of impact. Indeed, the NSP-IE collected information on 21 groups of indicators across five categories: access to utilities; economic welfare; governance; political attitudes; and social norms. Alternatively, Bacon's analysis of Liberian police reform highlights representation and responsiveness as ultimate goals and uses a variety of metrics to evaluate performance against these goals.

In assessing outcomes and impact, a key question concerns the appropriate baseline or counterfactual against which impact should be assessed.³⁸ Several approaches are typical. For one, analyses may focus on the state of affairs within a country ‘before' and ‘after' project or programme implementation. As other factors that occurred in the interim may influence outcomes, analyses further seek to isolate the impact of the intervention from these other influences. Experimental and quasi-experimental methods, such as employed by Beath et al. and Gramatikov et al. in this special issue, can draw on sophisticated quantitative techniques to control for such factors.

Non-experimental and qualitative work may also rely on counterfactual reasoning and inference.³⁹ Lepistö et al., for instance, consider whether the political transition in Somalia would have been the same absent Finn Church Aid's secondment. Their counterfactual is drawn largely on the basis of assessments by participants in the process, who reason that the transition process would still likely have proceeded without the secondment, but that it would have been less successful given that the Ministry of Constitutional Affairs and Reconciliation would have had lower capacity to pursue its core activities.

Another approach involves considering the impact of the intervention against that of other projects within the same country, which presumably are also subject to many of the same national influences. This approach is developed in Rosser and Bremner's analysis of the World Bank's health projects in Timor Leste. The impacts of these projects, they argue, rate comparatively well against the achievements of donor-supported educational projects in Timor Leste during the same period. As explored further below, this comparison also helps Rosser and Bremner to identify some of the unique contextual factors that supported more positive outcomes in health than in education.

In addition, outcomes may be assessed through comparative consideration of similar initiatives in other countries. As noted above, for instance, PFM reforms in Sierra Leone are identified as a relative success based on comparative study of PFM reforms in eight post-conflict countries.⁴⁰ One challenge in using such cross-country comparisons is that countries in fact may not be comparable, in the sense that other factors – beyond the initiative itself – may explain why projects in some countries show better outcomes than in others. For this reason, combining cross-country evaluation with within-case comparison is a promising strategy.

A further approach to evaluating outcomes is process tracing.⁴¹ While this approach receives relatively little attention in most work on impact evaluation, it is a standard qualitative tool for social scientists in evaluating causal hypotheses. If we consider, for instance, the hypothesis that Liberia's gender-sensitive police reforms had an impact on representation and responsiveness, Bacon's tracing of the trial and error process through which Liberia's reforms evolved to achieve these goals better, provides a loose example of process tracing. While none of the case studies in this special issue explicitly employ this method, it is a tool that could be further developed in future research.

Beyond Impact Evaluation: Considering Causal Processes and ‘Drivers' of Success

Assessment of development projects and programmes has in some circles become synonymous with impact evaluation. This special issue seeks to push beyond standard impact evaluation approaches towards greater focus on causal processes and ‘drivers' of success. Impact evaluations and RCTs also address causal hypotheses, but their focus has been different.⁴² For one, they tend to focus on establishing a (causal) link between an intervention and various outcomes, but not on the intermediate steps and process through which the intervention leads to these outcomes. A focus on causal processes and drivers also suggests more disaggregation in the sense that it may take note of how particular aspects of a project or programme – rather than necessarily the project or programme as a whole – led to outcomes. RCTs may also do this, but it is arguably less common. (Beath et al.'s analysis of Afghanistan's NSP offers several such examples. For instance, it considers separately the impact of NSP-funded projects in different sectors, showing different impacts for projects in water and electricity versus irrigation and transportation.⁴³) Analysis along these lines can provide new traction on policy by offering insight into whether specific changes or ‘tweaks' to projects and programmes may improve outcomes.

Causal processes and drivers can be explored in multiple ways. One approach – adopted by the majority of studies in this special issue – is to consider in turn the role of key factors identified in the literature as important for aid success and to explore how each played out in a given case. An example of this is Williams and Cummings's discussion of the role of local context, aid modalities, local ownership and programme design elements. They both explore how each is manifested in UNICEF's education programme in Somalia and tie these factors to specific effects. They find, for instance, that local ownership took the form of community ownership as it was the local communities that largely assumed responsibility for managing schools when the central government collapsed. While this community ownership had some drawbacks, it was valuable in the sense that it supported considerable resilience for activities at the local level, regardless of the fragility of the central government.

An alternative approach is to present a more comprehensive theory or argument that identifies the most significant factors and how they interact to create outcomes, and then to test or demonstrate the plausibility of this argument through the case study. Rosser and Bremner, for instance, adopt this approach in their study of health interventions in Timor Leste, arguing that ‘the HSRDPs benefitted from (a) a political economy context that was relatively conducive to aid effectiveness in general and (b) the fact that there was relatively little elite resistance to the World Bank's agenda in the health sector’.⁴⁴ The key factors noted above are then implicitly considered within the context of this central argument.

A variety of causal factors and explanatory arguments are identified in the case studies. Building both on the literature and a collective reading of the case studies, three sets of factors emerge as especially relevant in understanding more broadly the outcomes of development assistance for peacebuilding in fragile environments.

The first concerns the area of intervention and whether it requires extensive or weak involvement and reform of domestic state institutions.⁴⁵ This speaks to the inherent difficulty and likelihood of change in a particular domain. Fragile states, by definition, have extremely weak state authority, capacity and legitimacy.⁴⁶ Interventions that thus rely on the authority, capacity and legitimacy of the state to work – or that centre around strengthening of the state – thus face clear challenges. While all development interventions potentially involve the state, some might be structured to do so less than others. Discrete interventions to address specific health issues such as the international guinea worm eradication campaign, an aid-supported success, offer one example.⁴⁷ In the cases explored in this special issue, several initiatives work in whole or in part through collaboration with non-governmental or sub-national partners, sometimes alongside national state agencies. Examples include community-driven development like Afghanistan's NSP and Yemen's SFD, and UNICEF's collaboration on education with local communities. The Judicial Facilitators Programme in Nicaragua is another example. Although run by the Organization of American States (OAS) and the Nicaraguan judiciary, it is designed in recognition of the fact that the judiciary does not adequately function throughout the country and relies on the use of community leaders as local proxies. The underlying logic here is broadly similar to Paul Collier's recommendation that one promising strategy for donors to have impact in fragile states is to work outside the state by supporting independent public service agencies.⁴⁸

Likewise, it may be especially difficult to show progress from interventions in areas hinging on core functions of the state that are not easily privatized or handled by other actors, or in which reform of the state is the main objective. Examples from this special issue include support for political transition in Somalia (Lepistö et al.), policing and security sector reform in Liberia (Bacon) and reform of the public financial management (PFM) system in Sierra Leone (Tavakoli et al.). That said, even in these areas, the cases illustrate some room for manoeuvre. In Sierra Leone, for instance, Tavakoli et al. argue that one way in which civil service capacity limitations within government were addressed was through the use of a highly skilled ‘parallel public service' responsible for high-level policy development, planning and implementation. Similarly, Lepistö et al. argue that the use of ‘embedded' advisers within core government ministries can support greater functional state capacity.

That said, strengthening the authority, capacity and legitimacy of the fragile state is at the heart of statebuilding and thus ultimately core to peacebuilding. While circumventing the state may improve the chances that particular initiatives show results, development assistance for peacebuilding ultimately should support – or at the least not undermine – the building of a capable, legitimate and authoritative state. Indeed, for instance, as Beath et al. point out, Afghanistan's NSP was designed in part to extend the state's administrative reach. It also appears to have had effects on government legitimacy – and arguably by extension, state legitimacy – as data show a positive impact on perceptions of the favourability of government. This effect notably fades over time with the completion of NSP-funded projects, suggesting that impact on government legitimacy could be maximized through regularized and frequent project activity. Likewise, in Yemen, Al-Iryani et al. find that the SFD's broad, nationwide coverage and transparent, participatory processes contributed to the country's institutional development and state-building process.

A second set of factors falls broadly under the label ‘local context', in particular windows of opportunity, capacity and the existence of local supporters or champions. This set of factors speaks broadly to the ‘ripeness' of particular situations for intervention and reform. While donors for the most part cannot control these factors directly, they can take them into account in assessing whether to attempt a particular project in a given context and in adapting, as possible, to improve a project's viability. Major events such as the cessation of conflict or the election of new leaders can provide windows of opportunity to introduce institutional changes that might not otherwise be possible.⁴⁹ Changes in one area may also facilitate changes in others. As Bacon discusses, for instance, post-conflict security sector reforms in Liberia, which involved deactivating police officers and inviting them to re-apply for the new service, were an opportunity to bring in new recruits and build a new policing culture.

Weak civil service capacity – and by extension weak state capacity – are also highlighted in the case studies as important to the ability of donors to work through and with state actors. To a large extent, this is a chicken and egg problem: the success of interventions in supporting better state institutions is affected by the underlying quality of public institutions themselves. As discussed above, fragile states as a group suffer from relatively weak capacity in comparison to non-fragile states, but some clearly have weaker capacity than others. Capacity for instance is likely to differ between a post-conflict state that was previously a functional state before hostilities began, and one which was not.⁵⁰ Donors may also, over time, seek to influence change in capacity, for instance through training and capacity-building initiatives. However, such activities take time to show results.

The issue of local support and local champions is particularly interesting – and closely connected to the growing recognition in development circles that politics and political economy should be more fully considered in aid policy.⁵¹ Indeed, it is telling that some form of country ownership – whether at the level of national government, sub-national government or non-state actors – can be seen in all of the cases explored in this special issue. The majority of the cases explored here are in fact described as largely national reform processes or initiatives that are supported by foreign assistance, rather than donor-initiated or donor-led efforts.

Rosser and Bremner's article on health interventions in Timor Leste is one exception. Their discussion underscores the differences that may exist between the priorities of donors and those of local actors, as well the fact that local actors may back donor-supported initiatives for strategic reasons. Indeed, reliance on foreign assistance can provide strong incentives for local cooperation. Nevertheless, as the literature suggests, local support or ‘ownership' generally facilitates better results – perhaps even when it is tactical. In Timor Leste, Rosser and Bremner find that Fretilin and Falintil leaderships provided broad support to the World Bank's health projects, which greatly facilitated their relative success. The donor agenda on education, however, was less attractive for local leaderships to support as it was in conflict with their own nation-building efforts, which included use of Portuguese in schools – rather than Bahasa Indonesia or Tetun (an indigenous language), as favoured by some donors. This made for much weaker local support on education projects, which were also in turn relatively less successful.

Further, while the literature underscores that donors should take politics into account, it also notes the risks of engaging in politics directly. This point is supported in particular in Al-Iryani et al.'s analysis of the SFD in Yemen, which identifies its political neutrality and impartial mechanisms for delivering resources as a key to stakeholders' trust and the SFD's performance.

Finally, while local contextual factors are often treated as fixed constraints, they can change – and indeed the objective of some aid-supported initiatives is precisely such change. This special issue presents several examples in the area of gender. Gender-sensitive policing reforms in Liberia, for one, were a clear effort to change traditional cultural norms and the male-domination of the security sector. Likewise in Afghanistan, the NSP's policy of mandating female participation was purposely contrary to recognized local customs.

A third set of factors relates to project and programme design and management. Thus, they are largely within the control of development administrators and are the most clearly transferrable and scalable elements of ‘successful' projects and programmes. One of the truisms of recent discussions in development is that ‘one-size-fits-all' policies do not work and that context must be taken into account in the design of aid interventions. That said, it is clear in examining aid projects and programmes that some models could be a decent starting point for interventions across multiple contexts. Several examples highlighted in this special issue include community-driven development, community-based paralegal programmes and specific project components like the ‘teachers/school kits' and ‘pupil kits' employed in UNICEF's education programme in Somalia.

Another promising element of programme design and management identified in the cases relates to internal organizational mechanisms that allow for continued operations during periods of crisis. This point is highlighted in particular in Al-Iryani et al.'s discussion of how Yemen's SFD was able to continue its operations (including implementation of over 80 per cent of its projects) during Yemen's 2011 economic and political crisis. While some donors stopped funding during the crisis, Al-Iryani et al. find that the SFD's reliance on multiple sources of funding helped to make possible continued activities. In addition, in terms of implementation, continued operations were facilitated for instance by shifting interventions more towards cash-for-work and water projects implemented directly by beneficiaries using local materials, and with immediate impact on local livelihoods.

Broadly, design and management that is flexible enough to be responsive to local context, local priorities and an unstable environment is highlighted in many of the cases. This approach is similar to what Lant Pritchett, Michael Woolcock and Matt Andrews label ‘Problem Driven Iterative Adaptation' (PDIA) and advocate for as a promising way to do development that can address key weaknesses in how development has been done.⁵² PDIA is based on four principles: (1) addressing locally identified problems; (2) creating an authorizing environment for experimentation; (3) supporting iterative and rapid experiential learning through tight feedback loops; and (4) engaging broad sets of local agents to promote viable, legitimate interventions.⁵³

External Validity, Transferability and Scalability

The third set of factors identified above is largely transferrable and scalable, while the first two sets of factors are less so. In assessing whether a project or programme is likely to transfer successfully to another environment (or to be scaled up successfully), consideration of the first two sets of factors then is particularly important. Narrowly focused studies on the impact of specific projects do not provide this leverage as they cannot separate out the influence of local contextual factors, which may be necessary for initiatives to function as they do. Gramatikov et al.'s analysis of the scaling up of the Judicial Facilitators Programme in Nicaragua provides one illustration of how contextual factors – such as community-level trust – may provide a necessary condition for an initiative to function in one environment that may not be obvious until the programme is extended to other environments.

This then underscores the gaps in the literature on aid effectiveness left by a focus on impact evaluation using experimental methods, as discussed above. From a theoretical standpoint, external validity is important in theory-building and theory-testing. While an internally valid argument can tell us about a particular case, without some traction on external validity, we cannot use the particular case to speak to broader relationships and generalizable theories. From a more practical standpoint, if we do not know whether and how factors external to the intervention influenced results, we cannot say how the same intervention will function in a different environment. The approach adopted above – moving beyond a narrow approach to impact evaluation to focus more on explanation and causal processes – thus is one way to begin to address such concerns.

We might gain even more traction by pushing such analysis to engage and build more directly on broader theoretical frameworks.⁵⁴ A simple example is suggested by Gramatikov et al.'s discussion. In retrospect, if early assessment of the Judicial Facilitators Programme had identified social trust and the authority of community leaders as potential ‘drivers' of the success of the programme, consideration of theories of social capital might have helped to refine expectations about the transferability and scalability of the programme. Relevant theoretical and empirical work generally suggests lower levels of social capital in urban versus rural areas; thus, to the extent that the success of the programme is due to strong social capital, we can predict weaker impact for the same programme implemented in other regions, such as urban areas, with weaker social capital.⁵⁵ Further work along these lines remains one topic for future research.

Conclusion and Extensions

Among the core and continuing challenges for both scholars and practitioners of peacebuilding is the question of how positive change occurs, peace is built and fragile states become more resilient. As this article has argued, despite a large and rich literature, there are major gaps in existing work. New analysis of ‘good aid' performance, as presented in this special issue, thus can be useful, both from a practical and theoretical perspective. In terms of how positive change occurs, the discussion above underscores both limits and potential for external actors to support peacebuilding. In particular, external actors have direct control over project and programme design and management, but only limited influence over local contextual factors as well as the area of particular inventions and how much they necessarily involve domestic state institutions. In these latter areas, local actors and institutions and endogenous processes play a more decisive role.

One key area for future research is to elaborate further and build theory on experience with reform efforts and processes. One way to proceed is to apply more widely and consider implications based on existing models of policy making (including consideration of how such models might be adjusted for fragile states). A simple application of John Kingdon's agenda-setting model provides one example.⁵⁶

Kingdon's model posits three process streams – problems, proposals and politics. Windows of opportunity for policy change emerge when these three streams combine, and they may be predictable (such as due to a scheduled vote on legislation) or sudden. Such windows provide policy entrepreneurs with the opportunity to push forward their agendas and policy solutions. Within this model, development actors might be seen as one type of policy entrepreneur. They can also contribute to problem and proposals streams by bringing issues to the attention of policy-makers, advancing proposals, and creating incentives for policy-makers to identify issues as problems and to support donor proposals to address them. However, they generally do not act directly within the domestic politics stream. Likewise, domestic interest groups may play analogous roles. While they may have more limited resources and perhaps information than development actors with which to influence policy-makers in the problem and proposals streams, they may act directly and have additional leverage within the politics stream.

Applying Kingdon's model in this way, it should not be surprising that the influence of development actors on reforms and policy outcomes in developing countries is indirect and sometimes quite limited. However, the same might also be said of local actors. Extending this application thus would seem to offer some potentially interesting points of discussion with recent scholarly literature on peacebuilding, which has tended to highlight the importance of local over external actors in explaining outcomes. This recent focus in turn might be understood as a response to an opposite emphasis in earlier literature and policy discussions.

As this special issue also suggests, reality is likely somewhere in the middle. The case studies offer eight analyses of instances in which development assistance, through interaction with local actors and processes, supported some positive peacebuilding change. Still, none are unambiguous successes and all underscore that development assistance provides no easy fix for peacebuilding.

ACKNOWLEDGEMENTS

This special issue was supported by UNU-WIDER, and in particular presents work developed under its Research and Communication on Foreign Aid (ReCom) programme (2011–13). We gratefully acknowledge specific programme contributions from the governments of Denmark (Ministry of Foreign Affairs, Danida) and Sweden (Swedish International Development Cooperation Agency – Sida) for ReCom, as well as core financial support to the UNU-WIDER work programme from the governments of Denmark, Finland, Sweden and the United Kingdom. Shohini Banerjee provided valuable research assistance. For helpful comments, a special thanks to Tony Addison, Neil Cooper, Philip Cunliffe, David Gisselquist, Omar McDoom, Finn Tarp and all project participants.

ABOUT THE AUTHOR

Rachel M. Gisselquist is a Research Fellow with the UNU-WIDER. She works on the politics of the developing world, with particular attention to ethnic politics and group-based inequality, state fragility, governance and democratization in sub-Saharan Africa. Her work is published in various journals and edited volumes. She holds a PhD in Political Science from the Massachusetts Institute of Technology.

Notes

1. Stephen Baranyi and Marie-Eve Desrosiers, ‘Development Cooperation in Fragile States: Filling or Perpetuating Gaps?', Conflict, Security & Development, Vol.12, No.5, 2012, pp.443–59; Keith Krause and Oliver Jütersonke, ‘Peace, Security and Development in Post-Conflict Environments', Security Dialogue, Vol.36, No.4, 2005, pp.447–62; Peter Uvin, ‘The Development/Peacebuilding Nexus: A Typology and History of Changing Paradigms', Journal of Peacebuilding & Development, Vol.1, No.1, 2002, pp.5–24; Pieter van Houten, ‘The World Bank's (Post-) Conflict Agenda: The Challenge of Integrating Development and Security', Cambridge Review of International Affairs, Vol.20, No.4, 2007, pp.639–57; J. David Whaley, ‘Improving UN Developmental Co-ordination within Peace Missions', International Peacekeeping, Vol.3, No.2, 1996, pp.107–22; World Bank, World Development Report: Conflict, Security, and Development, Washington, DC: World Bank, 2011.

2. OECD, Fragile States 2014: Domestic Revenue Mobilisation in Fragile States, Paris: OECD, 2014.

3. OECD, States of Fragility 2015: Meeting Post-2015 Ambitions, Paris: OECD, 2015.

4. It is worth noting the considerable debate in the literature over the concept of fragility. See, e.g., Lars Engberg-Pedersen, Louise Andersen and Finn Stepputat, ‘Fragile Situations: Current Debates and Central Dilemmas', DIIS Report, Copenhagen: Danish Institute for International Studies (DIIS), 2008; Rachel M. Gisselquist, ‘Varieties of Fragility: Implications for Aid', Third World Quarterly, Vol.36, No.7, forthcoming 2015, doi: 10.1080/01436597.2015.1041104; Sonja Grimm, Nicolas Lemay-Hébert and Olivier Nay, ‘“Fragile States”: Introducing a Political Concept', Third World Quarterly, Vol.35, No.2, 2014, pp.197–209.

5. Stephen D. Krasner and Thomas Risse, ‘External Actors, State-Building, and Service Provision in Areas of Limited Statehood: Introduction', Governance, Vol.27, No.4, 2014, pp.545–67; Gearoid Millar, Jaïr van der Lijn and Willemijn Verkoren, ‘Peacebuilding Plans and Local Reconfigurations: Frictions between Imported Processes and Indigenous Practices', International Peacekeeping, Vol.20, No.2, 2013, pp.137–43.

6. From ‘The 10 Fragile States Principles' http://www.oecd.org/dacfragilestates/ (at: www.oecd.org/dacfragilestates/).

7. A few examples include Robert H. Bates, When Things Fell Apart: State Failure in Late-Century Africa, Cambridge: Cambridge University Press, 2008; Jack A. Goldstone, ‘Pathways to State Failure', Conflict Management and Peace Science, Vol.25, No.4, 2008, pp.285–96; National Academy of Public Administration, Why Foreign Aid to Haiti Failed, Washington, DC: National Academy of Public Administration, 2006; Robert I. Rotberg (ed.), State Failure and State Weakness in a Time of Terror, Washington, DC: Brookings Institution, 2003.

8. The OECD's 2012 report, ‘Think Global, Act Global', for instance, draws on analysis of the causes of conflict and fragility to identify nine ‘entry points' and ‘global initiatives for action' to confront these factors.

9. See, e.g., Peter Uvin, Aiding Violence: The Development Enterprise in Rwanda, West Hartford, CT: Kumarian Press, 1998.

10. See, e.g., Frances Stewart (ed.), Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies, Basingstoke, Hampshire: Palgrave Macmillan, 2008.

11. Rachel M. Gisselquist, ‘Aid and Institution-Building in Fragile States: What Do We Know? What Can Comparative Analysis Add?', Annals of the American Academy of Political and Social Science, Vol.656, No.1, 2014, pp.6–21.

12. Barbara Geddes, ‘How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics', Political Analysis, Vol.2, No.1, 1990, pp.131–50; Gary King, Robert O. Keohane and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research, Princeton, NJ: Princeton University Press, 1994.

13. Drew B. Cameron, Anjini Mishra and Annette N. Brown, ‘The Growth of Impact Evaluation for International Development: How Much Have We Learned?', Journal of Development Effectiveness, DOI: 10.1080/19439342.2015.1034156. Published online: 28 Apr. 2015.

14. Dean Karlan and Jacob Appel, More Than Good Intentions: Improving the Ways the World's Poor Borrow, Save, Farm, Learn, and Stay Healthy, New York: Plume, 2012.

15. Abhijit Banerjee and Esther Duflo, Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty, New York: Public Affairs, 2012.

16. Drew B. Cameron, Annette N. Brown, Anjini Mishra, Mario Picon, Hisham Esper, Flor Calvo and Katia Peterson, ‘Evidence for Peacebuilding: An Evidence Gap Map', Evidence Gap Map Report 1, New Delhi: International Initiative for Impact Evaluation (3ie), 2015.

17. See, e.g., Angus Deaton, ‘Instruments, Randomization, and Learning about Development', Journal of Economic Literature, Vol.48, No.2, 2010, pp.424–55; Rachel M. Gisselquist and Miguel Niño-Zarazúa, ‘What Can Experiments Tell Us about How to Improve Government Performance?', Journal of Globalization and Development, Vol.6, No.1, 2015, pp.1–45.

18. Macartan Humphreys, ‘Reflections on the Ethics of Social Experimentation', Journal of Globalization and Development, Vol.6, No.1, 2015, pp.87–112.

19. See, e.g., Rajeev Dehejia, ‘Experimental and Non-Experimental Methods in Development Economics: A Porous Dialectic', Journal of Globalization and Development, Vol.6, No.1, 2015, pp.47–69; Lant Pritchett and Justin Sandefur, ‘Context Matters for Size: Why External Validity Claims and Development Practice Don't Mix', Working Paper 336, Washington, DC: Center for Global Development, 2013.

20. See www.usaid.gov/results-data/success-stories; http://devtracker.dfid.gov.uk/; www.odi.org/programmes/development-progress; Punam Chuhan-Pole and Manka Angwafo, Yes Africa Can: Success Stories from a Dynamic Continent, Washington, DC: World Bank, 2011; James Manor (ed.), Aid That Works: Successful Development in Fragile States, Washington, DC: World Bank, 2007.

21. E.g. William Easterly, The White Man's Burden: Why the West's Efforts to Aid the Rest Have Done So Much Ill and So Little Good, New York: Penguin Press, 2006; Dambisa Moyo, Dead Aid: Why Aid Is Not Working and How There Is a Better Way for Africa, New York: Farrar, Straus and Giroux, 2009.

22. Judith Tendler, Good Government in the Tropics, Baltimore, MD: Johns Hopkins University Press, 1997.

23. Qualified contributors could not be found for several selected cases and these cases thus had to be dropped. The project included nine case studies.

24. The workshop was held 12–13 Dec. 2013.

25. Susan Wong, What Have Been the Impacts of World Bank Community-Driven Development Programs? CDD Impact Evaluation Review and Operational & Research Implications, Washington, DC: World Bank, 2012, p.4.

26. Ibid.

27. See, e.g., Ashraf Ghani and Clare Lockhart, Fixing Failed States: A Framework for Rebuilding a Fractured World, Oxford and New York: Oxford University Press, 2008; John A. Nagl, Andrew M. Exum and Ahmed A. Humayun, A Pathway to Success in Afghanistan: The National Solidarity Program, Washington, DC: Center for a New American Strategy, 2009; Robert B. Zoellick, ‘The Key to Rebuilding Afghanistan', Washington Post, 22 Aug. 2008 (at: http://www.washingtonpost.com/wp-dyn/content/article/2008/08/21/AR2008082103110.html), last accessed 24 June 2015.

28. See, e.g., Lawrence Chandy and Johannes F. Linn, Taking Development Activities to Scale in Fragile and Low Capacity Environments, Washington, DC: The Brookings Institution, 2011; Marcus Cox and Nigel Thornton, Managing Results in Conflict-Affected and Fragile States: A Stock-Take of Lessons, Experience and Practice, London: DFID, 2010.

29. Paolo de Renzio, Matt Andrews and Zac Mills, ‘Does Donor Support to Public Financial Management Reforms in Developing Countries Work? An Analytical Study of Quantitative Cross-Country Evidence’, Working Paper, London: ODI, 2011.

30. World Bank, ‘Public Financial Management Reforms in Post-Conflict Countries: Synthesis Report', Washington, DC: World Bank, 2012.

31. See, e.g., Laura M. Bacon, Building an Inclusive, Responsive National Police Service: Gender-Sensitive Reform in Liberia, 2005–2011, Princeton, NJ: Innovations for Successful Societies, Princeton University, 2012; Otwin Marenin, ‘Styles of Policing and Economic Development in African States', Public Administration and Development, Vol.34, No.3, 2014, pp.149–61; Karen Barnes Robinson and Craig Valters, Progress in Small Steps: Security Against the Odds in Liberia, London: Overseas Development Institute, 2015.

32. Margot Kokke and Pedro Vuskovic, ‘Legal Empowerment of the Poor in Nicaragua: Microjustice avant-la-lettre?', Social Science Research Network, 2010. In 2011, it received an ‘Innovating Justice' award (at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1674020), last accessed 24 June 2015.

33. Pamela Dale, Delivering Justice to Sierra Leone's Poor: An Analysis of the Work of Timap for Justice, Washington, DC: World Bank, 2009; Vivek Maru, ‘Between Law and Society: Paralegals and the Provision of Justice Services in Sierra Leone and Worldwide', Yale Journal of International Law, Vol.31, No.2, 2006, pp.427–75.

34. Ken Menkhaus, ‘State Failure, State-Building, and Prospects for a “Functional Failed State” in Somalia', The ANNALS of the American Academy of Political and Social Science, Vol.656, No.1, 2014, pp.154–72.

35. See, e.g., Chris Berry, ‘A Framework for Assessing the Effectiveness of the Delivery of Education Aid in Fragile States', Journal of Education for International Development, Vol.4, No.1, 2009 (at: http://www.equip123.net/jeid/articles/8/Berry-FrameworkAssessingtheEffectivenessDeliveryEducationAidFragileStates.pdf), last accessed 24 June 2015; Lawrence Chandy and Johannes F. Linn, Taking Development Activities to Scale in Fragile and Low Capacity Environments, Washington, DC: The Brookings Institution, 2011.

36. James Manor (ed.), Aid That Works: Successful Development in Fragile States, Washington, DC: World Bank, 2007.

37. Andrew Rosser and Sharna Bremner, ‘The World Bank’s Health Projects in Timor-Leste: The Political Economy of Effective Aid', International Peacekeeping, Vol.22, No.4, 2015, pp.435–51. doi: 10.1080/13533312.2015.1059731

38. Howard White, ‘A Contribution to Current Debates in Impact Evaluation', Evaluation, Vol.16, No.2, 2010, pp.153–64.

39. Philip E. Tetlock and Aaron Belkin (eds), Counterfactual Thought Experiments in World Politics: Logical, Methodology, and Psychological Perspectives, Princeton, NJ: Princeton University Press, 1996.

40. World Bank (see n.30 above).

41. David Collier, ‘Understanding Process Tracing', PS: Political Science & Politics, Vol.44, No.4, 2011, pp.823–30.

42. Note also that some argue that impact evaluation has moved towards greater focus on explanation. See, e.g., Elliot Stern, ‘Impact Evaluation: A Guide for Commissioners and Managers’, Report prepared for the Big Lottery Fund, Bond, Comic Relief and the Department for International Development, London: Bond, 2015.

43. Some impact evaluations are indeed designed to explore the impact of particular project components as well. See, e.g., Abhijit Banerjee, Esther Duflo, Daniel Keniston and Nina Singh, ‘Making Police Reform Real: The Rajasthan Experiment', Cambridge, MA: Massachusetts Institute of Technology, n.d.

44. Andrew Rosser and Sharna Bremner (see n.37 above).

45. A related issue concerns how sector characteristics may influence service delivery. Claire Mcloughlin and Richard Batley, ‘The Effects of Sector Characteristics on Accountability Relationships in Service Delivery', Working Paper 350, London: Overseas Development Institute, 2012. See also Marco Schäferhoff, ‘External Actors and the Provision of Public Health Services in Somalia', Governance, Vol.27, No.4, 2014, pp.675–95.

46. See, e.g., David Carment, Stewart Prest and Yiagadeesen Samy, Security, Development, and the Fragile State: Bridging the Gap between Theory and Policy, London: Routledge, 2010; Jörn Grävingholt, Sebastian Ziaja and Merle Kreibaum, ‘Disaggregating State Fragility: A Method to Establish a Multidimensional Empirical Typology', Third World Quarterly, Vol.36, No.7, forthcoming 2015.

47. Sandy Cairncross, Ralph Muller and Nevio Zagaria, ‘Dracunculiasis (Guinea Worm Disease) and the Eradication Initiative', Clinical Microbiology Reviews, Vol.15, No.2, 2002, pp.223–46; Frank O. Richards, Ernesto Ruiz-Tiben and Donald R. Hopkins, ‘Dracunculiasis Eradication and the Legacy of the Smallpox Campaign: What's New and Innovative? What's Old and Principled?', Vaccine, Vol.29, Suppl.4, 2011, pp.D86–D90.

48. Paul Collier, ‘How to Spend It: The Organization of Public Spending and Aid Effectiveness', WIDER Working Paper, Helsinki: UNU-WIDER, 2012.

49. See, e.g., Maria Koinova, Ethnonationalist Conflict in Postcommunist States: Varieties of Governance in Bulgaria, Macedonia, and Kosovo, Philadelphia, PA: University of Pennsylvania Press, 2013; Richard J. Samuels, Machiavelli's Children: Leaders and Their Legacies in Italy and Japan, Ithaca, NY: Cornell University Press, 2003.

50. Gisselquist (see n.4 above).

51. See David Booth, ‘Aid Effectiveness: Bringing Country Ownership (and Politics) Back In', Conflict, Security & Development, Vol.12, No.5, 2012, pp.537–58.

52. Lant Pritchett, Michael Woolcock and Matt Andrews, ‘Looking Like a State: Techniques of Persistent Failure in State Capability for Implementation', Journal of Development Studies, Vol.49, No.1, 2013, pp.1–18.

53. Matt Andrews, Lant Pritchett and Michael Woolcock, ‘Escaping Capability Traps through Problem Driven Iterative Adaptation (PDIA)', World Development, Vol.51, 2013, pp.234–44.

54. This approach is also a way to improve external validity in experimental work. Fernando Martel Garcia and Leonard Wantchekon, ‘Theory, External Validity, and Experimental Inference: Some Conjectures', The ANNALS of the American Academy of Political and Social Science, Vol.628, 2010, pp.132–47.

55. Jens F.L. Sørensen, ‘Rural–Urban Differences in Bonding and Bridging Social Capital', Regional Studies, DOI: 10.1080/00343404.2014.918945. Published online: 1 Jul. 2014.

56. John W. Kingdon, Agendas, Alternatives, and Public Policies, New York: HarperCollins, 1995.