DFID’s governance work in Nepal and Uganda

Score summary

DFID’s governance work in Nepal and Uganda is tailored to the country contexts and largely achieving its intended outcomes, but needs a stronger strategic orientation and approach to learning.


In both Nepal and Uganda, DFID’s approach to its governance work is relevant, with clear evidence that it is tailoring the initial design of its programming to country contexts. Less clear is how DFID governance programmes are intended to work together to achieve more transformational impact.

In both countries, DFID’s governance programmes have mostly achieved their intended objectives. We saw good examples of change at the local level, through DFID’s local governance and voice and accountability programming, and in public financial management. However, it is difficult to judge the portfolio-level performance of governance programming because of both deficits in monitoring and evaluation and a lack of country-level strategic objectives on governance. Some of the strongest results have been achieved through sustained engagement with particular thematic areas or institutions over multiple programme cycles, enabling DFID to develop its knowledge and relationships, and through flexible and adaptive approaches to programme delivery (especially in Nepal). This is in line with current thinking on good practice. However, DFID’s programme management practices do not necessarily demonstrate sustained engagement and flexible and adaptive ways of working. DFID governance advisers provide advice and support on the governance aspects of programming in other sectors, but could provide more ongoing support on managing complex institutional change.

While DFID has made a considerable investment in the analysis of country and sector contexts to inform programme design, it needs to focus more on capturing learning during programme delivery, to support effective adaptive management.

Executive summary

How countries are governed has a major influence on their progress on poverty reduction and development. More stable, effective and accountable governance is both an objective in its own right and a means of achieving broader development goals, including in areas such as economic development, health, education and climate change. In the UK aid programme, governance is both an area for dedicated programming and a cross-cutting theme.

In this performance review, we explore the relevance and effectiveness of DFID’s support to strengthening governance in two countries – Nepal and Uganda – since 2009. We look at how this work has supported DFID’s wider development objectives, including stability and economic development. We also consider how well DFID is using learning to adapt its approach to governance. We chose to focus the review on two countries in order to allow for more in-depth exploration of cumulative results over several cycles of programming. Our methodology included desk reviews of 36 governance programmes and samples of other sector and centrally managed programmes, a survey of DFID’s governance advisers and visits to both Nepal and Uganda.

How relevant has DFID’s governance work been to individual country contexts and the priorities of the UK aid programme?

DFID’s governance work in Nepal and Uganda is well aligned with the UK government’s priorities of increasing stability and promoting economic development. In both countries DFID’s programming has included a strong focus on post-conflict development, working to improve services and livelihoods in conflict-affected areas by making local governments more responsive to poor communities. DFID has provided long-standing support for national civil society, to build its capacity to hold governments to account and give voice to marginalised groups. It has also supported public financial management to tackle corruption and enable the partner countries to make better use of their own resources.

Over the review period, DFID has significantly increased its investment in studies and analysis of the governance situation in each country, producing a substantial amount of high-quality work on the national political context and the performance of public institutions. This analysis has helped it to tailor DFID’s programme designs and delivery arrangements to the local context.

While we found individual programmes to be relevant to the UK’s objectives and the country contexts, DFID lacks clear strategies at the country level to guide its governance portfolios. It is sometimes difficult to follow the logic underpinning the spread and balance of DFID’s investments across different areas of governance, or to identify how it has chosen to respond to challenges identified in its analytical work. For example, there has been a progressive shrinking of democratic space in Uganda, as the government has become more resistant to efforts by civil society to hold it to account, posing wider risks for the quality of governance. We found that DFID’s analytical work had not translated into a clear strategy for responding. A lack of emphasis on scenario planning and risk management has also made the country offices at times slow to respond to external events. In Uganda, for example, DFID has taken some time to articulate an alternative strategy for policy engagement with government, following the UK’s termination of general budget support. In Nepal, DFID has recently increased its focus on scenario planning linked to the country’s federal transition process, although a number of stakeholders suggest that this work has come quite late.

We also found some inconsistencies in decision-making on individual programmes. DFID has a strong record of supporting civil society in both countries, with a particular focus on gender equality and marginalised groups. However, following a decision by DFID to discontinue support to a long-running and largely successful multi-donor civil society fund in Uganda, we heard concern from a range of external partners and stakeholders that DFID was sending mixed signals as to whether it was still willing to work jointly with others in directly challenging the state on its accountability and responsiveness to citizens.

We have awarded DFID a green-amber score for the relevance of its governance work in the two countries, reflecting the considerable investment that has gone into improving its analysis, resulting in programming that is well tailored to the country contexts. The lack of a more explicit strategy for the governance portfolios in each country, however, left the rationale for the choice of programmes that were undertaken unclear.

How well has DFID delivered its governance work, and what difference has it made?

Measuring results in a robust manner is challenging for governance programmes and there has been underinvestment in evaluation, making it difficult to judge the overall effectiveness of the two portfolios. According to DFID’s own reporting, most completed programmes in our sample achieved or exceeded their expectations at outcome level and we were able to identify positive results in a range of areas. However, while individual programmes are performing well, it is difficult to assess their contribution at the portfolio level to promoting DFID’s larger strategic objectives.

At the programme level, we saw a range of positive results in Nepal and Uganda, including for example:

  • DFID’s work with local government and communities in conflict-affected areas has helped to improve access to public services in some of the poorest and most marginalised areas of both countries. In principle, this should help to alleviate this key cause of past conflict in both countries, although this has not been directly assessed.
  • In both countries we saw evidence of changes in social norms, including attitudes towards gender-based violence and maternal health. In Nepal, we also heard that some communities had abolished the practice of ‘untouchability’. By changing social norms and practices, DFID’s programming is helping to reduce marginalisation.
  • DFID’s support to civil society groups has helped to promote greater public engagement in policy debates and in the democratic processes.
  • Long-running investments in public financial management have helped both governments make more effective use of their budgets.
  • Some innovative work on economic governance in Nepal has led to promising early results by removing barriers to foreign investment.

We found that DFID’s work was most effective where it sustained its engagement with particular institutions or policy issues over several programme cycles. This allowed DFID to build up relationships of trust with national counterparts and to generate cumulative results over time. DFID’s management systems do not, however, facilitate this way of working and we found that DFID was not always well sighted on the broader impacts of its governance work.

We saw some strong examples, particularly in Nepal, of a shift towards more politically informed, adaptive programming, reflecting current thinking on how to support complex policy and institutional change. Within these adaptive programmes, DFID has demonstrated that small and well-targeted investments can produce catalytic results by supporting national reformers and campaigners. However, while DFID’s programme management systems permit flexible programming, we found that this way of working remains the exception in practice. DFID’s approach to value for money, for example, could usefully include more attention to monitoring the quality of engagement and the balance of risk and return across its governance portfolios.

As well as its designated governance portfolio, DFID pursues governance objectives within its broader programming, such as on health, education and climate change. The design of these programmes is usually informed by political economy and institutional analysis, with technical input from governance advisers. However, advisers in other sectors would welcome more input from governance advisers during programme implementation, especially on managing complex institutional change. We noted some positive innovations in Nepal designed to promote a more consistent approach to local governance issues across programmes. This included a pilot project on coherence and different ways of working at the field level.

We have awarded a green-amber score for effectiveness, reflecting an overall pattern of successful delivery and some significant achievements. We remain concerned, however, that DFID’s evidence base on results is not strong enough and that DFID could do more to track the cumulative impacts of its programming.

How well has DFID adapted its approach to governance in response to lessons learned?

DFID’s programme designs are generally informed by research and experience from past programming. However, there is less evidence that they adapt during implementation in response to lessons learned – although some of the recent programmes are more flexible.

While DFID undertakes significant amounts of political economy analysis, we are concerned that it has come to be seen as a product commissioned during design, rather than an ongoing process of engagement with national stakeholders, used to guide implementation. In both countries, external stakeholders are concerned that DFID has narrowed its range of interlocutors, favouring civil servants over politicians, and is at risk of being overly influenced by a narrow range of views and interests. In Nepal, DFID was much better at staying engaged with the politics through a number of different mechanisms, but concern over the range of its interlocutors was still valid.

While DFID governance advisers are well regarded, the skills profile they self-report is relatively narrow and the cadre would benefit from more practical experience with government. DFID reports that governance cadre has the largest number of advisory staff appointed in country within DFID. While there has been progress in the development of staff appointed in country in recent years, with a small but growing number of senior advisers, DFID could do more to make best use of the knowledge of these staff.

DFID lacks sufficient mechanisms for sharing experience between programmes and implementing partners. While lessons are documented in project completion reports, they are not compiled or shared across the department. However, DFID has some useful communities of practice in the governance area and staff share practical experience with each other through various informal mechanisms. DFID has underinvested in evaluation, particularly impact evaluation and portfolio-level evaluation, and we found that programme evaluations that were undertaken did not clearly influence decision-making. It has a substantial central research portfolio on governance, but no mechanism for tracking the uptake of this research in its programming. We have awarded DFID an amber-red score for learning, owing to concerns that it is not making the best use of its knowledge and experience.

Conclusions and recommendations

We find that DFID’s governance programming in Nepal and Uganda is relevant to DFID’s objectives and the country context, and is achieving some worthwhile results. We saw examples of effective programming in both countries, but we have some concerns that DFID’s programme management systems are in tension with good practice as they do not facilitate sustained engagement with particular issues or institutions or flexible and adaptive approaches. We would also encourage DFID to make better use of the capacities of its governance advisers and to do more to generate and make use of learning from its programming.

We offer the following recommendations to help DFID strengthen its approach to governance in Nepal and Uganda and, where relevant, across the wider department.

Recommendation 1

DFID should articulate in more detail its strategic approach to governance at country level, and make more use of scenario planning and risk management tools to support portfolio management in volatile contexts. Its strategies should consider and articulate the desired balance of risk and return across its governance portfolios.

Recommendation 2

DFID should identify areas where sustained engagement is likely to be required to generate the desired results, and invest in long-term relationships with key counterparts, while maintaining the flexibility to scale individual activities up and down as appropriate.

Recommendation 3

DFID should maximise value from its governance cadre by increasing the amount of time governance advisers spend on technical inputs (including into other sector programmes) and external influencing and engagement, rather than programme management and administration.

Recommendation 4

DFID should continue to develop the capacity of its governance cadre by: (i) making better use of the knowledge and experience of staff appointed in country; (ii) posting home civil service staff for longer periods in country; (iii) increasing diversity in terms of experience, backgrounds and local knowledge; (iv) placing more weight on practical delivery experience in its recruitment.

Recommendation 5

DFID should use evaluation more to test the validity of key propositions underlying its governance programmes and portfolios (for example that support for marginalised groups reduces conflicts). It should increase its investment in how learning takes place within programmes, and make sure the lessons are used to inform the management of country portfolios and programmes.

Introduction

Governance programming is designed to improve the quality of institutions and governance processes in developing countries. Better governance is both an objective in its own right, linked to values such as human rights and democracy, and a means of achieving other development goals. The UK recognises that more inclusive societies, with open, democratic and accountable institutions and peaceful political processes, are better able to sustain economic development in the longer term.


“‘Governance’ is about the use of power and authority and how a country manages its affairs… It concerns the way people mediate their differences, make decisions, and enact policies that affect public life and social and economic development.”

Governance, Development and Democratic Politics, DFID, 2007.

Government and Civil Society is a category of programming recorded separately in the aid statistics, constituting £537 million of DFID’s bilateral expenditure in 2016. It includes efforts to strengthen particular public functions and institutions, such as budgeting and public expenditure management, the electoral process, media, courts, police, local government and civil society. It aims to make governments more responsive and accountable, build organisational capacity to promote development, and to strengthen processes for the peaceful resolution of disputes. DFID has a cadre of advisers specialised in governance. As well as managing dedicated governance programmes, they advise on governance issues arising across other DFID programmes.

The link between improvements in governance and better development outcomes is well documented in the development literature. Economic growth is more likely to be pro-poor where governments are able to provide supportive policies and to invest in infrastructure and public services. Educational and health outcomes are improved if service providers are both competent and accountable to citizens.

This performance review assesses the relevance and effectiveness of DFID’s governance programming in Nepal and Uganda, and how well it adapts to learning. It explores what progress has been made towards improved governance in the two countries over a period of nearly a decade and how this has contributed to DFID’s higher-level objectives. It considers how well DFID has adapted its governance approach as a result of lessons learned. We review DFID’s dedicated governance programmes in the two countries (with total expenditure of £127 million and £94 million over the review period in Nepal and Uganda respectively), and how governance objectives have been pursued within sectoral programmes. Our review questions are set out in Table 1.

We chose to focus this review on just two countries in order to allow for in-depth exploration of the quality of assistance and results in a broad area of programming. We also explored results over a longer period (2009 to 2017) than is usual in an ICAI review, allowing us to cover several cycles of programming, in recognition that governance results may take time to emerge. While the two country studies display unique characteristics and are not representative of DFID’s governance portfolio as a whole, we expect that many of the issues explored here will have wider significance.

Our review is limited to DFID, which spends 79% of total UK governance aid globally (most of the rest is spent by the Foreign Office and the Conflict, Stability and Security Fund).

We focused primarily on DFID’s in-country work. We considered its global approach and its centrally managed activities (including its research portfolio) only where relevant to assessing country-level activities.

Table 1: Our review questions

Methodology

This review focuses on two countries in order to allow a deeper assessment of how well DFID’s governance programming is adapted to country contexts and to examine results over a long period. Nepal and Uganda were chosen as countries with significant governance expenditure and a wide range of programming across different themes and issues. They are not, however, intended to be representative of the global portfolio.

Our methodology for this review consisted of five main components that worked together to provide an evidence base from which we could triangulate findings:

  • Literature reviews: we undertook one general-level literature review and individual literature reviews for both Nepal and Uganda. Our general literature review considers DFID’s approach to governance within the current context and thinking around governance in development. Our country-level literature reviews introduce the governance context, provide an overview of the donor landscape and summarise how the UK has approached governance at the country level.
  • UK interviews and survey: we carried out interviews with DFID staff at corporate level, undertook a survey of governance advisers and held a focus group session at DFID’s governance professional development conference in December 2017.
  • Country portfolio analysis: we undertook desk reviews of DFID’s governance programmes and a sample of other sector programmes in Nepal and Uganda, collecting evidence of relevance, effectiveness and learning from DFID’s own project documentation.
  • Centrally managed programme analysis: we undertook desk reviews of a sample of centrally managed programmes that include governance work in Nepal and/or Uganda. Here we sought evidence of country-level engagement in the design and implementation of centrally managed programmes.
  • Country visits: we carried out two-week country visits to both Nepal and Uganda, including field visits to mid-western and central Nepal and northern Uganda, to see how current governance programmes are being delivered.

Both our approach and this report have been independentl peer-reviewed.

Figure 1: Summary of data collected

Background

Governance and development

‘Governance’ refers to the processes by which societies and groups of people organise themselves to make decisions and manage their affairs. It lies at the heart of the development process. Making policies, mobilising and investing public resources, organising services, ensuring public safety, regulating markets and empowering communities are all governance challenges that are integral to promoting development. In its 2017 World Development Report, Governance and the Law, the World Bank describes governance arrangements as “underlying determinants” of growth and poverty reduction, concluding: “Bad governance is the enemy of progress, and the number one cause of persistent poverty.”

Over the period covered by this review, thinking on the role of governance in development has evolved significantly, influenced by research and analysis at both the empirical and the theoretical level. There has been a shift away from approaching governance assistance as a purely technical process of transferring policy and institutional solutions from wealthier to poorer countries. Increasingly, development practitioners have sought to understand how politics shapes governance outcomes and to engage with local actors to find problem-focused and context-specific approaches. Governance assistance has also developed from focusing primarily on building the capacity of state institutions to encouraging them to be more responsive and accountable to citizens.

The World Bank’s 2017 World Development Report reflects this changed thinking, noting the centrality of political power to the development process and rejecting ‘one-size-fits-all’ solutions to institutional development. ‘Thinking and working politically’ has emerged as a guiding principle for governance programming, supported by DFID and like-minded donors, and emphasising good political insight, responsiveness to the domestic environment and flexibility in programme design and implementation. Donors put more emphasis on studying how incentives drive the behaviour of leaders and organisations (‘political economy analysis’). Practitioners have experimented with new ways of helping national actors to craft locally appropriate solutions. However, the change in approach has not necessarily proved easy for donors to implement.

DFID’s approach to governance

Governance work has long underpinned DFID’s approach to development assistance. A 2006 White Paper – Making Governance Work for the Poor – made the case that governance is central to poverty reduction. It introduced the ’capability, accountability and responsiveness’ framework which includes the key elements that national governance systems need in order to be pro-development.


“Whether states are effective or not – whether they are capable of helping business grow, and of delivering services to their citizens, and are accountable and responsive to them – is the single most important factor that determines whether or not successful development takes place. Good governance requires: capability – the extent to which government has the money, people, will and legitimacy to get things done; responsiveness – the degree to which government listens to what people want and acts on it; and accountability – the process by which people are able to hold government to account.”

Making Governance Work for the Poor, DFID, 2006

In 2010, the UK government adopted the idea of a ‘golden thread’ of institutions and governance principles to underpin the aid programme. The concept is that, to escape poverty, countries need “a golden thread that starts with the absence of war and the presence of good governance, property rights and the rule of law, effective public services and strong institutions, free trade, and open markets.” This idea of a direct causal link between institutions and economic development prompted DFID to explore which institutions are most likely to drive better development outcomes, with a particular focus on economic and political inclusion.

Although DFID has not updated its definition of governance since 2007, in recent years governance work in DFID has become more clearly associated with promoting stability and economic development. The 2015 national aid strategy “Tackling Global Challenges in the National Interest” has four strategic objectives, one of which is strengthening global peace, security and governance. Under this objective, the UK pledges to tackle the causes of instability, insecurity and conflict and to combat crime and corruption. The strategy linked spending of the UK’s official development assistance commitments to the recommendations of the 2015 National Security Strategy and Strategic Defence and Security Review. This committed the UK to working with partners to “reduce conflict, and to promote stability, good governance and human rights”.

DFID continues to identify security and governance as “the building blocks of sustainable, successful societies”. In 2015, the national aid strategy committed the UK to “continue to promote the golden thread of democracy, the rule of law, property rights, a free media and open, accountable institutions”. These aid strategy commitments were reiterated in the 2016 Bilateral Development Review.


“DFID’s governance advisers support improvements in good governance and politics that enable development, stability, and inclusive growth and improved opportunities for the poorest and marginalised groups. In doing so, they play an important role in helping DFID to tackle the underlying causes of poverty and instability, which is central to both the UK Aid Strategy and the Sustainable Development Goals.”

Technical Competency Framework Governance Cadre, DFID, 2016

The governance cadre is one of DFID’s largest professional groups, with around 170 advisers. Of these, 120 are working in governance roles and 50 are engaged in other areas but remain accredited to the governance cadre. DFID’s technical competency framework identifies the skills required of governance advisers, including familiarity with some or all of the following specialities:

  • security, justice and human rights
  • accountability and inclusive politics
  • public sector governance and service delivery
  • inclusive growth and economic development
  • public financial management and domestic resource mobilisation
  • anti-corruption.

Figure 2: DFID’s global governance spend

The OECD’s aid statistics record donor expenditure on governance in a number of different areas. We have used these categories (see Figure 2) to map DFID’s expenditure on governance. DFID’s global spend on governance increased steadily over most of the review period, rising from £492 million in 2009 to £628 million in 2015. DFID’s governance spending dropped to £537 million in 2016 (the latest year for which DFID has complete data), down 14.5% on 2015. The most significant increases in spending over the period were on anti-corruption, legislatures and political parties, and domestic resource mobilisation. The most significant decreases in spending over the period were on public financial management, democratic participation and civil society, elections and human rights.

In 2011, DFID undertook a portfolio review of its governance programming between 2004 and 2009 to identify the impact of its governance spending at country and global levels. This review identified key lessons for DFID’s governance programming, all of which remain true today (see Box 4).

Figures 3 and 4 set out some of the key characteristics of the two country contexts, including timelines of key political events and the countries’ ranking on some standard governance indices, and introduce DFID’s governance portfolios in each country.

Figure 3: DFID’s governance programme in Nepal

Figure 4: DFID’s governance programme in Uganda

Findings

This section presents the findings of our review, covering the relevance of DFID’s approach to governance in Nepal and Uganda, the effectiveness of delivery and the quality of associated learning.

Relevance: How relevant has DFID’s governance work been to individual country contexts and the priorities of the UK aid programme?

We begin by exploring the relevance of DFID’s governance work to the needs of Nepal and Uganda and how well it aligns to the UK’s overall development objectives in each country.

DFID’s governance programming in Nepal and Uganda is well aligned behind the UK’s global objectives of promoting stability and economic development

The UK government prioritises improving governance as a means of increasing stability and promoting economic development. In both Nepal and Uganda, we saw that clear links could be traced between DFID’s governance programmes and those two objectives, particularly in the area of post-conflict development. We found that the objectives of individual governance programmes were expressed in terms of the priorities set out in the UK aid strategy and, following their introduction in 2015, in National Security Council strategies for both countries.

According to UK government stakeholders, supporting Uganda’s stability within a volatile region is a high priority. The country suffered a long-lasting civil war in the north, and much of DFID’s governance work has been focused on restoring stability in conflict-affected areas. This has included work to strengthen the responsiveness and accountability of local governments and to improve the livelihoods of the most vulnerable. The Post-Conflict Development programme (2011-2016; £90.1 million), for example, included a focus on youth training and employment in disadvantaged areas of northern Uganda. It also invested in small-scale development projects, such as local infrastructure, using techniques that encouraged communities to come together to identify priorities and oversee the work. This type of programming is important for underpinning stability.

In 2017, DFID redirected part of its Uganda governance portfolio in response to the arrival of 1.4 million refugees from South Sudan. It chose to concentrate programming in the areas where refugees were located, to support both refugees and host communities in accessing essential services. This work is expected to help Uganda manage the high volume of refugees and reduce the risk of conflict.

In Nepal, stability was also identified as a key priority for the UK, following the end of a long-running civil conflict in 2006. The UK has strong historical ties to Nepal, not least through the British Army (the UK continues to pay army pensions and other benefits to Gurkhas and their families). DFID’s efforts to promote stability included building state capacity to deliver basic services, in particular by strengthening budgeting and financial management. DFID’s analysis showed that the political and economic exclusion of social and ethnic groups had been a major cause of the conflict. DFID has therefore worked to support a new constitutional settlement, based on a federal system. It engaged in a sustained effort to strengthen local government and make it more representative and accountable. As in Uganda, DFID redirected (and in this case provided additional) funds and programming in response to a major crisis: the 2015 earthquake, which resulted in an estimated 8,900 lives lost and economic losses of up to $10 billion (£7 billion).

In both countries, DFID’s governance programming has also been working to promote the conditions for economic development. There has been a strong focus on raising domestic revenue and spending more effectively through the budget. We found the approach to promoting economic governance to be stronger in Nepal than in Uganda. DFID Nepal had recently increased its investment, replacing the Centre for Inclusive Growth programme (2010-2014; £15.5 million) with the Accelerating Investment and Infrastructure in Nepal programme (2015-2020; £35.2 million). Its programmes are working to remove legal and investment barriers to economic growth, especially in the hydropower sector. DFID’s support to the Investment Board of Nepal has focused on addressing key institutional failures in government that have prevented foreign direct investment. In Uganda, we found the rationale for DFID’s economic governance work to be less convincing, particularly in the context of an underlying trend of deterioration in many key aspects of governance (see Figure 4).

DFID has increased its investment in research and analysis to inform programming

A clear change in DFID’s governance work over the past decade is that it now takes the national context, and specifically the political process, more into account in the design of its programmes. Across the department, more than 500 people have been trained on political economy analysis since 2009. In Nepal and Uganda, these have included both governance advisers and other specialists.

DFID has increased its investment in analysis and diagnostic studies over time, to provide, in particular, better understanding of the constraints to and opportunities for development; it is now routine for programme designs to refer to studies and contextual analysis. Of the 26 programmes in our sample designed before 2011, we assessed that 15 were based on an appropriate level of situation analysis while four had partial analysis. Since 2011, all programmes in our sample have done so. This major investment in analytical work to inform programme design in the face of dynamic country contexts has helped ensure that new DFID programming is relevant at the time of design. Challenges in maintaining relevance during implementation are discussed below in the section on learning.

DFID Nepal had generated a particularly large number of contextual studies, including political economy and drivers of conflict analyses, since 2009, with at least 45 substantial pieces of work. Specific studies included analysis of the civil war, and of constitutional changes and the move to a federal system of government. There had been studies addressing programme challenges (such as new local governance arrangements), sectoral challenges (for instance in education, health and forestry) and cross-cutting issues (such as anti-corruption). A useful example of a study demonstrating good political awareness is the 2011 political economy analysis of local governance in Nepal. Uganda had also carried out good diagnostic work, including the Uganda Fiduciary Risk Assessments undertaken in 2008, 2011 and 2017, consideration of the impact of an expansion of oil revenues, two horizon-scanning exercises to better understand the politics of development in Uganda and a regional political economy analysis of the Karamoja region in the north. These studies support DFID’s understanding of the context in which it is working and help country offices to make relevant programme decisions.

Such analysis has become hardwired into DFID’s business case processes for individual programmes and also (to a greater or lesser degree) into elements of DFID and the UK government’s country-level planning. At the country level, we saw that elements of political economy analysis have increasingly been included in DFID’s Country Poverty Reduction Diagnostic, Inclusive Growth Diagnostic and Country Governance Analysis, and the interdepartmental Joint Analysis of Conflict and Stability. DFID intends also to include political economy analysis as one strand of its newly introduced Country Development Diagnostics, which provide a multidisciplinary approach to country assessments. This analysis helps to inform DFID’s decisions at the programme level, and increasingly at country level.

One interesting piece of feedback from our interviews with national stakeholders in both countries was that DFID tended to outsource most of its political economy analysis to non-DFID staff. In Uganda this analysis had largely been provided by a limited number of experts. While nobody questioned that these individuals were well qualified, they noted that the lack of diversity of sources – particularly among the expatriate experts – might limit its usefulness in understanding the national context and misses opportunities to challenge established views. This is a challenge that DFID could usefully explore further.

Individual governance programmes were plausible responses to governance challenges in both countries

We were able to follow the logic of developmentally significant challenges identified in DFID’s analysis through into most of its programming choices at the country level. The problem and context summaries from country-level plans and additional studies were used to justify why, and in some cases how, programmes were being undertaken. This has helped to maintain the relevance of DFID’s governance programming in both countries.

In both countries, analysis of drivers of conflict and instability had informed DFID’s efforts to strengthen citizens’ opportunities to voice their opinions on local service delivery and to increase government accountability for the same. This work had focused particularly on marginalised groups and hard-to-reach areas. In Uganda, we saw an increased focus on decentralisation of decision-making on taxation and spending and on anti-corruption efforts over the period. This aligned with the government’s own initiatives and with growing concerns around corruption, especially after the 2012 budget support scandal which involved allegations of aid money being embezzled. In Nepal, to support the new constitutional settlement, DFID had increased its focus on federalism, for example through a workshop on federalism to help DFID and partners think through next steps for programming. DFID is now working with partners to design a new local governance programme to fit within this context and that builds on its earlier work in this area: for example Community Support Programmes Phase I (2005-2010: £17.8 million) and II (2010-2014: £19.7 million) and the Nepal Local Governance Support Programme (2013-2019: £68.1 million).

DFID also demonstrated an ability to tailor its implementation approach to the country context. DFID Nepal’s Centre for Inclusive Growth programme, for example, used contextual analysis to identify problems and work in a flexible way to remove barriers or build constituencies for change. As compared to designing projects around pre-defined activities and outputs, this allowed for greater adaptability and creativity in a dynamic context.

DFID lacks clear strategic direction to guide its approach to governance in each country

While the governance programmes in both countries were broadly aligned with high-level UK objectives, for the period of review there was no comprehensive explanation of the strategic choices informing the selection of programmes that make up DFID’s governance portfolio.

The only formal statement of DFID’s governance objectives is a few sentences in the country operational plans. These have improved more recently in the country business plans (see Table 2). The statements refer to some high-level objectives, such as promoting stability, reducing corruption and working with civil society. There are a few governance-related output targets (such as the number of independent audits or the existence of anti-corruption strategies), but these relate to a small subset of DFID’s governance programming in each country. We find these statements to be inadequate. We are not able to see a sufficiently clear link between DFID’s situation analysis, its choice of strategic objectives and a more detailed country-level strategic approach that guides portfolio-wide decision-making and programming. In Nepal, the current ambassador introduced in 2016 ‘six big things’ (initially five), which set out common mission objectives for the UK government in the country. These have helped to establish issue-based collaboration around joint UK government priorities but do not set the strategic direction for DFID’s governance portfolio.

Table 2: Governance objectives in DFID’s country plans

Governance work requires making difficult strategic choices, particularly where DFID’s objectives are not fully aligned with those of the national government. For example, in the face of some negative trends in Ugandan governance (see Figure 4), political economy analysis commissioned by DFID outlines three broad strategic choices:

  • to withdraw support for improved governance, in order to signal disapproval
  • to limit governance work to specific areas where the Ugandan government and DFID interests remain aligned (such as aspects of public financial management)
  • to support constituencies for change within Ugandan society, to help counteract the narrowing of the political space and promote longer-term change.

We found that DFID Uganda had, to some extent, pursued all of these options at different points during the review period. Furthermore, the risks of political instability increase as the 73-year-old president approaches the end of his term, calling for careful management by Uganda’s international partners.

While DFID’s political economy analysis does a good job of framing the strategic dilemmas, we find it difficult to reconstruct the decisions that DFID has made and how these have informed the balance of its governance programming. It is not clear to us to what role DFID Uganda expects to play in supporting improved governance in the country, and how it proposes to manage the risks that accompany an eventual political transition. A more explicit strategy for the governance portfolio would make it easier (both for DFID and ourselves) to assess whether the portfolio is coherent and well balanced across priority areas for governance reform.

In Nepal, the strategic logic of the governance portfolio is easier to follow. There is a strong strand of activity around promoting greater inclusion of previously marginalised groups within the political settlement, to alleviate marginalisation as a driver of conflict. There is also a good focus on supporting implementation of the new federal constitution, by helping to strengthen the quality of the political process and promote stronger engagement between communities and elected local authorities. We saw signs of DFID reflecting strategically on the implications of the new federal system for its governance work, including thinking through different scenarios for the federal transition. Nonetheless, given the continued uncertainties around how such political processes may play out, we similarly feel that DFID Nepal would benefit from a more detailed statement of its strategic approach to governance. This should include associated work around scenario planning and risk management more broadly.

One consequence of the lack of a more detailed statement of strategic approach is what happens when DFID has to react to external events that affect its programming. A key event in Uganda during the period under review was a corruption scandal in the Office of the Prime Minister that came to light in 2012, causing DFID to terminate its general budget support (funds given to government for spending through the budget on national development priorities). There was a sound rationale for this decision: in addition to the fiduciary risks, DFID’s reporting suggests that government and donor priorities had diverged and that the space for effective dialogue had declined. Phasing out general budget support later became DFID policy globally. However, the loss of budget support as a platform for dialogue with the Ugandan government amounted to a significant change in direction for the development partnership. We find that DFID was slow to articulate an alternative strategy for policy engagement with government.

A second consequence is inconsistencies in decision-making around certain types of programming and in the signals given to external stakeholders about DFID’s priorities. Through most of the review period, DFID has supported national civil society in both countries to represent marginalised groups within society and to engage in policy advocacy. In both countries, it has invested substantial sums in multi-donor programmes that help build the capacity of national civil society organisations to support citizens to assert their rights and engage in the democratic process. Some of this work is inevitably contentious: we saw examples of DFID supporting advocacy on behalf of gay rights in Uganda and on the inclusion of marginalised groups in Nepal. DFID has also had a strong history of engagement on gender equality, which is a departmental priority. We saw examples of work on violence against women and girls in northern Uganda, local voice and accountability work that prioritised marginalised women in Nepal, and support for women’s organisations in both countries. (Over the review period, DFID’s global spending on ‘women’s equality organisations and institutions’ increased from £5.9 million in 2009 to £36.7 million in 2016 – see Figure 5). National civil society organisations are also partners for DFID in other areas of governance work, such as strengthening local government and voter education.

However, we heard from stakeholders that, in recent years, DFID has become less willing to engage in areas of governance programming that directly challenge the state on its accountability and responsiveness to citizens. In Uganda, in particular, DFID decided in 2017 to discontinue its funding to the multi-donor Democratic Governance Facility, after a £28.3 million investment over six years. Yet a 2016 independent evaluation had concluded that the facility was successful and “highly relevant to the national and local political context”. It had achieved “measurable and lasting improvements in accountable, responsive, rights-based governance through effective citizen participation coupled with stronger institutions”. Internal DFID reporting was also favourable, noting clear achievements (particularly at local level) and that the management value for money exceeded expectations (see Box 7). DFID Uganda’s own political economy analysis has indicated that the quality of governance in the country will not change unless there is a change in the accountability relationship between citizens and the state. DFID has, however, continued to fund limited accountability programming through the Governance, Accountability, Participation and Performance programme (2014-2019; £12 million), although this is focused on seeking to improve accountability of local government for public service delivery.

The key explanation offered in country by DFID Uganda and by other sources for DFID’s withdrawal from the Democratic Governance Facility was that the management costs of the programme had become too high. Isolated examples of fraud had also been identified, although controls had been increased to prevent a recurrence. Both issues were problems that were able to be managed. Other donors expressed regret at DFID’s decision to withdraw from the Democratic Governance Facility, querying whether it reflected a change in DFID’s priorities or a decreased willingness to work jointly with others. Country office staff in both countries reflected that it had become harder to make the case within the department for this kind of work, compared to programming that delivers more direct and tangible results. We cannot tell from this review whether this example of withdrawal reflects a wider trend across DFID. However, we note that DFID’s global spending on democratic participation and civil society increased through the first half of the review period to a peak of £124 million in 2013, before falling back to £64 million in 2016.

There is some evidence to suggest that the lack of clear strategic direction and statements of priorities are leading to inconsistencies in decision-making. This is sending mixed signals to external stakeholders as to DFID’s intentions and priorities, which potentially affects the credibility of its governance programming.

Figure 5: DFID’s global spend on democratic participation and civil society

Conclusions on the relevance of DFID’s governance programming in Nepal and Uganda

DFID’s governance programming in both countries broadly reflects UK strategic priorities, particularly on promoting stability and economic development. Both countries have used a range of approaches to help the countries recover from civil conflict and have adapted in response to recent crises (refugee arrivals in Uganda and the Nepal earthquake). There is a good range of economic governance programming in Nepal, but the approach to economic development is less clear in Uganda. There have been long-running investments in building core state capacity and strengthening the capacity of national civil society to represent marginalised groups.

We find that DFID has made a strong investment in analytical work, which has helped to maintain the relevance of its programming in dynamic environments. Its most recent programming is consistently supported by political economy analysis and both the objectives and the approaches of its programmes are adapted accordingly.

However, the lack of overarching strategic approaches for its governance portfolios in each country makes it difficult to follow how DFID has responded to key strategic challenges, such as the decline in democratic space in Uganda. A more detailed articulation of its strategic approach, encompassing planning tools such as scenario analysis and risk management, would promote greater clarity on long-term priorities and stronger strategic management in a dynamic context. Weaknesses in strategic planning were particularly apparent in Uganda, where DFID’s governance portfolio appeared to us to be more reactive to events and lacking consistency. This has led some of DFID’s key external partners and stakeholders to question its priorities

Overall, we award DFID a green-amber score for the relevance of its governance work because of the considerable investment it has made in improving its analysis, which has resulted in governance programming that is well tailored to the challenges and opportunities at country level. However, we find that DFID would benefit from a strong approach to translating its analysis into strategic decision-making.

Effectiveness: How well has DFID delivered its governance work, and what difference has it made?

In this section, we examine how effective DFID’s governance work has been in Nepal and Uganda. To do this we explore how well DFID has delivered its governance work in the two countries, and whether the results are contributing to the achievement of its objectives in each country.

The majority of completed programmes met or exceeded their expected programme objectives

DFID’s own reporting suggests that most completed programmes in our sample met (48%) or exceeded (22%) their expected programme-level objectives. This data has been captured from completed programmes as DFID’s project completion reports provide a final assessment against the extent to which a programme has achieved its objectives. Project completion report scores were not available or required for five of the 17 completed programmes in Uganda, and for two of the 12 completed programmes in Nepal (see Figure 6). The data suggests that Nepal performed better overall: 90% of governance programmes either “moderately exceeded” or “met” their expected programme-level objectives, compared to just over half in Uganda. However, we note that programme scores in Uganda improved significantly from 2014 onwards.

Figure 6: Project completion report scores of completed governance programmes in Nepal and Uganda (2009 to present)

To gain a deeper understanding of programme effectiveness, we looked at achievement across clusters of related programmes in each country. This was based on a combination of our own desk review of narrative reports from the programmes and feedback from our interviews with government counterparts and other stakeholders in Nepal and Uganda, including donors, civil society and beneficiaries. We found a good range of meaningful results.

In Nepal, programmes had performed well in the related areas of service delivery, local government and inclusion of marginalised groups. The Social Inclusion Action programme (2006-2012; £2.7 million) had helped to make government more sensitive to issues of gender equality and social inclusion. The results included a greater focus on gender and inclusion in the government’s national Poverty Monitoring and Analysis System and in the national health strategy, as well as revisions to the legal framework on gender-based violence. At the community level, women’s organisations also became more successful at accessing village development funds, using the resources for a range of useful activities such as scholarships for Dalits (members of the lowest caste).

At the local level, DFID Nepal has long-standing investments designed to help communities engage with local governments and service providers, focusing on some of the country’s poorest and most conflict-affected areas. For example, it supported the formation and training of Citizen Awareness Centres at village and community levels, which help to make marginalised groups aware of and able to assert their rights as citizens. In our interviews with local officials and participants, we heard that social norms were changing as a result of this work, leading to, for example, reductions in violence against women and in caste-based discrimination. In some villages, we heard that the process of bringing marginalised castes together to mobilise for social change had led communities to abandon the practice of ‘untouchability’. The Community Support Programme Phase II (2010-2014; £19.7 million) had also helped communities to access and make better use of development funds and to advocate for improved public services. The programme reported a 2.5% increase in access to primary education in its target areas. Sanitation coverage had also increased from 39% to 62% and 237 villages were declared free of open defecation – with consequent benefits to public health and women’s safety and security.

In Uganda, at the central government level, DFID’s long-standing support on public financial management has “supported huge investments in policies, systems and capacity across the [public financial management] cycle.” Public Expenditure and Financial Accountability (PEFA) assessments (a standard methodology for assessing progress on public financial management used by multiple donors) carried out in 2012 and 2016 show improvements in Uganda’s systems in a number of areas, including:

  • new public financial management legislation, adopted in 2015
  • budget preparation, including macroeconomic and revenue forecasting
  • the introduction of an integrated financial management system
  • improved information on budget execution for policy-makers
  • improved monitoring of the operations and finances of off-budget statutory bodies and state-owned enterprises.

Overall, the PEFA report finds that the Ugandan budget allocates resources strategically, and that government shows good overall fiscal discipline. While public financial management in Uganda is clearly improving, it is a long-term undertaking and DFID has shown commitment to remaining engaged in this process. Out of 71 areas assessed under the PEFA framework, 21 had improved since 2012, six had deteriorated and 44 demonstrated no change.

In Nepal, we found some useful emerging results from DFID’s work on economic governance. DFID’s Accelerating Infrastructure and Investment in Nepal is a recent programme that is based on a strong understanding of the domestic politics affecting economic growth, and uses a problem-solving approach (see paragraph 4.45) to clear obstacles to infrastructure investment. It supported the passage of a long-delayed Special Economic Zone law through parliament, allowing the development of a special commercial zone to attract and promote foreign trade and revive manufacturing exports. DFID reports that the new regulatory framework has already cleared the way for investments, including a recent $360 million investment into a cement factory and two large hydropower projects expected to get underway in 2018.

Not all the programmes in our sample were successful: in particular, just over a third of the Uganda programmes were rated as “moderately below expectation”. There were various reasons for the shortfalls, with no clear pattern. For example, the two last budget support operations experienced declining performance before they were cancelled. This was partly due to reduced space for effective policy dialogue between donors and government and also the shift in government focus away from social service delivery in favour of advancing economic infrastructure. One public financial management programme scored poorly on a technicality, because it changed its priorities with good reason, but without updating its logframe, while a community development programme delivered through a government ministry was held back by weaknesses in national procurement systems.

There have been limited external evaluations of DFID governance programmes in the two countries to verify reported outcomes or examine longer-term impact (see paragraph 4.84 below in the section on learning). The literature suggests that under-investment in evaluation in governance work is a wider problem, in part because there are methodological difficulties with identifying sustainable results in a rigorous way.

Overall, we find that both countries have generally delivered their expected governance objectives at programme level. Furthermore, in areas where there has been a concentration of investments over time – such as local government in marginalised areas and public financial management – there is some evidence that programmes have been mutually reinforcing, helping to build up momentum for change.

While individual programmes are performing well, we found it difficult to assess how far these results have contributed to the larger strategic objectives of the UK aid programme or the extent to which the portfolio adds up to more than the sum of its parts. The link between its activities and higher-level indicators (such as the World Governance Indicators – see Figures 3 and 4) is not explicitly made. This is partly because of the absence of country-level governance strategic direction as noted in the previous section on relevance. In addition, the strategic case for the governance portfolios rests upon key assumptions – for example, that helping marginalised communities interact with local government reduces conflict risk, that investing in civil society capacity encourages democratic participation, or that building public financial management capacity reduces corruption and increases expenditure on development. While these assumptions are plausible, they have not been specifically tested in each country context and, as noted in the following section on learning, DFID has not put in place adequate mechanisms to evaluate them.

Good results are achieved through sustained engagement and relationship-building, but these are not prioritised in DFID’s management systems

Some of DFID’s more successful governance programming in the two countries comes about through sustained engagement with particular institutional partners, building up cumulative results over more than one programme cycle.

In Nepal, we saw strong continuity across four cycles of community development and local governance programming since 2004. The programming has focused on some of the poorest and most vulnerable communities, helping to build a critical mass of citizens with the knowledge and confidence to work together to pursue their rights. There is good reason to believe that the results have been cumulative and mutually reinforcing across different activities and programme cycles. We heard from DFID’s implementing partners and community representatives that many of those elected in the recent local government elections had previously been supported under DFID’s community development and local governance programming.

While the strongest parts of DFID’s governance portfolios show both continuity of programming and investment in relationships, this is not done systematically. DFID’s results system does not track outcomes across programme cycles or reflect on how results build over time. This means that there is less incentive to retain historical knowledge and to prioritise continuity in programming.

When working with government, continuity of programming helps DFID to build up relationships of trust with counterparts. This facilitates access to decision-makers and gives DFID a better understanding of the interests and power dynamics at play. This investment in knowledge and relationships makes DFID better placed to respond to opportunities to influence policies and institutional development as they emerge from time to time. We found that DFID Uganda had built a strong relationship with the Office of the Auditor General, which had made it well placed to support both its capacity as an organisation and its status and independence across the Ugandan government. The 2016 PEFA assessment notes: “External audit is an area of significant strength. The Office of the Auditor General is active in carrying out financial and compliance audits. It adopts auditing standards to govern its work with audit plans and strong staff development programs.”

Where governance programmes build up good relationships, they provide DFID with a platform to engage with government and other national stakeholders on key development policies and reform issues. We heard from stakeholders in Nepal and Uganda that the policy competence and technical skills of DFID governance advisers were highly valued. In Nepal, for example, DFID governance advisers were playing an important role in the design process for the government’s new local governance programme. In our assessment, DFID’s engagement with national stakeholders in and around governance programmes is an important component of their effectiveness.

However, we heard from some DFID staff that, while the department values policy influence, it does not always build in the staff time required to develop and maintain necessary relationships, especially when such relationships may not be immediately linked to access and influence. There is evidence that the governance advisers now have less time to spend on external engagement. In our governance adviser survey, 85 out of 92 respondents perceived external influencing and engagement as either extremely valuable (63%) or valuable (29%) (see Figure 7). However, two thirds reported spending less than 20% of their time engaging with policy-makers and influencers (see Figure 8). The main reason appears to be an increase in the programme management burden. In particular, the introduction of the Senior Responsible Owner role in 2014 has led to an increase in ‘compliance tasks’ – that is, ensuring that internal procedures are followed. However, this might mean that DFID advisers are spending too much time on managing risks associated with compliance rather than those associated with effectiveness (such as maintaining relationships).

Figure 7: Governance advisers’ perceptions of the value-add of work activities

We surveyed governance advisers to elicit feedback on their perceptions of the value-add across a number of different work activities they are engaged in.

Figure 8: Governance adviser proportion of work time spent with colleagues and stakeholders

We surveyed governance advisers to gain an understanding of how much time they spend with colleagues and various relevant stakeholders.

In both countries, we heard feedback from external stakeholders (including government counterparts, other donors and civil society representatives) that DFID governance advisers had become less visible in recent years in policy dialogue and coordination processes. In Nepal, we also heard from a number of stakeholders that DFID had narrowed its range of interlocutors in recent years, leading to a shrinking of its influence. This raises a question as to whether DFID is currently deploying its advisory staff in a way that maximises results.

There are good examples of flexible and adaptive programming, but this is not yet the norm

There is now widespread recognition in the development literature that solutions to governance challenges should be identified locally, rather than based on common international practice, and that programmes should adopt incremental and adaptive approaches rather than pursue pre-defined activities and targets. At the central level, DFID has adapted its programme management systems to allow more space for adaptive programming and produced internal guidance on the subject. The department’s Smart Rules now give staff more discretion to use their own judgment in adapting to local context.

In both countries, DFID has demonstrated the ability to be flexible and adaptive in its governance programming, but is not consistently doing so. We saw strong examples of DFID testing new approaches to adaptive programming and results management in Nepal. This included the Economic Policy Incubator (2016-2020; £6.1 million), a component of DFID’s Accelerating Infrastructure and Investment in Nepal programme. The Economic Policy Incubator works with a ‘live’ theory of change that is reviewed and updated quarterly. Logframe targets are agreed annually, rather than for the
duration of the programme, and out of a list of ten targets, traction on any three is accepted. This allows the programme the flexibility to respond to opportunities as and when they arise instead of working to meet priorities identified at the beginning of the programme that may no longer be as relevant.

DFID Nepal’s Enabling State Programme (2001-2014; £34.4 million) is a good example of how adaptive programming can lead to greater effectiveness. Taking an adaptive approach was a deliberate choice, right from the early years of the programme. This was highly unusual for DFID at the time. The programme was modestly funded over 13 years and had considerable flexibility over spending decisions for amounts up to £25,000, enabling it to respond quickly to changing circumstances and new opportunities. The project completion report for this programme states “piloting and incubating new ways of working has been key to the Enabling State Programme; in a number of cases, successful Enabling State Programme pilot projects have provided the basis for larger projects and programmes”. The Enabling State Programme’s work with the Ministry of Finance to consolidate multiple government bank accounts into a single treasury account in three districts, for example, was later piloted in 38 districts and eventually replicated across 75 districts in partnership with the World Bank. This improves accountability and reduces opportunities for public funds to be misused. Enabling State Programme-funded research has also helped to influence national policies. One study on corruption was a key input into the public debate that preceded the adoption of more stringent anti-corruption legislation. The programme’s work with excluded groups is also credited with encouraging a wider shift in donor programming towards gender and social inclusion. A 2014 study found that the programme “took a long-term perspective, working through a team of politically astute, well-connected Nepali staff and providing very flexible funding to a wide range of groups  including many hitherto unreached by aid programmes”.

Between the two country offices, we found that the culture of DFID Nepal was more open to adaptive governance programming. The office has been a pioneer within DFID of this way of working. This may be because the dynamic country context has required high levels of experimentation. It also reflects the skills and interests of key staff over the period. Adaptive programming typically involves higher administrative and management costs, relative to expenditure, but with the potential of yielding better results. It involves taking risks, building networks and relationships with a wider range of stakeholders and rethinking standard programme management tools and approaches. While this is possible within DFID’s programme management system, it calls for a high level of initiative and effort on the part of staff. As a result, the good examples of adaptive programming that we identified remain exceptions to the usual pattern.

DFID’s value for money approach is not well suited to its governance programming

Measuring and managing value for money in governance programmes is challenging. For programmes that provide technical assistance or support for capacity-building, the major input is consulting services. In these cases, the number or cost of the outputs produced by the consultants (such as reports and training sessions) are less important as indicators of value for money than the quality of the engagement. There is also significant uncertainty within theories of change in governance work, for example whether such engagement and influence will result in the desired changes in governance, given complex policy-making environments. We find that the techniques and approaches that DFID uses to ensure value for money in its programmes (see our ICAI review on DFID’s Approach to Value for Money in Programme and Portfolio Management) are not well suited to this type of programming.

DFID’s annual reviews and project completion reports describe a range of measures taken to ensure economy and efficiency. They provide evidence that DFID is keeping a close watch on costs by tracking staffing levels (including the balance of international and national staff), consultancy fee rates and administrative costs (such as office rent and travel). In more recent programmes, there are examples of DFID monitoring indicators of efficiency, such as management overheads or the proportion of funds spent on frontline delivery. These indicators are sometimes benchmarked against comparable programmes. There are assessments of the quality of financial management, including procurement of goods and services. Scrutiny at this level has increased over time.

As discussed, most programmes are still designed around predefined activities and outputs, often linking them to the payment of suppliers through milestone payments. The risk is that this incentivises the delivery of activities that may not prove to be the most useful ones for achieving the desired results. We found examples of DFID reworking programme logframes to reflect changes in the context or experience gained during the inception phase, but the management burdens of doing so are significant.

In the programmes in our sample, reviewers were having difficulty bringing assessments of effectiveness (or cost-effectiveness) into the value for money equation in a credible way. Measuring results in governance programming (and attributing them with confidence to an aid programme) is difficult, partly because the results may not be apparent until well beyond the life of the programme. We also saw examples where results targets had not been defined or baselines established. While annual reviews comment on progress towards outcomes and whether the theory of change remains valid, there is often limited evidence available to support that assessment. Thus, while DFID produces narrative assessments of value for money, the confidence in the assessment is not necessarily high enough to support decisions about the management of the programme.

In our assessment, the balance of risk and return is a key element in the value for money equation for governance programming. Programmes that aim to influence national policies, institutions or political processes are aiming for catalytic effect. If successful, they may have a wide range of consequential results that cannot easily be captured within a programme logframe. For example, DFID’s work with the Auditor General in Uganda could have a long-term impact on the integrity of all government expenditure, while its work on economic governance in Nepal could reduce barriers to economic growth. Potentially, small investments by DFID could produce much larger financial or economic returns for the partner country. However, this kind of programming is also inherently high risk. DFID is only one player in a complex political space. There is no guarantee of success, and setbacks and reversals are likely.

In this context, there are two areas where DFID’s approach to managing value for money falls short. First, at the programme level, monitoring is not sufficiently focused on the quality of engagement. Where programmes are designed to be informed by political realities and adaptive, monitoring, reporting and supervision should focus on whether the conditions are in place for effective engagement – focusing on questions such as the programme’s depth of understanding of the context, the quality of its relationships with stakeholders, its level of access to senior counterparts and so on. Although these are not easily captured through logframe indicators, they can be assessed through programme governance arrangements.

Second, at the portfolio level, DFID needs a stronger understanding of the balance of risk and return. At the moment, it is difficult to see a clear rationale for the scale of investment on different objectives, in relation to the potential return and the likelihood of success. Furthermore, while governance programmes report on how they have influenced national policies or institutions, the reporting does not capture their contribution to achieving DFID’s strategic objectives. A strong strategic framework for the portfolio would help DFID to assess the significance of what is being achieved.

Governance advisers are helping to support programmes in other sectors, but their role is not well defined

Governance is not only a programming area in its own right, but is also an important enabler (or barrier) to achieving development in other sectors. Most sector programmes, such as health, education and infrastructure programmes, involve elements of influencing policies, reforming institutions and building government capacity. We reviewed a sample of 13 non-governance programmes, selected for their links to governance issues, to understand how DFID promotes governance objectives through other sector programming. Governance-related outputs were clear in their logframes and included strengthening public financial management, tackling corruption or promoting voice and accountability.

DFID Nepal’s Health Sector Programme II (2010-2016; £78 million) had to grapple with the challenge of federalisation, with responsibility for health services shifting to regional authorities under the new federal constitution. This had potential consequences for the quality and reach of services in marginalised areas. In this case, DFID’s work on health system strengthening dovetailed with its governance work helping marginalised communities to assert their rights. In Uganda, DFID’s programme on Accelerating the Rise of Contraceptive Prevalence in Uganda (2011-2017; £31.8 million) identified the political climate for family planning as a key risk to effective implementation. While the Ugandan government has supportive policies on family planning, DFID noted the risk of opposition from some politicians. It sought to mitigate this risk through a separate programme, the UN Joint Programme for Implementation of the National Population Policy (2010-2015; £29.4 million), which worked to build political understanding and support for family planning services.

DFID governance advisers play an advisory role on other sector programmes. DFID’s technical competency framework for the governance cadre highlights that “governance advisers play an important role in the delivery of other DFID strategic objectives, including the institutional and political aspects of work on economic development, service delivery and climate change”. It provides some indication of what governance advisers should do to support economic development and public sector governance and service delivery, for example through diagnostics and analysis. It does not, however, provide guidance on how governance adviser support or engagement in other sector programmes should take place. This means that there is a variation of expectations (from both governance advisers and other specialists) on what level of service and support provision is appropriate. While we saw some good examples of governance adviser engagement in other sector programmes in both Nepal and Uganda (see below), there was not a consistent approach to how this is done and, as a result, engagement appears determined by the preference of individual advisers rather than what is in the best interests of effectiveness. The quality of governance advice and inputs therefore varies across different programmes.

In the 13 programmes that we reviewed, we were able to identify governance adviser input into the design of eight of them. In Nepal, we heard that DFID governance advisers were ‘red teaming’ (providing a challenge function) on other sector programmes. For example, in the earthquake reconstruction team, governance advisers had challenged other sector advisers to assess the risks of programme funds being captured by local political elites. In Uganda, a governance adviser had been allocated to advise at key stages of the Strengthening Education Systems for Improved Learning programme (2016-2022; £46 million). This had been formalised in this instance through the allocation of time in their individual performance management framework.

We found that most of the programmes in our sample were informed by governance-related analysis (especially political economy analysis). However, advisers from other sectors noted the need for more support from governance advisers on programme implementation, including on how to influence policy and support complex processes of institutional reform.

In Nepal, we saw two examples of innovations in portfolio management designed to promote consistency in the way sector programmes address governance challenges. In 2015, DFID conducted a series of joint studies with the government on local service delivery to explore who decides on the allocation of resources within a village or district and how. These studies identified that government and donor-funded development programmes (including DFID programmes) were working in silos, sometimes promoting inconsistent or parallel solutions to governance challenges, including on their approaches to community mobilisation and targeting poor and marginalised groups. In response, DFID is working with the Ministry of Federal Affairs and General Administration to pilot a new initiative in Humla, a remote area, to negotiate changes in the way local development programmes work, to promote a more coherent approach to local service delivery. While this is a recent initiative, the hope is that transformative results can be achieved by encouraging synergies across programmes. The Nepal Field Office for Reconstruction Support is another innovation with the potential to support consistency of approach to local governance challenges across DFID’s portfolio.

Conclusions on the effectiveness of DFID’s governance programming in Nepal and Uganda

We found a solid pattern of performance across DFID’s governance programmes in the two countries. DFID’s own reporting suggests that the majority of completed programmes have met or exceeded their expected programme objectives. In areas where there have been clusters of programmes, we saw a range of positive results. For example, DFID’s work with local government and communities had improved service delivery, with some evidence of reduction in marginalisation. The long-running investments in public financial management had helped the Ugandan government to make more effective use of its development resources, while the work on economic governance in Nepal is producing some promising potential results. DFID’s support to national civil society shows evidence of promoting greater public engagement in policy issues, local service delivery and the democratic process. While there is a pattern of worthwhile results, the lack of country-level governance strategic direction and investment in evaluation makes it difficult to gauge overall effectiveness and to determine whether the portfolio adds up to more of the sum of its parts. This is important since DFID does not have a mechanism to test whether its efforts are being targeted at activities that will, over time, result in the most significant change.

Within the two portfolios, we identified a number of factors that enhanced effectiveness. Sustained engagement with particular institutions and policy areas appears across programme cycles to enhance results, as does considered investment in building relationships with decision-makers and stakeholders. DFID Nepal’s investments in adaptive approaches reflect current thinking on good practice and show promising early results. We note, however, that aspects of DFID’s programme management systems (such as the approach to value for money and results management) are not necessarily supportive of these practices. We also found that levels of staff comfort and confidence with flexibility and risk-taking vary. The use of adaptive approaches is therefore not standard.

Ensuring value for money in governance programming is a difficult challenge, and we are not convinced that DFID’s current approach is sufficient. While it is diligent at controlling costs and encouraging efficiencies, these are unlikely to be the most important aspects of value for money in this type of programming. Given the difficulties of measuring programme outcomes over short time periods, we would welcome more attention to monitoring the quality of engagement and the balance of risk and return across the portfolio.

We found that while DFID also addresses governance issues through its sector programmes in useful and often successful ways, the approaches taken are not consistent. DFID governance advisers provide input into programme designs, which are generally informed by governance analysis. However, other cadres would welcome more sustained input from governance advisers on complex institutional challenges. We saw some positive examples of innovations in portfolio management in Nepal, designed to promote consistency in the way programmes address governance challenges. Governance advisers attribute this partly to their resources being stretched and partly to a need to encourage other cadres to further develop their skills and capabilities in governance and political economy.

Overall, we have awarded a green-amber score for effectiveness. This reflects an overall pattern of successful delivery of programmes in challenging circumstances, and some significant achievements in DFID’s priority areas. However, we remain concerned that DFID is not systematically testing its theories of change and verifying its longer-term impact through external evaluation at the programme or portfolio level.

Learning: How well has DFID adapted its approach to governance in response to lessons learned?

In this section, we explore the extent to which DFID’s governance work in Nepal and Uganda has adapted in response to learning. We assess how learning is generated, both centrally and at the country level, and how lessons are shared. More importantly, we look at how DFID is using this learning to make its governance programming more effective.

While programme designs are evidence-based, learning during implementation has been inconsistent

There are references to evidence and past experience in most of the programme design documents that we reviewed. Experience from earlier programmes in the same country is referenced in 83% of the documents and from other countries in 49%. Findings from research (whether done by DFID or others) appear in 64%. For example, the Local Governance and Community Development Programme incorporated key lessons from a performance-based funding system piloted under an earlier local governance programme. However, we also found cases where learning from earlier programmes had been disregarded. In Nepal, the Enabling State Programme had undertaken a series of lesson-learning exercises between 1997 and 2013 to identify what elements of the programme had worked effectively. However, this learning was not properly incorporated into the successor programme, the Governance Facility (2015-2018; £2 million), which had to undergo redesign when its implementation went off track.

It is now widely recognised that for governance programmes to be effective, they need to be designed and adapted in response to continuous (and politically informed) learning and analysis. While most programme designs were informed by some level of evidence, we saw that Nepal had more evidence of programmes responding to learning in real time through the implementation process than Uganda. Towards the end of the review period, however, DFID began to experiment in both countries with flexible designs that allow programmes to change their activities and theories of change during implementation, in response to experience. We saw examples of flexible programming in Nepal on economic governance and security sector reform, which were designed with the ability to respond to rapidly evolving political contexts. In Uganda DFID experimented by embedding a technical adviser in the district chairman’s office in the city of Gulu, northern Uganda, to help address local governance problems. This initiative was later scaled up to other regions under the Governance, Accountability, Participation and Performance programme.

DFID depends too much on commissioning studies for its political understanding, rather than regular dialogue with national stakeholders

As discussed above, DFID has commissioned a significant amount of political analysis. While this is a good foundation, we are concerned that governance advisers are not supplementing this by engaging with a sufficient range of national stakeholders, so as to collect different perspectives and build a more nuanced understanding. It appears that understanding politics depends too much on specific analytical products, rather than continuously engaging with political actors and domestic stakeholders. This was particularly evident in Uganda, where the analysis is mainly used to inform the design of programmes and programme delivery mechanisms, rather than to fully inform and guide their implementation. In Nepal, DFID has developed more innovative models for engaging with ‘real time’ feedback and understanding the political context, including through the Nepal Field Office for Reconstruction Support.

In both countries, external partners were concerned that DFID has narrowed its circle of regular interlocutors. This means that it is no longer as informed on political developments and risks being overly influenced by a narrow range of views and interests. It is also missing opportunities to influence key national stakeholders. In Nepal, for example, we found that DFID works primarily with civil servants (who tend to be conservative in outlook) and had missed opportunities to engage with politicians and others who are driving the constitutional change process.

In Uganda, stakeholders also questioned whether DFID had been sufficiently engaged with national political leaders and opinion-makers in recent years. As the country moves towards an inevitable political transition, the risks of instability increase. The refugee crisis, shrinking civil society space and future oil revenues are additional sources of uncertainty. Remaining engaged in a variety of domains and with a variety of actors (including politicians) is important for future-proofing DFID’s governance work.

In Nepal, DFID, the Foreign Office (FCO) and the Ministry of Defence (MOD) hold a widely attended joint meeting weekly that offers a useful forum for updating and testing DFID’s understanding of the political context. DFID Nepal also has a mechanism for receiving regular updates on local conditions across the country, through the Risk Management Office (funded jointly with the German development agency GiZ). In Uganda, although weekly meetings are held between DFID, the FCO and the MOD, processes for jointly monitoring and assessing the evolving political context are less advanced.

DFID is not adequately building or making effective use of the knowledge and skills of staff

The literature suggests that engaging staff with the right knowledge and skills is important for effective governance programming. It is helpful to have staff who have a deep understanding of the national context and the networks of relationships with national stakeholders that enable them to act as advocates and coalition builders. This requires staff to spend enough time at each posting. DFID’s current practice is to rotate home civil service staff, including governance advisers, every three years (or two years in the most fragile contexts such as Afghanistan and South Sudan). This is intended to ensure impartiality and to enable experience to be shared across country offices.

We are also concerned that the skill set and experience of DFID’s governance advisers may be too narrow. Stakeholders are in agreement that DFID staff are capable, committed and skilled. Our survey of governance advisers, however, suggests that they tend to have limited practical experience of working in government before working with DFID (see Box 10). Governance advisers also perceived their own expertise as limited in some thematic areas, particularly in the areas of tax, economic development and public financial management. It is notable that while the majority of governance advisers had received specialist training in political economy analysis (92%) and anti-corruption (85%), fewer governance advisers reported having received specialist training in democratisation (27%), economic development (36%) and tax (37%).

DFID reports that the governance cadre has the highest proportion of staff appointed in country of all technical cadres, at 25% (or 40% of country-based advisers). They offer a key source of knowledge on the country context, as well as institutional memory, as they are not usually rotated around country offices. However, of those responding to our survey, only 6% of the more senior A1-level posts were staff appointed in country, which DFID considered a significant improvement compared to the past. In both Nepal and Uganda, we identified a hierarchy between UK expatriate staff (termed home civil servants) and in-country staff that stands in the way of DFID making the best use of the knowledge and experience of staff appointed in country. In Nepal, we are aware of an instance where DFID had declined to follow the advice of in-country staff on a politically sensitive issue, and as a result had faced a backlash from stakeholders and the local media. Our assessment is that DFID’s governance programme would be strengthened by the department making better use of staff appointed in country.

Sharing and use of lessons is limited at country level and DFID underinvests in evaluation

In both Nepal and Uganda, we heard that during our review period DFID had not regularly brought its implementing partners together to share lessons and information between and across programmes. In Nepal, for example, DFID funds a substantial number of civil society organisations through various programmes, but has never sought to bring them together to share their experience (although it does hold occasional meetings to share information on DFID’s requirements as a funder). A recent DFID-funded study showed that a lack of coordination across donor programmes had resulted in inconsistent and contradictory approaches. DFID Nepal, in collaboration with DFID’s Better Delivery Department and the Overseas Development Institute, had recently commissioned work to help develop a new approach to learning, including a focus on adaptive programming.

In the absence of cross-programme learning initiatives from DFID, some of the implementing partners had taken to sharing learning between themselves. In Uganda, the two main implementing partners under DFID’s Strengthening Uganda’s Anti-Corruption and Accountability Regime programme (2014-2020; £29.2 million) had come together for a joint learning session on problem-driven iterative adaption. The implementing partner for the Governance, Accountability, Participation and Performance component of this programme had also brought together the participating civil society organisations to share their experiences.

DFID’s project completion reports are the main process for documenting lessons. DFID’s 2018 project completion review template requires reviewers to consider what went well and which lessons should influence future programmes. There is currently no process, however, for collating these lessons or sharing them across the department.

In the past, DFID central policy teams were responsible for capturing learning in the form of guidance material (especially ‘how to’ notes). This involved dialogue and exchange of experience across country teams, in order to generate a recommended approach. Several DFID staff commented that this process of discussion had been as valuable as the final written guidance. However, this practice has largely fallen away in recent years.

At the central level, we heard that there are established mechanisms for sharing learning across the governance cadre, for example through professional development conferences, seminars and ‘share and learn’ events. In addition, advisers are required to spend 50 to 100 hours a year on continuing professional development, and 10% of their time supporting governance programmes in other countries (for example by conducting annual reviews). As we found in our 2014 report on How DFID Learns, this direct exchange of experience among staff is one of the most important sources of learning.68 There is also a range of relevant communities of practice, such as on political economy analysis, tax, transparency, anti-corruption and political governance as well as regional governance networks. The Governance and Social Development Resource Centre has also been providing support to DFID on evidence and learning since 2001, including through topic guides and synthesis products. Its ‘helpdesk’ function has also provided a rapid response to specific research questions to inform governance programme needs. However, the relationship of this central investment in learning to country requirements, strategic priorities and challenges was not clearly articulated or evident.

While there is regular monitoring and reporting, we found that DFID’s approach to evaluating its governance work was inadequate for learning purposes. In Uganda, programme-wide evaluations had been conducted for four of the 21 programmes reviewed, with project- or activity-level evaluations for a further two. In Nepal, the proportion of evaluations was higher, with programme evaluations conducted for six of the 15 programmes reviewed and one additional activity-level evaluation. Where evaluations had been undertaken, we found that they did not clearly influence decision-making in some key programmes (for example, see para 4.22 and Box 7 above). There has been very little use of impact evaluations to assess whether DFID’s governance programmes have contributed to higherlevel strategic objectives, such as promoting peace and stability or reducing corruption. There was no evaluation of the overall impact of its governance portfolio in either country.

DFID increased its investment in governance research over the review period, but the impact on programming is not evident

DFID is a significant investor in research to understand what works in development assistance. Its 2016 Research Review set out plans to spend an average of £390 million, or approximately 3% of its annual budget, on research over the next four years.70 Staff from DFID’s Research and Evidence Division told us that DFID’s dedicated research budget for governance work was currently around £15 million per year, although a lot of governance research intersects with social development and conflict research, making it difficult to deduce a completely accurate figure.

DFID’s investment in governance research goes to a range of partners, including the Effective States and Inclusive Development Research Centre, the International Centre for Taxation and Development, the Abdul Latif Jameel Poverty Action Lab Governance Initiative and the Anti-Corruption Evidence programme, all of which deliver research that is publicly available.

DFID has no mechanism for tracking the impact of its research investments on programming, however. We heard of some specific examples: in Uganda, work undertaken by the Effective States and Inclusive Development Research Centre had been used to support in-country diagnostic work and influenced programme choices. However, there is no basis for assessing whether DFID has maximised the value of its research work.

Conclusions on learning in Nepal and Uganda

We found that DFID’s programme designs are generally informed by research and lessons from previous programmes. There is less evidence that they are adapted during implementation in response to experience, although this is more evident with recent programmes with a more flexible design. While DFID has invested in a significant amount of political analysis, it also appears to have narrowed its engagement with national stakeholders. Political analysis is too often seen as a commissioned product, rather than an ongoing process, particularly so in Uganda, and there is a risk that DFID may be overly influenced by a narrow range of views.

While DFID governance advisers are well regarded, we are concerned at the narrow range of practical experience in the cadre. In addition, regular rotation of advisers and under-use of the knowledge of staff from the partner country means that DFID is not drawing on deep understanding of the national context. While we recognise that the governance cadre has the highest proportion of staff appointed in country across all of DFID’s technical cadres, we feel it could still go further on how it uses and applies this knowledge.

We found that DFID lacks fully functional mechanisms for sharing experience across its programmes and implementing partners. It documents lessons in its project completion reports, but these are not disseminated across the department. There are some useful mechanisms for sharing practical experience, including communities of practice and the ‘10% cadre time’ pledge, but the relationship of this to country priorities and challenges was not evident. We find that DFID has underinvested in portfolio-level and impact evaluations and the use of programme evaluations in decision-making has been unclear. DFID has a substantial central research programme on governance, but no mechanism for tracking the uptake of this research in its programming.

Overall, we have awarded an amber-red score for learning, reflecting concerns that DFID is not making the best possible use of its knowledge and experience to inform its governance programming.

Conclusions and recommendations

Conclusions

Over the past decade, DFID has improved its understanding of the role of governance issues within the development process. Political analysis is now routine and informs programme design and operational choices. As a result, programmes have become better tailored to the local context. In Nepal and Uganda, we found that they were supportive of broader UK strategic objectives around promoting stability and economic development. However, we were concerned at the lack of a clear strategic framework for governance programming in each country. This makes it difficult for DFID to manage the portfolio in dynamic environments and ensure that programmes work together to drive transformational change.

While DFID’s evidence base on programme effectiveness is not particularly strong, the evidence that we were able to collect suggests a good pattern of performance, particularly in Nepal. We saw good results at the community level through voice and accountability programming, helping to improve service delivery in some of the poorest and most remote areas, and in some cases to challenge social norms and practices. Work with central government was helping to strengthen public financial management and build economic governance capacity.

Governance results were particularly strong where DFID had sustained its engagement in thematic areas or with institutions over an extended period. We also saw strong examples of flexible and adaptive programming, with DFID Nepal a leader in this area. However, this is not being fully demonstrated in practice. While DFID’s rules allow for more flexible programming, we found that levels of staff comfort and confidence with flexibility and risk-taking varied. Tensions therefore exist between DFID’s approach to promoting effectiveness and value for money and its latest thinking on good practice in governance programming.

Despite considerable investment in analysis, DFID is not getting the most value out of the knowledge it generates on how to do governance programming. We expected to see more evidence of evaluation of governance efforts and more application of the lessons it has already gathered, or has the potential to gather.

Recommendations

Recommendation 1

DFID should articulate in more detail its strategic approach to governance at country level, and make more use of scenario planning and risk management tools to support portfolio management in volatile contexts. Its strategies should consider and articulate the desired balance of risk and return across its governance portfolios.

Problem statements

  • DFID does not set out in sufficient detail its country-level strategic approach to governance and is not taking a portfolio approach to its governance work.
  • We found that this lack of clarity at the strategic level contributed to gaps in the portfolios and a lack of continuity in programming over time.

Recommendation 2

DFID should identify areas where sustained engagement is likely to be required to generate the desired results, and invest in long-term relationships with key counterparts, while maintaining the flexibility to scale individual activities up and down as appropriate.

Problem statements

  • DFID’s results system does not track outcomes across programme cycles or reflect on how results build over time. As a result there is less incentive to retain historical knowledge and to prioritise continuity in programming.
  • It is widely understood that governance changes take time. Results for governance work often emerge over several programme cycles, and are not captured in programme-level results frameworks.
  • In both countries, we saw instances where DFID’s engagement had not been consistent. DFID has not always prioritised building long-term relationships.

Recommendation 3

DFID should maximise value from its governance cadre by increasing the amount of time governance advisers have available to spend on technical inputs (including into other sector programmes) and external influencing and engagement, rather than programme management and administrative tasks.

Problem statements

  • Most DFID governance advisers perceive external influencing and engagement as valuable but report spending less than 20% of their time engaging with policy-makers and influencers.
  • There is no formal guidance that enables governance advisers or other sectoral specialists to have shared expectations for how governance support should be provided to other sectors.
  • In both countries, we identified a risk that the governance adviser role had become too narrowly focused on political economy analysis rather than supporting sectors with advice such as on institutional development.

Recommendation 4

DFID should continue to develop the capacity of its governance cadre, including by: (i) making better use of the knowledge and experience of staff appointed in country; (ii) posting home civil service staff for longer periods in country; (iii) increasing diversity in terms of experience, backgrounds and local knowledge; (iv) placing more weight on practical delivery experience in its recruitment.

Problem statements

  • National staff typically have strong contextual knowledge and good institutional memory, but this is not sufficiently utilised.
  • While 40% of DFID’s country-based advisers are nationals of the partner country, they make up only 6% of the more senior A1-level advisers.
  • DFID’s practice of rotating home civil service advisers every three years does not facilitate the depth of knowledge or relationships required for effective governance work.

Recommendation 5

DFID should use evaluation more to test the validity of key propositions underlying its governance programmes and portfolios (for example that support for marginalised groups reduces conflicts). It should increase its investment in learning within programmes, and make sure the lessons are used to inform the management of country portfolios and programmes.

Problem statements

  • We found that DFID had undertaken no evaluations of its country governance portfolios, there was limited use of impact evaluations and the use of programme evaluations to inform decision-making was unclear.
  • This makes it difficult to gauge overall effectiveness.
  • DFID does not routinely track the impact its central-level investment in learning has made on programming.
  • DFID does not sufficiently curate how learning is captured, disseminated and used across the organisation.

Annexes

Annex 1 Sample programmes

Part 1: Governance programmes in Nepal and Uganda

Part 2: Other sector programmes in Nepal and Uganda

Part 3: Centrally managed programmes with a governance focus in Nepal and/or Uganda