DFID’s Approach to Anti-Corruption and its Impact on the Poor

Executive Summary

The Independent Commission for Aid Impact (ICAI) is the independent body responsible for scrutinising UK aid. We focus on maximising the effectiveness of the UK aid budget for intended beneficiaries and on delivering value for money for UK taxpayers. We carry out independent reviews of aid programmes and of issues affecting the delivery of UK aid. We publish transparent, impartial and objective reports to provide evidence and clear recommendations to support UK Government decision-making and to strengthen the accountability of the aid programme. Our reports are written to be accessible to a general readership and we use a simple ‘traffic light’ system to report our judgement on each programme or topic we review.

This report reviews DFID’s work on anti-corruption and is the second of our reports on this subject. It focusses on DFID’s efforts to reduce corruption as experienced by the poor, rather than on the integrity of the use of DFID’s money. Corruption is a fundamental issue that afflicts the everyday lives of the very poorest and thwarts global efforts to lift countries out of poverty.

Overall

Assessment: Amber-Red

DFID recognises corruption as a critical development challenge and seeks to tackle it through direct and indirect activities. DFID has not, however, developed an approach equal to the challenge, nor has it focussed its efforts on the poor. While some programmes show limited achievements, there is little evidence of impact on corruption levels or in meeting the needs of the poor. DFID’s willingness to engage in programming that explicitly tackles corruption generally is often constrained by political sensitivity in-country. It is not capturing and applying lessons learned.

Objectives

Assessment: Amber-Red

DFID’s objectives at the global and overarching level are clear but do not address the poor. DFID has been a key participant in global initiatives that attempt to fight corruption. At the country level, DFID has created a varied portfolio of activities that can contribute to reducing corruption. Very few of these activities, however, explicitly tackle corruption or its impact on the poor. A lack of evidence of what works restricts the type of programming undertaken. DFID’s recent theory of change on corruption is coherent and realistic but it remains at a high level and is not sufficiently focussed on tackling corruption as it affects the poor.

Delivery

Assessment: Green-Amber

DFID makes good use of flexible programme structures. It aligns with government delivery when possible and uses knowledgeable external personnel for much of its work. The politically sensitive nature of corruption, however, may lead country offices to avoid anti-corruption programming too readily. The lack of co-ordination within DFID and with the Foreign and Commonwealth Office (FCO) at Whitehall has resulted in a lack of ambition toward tackling corruption. Staff turnover in country offices and an absence of experienced DFID personnel in some key positions have led to sub-optimal co-ordination and follow-through.

Impact

Assessment: Amber-Red

DFID’s anti-corruption activities have demonstrated certain achievements but have had little success in reducing the effects of corruption, especially as directly experienced by the poor. Understanding of impact is hindered by objectives and evaluation criteria that rarely relate to measures of corruption. Programmes are not often planned or managed around realistic intermediate goals and interim outcomes. DFID’s anti-corruption programming is not accountable to the poor: DFID does not engage with the poor during design and implementation and programming suffers from lack of continuous feedback loops with stakeholders, especially the poor. DFID’s anti-corruption activities can sometimes result in increased opportunities for corruption.

Learning

Assessment: Red

DFID has little understanding of what is working with respect to its anti-corruption activities. Further, DFID does not fully understand which of its activities are addressing corruption or how they will make a difference. It also has not sought sufficient evidence to understand whether its anti-corruption activities are having an impact on the poor. DFID has missed opportunities to use regional anti-corruption advisors to drive the corruption agenda. Strong analysis of specific country contexts regarding corruption has also been lacking.

Recommendations

Recommendation 1: DFID, in conjunction with the FCO and other UK Government departments, should articulate and implement a detailed plan setting out the level of ambition, commitment and positioning of the UK with respect to tackling corruption in its priority countries, including as experienced by the poor.

Recommendation 2: DFID should develop standalone anti-corruption country strategies and, in addition to its current activities, programming that explicitly tackles corruption and that extends over a 10- to 15-year time horizon with short-, medium- and long-term goals for reducing corruption, particularly with respect to the poor.

Recommendation 3: DFID should include in its expanded anti-corruption portfolio many more programmes which specifically target the everyday corruption experienced by the poor and educate the population about the ill effects of corruption.

Recommendation 4: DFID should gather and publish targeted and dynamic feedback from the stakeholders of its anti-corruption work, including the intended beneficiaries, to allow DFID to ‘spot check’ and correct its existing programmes and to inform new programming.

Recommendation 5: DFID should create an internal embedded centre of excellence explicitly to focus on anti-corruption and to gather evidence of effectiveness, disseminate lessons learned and cultivate expertise that will drive anti-corruption efforts globally.

1 Introduction

Background to this review

1.1 This review considers whether the Department for International Development’s (DFID’s) anti-corruption activities are reducing the negative effects of corruption, in particular as experienced by the poorest in society. It is the second of our reviews of DFID’s approach to anti-corruption. The first, published in 2011, focussed primarily on the effectiveness of DFID’s efforts to protect UK taxpayers’ money from the losses arising from the mismanagement of funds invested in programmes, which is termed as ‘fraud’.¹ The term ‘corruption’ can encompass those issues of fraud but, as used in this report, is intended to refer more particularly to the abuse of power for private gain.²

1.2 This is a thematic report which sets out to assess the coherence and effectiveness of DFID’s approach to tackling corruption at the international, national and sub-national levels. While our review was informed by DFID programming from around the world, our principal sources of evidence for DFID’s country-level efforts are the programmes in Nigeria and Nepal. We chose these countries because DFID engages in representative anti-corruption activities in them and because they offered contrasting but complementary contexts in which to review anti-corruption efforts.³

Corruption is a paramount development challenge

1.3 Corruption presents a formidable challenge to international development. It is prevalent in countries where the world’s poorest live, siphoning off precious resources from those most in need.⁴ It impacts the poor disproportionately and particularly women and girls, as highlighted in Figure 1.⁵

1.5 Corruption is not susceptible to quick or easy fixes. A paper by the U4 Anti-Corruption Resource Centre, sponsored by DFID and published in 2012, found a lack of research focussing on the effectiveness and impact of anti-corruption interventions. In the research that does exist, it found very little strong or conclusive evidence that the interventions that have been pursued have been effective.⁹

1.6 Corruption is often entrenched at the highest levels of government and society and is a politically sensitive issue for host governments. It can therefore be difficult for organisations such as DFID to take action to tackle corruption.

Corruption is recognised as an international priority

1.7 In recent years, tackling corruption has become an increasing international priority, as reflected in the widespread ratification by countries of the United Nations Convention Against Corruption.¹⁰ Anti-corruption has been a priority of the G20, as expressed most prominently in the activities of the G20 Anti-Corruption Working Group.¹¹

1.8 In addition, the Prime Minister and the Secretary of State for International Development, among others, have elevated anti-corruption on the UK’s international agenda.¹²

DFID’s anti-corruption activities

1.9 DFID recognises corruption and fraud as central development challenges. It engages in a variety of activities that seek to tackle corruption but mostly in an indirect way. That is, the activities do not principally or explicitly seek to impact corruption but are perceived to reap some anti-corruption benefit in the process of doing something else. DFID’s efforts in response to our earlier report

1.10 Our 2011 report included recommendations related to the effectiveness of DFID’s efforts to protect UK taxpayers’ money from the ill effects of fraud. The report gave DFID an amber-red rating.¹³

1.11 In response to that report, DFID has initiated a number of activities to advance its approach to counter-fraud and anti-corruption, including:

• the development by country offices of Anti-Corruption and Counter Fraud (ACCF) strategies under the guidance of regional anti-corruption advisors; and
• the creation and work of the Counter Corruption and Counter Fraud Group, known as Cx3 and its operational sub-group, Cx3(o).

1.12 Most of this effort has focussed on tackling fraud and particularly on protecting UK taxpayer funds. By contrast, DFID’s efforts on anti-corruption in response to our earlier report, including developing anti-corruption programming or strategies, have been more limited, although they too were an element of our 2011 recommendations.

1.13 In recent months, however, this has begun to change. Since our current review commenced, DFID has developed an overarching anti-corruption theory of change¹⁴ and it has begun to make efforts to understand better the type, scale and effectiveness of the anti-corruption work it does.¹⁵

DFID’s indirect anti-corruption work

1.14 At the global level, DFID’s anti-corruption activities include financial and technical support for multilateral anti-corruption initiatives. These include the implementation of the UN Convention Against Corruption;¹⁶ transparency and accountability initiatives, such as the Extractive Industries Transparency Initiative¹⁷ and the Medicines Transparency Alliance;¹⁸ and international efforts to combat illicit financial flows, such as the International Centre for Asset Recovery at the Basel Institute for Governance¹⁹ and the Financial Action Task Force.²⁰

1.15 DFID also supports a number of UK investigative and prosecution bodies that more directly target corruption. These include the Metropolitan Police Proceeds of Corruption Unit and the City of London Police Overseas Anti-Corruption Unit, which bring actions against money laundering and overseas corruption and seek to recover stolen assets, which may be invested or kept in bank accounts in the UK.

1.16 At the country level, DFID undertakes programmes to improve public financial management systems in government to make them more transparent, reduce the scope for discretionary decision-making and strengthen oversight by government and citizens over the use of public resources. For example, many such programmes rely on the assumption that making budgets open and transparent will result in citizens or other stakeholders reviewing those budgets and acting on them to hold public officials to account.

1.17 Other activities relate to building capacity in sectors particularly prone to corruption, such as the police and the judiciary. Underlying these activities is an understanding that by increasing accountability or by, say, improving relationships between police and communities, the opportunity for corruption in these sectors will be diminished.

1.18 These indirect efforts undertaken at the country-level — in the terminology used by DFID and the development sector more widely — are divided into two categories: ‘supply’-side activities and ‘demand’-side activities. Supply-side activities focus on improving the institutions and systems that provide governance or services. Demand-side activities focus on increasing the capacity of citizens or citizen groups to hold their governments to account and to demand better services.

DFID may be embarking on more direct measures

1.19 DFID has not historically undertaken extensive programming specifically designed to address petty corruption, the everyday, small-scale but endemic corruption that includes bribery and personal favours, which most directly affects the lives of poor people. Very recently, DFID has developed one such programme, Support to Anti-Corruption in Sierra Leone,²¹ which seeks to empower citizens in their ability to challenge petty corruption in health, education, water and sanitation and national power services delivery. This is the only DFID programme to focus almost exclusively on petty corruption.²²

1.20 DFID has recently also embarked on efforts designed to fill the evidence gap in anti-corruption interventions. These include:

• working with the U4 Anti-Corruption Resource Centre to identify proxy (or indirect) indicators for corruption;
• commissioning a paper by the Overseas Development Institute on factors driving corruption, the costs and impacts of corruption and anti-corruption interventions; and
• developing a research programme on effective anti-corruption efforts, to begin in early 2015.

DFID’s expenditure on anti-corruption activities cannot be accurately quantified

1.21 DFID does not currently have data on the entirety of its spending to tackle corruption, either in terms of programmes or staff costs. This is because DFID codes its spending according to Organisation for Economic Co-operation and Development (OECD) rules, which apply a dedicated anti-corruption code only to specific types of expenditure.²³

1.22 As a result, DFID was unable to tell us how much it spends on anti-corruption activities. We, therefore, could not form a view as to the relative weight given to anti-corruption work within DFID’s overall development assistance. Using the narrow definition of anti-corruption expenditure, DFID has increased its annual anti-corruption expenditure from approximately £3.5 million in 2007-08 to some £22 million in 2013-14, as reflected in Figure 2. This is in the context of DFID’s overall annual budget of approximately £10.3 billion. It significantly underestimates DFID’s efforts by excluding programmes that seek to affect corruption indirectly and DFID funds invested through agencies of various multilateral institutions, such as the UN Development Programme.

1.23 Under current projections, DFID expects to spend about £190 million on these narrowly defined anti-corruption activities during the years 2014-15 to 2017-18 (see Figure 2). £71 million of this expenditure has already been approved.²⁴

Our approach to the review and methodology

1.24 For this review, we undertook the following steps:

• a desk review of programme documentation, strategies and other documents related to DFID’s anti-corruption efforts;²⁵
• visits to two countries, Nigeria and Nepal, to witness the impact of a range of activities DFID is supporting to tackle corruption; and
• interviews with stakeholders, including DFID staff in Whitehall and East Kilbride, as well as with country offices and regional advisors, programme personnel, programme stakeholders (including government officials and intended beneficiaries) and independent experts.²⁶

1.25 Among the wider questions we covered — as set out in our Terms of Reference²⁷ — we focussed on the following core issues:

• the extent to which DFID is directly tackling corruption;
• how much DFID’s anti-corruption efforts are making a difference in the lives of poor people; and
• the extent to which DFID is listening to and learning from the views of its intended beneficiaries and other stakeholders.

1.26 This approach has allowed us to complement our first anti-corruption review, which focussed largely on DFID’s counter-fraud measures to protect UK taxpayers’ funds. It also enabled us to understand the wider impact of DFID’s efforts to stem corruption in its priority countries. With respect to counter-fraud issues, we undertook, among other things, a desk review of the audited financial statements of the civil society organisation (CSO) and private sector-implemented programmes that were the subject of our review.²⁸

Listening to intended beneficiaries and other stakeholders was an important part of our work

1.27 The central focus of this report is on the poor, particularly in the context of fragile states where corruption is endemic, rule of law is uncertain and populations are vulnerable. In these complex, rapidly changing environments, it is highly beneficial to collect continuous feedback from the intended beneficiaries and other stakeholders of programmes and to use this feedback to course correct programmes towards interim outcome targets. For the purposes of this report, we use the term ‘stakeholders’ to refer to the broad range of constituencies who are involved in or impacted by DFID’s programmes; we use the term ‘intended beneficiaries’ more narrowly to refer to the poor, who are the ultimate beneficiaries of DFID’s development assistance.

1.28 In order to broaden the base of evidence obtained during the course of our review, we administered targeted surveys to understand whether, and to what extent, DFID’s anti-corruption efforts were perceived by survey respondents to be making a difference.

1.29 Our surveys obtained the views of 5,789 people, the vast majority through face-to-face surveys, as described further in Annex A2. The respondents to the surveys comprised stakeholders of DFID’s programmes, including beneficiaries and also members of the public more generally.²⁹ These surveys allowed us to do two things. First, they generated data that serve as ‘spot checks’ on specific programme outputs and the effect of anti-corruption efforts. Second, they illustrate the potential for light-touch, continuous micro-surveys to enable citizen accountability and improved performance management in anti-corruption programming.³⁰ As discussed in Annex A2, micro-surveys of this kind could serve as a valuable complement to DFID’s existing approach to obtaining feedback.

We undertook visits to Nigeria and Nepal, two contrasting but complementary countries

1.30 We looked in detail at DFID’s anti-corruption work in two contrasting but complementary countries: Nigeria and Nepal, both of which are fragile states and have been or are affected by internal conflict. Annex A1 sets out the anti-corruption-related programmes that we reviewed in both countries, placing the programmes in their historical context and in the context of relevant predecessor and successor programmes. Figure 3 below and Figure 4 on page 7 provide additional information on the contexts of Nigeria and Nepal.

1.31 During our country visits, we observed a widespread acceptance of corruption by the populations of these countries, with a ‘democratisation of corruption’ having fully taken hold in Nigeria, and Nepal being currently at a tipping point.³¹ We witnessed DFID’s efforts to combat corruption in countries where corruption is at different but advanced stages of entrenchment.

1.32 In Nigeria, petty corruption touches virtually every aspect of life and is accepted throughout society as normal and necessary. We heard stories of parents paying bribes to teachers in order to educate their children; students paying bribes to administrators to take exams; workers paying bribes to get jobs and then to receive their salaries; and pensioners paying bribes to receive their pensions.

1.35 DFID also categorises Nepal as a fragile state but, in contrast to Nigeria, DFID provides sector budget support to the Government of Nepal. This degree of engagement with a host government can be challenging, given the extent to which issues of corruption touch the government.

2 Findings: Objectives

Objectives

Assessment: Amber-Red

2.1 This section considers whether DFID has clear and ambitious objectives for tackling corruption and its impact on the poor. In doing so, we assess the effectiveness of DFID’s ACCF country strategies and consider whether they are predicated on a robust theory of change.

DFID’s objectives at the global and overarching level are clear but not focussed on the poor

Global and UK-based anti-corruption efforts are extensive

2.2 Publicly acknowledging the link between corruption and development, the International Development Secretary commented in 2013 that ‘when corruption happens in developing countries, it is the very poorest people who foot the bill. It deters investment, cheats citizens out of the services and support they need to develop their economies and end aid dependency. Stamping out corruption, fraud, money laundering and tax evasion is a vital part of our work.’⁴⁴ Indeed, the absence of corruption has been identified by the UK Prime Minister as being part of the ‘golden thread’ of conditions that enable open economies and societies to thrive.⁴⁵

2.3 DFID shows leadership on important activities at the global level to tackle corruption. This includes global transparency initiatives, such as the Extractive Industries Transparency Initiative (EITI) and the Medicines Transparency Alliance, which are concerned with processes that address transparency levels in sectors particularly prone to corruption. DFID has provided support to the EITI since 2002 and its current chair is a former Secretary of State for International Development. Since DFID’s support for the EITI commenced, it has been implemented by 46 countries, including 12 DFID priority countries. In addition, DFID has supported the formation and implementation of the UN Convention Against Corruption – which seeks to create a robust framework of anti-corruption laws and promotes their implementation by governments – and supports the operation of its Review Mechanism, a process which analyses the adequacy of parties’ legal frameworks for tackling corruption.

2.4 In response to the recommendations of our 2011 anti-corruption review, DFID has taken positive steps to support efforts to enforce against the proceeds of corruption.⁴⁶ Globally, DFID has funded initiatives to pursue asset recovery through the International Centre for Asset Recovery at the Basel Institute for Governance, which supports international efforts to end safe havens for funds obtained through corruption and other illicit means.⁴⁷

2.5 DFID funds investigations of money laundering by foreign officials in the UK and of allegations of corruption overseas by UK citizens and companies. DFID is also working to return stolen funds to their country of origin. DFID is conscious, however, that the International Development Committee has advised it to help ensure that funds are returned for the benefit of the poor.⁴⁸ Resolving these conflicting demands requires detailed consideration amongst DFID, FCO, enforcement agencies and recipient governments.

2.6 While these extensive efforts acknowledge the link between corruption and development, they do not articulate or provide a line of sight to the way in which the lives of the poor will be improved as a result. Indeed, their eventual impact on the poor is not assessed by DFID. For instance, Nigeria is a signatory to the UN Convention Against Corruption and has sought to implement the Extractive Industries Transparency Initiative (EITI) locally. The Nigerian EITI was established in 2004 and reports that Nigeria has been compliant with the EITI since March 2011;⁴⁹ DFID has provided support to that implementation more recently. DFID and others, however, have been unable to establish a robust causal link to evaluate whether corruption, as experienced by the poor, has been reduced by these efforts.⁵⁰

DFID has very recently articulated an overarching theory of change to inform its anti-corruption work

2.7 Since our review commenced, DFID has sought to formalise a theory of change on tackling corruption in its priority countries. This sets out an extensive set of interventions to reduce corruption, in line with the interventions DFID currently undertakes in-country. Underpinning this theory of change is the assertion, consistent with the research literature, that three dimensions of behaviour drive corruption: the existence of opportunity; the existence of incentives; and the absence of sanctions to punish.⁵¹

2.8 The theory of change also points to the need to change social attitudes and political commitment to tackle corruption. It proposes a joint UK Government strategic approach on anti-corruption to facilitate political interactions with governments in DFID’s priority countries. Such a strategic approach does not appear to exist currently. We saw limited evidence of co-ordination between UK Government departments centrally on anti-corruption. In Nigeria and Nepal, DFID has made preliminary efforts to co-ordinate its approach on corruption with the FCO at a country level.

2.9 The theory of change appears to be coherent and logical. It remains, however, largely at the macro-level of strengthening institutions and enabling countries to achieve a sustainable exit from poverty, without discussing how to tackle corruption as experienced by the poor. We also found it to be insufficiently detailed in the nature of some of its elements and how they will lead to a reduction in corruption.

2.10 It does not, for example, set out a short-, medium- and long-term vision for tackling corruption with intermediate goals and with reference to the importance of aligning predecessor and successor programming over the long term. Sustained engagement to tackle corruption should be integrated into the theory of change in terms of ambition, planning and application of lessons learned. For instance, DFID has been engaged in the security and justice sector in Nigeria for more than ten years, through the Justice for All programme (J4A) and its predecessor, the Security, Justice and Growth Programme. We commend DFID for putting sustained effort into trying to tackle reform of the justice sector and enforcement against corruption. J4A worked to build on the lessons from its predecessor programme and has sought to scale up aspects of that programme. We saw little evidence, however, that the successes and challenges of its predecessor were taken into account in the design of J4A. We also saw limited evidence that J4A had taken into account the challenges of implementation of its predecessor programme.

Anti-Corruption and Counter Fraud strategies at the country level

2.11 Our 2011 anti-corruption review recommended that DFID should develop an explicit anti-corruption strategy in any country assessed as having a high risk of corruption. Going beyond our recommendation, in January 2013 DFID published ACCF strategies for 28 priority countries.⁵² DFID’s regional anti-corruption advisors provided support to its country offices to develop the ACCF strategies and specific anti-corruption activities.

2.12 All 28 priority country offices also have non-public ACCF strategies. These documents seek to incorporate sensitive political economy analysis, although this is generally not extensive. In Nepal, DFID was aware of – and had drawn on – one recent political economy analysis and had funded a fiduciary risk analysis of Nepal in 2013. With respect to the impact of corruption on the poor, DFID Nigeria, amongst others, told us that it is difficult to determine its effect, as corruption is often more readily linked to government activities than to the everyday patronage and bribery that the poor experience.

2.13 We also saw recent efforts by DFID to engage with FCO counterparts in-country to understand recent and prospective changes in the political economy of the countries in which they operate. In Nigeria, for example, discussions are under way to create a common UK Government strategy to engage against corruption following the Nigerian national elections that are due in 2015.

ACCF strategies focus on indirect approaches to tackle corruption

2.14 The ACCF strategies discuss the portfolio of DFID’s activities that seek to increase transparency and accountability, as an indirect means to reduce corruption. On the government ‘supply’ side, transparency and accountability programming includes strengthening financial systems and institutions to reduce the space for corruption. On the ‘demand’ side, accountability programming includes efforts to raise awareness of entitlements and mobilise campaigns of citizens, their political representatives and the media to obtain those entitlements from the government. These types of anti-corruption activities are consistent with those undertaken by other donors and discussed in academic literature.

2.15 In its programmes targeting the health and education sectors, DFID makes some more focussed efforts to reduce elements of corruption such as absenteeism and lack of provision of medicines or textbooks. For example, it has undertaken targeted patient surveys in connection with the Nepal Health Sector Programme and a public expenditure tracking survey on the delivery of textbooks in connection with the School Sector Reform Programme in Nepal.

In their current form, ACCF strategies are not effective documents

2.16 At the time of publication in 2013, DFID’s ACCF strategies appeared primarily to memorialise existing anti-corruption efforts. We found little evidence that country offices have subsequently adapted their efforts to tackle corruption on the basis of any assessment in their ACCF strategies.

2.17 The strategies do not appear to be dynamic documents. We saw little evidence that they are frequently updated to address changes in circumstances and new opportunities for anti-corruption programming. For example, while the Nepal ACCF strategy was updated just prior to our visit, the amended political economy analysis had not resulted in changes to DFID’s plans to tackle corruption. The strategy still referred to the limitations presented by ‘the absence of a parliament’, rather than the impact of the elections in 2013 and formation of a governing coalition.

2.18 Where the ACCF strategies do set out a clear picture of corruption challenges, the programmes that are undertaken in-country do not always appear to be designed to address those challenges. In Nepal, for example, DFID identified the police, judiciary, political parties and civil service as institutions particularly affected by corruption. DFID has, however, designed and undertaken very little programming to address corruption in those institutions, even indirectly.

2.19 DFID does have a tool, the Country Poverty Reduction Diagnostic (CPRD), which could place its ACCF strategies into a framework of development activities at the country level. Given that DFID’s global theory of change argues that the reduction of corruption leads to a more sustainable exit from poverty, it is surprising that DFID’s diagrammatic system for poverty reduction in the CPRD does not require country offices specifically to identify challenges related to corruption. DFID’s core guidance on the CPRD suggests that corruption might arise as a barrier to poverty reduction in two circumstances – as a matter of the political settlement or in social policy and service delivery. Neither DFID in Nepal nor DFID in Nigeria opted to identify corruption as a key issue in their interpretation of the diagrammatic system of poverty reduction. While the CPRDs for these countries acknowledge that corruption is a driver of poverty, they do not translate that understanding into proposed action. The CPRD does not yet, therefore, appear to have driven new thinking on tackling corruption. In our view, DFID should raise the challenges of corruption more clearly in its core guidance on the use of the CPRD and the diagrammatic system of poverty reduction. This would help ensure that country offices identify the country-specific issues of corruption and articulate how they can be addressed.

There is little explicit anti-corruption programming

There is scope for more direct anti-corruption work

2.20 In Nepal and Nigeria, we heard consistently from stakeholders that DFID could take a more overt and direct approach to corruption. Those stakeholders include intended beneficiaries, government officials, CSOs (those which seek to target corruption and others), the media and other donors. They suggested that this approach could take the form of funding explicit anti-corruption programming and advocating more robustly within government against corruption. They also suggested the provision of education, particularly to young people, to ensure they learn that corruption is a form of theft from society, that it is not to be condoned or aspired to and that a non-corrupt society, which is not reliant on patronage, is a viable and preferable alternative.

2.21 During our review, we heard of a work plan by the British High Commission in Kenya for a more direct approach to corruption along these lines. This approach combines strong diplomatic messages⁵³ with targeted development action undertaken by DFID to promote open government, transparency and civic awareness, including raising awareness that corruption deters foreign investment and defeats growth and development.

The politicised nature of corruption may deter DFID too readily from action

2.22 We appreciate that in order to have a mandate to operate in a country, DFID needs to maintain a relationship with the host government. Where issues of corruption are linked to individuals within specific ministries or embedded within the systems of host governments, DFID’s desire to tackle corruption can directly conflict with those relationships. In such circumstances, DFID has to decide whether it will engage with the host government at all. If DFID does decide to engage, it — together with other donors and arms of the UK Government — must maintain strong and continuous political dialogue to achieve real commitments and action from the host government to tackle corruption.

2.23 We were concerned that reluctance on the part of DFID to take a sufficiently strong stance with host governments on corruption within their own institutions may constrain and diminish DFID’s anti-corruption interventions in those countries.

2.24 We also saw instances of a hesitancy to react to changes in circumstances. In Nepal, for example, over a year after a head was appointed to one of the anti-corruption agencies, DFID was still considering the extent to which they would engage with that institution. In our view, DFID should have moved more quickly to assess fully this development and whether significant support should have been provided to the anti-corruption agency.

Without an explicit focus on corruption, the effectiveness of existing programmes is diminished

2.25 We found that programme documents related to DFID’s anti-corruption activities seldom use the word ‘corruption’. There are a few programmes which have specific anti-corruption components, such as J4A in Nigeria, which supports the anti-corruption agencies. Generally, however, programme documentation much more commonly speaks of enhancing transparency and accountability. Indeed, apart from the development by DFID of ACCF strategies, there appear to be limited ways in which the anti-corruption intents of programmes are formally identified, let alone catalogued.

2.26 Several programmes we reviewed, including the Nepal Health Sector Programme and the Nigerian State Accountability and Voice Initiative, set out their high-level impact statements and outputs by reference to the Millennium Development Goals. By focussing on those targets and failing to measure their effect on corruption, the programmes undervalue their actual achievements in tackling corruption.

2.27 Given the sensitivity regarding anti-corruption programming, particularly where it impacts on the vested interests of those in power, it is understandable that public documents are not open about the intended effect of programming. This can, however, lead to DFID losing sight of the importance of programmes from an anti-corruption perspective, thereby failing to recognise their wider implications, as we found with one programme in Nepal (see Figure 5).

DFID’s anti-corruption work does not focus sufficiently on the poor

2.28 We found that DFID has done little to tie its portfolio of anti-corruption activities to impact on the poor. Moreover, DFID is not sufficiently focussed on the experiences of corruption by the poor.

2.29 While DFID’s efforts at the global level potentially have important development benefits for poor countries, they are far removed from the poor, a fact which was acknowledged by a DFID representative to whom we spoke. The chain between strengthening systems at the macro-level that reduce the scope for corruption and the eventual reduction of corruption experienced by the poor is, inevitably, very long.

2.30 The chain is also, in certain respects, untested. For example, DFID expects to repatriate stolen assets which have been recovered through legal proceedings in the UK. In theory, these funds will be used by governments to support development in their countries.⁵⁴ We did not, however, hear of any mechanism that would ensure that this would occur and DFID has not tested the impact on the poor of such activities.

2.32 DFID appears to lack clear leadership in translating the UK Government’s high-level anti-corruption policy into programming that seeks to impact on the poor. DFID does have a working group, Cx3, to co-ordinate its anti-corruption and counter-fraud work.⁵⁶ It is intended to operate strategically across DFID to join up policy, thinking and actions around counter-fraud and anti-corruption as it relates both to misuse of DFID funds and DFID’s in-country efforts to stem corruption generally.

2.33 Much of the work of Cx3 and Cx3(o) has, however, focussed on the counter-fraud aspects of its mandate. It has focussed less on co-ordinating DFID’s anti-corruption work generally and even less on considering impact on the poor. In October 2013, Cx3 noted that country offices lacked engagement from senior management on corruption and, consequently, were unclear on where corruption sits against other priority issues. Progress since then to clarify and drive the agenda on corruption appears to have been limited.

2.34 The lack of focus on the poor is also true of DFID’s programming in-country. Very few of DFID’s activities we reviewed in Nepal, Nigeria and elsewhere explicitly focus on the everyday corruption experienced by the poor and we found that limited efforts have been made to find ways to measure or tackle it. Programme logframes make limited reference to the reduction of corruption and, therefore, it does not form a key element of programme targets.⁵⁷ There also appears to be no focus on the impact of corruption on women and girls specifically.⁵⁸

2.35 While data on the perspective of the poor may be generated at a programme level, it is not always apparent that it has practical use nor that it is seen and acted upon. In Nepal, for example, DFID is only now starting to collate feedback received from stakeholders in a systematic way. Further, it is not yet clear whether the feedback collated in Nepal will be in a form that can or will be used to guide and improve current programmes or to design and implement more effective indirect or direct anti-corruption programmes. In Nigeria, systematic collation and dissemination of feedback is not yet undertaken.

3 Findings: Delivery

Delivery

Assessment: Green-Amber

3.1 In this section, we examine the mechanisms by which DFID seeks to reduce corruption at the global level and, through the application of its ACCF strategies, at the country level. We consider whether DFID’s delivery partners are appropriate and whether DFID has the internal capacity to support its programmes effectively.

DFID’s choice of delivery mechanisms is appropriately tailored to the context

3.2 As reflected in its current programmes and in predecessor and successor programmes in this area, DFID has worked to create a varied portfolio of activities at the global, national and sub-national levels to tackle corruption, largely indirectly. The mechanisms it uses to deliver those activities are flexible in accordance with the circumstances. Examples of its few direct efforts to tackle corruption in other countries are described in Figure 7 on page 17.

Working with, through and alongside government

3.3 At the global level, DFID has been focussed and active in developing and supporting a wide range of multilateral efforts related to tackling corruption, ranging from international transparency initiatives, to asset recovery efforts, to promoting law and policy that address corruption. Building consensus among political leaders on the importance of tackling corruption, through such multilateral efforts, helps to provide a solid basis on which to tackle it at a country level.

3.4 The extent to which these various efforts are being joined up by DFID in a co-ordinated effort to fight corruption was, however, not clear to us. Indeed, we did not see evidence that the portfolio of DFID’s global efforts is assessed for its overall impact on corruption or, with the exception of DFID’s efforts on EITI, evidence of co-ordination between these global efforts and efforts at the country-level.

3.5 Where appropriate in-country, DFID aligns itself with the institutions of its host governments, providing budget support either for the purpose of a specific item of expenditure or for general use by a particular ministry.⁵⁹ Where the relevant government institutions and systems do not yet exist or require strengthening, DFID may choose to provide technical assistance. In other instances, DFID supports civil society and other non-state actors, technically and financially, to empower them in seeking transparency and accountability from government.

3.6 For example, in Nigeria, DFID does not provide funding through government systems, based on the assessment that the fiduciary risk of misdirection of funds is too great. Instead, substantial funding is provided to delivery partners to provide technical assistance and capacity-building to government and civil society.

3.7 In Nepal, by contrast, DFID delivers its programming through a combination of different mechanisms. It provides funding to government, directly and through multi-donor arrangements (such as the World Bank’s Multi-Donor Trust Fund for Public Financial Management) and through CSOs. It also provides technical assistance and capacity-building to civil society and government.

3.8 DFID faces some difficult decisions in assessing whether and how much financial support to provide, rather than technical assistance, and whether to provide that financial support through government systems. We noted in our visit to Nigeria that there are arguably junctures at which reforms that tackle corruption begin to take hold; however, they may need direct support to enable them to flourish. By limiting the practical financial support it provides, DFID may risk the progress it has made to engender anti-corruption reform.

3.9 In such situations, we consider that DFID may be helped in its decision-making by careful use of stakeholder involvement. We heard from judges and police who were participants in the J4A programme in Nigeria, for example, that they required financial support to buy equipment (such as IT systems and police cars) to enable them to dispense justice in a swifter and more effective way, now increasingly possible due to DFID programmes. Also in Nigeria, the anti-corruption agencies supported by DFID need protection for their staff from security threats, as their staff are otherwise coerced into ceasing their investigations and prosecutions. DFID could consider providing the practical financial support itself or strongly petition the host government to do so.

The balance between work on strengthening government and on accountability is largely good

3.10 In implementing its indirect anti-corruption activities, DFID needs to find the right balance between strengthening government systems (supply-side activities) and increasing accountability (demand-side activities). We found that DFID had established a clear division of responsibility between its supply-side and demand-side programming in Nigeria. At the state level, DFID’s work with civil society has been undertaken by the State Accountability and Voice Initiative (SAVI), while its work with government institutions has been undertaken by the State Partnership for Accountability, Responsibility and Capability (SPARC).

3.11 This separation has been beneficial in some respects, enabling DFID to continue to work with civil society, even when a government is not interested in receiving donor-funded technical assistance. We saw instances where SPARC is frustrated in its work by an inability to engage with local government but progress can still be made by SAVI with civil society. The programmes are, however, often understood by government to be linked. Figure 6 on page 16 describes a situation where this has caused issues for DFID’s programmes.

3.12 At a Nigerian federal level, the Federal Public Administration Reform Programme (FEPAR) incorporates elements of technical assistance to government. As the only accountability programme at the federal level, it also supports civil society within the same programme. We noted that this merging of ‘demand’- and ‘supply’-side programming into a single programme led to limitations on its ability to work with civil society.

3.13 We observed a reasonable balance between supply- and demand-side delivery mechanisms in Nepal. In common with Nigeria, DFID Nepal engages in supply-side programming in the health and education sectors and through public financial management support. This programming has been complemented by demand-side, community-led programming that has long been a part of DFID’s work in Nepal as a result of the years of armed conflict and political turmoil in that country.

3.14 There are important interdependencies between supply- and demand-side activities which need to be recognised in portfolio planning and delivery. In our view, these must be tackled simultaneously to bring about the reduction of corruption in a society. In Nigeria, for example, J4A has components relating to the police, the judiciary and the anti-corruption agencies and SAVI has components related to the media, civil society and government. We found, however, that insufficient attention is paid to the links between these elements. For example, SAVI in Kaduna organised a workshop to educate people about the anti-corruption agencies but failed to capitalise on the opportunity to co-ordinate this event with J4A, which separately works to strengthen those agencies.

DFID makes good use of flexible programme structures

3.15 Where corruption is endemic, opportunities to tackle corruption may arise suddenly and exist only briefly. Thus, the ability to move swiftly to instigate change appears to be a very effective counterpart to the institution building that longer term programmes provide. In Nepal and Nigeria, we saw examples of DFID using such flexible programme structures to good effect in its anti-corruption efforts.

3.17 In Nigeria, DFID’s Facility for Oil Sector Transparency (FOSTER) provides an effective umbrella programme by which targeted support can swiftly be given to promote transparency and accountability in the oil sector. Under the programme, support has been given to the Nigerian Extractive Industry Transparency Initiative and also to CSOs to demand and make use of disclosed information.

3.18 As another example of opportunistic interventions, FEPAR in Nigeria has recently initiated an approach to enable swift support for identified issues, so that ‘quick wins’ can be achieved that can be publicised and built upon. FEPAR has not yet, however, demonstrated substantial impact and care will be needed to ensure that the interventions cumulatively reduce the opportunity for corruption.

Where there is the opportunity, DFID co-ordinates with other donors

3.19 At the global level, DFID contributes to many of the international initiatives on corruption and, therefore, co-ordinates with other donors in that context. At a country level, the extent of DFID’s co-ordination with other donors is influenced by the presence and engagement of those donors.

3.20 Nigeria has a relatively limited and fragmented donor presence. DFID, therefore, has little interaction with other donors. The principal co-ordination we observed among donors was in determining ways in which they can limit overlap between their programmes, by focussing on different states and areas of intervention.

3.21 For anti-corruption activities, we heard from donors that co-ordination is likely to be most effective when donors are aligned and using their individual and collective strength to tackle corruption.⁶⁰ In countries where the scale of the corruption challenge is high, DFID — as an acknowledged leader in the donor community — is well-placed to seek synergies with other donor agencies and entities, to expand existing anti-corruption efforts and drive the anti-corruption agenda.

Delivery of DFID’s programmes is sound

DFID’s delivery partners are effective

3.22 The delivery organisations we observed demonstrated competence and effectiveness in implementing DFID’s programmes. Frequent use is made of international consultants to manage programmes and deliver work but there is also involvement of local experts in programmes in Nigeria and Nepal. The use of local experts can be particularly important in engendering the trust and support of government officials, so that anti-corruption interventions are not seen as a foreign imposition. This benefit was particularly apparent to us in northern Nigeria.

Financial management of the programmes we reviewed appears relatively sound

3.23 We undertook a desk review of the most recent audited financial statements of the majority of the programmes in Nigeria that were the subject of our review and the one programme in Nepal (CSP) that is implemented through CSOs. The audited financial statements were each for the financial year ended 31 March 2012 or later. They all evidenced spending in line with our experience of the scope and content of the programmes. One programme received a qualified audit opinion⁶¹ on the basis that evidence could not be provided for all expenses. DFID has agreed measures with the programme implementers to address the issues which caused the qualification.

DFID’s support of enforcement against corruption has been limited in-country

3.24 In our 2011 anti-corruption review,⁶² we recommended that DFID support more robust law enforcement activity to build transparency and accountability. Both globally and in the UK, DFID has taken positive steps to address this recommendation, as discussed in paragraphs 2.4-2.6. DFID appears to have had notable success in supporting enforcement against corruption in the Caribbean (as described in Figure 7) but has had less success in Nepal and Nigeria.

3.25 In Nigeria, we were disappointed to find no sign that DFID is undertaking new initiatives, contrary to what we would have expected as a result of our 2011 recommendation. Part of the J4A programme works to promote more effective and accountable police services in Nigeria through developing and providing technical assistance to a limited number of Model Police Stations. This programme, however, predates our 2011 review and, whilst some activities on policing and anti-corruption have been scaled up, no new initiatives to promote transparency and accountability in law enforcement appear to have been undertaken since that review.

3.26 DFID is supporting the anti-corruption agencies in Nigeria through J4A. As our stakeholder surveys reflect — set out in Figures A3-A8 of Annex A2 — their stakeholders have perceived limited improvement in the success of the anti-corruption agencies in tackling corruption. In part, this could be as a result of the security challenges experienced by members of those agencies, described above. Our surveys do also indicate a significant improvement of stakeholders’ confidence in understanding the investigation and prosecution of corruption cases, as a result of J4A’s support.

3.28 Limited technical assistance is being provided by DFID to the CIAA and to the National Vigilance Centre in Nepal, an entity under the supervision of the Prime Minister of Nepal, whose mission is to create awareness of corruption and to promote activities that help to reduce corruption. This assistance to the CIAA and the National Vigilance Centre is provided through a public financial management programme but its effectiveness will not be measured until 2015.

Institutional memory and continuity is lacking

3.29 Tackling corruption is a multi-generational undertaking which requires strategic patience.⁶⁴ A strong institutional memory and consistency of approach, including across predecessor and successor programmes, are vital in achieving this. The rapid turnover of staff has the potential to be particularly detrimental to the effectiveness of anti-corruption interventions.

3.30 DFID has a restricted pool of dedicated specialists who understand corruption and can advise on ways in which to tackle it. In country offices, however, we found that the task of advising on corruption frequently fell to governance advisors who may have some anti-corruption expertise but whose wide scope of responsibilities preclude them from paying sufficient attention to the strategic approaches required to tackle corruption.

Sustained anti-corruption programming exists

3.31 In both Nigeria and Nepal we saw examples of programmes that were long-lived, having been extended or replaced by successor programmes. In Nigeria we saw, in the context of a relatively stable society, the opportunity being taken for over ten years to penetrate institutions of justice through J4A and institutions of state through SPARC. In each case, this provided the opportunity to identify reform-minded individuals whose support to tackle corruption could be harnessed in the future. In Nepal, long-term programmes, such as the Enabling State Programme and successive Community Support Programmes, have allowed DFID to contribute meaningfully to the political stabilisation of the country.

3.32 These examples suggest that DFID recognises that institutional change, particularly to reduce corruption, requires a stable and lengthy investment. Nonetheless, it is not apparent that DFID has, in either country, consciously assessed the role each programme plays in tackling corruption over the long term, nor sought effectively to use the duration of its programmes to persistently tackle issues of corruption in a continuous way, actively building on past successes and learning from past failures in the course of programme implementation.

Dedicated staff are vital in creating institutional memory

3.33 Stability among the staff of DFID’s implementing partners in the private sector and NGOs can compensate somewhat for the regular churn of DFID staff, such as that visible in Nigeria and Nepal, by bringing experience and sophisticated understanding of the local context and challenges. In Nigeria, we found that private sector contractors had frequently been in the country for an extended period and, in some instances, had overseen several iterations of the same programme or moved between DFID-funded programmes. It was not apparent, however, that the learning and understanding of private sector contractors was effectively acquired by DFID staff.⁶⁵ In addition, while some steps have been taken to deepen governance advisor skills and capabilities on corruption since our last report, there is no dedicated cadre, or centre of excellence, focussed on anti-corruption at DFID centrally and in its priority countries. This means there is no organisational entity to compensate for the deficiency in institutional memory.

3.34 Long service of local DFID staff also has the potential to provide the institutional memory that non-local DFID staff do not provide. It was more apparent to us in Nepal than in Nigeria that the experience of local staff was being harnessed.

4 Findings: Impact

Impact

Assessment: Amber-Red

4.1 This section examines the impact of DFID’s programmes, focussing particularly on our case study countries of Nigeria and Nepal. We consider impact in terms of the programmes’ own measures – which largely do not relate to corruption – and in terms of their impact on corruption as experienced by the poor. We also consider the impact of DFID’s overall portfolios in-country.

Global initiatives have had limited observable impact on corruption, while UK initiatives have done better

4.2 At the global level, our impression was that DFID has involved itself in most of the major international efforts that seek to tackle corruption. We acknowledge the importance of those efforts in raising the issue of corruption and seeking to tackle it at the global level. We observed, however, that neither DFID, nor the international community more generally, has an understanding of what impact these efforts may be having on broader governance or anti-corruption metrics or on the poor. DFID staff told us that they consider the impact of their work from the perspective of intended beneficiaries, as reflected in the logframe outputs. These outputs are largely focussed, however, on the country level, so the links between DFID’s efforts and intended beneficiaries are highly attenuated. DFID has commissioned five impact assessments of transparency and accountability in the extractives sector as part of a new programme approved in July 2014. There is not yet, however, any directly measured impact on the poor.

4.3 We did see points of connection between the initiatives DFID is supporting at a global level and its priority countries. Nepal and Nigeria, for example, are both signatories to the UN Convention Against Corruption. In Nigeria, DFID is supporting the Nigerian Extractive Industries Transparency Initiative through the FOSTER programme. We found little indication, however, that domestic implementation of global initiatives is making a measurable difference on how these countries are tackling corruption.

4.4 The impact of the enforcement on financial crimes in the UK is more readily determined. Since 2006, the Metropolitan Police Proceeds of Corruption Unit has conducted eight successful prosecutions; approximately £120 million in illicit assets are currently being held, awaiting eventual repatriation. There are, however, limitations and risks to this approach, as discussed in paragraph 2.5. The City of London Police Overseas Anti-Corruption Unit has had a dozen successful prosecutions, with 25 cases currently under investigation. DFID supports both of these efforts.

Some of DFID’s efforts in-country have had modest impact

4.5 Very few of the programmes we reviewed directly measure their impact in reducing corruption. Only J4A in Nigeria includes explicit anti-corruption measures in its logframe.⁶⁶ We, therefore, consider in this section those instances where outputs of the programmes appeared to address issues of corruption through indirect means.

4.6 The programmes we reviewed in Nigeria and Nepal evidence modest achievements in delivering against their objectives. We noted, however, that the outputs contained in the programme logframes, business cases and project evaluations are general in nature. For example, SPARC and SAVI in Nigeria and the Nepal Health Sector Programme and School Sector Reform Programme include, in their logframes, goals related to achievement of the Millennium Development Goals. By measuring themselves against broad targets that do not relate to corruption, we feel that these programmes undervalue their actual successes.

Public financial management reforms have had some successes

4.7 Programmes have achieved positive steps, leading to fewer opportunities for funds to be improperly extracted. For example, in Kano State in Nigeria, the Treasury Department has closed 664 out of 756 accounts, initially with the support of DFID and then independently. This prevented leakage of approximately £18 million in 2010.

4.8 Through the work of the World Bank’s Multi-Donor Trust Fund, a treasury single account has been established in Nepal and approximately 13,717 out of 14,000 accounts have been closed in government offices. This has enabled increased control and more transparent monitoring of revenue collection and expenditure. The amount of money protected from leakage has not yet been quantified.

4.9 In Nigeria, SPARC has supported ten states in improving their financial management and developing and implementing state economic plans. It made particular progress with the State of Jigawa, where the incumbent Governor was receptive to the programme. At a Nigerian federal level, through FEPAR, CSOs have been provided with grants to stimulate demand for accountability. Progress has also been made in establishing positive relationships with civil servants to form a basis for future work.

4.10 Our interviews suggested that many of the public financial management reforms are hostage to political changes. In Jigawa and Enugu, we were told that the impending election of new governors may lead to a reversal of some of the reforms that have been implemented. SPARC, therefore, has worked to embed reforms in the systems of government where DFID hopes they will be difficult to displace.

4.11 During the period of conflict and unrest in Nepal and in response to the political power vacuum that followed, the Government of Nepal, through the Local Government and Community Development Programme (LGCDP), successfully introduced local representative bodies in the country. Ward Citizen Forums and Citizen Awareness Centres, covering 97.5% of the country, have provided communities with local representatives to influence spending decisions of local government on infrastructure projects.

Civil society programmes, particularly with the media, have had positive impact

4.12 We found community implementation and monitoring of projects, under the guidance of NGOs, to be an effective way of tackling corruption in infrastructure projects.⁶⁷ CSP in Nepal, for example, has helped to create transparency and accountability to communities through the project cycle by posting project budgets and related public audits on signs visible to the community. Communities told us they benefited and learned from the approach taken by CSP.

4.13 Some progress has been made in raising awareness on citizens’ entitlements. In Nepal, the small-scale pilot of the Programme for Accountability in Nepal (PRAN), which is funded by the World Bank’s Multi-Donor Trust Fund, appears to have been successful in raising awareness and activity on the allocation of social security payments. In Nigeria, SAVI has also worked to increase budget awareness. We saw a number of organisations in Nigeria lobbying for budget allocations for certain interest groups and, as reflected in the results from our stakeholder surveys in Kaduna State, set out in Figure A10 of Annex A2, we found more debate about the budget among media and communities.

4.14 The impact of ‘demand-side’ accountability appeared to be particularly strong where civil society and media work together. Success was more likely where the media persist in investigating and publicising one incidence of corruption, rather than dropping it in favour of a new scandal, as is frequently the case in Nigeria. We heard of instances with SAVI where media coverage led to the exposure of corruption in infrastructure projects in Lagos and the punishment of judicial officials in Kano. Of course, care must be taken to ensure that the media is not politicised and is not selectively publicising allegations of corruption.

4.15 We found occasions where the media widely publicised their successes in tackling corruption – on the advice of SAVI – and we heard this gave civil society the confidence that corruption can effectively be challenged.⁶⁸ Where the sense of impunity is pervasive, people increasingly perceive that corruption can be indulged without retribution. Publicising success stories of tackling corruption can have a deterrent effect. It can also help the populace to understand that acting in a non-corrupt way is a viable option.

Justice reforms have also had some success

4.16 Some impact has been achieved in the justice sector in Nigeria, which is reflected in our stakeholder surveys as described in Figures A3-A7 in Annex A2. The prosecutors who received training from J4A overwhelmingly felt that their capacity to prosecute corruption cases had improved. The same was true for investigators who received training on how to investigate corruption. The judges who hear corruption cases observed that the quality of evidence at their trials had improved. Citizens with whom we spoke and SAVI stakeholders whom we surveyed indicated, however — as described in Figure A11 in Annex A2 — that the anti-corruption agencies are not perceived as being more effective.

4.17 DFID programmes have also benefitted from the political support they have received from local government. For instance, DFID’s justice sector reform efforts in Jigawa State were supported by the State Governor there; as a result, DFID was able to help the state to establish the Justice Sector and Law Reform Commission, based on a model provided by J4A. In view of successes like this, the Jigawa State Government has provided approximately £62,000 to co-fund justice sector reform.

Some activities have increased the opportunities for corruption or have had no impact on corruption

In Nepal, we saw a programme increasing the opportunities for corruption

4.18 Disappointingly, we found that at least one programme supported by DFID appears to have increased the opportunities for corruption in society. Instances, such as those described below, raise concerns that the general principle that aid should first ‘do no harm’⁶⁹ has been breached.

4.19 In our view, it is highly problematic for DFID to support government systems and structures that are known to be corrupt, especially if they increase the perception among the poor that corruption is a necessary and legitimate way of acting. Choosing to proceed in such circumstances clearly risks breaching the principle that development activity should first do no harm. Before proceeding with such support, DFID and the FCO should redress the balance of their relationship with the host government to attempt to ensure that DFID-supported activities will not increase the risk of corruption, especially among intended beneficiaries.

4.20 In Nepal, we heard from citizens that the Ward Citizen Forums, established by LGCDP⁷⁰ to be locally representative bodies, remain subject to political capture. Political elites are often able to use their status to influence the direction of government funds towards their preferred projects. As reflected in our stakeholder surveys shown in Figures A20 and A24 in Annex A2, more than 70% of citizens surveyed and more than 60% of Village Development Committee Secretaries surveyed believe that without influential community members, they are unlikely to receive local government funding.

4.21 We were told by DFID in Nepal and multiple stakeholders that the Ministry of Federal Affairs and Local Development, which is responsible for community projects in regional areas, is one of the ministries in Nepal most afflicted by corruption through some of its staff. Indeed, concerns regarding the implementation of projects by local government have been highlighted in the international media and publicly responded to by DFID.⁷¹

4.22 To mitigate the risks of corruption in government systems, LGCDP decided to restrict funding to local government bodies unless they could show that certain minimum conditions were met. This meant, however, that some local government bodies that most needed funds were unable to access them. Also, we were told by many citizens and other stakeholders that some government officials produce false documentation or documentation promising compliance with the minimum conditions that officials know will not be complied with in practice. Donors, including DFID, seem to be aware of these issues. It has led to a proposal that LGCDP’s successor programme should require the minimum conditions to be results-based.

4.23 We were particularly concerned to hear of the way in which the poor can be pushed towards corrupt practices through government projects such as those supported by LGCDP. We heard about citizens needing to pay bribes to government officials or to forge documentation in order to receive funding for projects supported by LGCDP, as is demonstrated through our stakeholder survey (reported in Figure 8) and our findings on issues with government projects (described in Figure 5 on page 12). This issue particularly arises with government projects, where delays in budgeting and procurement processes lead to funds being released late in the fiscal year. It is, therefore, of little surprise that only 32% of citizens we surveyed prefer to work with government-funded projects than projects funded by donors or NGOs.⁷²

4.24 DFID and other donors were aware that funds were frequently released as late as 11 months into the fiscal year, when only one month, during monsoon rains, remained for citizens to undertake and complete projects. Whilst DFID and other donors may not have known of the specific knock-on effects of citizens paying bribes and forging documentation, they knew the late release of funds placed citizens in untenable positions for completing their projects. More frequent and direct interaction between DFID and the stakeholders of its programmes could have ensured that DFID was aware of the impact of lapses in project funding.

In Nigeria, we challenged the belief of successful impact

4.25 J4A has worked to establish Model Police Stations, in which police have been trained and empowered in community engagement. In Kano State, staff at the Model Police Station at Sabon Gari and the area’s voluntary police service each told us that the police had reformed significantly, ameliorating corruption and bad practice. This was confirmed by experienced DFID programme staff. According to J4A perception surveys in Lagos and Enugu, after the commencement of MPS work in the selected sites, satisfaction with police services rose by 16%.

4.26 To corroborate J4A’s impact, we undertook a survey of 1,500 citizens, 300 around each of five police stations. Unfortunately, we found not only a continuing lack of trust in the police but also a perceived increase in the experience of requests for bribes by the police before they will help citizens, as indicated in Figure 9 on page 24. Issues of violence and corruption remain very high. There is no clear difference between Model Police Stations and others. The results in Kano State are as bad as — if not worse than — other police stations. Our further findings are set out in Annex A2.

4.27 Our findings that the Model Police Station in Kano does not appear to have addressed issues of trust and corruption provide a good basis for rapidly identifying ways for correcting the programme. This should be tackled before the model is considered for expansion to more areas.

Planning around interim outputs and outcomes is beneficial

4.28 A number of the programmes we reviewed had clearly made some progress in reducing corruption or, at least, in creating an environment where DFID can take the next steps to tackle corruption. This includes space to discuss corruption openly and the fostering of individuals who have the power and will to tackle corruption.

4.29 In line with this, SAVI in Nigeria articulates a theory of change which focusses on the receptiveness of states to change. The stages of development it recognises include ensuring that advocacy for reform is under way; that individual capacity and relationships are under development; that a constructive supply and demand discussion has been initiated; that successes are achieved and replicated; and that a critical mass of reforms is identified, developed, pushed forward, accepted and embedded.

4.30 In our view, the intermediate goals and the tailored activities that need to be undertaken to achieve progress as envisaged in SAVI’s theory of change are realistic and reflect the positive achievements — such as that set forth in Figure 6 on page 16 — we saw on the ground. The SAVI framework, however, is not clearly mirrored in the assessment of performance set out in the logframe or elsewhere. Whilst SAVI has gathered some survey data on citizen perceptions of accountability, it is unclear whether this is linked back to the theory of change, used to course-correct the programme, or followed up with additional survey work. The potential benefit of this tailored and realistic framework is, therefore, lost.

Limited visibility on value for money

4.31 Despite the significant amounts of money that DFID is spending on its anti-corruption efforts, we conclude that its expenditure does not appear to achieve significant impact in reducing corruption. As set out in the discussion above, in some instances, programmes do not address their negative impact on corruption and for others DFID is unaware that they are not reducing corruption at all. In other instances, DFID does not build on its earlier successes and does not indicate a consistent direction of travel, such as with the interventions with the anti-corruption agencies in Nepal. The failure by the Government of Nigeria or another donor to adopt J4A’s Model Police Stations is also reflective of the limited success of that programme.

4.32 Further, programmes indicate a mixed ability to assess their own value for money. Some have had a large impact for a relatively limited expenditure (including support by FOSTER of advocacy in the oil sector and the reduction of bank accounts by SPARC in Kano State). Others, such as the Enabling State Programme in Nepal, pragmatically state that it is not possible accurately to monetise their benefits due to attribution issues and the time-lag for effects to materialise. On their own terms and assessing them individually, it is, therefore, difficult to say whether these interventions represent value for money.

4.33 Viewed as a portfolio, DFID’s programming that targets corruption has not demonstrated impact greater than the sum of its parts. So it is difficult to conclude that the portfolio of programmes we saw represents value for money. The overall impression is that DFID is undertaking a range of independent activities which are producing little lasting reduction in corruption, exemplified by the limited co-ordination between programmes at the global level; between global-level programmes and country-level programmes; and between programmes in-country, such as the state-level programmes in Nigeria.

5 Findings: Learning

Learning

Assessment: Red

5.1 This section considers whether DFID is undertaking appropriate learning to inform its anti-corruption efforts. It assesses DFID’s central knowledge sharing and the level of adaptation in DFID programmes in response to lessons learned.

The quality and quantity of the evidence base is poor

5.2 There is a lack of research on the effectiveness and impact of anti-corruption interventions. DFID has a limited number of dedicated staff advising on corruption. Regional advisors on corruption have spent more time advising on fraud risks to programmes than originally envisaged and their influence on anti-corruption policy and action has been limited. This has been influenced by Cx3 focussing its attention on tackling fraud. As almost all of DFID’s programmes seeking to tackle corruption in its priority countries do so indirectly, limited efforts have been made to learn of their effect on corruption. Thus, it is difficult for DFID to justify the large amounts of money spent and forecast on programming to tackle corruption indirectly.⁷⁵

Evidence of effectiveness of anti-corruption efforts is limited

5.3 DFID has taken some steps to address the gap in academic evidence. For example, it was one of the founders of the U4 Anti-Corruption Resource Centre,⁷⁶ which runs a website that collates and disseminates knowledge on efforts to tackle corruption. DFID has also funded research papers in 2011⁷⁷ and 2012⁷⁸ to ascertain better the evidence base for effective anti-corruption programming.

5.4 Very recently — in late 2013 — a further research paper has been commissioned by DFID on the effects of corruption and measures to tackle it and the paper is currently being finalised.⁷⁹ DFID is also developing a £10.5 million anti-corruption research programme which will aim to address further the gap in the evidence base for anti-corruption interventions and their inter-dependence.

5.5 Each of the research papers, commissioned by DFID, has shown that evidence of the effectiveness of anti-corruption interventions is weak in terms of quality and quantity. Where the evidence is strong, the findings are frequently contested as to how effective that programming is. We found that there is also very little focus on the poor in these research papers which, instead, concentrate on the standard range of interventions employed to tackle corruption at a macro-level.

5.6 DFID is also aware that there is little research that documents the interdependencies among anti-corruption interventions, host government institutions and different sectors or the ways in which different interventions can be co-ordinated so that they work to tackle corruption holistically. There is, therefore, little to guide what needs to be done in parallel or in sequence to ensure that DFID’s efforts to tackle corruption are greater than the sum of their parts.

5.7 Internally, DFID has developed its capacity to advise on how best to tackle corruption but still has insufficient expert resources in this area. The core anti-corruption team has had a maximum of five centrally located staff supplemented, for a period of two years, with four specialised anti-corruption country support regional advisors. DFID also has a growing cadre of governance advisors for whom corruption is one of several requested technical competencies. To support them DFID has, since the previous review, appointed only a part-time senior research fellow to advise on corruption and, also for a period of two years, three specialised regional anti-corruption advisors.

Learning from in-country analysis is important

5.8 Corruption is rightly considered by DFID to be a difficult phenomenon to observe and assess. Much corrupt activity is hidden and participants in it are resistant to being studied. This is exacerbated by the intangible nature of some types of corruption, such as patronage, abuse of power and conflict of interest.

5.9 In the absence of strong global evidence on the effectiveness of anti-corruption efforts, we would expect DFID to make extensive efforts of its own to understand corruption in its priority countries and to justify the anti-corruption measures it is undertaking. We noted in our review of How DFID Learns that learning during implementation is substantially different from learning from research and from traditional evaluations but that it is a vitally important tool.⁸⁰ We found, however, that DFID’s learning for improvement of its anti-corruption work globally or in-country is quite limited. Indeed, much of DFID’s analysis of corruption is based on limited political economy analyses and fiduciary risk assessments, which are not often updated. DFID also uses general perception indices, such as the Corruption Perception Index, produced by Transparency International.⁸¹

DFID’s country context analysis is limited

5.10 The nature of the challenge that corruption presents in a country is strongly influenced by the political, geographic, social and economic factors in which corruption occurs. In order to tackle corruption effectively, it is vital to design interventions that are specific to the context.

5.11 In Nepal, many different stakeholders told us of the different challenges that corruption presents in the southern Terai region, bordering with India, than in the central hill or northern mountainous regions. In Nigeria, issues of elite capture arise across the country but there are additional challenges in the South, where corruption arises in relation to the extraction of oil. The willingness of individual Nigerian State Governors to engage in reform is also a significant factor. The ACCF strategies do not evidence action DFID may have taken to understand and tackle these variances in corruption and its drivers within a country.

5.12 We also found little evidence of DFID working to share information about corruption and anti-corruption strategies among the donor community in a particular country. In Nepal, the donor community has only recently formed an anti-corruption sub-committee, under the International Development Partners Group, which meets to discuss political and development themes to be taken up with the Government of Nepal. We did not observe activity of this sort of strategic thinking and action in Nigeria, despite the existence of a justice sector donor coordination group.

Programmes have made limited attempts to analyse their anti-corruption effectiveness

5.13 As the vast majority of DFID’s programmes to tackle corruption do not make explicit their anti-corruption intent, it is perhaps not surprising that DFID undertakes relatively little activity to measure the reduction in corruption achieved as a result of its efforts, either at a macro country-wide level or at the more detailed programme level. The focus of measurement is on activities directly undertaken by the programme.

5.14 We were able to identify some progress by programmes in tackling corruption, described in Figure 10 on page 28. We are of the view that these sorts of interim achievements can and should be assessed and can form the basis for intermediate goals and funding decisions for future programming.

5.15 DFID now seems to recognise that limitations in its internal coding system make it difficult to obtain a clear picture of its overall spending on anti-corruption. It is currently considering alternative approaches to coding and categorising its anti-corruption activities.

Dynamic use of learning from intended beneficiaries and other stakeholders would allow ‘course correction’

5.16 Tackling corruption requires a combination of patient, planned approaches over a long period and opportunistic interventions when an enabling environment suddenly appears. Each of these approaches requires continual evidence of programme performance, on the basis of which corrections can be made to programmes and short-term activities can be undertaken.

5.18 Continual stakeholder monitoring also provides opportunities to track spending on projects effectively. In Nepal, beneficiary monitoring is used in a number of programmes. In CSP, we heard of very positive experiences of communities undertaking audits to reduce the misuse of funds.

5.19 While we saw that feedback obtained through DFID’s surveys is used in the design of new programmes, we did not see evidence that feedback obtained in the course of programmes actively informs DFID’s evaluation of the impact of its anti-corruption activities or is used to course-correct existing programming. There was limited evidence that successes and failures in tackling corruption from predecessor programmes were carried through to subsequent programmes.

5.21 The fiduciary risk reduction action plan has not yet, however, been fully implemented and the survey, which was originally stated to be an annual exercise, will now be repeated only every two or three years, to allow for changes to take effect and be measurable. Further, while public expenditure tracking surveys are now stated to be part of the successor programme to LGCDP, delays in the implementation of other public financial management programmes in Nepal have meant that no such additional surveys have yet been undertaken. It remains to be seen whether these projects will be carried out and used to inform programming.

5.22 On the basis of the surveys we undertook for this review set out in Annex A2, we found that, by using small-scale, frequent, inexpensive surveys, it is possible to bolster the observed experiences of programme staff and lend additional evidence of the effectiveness of programmes. Where results of those surveys challenge the prevailing thinking about the successes or benefits of an intervention, they can be used as the basis for further investigation or course-correction for aspects of the programmes or even entire programmes.

Sharing of learning between countries and DFID centrally should be improved

5.23 In the absence of extensive evidence of the effectiveness of its anti-corruption efforts, it is vital that DFID shares and learns from the evidence that emerges from its various interventions. We found, however, that after more than two years of Cx3’s existence, DFID has undertaken very limited efforts to learn lessons about anti-corruption activities from country offices or from its global initiatives.

5.24 As part of its global Evidence and Programme Exchange initiative, DFID does have an internal website on anti-corruption, which has been running for over a year and operates as a repository to access evidence on corruption and on anti-corruption interventions as well as examples of programmes implemented by field offices. It is not, however, accessible to non-DFID programme staff who are at the front line of many programmes that seek to tackle corruption.

5.25 The internal website provides some literature and research on corruption and information on a selection of programmes. It does not, however, contain a holistic set of DFID’s programmes that tackle corruption. We could not form a complete view as to how much the site is used by DFID staff, as basic site statistics are the only method of determining who is viewing information on it. Its existence was not mentioned in our interviews with DFID staff in-country.

5.26 The lack of a knowledge management system that is accessible and widely utilised by DFID and programme personnel regarding its anti-corruption efforts represents a missed opportunity for learning.

5.27 The appointment of regional anti-corruption advisors, following our 2011 report, should have provided an opportunity for DFID to develop centres of excellence. The advisors appear, however, to have focussed their efforts on advising country offices on the formulation of ACCF strategies and on tackling fraud in the delivery of DFID programmes rather than on anti-corruption. The roles were only intended to be for two years and are now winding down. The posts of anti-corruption advisor for West Africa and for Middle East and North Africa have been vacant for the entirety of this anti-corruption review.

5.28 In these two years, the advisors’ ability to fulfil a remit beyond the development of the ACCF strategies has been limited. In a few instances, they have been called on by country offices to advise on the development of new anti-corruption efforts. The full extent of advisors’ expertise and skills, therefore, has been underutilised during their tenure. Had they been seen and used as a cadre of experts, responsible for advising on corruption, their impact may have been greater.

5.29 As things stand, DFID’s learning on corruption remains limited and has not been used effectively. This is disappointing, given the recommendations in our 2011 report, the implementation of the ACCF strategies in January 2013 and Cx3 now having been in existence for more than two years. The presence of this co-ordinating body also appears to have had little success in leading learning, both within DFID and with other departments of the UK Government, on how best to tackle corruption as it touches the UK and developing countries.

6 Conclusions and Recommendations

Conclusions

6.1 DFID recognises corruption in the countries where it operates as a critical development challenge which affects the everyday lives of the poor, restricts foreign investment and thwarts efforts to lift poor countries out of aid dependency. DFID, however, has not developed an approach to anti-corruption that is equal to this challenge.

6.2 DFID funds programmes that aim to constrict the environment where corruption flourishes, such as programmes that improve the effectiveness of the police and judiciary or local government institutions. Research on the effectiveness of anti-corruption interventions suggests that some of this programming — particularly programmes supporting improvements in public financial management — can have some impact on the corruption environment, although this is not systemic. We found, however, little evidence of impact on corruption from the majority of DFID’s anti-corruption activities.

6.3 Some of DFID’s programmes that touch on corruption have produced achievements, even if those achievements have not resulted in proven reduced levels of corruption. That said, many of the programmes have objectives and evaluation criteria that rarely relate to measures of corruption. If corruption is not adequately measured, programme evaluations can be unrepresentative and programme trajectories can be distorted. Further, DFID has not developed a systematic approach to identifying the corruption challenges it wishes to confront or to understanding how its anti-corruption work might benefit the poor.

6.4 We accept that the politicised nature of corruption makes action in this area difficult. We found, however, that DFID’s anti-corruption efforts can be constrained by its relationships with host governments that often fail to take responsibility for the corruption challenges in their own countries.

6.5 The specific purpose of this review is to focus on DFID’s work to reduce the negative effects of corruption as experienced by the poor. We saw very little evidence that the work DFID is doing to combat corruption is successfully addressing the impact of corruption as experienced by the poor. Indeed, there is little indication that DFID has sought to address the forms of corruption that most directly affect the poor, so called ‘petty’ corruption. This is a gap in DFID’s programming that needs to be filled.

6.6 DFID is involved in a range of global anti-corruption initiatives and UK-focussed efforts to mitigate corruption. We found, however, that little evidence has been sought or obtained by DFID to establish that these efforts have had or will have an impact on corruption as experienced by the poor. We did not see evidence that the impact of these efforts is being enhanced by co-ordination of them with DFID’s in-country anti-corruption portfolios.

6.7 This review also undertook a limited follow-up to our earlier report on anti-corruption. We conclude that DFID has made positive strides in preparing a coherent response to the counter-fraud elements of that review. It has developed the ACCF strategies at the country level and has created Cx3. We note, however, that the emphasis in these efforts, to date, has largely been on counter-fraud. The ACCF strategies at the country level mainly set out a description of the existing programming to tackle corruption. Very little progress has been made so far on developing a vision for tackling corruption in its priority countries. In particular, DFID’s medium- and long-term goals and ambition to tackle corruption, including those areas DFID proposes to tackle in the future, are missing from the strategies.

6.8 DFID uses a variety of mechanisms to deliver its anti-corruption efforts. Delivery organisations are generally of high quality and the programmes are run by personnel with substantial in-country and sector experience. We found, however, that DFID lacks experienced in-country personnel in some critical roles, particularly in terms of the co-ordination of the full portfolio of anti-corruption programmes. Also, while DFID has taken a role in tackling corruption in some of its priority countries, more could be done to leverage senior personnel and to co-ordinate well with other departments in the UK Government.

6.9 DFID’s learning in this area has been particularly weak. It is hampered by the general absence of clear research evidence on what works and does not work in anti-corruption and therefore has little understanding of what is and is not working with respect to its anti-corruption activities. Further, DFID does not fully understand which of its activities are even addressing corruption. This is not helped, as a result of very restricted definitions in DFID’s finance system, by DFID not knowing how much it is spending on anti-corruption overall. Cx3, to date, has focussed very little on anti-corruption.

6.10 Central learning on effective programming has not been matched by strong analysis of corruption and appropriate interventions tailored to specific country contexts. Anti-corruption activities frequently span multiple programmes but worthwhile achievements are not appropriately recognised in predecessor programmes and not sufficiently built upon in successor programmes. Cross-country and in-country learning is weak, where it exists at all.

6.11 We also witnessed negative impacts of programming, where government systems that lead the poor to have to pay bribes and become the victims of corruption were perpetuated and not tackled by programmes. DFID and other donors knew — or should have known, given the media coverage — of these negative consequences for the poor. Instances, such as those in Nepal in the context of LGCDP and CSP, raise concerns that the general principle that aid should first ‘do no harm’ has been breached. We also noted instances where DFID did not seem to realise that its programmes were simply not reducing corruption as experienced by the poor. We were concerned that DFID had not yet established a basis for scaling up its programmes, as was the case with the Model Police Stations in Nigeria.

6.12 It is clear that the will for change on the part of host country governments impacts DFID’s ability to act against corruption. One rational response to such a context would be to give up fighting corruption altogether and to concentrate, instead, on issues more susceptible to immediate progress. In addition to the moral issues such an approach would raise, the development community acknowledges that corruption negatively affects the wider environment into which all DFID’s other development efforts flow. As a result, successful action to reduce corruption will help to increase the effectiveness of DFID’s aid more generally.

6.13 Indeed, a more joined up approach by DFID and the Foreign and Commonwealth Office — one which works to deploy all the foreign policy levers available to the UK Government — may present the UK with an opportunity to develop an approach to tackling corruption that is better aligned with the complex, multi-faceted nature of the challenge that corruption presents. This would allow the UK to further establish itself as a leader in the global effort to fight corruption.

Recommendations

6.14 This section contains formal recommendations for DFID.

Recommendation 1: DFID, in conjunction with the FCO and other UK Government departments, should articulate and implement a detailed plan setting out the level of ambition, commitment and positioning of the UK with respect to tackling corruption in its priority countries, including as experienced by the poor.

6.15 Corruption presents a challenge and an opportunity for DFID and for the UK Government more generally. Because corruption is a development challenge and a political challenge, it requires a response that brings together the full range of a country’s foreign policy influence.

6.16 Were the UK to take an ambitious stance with respect to tackling corruption around the world as experienced by the poor, DFID could develop an approach to fighting corruption that would be an integral part of the UK’s wider efforts. DFID, in partnership with the FCO, could be a beacon for anti-corruption efforts in the development sector.

Recommendation 2: DFID should develop standalone anti-corruption country strategies and, in addition to its current activities, programming that explicitly tackles corruption and that extends over a 10- to 15-year time horizon with short-, medium- and long-term goals for reducing corruption, particularly with respect to the poor.

6.17 Given the scale of the corruption challenge and the extended timeframe in which change against corruption can be realistically effected, DFID should begin to conceive of its anti-corruption programming in 10- to 15-year timespans. In doing so, programmes should be designed with a clear set of realistic intermediate goals and interim outputs, upon which funding decisions can be based and achievements can be assessed and built upon. This will allow DFID to plot better the progress of its programming in alignment with the nature of the challenge. DFID should also explicitly identify the challenges of corruption more clearly in its core guidance on the use of the country poverty reduction diagnostic and the diagrammatic system of poverty reduction.

6.18 In the short term, DFID should seek opportunistically to arrest the spread of corruption by focussing on targeted interventions into areas susceptible to change, with a particular focus on enforcement where possible. In the medium term, it should undertake direct anti-corruption programming that strategically targets the linkages between host government institutions and systems and the reduction of corruption. Such programming should also seek to provide moral and practical support for indigenous anti-corruption efforts. In the long term, DFID should ambitiously seek to produce changes in attitudes with respect to corruption, as discussed in Recommendation 3.

6.19 To help combat widespread forms of corruption and the endemic corruption that affects people’s everyday lives, such programming should leverage emerging technologies and social networks that have a possibility to bring about behaviour change on a mass scale.

6.20 DFID currently has combined anti-corruption and counter-fraud strategies. To date, these strategies have led DFID to place greater emphasis on counter-fraud and the protection of UK taxpayers’ money in the context of its programming. They have not led DFID to increase or refine its focus on tackling corruption in-country.

6.21 To rectify this and to help to ensure that these strategies contribute to DFID’s efforts to tackle corruption, DFID should develop standalone anti-corruption country strategies. These should be developed from the same comprehensive contextual and political economy analysis as is used to tackle fraud. The strategies to tackle corruption should be complete and comprehensive to identify issues of corruption in-country and methods by which to tackle them. DFID should ensure that these strategies are dynamic, context-specific documents that are updated rapidly and as necessary to reflect particular corruption challenges and to highlight emergent opportunities at the national and sub-national level. These strategies should drive DFID’s anti-corruption programming.

Recommendation 3: DFID should include in its expanded anti-corruption portfolio many more programmes which specifically target the everyday corruption experienced by the poor and educate the population about the ill effects of corruption.

6.22 Our surveys with intended beneficiaries and other citizens, as well as our direct contacts with many DFID programme stakeholders and government officials, demonstrated to us a strong, pent-up desire for aid programmes directly to target corruption in their everyday lives. This could include programming that trains media to investigate and expose the unique effects of corruption on the poor, whether the corruption occurs at the national, state, local or community level, or that facilitates the reporting of, and taking action on, specific incidences of corruption experienced by the poor.

6.23 The poor have become the victims of corruption to such an extent that they are habituated to paying bribes to go about their daily lives. Given the pervasiveness of corruption throughout these societies, additional steps are required to seek to change perceptions of the ill effects of corruption. It was particularly apparent to us that there is a desire to educate young people about the ill effects of corruption, so that they clearly understand that it is a form of theft from society, and to support them in their efforts to resist and fight back against corruption. This sort of education is part of building a self-initiated cultural shift away from corruption. DFID has done little to develop such programming, and as a result, has missed an important opportunity.

Recommendation 4: DFID should gather and publish targeted and dynamic feedback from the stakeholders of its anti-corruption work, including the intended beneficiaries, to allow DFID to ‘spot check’ and correct its existing programmes and to inform new programming.

6.24 In the face of widespread, deeply entrenched challenges, such as corruption, the absence of real-time evidence that would enable iterative course correction has limited the effectiveness of DFID’s efforts. DFID, therefore, should engage regularly with all the stakeholders of its anti-corruption work, including intended beneficiaries, to understand better the impact its anti-corruption activities are having. DFID should publish that information to foster the empowerment of citizens.

6.25 By utilising continuous, targeted micro-surveys, and responding to the data, DFID could transform feedback from the poor and other stakeholders into a system of learning and performance management, as described in further detail in Annex A2. This would enable it to check how well its existing programmes are operating and undertake iterative corrective actions to them, as required. Those micro-surveys would complement DFID’s existing approach to obtaining feedback and would also allow DFID to generate ideas for the development of new anti-corruption programming and their publication will foster accountability to and empowerment of citizens.

Recommendation 5: DFID should create an internal embedded centre of excellence explicitly to focus on anti-corruption and to gather evidence of effectiveness, disseminate lessons learned and cultivate expertise that will drive anti-corruption efforts globally.

6.26 DFID does not currently have an embedded organisational entity explicitly focussed on anti-corruption efforts in-country. As a result, anti-corruption work is often not prioritised. This also limits DFID’s ability to gather intelligence and evidence of effectiveness with respect to anti-corruption and negatively affects DFID’s ability to learn, as well as the cultivation of organisational memory.

6.27 Such an entity should also liaise with other donors to continue to gather and share anti-corruption learning from other sources.

Annex

This Annex provides more detailed background information to the review. It comprises the following:

1. Programmes reviewed and country maps (Annex A1);
2. Stakeholder feedback (Annex A2);
3. Bibliography (Annex A3); and
4. List of consultations (Annex A4).

Abbreviations

• ACCF – Anti-Corruption and Counter Fraud
• Cx3 – Counter Corruption and Counter Fraud Group
• Cx3(o) – Operational Group of Cx3
• CIAA – Commission for the Investigation of the Abuse of Authority
• CPRD – Country Poverty Reduction Diagnostic
• CSO – Civil society organisation
• CSP – Community Support Programme
• DDC – District Development Committee
• DFID – Department for International Development
• ESP – Enabling State Programme
• FCO – Foreign and Commonwealth Office
• FEPAR – Federal Public Administration Reform Programme
• FOSTER – Facility for Oil Sector Transparency
• ICAI – Independent Commission for Aid Impact
• J4A – Justice For All
• LGCDP – Local Government and Community Development Programme
• MPS – Model Police Stations
• NGO – Non-governmental organisation
• NHSP – Nepal Health Sector Programme
• OECD – Organisation for Economic Co-operation and Development
• PETS – Public Expenditure Tracking Survey
• PRAN – Programme for Accountability in Nepal
• SAVI – State Accountability and Voice Initiative
• SPARC – State Partnership for Accountability, Responsibility and Capability
• SSRP – School Sector Reform Programme
• UK – United Kingdom
• UN – United Nations
• UNCAC – United Nations Convention Against Corruption
• VDC – Village Development Committee
• WCF – Ward Citizen Forum

Footnotes

1. ICAI, The Department for International Development’s Approach to Anti-Corruption, November 2011, http://icai.independent.gov.uk/wp-content/uploads/2010/11/DFIDs-Approach-to-Anti-Corruption.pdf.
2. We – as does DFID – use Transparency International’s definition of corruption as ‘the abuse of entrusted power for private gain’. Hence, corruption is not limited to activities involving a government official.
3. Annex A1 describes the programmes we reviewed in detail for this report and also includes maps of Nigeria and Nepal, showing where we focussed our review.
4. World Bank, Prosperity for All: Ending Extreme Poverty, April 2014, http://siteresources.worldbank.org/INTPROSPECTS/Resources/334934-1327948020811/8401693-1397074077765/Prosperity_for_All_Final_2014.pdf; Transparency International, Corruption Perceptions Index, 2013, http://www.transparency.org/cpi2013/results; the ONE Campaign recently issued a report suggesting that ‘at least $1 trillion is being taken out of developing countries each year through a web of corrupt activity….’ ONE Campaign, The Trillion-Dollar Scandal, September 2014, http://www.one.org/international/blog/exposed-the-trillion-dollar-scandal/.
5. International Monetary Fund, Does Corruption Affect Income Inequality and Poverty? IMF Working Paper WP/98/76, May 1998, http://www.imf.org/external/pubs/ft/wp/wp9876.pdf. This study estimates that an increase of 0.78% in the growth rate of corruption in a country reduces the income growth rate of the poorest 20% by 7.8% every year.
6. Nawaz, F., State of Research on Gender and Corruption, U4, June 2009, http://www.u4.no/publications/state-of-research-on-gender-and-corruption/downloadasset/403; Hossain, N., C.N. Musembi and J. Hughes, Corruption, Accountability and Gender: Understanding the Connections, United Nations (UN) Development Programme and United Nations Development Fund for Women, 2010, http://www.undp.org/content/dam/aplaws/publication/en/publications/womens-empowerment/corruption-accountability-and-gender-understanding-the-connection/Corruption-accountability-and-gender.pdf; and Who Answers to Women? Gender and Accountability, United Nations Development Fund for Women, 2008, http://www.unifem.org/progress/2008/media/POWW08_Report_Full_Text.pdf. Women and girls can also be required to pay bribes when they come into contact with government officials, for example at border crossings; see: ICAI, DFID’s Trade Development Work in Southern Africa, December 2013, paragraph 2.97 http://icai.independent.gov.uk/wp-content/uploads/2013/12/DFIDs-Trade-Development-Work-in-Southern-Africa-Report.pdf.
7. Organisation for Economic Co-operation and Development, The Rationale for Fighting Corruption, Brief, CleanGovBiz, 2014, http://www.oecd.org/cleangovbiz/49693613.pdf (provides a comprehensive overview of widely accepted estimates of the cost of corruption).
8. Organisation for Economic Co-operation and Development, The Rationale for Fighting Corruption, Brief, CleanGovBiz, 2013, http://www.oecd.org/cleangovbiz/49693613.pdf.
9. Johnsøn J., N. Taxell and D. Zaum, Mapping evidence gaps in anti-corruption: Assessing the state of the operationally relevant evidence on donors’ actions and approaches to reducing corruption, Anti-Corruption Resource Centre, 2012, http://www.u4.no/publications/mapping-evidence-gaps-in-anti-corruption-assessing-the-state-of-the-operationally-relevant-evidence-on-donors-actions-and-approaches-to-reducing-corruption/.
10. As of September 2014, 172 countries have ratified the UN Convention Against Corruption. See https://www.unodc.org/unodc/en/treaties/CAC/signatories.html.
11. https://www.g20.org/g20_priorities/g20_2014_agenda/fighting_corruption.
12. Joint UK Ministerial Statement on International Anti-Corruption Day, 9 December 2013, https://www.gov.uk/government/news/joint-uk-ministerial-statement-on-international-anti-corruption-day.
13. ICAI, The Department for International Development’s Approach to Anti-Corruption, November 2011, http://icai.independent.gov.uk/wp-content/uploads/2010/11/DFIDs-Approach-to-Anti-Corruption.pdf.
14. This is discussed in more detail in the Objectives section.
15. This is discussed in more detail in the Learning section.
16. See https://www.unodc.org/unodc/en/treaties/CAC/.
17. See http://eiti.org/.
18. See http://www.medicinestransparency.org.
19. See http://www.baselgovernance.org/icar/.
20. See http://www.fatf-gafi.org.
21. See http://devtracker.dfid.gov.uk/projects/GB-1-203881/.
22. We are aware that some DFID programmes include components that seek to tackle petty corruption as part of a holistic approach.
23. The OECD classifies anti-corruption expenditure as that which supports ‘specialised organisations, institutions and frameworks for the prevention of and combat against corruption, bribery [and] money-laundering … e.g., anti-corruption commissions and monitoring bodies, special investigation services, institutions and initiatives of integrity and ethics oversight, specialised non-governmental organisations (NGOs), other civil society and citizens’ organisations directly concerned with corruption’, see, http://www.oecd.org/dataoecd/18/55/41992632.xls.
24. See OECD, List of CRS Purpose Codes, http://www.oecd.org/dataoecd/18/55/41992632.xls.
25. Annex A3 provides a list of the types of documents we reviewed for this report.
26. Annex A4 provides a list of our consultations, excluding our surveys, for this report.
27. ICAI, DFID’s Approach to Anti-Corruption and its Impact on the Poor – Terms of Reference, http://icai.independent.gov.uk/wp-content/uploads/2014/04/ICAI-ToR-Anti-Corruption-Second-Review.pdf.
28. This is described further in the Delivery section.
29. Our survey work followed an open source methodology known as Constituent Voice. That approach was developed by Keystone Accountability and published under the open source Creative Commons Attribution-ShareAlike 4.0 International License, http://creativecommons.org/licenses/by-sa/4.0/.
30. Annex A2 details the stakeholder surveys we undertook for this review, including what we did for each of the surveys we conducted, as well as the principal insights gained from this work and the limits thereon.
31. The term ‘democratisation of corruption’ is found in the literature relating to corruption. We use it here to express the mass spread and acceptance of corruption in a society, as well as its political, moral and social ramifications.
32. World Bank, 2013 figure, http://data.worldbank.org/indicator/SP.POP.TOTL.
33. World Bank, Nigeria Overview, http://www.worldbank.org/en/country/nigeria/overview.
34. UK Government Guidance, Nigeria: country information and guidance, 20 May 2014 https://www.gov.uk/government/publications/nigeria-country-information-and-guidance.
35. Transparency International, Corruption Perceptions Index, 2013, http://www.transparency.org/cpi2013/results. See also the World Bank’s Worldwide Governance Indicators for Nigeria, http://info.worldbank.org/governance/wgi/index.aspx?fileName=c161.pdf#countryReports. This sets out Nigeria’s aggregate score, based on a number of different governance measurements, for issues ranging from rule of law, voice and accountability, government effectiveness and control of corruption.
36. Cited in a 28 August 2012 speech by former Vice President of the World Bank for Africa, Dr. Obiageli ‘Oby’ Ezekwesili, entitled Corruption, National Development, the Bar and the Bench, Part 1, http://ynaija.com/oby-ezekwesili-corruption-national-development-the-bar-and-the-bench-part-3/
37. DFID defines fragile states by reference to the World Bank’s Country Policy and Institutional Assessment indicators, the Fund for Peace’s Fragile States Index and Uppsala University’s Uppsala Conflict Data Program. See DFID, Results in Fragile and Conflict-Affected States and Situations, 2012, https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/67437/managing-results-conflict-affected-fragile-states.pdf.
38. World Bank, 2013 figure http://data.worldbank.org/indicator/SP.POP.TOTL.
39. World Bank, Nepal Country Profile, http://data.worldbank.org/country/nepal.
40. UN Development Programme, 2013 Human Development Report, The Rise of the South: Human Progress in a Diverse World, http://hdr.undp.org/en/2013-report.
41. Transparency International, Corruption Perceptions Index, 2013, http://www.transparency.org/cpi2013/results.
42. Transparency International, Corruption Perceptions Index, 2012, http://www.transparency.org/cpi2012/results.
43. Transparency International, Global Corruption Barometer, 2013, http://files.transparency.org/content/download/604/2549/file/2013_GlobalCorruptionBarometer_EN.pdf.
44. FCO and DFID, Joint UK Ministerial Statement on International Anti-Corruption Day, 9 December 2013, https://www.gov.uk/government/news/joint-uk-ministerial-statement-on-international-anti-corruption-day.
45. David Cameron, Combating Poverty at Its Roots, 1 November 2012, http://online.wsj.com/news/articles/SB10001424052970204712904578090571423009066.
46. This is described further in the Delivery section.
47. The International Centre for Asset Recovery works with countries to recover stolen assets by providing strategic case advice, technical assistance and capacity building, and participates in policy discussions at the global level on setting standards in asset recovery.
48. These considerations have arisen in relation to the settlement payment by BAE Systems to the Government of Tanzania. See International Development Committee, Financial Crime and Development, Eleventh Report of Session 2010–12, 30 November 2011, http://www.publications.parliament.uk/pa/cm201012/cmselect/cmintdev/847/847.pdf.
49. Extractive Industries Transparency Initiative, http://eiti.org/Nigeria.
50. Mejía Acosta, A. The Impact and Effectiveness of Accountability and Transparency Initiatives: The Governance of Natural Resources, 2012, http://r4d.dfid.gov.uk/PDF/Outputs/Mis_SPC/60827_DPRMejia_Preprint.pdf. EITI was subject to an independent evaluation in 2011, which concluded that ‘little impact at the societal level can be discerned … largely due to [EITI’s] lack of links with larger public sector reform processes and institutions.’ EITI responded to this evaluation by redesigning the EITI standard and country implementation processes to enhance impact. Countries are beginning to adopt the new standard. See http://eiti.org/eiti/history.
51. The principal elements considered to allow corruption to flourish are an adaptation of Robert Klitgaard’s ‘corruption formula’, International Cooperation Against Corruption, International Monetary Fund, Finance and Development, March 1998, http://www.imf.org/external/pubs/ft/fandd/1998/03/pdf/klitgaar.pdf.
52. The public ACCF strategies are available at https://www.gov.uk/government/collections/anti-corruption-strategies-by-country.
53. For an example of this, see the combined statement of embassies of foreign governments in Kenya, including the UK, Graft is Kenya’s Achilles’ heel, Daily Nation, 12 April 2014, http://www.nation.co.ke/news/politics/Graft-is-Kenyas-Achilles-heel-/-/1064/2277652/-/m0km0u/-/index.html.
54. The limitations and risks of this policy are described further in paragraph 2.5.
55. The government-led programme and DFID’s approach to the risks related to it are described further in the Impact section.
56. This is described further in paragraph 1.11.
57. Other aspects of the limited nature of the logframes are discussed in Figure 10 on page 28.
58. This is described further in Figure 1 on page 2.
59. We reviewed DFID’s provision of budget support in our report, The Management of UK Budget Support Operations, ICAI, May 2012, http://icai.independent.gov.uk/wp-content/uploads/2012/05/ICAI-Budget-Support-Final-Report-32.pdf.
60. This is consistent with academic reviews. See M. de Vibe, N. Taxell, P. Beggan and P. Bofin, Collective donor responses: Examining donor responses to corruption cases in Afghanistan, Tanzania and Zambia, U4, Report 1. November 2013, http://www.u4.no/publications/collective-donor-responses-examining-donor-responses-to-corruption-cases-in-afghanistan-tanzania-and-zambia/downloadasset/3289.
61. A qualified audit opinion is provided by an auditor if it is unable to determine that an entity’s financial statements present fairly, in all material respects, the financial position and results of operations of that entity.
62. DFID’s Approach to Anti-Corruption, ICAI, November 2011, http://icai.independent.gov.uk/wp-content/uploads/2010/11/DFIDs-Approach-to-Anti-Corruption.pdf.
63. The reasons for this project stalling and being cancelled are described in Figure A4 in the annex of our report, entitled DFID’s Peace and Security Programme in Nepal, February 2013, http://icai.independent.gov.uk/wp-content/uploads/2013/02/ICAI-report-DFIDs-Peace-and-Security-programme-in-Nepal2.pdf.
64. Bertucci, G. and Yi Armstrong, E., Why Anti-Corruption Crusades Often Fail to Win Lasting Victories, 2000, http://unpan1.un.org/intradoc/groups/public/documents/un/unpan010748.pdf.
65. This is a finding we also noted in our prior reports: DFID’s Use of Contractors to Deliver Aid Programmes, ICAI, May 2013, http://icai.independent.gov.uk/wp-content/uploads/2013/12/ICAI-REPORT-DFIDs-Use-of-Contractors-to-Deliver-Aid-Programmes.pdf; and How DFID Learns, ICAI, April 2014, http://icai.independent.gov.uk/wp-content/uploads/2014/04/How-DFID-Learns-FINAL.pdf.
66. Through the J4A programme, DFID seeks to assess the improved effectiveness of the anti-corruption agencies by virtue of the number of reforms being implemented at those institutions, value of assets recovered, level of information-sharing and level of public awareness of corruption prevention. There is very limited measurement of how these measures of tackling corruption actually impact on corruption as experienced by the poor.
67. See also our findings on the Kalondolondo programme in Malawi in ICAI, DFID’s Empowerment and Accountability Programming in Ghana and Malawi, October 2013, http://icai.independent.gov.uk/wp-content/uploads/2013/12/Empowerment-and-Accountability-081013-FINAL.pdf.
68. This is a view consistent with research findings on corruption — that publication of incidences of corruption is effective, provided that the corruption that is highlighted is actually punished; see Johnsøn, J., N. Taxell and D. Zaum, Mapping evidence gaps in anti-corruption: Assessing the state of the operationally relevant evidence on donors’ actions and approaches to reducing corruption, Report 7, October 2012, http://www.u4.no/publications/mapping-evidence-gaps-in-anti-corruption-assessing-the-state-of-the-operationally-relevant-evidence-on-donors-actions-and-approaches-to-reducing-corruption/downloadasset/2979.
69. See Department for International Development, Working effectively in conflict-affected and fragile situations: Summary note, 12 March 2010, https://www.gov.uk/government/publications/working-effectively-in-conflict-affected-and-fragile-situations-summary-note.
70. The Local Government and Community Development Programme in Nepal.
71. Aid and corruption in Nepal: Low road through the Himalayas, The Economist, 31 May 2011, http://www.economist.com/blogs/banyan/2011/05/aid-and-corruption-nepal.
72. See Figure A16 in Annex A2.
73. Source: Survey undertaken on behalf of the ICAI review team. Additional details on the survey are set out in Annex A2 and additional findings are set out in Figures A15-A20 in Annex A2.
74. Source: Survey undertaken on behalf of the ICAI review team. Additional details on the survey are set out in Annex A2.
75. See Figure 2 on page 5 for details of the actual and forecast expenditure to tackle corruption.
76. The U4 website can be seen at http://www.u4.no.
77. Hanna, R., et al., The effectiveness of anti-corruption policy: What has worked, what hasn’t and what we don’t know, Systematic review, Technical Report. Social Science Research Unit, Institute of Education, University of London, July 2011 http://r4d.dfid.gov.uk/PDF/Outputs/SystematicReviews/Anti_corruption_2011Hanna.pdf.
78. Johnsøn, J. N. Taxell and D. Zaum, Mapping evidence gaps in anti-corruption: Assessing the state of the operationally relevant evidence on donors’ actions and approaches to reducing corruption, Report 7, October 2012, http://www.u4.no/publications/mapping-evidence-gaps-in-anti-corruption-assessing-the-state-of-the-operationally-relevant-evidence-on-donors-actions-and-approaches-to-reducing-corruption/.
79. Overseas Development Institute, Corruption Evidence Paper, forthcoming 2014.
80. ICAI, How DFID Learns, April 2014, http://icai.independent.gov.uk/wp-content/uploads/2014/04/How-DFID-Learns-FINAL.pdf.
81. Transparency International, Corruption Perception Index, http://www.transparency.org/research/cpi/.
82. The impact of the late release of funds is described further in Figure 5 on page 12.