Strengthening Tanzania’s Anti Corruption Action Business Case - Intervention Summary What support will the UK provide? The UK will provide £11m over four years (2011/12 – 2014/15). Why is UK support required? Tanzania has taken important steps in the past decade to tackle corruption, out performing most of its neighbours in East Africa in international rankings (with the exception in recent years of Rwanda). Nonetheless, corruption is still prevalent. Tanzania has a relatively comprehensive anticorruption legislative and institutional framework, but implementation is weak, and there is a low likelihood of being held to account for corrupt behaviour. Absolute levels of corruption are hard to assess with a high degree of accuracy, as are overall trends. After an extended period of progress on Transparency International’s Corruption Perceptions Index (CPI), Tanzania’s score declined between 2007 and 2010 (3.2 to 2.7). Public perceptions, although not a reliable measure of absolute levels of corruption, reflect a growing awareness of, and concern about corruption across key sectors of government1. Corruption in Tanzania threatens to undermine development gains by diverting scarce revenue from productive use, weakening essential services such as healthcare and law enforcement, and deterring private investment2. It also poses a risk to the UK’s substantial aid investment in Tanzania: DFID plans to channel over half of its support to Tanzania through the government system in the next four years. DFID’s wider engagement in Tanzania already tackles various aspects of corruption through a variety of different interventions; for example through improved management of public finances including procurement, parliamentary scrutiny of spending, improved human resource management in the civil service, and empowering civil society to hold government to account. This programme will focus on improving the performance of the institutions most directly involved in tackling corruption (Preventing and Combating Corruption Bureau – PCCB, Directorate of Public Prosecutions – DPP, Financial Intelligence Unit – FIU, and the National Audit Office – NAO). It will help them to work more effectively together on the identification and handling of corruption risk and cases. It will also address corruption in the police and judiciary, which the Tanzanian public perceive as the most corrupt government institutions. The programme will also provide demanddriven advisory inputs to other sectors that are at high risk of corruption and/or where significant investments are planned by DFID (likely to include Tanzania Revenue Authority and Ports Authority, education, health, water and roads). Finally, this programme will help increase national and international incentives for Tanzania to take action against corruption, though an Integrity Fund. This programme will be delivered through a managing agent to be competitively procured. The programme will be overseen by the Tanzanian government through the National Criminal Justice Forum. What are the expected results? The programme’s intended impact is reduced impact of corruption on the poor in Tanzania (directly in terms of bribery and improved access to services and indirectly through reduced diversion of public resources). The programme’s intended outcome is more active implementation of anticorruption measures, particularly in high risk/high impact sectors. Success will be measured through improvements in relevant Global Integrity Index scores and National Corruption Surveys, and through a reduction in the proportion of the public who pay a bribe when they encounter the police or the judiciary . 1 2 East African Bribery Index 2010, Afrobarometer 2008 Public Goods, Rents and Business in Tanzania’ ODI 2011, ‘Informal Pay and the Quality of Healthcare’ CMI 2007 The outputs will be: a. Oversight agencies improve identification and handling of corruption risks and cases Success will be measured through increases in numbers / percentages of corruption cases reported and accepted for investigation and prosecution, faster average handling speeds, improvements in follow up on corruption findings in NAO reports, and Tanzania being taken off the Financial Action Task Force watch list for deficiencies in anti money laundering compliance. b. Accountability mechanisms strengthened in the police and judiciary, and anti-corruption measures incorporated into major DFID sector programmes Success will be measured through improved corruption perceptions of traffic police and compliance with police front desk, police criminal investigation and primary court procedures, increases in corruption complaints received / addressed by the police and judiciary, and active management of corruption risk in DFID sector programmes including through special audits c. Government has increased domestic and international incentives to tackle corruption Success will measured through increases in public willingness to report corruption / numbers of corruption complaints, results achieved through NGO and private sector advocacy efforts, and results achieved through enhanced joint donor dialogue with government on corruption. The above results will be measured using a mixture of: international and country-specific corruption surveys, government reports and data sets, external reviews and evaluations. 1. Strategic Case A - Context and need for DFID intervention Corruption challenges 1.1 Corruption is a brake on economic development, a burden for the poor and a waste of valuable and limited resources. It poses a serious development challenge in Tanzania. The Preventing and Combating of Corruption Bureau (PCCB) conducted an extensive governance and corruption survey in 2009, which found that 87% of people regarded corruption as one of the main problems facing Tanzania, with the police, judiciary and health services cited as the worst offenders. The survey also identified corruption as being a significant impediment to the growth and operation of businesses. Based on recent surveys and reports, DFID’s 2011 Fiduciary Risk Assessment (FRA) also assessed the risk of corruption (mainland Tanzania and Zanzibar) as substantial. 1.2 Corruption impedes DFID’s objectives in Tanzania in a number of ways including: (i) distorting the allocation of scarce public resources away from productive to non-productive activities3; (ii) by undermining the impact of services4; iii) discouraging private sector investment and development5 and (iv) through its disproportionate impact on the poorest.6 It also risks reducing the value for money of UK aid in Tanzania. 1.3 During most of the past decade, Tanzania has shown progress in tackling corruption, and outperformed most of its neighbours in Africa. This has been reflected in an overall rise on the main international anti-corruption indices (up from -1.09 (2000) to -0.42 (2009) on the World Bank Control of Corruption Index, and from 2.2 (2001) to 2.7 (2010) on Transparency International’s Corruption Perception Index (CPI)). However, in recent years some of these gains have been lost, with Tanzania’s score declining in the Transparency International CPI between 2007 and 2010 (see Chart 1 below). Chart 1 – Tanzania’s score on Transparency International’s Corruption Perceptions Index (compared to other DFID Africa Partner Countries) 4.5 Tanzania DRC Ethiopia Ghana 4 3.5 3 2.5 2 1.5 1 2005 3 2006 2007 2008 2009 2010 Kenya Malawi Mozambique Nigeria Rwanda Sierra Leone Sudan Uganda Zambia Zimbabwe e.g. by prioritising spending in areas where money can most easily be diverted, inflating the cost of contracts, or extorting bribes from businesses. See Systematic Review on Growth and Corruption, Ugur 2011 4 Public Goods, Rents and Business in Tanzania’ ODI 2011, ‘Informal Pay and the Quality of Healthcare’ CMI 2007 5 The World Bank Institute produced research in 2004 stating that countries that tackle corruption and improve their rule of law may ‘increase their incomes by as much as four times in the long term, with child mortality falling by as much as 75 percent’. 6 See Transparency International Global Corruption Barometers 2003-2009, the last of which shows that the percentage of people in the lowest income quintile paying bribes for services exceeds those in the highest income quintile in all but one of the services surveyed 1.4 Other surveys capturing Tanzanian perceptions and experiences of corruption reflect a growing awareness of corruption. 17.4% of businesses identified corruption as the most problematic factor for doing business in Tanzania in the 2010/2011 Global Competitiveness Report, compared with 12.4% in 2008/09. According to the Afrobarometer, the proportion of Tanzanians rating government as corrupt rose between 2005 and 2008 (see Table 1 below), while the proportion rating government as doing badly fighting corruption rose from 26% to 39% over the same period. Over 50% of Tanzanians surveyed said they would fear punishment if they complained about misuse of public funds (compared to 34% in Kenya, 33% in Uganda, and 27% in Mozambique). Table 1: Tanzanian perceptions of corruption in government Taken from Afrobarometer Surveys for Tanzania (2005 and 2008) Government entity President and officials in his office Members of Parliament Elected local government councillors Government officials % saying some, most or all are corrupt (2005) 29 38 44 45 (national/local) % saying some, most or all are corrupt (2008) 54 70 71 73 % increase between 2005 and 2008 86% 84% 61% 62% 1.5 The sectors most affected by corruption, according to public perception, are the police and judiciary. Between 2009 and 2010, the proportion of Tanzanians who had to pay bribes when dealing with the Tanzanian Police Force (TPF) or judiciary rose significantly, from 40.9 to 53.8% for the police, and 28.5 to 46.6% for the judiciary7. Analysis of corruption in the police and judiciary highlights problems ranging from petty bribery in the traffic police, through to extortion from businesses, wrongful arrest, and interference with due process in remand and trial of suspects particularly at primary court level (eg manipulating bail conditions, ‘mislaying case files’, and influencing case hearings). 1.6 As a low income country, and in a context of severe budget constraints8, the Tanzanian government cannot afford the misuse or diversion of public funds. The pervasiveness of corruption (or at least perception of pervasiveness) risks undermining reforms already achieved. Ultimately, it is the poor and vulnerable who are most affected because they are the most dependent on government services, and can least afford to pay petty bribes or challenge corrupt behaviour. What has already been done to tackle corruption 1.7 Tanzania has put in place many of the fundamentals required to tackle corruption successfully. Much of the momentum generated during President Kikwete’s first term (2005-2010) was built upon Tanzania’s credible anti-corruption legal framework. The Prevention and Combating of Corruption Act 2007 has helped Tanzania to meet most of the legal standards required by the UN Convention Against Corruption (UNCAC) and has also strengthened the PCCB, although it is still not independent. More recently Tanzania also adopted the Election Expenses Act (2010) to achieve more transparency in campaign finance. Tanzania will be reviewed on UNCAC compliance in 2012, which may drive further improvements in its institutional and legislative frameworks, eg the adoption of Access to Information legislation. 1.8 Considerable donor support has been given to the PCCB, which has increased its capacity significantly in recent years (as of April 2010, 471 corruption cases had commenced prosecution, more than in the whole of 2009). The Director of Public Prosecutions (DPP) has also reported strengthened capacity in terms of number of cases handled and files closed per year. In May 2010 and 2011, Tanzania secured its first grand corruption convictions and the recent informal release of the PCCB Governance and Corruption Survey has led to greater levels of media focus on anticorruption. In 2008, the Ministry of Finance established a Financial Intelligence Unit (FIU) to investigate or report cases of suspicious financial transactions. The FIU is now scaling up to 7 East Africa Bribery Index 2010 Tanzania’s fiscal deficit has significantly increased in the past few years from 1.6% in 2007/8 to an estimate of 6.5% of GDP in 2010/11; i.e. from TSh. 366.0bn to 2241.0bn – Economist Intelligence Unit 8 improve compliance with international standards on Anti Money Laundering (AML). As of Oct 2011, Tanzania was at Tier 1 of the FATF watch-list of countries that are non-cooperative and/or have serious AML deficiencies. If it does not make sufficient progress by Feb 2012, it will drop down to Tier 2. It then has 12 months at Tier 2 before it could be subjected to counter-measures such as loss of international banking relationships. 1.9 Extensive donor support has also been provided to the NAO, which is now close to full AFROSAI (African Organisation of Supreme Audit Institutions) Level III accreditation. Progress includes the production of more comprehensive and detailed audit reports, and broadening the NAO’s mandate to forensic audits and reviews of operational systems to identify and mitigate potential risk areas for fraud. Donor support has also helped to strengthen the Public Procurement Regulatory Authority (PPRA) resulting in improved compliance with procurement legislation. Parliament has also intensified its debates on the misuse of public funds (as identified in audit reports and public expenditure tracking surveys). An Internal Auditor General (IAG) office has also been recently established within the Ministry of Finance to strengthen internal audit functions throughout Government. Prompted by pressure from budget support donors, the Government has initiated a high-level dialogue on corruption. What more is needed to tackle corruption 1.10 Despite the above progress, perceptions of corruption are worsening. This may be partly due to increased awareness of corruption, and improve ability of government systems to detect it. However, surveys clearly show experiences of corruption are also on the rise in certain institutions, such as the police and judiciary. Several factors may be responsible. The Global Integrity Index Tanzania Report for 2010 finds that Tanzania’s problem is not so much the quality of its institutional framework, but the weak implementation of that framework. This manifests for example in limited ability of the PCCB to tackle powerful interests, a reactive rather than proactive approach to investigations, low willingness of citizens to complain on corruption, and poor responsiveness to corruption complaints. A further challenge is translating institutional improvements at the central level into sharper anti-corruption controls in government services. The judiciary and law enforcement officials are notable for their lack of accountability. Other high spend/high risk sectors such as health, ports, tax administration and local government also rate poorly in the East Africa Bribery Index9. DFID’s Objectives and Comparative Advantage 1.11 Tackling corruption is integral to DFID’s overall Business Plan, which commits that up to 5% of all budget support should go to accountability institutions. Corruption remains a ministerial priority. It is also a major factor affecting public support for the aid agenda10. Addressing corruption is a core element of HMG’s objectives in Tanzania, which aim to support stronger, more transparent and democratic oversight institutions and hold institutions to account by empowering civil society. Reduced corruption in the public sector will help DFID to achieve its broader goals, since over 50% of DFID resources for Tanzania will be channelled through the government over the next four years. It will also improve conditions for private sector led growth. 1.12 DFID already engages in key areas that are important to the prevention of corruption. These include public financial management reform11, public sector reform12, deepening democracy13, and 9 Approximately one quarter of those surveyed that sought services from the Tanzania Revenue and Ports Authorities paid a bribe; and one fifth of those seeking services from a hospital. East Africa Bribery Index 2010 10 57% of the UK public believe that corrupt leaders, bureaucratic wastage and poor use of funds make it pointless donating money to help reduce poverty. 11 The government’s PFM reform programme addresses areas such as more realistic revenue forecasting and improved financial reporting; simplification of financial management information systems; strengthening of payroll controls; procurement compliance audits; strengthened action by parliamentary accounts committees; and harmonisation of tax laws. 12 The Public Sector Reforms Programme II (PSRP II) aims to enhance capacity, performance and accountability of Ministries, independent government Departments and Agencies (MDAs) in the use of public resources and service delivery to levels consistent with timely and effective implementation of the strategic and priority programmes under the National Strategy for Growth and Reduction of Poverty (MKUKUTA II). Specifically the programme addresses accountability and responsiveness mechanisms which provide opportunity to citizens to be aware and proactively demand quality services; and systems for managing public servants and leaders to instil integrity and work ethics. empowering citizens to hold government to account. DFID is also engaged in high level dialogue on corruption through General Budget Support fora,. This programme will complement and these efforts, by supporting more targeted action on anti-corruption enforcement and more active implementation of anti-corruption measures at sector level. For more details on DFID’s overall anti-corruption engagement in Tanzania, see the diagram at Annex 5. 1.13 DFID will remain one of the largest bilateral supporters of Tanzania for the next five years, contributing over 10% of total foreign aid. The UK has influence beyond its spending, eg through the UK’s broader bilateral relationship, DFID’s key positions in the Government–donor dialogue structure, and through the multilateral organisations that the UK finances such as the World Bank, UNDP, AfDB and EC. These place it in a good position to make linkages and lever collective efforts. B – Impact and Outcome 1.14 The programme’s intended impact is reduced impact of corruption on the poor in Tanzania (directly in terms of bribery and improved access to services, indirectly through reduced diversion of public resources). The programme’s intended outcome is more active implementation of anticorruption measures, particularly in high risk/high impact sectors. Success will be measured through improvements in relevant Global Integrity Index scores, and through a reduction in the proportion of the public who pay a bribe when they encounter the police, judiciary or magistrates. 1.15 The outputs will be: a. Oversight agencies improve identification and handling of corruption risks and cases Success will be measured through increases in numbers / percentages of corruption cases reported and accepted for investigation and prosecution, faster average handling speeds, improvements in follow up on corruption findings in NAO reports, and Tanzania being taken off the Financial Action Task Force watch list for deficiencies in anti money laundering compliance. b. Accountability mechanisms strengthened in the police and judiciary, and anti-corruption measures incorporated into major DFID sector programmes Success will be measured through improved corruption perceptions of traffic police and compliance with police front desk, police criminal investigation and primary court procedures, increases in corruption complaints received / addressed by the police and judiciary, and active management of corruption risk in DFID sector programmes including through special audits. c. Government has increased domestic and international incentives to tackle corruption Success will measured through increases in public willingness to report corruption / numbers of corruption complaints, results achieved through NGO and private sector advocacy efforts, and results achieved through enhanced joint donor dialogue with government on corruption. 13 The overarching objective of Deepening Democracy Programme is to strengthen the democratization processes by introducing interventions that will support efforts to advocate legal reforms (and, possibly, constitutional amendments) for a more liberalized political environment; strengthen and entrench the human and material elements of existing democratic practices and institutions, making them more robust, responsive effective and efficient in their operation; and enhance democratic beliefs and culture and intensify understanding of and respect for democratic principles, values and culture. 2. Appraisal Case A - Feasible options 2.1 This section presents a theory of change for how to achieve more active implementation of anti-corruption measures in Tanzania. It then considers the feasible options for DFID’s anticorruption programme (taking into account what is already being supported by DFID and others, and evidence of what is most likely to have impact), outlines in more detail what the programme will deliver, and identifies the optimum delivery channels. Theory of Change 2.2 There is no one widely accepted international theory of change on tackling corruption that can be applied in all contexts. There are multiple views about what drives corrupt behaviour at the individual level and why anti-corruption efforts are often so weak. Evidence of what works on anticorruption is limited and often highly country specific. However to the extent possible, the theory of change for this programme draws on relevant international evidence and academic literature. 2.3 The corruption problem identified in the Strategic Case is as follows: Although Tanzania has put in place most of the institutional structures necessary to tackle corruption, these are not fully deployed, and corruption levels remain significant, weakening government services and prospects for long term growth and development. 2.4 Analysis suggests three factors lie behind this. The first, which concerns individual behaviour, is simply that the rewards of corruption presently outweigh the risks. Despite Tanzania’s progress on anti-corruption, there remain plentiful opportunities to commit corruption, and a low likelihood of being caught or sanctioned. The realities of a cumbersome bureaucracy with limited capacity to meet service demand, coupled with comparatively low pay on the part of junior government officials14, help drive this equation. Monopoly and discretion strengthen the hands of government officials to extort payments. Despite doing relatively well overall in tackling corruption relative to its East African neighbours, the 2006 World Bank Enterprise Survey for Tanzania found that 49.47% of firms expected to pay informal payments to public officials to get things done, compared to 36% for the region. This leads to a collective action problem: the more pervasive corruption is, the more it defines the ‘rules of the game’ and the greater the cost to the individual to challenge it15. 2.5 The second factor is the weak capacity of the Tanzanian government system to actively implement anti-corruption and wider accountability measures16. Although several oversight institutions have improved their performance, major weaknesses remain, especially in terms of the disconnect between oversight institutions (as evidenced by low follow up on audit reports, and slow passage of corruption cases between institutions), inadequate skills and lack of proactivity in the detection and handling of complex corruption cases, and weak implementation of anticorruption measures at the sector level (eg the absence of police supervision). Historically, analysis shows that those developing countries that achieved significant improvements in international corruption rankings have had the bureaucratic ability to properly impose and enforce anti-corruption measures17. 2.6 The third factor, and possibly the most important in explaining recent trends, is inconsistent and uneven levels of political commitment to tackle corruption. This is the case in high politics but also at an institutional level. A recent study in Tanzania finds that transparency and accountability Graduate entry scale into the general civil service commences at Tsh 156,000 per month. The Tanzania Trade Union’s Congress recommends that the minimum national monthly salary should be Tsh 350,000 (though this figure is not supported by the GoT); the general view on low pay is that it can be a driver of petty corruption, but improving pay does not by itself reduce corruption ie it’s a necessary but not sufficient condition 15 See ‘The Failure of Anti Corruption Policies’, Persson, Rothstein and Teorell 2010, and A Vicious Circle of Corruption and Mistrust in Institutions in sub-Saharan Africa, Cho and Kirwin 2007 16 See Global Integrity Report 2010 17 See ‘Governance and Anti Corruption Reforms in Developing Countries: Policies, Evidence and Ways Forward’, Mushtaq Khan 2006, and ‘Developmental Patrimonialism? The Case of Rwanda’, Booth and Golooba-Mutebi 2011. 14 initiatives supported by international donors, are ‘systematically undermined by...informal processes of patronage and rent seeking’ (Cooksey 2011). A major driver for these processes is the government’s desire to maintain political support and to manage powerful constituencies. Various factors may intensify patronage and rent seeking, for example disunity in the political system (as happened in the CCM in the run up to the 2010 general elections) and new access to ‘easy’ rents (such as the anticipated gas reserves in Tanzania). The challenge is how to support a transition from a political system driven by narrow powerful constituencies, to one in which resources are used in the wider public interest. 2.7 This leads to the following hypothesis: Corruption levels in Tanzania will reduce if there is more active implementation of anti-corruption measures, and this in turn will occur if there is: (a) improved technical capacity and joint working among oversight agencies – leading to more proactive detection and efficient handling of corruption risks and cases, and stronger collective action / less scope for political interference (b) fewer opportunities and incentives to be corrupt - through government reforms that reduce opportunities for corruption (eg simplified financial management and bureaucratic systems, clearer performance standards); also through increased likelihood of detection and sanction if committing corruption (eg more functional complaints mechanisms, improved performance monitoring, more in-depth auditing). These also need to be applied at sector level if they are to have a tangible impact on public experiences of corruption (such as paying bribes to police). (c) increased domestic and international incentives for government to tackle corruption – strengthening the voice and influence of domestic actors that have an interest in how public resources are used in Tanzania, to broaden the government’s constituency base beyond political elites, and increase public confidence that it is worthwhile acting against corruption. Using international levers to motivate government action, eg donor dialogue on budget support, international corruption investigations, and compliance with international standards. 2.8 These three areas are mutually reinforcing, as the diagram below shows: Diagram 1: Theory of Change for Strengthening Anti Corruption Enforcement in Tanzania 2.9 There are some key assumptions undermining this theory of change. On A – Stronger oversight system – these are that: there is sufficient funding for the oversight institutions to deliver their roles; there is a coherent institutional framework allowing institutions to work effectively together rather than in competition with each other; and that the institutions have sufficient independence to be able to operate without too much political interference. On B, the assumptions are that: there will be sufficient increases in the proportion of corruption cases identified and sanctioned, and in the degree of oversight and monitoring within individual government institutions, to alter behaviour. On C, the assumptions are that: the Tanzanian government would be responsive to increased incentives to act; that civil society and the private sector would make reasonable demands; that there are international levers that touch sufficiently on national government interests to motivate action; and that development partners have sufficient common interests to be able to engage jointly on anti-corruption. 2.10 Clearly, tackling high levels of corruption is a long term proposition. Sustained improvements in absolute levels of corruption may therefore not be seen during the timeframe of a four year anti-corruption programme. Nonetheless intermediate outcomes are attainable, and should, over time, reduce corruption. Broad options for DFID programme 2.11 Perhaps the main challenge with a system-wide approach of the kind outlined in the diagram above is where to draw the line. Much of the wider system-strengthening on resource management and accountability in Tanzania is already being tackled through other programmes (civil society strengthening, public financial management, public sector and legal sector reform). In particular, much of the government’s public financial management reform work – which DFID supports - is directly relevant to anti-corruption. For example, the government intends its fourth PFM reform programme to be “instrumental in the fight against waste and unnecessary spending and corruption, as it improves access to financial data, increases audits relating to risks and results and promotes more open and competitive procurement processes and commitment controls”. Specific actions will include: More realistic revenue forecasting Strengthened centralised internal audit Simplification of financial management information systems Strengthened payroll controls Improved quality of financial recording and reporting Improved procurement compliance Greater use of risk-based auditing Strengthening of Parliamentary Accounts Committees Harmonisation of tax laws 2.12 These wider cross-government reforms do not need to be duplicated in this programme. Instead, it should focus on those elements of the theory of change that are not presently addressed elsewhere. This suggests three main options for DFID: i) Focus only on oversight institutions, since strengthening them would impact on all three factors above (stronger oversight, incentives not to be corrupt, and internal pressure on government) - A) in above theory of change) ii) Take a broader approach, which also addresses high risk sectors and external pressures on government – A), B), and C) in theory of change. iii) ‘Do nothing’, resulting in only indirect DFID support for anti-corruption through the existing reform programmes mentioned above, combined with any assistance that other development partners such as UNDP might provide on anti-corruption. B. Evidence of likely impact of options 2.13 The latest evidence on the effectiveness of anti-corruption measures – and thus, which of the above options would be most desirable - is medium. There are two relevant studies that have been carried out recently. The first, a DFID commissioned systematic review of the effectiveness of anti-corruption approaches, tentatively concludes that prospects for success are greatest when oversight is combined with strengthened incentives not to be corrupt (such as performance related pay), and that oversight is most powerful when it is capable of both detecting and sanctioning corruption18. The second is a 2011 Joint Evaluation of Anti Corruption Effectiveness (JEACE) sponsored by DFID and other partners. This draws on five country case studies and is more of a qualitative evaluation, examining a variety of programmes,. The JEACE suggests that a comprehensive, system-wide approach to tackling corruption has the greatest prospect of success, and identifies the following as important for impact: stronger working across oversight institutions; the incorporation of anti-corruption in sectors and wider reform programmes such as public financial management; greater engagement of civil society; and more use by donors of international analysis and dialogue with government. Both studies make clear the paucity of evidence on impact of anti-corruption measures. Most success stories are case specific and based on qualitative evidence of impact. They also note the challenge of translating successful project outcomes into impact on overall corruption trends. 2.14 Within Tanzania, efforts so far have focused on oversight institutions, together with wider system strengthening. As the Strategic Case makes clear, this has resulted in some gradual improvements in the performance of oversight institutions and wider public financial management (particularly auditing and procurement) but does not appear to have had much tangible impact on sectors, where the public are most directly affected, such as the police and judiciary. Nor has this been sufficient to sustain Tanzania’s ratings on international corruption indices (although this may be outside the scope of any donor programme). Environment and climate change effects 2.15 The environmental and climate change effects of the first two programme options above would be largely positive, since they would result in improved implementation of anti-corruption measures, which would indirectly benefit the management of natural resources and environmental services. The third option – ‘do nothing’ – would be the worst because without targeted action there may be less likelihood of offenders being held to account for corruption, eg in areas like award of contracts to logging companies. This is summarised below. Option Evidence Climate and Environment Category – risk and opportunity (A, B, C, D) 1. Support to oversight institutions Medium C risk and B opportunity. The poor and vulnerable rely heavily on environment for their livelihoods (micro) and contribution to the economy (macro). Direct impact is limited but indirect impact is substantive both on environmental services and in tracking revenues of natural resources such as forestry, land, water and mining 2. Oversight institutions, sectors and external influencing Medium C risk and B opportunity. See above – with possibility of even greater impact if external influencing leads to greater action on corruption in related sectors. 3. ‘Do nothing’ - support through existing system reform programmes and others’ anti corruption efforts Medium B risk and C opportunity. Without targeted action there may be less likelihood of offenders being held to account for corruption, and therefore greater risk of environmental abuse. Hanna, R., Bishop, S., Durlacher, K., Nadel, S., Scheffler, G. 2010. ‘The Effectiveness of Anti-Corruption Policy: What has worked, what hasn’t, and what we don’t know – A Systematic Review Protocol,’ DFID Systematic Review. 18 C. Appraisal of each feasible programme option 2.16 The benefits and disadvantages of each programme option will now be appraised. Option 1 - Taking a narrow focus on oversight institutions (estimated cost £5.15m) would keep the programme structure simple, and allow for concentrated effort but therefore limited impact. It would fail to tackle significant corruption vulnerabilities at the sector level – in particular in the police and judiciary – and would also fail to strengthen external demands on government, which are a critical part of motivating action. Option 2 - Taking a broad focus (estimated cost £11m) to include high risk sectors and external demands on government would be the most complicated to design and manage. However, it would be addressing more of the factors identified as important in the theory of change. It would also spread programme risk, so that if one area was making limited progress (such as working with the PCCB and DPP), more attention could be given to others that were showing more traction (such as external demands on government, or incorporation of more robust anti-corruption measures into the Tanzania Police Force). Option 3 - The ‘do nothing’ approach (financial cost £0) is not credible in the context of continuing substantial levels of corruption, the growing awareness of the levels of corruption and its negative impact amongst Tanzanians, and DFID’s continued investment through GoT systems. Working through existing reform programmes would be time consuming and the impact limited, because they have much broader goals and do not include all the institutions most relevant for anti-corruption efforts (such as the FIU). If DFID did not go ahead with a new anti-corruption programme, some modest assistance would continue through other donors: principally UNDP (probably focused on the PCCB through the next National Anti Corruption Strategy and Action Plan), plus some ad-hoc technical assistance from the US, Canada and others to the Ethics Secretariat, FIU and DPP. However, this would be on a much smaller scale and without taking a system wide approach, and it would represent a net loss of action on anti-corruption since DFID’s existing Tackling Corruption Programme has been brought to an end and would not be replaced by new anti-corruption specific programming. 2.17 The table below summarises this analysis, rating each option high, medium or low: Option Costs Benefits Risks 1. Narrow: Oversight institutions £ 5.15 million Medium Oversight institutions improve their identification and handling of corruption risk and cases Medium Action will be limited, because of lack of external pressure on government to act on corruption, and therefore lack of clear mandate for oversight institutions to do their job 2. Broad: Oversight institutions, sectors and external influencing £ 11.0 million High Oversight institutions improve their identification and handling of corruption risk and cases Reduced corruption in the most vulnerable sectors (e.g. police, judiciary) Reduction in citizens’ direct experience of corruption Greater external demands on government should build commitment to act Having a broad approach will build faster momentum, and will also allow more flexibility to scale up or down different components depending on where there is most traction Medium Broader programme, risks spreading funding and efforts too thinly. 3. Do nothing £0 Low/Medium Continued minor improvements in the performance of oversight agencies – eg agencies acquire new specialist skills Wider reforms eg on public financial management will indirectly and gradually reduce opportunities to commit corruption High Drop in support for anti-corruption and lack of system wide approach will reduce momentum for action among oversight institutions Fiduciary risk for DFID for lack of specific programme on anti-corruption given scale of bilateral investment in Tanzania and the present corruption indicators 2.18 From the table above, Option 3 is eliminated. The question is whether the benefits of Option 2 are worth the additional financial investment compared to Option 1. This appraisal concludes that they are, because of the greater likelihood of overall success on anti-corruption (based on the research listed in para 2.13 above), as well as the increased momentum that would be achieved through a major push of this kind. More detailed appraisal of Option 2 2.19 During the appraisal period, DFID Tanzania explored the components that could be included in Option 2, through consultations with the oversight agencies and other development partners, lessons learned on DFID’s previous anti-corruption programme, and specific analytical work. Below is a brief summary of findings: Support to oversight agencies needs to go beyond donating resources and one-off training courses, to include hands-on capacity strengthening, longer-term provision of high quality expertise, and more sustainable transfer of technical skills. There are weaknesses in handling complex corruption cases, and proactive intelligence gathering. Facilitating linkages between oversight institutions is critical: the NAO is producing findings that are not systematically followed up; the PCCB is submitting cases to the DPP that are not routinely processed; the FIU works in relative isolation and needs to be better connected to the other oversight bodies. In-depth support to vertical sectors should focus on the police and judiciary, because they are the worst ranking institutions in all corruption perceptions surveys on Tanzania. The risks of engaging are high: there are only a few international examples of successful anti-corruption efforts in the police and judiciary19, and previous reform efforts with the judiciary in Tanzania have had limited traction. However, there is a window of opportunity with the police, which has an Inspector General that is keen to tackle corruption, and also a new reform programme that would help underpin any anti-corruption efforts. Similarly, the judiciary has a new Chief Justice who wishes to tackle corruption in his institution20. Beginning work at primary court level (where both police and judiciary are involved) may be less contentious and stand more chance of success. The Canadian-led Legal Sector Reform Programme (LSRP) provides a context and dialogue mechanism that could be used to support this work (for example the LSRP is already engaged in supporting aspects of case management and records management). In addition, the appraisal found openings to provide demand-driven analytical and advisory support to other high risk/high spend sectors where DFID is engaged. The purpose would be to influence key design and review points in sector programmes, to ensure effective incorporation of anti-corruption measures. These would be light-touch inputs, with the responsibility for implementation lying with the sector programmes. At this stage of the programme, it is difficult to be specific about the precise support that would be provided through this demand-driven mechanism. However, the appraisal looked at six high risk sectors where DFID plans major spend, where there would be potential merit in analytical and advisory support beyond the normal corruption risk management / due diligence that is built into all DFID programmes: i) 19 Tanzania Revenue Authority – ranked as the 5th worst institution for bribery in Tanzania by the East African Bribery Index (EABI) 2011, with 26% of respondents that had contact with the TRA paying a bribe. DFID’s existing programme is coming to an end, and a new programme is likely from 2012. There would be an opportunity to influence the design of this programme, learning lessons from past anti-corruption efforts with the TRA. Benefits of improved anti-corruption would include increased government revenue, and more predictability for businesses and private individuals paying tax. Structural reform and rebranding led to a significant reduction in street level police corruption in Georgia (di Puppo, 2010). In Rajasthan, simple measures such as duty rotation at station level, and the involvement of community observers, were found to have an impact on corruption (Banerjee, Chattopadhyay, Duflo and Keniston, 2009. In Guatemala City, the establishment of a Clerk of Courts office virtually eliminated the problem of lost files and reduced court congestion. In Serbia, a new automated case management system that selected judges at random and allowed citizens to track the progress of their cases online enabled a 24% reduction in pending cases within two years (‘Reducing Corruption in the Judiciary’, USAID 2009) 20 Following his appointment, on 27 Dec 2010 Newspapers and TVs quoted the Mr Justice Othman who became Tanzania's seventh Chief Justice since independence outlining his priority areas as to address; one, the delay of cases and judgments, two, corruption, etc... (New CJ Othman outlines priority areas) ii) Tanzania Ports Authority (TPA) – ranked as the 4th worst institution by in Tanzania by the EABI 2011, with 24% of respondents that had contact with the TPA paying a bribe. DFID supports the TPA through the Trademark East Africa programme, which is planning a US$21m investment on port connectivity, increasing the number of port concessionaries and looking at revenue streams. This programme could provide analytical inputs to help strengthen anti-corruption measures. Benefits would include improved predictability for importers and exporters, reduced tax evasion and reduced scope for smuggling. iii) Health – 84.9% of respondents in the PCCB 2009 Anti Corruption Survey regarded health institutions as very or fairly corrupt. DFID is planning a major new health sector programme (investment level to be determined) that will focus on quality of basic services, accountability and demand. Corruption challenges in the health sector are multiple, affecting areas such as drugs procurement and supply, petty bribery, and absenteeism. This programme could provide analytical inputs into the design of the health programme, and provide a challenge function during programme implementation. iv) Education – DFID plans major spend on education, comprising £30m through sector budget support, £30m on girls’ education, £30m on improving education quality, and an imputed share of £15m of General Budget Support (since the government spends 18% of its budget on education). Although education ranks better than other services in perceptions surveys on corruption (eg ‘just’ 61% of respondents in the PCCB 2009 Anti Corruption Survey regard education institutions as fairly or very corrupt), the sector is, nonetheless, affected. At the service delivery point, the main complaints relate to demand for unofficial payments. At the higher level, risks and vulnerabilities include: the transfer of capitation grants to local government level, the lack of separate audits for the education sector (because it cuts across several government ministries), and procurement. This programme could provide analytical and advisory resources to ensure the robust assessment and mitigation of corruption risk in the new programmes. v) Water – This sector is identified not so much because corruption risk is particularly high in the water sector in Tanzania, but because of the nature of investment that DFID is planning (£30m on a rural water programme, probably through a basket fund, entailing large scale construction and procurement). DFID has already been improving financial management in the water sector, for example introducing cost norms that brought down per capita expenditure from US$90 to US$58. This programme could assist in the design of the new programme, by analysing the technical audits that have been carried out in the water sector, and mapping out corruption risks and mitigation measures in areas such as management of water at the user level, procurement of civil works and associated design, supply and consultancy services. vi) Rural roads – to date, investments in Tanzania have focused on trunk roads connecting major cities with neighbouring countries. There has been very limited investment in rural roads. DFID is considering a major scale up on support to rural roads, with spend somewhere between £50m and £150m depending on how resources are allocated across the DFID programme. Again, this will involve multiple large scale construction and procurement activities at high risk for corruption, potentially to be spent through local government and managed at district level, where scrutiny of expenditure is relatively weak. Common corruption risks in the road sector include: exaggeration of needs, manipulation of allocation formula, phantom projects, loose specifications, over-design of projects, collusion and bribery in procurement, and diverting funds under the cover of expenditure such as staffing, equipment, purchase of supplies and quality control. This programme could provide analytical inputs to assist in robust assessment and mitigation of risk in the road sector, helping to maximise the value for money for DFID’s investment. The appraisal also considered whether natural resource management and the Extractives Industry Transparency Initiative (EITI) merited targeted support through this programme. It concluded against this. The EITI process is already well supported by the World Bank, Norway and (to a lesser extent) Canada. Tanzania is progressing through its EITI candidacy, and is due to submit a second report to the EITI Secretariat, which – if accepted – will pave the way for EITI membership. The African Development Bank will then assist with the establishment of EITI as an institution in country. Norway is also engaged in wider natural resource management issues – firstly in the mining sector, where it has been advising on mining tax models, terms and conditions of contracts, and auditing of mining companies; and secondly timber, forestry and fisheries, where it is considering a new programme on effective approaches to manage finances and enhance non-tax revenue collection, and also supporting NGOs to research and monitor forestry cover under the Reduced Emissions from Deforestation and Forest Degradation (REDD) climate change initiative. Given the importance of natural resources, and particularly new oil and gas revenue, to Tanzania’s overall governance and financial management, DFIDT will be engaged generally in the sector but in terms of specific anti-corruption interventions, there are no immediate entry points. Opportunities may emerge when the second EITI report is submitted, and DFID will assess the scope for engagement at that point. In terms of increasing incentives on government to act, the appraisal found that the media and parliament already play an active role in pressing for action on corruption21. Demand from civil society, academia and the private sector has been less prominent, and should be strengthened, although it will be important not to duplicate existing efforts22. There is room for more coordinated and concerted dialogue by international partners on anti-corruption, along the lines already in place in Uganda (which entails improved analysis and agreement on corruption priorities (systemic and case-based), use of a core script, and agreed graduated responses in the event of deteriorating circumstances). More use could be made of international instruments such as UNCAC and international AML standards to lever action (eg Tanzania’s place on the Financial Action Task Force watchlist). What the DFID programme will deliver 2.20 Based on the above analysis, DFID will deliver the following outputs in this programme: Output 1. Oversight agencies work together to tackle corruption more effectively Detection: Assist the FIU to scale up and strengthen monitoring of suspicious financial transactions, and meet international AML standards. Support the NAO to strengthen its statutory financial audits in high risk areas, to increase the probability that fraud will be detected, and enhance the NAO’s capacity to conduct forensic audits to ensure that preventative measures are in place. Both audit types will also support the gathering of evidence to be used in investigations/sanctions Investigation and prosecution: Support the PCCB and its Zanzibar equivalent (once established) and the DPP and its Zanzibar equivalent in sustainable upgrade of skills for handling complex corruption cases; Assist the PCCB in ensuring Tanzania meets international standards relating to UNCAC, particularly during the 2012 review process Other sectors: Provide anti-corruption ‘back-stopping’ interventions on a demand-driven basis to specific sectors. This might include identifying corruption vulnerabilities and designing mitigating measures for incorporation into ongoing/new DFID programmes supporting the Tanzania Revenue Authority, roads, water and health. Monitor impact. Cross-linkages: Support joint working between the above institutions to improve management of corruption cases, and increase the proportion that result in prosecution or other sanction or satisfactory resolution; Facilitate a more productive working relationship between the NAO and PCCB to better identify and mitigate common corruption risks identified as a result of corruption cases uncovered 21 This helped to precipitate the resignation in 2008 of the Prime Minister and three Cabinet Ministers, as well as the dismissal of the Governor of the Bank of Tanzania 22 DFID’s Accountability in Tanzania Programme already supports the role of civil society organisations in accountability, eg: increasing access to information and engagement in budget oversight; better monitoring and advocacy on corruption in the health and natural resource sectors; and investigative journalism. Output 2. Accountability mechanisms strengthened in the police and judiciary, and in four DFID sector programmes: customs and revenue, roads, water and health Police: Strengthen oversight of the police through the Ministry of Home Affairs and within the TPF. To include central referral of complaints against the police, training for investigators dealing with police misconduct, and improved supervision and management structures for high risk police units; Increase role of communities in working with police on anti-corruption Judiciary: Improve judicial accountability systems through introduction of a performance management system, and improving the Judicial Service Commission’s capacity to respond to and investigate complaints; Strengthen public awareness of court procedures, introduce a service charter and court users’ manual and improve collection and analysis of court statistics Other sectors: Provide demand-driven anti-corruption analytical/advisory inputs to specific sectors. This might include identifying corruption vulnerabilities and designing mitigating measures for incorporation into ongoing/new DFID programmes supporting the Tanzania Revenue Authority, Tanzanian Ports Authority, rural roads, rural water, health and education. Monitor impact. Output 3. Increased domestic and international incentives for Governent to tackle corruption Domestic stakeholders: Improve understanding of public barriers to complaint about corruption, and address issues identified (likely to include improving complaints channels); Support better evidence collection and dissemination on corruption levels and trends. Use this to facilitate more productive dialogue between government and civil society groups/private stakeholders Other sectors: Facilitate enhanced coordination and dialogue by the international community on anticorruption; Analyse international levers impacting on corruption, and maximise engagement on these – to include compliance with UNCAC and Financial Action Task Force AML 2.21 DFID is also considering support for the NAO through the Government’s new five year Public Financial Management Reform Programme (2012 – 2016). Care will be taken to ensure that the two support lines are coherent and maximise synergies as best as possible. For all three outputs, frequent monitoring of impact will be critical. As mentioned above, one of the benefits of a comprehensive programme of this kind is the flexibility to scale up or scale back individual interventions depending on where greatest progress is being made and upon regular assessments of political will. D. Appraisal of feasible programme delivery options 2.22 Initially, in the interests of getting the programme initiated quickly, it will be managed by DFID advisors and programme staff. DFID will oversee the disbursement of funds to implementing organizations throughout the life of the programme. However, due to limited resource within DFID general management of the programme is not sustainable long-term. Three decisions are needed about the long term delivery of the programme. These are: (i) how to handle programme management; (ii) what channel to use for support to government; and (iii) what channel to use for external dialogue, analysis and advice. Each is considered in turn below. (i) Options for overall programme management 2.23 The programme could be overseen either by DFID itself, or by a managing agent. If the managing agent route is chosen, there is the option of either channelling support through UNDP (which is developing a new anti-corruption programme and already works closely with the PCCB), or through a private agent to be competitively procured. Cost-benefit criteria compared are: overhead costs; speed with which the arrangement can be established, and acceptability to stakeholders (including track record so far). Each is scored 1 – 5 with 5 being the best score. Option Overhead costs How fast to establish Acceptability to stakeholders Score A. DFID 1. Would represent big transaction cost for DFID given complexity of programme, and degree of influencing required. Would require one full time adviser plus local consultancy (6 – 7% of programme cost). 5. Fast. A2 advisers on board that could initiate work while additional advisory expertise was recruited in 5. Good. Should be acceptable. 11 B. UNDP 3. Standard charge is 7 – 9% for NEX+, ie where UNDP channels finances to government on behalf of DFID, manages all procurement etc. Would entail some transaction costs to ensure UNDP able to deliver the programme to the standard required 3. Medium. UNDP would have to design a proposal. It could then bring in a team within 3 months of approval 2. Medium to low Government would approve. Donors have concerns about UNDP performance, based on past delivery of weak programmes with mixed calibre staff 8 C. Private agent 3. Standard management agent fees are around 10% based on what KPMG charges for AcT, and Ernst & Young standard quotes for similar programmes. Would entail some transaction costs to handle tendering and get consultancy team operating 4. Medium to Fast if use DFID advisory expertise to get programme off the ground while going to full competitive tender (which would take 9 months) 4. Medium to good. Depends on agent selected and calibre of staff. Some more acceptable to government than others. Acceptable to external stakeholders because of perceived higher efficiency and neutrality. 11 2.24 Options A and C score the same. On balance, Option C is preferred because the transaction costs for DFID to fully manage this programme would be so high, and because it would be more difficult for DFID to play a steering/strategic oversight role if it were involved in day to day management. The private agent would do more than programme administration: it would also have a critical role in networking between actors involved in anti-corruption, influencing/dialogue, and bringing in world class international analysis and advice. (ii) Options for channelling support to government 2.25 Most technical support and training for government will be directly procured through the managing agent. Therefore this delivery channel will be used to: oversee progress against workplans; and promote more effective horizontal linkages23. Aid effectiveness principles argue against the creation of a new project implementation unit. Therefore, the options are: the National Anti Corruption Strategy and Action Plan coordinated by the PCCB; the Legal Sector Reform Programme; and the National Criminal Justice Forum, hosted by the DPP. Given that the financial cost of each channel would be similar, cost-benefit criteria are: capacity to manage resources and workplans, and ability to convene and forge closer working links between relevant agencies. As above, these are scored 1 – 5. Analysis of delivery options for support to government Option Management of workplans Management resources NACSAP III 2. Low-Medium. NACSAP II is a very broad strategy. NACSAP III has yet to be drafted but it is likely that DFID interventions would be a small part of a much broader process, and therefore receive limited attention Legal Sector Reform Programme National Criminal Justice Forum 23 of Convening power # 4. Medium-High. UNDP would provide the interface for this. Therefore, there would be reasonable resource and programme management capacity. 4. Medium-High. PCCB hosts regular NACSAP meetings attended by all relevant agencies. However as discussions are very broad, may not get specific agency representatives that need 10 2. Low-Medium. LSRP is broad and DFID interventions would be secondary to long-established objectives. Not clear what replaces LSRP post 2012. 2. Low-Medium. World Bank audit found concerns about LSRP financial management. EC audit more positive. 1. Low. LSRP coordinated through Ministry of Justice and Constitutional Affairs. Has limited leverage over PCCB and DPP. 5 5. Good. A Committee would be created under the NCJF dedicated to this work. 3. Medium. Funding would be disbursed via Treasury ensuring fiduciary assurance such as audit oversight. NCJF would provide oversight and approval mechanism for funding flows. 4. Medium-High. NCJF is a statutory body therefore can convene oversight bodies for this work. However, PCCB would need a leading role, e.g. co-chair or there may be institutional rivalries 12 Actual disbursement of funds to Government institutions via the Treasury with be undertaken by DFID (see diagram 2 below) 2.26 At this stage, co-ordination through the National Criminal Justice Forum is preferred as the most credible option. Participation of PCCB will be handled by representation at Director of Investigations level, and options for co-chairing and hosting meetings at the PCCB as well as DPP will be explored. During the inception phase of the programme, this coordination mechanism will be monitored closely (a) to ensure that the PCCB plays an active role, and (b) to verify how NACSAP III is developing and how far programme objectives can be aligned with and/or jointly delivered. If NCJF remains the main delivery channel, MOUs will be needed with LSRP (before it closes in December 2012) and NACSAP III (once operational) to ensure no duplication of efforts. (iii) Options for supporting external dialogue, analysis and advice 2.27 There are no suitable existing mechanisms to support the non-government components of the programme, and therefore an Integrity Fund will be set up. This will be used to support the advisory and analytical inputs into sectors, broader corruption analysis and trend monitoring, joint donor dialogue with government on anti-corruption, and more concerted engagement by civil society and private sector – to include issue-based advocacy with government. The Integrity Fund will be supervised by the managing agent (see Management Framework at Annex 8). Links will be established with DFID’s Accountability in Tanzania Programme, and to Tanzanian think-tank organizations such as Research on Poverty Alleviation (REPOA) and Economic and Social Research Foundation (ESRF) 2.28 Diagram 2 below summarises the programme delivery structure that will be used. Diagram 2: Summary of programme delivery structure E. Value for Money 2.29 The following measures will be used to assess the value-for-money of the programme, noting that in many cases, baseline data has yet to be determined: Effectiveness (how well our outputs will achieve our desired outcome): % of low income households paying bribes and/or the % of the public that encounter police, judiciary or magistrates and pay a bribe; % of targeted citizens who believe that Government is committed to tackling corruption in the public sector and/or % that are willing to complain about corruption; Number of corruption cases reported and/or % of reported cases investigated by PCCB and/or & % of investigated cases that are accepted for prosecution Number of complaints concerning the police and judiciary that are received and acted on; Efficiency (how our inputs are converted into outputs) increase in the number of fraud cases detected by NAO as a result of staff trained on fraud detection, new fraud detection system being installed and forensic/special audits conducted; increase in number of complex corruption cases detected and prosecuted as a result of PCCB, FIU and DPP staff being trained decrease in perceptions of corruption in traffic police in pilot areas as a result of measures introduced by programme numbers of sector programmes with robust anti-corruption measures as a result of advisory and analytical inputs provided Economy (inputs being purchased at the appropriate quality and right price) Overhead costs of the managing agent will be monitored. These should not exceed 10% of the cost of the programme, as benchmarked against programmes with similar management structures; Unit costs for the provision of technical assistance and training will be monitored to ensure high quality assistance at a reasonable price. The managing agent will be asked to benchmark this against the equivalent provided in Tanzania and internationally. 2.30 Finally, we will also monitor qualitatively the extent to which the interventions funded leverage outcomes (i.e. catalyse wider reform efforts) beyond DFID’s own contribution. 3. Commercial Case A. Clearly state the procurement/commercial requirements for intervention 3.1 This programme has several components, and will involve both direct and indirect procurement. The diagram below summarises the procurement structure that will be used. Diagram 3: Procurement Structure for Programme ` 3.2 The direct procurement will be in the form of a contract competitively tendered to a managing agent. The agent will procure analysis, technical assistance and advocacy work on behalf of DFID. The technical assistance will be provided both to government and external bodies. The agent will also be responsible for managing overall programme reporting. 3.3 The money channelled through the government via the Treasury and National Criminal Justice Forum (NCJF) will be used to support coordination of oversight institutions, and to fund selected key results in the beneficiary institutions’ workplans. The proposed design means that the indirect procurement arrangements for the intervention will follow the Government of Tanzania procurement systems. A standard Memorandum of Understanding will be signed by the Treasury (Ministry of Finance) and the recipient Institutions. The Treasury will disburse funds according to requests upon approval of request by the NCJF and DFID, and will subject those funds to the same scrutiny as other Government of Tanzania expenditure, including auditing. B. How does the intervention design use competition to drive commercial advantage for DFID? 3.4 For the direct procurement element, DFID will conduct a competitive tender under full European Union competition procedures (see Section E). In order to get the best value for money, the Terms of Reference (TOR) will set out clear expectations of the supplier, with outcome-focused specifications. The contract will also adhere to DFID standard terms and conditions (eg relating to travel and subsistence), thus keeping costs down. Milestone payments will be used to pay the contractor upon satisfactory completion of the outcomes set out in the TOR. 3.5 For the indirect procurement through government, DFID will make an initial payment, then make all future payments dependent on achievement of key results in the oversight agencies’ workplans. This ‘payment on results’ will increase the ownership and commitment of government to achieving the programme outcomes. The results will be clearly linked to the programme’s intended inputs and outputs, and will be both measurable and attributable to DFID’s support. DFID will pay for the DPP to recruit a procurement agent (who could be seconded from the managing agent, or sourced within government) to ensure that all procurement of goods or services under this project represents value for money and complies with Government of Tanzania procurement standards. All procurement under the programme will be subject to spot audit checks to verify compliance. DFID will also pay for the DPP to recruit a reporting officer (seconded from within government, or through the managing agent) to help track programme progress and submit reports on results and impact. 3.6 The default position is that DFID funding will not be used to pay for per diems, workshops or study tours unless explicitly agreed in advance with the managing agent and DFID, and linked to clear deliverables. C. How do we expect the market place will respond to this opportunity? 3.7 The appraisal process has already shown that government is ready to collaborate on this programme. For the direct procurement, several companies are likely to be interested in the managing agent role. These include the main management consultants – KPMG, Ernst and Young, Deloitte, PWC and Adam Smith International – as well as the British Council which has run such multi-component anti-corruption programmes in a number of other countries (eg Sierra Leone and Nigeria), and has expertise in the police and judiciary sectors as well. The bigger firms may wish to partner with specialist organisations focusing on police and judiciary matters, eg Blue Light Consulting, and South Africa’s Institute for Security Studies. 3.8 There are also likely to be several Tanzanian civil society and private sector organisations interested in carrying out analytical and advocacy work through the Integrity Fund. These include REPOA, ESRF, the Uongozi Institute, and other Non State Actors such as the Trade Union Congress of Tanzania (TUCTA), Tanzania Private Sector Foundation (TPSF) and Tanzania Association of NGOs. D. What are the key underlying cost drivers? How is value added and how will we measure and improve this? 3.9 The key underlying cost drivers relate to: personnel (the overheads for the managing agent, the costs of technical assistance provision, and the support staff in the DPP), procurement, and delivery of activities in workplans (government, bodies funded through the Integrity Fund). 3.10 On personnel, the key is to bring in the highest quality people possible at a reasonable price. Clear criteria will be specified for the quality of personal required, and their costs benchmarked against similar providers. Pay will be directly linked to performance and outputs. The managing agent’s other overhead costs (such as transport, communications) will be scrutinised to ensure they are comparable with costs that DFID itself would incur. 3.11 On procurement, the programme will need to ensure timely procurement of good quality goods and services that can deliver the tasks required at the best possible cost. This will be done by requiring the managing agent to use DFID (or equivalent) standards for procurement, and by ensuring that the government complies with its own standards for procurement. Compliance will be verified through spot checks. Any large value procurement items (over £100,000) may only be signed off with the approval of DFID project manager with the delegated authority, unless if it is a sensitive procurement. 3.12 The programme will need to ensure workplan activities are high quality and likely to achieve their desired impact, that they are linked to clear and measurable results, and that they are delivered at a reasonable cost. This will be achieved by DFID and the managing agent working with government counterparts and external stakeholders to agree workplans with realistic and measurable results, and by benchmarking the cost of activities against similar activities carried out in Tanzania and elsewhere. DFID’s funding for government and for the managing agent will be in the form of ‘payment by results’, ie providing the next tranche of funding when the previous sets of milestones have been met. E. What is the intended Procurement Process to support contract award? 3.13 All contracts awarded by this project will be in line with DFID procurement Guidelines. Exceptions will be applied only when absolutely necessary, in which case appropriate waivers will be sought. 3.14 DFID will go out to competitive tender to contract the managing agent. DFID will do this following the European Union Restricted Procedure for tendering, namely: The contract requirement (full TOR and evaluation criteria) will be advertised in the Official Journal of the European Union. All interested suppliers will be invited to submit expressions of interest to DFID Tanzania. This will take 35 days. Any companies that are known to have particular competency in this area will be alerted to the upcoming Request to Participate. DFID will then select a shortlist of suppliers who will be invited to tender. Finalised TOR and evaluation criteria will be shared with the suppliers. The submission of full tenders will take another 35 days. DFID will then evaluate the tenders together with a reference group comprising the DPP, PCCB, police and judiciary, and one representative each from civil society and the private sector. Based on the result of this evaluation, the preferred supplier will be selected. 3.15 This tendering process is anticipated to take four months, with a further two months for the chosen supplier to prepare to take up the contract. The chosen supplier will therefore be expected to commence work at April 2012. F. How will contract and supplier performance be managed through the life of the intervention? 3.16 Contract and supplier performance will be managed through existing established mechanisms from DFID Headquarters Procurement Group and DFID Tanzania procurement unit for the managing agent. Managing performance is covered by the monitoring and evaluation section in the Management case. 4. Financial Case A. How much it will cost 4.1 The estimated cost of the programme is £11 million over 4 years. It is also expected that Government Oversight Institutions will contribute around £1 million in total through their usual activity budgets in their respective Medium Term Expenditure Frameworks (MTEFs), plus significant resourcing in terms of staff time and running costs of the institutions. 4.2 An indicative breakdown of costs for the programme is in the table below. This is likely to be subject to change once detailed workplans have been developed with the oversight institutions, police and judiciary, and the contract has been awarded to the selected managing agent. Programme intervention 2011/12 2012/13 2013/14 2014/15 Total Support through government Coordination Tanzania Police Force Judiciary (Primary Courts) DPP PCCB NAO FIU Sub-Total 40,000 200,000 100,000 300,000 300,000 100,000 75,000 1,115,000 85,000 1,000,000 400,000 400,000 500,000 250,000 150,000 2,785,000 75,000 1,000,000 400,000 400,000 500,000 250,000 150,000 2,775,000 75,000 800,000 300,000 300,000 400,000 200,000 75,000 2,150,000 275,000 3,000,000 1,200,000 1,400,000 1,700,000 800,000 450,000 8,825,000 100,000 70,000 120,000 50,000 335,000 100,000 65,000 90,000 50,000 305,000 300,000 205,000 400,000 150,000 1,025,000 Integrity Fund Corruption trend analysis Civil society/private sector dialogue Analytical/advisory inputs in sectors Donor joint response/dialogue Sub-Total 0 0 40,000 0 50,000 50,000 100,000 70,000 120,000 50,000 335,000 40,000 40,000 315,000 65,000 380,000 320,000 65,000 385,000 245,000 100,000 345,000 920,000 230,000 1,150,000 1,205,000 3,500,000 3,495,000 2,800,000 11,000,000 Programme management, monitoring and evaluation Programme Management Monitoring and evaluation Sub-Total Grand Total 4.3 In addition, the Government of Tanzania will contribute an estimated £1m through its core funding for the government institutions identified above, for related work. Adequate funding is available within DFID Tanzania’s pipeline for this programme, and it is within the delegated authority of the Head of Office to approve. Funding is not considered to be contentious, as the content of this programme has already been discussed with the Government of Tanzania and is similar to programmes being supported in other countries, such as Zambia and Nigeria. 4.4 As this programme is high risk and has multiple components, six monthly reviews will be built in to assess how far each component has traction and is achieving milestones, and funding may be moved between programme interventions as appropriate. A monitoring and evaluation budget line has been included to allow for frequent independent reviews of progress. More detailed monitoring and evaluation will be built into the oversight institution programme lines. B. How it will be funded: capital/programme/admin 4.5 The intervention will be funded from programme resources (RDEL), under the DFID Tanzania framework and has been budgeted for in the Operational Plan for DFID Tanzania. While funding may be shifted between different programme intervention lines (see 4.4 above), the overall totals to be spent each year will remain consistent with the table above. There should not therefore be any higher level impact on DFID’s forecast of spend in Tanzania. C. How funds will be paid out 4.6 The diagram at the beginning of this section illustrated the delivery channel for funding: one to government (indirect procurement), and one to the managing agent for the remainder of activity. 4.7 The first tranche of funds will be paid to the Government of Tanzania for the oversight institutions, police and judiciary, on receipt of satisfactory work plans with clear and measurable results and milestones. Subsequent tranches will be paid on results (i.e. achievement of milestones). Payments will be made annually. 4.8 The first tranche of funds will be paid to the managing agent on receipt of a satisfactory programme action plan. Subsequent tranches will be paid on achievement of contract milestones, and will require submission of valid claims, invoices and proof of completion of work. 4.9 In the event that it is deemed necessary to make a payment in advance of need, DFID will contact the Finance, Corporate and Procurement Department to justify the request and seek approval. D - What is the assessment of financial risk? 4.10 The main financial risk associated with the project relates to channelling funds through government. DFID’s 2011 fiduciary risk assessment for Tanzania identifies the overall level of risk as substantial. According to the FRA, the main areas of risk include: (i) limited independence and effectiveness of external audit; (ii) external pressures and weak commitment control systems for large contracts; (iii) a weak internal control environment; (iv) the lack of transparency and oversight of the internal transfer system to public authorities and other bodies; and (v) poor value for money in procurement practices 4.11 These risks will be managed by (a) limiting the level of funding to be directly managed by government institutions, e.g. providing technical assistance in kind; (b) paying funds to institutions through the Treasury thus ultimately achieving more GoT audit coverage; (c) requiring separate audit of DFID funds provided; (d) requiring spot checks of procurement carried out with DFID funding, as well as DFID/managing agent co-signature of large procurement items; and (e) linking payment to achievement of milestones. 4.12 There is also a risk that the managing agent itself, or the recipient institutions of the Integrity Fund, misuse funds. Due diligence will be conducted on the managing agent before funds are disbursed, to ensure adequate checks and balances are in place, and the managing agent itself will also conduct due diligence checks before disbursing funds to partners. In addition to annual audits, high risk areas of expenditure such as procurement and workshops will be subjected to additional spot-checks to verify their legitimacy. E - How expenditure will be monitored, reported, and accounted for 4.13 Annual audited accounts, six monthly financial statements and budget forecasts will be required for each institution receiving DFID funding. Further funding will only be provided once prescribed milestones have been met and previous funds accounted for. Agreements with partners will include an exit strategy where fraud is identified or the expected results are not being achieved. 4.14 The management agent will compile annual assets returns for DFID on behalf of government and others receiving DFID support. Assets purchased by government will be subject to Government of Tanzania regulations and will be disposed of or transferred to the individual institutions when the programme comes to an end. Assets purchased by the managing agent and other institutions will be similarly disposed of at the end of the programme, in agreement with DFID. 5. Management Case A - Oversight 5.1 DFID will be responsible for oversight of this programme, through the Public Sector Advisor. The Deputy Head of Office will quality assure this programme and ensure it delivers the quality outcomes intended. The managing agent will retain oversight of the workplan and inputs provided. 5.2 Co-ordination of support to oversight institutions, the police and judiciary will be through a subcommittee of the National Criminal Justice Forum, which will steer their overall direction of work, and review impact. The DPP will co-chair the sub-committee, and all the key oversight institutions will be represented. DFID and the managing agent will be observers to the group, and their consent will be required for sign off of workplans and budget requests. Any technical assistance provided by the managing agent to the oversight institutions will be in line with the workplans and approved by the NCJF. 5.3 The Integrity Fund, which will be administered by the managing agent, will be the vehicle for support to external bodies – civil society and academia, the private sector, and coordination of international partners. It will have a small steering committee comprising representatives from each of the above, as well as DFID and the managing agent, to help shape the overall direction of the Fund and review its impact. B - Management 5.4 DFID will have responsibility for overall programme management, led by the Public Sector Adviser, with support from the Deputy Programme Manager on issues such as contracts, payments and reporting. 5.5 For the first six months of the programme, DFID will lead on programme start-up with the help of locally recruited consultants, while the tendering process proceeds for selection of the managing agent. This will comprise: developing workplans with the respective oversight institutions and carrying out initial disbursements to government, as well as initial analytical and advisory work through the Integrity Fund (eg to input into sector level work). Thereafter, day to day programme management will be handled by the management agent, whose duties will include: programme planning, managing short-term consultancies, making sure programme management is carried out in line with standard DFID procedures; budgeting and forecasting; and submitting reports for the project. The agent will also carry out dialogue and influencing with partners, and manage productive working relations with them. Within the NCJF, a reporting officer will be responsible for overseeing workplan progress by the oversight institutions, and submitting narrative and financial reports to the NCJF sub-committee and managing agent. C - Conditionality 5.6 None. D - Monitoring and Evaluation 5.7 The results framework for this programme is attached at Annex 3. This will be supplemented by more detailed workplans for the managing agent and for each of the government beneficiary institutions. These will have clear, measurable results that are time-bound, and will form triggers for subsequent disbursement. The main anticipated activities to be included in each workplan are attached at Annex 4. DFID will carry out an annual review of the programme each year, during which milestones and indicators will be refreshed as needed. 5.8 Earlier this business case stated that the evidence for this intervention is medium. Because of the limited international evidence on effectiveness of anti-corruption measures, as well as the scale (£11m) and degree of innovation and risk in this programme (especially the support to police and judiciary), an independent impact evaluation will be carried out at the end, including a detailed review of the support to the police and judiciary. This will draw on the indicators set out in the attached results framework and the theory of change. It will use evidence gathered through programme monitoring: a monitoring plan will be put in place for each beneficiary institution, with clear baselines, milestones and targets, to allow for assessment of progress. It will report on the extent to which programme outcomes have been achieved, lessons learned in the provision of support, and scope for replication in other countries. A more detailed evaluation plan will be developed by DFID in the first six months of the programme. F. Risk Assessment Risk Description Impact on Success (L,M,H) Probability (L,M,H) Mitigating Actions Risk 1 Programme relationships negatively impacted by wider pressures building in GBS/SBS relationship around underlying principles etc. H M - Strong communication of results - Strong relationship with the heads of implementing institutions Risk 2 The UK seen as not being the right bilateral donor to lead on an anti-corruption programme because of its own recent corruption record in relation to Tanzania M M - Provide world class expertise – if the product is useful, this will overcome concerns - The recent introduction of the UK Bribery Act may help to counter the negative impressions of the UK on corruption Risk 3 The Judiciary is not fully committed to working with the UK on tackling corruption and may therefore block work in this area M L - programme of support to the judiciary will only take place under strict conditions including clear commitments from the judiciary themselves (to be initiated through early high-level engagement with the Chief Justice) - the programme’s focus on the primary courts may offer a less contentious entry point than seeking to tackle higher level corruption Risk 4 Programmes supporting police reforms or working across the security and justice sectors are normally categorised as high-risk. The possibility of high level political/public criticism is present especially in the event of allegations of a human rights breach. M M - Capitalise on the Inspector General of Police’s commitment to act. Ensure that interventions supported have a clear benefit to the police and judiciary - Closely monitor progress and be prepared to alter the programme or cease support if there is no traction - monitor closely (through CGA and other mechanisms) HR record of police Risk 5 Consultancies do not deliver the high quality work needed M M - clear specification of consultancy requirements and outputs will help mitigate this risk - use of results based management, checking that milestones have been achieved before disbursing next payment. Risk 6 Corruption is uncovered in the programme, undermining its credibility and forcing DFID to suspend funding M H - due diligence of NCJF and project partners before disbursing funding - development of clear workplans with measurable results, and payment by results - annual audits with in-depth checks on high risk spend such as procurement and workshops - DFID will authorise any procurement over £100,000 - annual reviews to look at corruption prevalence and risk, and whether managing risk adequately - publish information on support given to each institution, and establish complaints mechanism to facilitate reporting of any abuses. Acronyms ADB AML CAG CIDA CJ CJF CPI FC CSOs DFID DPP DPs EITI ES EU FATF FBO FIU GAAP GOT HQ I/E IGP IEC KPIs LSRP M&E MDGs MIS MKUKUTAII MoHA MOU MTEF NACSAP NCJF NGOS PCCB PPRA PRBS TA TRA UNCAC UNDP African Development Bank Anti money laundering ( Control and Auditor General (), the Canadian International Development Agency Chief Justice Criminal Justice Forum Transparency International’s Corruption Perception Index ( Feasibility Criteria Civil Society Organizations Department for International Development The Director of Public Prosecutions () Development Partners Extractive Industries Transparencies Initiative Ethics Secretariat European Union Financial Action Task Force on Money Laundering Faith-based organisation Financial Intelligence Unit (the Generally Accepted Accountancy Principles Government of Tanzania Headquarters Income/expenditure Inspector General of Police Information, Education, and Communication Key performance indicators Legal Sector Reforms Programme Monitoring and evaluation Millennium development goals Management information system Tanzanian poverty reduction strategy Ministry of Home Affairs Memorandum of Understanding Medium-term expenditure framework National ANTI Corruption and Strategic Action Plan National Criminal Justice Forum Non-governmental organisation Preventing and Combating of Corruption Bureau Public Procurement Regulatory Authority Poverty Reduction Budget Support Technical Advisor Tanzania revenue Authority UN Convention Against Corruption United Nations Development Programme