Rakner/Wang Draft December 2006 The Parliamentary Bottleneck: External Audits in Malawi, Tanzania and Uganda Lise Rakner and Vibeke Wang Draft prepared for the Norwegian Political Science Association’s Annual Meeting, Trondheim 3-5 Jan. 2007. The combined process of economic and political reforms that have influenced the ‘good governance agenda’ of the international donor community since the early 1990s has strongly emphasised economic accountability. As a result, sub-Saharan African countries have been grappling with economic accountability to make sure that allocations from the public purse go where they are supposed to and that spending complies with policy objectives. Central institutions in this regard are the supreme audit institution (SAI) and parliament.1 The SAI is the national body responsible for scrutinising public expenditure. Parliament is tasked with ex post scrutiny of the SAIs audit findings to ensure that the national budget is implemented fully, efficiently and effectively. The bulk of this work takes place in the parliamentary committees. In countries modelled on the Westminster tradition of external auditing, the legislature is the primary audience of the SAI. Legislatures become part of the audit cycle at the oversight stage of the budget process and it is at this stage legislature-SAI interaction occurs. Notwithstanding the fact that economic accountability demands a well-functioning legislature-SAI relationship, the nexus between the two institutions in newly democratising states has not been subject to extensive research. A central problem, in our view, is that the literature on parliament and SAIs is largely separate and there is little focus on the interaction between the two institutions. The ongoing national level poverty policy processes with a high level of influence by international development partners have only enhanced the importance of the interdependence of the legislature and SAI in many sub-Saharan African countries. Still, aid assessments made only rarely take into account the relations between the SAI and parliament, but instead focus on the individual institution in isolation. The result is two literatures and two aid streams which by and large are disconnected. This article lifts the veil on an important feature of today’s emerging democracies, namely economic accountability and the linkages between the SAI and parliament by undertaking an empirical analysis and comparison of the interplay between the SAIs and the legislatures in Uganda, Tanzania and Malawi. 1 Rakner/Wang Draft December 2006 The countries display a high degree of donor dependency and there are ongoing poverty reduction strategy processes in all three countries. These processes provide the basis for assistance from the World Bank and the International Monetary Fund, debt relief under Heavily Indebted Poor Country initiative, and also largely guide the aid provided by the other members of the international aid community (both multilaterals and bilaterals). The poverty reduction strategy paper (PRSP) is intended to be a steering document for poverty reduction and constitutes a point of departure for developing countries’ dialogue with development partners. With a ‘home grown’ PRSP Uganda is the first African country to embark on a third PRSP while Tanzania developed its second generation PRSP in 2005. Malawi, being a relative latecomer, is still in the process of implementing its first PRSP. Much due to the international trend towards a shift in aid modalities from project aid to various forms of programme aid, and recently also general budget support economic accountability is high on the three countries agenda. This entails a greater amount of aid being ‘on budget’ and thus allocated in the national budget process. Their system of governance also sets them apart. Uganda stands out with its no-party National Resistance Movement (NRM) system which was formally transformed into a multiparty system before the Feb. 2006 presidential and parliamentary elections. Both Tanzania and Malawi introduced multiparty systems in the early 1990s, but so far no major political shifts have taken place in Tanzania. The top-down transition was guided by the ruling party Chama Cha Mapinduzi (CCM), and the CCM has enjoyed an overwhelming majority in parliament since the first multiparty elections in 1995. In contrast, the three democratic elections in Malawi (1994, 1999, 2004) have not resulted in a dominant party system. The United Democratic Front (UDF) defeated ‘president for life’ Hastings Banda and the Malawian Congress Party in the first multiparty elections in 1994. While UDF retained power in the two subsequent elections, the party has had to rely on support from the opposition through various coalitions. This has necessitated negotiations with sections of the opposition to facilitate the passing of legislation. In our analysis of the legislature-SAI relationship in Malawi, Tanzania and Uganda we find that all of the countries struggle with basic issues of economic accountability that are accentuated by weaknesses in the institutional bonds between the two institutions. 2 The countries may be placed on a continuum according to their degree of institutionalisation of the legislature-SAI relationship. Malawi which is marked by institutional instability and unwarranted donor influence is found in one end of the scale. Tanzania is in a middle position. Relations are relatively well institutionalised, but mechanisms of accountability are 2 Rakner/Wang Draft December 2006 contested by the well entrenched hegemony of the ruling party CCM. Of the three countries Uganda displays the most institutionalised system of economic accountability. However, in the wake of the introduction of a multiparty system there are indications that the room for dissenting views gradually has been curbed by the Museveni led NRM/NRM-O3 that has been eager to confirm its position at the top of the political system. In the first part of this article we succinctly review the literature available on African parliaments and SAIs before outlining a conceptual framework for an analysis of the interplay between the legislature and the SAI in the audit process. The article proceeds in the second and main part to analyse this relationship in Uganda, Tanzania and Malawi. Variables looked into to assess this relationship are the support to the committee system, party discipline and follow up of audit findings. The mandate, capacity and autonomy of the SAIs are also briefly considered to illustrate the limitations placed on their operations. Institutional Interdependence under the Westminster Model The SAIs in Uganda, Tanzania and Malawi are modelled on the Westminster system of external auditing,4 which is cyclic in nature. The audit stages may be seen as interdependent and for the audit to be successful each stage must function satisfactorily. First the government accounts for the previous financial year are prepared by all government departments and public bodies and made available on a prearranged time for the SAI to audit. The SAI audits the accounts and produces its report(s) to parliament. The Westminster model is typically marked by SAIs interactions with parliament. The SAI is a core element of parliamentary oversight, and there are close relations between the SAI and the parliamentary committees. In parliament the audit report is referred to the relevant parliamentary committees and they examine and debate the SAI report and normally conduct public hearings. When finalised, a report with recommendations and comments is handed over to the plenary, where it is debated and voted on. In the event of approval the parliamentary report is conveyed to the executive. In Malawi, Tanzania and Uganda the executive is required to respond to the parliamentary report in the form of a treasury memorandum. If the executive decides not to address and act upon a certain issue raised in the report this should be accounted for in the treasury memorandum as well the measures taken to implement other committee recommendations. Throughout, the process is to take place within certain timeframes. When delays or complications occur at one stage, this will impact on the efficiency and effectiveness of the work at the remaining stages. 3 Rakner/Wang Draft December 2006 Part I: Auditing in Theory: The Role of Parliament and SAI On the whole little research exists on parliament and SAIs in African countries, and particularly there is a void with respect to the nexus between parliament and SAI. This can in part be attributed to lack of interdisciplinary dialogue as the work relating to the two institutions is dominated by scholars from different ‘camps’. Political scientists have traditionally focused their attention on parliaments while economists, auditors and accountants have been occupied with SAIs. This has reinforced an artificial divide and contributed to largely separate bodies of literature. This dissuades perspectives explicitly linking the work of the two institutions. Supreme Audit Institutions No large body of literature on SAIs exists and the approach to studying SAIs has generally been very technical in nature and dominated by professionals in the field of audit. Current debates encompass for instance trends within auditing and different auditing techniques. A great deal of the literature originates from the International Organisation of Supreme Audit Institutions (INTOSAI) and international professional accountancy bodies and is westernbased. It should be noted, however, that there is an emerging literature on SAIs in Eastern and Central European countries,5 and a few attempts have been made to study SAIs in developing countries. The latter generally comes as a result of donor funded programmes and typically is part of larger assessments of fiduciary risk.6 Only few authors draw on an institutional approach when studying SAIs and only some of these specifically deal with relations between SAIs and parliamentary committees.7 What these institutionally centred authors have in common is the focus on independence as a cornerstone to a successful audit institution.8 Independence relates to the constitutional and legal status as well as the appointment and removal procedures. Dewar 9 also points to the relationship to parliament (SAI should not be subject to parliamentary control on operational matters); freedom from staffing controls and autonomy in budget and financial matters; operational independence (access to necessary information); and finally, reporting freedoms. These criteria are also included in Dye and Stapenhurst’s10 outline of conditions. They are categorised into: supportive environment, clear mandate, independence, adequate funding, facilities and staff, and sharing of knowledge and experience. Others use INTOSAI’s 1977 Lima Declaration of Guidelines on Auditing Precepts as a point of departure. This has become an internationally accepted standard which SAIs generally employ as it contains a 4 Rakner/Wang Draft December 2006 comprehensive list of all goals and issues relating to government auditing. Franz Fiedler,11 secretary general of INTOSAI distinguishes organisational independence (relates to Constitutional mandate, removal and appointment procedures, human resources), functional independence (freedom from constraints and/or third party conditions in its activity) and, financial independence. Wojciechowski12 applies the same categorisation but adds operational independence (the SAI should enjoy a high degree of autonomy and freedom of initiative) to the list. Although the wording and categories used may be different, the literature largely highlights the same factors as conducive for successful audit institutions. Parliament A substantial volume of literature exists on the role of legislatures in established Western democracies and the US13 but there are comparatively fewer studies on African parliaments, this pertains to both theoretical and in-depth empirical studies. There are, however, a few notable exceptions. The literature on African parliaments may roughly be divided into independence, post-independence and ‘third wave’ literature. 14 The independence literature was mostly concentrated on the weaknesses of the legislatures in policy and lawmaking. A few of the studies, however, also referred to alternative functions of the legislature. 15 In the post-independence period generally little attention was paid to the role of legislatures 16 partly reflecting the 1970s ‘modernisation school of development’ thesis claiming that a government unrestrained by a strong legislature is necessary for development in developing countries,17 and/or the view that parliament operating within the confines of a single-party state plays a largely insignificant role. As the third wave of democratisation swept the African continent in the 1990s scholars drew attention to the conceptual framework of accountability 18 and a revitalised interest in African parliaments emerged. There was a gradual acceptance of the role of the legislature in holding the executive to account, also in the South. One of the outcomes was a growing focus on African parliaments by academia as well as development partners. Due to the latter an increasing number of commissioned studies in which the role of the legislatures also is assessed as part of the overall framework of governance has materialised in addition to a small body of scholarly literature.19 It has been observed that the most common comparative statements about legislatures have been made in relation to the importance of the legislature in the policy-making process,20 relative to the importance of the executive. 21 Consistent with our focus on the interaction between SAIs and parliament, considerable scholarly attention has more recently been paid to parliamentary committees and a growth of the centrality of committees as a global 5 Rakner/Wang Draft December 2006 phenomenon has been noted.22 Strong committees must be valued to be at least a necessary condition for effective parliamentary influence in the policy-making process. However whether they are sufficient is not as evident.23 The implication is that the more specialised a legislature is, i.e. the more elaborate its committee system, the greater viscosity it has.24 This also will impact on its ability to fulfil its role in the audit cycle. There is a strong link between committees and political parties.25 The prevailing assumption is that the more important the parties are, the less important the committees and vice versa.26 The same logic applies to the degree of party discipline imposed and the frequency with which the legislature constrains the government. This is irrespective of how large majority the governing party has in the legislature.27 A common denominator cutting across the different theoretical strands outlined above is a predominant focus on the West. As SAIs and parliaments in sub-Saharan Africa to a large extent are faced with context specific challenges it is necessary to review the theoretical perspectives in light of this and avoid uncritically applying them in a developing country context. We find that this particularly has proven to be a challenge with regard to the standards against which SAIs are evaluated. Moreover these are two rather separate traditions dominated by scholars largely originating from different disciplines. This in turn may partly explain the lack of interdisciplinary collaboration and inadequate focus on legislature-SAI interaction. Why and When Do Audits Succeed? The interplay between the SAIs and parliaments in Malawi, Tanzania and Uganda in the audit cycle will be analysed on the basis of what may be termed the SAIs relational resources. Partakers playing into the audit cycle are parliament, development partners, civil society organisations, other special institutions of constraint, and the media. Notwithstanding the importance of these actors the SAI particularly relies on parliament to be its primary audience. While our main focus is the relational and dynamic aspects affecting the performance of the SAI and parliament, the general conditions under which the SAIs operate will first be illustrated by briefly looking at their mandate, capacity, and autonomy. 28 This is to display the basic limitations influencing their institutional capabilities. A SAI’s mandate refers to its formal independence as laid down in the legal framework such as the constitution, laws, and regulations. The nature of the SAI mandate involves who it itself is accountable to and reports to. It is also composed of other aspects e.g. scope of responsibilities. The capacity of a SAI depends on its financial and operational independence. Of relevance in this regard are the 6 Rakner/Wang Draft December 2006 organisational resources (human resources and organisational structure), finance (resources at its disposal/overall scale of funding), and infrastructure (access to computers, offices, training etc.). A SAI’s autonomy points to the extent to which it is able to carry out its mandate without any interference. Critical factors include appointment, tenure, independent sources of funding, and access to information. There is no definite ‘best practice’ with respect to all of the factors listed. Tenure arrangements, for instance inevitably face a trade-off between independence, continuity, and the need to ‘guard the guardians’. In principle appointments ‘for life’ provides the best security of tenure. However, a problem may occur if the person in office turns out not to be of a high moral standard and integrity. Serving on contracts may act as a safeguard against this. But this model may create a window of opportunity for the executive to exert undue influence, especially at the time of renewal of the contract. In this perspective a single, nonrenewable term is favourable. The Capability of Supreme Audit Institutions to Check Public Finance Overall, it may be argued that there are weaknesses in the SAIs’ mandate in Tanzania and Uganda. Executive dominance is a particular threat in the reporting procedures in Tanzania as it reports to parliament through the ministry of finance whilst the Uganda and Malawi SAI reports directly to parliament. In Malawi and Tanzania the de jure framework allows for comprehensive audits. New legislation has increased the scope of auditing and reduced the level of expenditure outside the mandate of the SAIs in all three countries. However, the large amount of expenditure labelled ‘classified’ and kept outside the jurisdiction of the SAI is a concern in Uganda. Audits have until recently not covered state house, the president’s office and the ministry of defence which traditionally receive a significant amount of supplementary funds throughout the year.29 For all three countries, the scale of funding is inadequate and serves as a constraint on their capacity. 30 Furthermore infrastructure and human resources are a problem. On a comparative scale, Malawi has made less progress in ensuring that the staff receives formal training in accountancy than in neighbouring Tanzania and Uganda.31 The challenges facing the SAIs are exacerbated by the introduction of an integrated financial management system (IFMS) in government, a multiplication of auditees and performance audits in all three countries. IFMS capacity building has been centred on government ministries and SAIs have been sidelined, illustrating that some of the initiative within the general framework of poverty reduction strategy programmes have largely ignored the SAIs and thus contributed to 7 Rakner/Wang Draft December 2006 marginalise them. Overall, however, the Uganda SAI has better basic facilities than the Malawi and Tanzania SAIs. 32 Significantly also, the SAIs’ workload has been steadily augmented in the wake of local government reforms and a restructuring of the financial framework. This has added the responsibility of auditing statutory corporations and local authorities. In accordance with the international trend in auditing, a new type of audit, performance audit, was included in the offices’ responsibility in 2003 (Malawi), 2001 (Tanzania) and as early as in 1995 (Uganda). Development partners generally have endorsed this development as it is in agreement with international standards of auditing. Performance audits represent quite a different approach to audit than the traditional types such as financial and compliance audits and are cost demanding since time and resources must be invested in training and building of competence. Performance audits were not taken into account when the institutional structure was established in any of the countries. At the moment the three countries’ audit systems do neither have the necessary expertise nor the financial resources to carry out adequate performance audits. This leads us to question the rationale behind introducing performance audits in countries struggling to carry out the basic types of audit. SAI autonomy is compromised in all three countries by appointment procedures, insecurity of tenure, insufficient budgets, and access to relevant and timely information. The president appoints the AG posing a challenge to his independence. Grounds for removal of the AG also leave ample room for exercise of presidential discretion.33 In Malawi and Uganda the AG serves on 5 year contracts which can be renewed once,34 while in Tanzania the AG serves until reaching retirement age. In all countries we note that an insufficient budget forces the AG to spend time lobbying donors and the ministry of finance for money. 35 This also makes them susceptible to the informal sphere of politics, putting at risk the independence of the office as well as the AGs personal integrity. The autonomy of the SAI is further indicated by whether it has access to the relevant information—and on time. The quality of internal audit within spending units is, however, weak in all three countries. It adds to the workload of the SAIs and diminishes the effect of external audits. Despite improvements, especially in Uganda and Tanzania, the statutory national financial reporting deadline is repeatedly not met by a number of units, further constraining and delaying the work of the SAIs. This diminishes the practical relevance of the audit report in the budget and financial management process. While Uganda 36 is nearly up to date with most of its reports, and Tanzania 37 has made significant progress, the Malawi SAI was a year behind in reporting to parliament in 2003.38 This short sketch of the institutional capabilities of the SAIs in Malawi, Tanzania and Uganda indicates that there are significant shortcomings both in terms of the mandate, 8 Rakner/Wang Draft December 2006 capacity and autonomy of the audit institutions. The Malawi SAI lags behind both Uganda and Tanzania in terms of ensuring that its’ staff receives formal training enabling them to complete the audit plan, while the Uganda SAI has better basic facilities than the Malawi and Tanzania SAIs. Overall the capacity problems appear far grater in Malawi than in Tanzania and Uganda. The Uganda SAI have come the furthest in terms of sticking to the critical deadlines. Context specific issues such as the external presence of donors play into and complicate the situation in all countries. Adding the responsibility of conducting performance audits to the SAIs responsibilities could be seen as overly ambitious. At the moment it just seems to widen the gap between the de jure and de facto framework of audits in all three countries. Against this backdrop of limited institutional capabilities the inter-linkages between the SAI and parliament will be explored. Part II: The Impact of SAIs: Interactions with Parliament Economic accountability under the Westminster model of auditing ultimately relies on a good working relationship between the SAI and the legislature. The legislature should hold government agencies accountable and take appropriate action when presented with the SAIs findings – unless audit findings are followed up and acted upon the problems will persist. Accordingly the legislature must be diligent in its effort to ensure that the government implements its recommendations or at least explains why changes have not taken place. An actively engaged and knowledgeable national assembly is thus required for the system to function properly. Parliamentary scrutiny of audit opinions and reports takes place in the committees. Each of the case study countries has a public accounts committee (PAC) to scrutinise audit reports on government ministries and departments and separate committees have been assigned the task of dealing with the audit reports on local government accounts. Uganda is the only country in which a special committee examines the audited accounts of publicly owned companies. All of the committees are according to the legal framework empowered to summon officials to stand before them (conduct hearings) and all of them exercise their right to do so. Nonetheless there are considerable differences in the way they are functioning. In the following we will assess the legislature-SAI relationship by looking at the support to the committee system, party discipline and follow up of audit findings. 9 Rakner/Wang Draft December 2006 The relative strength of the legislatures in acting upon SAI reports The PAC of the Uganda parliament is renown for its vigorousness. It has summoned various departmental accounting officers and ministers to stand before it and has on occasions even locked accounting officers up in the parliamentary cell for showing contempt of the committee. 39 The local authorities’ accounts committee (LAAC) 40 and the committee on parastatals have been less effective and visible in their work. This is partly explained by the fact that these are recent establishments being formed when the 7th parliament (2001-2006) was inaugurated. In parallel to the Uganda PAC the Tanzania PAC has been complimented for its will to act, perhaps increasingly so under the leadership of Hamad Rashid Mohamed (Civic United Front). The LAAC is less high profile than the PAC but both committees commonly call people to stand before them. The committees have gained a modicum of strength in matters of oversight 41 On a more negative note there are several mechanisms constraining the committees’ oversight function. The director of a parliamentary technical support program comments that, ‘Due to the strong history of one-party rule and the strong one-party system…….. MPs are not sure of their own role and are not sure of how far they can push. They are aware but not of the scope and breath of oversight’.42 The Malawi PAC is not as highly profiled as its Tanzania and Uganda counterparts. It meets infrequently and on an irregular basis, limiting its activity. Its lack of vigour is also related to the fact that five years ago there was no functioning committee system in the Malawi national assembly, in practice meaning that the legislature was unable to perform any oversight function on the executive. Despite unfavourable working conditions, the Malawi PAC has examined cases of top-level corruption. However, indicating the level of executive dominance, the committee chair (from the opposition party Malawi Congress Party) was fired, allegedly because of ‘treading on sensitive ground’.43 This particular case clearly points to diffusion rather than separation of the branches of power of government in Malawi. Comparing the support to the committee system Without the necessary technical backing, funding and human resources parliament’s oversight function is diminished. This negatively affects their ability to work with the SAI and their impact on the audit cycle. The technical assistance provided to the Uganda committees has recently been improved as a parliamentary budget office was established in 2001 to provide assistance to the MPs in budget related work and there is also a research department. Still there is a considerable backlog in the three committees dealing with the SAI reports. 44 In July 10 Rakner/Wang Draft December 2006 2004 the PAC examined the 2001/2002 report and the LAAC was still working on the report for 2000/2001.45 The committee on parastatals was looking at the SAI reports on the Uganda Revenue Authority from 1998/1999/2000. 46 This delay considerably constrains the effectiveness of external audit. Some of the ineffectiveness in the handling of reports is still due to lack of research assistance and other necessary facilities such as office space. 47 Committee members may constitute informed members by virtue of their parliamentary experience and/or their professional background. The turnover of MPs was high from the 6th to the 7th Parliament and has limited the building of institutional memory and thus expertise.48 PAC members have received some training while members of the parastatals committee have not. Its workload is considerable and few of the committee members have relevant backgrounds. 49 On the other hand it is a strength that the membership of the standing committees lasts for the entire tenure of parliament. All the same few MPs examine the SAI report. A PAC member avers that ‘When the report is debated most have lost or forgotten it if you know what I am saying’.50 In Tanzania the relevant committees like the PAC and the LAAC seem to do less well than the oversight committees in Uganda, particularly in terms of technical assistance. The Tanzanian committees also have inadequate human and financial resources 51 to the extent that they are unable to carry out proper oversight.52 The auditor general has on several occasions referred to this as a problem. 53 Few of the MPs serving on the committees have the professional backgrounds and skills required to scrutinise the SAI reports and their expertise is further weakened by the arrangement whereby committee membership only lasts for 2 1/2 years.54 The parliamentarians also suffer from a basic lack of researched information. The MPs do not have adequate working tools – they lack clerks, offices and access to computers. The research unit is understaffed and at the Bunge office in Dar es Salaam where most of the MPs do their research there is not even computers with Internet access at the MPs disposal. This clearly impedes the work of the PAC and the LAAC. The meetings of the committees are often held in the two weeks before a parliamentary session starts and take place in Dar es Salaam which situated on the coast. This is where most ministries are found. The sessions, however, are held at the Bunge building in Dodoma, located at the centre of mainland Tanzania and a six-hour road journey from Dar es Salaam. Due to this the MPs spend much of their time commuting between Dar es Salaam, Dodoma and their constituencies. The Malawi PAC suffers from the same shortcomings as the oversight committees in Uganda and Tanzania, but is even more constrained by resource deficits. Government funding of parliament only covers plenary work resulting in a situation where the committee system is 11 Rakner/Wang Draft December 2006 sustained by donor funding. The result is a donor directed committee system where only the committees receiving funding meet. 55 The speaker of parliament lamented that ‘most parliamentary committees are failing to meet. Donors cherry pick which ones to fund and they all pick the same ones.’56 PAC is one of the privileged committees and is reported to meet every month. While this benefits the oversight function of parliament in the audit cycle it at the same time confuses the patterns of accountability. Who is the PAC really accountable to? The tenure of MPs on committees does not necessarily last for the entire term of the legislature as political parties want to maintain tight control over their members and therefore is not willing to grant the parliamentarians this security and confidence. Despite a certain upgrading of parliamentary facilities and support functions since 1999, the 17 committees only have two researchers and four clerks to assist them and no computer or internet facilities. They also lack meeting rooms.57 In practice this means that the research assistance available is provided by donors. Coupled with weak capacity of MPs to understand the SAI reports, the oversight function of the PAC is significantly hampered. The effects of party discipline on the accountability function of parliament The mechanisms of sanction at the committees’ disposal are of prime importance for parliament’s ability to follow up on audit findings and monitor the implementation of committee recommendations. The ‘party’ groups of the ruling ‘parties’ in Uganda, Tanzania and Malawi exercise considerable control over their respective parliamentarians. Uganda’s ‘no-party system’ of politics meant that there were no official party groups in parliament in the 6th (1996-2001) and 7th (2001-2006) Parliaments but there were a number of caucuses such as the Parliamentary Advocacy Forum, the Young Parliamentarians Association, and the Acholi, Teso and Lango parliamentary groups that have voiced criticism against the government.58 Not least there was a very active Movement caucus which at times was hard to distinguish from a parliamentary party caucus59 as it was active and successful in disciplining and gathering support when important and prestigious legislation is about to be voted over in parliament.60 Irrespective of the coalition formation and lobbying going on in parliament, there was in general little suspense attached to the outcome of issues considered important to the executive since the Movementists largely outnumbered the opposition in parliament. One of the members of the LAAC asserts, ‘You are perhaps open minded in the caucus, but when an issue has been decided here you have to abide.’61 An effective way of silencing critical MPs has been to co-opt them into the government. 62 The Movement is known to be selective in its sponsoring of candidates and expects loyalty in return for 12 Rakner/Wang Draft December 2006 electoral support.63 These are all factors that constrained the Uganda oversight committees in their work and induced strong Movement discipline. The following statement made by a Ugandan MP is revealing of the state of affairs, ‘In relation to voting there is a lot of manipulation going on. You vote to show that you belong and that you support. People compete to be recognised in the system so most people are not independent minded’. This obviously impacts negatively on the oversight function of parliament. There is a noticeable difference between the 6th and 7th Parliaments in this regard. The latter made fewer efforts to curb the executive.64 Suggested explanations of are that the MPs were protecting access to future patronage and also a general change in the atmosphere in parliament as the 2006 elections drew near.65 In Tanzania the parliamentary standing orders stipulate that the chairs of the PAC and the LAAC should be from the opposition. Ensuring a sufficient number of opposition members within these committees have proved difficult and in July 2004 only the PAC was chaired by an opposition MP. 66 The CCM party group organisation is considerably more institutionalised and advanced than that of the opposition. The whip system and the CCM party caucus have become forceful organs of making the CCM MPs toe the party line. A senior CCM MP sums up the quandary the MPs face in a nice way: Toeing the party line is very important. You would not go against it because you are also serving the party. This is very detrimental to parliamentary oversight. You can see that this is wrong but what can you do?67 The president has on several occasions made it clear in meetings of the CCM party caucus that MPs opposing particular pieces of legislation in parliament will not be able to stand for re-election and according to Tanzanian electoral law you cannot cross the floor in parliament.68 This provides the parties with an effective disciplinary mechanism. Expulsion from the party in practice means that you have to resign your parliamentary seat. Since there is a lot at stake for an MP in such a situation, the incentive for acting according to party discipline is strong. Illustrative of the situation, a former PAC member simply states ‘In Tanzania you have separation of powers when it seems to serve the government’s powers and only then’.69 The MPs from the ruling party is not expected to query party policy, at least not when it has been introduced in the plenary. The speaker of parliament Pius Msekwa 70 when discussing the party discipline in the Tanzanian parliament is crystal clear in his statements, claiming that ‘It is a moral obligation for them [the majority ruling party members of parliament] to support the government of their party on the floor of the House.’ In this way he underscores the lack of checks and balances in a system where the CCM held nearly 88 per 13 Rakner/Wang Draft December 2006 cent of the seats in parliament after the 2000 general elections, and about 85% after the 2005 elections.71 In contrast to Uganda and Tanzania there are shifting coalitions in Malawi and no dominant party. This has, however, proved to be a strong breeding ground for executive dominance as every vote counts for the government – in fact one vote could be decisive for its ability to rule. This has reinforced a system of patronage and a number of the MPs claim that they are unable to deliberate freely in the House and they perceive the party caucuses as a means for the parties to enforce their views on them.72 This is confirmed by our interviews with Malawian MPs who felt precluded from articulating national interests due to party loyalty. The Public Affairs Committee, a Malawi NGO focused on civic affairs, in relation to this comments that: The UDF and Aford continue to harass and intimidate MPs who are not in favour of third term Constitutional Amendment. …we are convinced that one major strategy the ruling party has embarked on is to perpetuate harassment and intimidation among those people who wisely reject the third term bid….73 It is reasonable to believe that not abiding by the party line could have severe consequences.74 The MPs refer to extreme reactions such as party expulsion or demotion to insignificant positions. 75 The coalitions between the United Democratic Front and the opposition have also strengthened the hand of the executive. Interviews with MPs revealed that both for MPs from the government party and the opposition there were repercussions for speaking up against the government as they risked reductions in developing funds to constituencies with ‘critical MPs’. Follow up of audit findings In Malawi, Tanzania and Uganda lack of follow up of audit findings poses a serious problem and threatens the integrity of the audit cycle. The Uganda oversight committees face many of the same problems identified in Malawi and Tanzania but possibly to a lesser extent. Particularly the Uganda PAC has been able to better exploit the opportunities the committee has to hold the executive accountable. An explanation is that there is considerable openness of committee proceedings as opposed to in Tanzania. In Uganda PAC sessions are open to the press and officials from the SAI, the accountant generals office, and the criminal investigation department. The latter is invited because, in the words of the committee chair ‘we would like to include prima facie cases in the event of embezzlement of funds to ensure prosecution’.76 This has ensured a quick response and that action is taken in cases of outright fraud. While the 14 Rakner/Wang Draft December 2006 PAC has been reasonably effective in this regard the LAAC has a less impressive track record. The committee has held some hearings and chief administrative officers have been taken to court as a result, but the committee had by mid-2004 not presented a single report in the plenary. Allegedly because the speaker had been unable to allocate time to discuss it.77 The PAC does not have the same backlog of accounts as the LAAC and the committee on parastatals, but insufficient time has been allocated to discuss its reports in parliament, seriously weakening its scrutiny work.78 The available disciplinary measures are inadequate. The treasury memorandums are not on time and cannot be when the committee reports are delayed. As a rule they are not made use of by any of the committees. The LAAC had not yet received a treasury memorandum in July 2004 due to the fact that it had not presented a report to parliament. The committee on parastatals does not pay much attention to them (if it receives them at all) – as a prominent member of the committee reveals ‘I do not think we receive the treasury memorandums. To me I do not see the way we follow up. This one is missing’. 79 The PAC at least receives the treasury memorandums but does not have the time to examine them.80 The Tanzania PAC is hampered in its work by lack a lack of effective enforcement mechanisms and disciplinary measures for non-compliance. 81 The committees have the possibility to go deeper into issues in the audit reports by means of establishing select or probe committees to conduct inquiries. However, this entails a rather elaborate procedure. A private member motion has to be moved and the motions are frequently turned down. Part of the reason is lack of funding but the interests of the ruling party are also of relevance, ‘the party has to agree to the private member motion to establish such a committee. You must think about whether it will be detrimental to the government. The ruling party will often not accept…For the opposition there is no point in introducing private member motions – it will not go through anyway’. 82 There are few incentives on the executive to act in accordance with PAC recommendations.83 The committee meetings takes place ‘in camera’ which is in contravention of internationally acceptable standards since it diminishes transparency and possibly also the effect of audit findings. In conformity with the committees in Uganda there is also a substantial problem on the follow up side in Tanzania. Already in 2002 it was pointed out that the Paymaster General lacks authority.84 He is at the same level as other Ministries’ chief accounting officers, and has not been adequately empowered to ensure that they implement the changes recommended by the PAC and the LAAC.85 Changes are now underway to rectify this problem.86 With the finalisation of the penalties and surcharges regulations towards the end of 2005 new 15 Rakner/Wang Draft December 2006 regulations for the sanctioning of civil servants for financial mismanagement will be in place.87 The executive’s reporting back to parliament in the form of treasury memoranda is another weak area.88 Most of the MPs referred to the scrutiny of the SAI report as practically the only way to check on the implementation of committee recommendations. It is also a considerable concern that five PAC reports, including two special reports from 2004 have not been debated by parliament. Accordingly the reports cannot be acted upon.89 The audit cycle is in this way stalled and financial accountability is undermined. The story is by and large repeated in Malawi. Controlling officers are reported to ‘rarely bother to provide replies to the findings of the auditor general’s office’ and PAC recommendations. 90 Overall the auditees’ effort to comply with recommendations and required action is at best unreliable. According to members of the PAC, the main weakness is that the system provides few openings for sanctioning controlling officers. So far instances of disciplinary measures taken on permanent secretaries (controlling officers) are non-existent. Habitually, defaulting officers are moved to other ministries with no other legal actions. Since the committee does not itself effect sanctions, its recommendations can easily be ignored by those empowered to take such action. The MPs incentive to pressure and follow up on government action or inaction is also limited. Committee members interviewed complained that increased committee activity (resulting from donor finances) kept them away from their constituencies and that the oversight function of the office was in direct conflict with the direct (vertical) accountability function vis-à-vis the constituencies.91 Treasury memorandums are rarely received by the committee and even if these had been produced lack of capacity, time and incentives on the part of the MPs to examine them would most likely render them unimportant. At the moment what exist of follow up on enforcement of PAC advice takes place when a new audit report is scrutinised. Although non-compliance to SAI recommendations may be included in the coming SAI report waiting a whole year contributes to considerable delays and less effective follow up in all three countries. The outcome is that enforcement ultimately depends on the will of the executive to take action which is hardly a recipe for success. In sum we have found that particularly the Uganda PAC but also the Tanzania PAC has gained a modicum of strength in acting upon the SAI reports. The Malawi PAC is less active than its Tanzania and Uganda counterparts. The committee performance is in part explained by the support to the committee system. The Malawi committees’ are more constrained by resource deficits than the committees in the other two countries and is marked by exceedingly undue donor influence. The relevant Ugandan committees seem to do the best 16 Rakner/Wang Draft December 2006 in terms of technical assistance and funding while the Tanzanian committees take a middle position. While we find that the parliaments increasingly show commitment to the oversight process serious problems exist on the follow up side. Despite their different systems of governance a common denominator is that executive dominance, party discipline and weak opposition parties to various degrees constrain the accountability function of parliament in all three countries and thus its ability to effectively work together with the SAI and fulfil its role in the audit cycle. There are numerous instances of audit findings that never have been acted upon in the three countries. As one of the directors of audit in Uganda readily admitted ‘I would say that about 30-40 per cent of the process is cosmetic. The net cannot catch the big fish because of the positions they are in’.92 Concluding Remarks Overall, our analysis shows that the nexus between the SAI and parliament play into and hamper the economic accountability performance in all three countries. Our findings suggest an informal ‘ranking’, in which the Uganda SAI and parliament, in tandem, performs slightly better than its Tanzanian counterparts, with Malawi lagging behind. When assessing the interaction between the SAI and the legislature we find that although the Uganda PAC and to some extent the Tanzania PAC have displayed some degree of independence in the scrutiny of audit reports, the real weakness in the audit cycle is to be found in the area of follow up. The same problems are repeated in the audit reports year after year underscoring the limits of enforcement mechanisms and lack of incentives to impose sanctions, partly rendering the audit process a cosmetic exercise. This is reflective of the general tenor of legislature-executive relations in the three countries. The Malawi system of governance displays lack of institutionalisation and is largely donor driven. The national assembly has been seen as a rubber stamp of the executive’s decisions and there is little distinction between the government and the incumbent party UDF. 93 Weak party identities and lack of ideological differences between parties have only increased executive dominance.94 In Tanzania relative institutional stability is challenged by the long standing CCM reign which permeates all branches of the political system from top to bottom. The Tanzanian executive is commonly held to experience little pressure from parliament. 95 The party discipline is so strong that it negatively affects parliament’s ability to hold the executive accountable. 96 Of the three countries Uganda has commonly been held to have the most institutionalised system of checks and balances. Our findings to some extent confirm this 17 Rakner/Wang Draft December 2006 view. In the 6th parliament there was accept for dissent. A vocal parliament flexed its muscles on several occasions – even in relation to a few controversial cases. Since 2001, however, a shift towards a more executive driven system has been witnessed. With the change to a multiparty system there is growing uncertainty as to how the system will develop in the future. The 7th parliament was more subdued than its predecessor. In light of this development and the recent 2006 parliamentary and presidential elections, it seems reasonable to expect the 8th parliament to carry on the trends of the 7th parliament. The Ugandan oversight institutions and functions were established within the Movement system, where presumably ‘everyone agreed on the basic principles’. Now moving to a competitive and more conflictual political situation the government commitment to the oversight institutions is clearly challenged. 1 Representative assemblies are designated by various names. Here the concepts of parliament, legislature and national assembly will be used interchangeably and as synonyms with the legislative branch of government. 2 The analysis utilises a comparative, qualitative case study approach and its findings are based on four different sources: Key informant interviews, government documents, reports and academic literature. Fieldwork for this article was conducted in Tanzania and Uganda in June/July 2004 and in Malawi during the first half of 2004. Key informant interviews were carried out with policy makers, representatives from the private sector, the Government, civil society, and the donor community. In Tanzania 31 interviews were carried out while the number was 34 and 35 in Uganda and Malawi respectively. In Malawi the number also includes focus group discussions. Numerous people helped us during these visits. Deserving of special mention are Donna BugbySmith, Martin Wilcox and Carl Åke Gerdén. 3 ‘NRM-O’ refers to the political party National Resistance Movement Organisation which was formally registered on 31 Sept. 2003. It was the first party to register in front of the Feb. 2006 parliamentary and presidential elections. 4 Public external audits are commonly divided into three main types: the Judicial, the Board and the Westminster. For a more detailed outline of these models, see, DFID, ‘Characteristics of different external audit systems’, DFID Briefing Paper. Policy Division Info Series (2004). 5 Most of this literature is a result of the Support for Improvement in Governance and Management in Central and Eastern European Countries (SIGMA) which is a joint initiative of the EU and OECD. See, for instance, J. Mazur and B. Vella ‘Relations between Supreme Audit Institutions and Parliamentary Committees’, report prepared by SAIs of the Central and Eastern European Countries, Cyprus, Malta and the European Court of Auditors (Limassol: SIGMA, 2001); SIGMA ‘Achieving High Quality in the Work of Supreme Audit Institutions’, Sigma paper no. 34 (OECD and the EU, 2004). 6 With the growth of direct budget support assessments of the effectiveness of the central government systems and systems of accountability outside the executive such as those of the SAI and parliament have become common. See, for instance the DFID practice on this, DFID ‘Managing Fiduciary Risk when providing Poverty Reduction Budget Support’, DFID How to note, Policy Division Info Series (2004). 7 See, J. Mazur and B. Vella ‘Relations between Supreme Audit Institutions and Parliamentary Committees’, report prepared by SAIs of the Central and Eastern European Countries, Cyprus, Malta and the European Court of Auditors (Limassol: SIGMA, 2001); J. Whener ‘Best Practice of Public Accounts Committees’ paper for the Handbook for Public Accounts Committees commissioned by the Association of Public Accounts Committees (APAC) in South Africa (2002), www.internationalbudget.org/auditorgeneral.htm; SIGMA ‘Relations Between Supreme Audit Institutions and Parliamentary Committees’, SIGMA paper no. 33 (OECD and the European Union, 2002);D.G. McGee, The Overseers: Public Accounts Committees and Public Spending. (London: Commonwealth Parliamentary Association and Pluto Press, 2002); W. Krafchik, ‘What Role can Civil Society and Parliament Play in Strengthening the External Auditing Function?’, paper presented at the World Bank Institute workshop, Towards Auditing Effectiveness, Ethiopia (12-15 May 2003), www.internationalbudget.org/auditorgeneral.htm; V. Wang and L. Rakner, ‘The Accountability Function of Supreme Audit Institutions in Malawi, Uganda and Tanzania’, CMI Report No. 4 (Bergen: CMI, 2005). 8 See, for instance, D. Dewar, ‘Independence of State Audit’, International Journal of Government Auditing, 15/3 (1998), 10-12, 20; R. Stapenhurst and J. Titsworth. ‘Features and functions of supreme audit institutions’, PREMnotes 59 (The World Bank 2001); K. M. Dye and R. Stapenhurst, Pillars of Integrity: The Importance of 18 Rakner/Wang Draft December 2006 Supreme Audit Institutions in Curbing Corruption, (The Economic Development Institute of the World Bank, 1998). 9 Dewar, ‘Independence of State Audit’. 10 Dye and Stapenhurst, Pillars of Integrity. 11 F. Fiedler, The Independence of the Supreme Audit Institutions (INTOSAI, 2003). 12 Janusz Wojciechowski, Recommendations concerning the Functioning of Supreme Audit Institutions in the Context of European Integration (The Polish Supreme Chamber of Control, 1999). 13 Note, however, the growing literature on Eastern and Central Europe. See, for instance, A. Ágh (ed.), The Emergence of East Central European Parliaments: The First Steps (Budapest: Hungarian Centre of Democracy Studies Foundation, 1994); T. F. Remington, Parliaments in Transition: The New Legislative Politics in the Former USSR and Eastern Europe (Canada: Harper Collins, 1995); D. M. Olson and P. Norton (eds.), The New Parliaments of Central and Eastern Europe (London: Frank Cass 1996); A. Ágh and G .Ilonszki, Parliaments and Organised Interests: The Second Steps (Budapest: Hungarian Centre for Democracy Studies, 1996); J. M. Ostrow, Comparing Post-Soviet Legislatures. A Theory of Institutional Design and Political Conflict (Columbus: Ohio State University Press, 2000); D. M. Olson and W. E. Crowther (eds.), Committees in Post-Communist Democratic Parliaments. Comparative Institutionalization (Columbus: Ohio State University Press, 2003). 14 This division draws on an article by L. Nijzink, S. Mozaffar and E. Azevedo, ‘Parliaments and the Enhancement of Democracy on the African Continent: An Analysis of Institutional Capacity and Public perceptions’, The Journal of Legislative Studies, 12/3-4 (2006), pp. 311-335. 15 This literature includes the work of e.g. R. A. Packenham ‘Legislatures and political development’, in A. Kornberg and L. Musolf (eds.) Legislatures in Developmental Perspective (Durham: Duke University Press 1970) pp. 521-82; M. Mezey, ‘The Functions of Legislatures in the Third World’ in G. Loewenberg, S. C. Patterson and M. E. Jewell (eds.) Handbook of Legislative Research (England: Harvard University Press) pp. 733-772. 16 See, however, R. F. Hopkins, ‘The Role of the M.P. in Tanzania’, American Political Science Review LXIV (1970) pp. 754-771; R. F. Hopkins, ‘Political Roles in a New State’, (New Haven and London: Yale University Press, 1971), J. Barkan, ‘Bringing Home the Pork: Legislator Behavior. Rural Development, and Political Change in East Africa’, in J. Smith and L. D. Musolf (eds.), Legislatures in Development: Dynamics of Change in New and Old States (Durham: Duke University Press, 1979) pp. 125-54; H. Kjekshus, ‘The Question Hour in Tanzania’s Bunge’, The African Review 2/3 (1972) pp. 351-79; H. Kjekshus, ‘Perspectives on the Second Parliament, 1965-70’, in The Election Study Committee, Socialism and Participation (Dar es Salaam: Tanzania Publishing House, 1974) pp.60-95; H. Kjekshus, ‘Parliament in a One-Party State – The Bunge of Tanzania, 1965-70’, The Journal of Modern African Studies 12/1 (1974), pp. 19-43. 17 See, R. A. Packenham ‘Legislatures and political development’, p. 82. In short the essence of this thesis is that a vigorous executive not held back by a relatively strong parliament will act on behalf of the wider community and make decisions favourable to economic development while an influential parliament is seen as a potential obstacle for executive decision-making and, hence, development. 18 G. O’Donnell ‘Delegative Democracy’, Journal of Democracy, 5/1 (1994), pp. 55-69; R. Sklar ‘Developmental Democracy’, Comparative Studies in Society and History, 29/4 (1987), pp. 686-714; G. O’Donnell ‘Horizontal Accountability in New Democracies’, in A. Schedler, L. Diamond and M. F. Plattner (eds.), The Self-Restraining State. Power and Accountability in New Democracies (London: Lynne Rienner), pp. 29-51; R. Sklar ‘Democracy and Constitutionalism’, in A. Schedler, L. Diamond and M. F. Plattner (eds.), The Self-Restraining State. Power and Accountability in New Democracies (London: Lynne Rienner), pp. 53-58; P. C. Schmitter ‘The Limits of Horizontal Accountability’, in A. Schedler, L. Diamond and M. F. Plattner (eds.), The Self-Restraining State. Power and Accountability in New Democracies (London: Lynne Rienner), pp. 59-62; A. Schedler ‘Conceptualising Accountability’, in A. Schedler, L. Diamond and M. F. Plattner (eds.), The SelfRestraining State. Power and Accountability in New Democracies (London: Lynne Rienner) pp.13-28; M. F. Plattner ‘Traditions of Accountability’, in A. Schedler, L. Diamond and M. F. Plattner (eds.), The SelfRestraining State. Power and Accountability in New Democracies (London: Lynne Rienner) pp. 63-67. 19 A selection of this literature is: M. A. M. Salih (ed.) African Parliaments. Between Governance and Government (New York: Palgrave Macmillan, 2005); P. Burnell ‘Financial Indiscipline in Zambia’s Third Republic: The Role of Parliamentary Scrutiny’, The Journal of Legislative Studies, 7/3 (2001), pp. 34-64; P. Burnell ‘Parliamentary Committees in Zambia’s Third Republic: Partial Reforms; Unfinished Agenda’ Journal of Southern African Studies, 28/2 (2002), pp. 291-313; P. Burnell ‘Legislative-Executive Relations in Zambia: Parliamentary Reform on the Agenda’, Journal of Contemporary African Studies, 21/1 (2003), pp. 47-68; N. Kasfir and S. H. Twebaze ‘The Limits of Institutionalization of a Legislature without Parties: The Ugandan Parliament’, Paper prepared for delivery at the Annual Meeting of the APSA, Washington D.C., 2005; R. Nakamura and J. Johnson, ‘Rising Legislative Assertiveness in Uganda and Kenya 1996-2002’, Paper prepared for delivery at the 19th International Political Science Association World Congress, Durban South Africa, 2003; 19 Rakner/Wang Draft December 2006 N. Patel and A. Tostensen, ‘Parliamentary-Executive Relations in Malawi 1994-2004’, CMI Working Paper WP 10 (Bergen: CMI, 2006); I. Amundsen, C. Abreu and L. Hoygaard, ‘Accountability on the Move: The Parliament of Angola’, CMI Working Paper WP 11 (Bergen: CMI, 2005); M. A. Thomas and O. Sissokho, ‘Liaison Legislature: The Role of the National Assembly in Senegal’, Journal of Modern African Studies, 43/1 (2005), pp. 97-117; J. D. Barkan , L. Ademolekun and Y. Zhou, ‘Emerging Legislatures: Institutions of Horizontal Accountability’, in B. Levy and S. Kpundhe (eds.), Building State Capacity in Africa (Washington: World Bank Institute, 2004), pp. 211-56. For a more detailed review of the literature available on African legislatures, particularly in a comparative perspective, see L. Nijzink, S. Mozaffar and E. Azevedo, ‘Parliaments and the Enhancement of Democracy on the African Continent’. 20 In line with Copeland and Patterson we acknowledge that a functional definition of parliament is problematical since legislative institutions are very diverse and there is no function or set of functions which are exclusively parliamentary or universally undertaken by parliaments. Nevertheless, legislatures do engage in recurring activities that have systematic consequences. While authors like Packenham, Copeland and Patterson, and Pridham differentiate between different categories of functions the functions outlined generally correspond to each other, and all point to elements that can be subsumed under the heading decisional/influence. As several authors have suggested (e.g. D. Judge, R. A. Packenham and D. M. Olson) we find that attention should be focused more broadly than simply upon the contribution to law-making. Especially this is more productive in a developing country context. Impacting on the policy/decision making process can be seen as one of parliament’s main tasks and is arguably the most explicitly decisive and relevant for the legislature’s ability to impose accountability. See, G. W. Copeland and S. C. Patterson, ‘Parliaments and Legislatures’ in G. T. Kurian (ed.), World Encyclopedia of Parliaments and Legislatures vol. I (Washington D.C.: Congressional Quarterly Inc., 1998) pp. 19-32 ; Packenham ‘Legislatures and political development’ p.95; G. Pridham, ‘Political Parties, Parliaments and Democratic Consolidation in Southern Europe: Empirical and Theoretical Perspectives’, in U. Liebert and M. Cotta (eds.), Parliament and Democratic Consolidation in Southern Europe: Greece, Italy, Portugal, Spain and Turkey (London and New York: Pinter Publishers, 1990), pp. 225-48; D. Judge, ‘East Central European Parliaments: The First Steps’, in Ágh, Attila (ed.), The Emergence of East Central European Parliaments: The First Steps (Budapest: Hungarian Centre of Democracy Studies Foundation, 1994), pp.22-32, p.29; D. M. Olson, The Legislative Process: A Comparative Approach (Cambridge: Cambridge University Press, 1980) p.12. 21 M. Mezey, Comparative Legislatures (Durham: Duke University Press 1979) p. 23. Scholars like Blondel, Polsby, Weinbaum, and Mezey all fall within this category. They incorporate Third World legislatures into their theoretical considerations and apply rather broad perspectives on legislative influence and decision-making authority. In more recent years Philip Norton, together with Nizam Ahmed and David M. Olson have drawn on Mezey’s schema for classification and on the basis of this hypothesised about the relative importance of different variables on the legislature’s capacity to affect the policy process. However, these studies do not specifically target Africa. See, J. Blondel et al., ‘Legislative Behaviour: Some Steps toward a Cross-National Measurement’, Government and Opposition 5/196 (1969-70), pp. 67-85; J. Blondel, Comparative Legislatures (New Jersey: Prentice Hall, 1973); N. W. Polsby ‘Legislatures’, in F. I. Greenstein and N. Polsby (eds.), Handbook of Political Science: Governmental Institutions and Processes (Reading, MA: Addison-Wesley, 1975), pp. 257-319; M. Weinbaum, ‘Classification and Change in Legislative Systems: With Particular Application to Iran, Turkey and Afghanistan’, in G. R. Boynton and C. L. Kim (eds.), Legislatures in Developing Countries (Durham, North Carolina: Duke University Press, 1975), pp. 31-68; P. Norton and N. Ahmed (eds.) Legislatures in Asia (London: Frank Cass, 1999); P. Norton (ed.) Parliaments and Governments in Western Europe vol. 1 (London: Frank Cass, 1998); P. Norton and D. M. Olson, The New Parliaments of Central and Eastern Europe (London: Frank Cass, 1996). 22 L. D. Longley and R. H. Davidson, ‘The New Roles of Parliamentary Committees’, The Journal of Legislative Studies 4/1, (1998) pp.1-19, p.2; M. Shaw, ‘Parliamentary Committees: A Global Perspective’, Journal of Legislative Studies, 4/1 (1998) pp. 225-51, p 230; I. Mattson and K. Strøm, ‘Parliamentary Committees’ in H. Döring (ed.), Parliaments and Majority Rule in Western Europe (New York: St. Martin’s Press, 1995) pp. 249307, p.249; J. Blondel, Comparative Legislatures, p. 249. 23 K. Strøm, ‘Parliamentary Committees in European Democracies’, Journal of Legislative Studies 4/1 (1998), pp. 21-59, p.47, Mattson and Strøm, ‘Parliamentary Committees’, p. 250. 24 ‘Viscosity’ is characteristic of the degree to which legislatures are ‘compliant’ or ‘free’. It refers to the capacity of the legislature to resist legislation initiated by the executive. The concept was coined by Blondel and his associates. See J. Blondel et al. ‘Legislative Behaviour’ 25 M. Shaw, ‘Committees in Legislatures’, in P. Norton (ed.) Legislatures (Oxford: Oxford University Press, 1979), pp. 237-267; M. Shaw, ‘Parliamentary Committees: A Global Perspective’ 20 Rakner/Wang Draft December 2006 M. Shaw, ‘Conclusions’ in J. D. Lees and M. Shaw (eds.) Committees in Legislatures: A Comparative Analysis (Oxford: Martin Robertson 1979) pp.; M. Shaw, ‘Committees in Legislatures’; D. M. Olson, The Legislative Process: A Comparative Approach. 27 M. Mezey, ‘Executive-Legislative Relations’ in G. T. Kurian (ed.) World Encyclopedia of Parliaments and Legislatures Vol. 2 (Washington D.C.: Congressional Quarterly Inc., 1998) pp.780-86, p.782. 28 For more on this analytical framework see S. Gloppen and L. Rakner, ‘Checking the state: The Role of Special Institutions of Restraint in Africa’s New Democracies’, CMI Working Paper WP (CMI: Bergen, Forthcoming). 29 With the adoption of the 2003 Public Finance and Accountability Act the Uganda SAI is allowed to audit classified expenditure under specified conditions. Still, the question of who should decide whether a subject is confidential or not may be raised. If required, issues relating to national security or other interests will be exempted after consultation with the secretary to the treasury. What amounts to ‘required’ is unclear and although exceptions should be consistent with best practice this is problematic as auditees can be expected to protect their own interests and try to restrain the information. 30 Cowater International ‘Institutional Reform at the NAO. Malawi’, Final report. Financial Restructuring and Deregulation Programme’. (Canada: Cowater International Inc., 2003) p.40; MEJN ‘Response to the Proposed National Budget 2004/5’. (Lilongwe: Malawi Economic Justice Network, 2004) p.10; OAG ‘Corporate Plan 2003-2006’ Office of the Auditor General, Republic of Uganda (2003) p.5; L. Rutashobya, ‘Fiduciary Risk Assessment for the Provision of Direct Budgetary Support in Tanzania’. Summary Focus of Areas of High Risk’, Report submitted to DFID Tanzania. (Tanzania: DFID, 2004) p. 21. 31 Interview Auditor General, Malawi 17/03/04; SNAO and MNAO, ’Institutional Development Co-operation between the National Audit Office of Malawi and the Swedish National Audit Office. Project Document’. Phase II (2003) p 9; S. Tax ‘A Review of Budgetary Processes and Economic Governance in Tanzania’, Report submitted to SEAPREN (2004) p. 38; L. Rutashobya, ‘Fiduciary Risk Assessment’ p. 21; Interview Auditor General, Tanzania 06/07/04; Interview Auditor General, Uganda 14/07/04. 32 Rakner and Wang, ‘The Accountability Function of Supreme Audit Institutions’ 33 See Article 144 Tanzania Constitution, Article 163 (10) Uganda Constitution, and Article 184 (6) Malawi Constitution. 34 NOTE this has possibly changed with the enactment of the 2005 Constitutional Amendment Act (section 33). The AG can now serve until the age of 60. The former AG John Muwanga resigned in May 2006. 35 Interview Auditor General, Malawi 17/03/04; Interview Auditor General Tanzania 06/07/04; Interview Auditor General; Interview Auditor General 14/07/04. 36 UNAO, ‘The Report of the Auditor General to Parliament on the Public Accounts of the Republic of Uganda for the Year Ended 30th June’, (The Republic of Uganda, 2003). 37 PRBS/PRBC Review, ‘Tanzania. Joint PRBS and PRBC Review’. Review Report, 2004, p. 52. 38 SNAO and MNAO, ’Institutional Development Co-operation between the National Audit Office of Malawi and the Swedish National Audit Office,’ p. 5; L. Rakner et al. ‘The Political Economy of the Budget in Malawi. The Budget as a Theatre – the Formal and Informal Institutional Makings of the Budget Process in Malawi’, Report submitted to DFID Malawi (2004) p.18; N. Khembo ‘National Integrity Systems’ TI Country Study Report Malawi. (Transparency International, 2003) pp. 41-42. 39 H. Muloopa, ‘National Integrity Systems. TI Country Study Report Uganda’ (Transparency International, 2003) p. 41. 40 In Uganda the official name is the committee on local government accounts, but the committee will here be referred to as the local authorities’ accounts committee. 41 L. Rutashobya, ‘Fiduciary Risk Assessment’ 42 Interview donor representative, Tanzania 06/07/04. 43 Khembo, ‘National Integrity Systems’, p. 23. 44 WB, ‘Republic of Uganda’, p. 49. 45 Interview MP, Uganda 26/07/04. 46 Interview MP, Uganda 20/07/04. 47 Interview MP, Uganda 26/07/04; Interview MP, Uganda 16/07/04. . 48 Laure-Hélén Piron and Andy Norton. ‘Politics and the PRSP Approach: Uganda Case Study’. Working Paper 240. (London: Overseas Development Institute, 2004) p. 27; Joy Moncrieffe. ‘Uganda's Political Economy: A Synthesis of Major Thought. Report prepared for DFID Uganda’. (London: Overseas Development Institute, 2004) p. 21. 49 WB, ‘Republic of Uganda’, p. 49. 50 Interview MP, Uganda 19/07/04. 51 PER, ‘Public Expenditure and Financial Accountability Review 2005. Review of Fiscal Developments, Public Expenditure Management and Financial Accountability Issues 2003/04 – 2004/05’, Joint Evaluation Report (Dar es Salaam, 2005), p. 26. 26 21 Rakner/Wang Draft December 2006 L. Rutashobya, ‘Fiduciary Risk Assessment’, p. 28. J. Titsworth, L. Rutashobya, and R. Mushi. ‘Public Expenditure Review External Evaluation. The Budget Process and the Role of the Legislature’. (Dar es Salaam, 2004), p. 15. 54 The rotation of committee membership is about to change. The Speaker has recently accepted that committee members should serve five years on a committee. See, DAIMA, and ODI, ‘Joint Evaluation of General Budget Support: Tanzania 1995-2004’, Report to the Government of Tanzania and to the Poverty Reduction Budget Support Development Partners (Dar es Salaam, 2004). 55 Patel and Tostensen ‘Parliamentary-Executive Relations in Malawi 1994-2004’ 56 Interview Speaker, Malawi 15/03/04. 57 Patel and Tostensen ‘Parliamentary-Executive Relations in Malawi 1994-2004’ 58 Moncrieffe, ‘Uganda's Political Economy’, p. 22. 59 Piron and Norton, ‘Politics and the PRSP Approach’, p. 26. 60 Interview MP, Uganda 13/07/04; Interview MP, Uganda 14/07/04; Interview MP, Uganda 16/07/04. 61 Interview MP, Uganda 16/07/04. 62 Piron and Norton, ‘Politics and the PRSP Approach’, pp. 26-27. 63 Interview MP, Uganda 16/07/04. See also N. Kasfir and S. H. Twebaze ‘The Limits of Institutionalization of a Legislature without Parties’. 64 Kasfir and Twebaze ‘The Limits of Institutionalization of a Legislature without Parties’. 65 Kasfir and Twebaze ‘The Limits of Institutionalization of a Legislature without Parties’. 66 DAIMA and ODI, ‘Joint Evaluation of General Budget Support’ p. 94. 67 Interview MP, Tanzania 09/07/04. 68 Vibeke Wang, ‘Parliament as Machinery for Political System Control: The Inner Workings of Bunge, Tanzania’, in M. A. M. Salih (ed.), African Parliaments: Between Governance and Government (London: Palgrave Macmillan, 2005) pp. 183-200. 69 Interview MP, Tanzania 08/07/04. 70 Pius Msekwa. Reflections on Tanzania's First Multi-Party Parliament: 1995-2000 (Dar es Salaam: DUP(1996)LTD, 2000), p.76. 71 Numbers based on data from www.africanelectionsdatabase.tripod.com/tz 72 Patel and Tostensen, ‘Parliamentary-Executive Relations in Malawi 1994-2004’. 73 PAC statement, October 22, 2002, cited in Dulani and van Donge ‘A Decade of Legislature-Executive Squabble in Malawi’, p.24. 74 See also B. Dulani and J. K. van Donge, ‘A Decade of Legislature-Executive Squabble in Malawi, 19942004’, in M. A. M. Salih (ed.), African Parliaments. Between Government and Governance (New York: Palgrave Macmillan, 2004) pp. 201-24, p. 24 75 Patel and Tostensen ‘Parliamentary-Executive Relations in Malawi 1994-2004’p.16-17; Khembo ‘National Integrity Systems’ p. 35. 76 Interview MP, Uganda 19/07/04. 77 Interview MP, Uganda 19/07/04. 78 WB, ‘Republic of Uganda’, p. 48. In 2004 neither the 1998/99, 1999/00 reports, nor the 2000/01 report on central government accounts, had yet been presented to parliament. 79 Interview MP, Uganda, 20/07/04. 80 Interview MP, Uganda 19/07/04. 81 F. Naschold and A. Fozzard, ‘How, When and Why Does Poverty Get Budget Priority? Poverty Reduction Strategy and Public Spending in Tanzania’, Working Paper 165 (London: Overseas Development Institute, 2002) p. 46; Rutashobya, ‘Fiduciary Risk Assessment for the Provision of Direct Budgetary Support in Tanzania’, p.30. 82 Interview MP, Tanzania 09/07/04. 83 Interview MP, Tanzania 09/07/04; Interview MP, Tanzania 09/07/04; Interview MP, Tanzania 09/07/04; Interview 08/07/04. 84 Naschold and Fozzard. ‘How, When and Why’, p. 37. 85 Handouts from seminar on ‘The Swedish NAO Project’ by Martin Wilcox and Carl Åke Gerdén 05/07/2004 at the Royal Swedish Embassy, Dar es Salaam. 86 DAIMA and ODI, ‘Joint Evaluation of General Budget Support’, p. 94. 87 PRBS/PRSC MYR, ‘Tanzania. PRBS/PRSC Mid Year Review’, (2005), p. 31. 88 Rutashobya, ‘Fiduciary Risk Assessment’. 89 PER, ‘Public Expenditure’, pp.24-25. 90 Khembo, ‘National Integrity Systems’, p. 47. 91 Focus group interview with members of the public accounts committee and the budget and finance committee of the Malawi parliament 18/03/04. 52 53 22 Rakner/Wang Draft December 2006 92 Interview SAI official, Uganda 14/07/04. Patel and Tostensen, ‘Parliamentary-Executive Relations in Malawi 1994-2004’; Rakner et al. ‘The Political Economy of the Budget in Malawi’. 94 L. Rakner, L. Svåsand and N. Khembo, ‘Fissions and Fusions, Foes and Friends: Party System Restructuring in Malawi in the 2004 General Elections’, CMI Working Paper WP 6 (Bergen: CMI, 2005). 95 Wang, ‘Parliament as Machinery for Political System Control’; A. Lawson and L. Rakner, ‘Understanding Patterns of Accountability in Tanzania’, Synthesis Report. (Bergen/Oxford/Dar es Salaam: CMI, OPM REPOA, 2005); B. Frantz, ‘General Budget Support in Tanzania. A Snapshot of its Effectiveness’ (Dar es Salaam: USAID: 2004) p. 14. 96 Wang, ‘Parliament as Machinery for Political System Control’. 93 23