Signalling political autonomy

The most doctrinal answers are rooted intellectually in the New Institutional Economics and the New Public Management, and expressed in terms of concepts like ‘credible commitment’ and ‘signalling’ (Persson & Tabellini Reference Persson and Tabellini1994). The core problem is believed to lie in the capacity both of states, as legitimate public authorities, and of tax collectors, as corrupt abusers of positions of public authority, to extract money from taxpayers without adequate safeguards. There is a wealth of literature suggesting that, when faced with a corrupt tax collector, individual taxpayers will be better off if they pay a bribe, rather than refuse or try to join other taxpayers in protective joint political action (Chand & Moene Reference Chand and Moene1999; Fjeldstad & Tungodden Reference Fjeldstad and Tungodden2003; McLinden Reference McLinden, de Wulf and Sokol2005; Mookherjee Reference Mookherjee1997; Svensson Reference Svensson2003; Zuleta et al. Reference Zuleta, Leyton, Ivanovic, Campos and Pradhan2007). Taxpayers are largely defenceless in the face of retaliatory penalising harassment from the taxman.

A similar logic is applied to the capacity of governments (a) to use their taxation powers in a discretionary way to harass political opponents and their financial supporters and to reward their own allies, and (b) to renege on any agreements they might have made to tax citizens after they have seduced them into revealing their actual income and wealth. In short, the point of departure is the belief that neither governments nor their tax-collecting agents can be trusted with powers over the taxation process. To the extent that governments hand over these powers, in a binding and non-reversible way, to some independent authority that in turn can be trusted not to abuse them, and to abide by correct procedure and the law, taxpayers will have less to fear from the tax agency and its staff, and be more willing to declare their real income and wealth. More tax revenue will be forthcoming, less will leak into the pockets of the collectors, and the government itself will become more legitimate.

The prescription corresponding to this diagnosis is to increase – perhaps even to maximise – the degree of autonomy that the revenue authority has in relation to politicians and government. Governments are urged to do this in their own self-interest. It signals to business people and to potential investors that the power to tax will not be abused. Government has, according to a widely used term, ‘tied its own hands’ (Root Reference Root1989). To use more technical jargon, it has made ‘credible commitments’ to taxpayers about the integrity of future tax arrangements (Taliercio Reference Taliercio2004a, Reference Taliercio2004b). There is no single formula for doing this; the word ‘autonomy’ is abstract and does not translate directly into specific legal, procedural and organisational arrangements (Therkildsen Reference Therkildsen2004). There is, however, a relatively coherent package of formal measures that is likely to contribute to achieving this goal, provided that informal power relations do not completely override formal arrangements:

• • Give the revenue agency a separate legal status, as a corporate body with clear legal responsibilities and duties, and wide powers to own assets, borrow money etc.

• • Put it under the control of a management board whose members are independent of government by virtue of (a) being nominated from a diversity of sources, both inside and outside government;Footnote ⁹ (b) having relatively long, fixed periods of tenure, revocable only on clear criteria and through open and legal processes; and (c) having remuneration arrangements that cannot be affected by the current government.

• • Place all staff clearly and directly under the authority of the chief executive, who will in turn be chosen by and answerable only to the management board.

• • Provide an operational budget that is independent of the normal annual national budgeting process, either through constitutional provisions or by allowing the authority to fund itself by appropriating a fixed share of the revenues it collects.

When they established ARAs, were the African governments keen to signal substantive ‘political autonomy’? Did they institute governance arrangements for ARAs likely to persuade (business) taxpayers that politicians would in future play a much reduced role in the implementation of tax policy? The answer is a clear ‘no’. One of the most direct indicators relates to sources of finance. Outside sub-Saharan Africa, it is the norm that ARAs have access to sources of funding other than, or in addition to, the normal annual budget appropriations initiated by the executive and approved by the legislature. These alternative funding sources mainly take the form of entitlements to a proportion of revenue raised, and may also include some kind of performance bonus. Most ARAs in Africa do not have access, even in principle, to any funding beyond the annual budget appropriations. They are thus under the direct financial control of government.Footnote ¹⁰

Another useful indicator of reformers' intentions is the identity of the agencies responsible for appointing the chairs and members of ARA supervisory/management boards. In sub-Saharan Africa, it is either the President or the Minister of Finance in almost every case. While almost every ARA board includes a private sector representative, in every case s/he is chosen by government rather than by any independent organisation (Kidd & Crandall Reference Kidd and Crandall2006: 86–8). Similarly, the chief executives of ARAs are appointed either by the government alone (e.g. Ethiopia, South Africa, Tanzania and Zambia) or jointly by the government and the supervisory/management board (e.g. Botswana, Kenya, Lesotho, Rwanda, Uganda and Zimbabwe). Only in Mauritius does the supervisory/management board alone have sole formal power of appointment.Footnote ¹¹

It is true that a significant number of the people appointed to run ARAs in Africa have been expatriates or nationals believed not to be deeply embedded in local politics or networks of corruption.Footnote ¹² But such appointments are made, and terminated, very much at the will of presidents. Close relationships between the chief executive of the ARA and the head of state are the norm. For example, the chief executive of the South African Revenue Services (SARS) has, since soon after its establishment in 1997, been a very senior member of the ruling African National Congress (Hlope & Friedman Reference Hlope and Friedman.2002). The Rwanda Revenue Authority (RRA) has been able to count on the personal support of the president, who has played a major role in the campaign to change public attitudes towards paying taxes and corruption (FIAS 2006a). In Uganda, President Museveni and people close to him personally are widely believed to intervene directly in the Uganda Revenue Authority (Therkildsen Reference Therkildsen2004). In Zambia, during Chiluba's presidency (1991–2001), the management of the Zambia Revenue Authority found it difficult to maintain operational autonomy and to prevent political interference (von Soest Reference von Soest2006). For instance, the government instructed the agency not to tax certain businesses, including enterprises owned by ruling party politicians that allegedly had never been subject to assessments or paid tax. On the other hand, opposition politicians and former government members were subject to frequent tax audits and harassment from the tax authorities (Afronet Reference Afronet2002: 27).

These close relationships between heads of ARAs and heads of state have been used both to protect the tax collection process from the corrosive routine pressures of corruption and politicking, and to advance the immediate political interests of the head of state. One agenda can transmute quickly into the other.Footnote ¹³ Business people know that. If the people who established ARAs in sub-Saharan Africa seriously intended to send a credible signal to large taxpayers that tax collection was to be given immunity from routine politics, they would have aimed at governance arrangements for ARAs such as those listed above: as much operational autonomy as is consistent with ultimate accountability to the legislature; wide legal powers and financial authority; freedom to develop self-financing mechanisms; and significant independent representation of business interests on the supervisory board. These are also the governance arrangements one would expect if the creation of ARAs had been intended as a step on the road to the privatisation of tax collection under contract.

Creating managerial autonomy

The notion of ‘managerial autonomy’ is more precise than that of ‘political autonomy’.Footnote ¹⁴ In this context, it refers to the extent to which the managers are able to dispense with standard public service rules – about staff recruitment, deployment, promotion and remuneration, procurement, and operating procedures – in favour of some mixture of their own organisation-specific rules and managerial discretion. This is the normal explanation among taxation professionals worldwide for the establishment of ARAs (Hadler Reference Hadler2000; Kidd & Crandall Reference Kidd and Crandall2006; Mann Reference Mann2004; Terpker Reference Terpker1999). It is argued that the problem with the conventional arrangement – direct collection by departments within ministries of finance – is that standard public service rules and procedures make it impossible to run a revenue operation as it should be run.

Much of the argument can be lifted straight from orthodox New Public Management texts (Manning Reference Manning2001). First, managers need to be able to deploy resources flexibly to meet the particular needs of the tasks they are trying to achieve and the sectors in which they operate. Second, managers need to be able actually to manage their subordinate staff: to find ways of motivating them, rewarding them according to performance, and disciplining them if necessary. Standard public service rules and procedures – and sometimes the public service trade unions believed to lie behind them – are understood to be the main obstacles. These arguments, and especially those relating to personnel management, do have particular plausibility in relation to revenue-raising. On the one hand, tax collectors regularly interact with highly paid private sector professionals: accountants, consultants, tax lawyers, and tax advisers.Footnote ¹⁵ It makes sense that the collectors should be paid enough, and otherwise motivated, to ensure a high quality cadre who will not be outwitted by people on the other side of the fence. Perhaps more important, the collectors should not all leave to serve on the other side of the fence, or otherwise take their scarce finance-related skills into the private sector. Further, since tax collectors are especially vulnerable to temptations to corruption, there is a strong case for a distinct organisational form to make it possible to effectively monitor staff behaviour and exercise discipline.

We should be wary of how we interpret the prevalence of ‘managerial autonomy’ explanations and justifications for establishing ARAs. At least on the part of tax collection staff themselves, there is a clear self-interest involved. The establishment of ARAs has generally been followed by substantial formal salary increases.Footnote ¹⁶ However, had strong beliefs in New Public Management principles been the dominant motivation, two complementary measures would have been introduced. First, many if not all of the existing staff of tax collection units in ministries of finance would have been obliged to apply for jobs in the ARA, and been subjected to a rigorous selection procedure. This was the procedure followed at the establishment in 1991 of SUNAT in Peru, one of the earliest and most iconic ARAs (Fjeldstad & Moore Reference Fjeldstad, Moore, Bräutigam, Fjeldstad and Moore2008: 252). There has been little of this in Africa. Existing tax collection staff have generally been transferred directly into ARAs.Footnote ¹⁷ Second, disciplinary procedures would have been tightened up within ARAs, and more staff would have been penalised for the corruption that is widespread in many of them. However, the use of dismissals in the initial phases of some ARAs has not been sustained. In Tanzania, for instance, annual dismissals have in later years dropped to less than 2% of the staff total. Generally, the annual turnover of staff is low. In the Lesotho Revenue Authority, for example, turnover is around 1–2% per year (FIAS 2006b).Footnote ¹⁸

In sum, arguments about the need for managerial autonomy did play big roles in motivating and justifying the creation of ARAs. All have received some managerial autonomy. However, since the motivating arguments have been developed and deployed within a global epistemic community of tax professionals, spanning international financial institutions, consultants and tax administrators (Stewart Reference Stewart2002), it is not surprising that those aspects of the New Public Management case for ARAs that would tend to inflict pain on the tax collectors themselves – for instance, reselection on merit and strengthened anti-corruption procedures – have been rather neglected.

Facilitating reform

The most pragmatic argument for establishing an ARA is that, in two distinct ways, this facilitates other reforms in tax administration. First, it sends a signal to external supporters of reform, especially to aid donors and international financial institutions, that the government is (a) serious about reform, and (b) is willing, by adopting an iconic idea, broadly to accept the reform agendas that they offer. Second, this decision impacts on internal constituencies, especially sources of potential resistance to reform within ministries of finance. It signals a clear commitment to change and willingness to recruit external support to ensure that this happens. The establishment of ARAs both reflects, and has helped create, a broad consensus about tax reform in the global epistemic community mentioned in the previous paragraph.

Over the last two or three decades, a strong international consensus about tax reform has emerged. The key elements are the introduction of broad-based value-added taxes on consumption, simplified tax design and improved tax administration (Fjeldstad & Moore Reference Fjeldstad, Moore, Bräutigam, Fjeldstad and Moore2008). None of this requires the creation of autonomous tax agencies. But radical organisational reform facilitates many of the other more specific organisational changes that have become the norm: (a) introducing unique identification numbers for each individual taxpaying unit; (b) moving from a system organised around different taxes to one organised around localities and/or industries, such that individual taxpayers have to deal with fewer tax officers; (c) establishing separate offices and procedures for different categories of taxpayers, typically starting with the creation of Large Taxpayer Units focusing on big companies; (d) beginning to physically separate the ‘back office’ functions of assessing tax liabilities, and auditing and cross-checking records, from the ‘front office’ functions of actual tax collection, to reduce the scope for direct extortion and bribery; (e) trying to make the process more ‘user-friendly’;Footnote ¹⁹ and (f) generally exploiting the potential of new information and communication technologies.

It is hard to assess the strength of such instrumental motivation. The key players are unlikely to be fully transparent, or even necessarily to be fully and critically aware of the ways in which their own motivations and constructions of reality change over time. The only quantitative information we have on the issue comes from the 2006 IMF survey of ARAs (Kidd & Crandall Reference Kidd and Crandall2006). Respondents were given eight possible reasons why their ARA was originally established. The explanation that received the highest ranking was a rather anaemic claim about a general need for reform (‘low effectiveness of tax administration and low levels of compliance’). The second-ranked explanation was an assertion of instrumentality: ‘need for a catalyst to launch broader revenue administration reform’ (ibid.: 90). This accords with our own impressions from experience. ARAs have been established in sub-Saharan Africa in large part in order to advance other, more specific, reforms. It is partly for that reason that, relative to ARAs in other parts of the world, those in Africa are very similar to one another in basic form.Footnote ²⁰ Because their adoption had major signalling and symbolic dimensions, there was little motivation to extensively re-engineer the designs adopted in neighbouring countries with which there was a great deal of interaction. There is substance behind the listing on the home web page of the Rwanda Revenue Authority (www.rra.gov.rw) of links to ‘sister revenue authorities’ in Uganda, Kenya and Tanzania.

Justifying ARAs

The material presented above suggests four main conclusions about the motivations behind the creation of ARAs:

• • The more ideological justifications for creating new organisational forms, such as signalling political autonomy (or credible commitment), played little independent role.

• • The dominant motivation was that visible, substantial changes in organisational forms could be used as a vehicle for introducing a wide range of more specific changes in tax administration that had proved their worth elsewhere in the world.

• • Aid agencies and international development organisations played a major part.

• • The main ultimate objective of both African governments and the international actors was to increase tax revenues.

We conclude this section with further evidence for these general conclusions. First, ARAs have generally been established in regions of the poor world suffering from relatively severe fiscal stress, where governments were especially eager to improve their tax collection mechanisms. They are found mainly in sub-Saharan Africa and Latin America, the two regions of the world that suffered from low or declining incomes in the 1980s and 1990s in particular.Footnote ²¹ The first ARAs in those two regions – Ghana in 1985 and Peru in 1991 – were both established in situations of acute fiscal crisis, with government revenues reduced to about 4% of GDP in both cases (Chand & Moene Reference Chand and Moene1999; Joshi & Ayee forthcoming; Mahon Reference Mahon2004; Moore & Schneider Reference Moore and Schneider2004; Terpker Reference Terpker1999).

Second, within sub-Saharan Africa, ARAs are found mainly in the countries most directly influenced by (a) the international agencies that have promoted the ARA model (DfID, the World Bank and the IMF), and (b) the New Public Management ideas popular in Britain in the 1980s and 1990s. African ARAs are located either in anglophone Commonwealth countries (Ghana, Uganda, Zambia, Kenya, Malawi, Mauritius, Tanzania, South Africa, Zimbabwe, Sierra Leone, Lesotho, The Gambia) or in other countries where elites have become substantially anglophone in the post-colonial period as a result of their aid and other relationships with anglophone international organisations (Ethiopia and Rwanda). ARAs are absent from almost all of francophone Africa. This anglophone bias has been strengthened by the exemplary and consultancy role of the South African Revenue Service (SARS). Motivated in large part by the political commitment to raise more tax revenue to redeem the social debt of apartheid, SARS has been the most consistent success story among ARAs in sub-Saharan Africa. Its staff provide advisory services to many other countries in the region.

Third, the international agencies that have promoted ARAs have clearly been concerned to help increase tax revenues. The IMF, the World Bank and DfID generally work closely together over taxation and fiscal reform issues.Footnote ²² One of the major roles of the IMF is to safeguard the global financial system. It has an interest in ensuring the fiscal health of the governments of poor countries, so that they can repay existing loans and easily contract new ones. This in turn gives it a natural bias in favour of encouraging and supporting governments to increase their tax revenues (Mahon Reference Mahon2004). It has frequently been criticised for this, and for encouraging African governments in short term ‘campaigns’ to meet revenue targets that may impact adversely on the possibility of developing more consensual taxation systems in the long term. Neither the World Bank nor DfID have the same direct institutional interest in raising revenues, but there is no doubt that this has been the primary goal of their tax administration projects. For example, the recent independent evaluation of World Bank support to public sector reform concludes that ‘The Bank's entry point for tax administration reform has typically been the need to increase revenues’ (IEG 2008: 57). The 2002 synthesis of a review of more than forty DfID-funded revenue projects, mainly in sub-Saharan Africa, makes it clear, from the first page of the executive summary, that increasing the ratio of tax collection to GDP was a primary objective and measure of achievement (DfID 2002).