6 - Fiscal Decentralisation as an Example of Institutional Ineffectiveness

Summary

This chapter focuses on state coordination brought about by the relationship between central and local governments, using local tax collection as a practical example. The dominant fact discussed in this chapter is the vacillation of the Tanzanian government between deconcentration, with centrally appointed civil servants in charge of local affairs, and full devolution to local governments. An ambitious local government reform was voted through in 1998, but never fully implemented. Responsibility for the collection of the property tax has changed three times in the last decade. As capacity is clearly missing at a local level, tax collection should optimally be under the central government for the time being. It is suggested that oscillating responsibilities in the recent past reflect hidden rent-seeking competition between local politicians, central government tax collectors, and ruling party members. In effect, corruption has been observed with both central and local tax collection. An important conclusion that comes out of the chapter and the discussion by Jan Willem Gunning is the role of capacity in the design of institutional structures in developing countries.

I Introduction

Decentralisation involves political, administrative, and fiscal reforms aimed at increasing the decision-making capacity and development efficiency of local administrations through the redistribution of powers and resources between administrative levels. The different dimensions of decentralisation can vary in importance and can be rolled out in different sequences. Decentralisation reforms very often target public service delivery (such as health, education, transport, water, and sanitation) in ways that may relate primarily to the administrative or the fiscal dimension. This may be because of technical and pragmatic concerns about appropriate sub-national government functions, but it may also reflect powerful political and institutional dynamics (Eaton et al., 2010).

Decentralisation has been the objective of important reforms in many developing countries and a major focus of the considerable support provided by development partners. Such reforms have swept across the world since the mid-1980s a trend seen by some observers as being influential for good governance and for improving the lives of ordinary citizens. African governments and international donors alike have indeed embraced the idea that decentralisation can promote development and good governance as local governments are more likely to be responsive to local needs, even though the record is mixed on several fronts. In any case, local governments’ share of public expenditure has more than doubled in many countries, and they now often play the leading role in the delivery of local public services. Academics are increasingly interested in evaluating the consequences of the change this evolution entails for the institutional relationship between levels of government, particularly for fiscal transfers (Falleti, 2005).

The focus of this chapter is on fiscal decentralisation, dwelling mainly on the administration of local revenue mobilisation given the centrality of financial resources in empowering local authorities to deliver on their mandate and improve their performance. Effective mobilisation of local revenues calls for a proper coordination of the local/central government mechanism and an administrative system with sufficient capacity to collect and analyse information, and plan and execute such proposals. In the case of Tanzania, this fiscal dimension was chosen to demonstrate the weaknesses of state coordination and the critical challenges involved in setting up an institutional arrangement addressing such weaknesses.

The chapter begins with an overview of the theoretical considerations behind the growing global trend towards decentralisation. It then summarises how the relationship between central and local government has evolved in Tanzania since pre-colonial times. It explains why, despite a recent reform programme, the current legal framework remains complex and confusing, impacting negatively on efficiency.

The capacity to collect revenue at local level is extremely limited in Tanzania. Expanding on previous studies (in particular Tanzi, 2000; Fjeldstad, 2001; Fjeldstad et al., 2010; Fjeldstad and Semboja, 2011; Masaki, 2018), the chapter identifies five key reasons for this: ambiguity in defining the roles and responsibilities of different state organs, leading to overlaps and conflicts of interest; arbitrariness, inconsistency, and unpredictability in government decisions and actions; weak institutional capacity for effective fiscal decentralisation; overdependence on the central government for financial transfers; and transparency and accountability asymmetry, with institutions reporting mostly to the central authorities. Practical consequences in terms of revenue are dramatic, including frequent cases of tax evasion, corruption, and even embezzlement of revenues, and constant political tension between local and central governments.

At a more general level, the chapter also considers how important fiscal decentralisation is for the success of decentralisation overall, and concludes by identifying three key directions for future reform in Tanzania.

II The Theory of Central–Local Government Relationships

Every country has different layers of government with different functions, based on their particular circumstances and experiences (van der Dussen, 2008). Consequently, decentralisation processes are initiated for different reasons. Some countries want to make the public sector leaner and more efficient. Others are motivated by disenchantment with the performance of centralised policies. Decentralisation may be motivated by a desire to contain or appease local demands for greater cultural and political autonomy. It may reflect an awareness of the global trends in institutional reform and a desire to not be left behind. Governments do not generally decentralise with the aim of pursuing greater macroeconomic stability and growth, though this may be an outcome (Martinez-Vazquez and Vaillancourt, 2011).

Decentralisation generally refers to the devolution of decision-making powers from the central government to local or sub-national governments. A related idea is de concentration, in which central governments retain decision-making power but diversify and customise the provision of public services to lower levels of government. According to the sequential theory of decentralisation, the extent to which decision-making power is devolved in practice depends on the sequencing of political, administrative, and fiscal decentralisation (Falleti, 2005).

Regarding fiscal decentralisation, there are four basic approaches: empowering local governments to set up their own tax systems; central retention of all taxes, with proceeds shared with local governments through intergovernmental transfers; assigning selected taxes exclusively to local governments; and sharing revenue from specific centrally collected sources with local government. Many systems are hybrids of these approaches, with the choice depending on considerations that may be technical, historical, demographic, economic, geographic, or political (Martinez-Vazquez and Vaillancourt, 2011).

In principle, decentralisation is often considered to be a desirable aim. Economists such as Oates (1972), who first developed the theory of fiscal federalism, argue that decentralisation should increase citizens’ welfare because service providers will have better information about diverse needs and preferences, and greater flexibility to address them. While such theories assume that governments are benevolent, a growing literature on public choice theory – which assumes that officials are selfish – also often favours decentralisation. A branch of this literature known as market-preserving federalism holds that decentralisation can incentivise good behaviour among government officials, control the intrusiveness and expansiveness of the public sector, and support effective private markets (Weingast, 1995; McKinnon, 1997).

It is theorised that decentralisation should reduce corruption, as accountability, information, and transparency should be greater at the local level; so should possibilities to encourage collective action and build social capital, which would lead to a higher probability of corruption being detected and punished (Boadway and Shah, 2009). If individuals and businesses are mobile, fiscal decentralisation should also help to constrain government misbehaviour by opening up the possibility of competition among jurisdictions.

There are counter-arguments, however. Decentralisation could create opportunities for rent-seeking by weakening central agencies’ scope for monitoring, control, and audit. By involving a larger number of officials in dealing with potential investors and revenue sources, political decentralisation can also create more opportunities for corruption and clientelism. The risks are especially high when elites dominate the local political scene. Incentives for corruption at the local level may also be higher owing to poorer compensation, lower career prospects, and lower morale (De and Prud’homme, 1994).

What does the evidence say? Based on cross-country comparisons, Huther and Shah (1998) find that fiscal decentralisation is associated with greater citizen participation, more political and democratic accountability, social justice, and improved economic management and operational efficiency; it is also found to have a positive effect on institutional quality and the quality of government (De Mello, 2011). There is strong evidence that fiscal decentralisation increases the share of education and health expenditures in total government expenditures, especially in developing countries (Shelton, 2007). Working on Bolivia, Faguet (2004) finds evidence that fiscal decentralisation increases investment in social sectors, such as education, urban development, water and sanitation, and health care.

Based on case studies, decentralisation has also been found to positively impact education outcomes such as literacy rates, years of schooling, dropout rates for primary and secondary education, public school enrolment, and test scores (Faguet, 2004; Peña, 2007, among others). In the health sector, positive impacts include decreasing infant mortality. Decentralisation has been found to increase access to water and sewage services and deliver better quality infrastructure at lower costs than in centralised settings, mainly where a community-driven development approach is used.

While generally positive, however, the evidence is mixed and incomplete – not least because it is difficult to isolate the effect of decentralisation on development from other processes such as economic growth and institutional changes in the public sector. There remain open questions about how diversity, complexity, proximity of local officials, political constraints, accountability, incentives, corruption, rent-seeking, and state capture by local elites affect the success of decentralisation – and about whether deconcentration can be as efficient as decentralisation.

Capacity may be the key factor in determining the extent to which decentralisation succeeds: services may improve when decentralised to high-capacity local governments and deteriorate when decentralised to low-capacity local governments. For instance, theories of public finance often tend to assume that local governments will have fiscal capacity, defined as the ability to raise tax revenues ‘given the structure of the tax system and its available powers of enforcement’ (Besley and Persson, 2013). However, in practice local governments in low-income countries tend to lack fiscal capacity. Africa has performed particularly poorly compared with the rest of the world in terms of the level of local revenue generation and service delivery, with local governments depending heavily on central government grants to finance their budgets.

The diverse and complex political economy challenges that underlie lack of capacity at the local level rarely receive sufficient attention (McLure, 1998). They include the incentives and behaviours of national-level politicians and bureaucrats, who shape the rules of the intergovernmental fiscal game and how they are implemented, and the local-level political economy dynamics among elected local councillors, local government staff, and citizens. When local governments lack capacity, an appropriate balance needs to be found between central oversight and local autonomy.

III The Evolution of Local Government in Tanzania

Tanzania’s history of local government dates back to the chiefdoms of the pre-colonial era, as summarised in Table 6.1. During the first decade of independence, 1961–71, the government replaced native authorities with local officers who were democratically elected, in common with other newly independent African states, with the aim of improving the delivery of public goods and services. This was partly the result of the independence euphoria, but also reflected the genuine determination of the new government to bring fundamental changes to the citizens. The leadership’s reflection on a strategy for national social and economic development led to the Arusha Declaration of 1967 that committed Tanzania to a development strategy based on ‘socialism and self-reliance’. The emphasis of the Second Five Year Development Plan (1969–74) was on rural development, which required further administrative reforms – at the local level – in order to improve the capacity and effectiveness of the machinery of government in carrying out the new rural development effort (Collins, 1974). However, local governments remained closely supervised by, managed by, and accountable to the central government. This reflected in part the British system of government the country inherited, in part the aim of strengthening national unity, and in part the fear – not publicly acknowledged – that local authorities, just like the independent cooperative unions, could become a source of opposition (Mnyasenga and Mushi, 2015).

Table 6.1 The evolution of local government in Tanzania

Period	Type of local governance
Pre-colonial era	Chiefdoms, and councils of elders.
German era (1884–1917)	Mainly direct rule but also limited urban authorities.
British era (1917–61)	Native authorities encouraged since 1926 (indirect rule); township authorities for large urban areas; Municipalities Ordinance 1946; Local Government Act 1953.
First decade of independence (1961–71)	Chiefdoms abolished; inclusive local authorities encouraged; local governments overwhelmed by duties, with limited resources; rural authorities abolished 1972, urban authorities abolished 1973.
Deconcentration (1972–82)	A system of deconcentration of central government replaced the comprehensive local government system that had existed for a decade.
Reinstitution of local government (1982–95)	Urban Councils (Interim Provisions) Act 1978 required that town and municipal councils be re-established from 1 July 1978; 1982 comprehensive local government legislation passed; 1984 comprehensive system of local government re-established.
Local government reform (since 1996)	Comprehensive programme of reforming local governments to make them efficient, effective, transparent, and accountable.

Source: History of Local Government of Tanzania by United Republic of Tanzania President’s Office, Regional Administration and Local Government

It quickly became clear that local authorities were failing to achieve the expected results owing to, among other factors, expansion of services that did not match the available financial resources, lack of competent personnel, and rampant mismanagement of funds (both local and grants from the central government), leading to poor social and economic performance. The period also witnessed an ascendancy of politics and politicians over the bureaucracy that led to a loss of consistency in policies and operations at both central and local government level.

In 1972, local governments were abolished, and the government created new regional and district committees (in place of district councils), which were given responsibility to coordinate both economic and social development activities while reporting to the Prime Minister’s Office. By then, Tanzania had thus opted for a deconcentration rather than a devolution type of decentralisation. In effect, the central government started to directly manage the local development process and provision of social services.

There was, however, a lack of preparedness in the implementation of this reform that showed up in low human capacity, lack of resources, and inherent disincentives for task compliance in the whole administrative system. The social services infrastructure collapsed in the severe economic crisis of the late 1970s and early 1980s, which, under the strong pressure of the donors, led a few years later to a complete change of development strategy, from a socialist to a market economy. Notable at that time was the overextension of the state, which placed great pressure on its capacity, while the heightened ideological content and politicisation of the government decision-making process eroded the authority and self-confidence of the bureaucracy. Faced with that erosion in the capacity to carry out the economic management tasks of government, donors increasingly pressed for administrative reforms, including the reinstitution of local government institutions. The reintroduction of urban authorities had taken place in 1978. Then a series of laws on local government were passed in 1982 and there was a constitutional amendment in 1985.

These measures proved to be flawed. They did not clearly define the relationship between central and local government – in practice, the centre retained strong powers of control and supervision, and the structure of local government authorities (LGAs) overlapped with that of the ruling party (Mnyasenga and Mushi, 2015). LGAs were given only limited power to mobilise their own human resources, implement their own plans and strategies, and raise revenue, borrowing, and spending. From the early 1990s, various studies, commissions, workshops, and seminars pointed out the complexity, ambiguity, and fragmentation of the legal framework, with overlaps and conflicts among legislation, circulars, standing orders, and other regulations from ministries responsible for health, education, extension services, water supply, and rural roads. At a higher level, it also became evident that fundamental political, administrative, and economic reforms were imperative for the government to improve economic efficiency and effectiveness. Several far-reaching economic and political reforms were thus introduced during this period, including macroeconomic stabilisation and fiscal restraint, market liberalisation, and privatisation on the economic side, and also the establishment of multiparty democracy in 1992 on the political side.

Tanzania consequently embarked on a new Local Government Reform Programme (LGRP) in 1996, accompanied by the decentralisation by devolution (D by D) strategy, in which LGAs were supposed to be largely autonomous institutions, free to make policy and operational decisions consistent with the country’s laws and policies, and with the power to possess both human and financial resources. Reforms were aimed at downsizing central government, reforming local governments, and decentralising more powers to them. It was expected that the D-by-D strategy would yield, among other outputs, the delivery of quality services to the people in a participative, effective, and transparent way. There was, however, little analysis and documentation of the implementation challenges at both national and local authority levels. The LGRP was to be implemented in two phases – a stand-alone programme from 1998 to 2008, and integration into the government system from 2009 to 2014. It set out to address five dimensions:

1. 1. Financial: giving local authorities more sources of revenue, including conditional and unconditional grants from the central government;

2. 2. Administrative: de-linking centrally controlled personnel from sectoral ministries and integrating them in the local government system;

3. 3. Central–local relations: limiting the roles of central government to policymaking, support and facilitation, monitoring, and quality assurance;

4. 4. Service function: decentralising the management and provision of public services, with the aim of enhancing their quantity and quality; and

5. 5. Democratic: strengthening local democratic institutions, enhancing public participation and bringing control to the people.

By the end of the first phase in 2008, however, only four pieces of legislation had been partially amended,¹ and a legal harmonisation task force had only just started to review sector laws and policies (Mnyasenga and Mushi, 2015). Rather than clarifying overlaps in responsibility, some of these amendments actually exacerbated ambiguity: for example, Act No. 6 of 1999 and Act No. 13 of 2006 introduced a provision that the central government could do ‘any such other acts and things as shall facilitate or secure the effective, efficient and lawful execution by the District Authorities of their statutory or incidental duties’. By the end of the second phase in 2014, neither a comprehensive local government law nor harmonised central and sector legislation were in place. This remains the case today (April, 2023).

In summary, government decentralisation in Tanzania has gone through four phases: first, active decentralisation was pushed by the national-level bureaucratic elite that swiftly emerged following independence; second, the consolidation of that process was impeded by the major disruption that followed the Arusha Declaration and the increasing state control over the whole economy, at a time most able civil servants were transferred to manage the new parastatals, spreading available talent very thinly (Van Arkadie, 1995); third, an attempt at reverting the process took place some ten years later, with the major institutional adjustment process that followed the crisis of the early 1980s and led to the re-establishment of a market economy, but, for various reasons, the economy remained de facto essentially centralised; and, fourth, under the pressure from donors using economic arguments, including the need to reduce the role of the central government (World Bank, 2004) and improve the delivery of public services as well as the participation of citizens (Smoke, 1994; Manor, 1999; World Bank, 1999; Olowu, 2000), decentralisation, in its devolution definition, is again posted as a major reform objective (LGRP laws). How far has this reform gone?

IV Local and Central Governance in Tanzania Today: A Complex and Confusing Legal Framework

The legal framework governing relationships between central and local government in Tanzania is complex and confusing. For example, local authorities are legally mandated to make and implement their own development plans, finding their own sources of revenue – but central ministries are also legally empowered to determine the sources of local government revenue, and can veto decisions made at the sub-national level. Sector ministries are also legally empowered to intervene in the functions of LGAs.²

The overwhelming power of the minister responsible for local government is suggested by the sheer number of mentions in the relevant legislation: according to Mnyasenga and Mushi (2015), the minister is mentioned 95 times in the 156 sections of the Local Government (District Authorities) Act, 1982 [CAP 287 R.E. 2002]; 80 times in the 111 sections of the Local Government (Urban Authorities) Act, 1982 [CAP 288 R.E. 2002]; and 60 times in the 65 sections of the Local Government Finance Act, 1982 [CAP 290 R.E. 2002]. Most of these mentions are concerned with the control and supervision of local government powers, functions, and finance through approval powers; appellate power; issuance of guidelines, regulations, directives, orders, and direct interventions; appointment and transfer powers of local government staff; disciplinary powers over local government staff; variation of local government functions; and powers to dissolve local government councils. Most of these powers are discretionary and can be delegated by the minister to any public officer.

In practice, research indicates that the central government indeed exercises tight control over LGAs. Studies carried out by REPOA (2008), Tidemand and Msami (2010), and Kunkuta (2011) reveal the most frequently used mechanisms: issuing policy statements and guidelines; giving directives and commands that direct the LGAs to perform or not to perform certain activities; issuing circulars; discipline and transfer of local government staff; and setting budget ceilings. In the opinion of 87.4 per cent of those asked by the researchers, the minister’s power negatively influences the autonomy of LGAs.

The same studies also observed that Regional Administrative Secretariats negatively affect the autonomy of LGAs. In theory, these regional authorities should play a facilitating role, providing technical advice, support, and supervision.³ In practice, they put heavy pressure on local authorities, frequently issue directives, and veto development plans and programmes that are deemed to be inconsistent with national policies. While this can sometimes be justified, experience suggests these powers are exercised excessively – in particular, political tensions emerge in constituencies dominated by opposition parties when LGAs are pressured to prioritise implementing the party manifesto above their own plans.

Figure 6.1 depicts how LGAs receive directives, guidelines, circulars, memoranda, codes of conduct, and so on from a wide variety of other governmental bodies: the Ministry of Finance and Planning; the President’s Office – Regional Administration and Local Government; sector ministries, such as education and health; and regional and district authorities. LGAs lack the capacity to implement them efficiently, or to comply with these varied stakeholders’ different reporting requirements and formats. This results in data being unreliable: there are, for example, substantial variations between budget figures presented to local councils and to the parliament, information on expenditure compiled by the Prime Minister’s Office Regional Administration and Local Government (PMO-RALG), and what appears in the audited final accounts (Fjeldstad et al., 2010).

Figure 6.1 Interactions between central and local government

Source: Construct by authors

Overall, the general feeling about the 1998 reform of the functioning of LGAs and their relationship with the central government is that it is unfinished business. Progress seems to have taken place in the volume of services delivered and in their quality. It is also the case that LGA expenditures and employment weigh more in public spending and the civil service today (April 2023). Yet the control of LGAs over their staff and their total spending is limited, owing to what two evaluators call a ‘dual level of authority’. The same evaluators concluded that, up to 2008, the end of the first stage of the reform, LGAs were not more powerful than they were in 2000. To date (April 2023), the situation remains more or less the same, or may have slightly deteriorated with the increased control by the central government under the fifth phase, with the recent decision to transfer key staff in departments of land and water to the central government, including further transfer of local government sources of revenue. Thus, despite the significant devolution of authority and resources to sub-national levels of government, persisting capacity deficits, increased financial dependence on the central government, and political and institutional constraints impact negatively on the pace of reforms, and also mean that the achievements have fallen short of the intentions of the reform agenda.

Among the weaknesses underscored by observers, the issue of fiscal decentralisation ranks high. We now turn the spotlight on it.

V Fiscal Decentralisation: The Challenges of Revenue Collection

The LGRP aimed to ensure discretionary powers for local councils to levy local taxes and fees and pass their own budgets, reflecting their own priorities alongside the obligation to meet nationally mandated standards in the delivery of the public services for which they are responsible. The bulk of the funding for these services – which include primary education, primary health, local roads, potable water, sanitation, and agricultural extension – comes from central government, as do the salaries and emoluments of council civil servants. Transfers are allocated according to a formula that takes into account socioeconomic factors such as the size of population, area, poverty, and access to health facilities.

There is ample evidence that the reforms have not been effective in increasing LGAs’ fiscal autonomy. Only a few large urban councils in Tanzania can finance a substantial share of their expenditure from their own revenue sources. Between 2000 and 2007, revenue collected in urban LGAs increased by 36 per cent, but declined by 4 per cent in rural LGAs; this is attributed to the central government abolishing certain ‘nuisance taxes’ in 2003/4, inappropriate tax design and poor collection systems (REPOA, 2007; Fjeldstad et al., 2010). In the 2006/7 financial year, LGAs collected about TZS 60 billion in local taxes, representing only about 7 per cent of total LGA expenditure (Tidemand and Msami, 2010). In 2012/13, transfers from the central government accounted for 85 to 90 per cent of local budgets – on a par with corresponding numbers from other African countries, such as Lesotho (90 per cent), Uganda (88 per cent), and Ghana (69 per cent). The share of total national tax revenues collected by local governments – about 6 per cent – remained almost unchanged since 1996 (Fjeldstad, 2003).

Table 6.2 lists the main sources of revenue for local governments. The most important are classified as non-tax, including produce taxes, market fees, service levies, licences and permits, property tax, and fines and penalties. Collecting such revenues creates opportunities for rent-seeking; doing so efficiently requires robust monitoring and enforcement systems to ensure transparency and accountability, along with skilled staff who are costly to employ and maintain at the local level (Besley and Persson, 2013). Guidance from central and sectoral ministries is lacking; indeed, political interference with local revenue collection is prevalent.

Table 6.2 Revenue sources for local governments

Taxes on property	Other taxes on permission to use goods
Property rates	Forest produce licence fees
	Building materials extraction licence fee
Taxes on goods and services	Hunting licence fees
Crop tax (maximum 3% of farm gate price)	Muzzle-loading guns licence fees
Forest produce tax	Scaffolding/hoarding permit fees
Taxes on specific services	Turnover taxes
Guest house levy	Service levy
Business and professional licences	Entrepreneurial and property income
Commercial fishing licence fee	Dividends
Intoxicating liquor licence fee	Other domestic property income
Private health facility licence fee	Interest
Taxi licence fee	Land rent
Plying (transportation) permit fees
Other business licence fees
Motor vehicle and ferry licences	Other local revenue sources
Vehicle licence fees	Administrative fees and charges
Fishing vessel licence fees	Fines, penalties, and forfeitures

Source: Construct by authors

Political economy reasons, or more exactly unaligned incentives and disincentives, are among the critical forces that help to maintain such a complex and inefficient system. Important stakeholders, including bureaucrats and politicians, as well as powerful taxpayers, resist changes in an attempt to protect their influence and control of the local tax system. In the case of Tanzania, Fjeldstad (2001) maintains that such an environment offers informal incomes for civil servants and their social network members and provides a visible arena for local councillors to play out their political aspirations vis-à-vis their constituents. It also provides incentives for some powerful taxpayers, in particular businesspeople, landowners, parastatals, and the cooperative unions, to seek to retain the status quo, since it facilitates evasion.

A Property Tax: A Case Study

Property tax provides a case study of the challenges. While property valuations are based on the number of storeys and type of floors, walls, and roof, there is no clear and consistent methodology: more accurate valuations would require financial skills, infrastructure, and documentation, which are generally lacking. Valuations are often arbitrary and highly disputed, and in 2017 the central government proposed to replace them with flat lump sums depending solely on the number of storeys and urban or rural location.

Responsibility for collecting property tax has changed three times in a decade. Before 2008, when it lay with municipalities, revenue collection was poor, corruption was rife, and local politicians often interfered (Fjeldstad et al., 2010; Fjeldstad, 2015). As a result, the system was partially centralised: in 2008, the Tanzania Revenue Authority (TRA) was given the responsibility of collecting property tax on behalf of municipalities in Dar es Salaam, which retained the power to declare an area rateable, set rates, and grant exemptions.

However, mutual mistrust impeded cooperation between the TRA and municipal authorities. Imperfect information flowing to central operators created opportunities for corruption, contrary to what was expected from this re-centralisation decision, and negatively impacted revenue collection. The World Bank became concerned that the move indicated lack of government commitment to the decentralisation process, and temporarily stalled funding of the valuation and assessment of properties in Dar es Salaam. In February 2014, the system was thus re-decentralised. An elected councillor in one of the municipalities said:

Re-decentralisation of property tax administration is a perfect move. From the time TRA started to collect property tax, revenue deteriorated. I strongly believe that the collection by municipality will be far better than that of TRA. First and foremost is that the municipality knows that it is collecting the money to finance its budget, so all efforts will be instituted to meet the target.

(Fjeldstad et al., 2019, p. 14)

The TRA, in contrast, reacted with resignation and frustration. A senior officer argued:

All municipalities are very happy about re-decentralisation of property tax collection because right from the start when TRA took over they were disappointed […] [they] have been trying to make tricks so that TRA is perceived inefficient. For example, when TRA took over, all municipalities set larger targets to TRA year after year despite the fact that the tax base remained the same.

(Fjeldstad et al., 2019, p. 14)

In July 2016, the central government again made collection the responsibility of the TRA – a decision that took municipalities by surprise, as it appeared not to have been based on a comprehensive assessment of the challenges experienced between 2008 and 2014 or the performance of municipalities since 2014 (Fjeldstad et al., 2017). Yet the failure to establish a stable and predictable regime is reflected in a significant proportion of potential revenue going uncollected: Budget Execution reports for the past three fiscal years indicate an average collection of local revenue of between 47 and 53 per cent of projections.

B Lack of Local Capacity for Revenue Collection

The vacillation on property tax collection indicates a more general dilemma: while decentralising tax collection makes sense in principle, LGAs in Tanzania have always lacked the administrative, institutional, and fiscal capacity to collect local taxes. This is a problem common to most African countries, particularly in rural areas (Fjeldstad et al., 2014). A study conducted by REPOA for the PMO-RALG found that executive officers at the ward and village levels in Tanzania are typically educated only to primary- or secondary-school level and have minimal skills to handle the functions their posts require (REPOA, 2007).⁴ Although there are many unemployed graduates, they are not attracted by the status, remuneration package, and working environment at ward and village level. The councils studied did not have sufficient staff trained to collect, process, and manage fiscal data or conduct quantitative analyses to guide policymaking.

Some local governments tried to improve capacity by outsourcing revenue collection to private agents: property taxes, bus stand and parking fees, forestry levies, and market fees. However, arrangements were often poor owing to a lack of knowledge about the local tax base or clear methods of establishing charge rates. Assessment of revenue potential was generally ad hoc and based on outdated figures, suggesting corruption and inefficiency. For example, collection of fees at the Ubungo Bus Terminal in Dar es Salaam was outsourced to a private agent owing partly to concern about fraud among council officials – but the private agent then retained most of the revenue collected. A conservative estimate by the Chr. Michelsen Institute and REPOA researchers is that the city council received only about 44 per cent of revenue collected between 2002 and 2006.

Local government capacity is usually augmented by staff from central government institutions. However, these institutions themselves have shown limited capacity for designing, developing, and implementing measures to strengthen local government. Most of the staff are not accountable to the LGAs but to the Local Government Service Commission and/or parent sectoral ministries such as education and health in central government. Their effectiveness is limited by lack of knowledge of local conditions, with fragmented management of staff at the local authority level exacerbated by under-financing and subterfuge. Asymmetries in reporting and accountability create significant potential for overlaps and inefficiencies.

These problems are compounded by limited use of technology for planning and reporting, particularly fiscal planning and accountability. The International Monetary Fund considers Tanzania to have one of the best public financial management systems in sub-Saharan Africa (Nord et al., 2009), and most district councils have computerised budget and accounting systems – but the REPOA research team found that limited staff capacity means these systems are often not actually used. Most councils still carry out budgeting and accounting manually, with huge implications for fiscal management and operational efficiency in general.

When weak capacity at the local government level leads to inefficiency, as planned activities cannot be properly implemented, this fuels perception of corruption: survey data indicate that 72 per cent of citizens viewed corruption as a serious problem in councils in 2013, up from 59 per cent in 2003 (Fjeldstad et al., 2008).

C Can Central Transfers Be Used to Build Fiscal Capacity?

Local revenue collection is important for fiscal autonomy. It creates a compelling sense of local ownership of resources and builds a strong basis for local oversight of resource use. It is thus considered important as it increases public officials’ accountability to their constituents, even though this is not the only way of making progress on that account.⁵ It also incentivises efficiency and limits the pressure for ever more central transfers and public debt (McLure, 1998). A study of local budgets in East African countries found that collecting more local revenue led to a higher share of expenditure on service delivery, while dependence on intergovernmental transfers and development aid was associated with a higher budget share for administrative costs and employee benefits (UN-HABITAT, 2015).

However, decentralising revenue collection creates high potential for mismanagement and corruption: local governments may be better at eliciting people’s preferences, but they have a higher chance of being captured by local elites and politically powerful groups. When enforcement and monitoring systems lack capacity, the cost of collecting local revenue can be a significant proportion of the revenue collected – sometimes even exceeding it. It can be argued that when levels of administrative capacity are low, it makes sense to entrust the collection of sub-national taxes to the central tax administration and establish an elaborate arrangement for provision of capacity-enhancing fiscal transfers to LGAs. Almost twenty years ago, Fjeldstad (2001) observed that it was unrealistic to expect that the administration in many local governments in Tanzania would have adequate capacity and the required integrity to manage increased fiscal autonomy. He concluded that ‘In fact, there is a real danger that, in the absence of substantial restructuring of the current tax system combined with capacity building and improved integrity, increased autonomy will increase mismanagement and corruption.’ The situation has barely changed.

Is it possible for grants from the central government to build capacity to collect revenue at the local level? Some have argued that grants from the central government crowd out local revenues, sapping the incentive for LGAs to collect their own dues (Shah, 2006; Masaki, 2018). However, the evidence mainly comes from studies in countries with sound fiscal institutions: analysis by Masaki (2018) strongly suggests that intergovernmental transfers can help expand local revenues in Africa, especially in rural areas. In urban areas, the marginal positive effect is lower as there tend to be more robust fiscal institutions and higher political costs associated with increasing taxes (Resnick, 2012). Evidence also shows that when fiscal transfers facilitate the provision of public goods, this improves voluntary tax compliance (Masaki, 2018).

The question is how long it will take for central government transfers to build and strengthen local capacity so that they can have sufficient capacity to mobilise and manage their own revenue collection. And what incentives will be needed to enable LGAs to build such capacity? There is a genuine issue of capacity at the local government level, which necessitates continued central government support and balancing of the central government role at that level if there is genuine desire to see fiscal autonomy take root. The most important thing is to have the proper incentives in place and effective monitoring from the central government and its agencies. The Tanzania Social Action Fund programme has been able to build capacity of local communities through engagement of local citizens in direct implementation of local projects in the health, water, education, and roads sectors. On the other hand, it should be stressed that some local communities in Tanzania exhibit fairly satisfactory performances in terms of local governance and financial management, suggesting that capacity building in other communities might indeed achieve much (see Boex and Muga, 2009; King, 2014).

As earlier noted, efforts to build local capacity to collect revenue need to be accompanied by measures to ensure that citizens have the information on local government revenue, budgets, and accounts that they need to hold their leaders to account. This challenge is most acute when formal accountability institutions, such as audits and legislative reviews, are weak owing to limited knowledge of what is happening at the local level, as is common in most local authorities in Tanzania (Msami, 2011). LGAs publish the financial information required by the central government and development partners, but researchers have found that much of this information does not reach the public. Only a small minority of people are aware of basic budget information (Fjeldstad, 2004, 2006), and public notices are often too technical for ordinary citizens to understand (REPOA, 2007).

Given the objectives of decentralisation reforms of fiscal autonomy for local government, measures should be taken to build the necessary capacity and create the necessary environment for expenditure efficiency and accountability of local officials – promoting more effective coordination, stability of the system, policy consistency, and predictability of the decision-making process.

VI Conclusion: Directions for Reform

The centre has an important role to play in the quest for local autonomy. Future reforms in Tanzania should address the complex and confusing administrative relationship between central and local government, along with limited fiscal capacity at the local level and the frequency of centrally imposed changes in revenue regimes, which make it harder to develop sustainable fiscal capacity at the sub-national level.

A Revisit the Fiscal Decentralisation Agenda

Effective collection of non-tax revenue hinges on a constructive working relationship between central and LGAs to create sound fiscal institutions and accountability to local taxpayers. There are certainly costs to local taxation as its administration needs more tax evaluators and collectors, and greater capacity to monitor and penalise non-compliance. Given limited fiscal capacity at the sub-national level, it makes sense, as an assured way to build the requisite institutional capacity at that level, to have the central government collect revenues and establish clear legal mechanisms to transfer part of those revenues to LGAs based on recognised resource endowment, the need in terms of public services to be provided, and a fiscal capacity-building component.

B Address Institutional Set-Up to Create Efficiency

The profusion of conflicting laws leads to haphazard influence on LGA operations from central and sectoral agencies and regional and district leadership. This is counterproductive. A reform agenda should address harmonisation of laws and create a framework for centre–local interactions on policy and revenue mobilisation that prevents abuse and promotes efficiency.

C Address Unpredictability of Government Decisions

There has always been the question as to what motivates the central government officials to give up powers and resources to sub-national governments. Any decentralisation measure tends to reduce the power and authority that national politicians enjoy relative to sub-national actors. Yet the same political personnel recognise at the same time the efficiency and governance gains to be expected from decentralisation. This could be the basis for unpredictable reform behaviour as the incentive schedule changes over time and context. Unpredictability of government decisions and actions deters investment, slows economic activity, and has negative implications for decentralisation reforms. A reform agenda should rationalise the conduct of discretionary decisions and actions by the central government and set up a consultative forum to engage local authorities in policy discussions, with the Association of LGAs in Tanzania playing a role.

Tanzania’s government has already expressed the desire, through the LGRP, for full-fledged fiscal decentralisation – but there has, as yet, been limited realisation. A renewed reform agenda is needed, with central transfers allowing for smooth operations at the LGA level until the requisite fiscal capacity is built.