Report No. 34900-TZ Local Government Taxation Reform in Tanzania: A Poverty and Social Impact Analysis (PSIA) Report on Economic and Sector Work The World Bank Social Development Department January 2006 Copyright © 2006 The International Bank for Reconstruction and Development 1818 H Street, N.W. Washington, D.C., 20433, USA All rights reserved Manufactured in the United States of America First Printing December 2005 This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, and e-mail pubrights@worldbank.org. TABLE OF CONTENTS Acknowledgments ......................................................................................................................................iv Abbreviations and Acronyms ....................................................................................................................v Executive Summary...................................................................................................................................vi Chapter 1: Introduction ............................................................................................................................1 1 Objectives of the PSIA .......................................................................................................1 2 Background to Local Tax Reform ......................................................................................3 3 Socioeconomic Context ......................................................................................................4 4 The Decentralization Debate...............................................................................................6 Chapter 2: Research Instruments And Methodology.............................................................................8 1 Site Selection ......................................................................................................................8 2 Secondary Data Collection (Council Level).......................................................................9 3 Key Informant Interviews (Council Level).........................................................................9 4 Household Surveys, Urban and Rural Councils..................................................................9 5 Business and Market Surveys, Urban and Rural Councils .................................................9 Chapter 3: Distributional Impacts on Households ...............................................................................11 1 Profiles of Selected Councils............................................................................................11 2 Characteristics of Sample .................................................................................................11 3 Impact by Council and Income Group..............................................................................12 Chapter 4: Distributional Impacts On Businesses................................................................................18 1 Sample Characteristics......................................................................................................18 2 Impacts..............................................................................................................................18 Chapter 5: Stakeholder Perceptions and Suggestions: Households and Businesses..........................23 1 General Views on Taxation Reform .................................................................................23 2 Standards of Local Government .......................................................................................23 3 Demands for Improvement in Local Services ..................................................................25 4 Other Fees and Charges ....................................................................................................26 5 Who Should Pay?..............................................................................................................26 6 Attitudes toward Property Tax..........................................................................................27 7 User Fees...........................................................................................................................28 8 Conclusions.......................................................................................................................28 Chapter 6: Stakeholder Perceptions and Suggestions: Local Government Officials ........................30 1 Perceptions of Officials, Key Informants, Business, and Citizens ...................................30 2 Unfavorable Outcomes .....................................................................................................32 3 Favorable Outcomes .........................................................................................................32 4 Strategies of Adjustment to Tax Reform ..........................................................................32 5 Property Taxes..................................................................................................................34 Chapter 7: Local Government Taxation in Tanzania: Conclusions, Policy Dialogue, and the Way Forward .....................................................................................................................................................37 1 Summary of Main Findings ..............................................................................................37 2 Establishing Policy Dialogue............................................................................................37 3 Next Steps.........................................................................................................................38 4 Modeling Local Taxation Options....................................................................................39 5 Other Analytic Tasks ........................................................................................................39 Bibliography..............................................................................................................................................41 List of Tables: Table 1: Consolidated LGA Own Revenue Collection Data (Tshm/-)........................................................ 5 Table 2: LGA Financial Resource Flows, 2001­02 to 2003­04 (Tshm/-)................................................... 5 Table 3: Central Government Grants to LGAs, 1999 to 2005 (Tshm/-)...................................................... 6 Table 4: Criteria and Justifications for Selection of Field Study Locations ................................................ 8 Table 5: Sampling per Council or per Market ........................................................................................... 10 Table 6: Profiles of Sample Survey Councils ............................................................................................ 11 Table 7: Local Taxes Paid before and after the Tax Reforms by All Households..................................... 13 Table 8: Taxes Paid before and after the Tax Reforms: Households Stratified by Income....................... 14 Table 9: Taxes Paid by Sample HHs Before and After the Tax Reforms by Council............................... 16 Table 10: Moshi Municipal Council Own Revenue Collection (Tsh) ...................................................... 17 Table 11: Business Survey: Key Characteristics of Respondents (N=58)................................................. 19 Table 12: Summary View of Local Taxes Paid by Businesses, Sample Surveys...................................... 20 Table 13: Business Survey: Average Local Taxes Paid before and after the Reforms, Stratified by Turnover ............................................................................................................. 21 Table 14: Business Survey: Total Local Taxes Paid before and after the Reforms, by Council (Tsh `000) .................................................................................................................. 22 Table 15: Responses to Whether Tax Collection Is Fair............................................................................ 23 Table 16: Responses to Whether Taxes Collected Are Receipted............................................................. 23 Table 17: Taxes Stated by Household Respondents As Providing Receipts for Payment (No. of Affirmative Responses) .............................................................................................. 24 Table 18: Requests for Unofficial Payments and Amounts (Percent, Tsh) ............................................... 24 Table 19: Rating of Quality of Services Provided by Local Councils....................................................... 25 Table 20: Priority Ranking of Services by Survey Respondents............................................................... 26 Table 21: Other Fees and Taxes Mentioned in the Kilosa Household Survey 2005 ................................. 26 Table 22: Responses on Relative Wealth and Payment for Services......................................................... 27 Table 23: Attitudes toward Residential Property Taxes ............................................................................ 27 Table 24: Views on Preferred Taxes to Keep or Introduce........................................................................ 28 Table 25: Responses to How Services Should Be Paid For....................................................................... 28 Table 26: Kilosa Council: Trends in Tax Revenue, 2000­04.................................................................... 31 ii List of Boxes: Box 1: Impact on Business, Kilosa (Key Informant, District Executive).................................................. 21 Box 2 Moshi Businessman's Views on Tax Reform ..................................................................................29 Box 3 Dissenting View on Tax Impact, Kilosa (Official, Kilosa Council) ................................................ 32 Box 4: Property Tax Exclusions in Moshi Town....................................................................................... 34 Box 5: Thoughts on Property Tax, Kilosa (Key Informant, District Executive) ....................................... 34 List of Figures: Figure 1: Change in Taxes Paid by Socioeconomic Groups...................................................................... 13 Figure 2: Change in Taxes Paid by Scale of Business............................................................................... 20 iii Acknowledgments The study was undertaken by Sholto Cross (Lead Researcher) and Frank Ellis (Economist) of the Overseas Development Group (ODG), University of East Anglia, along with Joseph Semboja (Executive Director), Joanita Magongo and Lucas Katera (Researchers), Erasto Ngalewa (Capacity Building Coordinator), and Mary Kombe and Samuel Ebeneri (Field Supervisors) of Research on Poverty Alleviation (REPOA), Dar es Salaam. Alexandra Winkels of ODG provided research support. The World Bank team consisted of Paul Francis (Task Team Leader, Senior Social Development Specialist, AFTS2), Matt Glasser (Senior UrbanDevelopmentSpecialist, AFTU1) and SabineBeddies(SocialScientist, SDV).1 The study team wishes to thank the many institutions and individuals who have contributed their time and support to this study. In particular we wish to acknowledge the support and cooperation of Tanzania's Ministry of Finance (Joseph Mallya, Mr Kamugisha, Moses Mwampeta, Simon Lapper, and Arthur Mwakapugi), PO-RALG (Richard Musingi), and the Vice President's Office (Servacius Likwelile). Local government officials in Ilala, Kilosa, and Moshi were most cooperative, as were a number of other respondents in these councils. The respondents to the household and business surveys, and the many individuals who assisted in the fieldwork, without whose cooperation this study would not have been possible, also deserve our thanks and acknowledgement. Gerard Howe, Kevin Quinlan, and Carl Kalapesi of DFID Tanzania, Clemens Beckers of GTZ Tanzania, and Lucie Luguga and Daimu Mkwawa at UNDP were a source of support and advice on the design of the study and subsequent dissemination. A number of other individuals who have been working in the field of local government supplied helpful information and advice, and we should like to thank in particular Per Tidemand, Ole Therkildesen, Dan Kobb, Brian Cooksey, and Geir Sundet for their support. Jamey Boex and Philip van Ryneveld of the GSU team provided useful cooperation. The data, analysis, and conclusions of this research study are of course solely the responsibility of the study team. Within the World Bank, we are particularly grateful to our two peer reviewers, Roland White (Senior Public Sector Management Specialist, EASPR) and Meskerem Brhane (Social Development Specialist, MNSRE), as well as to Robert Utz (Senior Economist, AFTP2), Judy O'Connor (Country Director, Tanzania), Karen Brooks (Sector Manager, AFTS2), Melanie Marlett (Country Program Coordinator, AFCTZ), Robert Townsend (Senior Economist, AFTS2), and Maria Correia (Lead Social Development Specialist, AFTS2) for their advice and guidance. We also wish to express appreciation for the support of the World Bank office in Dar es Salaam, especially Philip Mpango.2 We specifically thank Ernestine Ngobo-Njoke (AFTS2) and Danielle Christophe (SDV) for their editorial and logistical support. The study was supported by several trust funds, generously provided by the Governments of Germany and Belgium. In addition, DFID Tanzania provided additional resources to deepen fieldwork and technical support, while GTZ Tanzania assisted with the costs of the Tax Reform Workshop held at REPOA on 20 April 2005. 1World Bank Africa Region acronyms: AFTS2 = Africa Region, Environmentally and Socially Sustainable Development 2; AFTU1 = Africa Region, Water and Urban 2; AFTP2 = Africa Region, Poverty Reduction and Economic Management 2; AFCTZ = Tanzania Country Team. Other World Bank acronyms: SDV = Social Development Anchor; EASPR = East Asia Poverty Reduction and Economic Management Sector Department; MNSRE = Middle East and North Africa, Rural Development, Water and Environment Group. 2AF iv Abbreviations and Acronyms CC - City Council CMI - Chr. Michelsen Institute DC - District Council DFID - Department for International Development GSU - Georgia State University GTZ - Deutsche Gesellschaft für Technische Zusammenarbeit (German Technical Cooperation Agency) HH - Household IOB - Policy and Operations Evaluation Department (The Netherlands Ministry of Foreign Affairs) LADDER - Livelihoods and Diversification Directions Explored by Research LGA - Local Government Authority LGFA - Local Government Finance Act LGRP - Local Government Reform Programme LGSP - Local Government Support Project LLG - Lower Levels of Government MC - Municipal Council NFU - Norwegian Association for Development Research NIBR - Norwegian Institute for Urban and Regional Research ODG - Overseas Development Group PETS - Public Expenditure Tracking System PO-RALG - President's Office: Regional Administration and Local Government PRSC - Poverty Reduction Support Credit PSIA - Poverty Social Impact Analysis REPOA - Research on Poverty Alleviation TRA - Tanzania Revenue Authority UAPP - Urban Authorities Partnership Programme UNDP - United Nations Development Programme USAID - United States Agency for International Development v Executive Summary The 2005 Tanzania Poverty and Social Impact Analysis (PSIA) on local government tax reform was designed to examine the intended and unintended consequences on poverty reduction and growth in Tanzania of the tax reforms implemented in June 2003 and 2004. The main elements of the reform were the abolition of the flat rate development levy in 2003 along with "nuisance taxes," and the abolition of business license fees for enterprises below a certain size--and capping of those fees for larger enterprises--in 2004. This PSIA had two principal aims: (a) To assess the distribution of the tax burden across different social and income groups and small businesses before and after the reforms (b) To inform other initiatives directed at fiscal policy reform in the context of Tanzania's decentralization This was a rapid study undertaken on a limited budget. Because Tanzania is undergoing such rapid urbanization, and because the base for future taxation is largely urban, the study paid particular attention to urban areas. Research was undertaken in just three councils: Ilala Municipal Council (MC) in Dar es Salaam (urban), Moshi town (urban), and Kilosa (rural). In each council, stratified sample surveys were conducted of better-off, middle income, and poorer households, and on medium-size, small, and micro- businesses. Altogether 270 households were interviewed (90 in each location, and, within each location, 30 in each well-being category). In addition, 58 business owners were interviewed, selected to represent three enterprise size categories. Questionnaires focused on taxes paid (classified by category) by households and businesses before and after the reform, as well as on attitudes toward the tax reform itself, and to taxation policy more widely. In addition, the views of key informants from local government and the business community were sought in some depth regarding the way forward for local government taxation in Tanzania. While overall sample size was small and is not claimed to be representative of the entire country, the basic methodology was sound and the data collection phase closely managed, such that quality of the data can be considered very high. The districts and councils studied reveal much about the course and impact of the reforms and stakeholder attitudes toward them that are likely to apply more widely in Tanzania. It is worth being reminded that ever since Adam Smith's Wealth of Nations, economists have shared an understanding of the basic principles of a good tax system. Such a system should avoid distorting relative prices (by imposing different tax rates on different items), avoid inhibiting investment and enterprise (by taxing start-up businesses out of existence), ensure that collection costs are as small a proportion of the tax collected as feasible, be perceived as fair by society, be transparent to the tax-paying citizenry with respect to amounts collected and the uses to which revenues are put, and emphasize the links between tax paid and services delivered. The pre-reform local taxation system in Tanzania could be interpreted as violating many of these principles. Variable market fees and dues distorted relative prices, small start-up businesses were taxed arbitrarily, collection costs were high relative to amounts collected, taxes were patently not fair (the flat rate development levy was self-evidently regressive), there was little transparency regarding amounts collected and disbursed, and citizens were unable to perceive links between the public services they received (or failed to receive) and the majority of taxes that they paid. These features have been extensively documented in previous literature as well as in recent reports by international agencies vi undertaken in the context of the Local Government Reform Programme (LGRP) and Local Government Support Project (LGSP). Multiple and complex taxes are much more likely to violate the basic tax principles outlined above than single and simple ones, and should be avoided where there are feasible alternatives. This is because neither the taxing authority nor the citizen is able to gauge the ramifications and impacts of complex tax systems--their incidence, their fairness or otherwise, their effects on prices and enterprise, the degree of leakage in their collection, and so on. Moreover, clearly the more complex the tax system, the higher is the collection costs per unit of tax collected. Disapproval by ordinary citizens and businesses of the complexity of tax systems emerged clearly from this study, as it has from other investigations that have elicited people's views on the matter. For these reasons, the tax reforms would seem a move in the right direction for equity in local government revenue collection policy in Tanzania, irrespective of some uncertainties that subsequently occurred--concerning the compensating payments to councils from the central government, and the steps that councils were permitted to take themselves to develop new revenue alternatives. The extent to which service delivery was compromised by the reform is debatable, especially for rural councils where numerous studies have shown that most locally-generated revenue was spent on administration and councilors' sitting allowances rather than on service delivery. In any event, more equitable taxes are not necessarily the sole objective. Taxes must also be stable, predictable, and adequate to the mandates they are intended to fund. Future work will be needed to reconcile these objectives, and synthesize a local taxation policy that accommodates both equity and the other requisites of a sound local tax regime. The PSIA must be placed in a context of the commitment of the Tanzania government to decentralized local government, and to fiscal decentralization within that commitment. Local councils are responsible for the delivery of public services--primary education, primary health, local roads, potable water, sanitation, agricultural extension--the bulk of whose funding is provided by central government, as also are the personal emoluments of council civil servants. According to this pattern, locally generated revenues are largely applied to the operational costs of councils and the sitting allowances of elected councilors. If sufficient, locally generated revenue could also be used to fund local service priorities that are not covered by central grants; however, administrative and councilor costs are bound to have first call on these revenues, since there is no other source of funds for these purposes, and otherwise councils would be unable to discharge their duties with respect to their primary service delivery responsibilities. Before summarizing the findings of the sample surveys of household and businesses some cautionary remarks about the clarity and interpretation of data would be in order. These apply particularly to more disaggregated levels of analysis, by councils, or social groups, or tax types. First, the reforms took place in two waves, over a year apart: June 2003 (when the flat rate development levy and a range of nuisance taxes were abolished) and July 2004 (when business licenses below a threshold turnover were abolished, and above the threshold reduced to a flat rate license fee). The surveys were conducted in early 2005, implying that pre-reform recall for households and businesses refers to 2002, while post-reform experience refers to 2004 during which significant further changes were taking place. In such a situation, it is inevitable that respondents will experience some difficulties in recall. Second, there are overlaps in revenue jurisdictions over certain types of fees or dues (for example, plying and parking fees recorded by respondents in Ilala MC are collected by Dar es Salaam City Council (CC), not by the MC). The impacts of the reforms varied between groups, but were broadly progressive. On average, households recorded a very small decrease (0.8 percent) in overall tax payments. Nevertheless, the impact of the reforms across social groups was steeply progressive. Total taxes paid by the better-off increased by 14 percent, decreased by 2 percent for the middle well-being category, and decreased by 34 percent for the poorest. Businesses recorded a 14 percent decrease in tax burden overall. Within this, vii medium businesses recorded 11 percent less tax, small businesses 36 percent less tax, and micro- businesses (under Tsh 54,000 turnover) 11 percent more tax. The increased payment by micro-businesses (an exception to the otherwise progressive trend) probably results from their nonpayment of previous business license fees, coupled with the wider use of other taxes (such as billboard fees) by councils after the reform: these were all imposed on micro businesses with greater vigor than were the defunct license fees. The reforms' progressive impact on households and their reduction in the tax burden of businesses indicate broadly desirable outcomes for equity and growth. By council, residents of Kilosa District Council (DC) seem to have benefited most from the abolition of the development levy and "nuisance" taxes (a reduction of 28 percent in total taxes paid). Ilala MC residents bore steep rises in plying and parking fees paid to Dar es Salaam CC, though these were not directly related to the tax reforms per se, and residents of Moshi MC recorded rises in property taxes paid to the town council. All councils were enterprising in replacing income lost from the development levy and market dues by intensifying collection of taxes that remained on the permitted schedule and were scarcely used before. These include parking, billboard, and bus-stand fees. There is a danger that this strategy of imaginative diversification will reproduce the previous multiplicity of taxes with a new array of minor (though permissible) "nuisance" taxes. While councils are required to seek Ministry of Finance permission before implementing new taxes, this applies only to new sources of revenue not envisaged in previous or current local tax law. The study explored the feasibility and acceptability of making greater use of property taxes as a potential way forward for councils to avoid tax proliferation, by rolling up all local taxes into a single property- based tax regime. Household respondents were asked whether they thought a tax on residential property was a good idea. The results seemed to provide quite a powerful mandate for residential property taxes, with two-thirds of the respondents in favor and less than one-third against (about 5 percent had no preference). Business respondents were 62 percent in favor of a property tax, and discussion showed these individuals to be even more enthusiastic if this tax were combined with the abolition of other fees and dues that they currently pay. Other findings were consistent with this. Combined, property taxes and land rents were the first preference of 61 percent of households for taxes to be retained, the next highest preference being business taxes at 15 percent. In addition, 66 percent of households and 71 percent of businesspersons stated that they thought wealthier people should pay more tax for local services than poorer people. The study thus reveals that a progressive property tax would be popular among Tanzania's aware of the need to pay taxes in order to obtain improved services, and would like that link to be further strengthened. The privatized garbage collection fee in Ilala MC exemplifies this awareness, with citizens voicing clear opinions simultaneously about both the necessity for such a fee and the performance that they expected in return. To summarize the argument thus far: the 2003 and 2004 tax reforms moved Tanzania's local government citizens and business folk, with the proviso that this would need to be decisively put forward as a substitute for, rather than an addition to, multiple other taxes and dues and levies. There were moderately encouraging replies to a wide variety of questions concerning the tax reform, service delivery, governance, and citizens' attitudes to local government. The abolitions of the development levy and nuisance taxes were manifestly popular in rural areas. Respondents in Kilosa district were especially full of praise ("We would like to thank the government" was a common response). Citizens' views were that by and large taxes were collected fairly, that receipts for taxes were issued, and that unofficial payments were rare in their direct experience (though they were more skeptical in their viii views about corruption levels more generally). Citizens are taxation system toward greater equity. At the same time, the mode of implementation of the reforms disempowered and frustrated local councils, whose officials subsequently felt obliged to resort to the remaining list of permissible taxes in order to restore their revenues. For rural councils, filling the gap left by the abolished taxes is more problematic than for urban areas, which have a broader tax base to exploit. Compensation of districts by the center may not solve this problem, both because it cannot necessarily be relied upon indefinitely and because it deepens the councils' dependency on central government, undermining the sprit of decentralization. Moreover, the reforms may have adversely impacted other important taxation objectives, such as stability, adequacy, and predictability. They may have also weakened the social contract at the local level. The reforms were progressive in their income distribution effects, and citizens were found to be broadly in favor of progressive taxes and property taxes in particular, as well as emphasizing the link between tax paid and services delivered. Citizens took a forward-looking view of tax collection as potentially leading to improving services (a finding confirming that of other studies). Citizens' expressed service priorities, in all three study locations, were, in order of priority, potable water supply, health, education, garbage collection, and roads. The PSIA suggests or reaffirms tax policy directions that would strengthen the legitimacy of local councils, improve the links between tax collection and service delivery, obey fundamental principles of good taxation, and be acceptable and fair to the citizenry at large. Some key policy options for consideration include: (a) a simplification and consolidation of currently fragmented local taxes (b) the possibility of a reformed business tax (c) the amplified use of property tax; property valuation and/or updates hereof and improvements in property tax coverage and collection (d) exploration of combining property tax and land tax regimes (e) enhanced service delivery links (f) further expansion of user fees (g) a concerted effort among national and international partners to develop a strategic approach in regard to the expected amount of local own-source revenues, the purposes to which these should be put, and the relationship to central government funding More systematic modeling will be needed to assess the extent to which new tax options--that is, (a) to (c) above--can substitute for the abolished taxes as source of revenue, taking into account the differential impact between urban and rural councils. Regarding user fees, although these emerge as an effective means of recovering operating costs, it is not yet clear whether they are a sufficient mechanism to widen access to goods and services, especially for the poor. This PSIA does not endorse any of these options. Rather, it suggests the need to develop a clearer strategic consensus on the respective fiscal and developmental roles of local and central government within the context of decentralization. In particular, greater clarity would be beneficial with respect to the following: (a) objectives of local revenue generation (b) respective responsibilities of central and local government for service provision (c) levels of central government support to local governments (including equalization policies and criteria) ix (d) the amount required to be collected per capita to achieve those objectives (that is, the target amount) (e) the most efficient methods for realizing the target amount in terms of unit costs of collection (f) ensuring a progressive income distribution effect of tax collection so that the poor are taxed less than the better off, and the extreme poor--those who are prone to requiring safety net support in bad years--are excluded from the tax net altogether Addressing important strategic fiscal decentralization issues will clearly require an overall set of principles that can then be varied for councils of particular sizes and types in terms of the economic activities of their citizens, their infrastructure, and the mix of central and local government funding of public services that is appropriate to them. Georgia State University (GSU) researchers are reviewing the policy framework underpinning the structure of the financing of local government and local government revenues. In addition, future work is to be undertaken with World Bank support to identify viable sources of revenue and to model the impact of various scenarios. This work includes modeling local taxation options: a range of simple models of households and businesses may be used to explore various scenarios, such as the imposition of residential and business property tax at different levels, and assess the impact of each. Drawing on income and asset information, where available, would enable the income distribution implications in each model to be assessed. The information generated by such simple models could inform central and local government decision makers regarding the implications of the options, as well as to engage the public in the debate, illustrating what the implications for tax assessments of each option would be (which may, if spread widely and progressively, turn out to be relatively modest). There are a number of other data and analytic tasks that could inform policy dialogue as it progresses: · Distributional data on current local tax impacts at the top of the income scale and for the largest firms. Such data seem critical to productive dialogue about future local tax policy, especially with the Ministry of Finance and the Tax Reform Task Force. · Further comparison of the profile of receipts reported by Local Government Authorities (LGAs) with the reported profile of payments by taxpayers. This would indicate the degree and location of any leakage of funds. · Gathering data from additional LGAs to remedy the small sample size of the present study. · Research on the aggregate impact burden of taxes (and other revenue-raising measures), up and down the income and firm-size scale, including both local and national taxes and revenue-raising measures. This list is merely an indicative one; more important than the themes of the research themselves will be the emergence from a process of policy dialogue of demands for analytic work that can contribute to the consultative development of a simple, fair, transparent, equitable, and sustainable tax system. This policy dialogue would need to bring together central and local government, representatives of businesses and citizens, of taxpayers and service users, as well as international development partners. Drawing on such wide participation, it would likely best be facilitated by the organizations charged with taking forward local government frameworks (that is, LGRP and LGSP). Since there is considerable overlap between the property/business-owning class and that of political decision makers, debate will x need to broach frankly the divergent vested interests served by the present system--including the political and social barriers to greater emphasis on property taxation, and to a more legitimate business tax--as well as the transparency of the revenue collection system. Policy options will need to be championed from within relevant institutions, rather than parachuted into the policy discussion from the outside. Accordingly, the best way forward may be to work with a small number of councils that would provide detailed case studies in the first instance, ensuring a more general approach to creating an equitable and sustainable revenue system for local governments. xi Chapter 1: Introduction 1 OBJECTIVES OF THE PSIA Poverty and Social Impact Analysis (PSIA) is the analysis of the distributional impavct of policy reforms on the well-being or welfare of different stakeholder groups, with particular focus on the poor and vulnerable. PSIA has an important role in the elaboration and implementation of poverty reduction strategies in developing countries. It promotes evidence-based policy reform options (World Bank, 2003).3 PSIAs typically seek to examine the assumptions underlying a specific reform, evaluate the potential poverty and social impact, assess the short- and the long-term aspects of reform, provide policy options that maximize the benefit of a reform, and involve a broad range of stakeholders in a debate of the reform being considered. PSIA is often applied to contentious or sensitive reforms, or to reforms expected to have a large impact. The main objectives of this PSIA on local tax reform in Tanzania were to: (a) assess the distribution of the tax burden across different social and income groups and small businesses before and after the 2003 and 2004 reforms (b) inform other initiatives directed at fiscal policy reform in the context of Tanzania's4 The findings of this study are intended to support the development of fiscal policy for decentralization, and in particular contribute to the implementation of the Local Government Support Programme (LGSP), as well as to the monitoring and design (respectively) of the Second and Third Poverty Reduction Support Credits (PRSCs). This study is particularly timely given current interest and debates on new initiatives for local revenue generation in the context of decentralization in Tanzania. Two longer-term ongoing studies are relevant. First, a four-year program (2002­05) by Research on Poverty Alleviation (REPOA), the Chr. Michelsen Institute (CMI), and the Norwegian Institute for Urban and Regional Research (NIBR) is examining the governance, financial, managerial, and service delivery aspects of local government. Second, a study by Georgia State University (GSU) with REPOA for the Local Government Reform Programme (LGRP) is undertaking a "comprehensive review of the policy framework underpinning the structure of the financing of Local Government and Local Government Revenues." This PSIA complements these longer-term studies by providing a detailed comparative illustration of potential winners and losers from the recent reforms, focused on actual tax collection practices, their social impact, and the distribution of the tax burden on different social and income groups in the sample Local Government Authorities (LGAs). 3The "PSIA User Guide," available on the World Bank's Web site, provides further in-depth information. 4 The initial proposal for the PSIA included two other objectives: assessing the reforms' impact on local government revenues and expenditures on service delivery and maintenance of infrastructure; and proposing viable options for change toward a buoyant and equitable system of local taxation. During the design phase, it became clear that, given the limited availability of reliable data, especially at the district level, and the financial resources available for primary data collection, these two additional objectives were not feasible within the scope of the study. Specifically, the impact on local government revenues was hard to trace accurately, mainly because (a) the reforms were in two waves across two financial years, with the PSIA started before the second of those waves had ended, and (b) own tax revenue returns by councils were not up-to-date and contained gross inaccuracies for particular revenue flows. 1 A final introductory word on the PSIA approach may be appropriate. PSIA aims to establish not simply a static picture of policy impacts on the poor, but rather to understand the dynamics whereby future new policies may improve livelihood opportunities for this category of the population. One justification for the reform measures, in addition to their manifest political return, is the supposed negative impact of certain types of local taxation on decisions to invest and produce, and on the poor and vulnerable. This view implies that the objective of raising healthy local revenues is necessarily in tension with economic development and poverty alleviation, and that a trade-off must be reached between the two. This may be phrased as a balance between, on the one hand, retaining money in the pockets of the poor, and on the other, finding the resources for local governments to deliver basic services that benefit the poor as much or more than others. But these goals are not necessarily inconsistent. A healthy, equitable system of local taxation should be able to provide adequate revenues without stunting economic development or worsening poverty. Such local fiscal viability has a number of essential attributes: (a) Taxes should not distort relative prices (as market dues tended to do in the past), causing inefficient production decisions to be made. (b) Taxes should not create barriers to investment and to increasing productive capacity, for example, by taxing start-up small businesses out of existence. (c) Taxes should be broadly considered across the communities that pay them to be fair and this typically means lessening the burden on those below the poverty line (or removing them from the tax net altogether), and relating tax at least partly to ability to pay (in other words, a desirable local government tax system should be progressive in character). (d) To these considerations must be added standard "principles of good taxation," namely: · Taxes should be simple to understand and low cost to administer. · Cost of collection should be a small proportion of total tax collected. · There should be transparency in the total tax that has been collected, and its allocation (for example, through posting budgets and expenditures outside town halls or in village community centers). · There should be visible links between taxes collected and services delivered (for certain services such as water supply, garbage collection, this is relatively easy to achieve; for general taxes such as property tax, the link must be made by publicizing the tax's intended uses and the results achieved). In reality, the connection between taxes paid and the provision of tangible services, and the social contract implied by it, may be broken by a number of factors. These include: (a) the application of locally generated revenues for purposes that do not lead directly (or perhaps even indirectly) to the provision of visible services (for instance, an excessive proportion of income spent on administrative costs and perquisites) (b) leakage of funds occurring either before or after these enter official accounts (c) excessive cross-subsidization from wealthy to poor areas, so that the wealthy see no benefits in paying local taxes and develop resistance or avoidance strategies 2 with respect to them (this is not an argument against cross-subsidization, just a caution that the wealthy also need to see services improved in return for their taxes) Where a social contract linking taxation to public service delivery has not been established in practice, as when a political culture of patrimonialism predominates, simply increasing the effectiveness of tax collection may not result in more efficient services. In fact the intensification of tax collection efforts may lead to the further alienation, rather than the involvement, of citizens (Fjeldstad 2001a, 2001b). The points discussed above emphasize the importance of ascertaining who pays and how much (in other words, fiscal incidence by socioeconomic group), before and after the reforms, as well as for any reforms proposed for the future. However, it is clearly equally important to explore how funds collected locally are spent, and to understand the wider context of governance within which taxation takes place, including the transparency of the processes of tax collection and expenditure, the channels of information available to citizens, and the attitudes of taxpayers toward local government. 2 BACKGROUND TO LOCAL TAX REFORM The early history of local government in Tanzania is not untypical of the region, with the pragmatic development of decentralized administration under the colonial rule. The new directions introduced by President Nyerere saw the rapid replacement of this system with a top-down, centralized structure based on regional administrations. In subsequent years, resource constraints led to a deterioration of local service provision. The de-concentrated system of local government that replaced this was enshrined in the Local Government Finance Act of 1982 (LGFA). This allowed district-level LGAs to establish their own sources of revenue, which varied from council to council, in competition with stronger regional administrations (Mawhood 1983; cf. Steffensen, Tidemond, and Mwaimpopo 2004). By the mid-1990s the shortcomings of this mixed de-concentrated system were so apparent that the LGRP was launched, aiming to reorganize local government based on more devolutionary principles over the period 1998­2011. Central to this was a commitment to fiscal decentralization. It may be noted that one consequence of this accretionary approach to local government reform is that "over time... the Acts contain a series of repetitive, duplicative and in some cases contradictory clauses" (Government of Tanzania 2004c). There was no systematic reporting system on local government budgets; widespread shortcomings characterized the range, type, and administration of local government taxes (Kobb 2001a­d; Fjeldstad 2001b, 2003), many of which were deservedly extremely unpopular. While the share of total revenues for LGAs of own revenues collected was only 3­5 percent in rural councils, and typically in the range of 10­20 percent of all revenues in urban councils (which have a much higher local revenue base), these taxes made a heavy impact on taxpayers. The development levy, a major source of revenue, was a highly regressive head tax which led to disruptive evasive behavior, and contributed to even greater civil disturbance (Kelsall 2000); a range of cesses and fees could lead to multiple taxation; taxes were numerous, fragmented, and in many instances cost ineffective (Fjeldstad et al. 2004, 15). Tax compliance was extremely low (Kobb 2003), and accusations of corruption abounded (Fjeldstad 2002b), with no obvious link between taxation and service delivery (Fjeldstad and Semboja 2001; Fjeldstad and Nygaard 2003). As an illustration of just how weak control and accountability of funds in local councils are likely to be in practice, reference may be made to the Public Expenditure Tracking System (PETS) for education (cited by Sundet 2005, 9) in which it was found that only 54 percent of centrally delivered funds based on a per student funding of US$10 reached the schools. 3 In June 2003, in response to concerns over the multitude of local taxes, the Government of Tanzania eliminated a range of local government taxes, the most important being the development levy. The crop cess was capped at 5 percent of farm gate price, multiple taxation was in principle banned, and a range of nuisance taxes scrapped. In the following fiscal year, further measures were announced to improve the business environment. These abolished the business license fee for health centers operated by religious organizations, and also for businesses with turnover under Tsh20 million per year and was capped at a maximum tax take of Tsh20,000 above this threshold. Government agreed to compensate any resulting revenue shortfall in local government revenues with grants from the equalization fund. The impact of these reforms on local governments' own revenue sources was considerable, with income from the development levy falling from 19.4 to 6.6 percent between 2002 and 2003 (table 1). For lower local government levels--at the ward and mitaa level--the consequences were a 90­100 percent loss of local revenue by late 2003, according to a study on 28 Lower Levels of Government (LLG) (PO-RALG 2003, cited in Government of Tanzania 2004a). Generally, the share of own revenue in the total resources available to LGAs declined by a third (table 2), while, as table 3 shows for our sample councils, government compensation for these loses was substantial. The reforms affected rural and urban councils in different ways, resulting in a further deepening of the fiscal disadvantages rural areas faced. While a rural council might rely on the development levy for 40 percent or more of own revenue, an urban council only received 10 percent from this source; in 2001 and 2002 the wealthiest LGAs were able to collect 40 times the own revenue per head of population as the poorest, and in 2003 this ratio increased to 100 (Government of Tanzania 2004c, 32). The reduction in business licenses and fees in 2004--a far more significant revenue generator in the urban than in the rural areas--is unlikely to have changed these ratios greatly, given the greater flexibility of urban councils in finding new sources of revenue. A number of questions thus arise as to the impact of the 2003­04 reforms. How have councils--and LGAs and LLGs--differentially been affected? Has central government compensation been effective and with what consequences? What have been the consequences for the loss of revenue on taxpayers, including households and businesses? Are the poor net gainers or losers in light of adjustment strategies adopted by councils? In setting out to address these questions, this PSIA provides a vignette from 3 out of Tanzania's 114 councils, and thus has no pretension to cover the whole national spectrum of change and response. However, by examining the issue in both quantitative and qualitative detail in three districts, the study can provide important insights into the impact and perceptions of reform at the national level.5 The changing social and economic circumstances of Tanzania provide a significant backdrop to this study. As discussed in chapter 2 on methodology, there was a purposive emphasis on urban locations in the study. This reflects both the important demographic shifts that have been occurring in Tanzania since the 1980s and the importance of urban areas as potential tax bases in the future. For these reasons, two urban councils and one rural council were sampled. 3 SOCIO-ECONOMIC CONTEXT Tanzania is classically thought of as a mainly rural society, with agriculture as its predominant sector. Despite the low world prices for its major estate crops, this sector still dominates the economy, 5According to the study, there seems to have been no uniform compensation system, but rather ad hoc arrangements where some councils were able to put pressure on government, for instance, urban district councils in general found their own ways of retrieving their revenue position, while rural councils unevenly obtained some transitional resources during the reforms and the postreform period. 4 responsible in 2002 for 45 percent of total production, 75 percent of exports, and 80 percent of employment (World Bank 2002b, figures 2 and 3). Yet perhaps the most significant macro aspect of Tanzania's society is that the country is experiencing urbanization at a rate that is among the highest in the world: the urban population grew from 14.8 percent in 1980 to 34.2 percent in 2002. With urbanization proceeding in secondary towns at twice the 3 percent national population growth rate, and in Dar es Salaam at three times this rate, the current population share of urban areas is probably about 40 percent. Sixty-five percent of Dar es Salaam's population consists of new immigrants, and 70 percent live in unplanned settlements (World Bank 2002a). The rural sector is profoundly affected by these changes, as the financial transactions of rural producers are deeply embedded in their social relationships with small traders operating out of secondary and tertiary urban nodes (Bryceson 1999; Bah et al. 2003). The fiscal consequences of this shift are major. While the average council in Tanzania may raise only 7 percent of its own revenue from property taxes, secondary urban councils raise two to three times this amount, and Dar es Salaam as a whole may be approaching the Nairobi level of 60 percent. This helps to explain why the per capita tax collection of urban councils is some four times that of rural councils (Government of Tanzania 2004c). Although Government institutions and a number of other institutional bodies are exempted from property taxes, the potential revenue for urban district councils is even greater. These developments bear centrally on the debate about the fiscal dimension of decentralization. Table 1: Consolidated LGA Own Revenue Collection Data (Tshm/-) % change % change Tax Type 2001 % share 2002 % share 2001­02 2003 % share 2002­03 Development levy 11.0 21.5 11.2 19.4 1.8 3.2 6.6 -71.4 Property tax 3.2 6.2 3.5 6.1 9.4 3.2 6.6 -8.6 Agricultural cess / livestock levy 9.3 18.2 9.2 15.9 -1.1 9.0 18.6 -2.2 Ind. cess/service levy 5.2 10.2 9.1 15.7 75.0 7.7 16 -15.4 Land rent share 0.6 1.2 0.7 1.3 16.7 0.8 1.6 14.3 Licenses and fees 10.2 19.8 11.1 19.1 8.8 12.0 24.9 -8.1 Charges 3.1 6.1 4.0 7 29.7 5.0 10.4 25.0 Other taxes and levies 8.6 16.8 8.9 15.4 3.5 7.4 15.3 -16.9 TOTAL: 51.2 100.0 57.8 100.0 12.8 48.2 100.0 -16.6 Source: Derived from Government of Tanzania (2004c, table 3.1) Table 2: LGA Financial Resource Flows, 2001­02 to 2003­04 (Tshm/-) Revenue source 2001­02 % share 2002­03 % share 2003­04 % share Central government grants 201.1 78.9 247.0 80.5 291.0 85.2 Own revenue 53.9 21.1 59.9 19.5 50.4 14.8 Total 255.0 100.0 306.9 100.0 341.4 100.0 Source: Derived from GoT Fiscal Review (2004, Table 1.1) 5 Table 3: Central Government Grants to LGAs, 1999 to 2005 (Tshm/-) 1999­ 2000­ % 2001­ % 2002­ % 2003­ % 2004­ % 2000 01 change 02 change 03 Change 04 change 05 change Moshi MC 904.6 1243.9 37.5 2629.9 111.4 1818.7 -30.8 2012.8 10.7 2469.8 22.7 Kilosa DC 1665.6 2537.7 52.4 2673.7 5.4 3385.1 26.6 3863.7 14.1 5095.4 31.9 Ilala MC -- 3141.4 -- 3393 8.0 4618.7 36.1 5221.3 13.0 6337.5 21.4 Source: Approved Budget Estimates. --. Not available. 4 THE DECENTRALIZATION DEBATE This PSIA study took place against the background of an ongoing debate on the merits and demerits of decentralization in African states in general (Cross and Kutengule 2005; Adamolekun 1999), and in particular the role of fiscal decentralization which lies at the heart of it. This is not the place to rehearse these arguments in any detail, but some key points should be mentioned. The raising of revenue locally is generally considered an essential element in local development, not only because this contributes to available public resources at that level, but also because it promotes public ownership, accountability, and good governance. Citizens, it is argued, will be more likely to hold their local government to account for its performance if it is spending their own tax money (Rakner and Golpen 2003). Much of this argument requires clear links to be established in citizens' minds between taxes collected and the quantity and quality of public services delivered. From this viewpoint, the move toward central grant financing entailed in the 2003 reforms could have detrimental long-term consequences by increasing local government dependency on intergovernmental transfers from the center. The loss of own revenue undermines the ability of local governments to meet their responsibilities for local service delivery, weakens the direct link between taxpayer and local government, and therefore hinders meaningful decentralization and local governance. The consequences could be either service cutbacks or the shifting of the tax burden from the local to the national or international (that is, donor-supported) levels. A counter-argument to this is the pragmatic one that most LGAs in Tanzania lack the capacity to raise and spend revenue effectively and rationally, and it is only in areas where the rich and articulate are located and where civil society organizations are at their densest that the perspective on the links among democracy, accountability, and fiscal decentralization hold weight. According to this perspective, the task is to make LGAs accountable for expenditure rather than for generating revenue, and to build the necessary capacity to administer local systems of tax collection, including undertaking cadastral surveys, record keeping, and so on. Allied to this is the observation that for local government the power to demand taxes may be seen as politically functional--in the sense that it grants status to the officials concerned--rather than administratively rational or economically functional. The evidence gathered by this study counters these views in that it reveals citizens' voices to be in favor of rational, progressive taxes that are simplified and consolidated. Citizens understand the link between tax payments and service delivery, and have strong views on the modalities of both of these elements (chapter 5). It must also be noted that the service delivery areas concerning which the strongest views were expressed (for example, education) remain central competencies controlled by dominant sectoral ministries (as is commonly the case in decentralizing states in Sub-Saharan Africa). Defining an appropriate and effective space in these sectors where LGAs can assert their profile and legitimacy and local taxpayers channel their voice, is a challenging policy issue. 6 The reforms have had a clearly differential impact on different LGAs. The effect has been toward a strengthening of the implicit "social contract" between government and society that tax-paying may be said to represent in urban areas (Moore and Rakner 2002; Brautigam 2002), whereas that in rural areas may be weakening. Given the widely varying levels of both tax base and administrative capacity in Tanzanian LGAs (and LLGs), close attention must be given to matching the degree of autonomy to the capacity of the unit in question. 7 Chapter 2: Research Instruments And Methodology 1 SITE SELECTION Fieldwork was conducted in two urban council areas (Ilala MC, Dar es Salaam; Moshi MC, Kilimanjaro Region) and one rural council area (Kilosa DC, Morogoro Region). The choice of these councils reflected the urban emphasis of the study, the availability of substantial empirical evidence on local taxation, and logistical considerations (table 4). In Kilosa district, research on rural households conducted in 2001 by the LADDER project6 has provided information on the activities and incomes of households and firms and businesses against which tax types and rates of various kinds can be compared (Ellis and Mdoe 2002; James, Mdoa, and Mishili 2002). Table 4: Criteria and Justifications for Selection of Field Study Locations City locations In favor Not in favor Comment Ilala MC Best data in ongoing REPOA/CMI study; slices through city with valuable property base atypical cityscape includes Temeke MC Good mix of residential, commercial, and middle to poor peri-urban poorest residents residential poor Kinondoni MC Very proactive municipal director socioeconomic affluent to middling mix residential Other urban locations Arusha CC Study by Kelsall of tax revolt (2000) very poor performance Moshi CC Included in ongoing REPOA/CMI study; active local economy, good performers on transparency accessible Iringa CC Good performance and data low-income small town low accessibility Rural locations Included in ongoing REPOA/CMI study Kilosa DC and LADDER; good management and part of REPOA study data Morogoro Rural DC LADDER site The components and instruments of the field research varied somewhat between urban and rural areas, although much of the methodology was common across all fieldwork locations. The main components of fieldwork were: · secondary data collection, council level · key informant interviews, council level · household surveys, urban and rural councils · small business and marketplace surveys, urban and rural councils 6A study of livelihoods diversification undertaken in Tanzania, Uganda, and Kenya by the University of East Anglia School of Development Studies with DFID support. 8 2 SECONDARY DATA COLLECTION (COUNCIL LEVEL) Lists of tax rates, business taxes, license fees, and qualitative data on how taxes are collected in practice are available for the pre-reform period from the REPOA/CMI studies conducted during 2001­02 to the extent that these have been published. The main task of the secondary data collection has been to obtain this information for the postreform period. It has not always proved possible to separate the 2003 and 2004 data, and it is likely that some informants tended to conflate these two years. Phased changes in policy such as in this case may make it difficult to avoid confusion among respondents. Data collected at the level of the council revenue office have comprised where possible: · revised lists of types of taxes that are applied, postreform (in 2004 and proposed for 2005) · revised lists of tax rates (flat rate taxes or percentage taxes) as applied in 2004 or proposed for 2005 · local revenue generated (total and broken down by tax categories) for 2003 and provisional data for 2004 · expenditure of local revenue, broken down by uses, for 2003 and provisional data for 2004 3 KEY INFORMANT INTERVIEWS (COUNCIL LEVEL) 3.1. The views of council officials, local councilors, and local business leaders on tax reform and related PSIA concerns were elicited using a structured set of questions, given separately to two council officials (one of them the council revenue officer), one council politician, and two local businessmen in different types of business. 4 HOUSEHOLD SURVEYS, URBAN AND RURAL COUNCILS This survey was designed to answer the central question of how much tax was and is paid by households of different socioeconomic levels before and after the tax reform. The household is thus the unit of investigation. To generate a representative sample, the survey was administered in three distinct areas (subdistricts) of each sample council: better-off, middle-income, and poorer subdistricts. A senior and knowledgeable council official was asked to advise on the appropriate locations. In each subdistrict, 30 households were randomly selected. During analysis, each of the 270 household questionnaires was allocated into one of the three categories above, by applying criteria to an assessment derived from a Household Asset Checklist (section 12) in the Household Survey Forms, as validated by council officials and knowledgeable local community leaders. 5 BUSINESS AND MARKET SURVEYS, URBAN AND RURAL COUNCILS The small and medium-size business and market survey was designed to discover how businesses and traders are currently affected by taxes and fees and permits and licenses, and how the situation has changed since the reforms. The survey targeted those medium-size, small and micro-businesses affected by taxes, fees, permits, and licenses according to selection criteria as advised by the council trade officer and the ward executive officer. The selection of markets was based on the market's revenue contribution to the council. Table 5 shows the sample sizes for each category of business. 9 Table 5: Sampling per Council or per Market Medium-size Small Micro enterprises enterprises Market traders Market suppliers and truckers Ilala 5 10 5 3 Moshi 5 10 5 3 Kilosa 0 10 5 3 10 Chapter 3: Distributional Impacts on Households 1 PROFILES OF SELECTED COUNCILS The three council areas in which the PSIA surveys were conducted vary considerably in area, demography, local socioeconomic systems, and notably in average local tax payments per household (table 6). Ilala MC covers a mixed use area of Dar es Salaam, ranging from government and business offices to small and medium-size enterprises, industry, export and import activities, petty retail and artisanal activities, and some agriculture and livestock in the peri-urban fringe. Moshi MC has limited but greater agriculture and livestock activities, with trade, industry, and tourism dominant. Both of these urban areas are rapidly growing and becoming more densely settled. Kilosa is a mid- to low-income rural area where livestock and crop production are predominant, with limited small industries (food and crop processing), forest products, and some intensive irrigation schemes (rice). Access to Morogoro town takes two to three hours on a gravel road. Table 6: Profiles of Sample Survey Councils Characteristic Ilala MC Moshi MC Kilosa DC Area (sq km) 210 58 14,245 No. of wards 22 15 37 Population 637,565 144,336 490,200 % male 49.5 49.2 49.8 % female 50.5 50.8 50.2 Household heads 148,383 35,598 105,748 Local tax per HH prior 65.0 130.1 20.0 Local tax per HH post 71.9 114.3 14.8 % change 10.6 -12.1 -26.0 Source: Derived from National Census (2002) and PSIA Household Surveys. HH. Household Over the three years prior to local tax reform, own revenue as a percentage of total revenue averaged 61.4 percent for Ilala MC, 10 percent for Moshi MC and 14.2 percent for Kilosa DC (Fjeldstad et al. 2004), indicating that of these only Ilala has any real level of fiscal autonomy. Own revenue for Kilosa was very largely absorbed in tax administration and council overhead expenditures (Fjeldstad and Semboja 2000; Ellis and Mdoe 2002). Figures for Moshi are not available, but with its low average level of own revenue the council is unlikely to be significantly different from Kilosa in this respect. Kilosa DC has benefited from a long partnership with Irish Aid, and has made significant improvement in communication with citizens, and would appear to have outperformed Ilala and Moshi MCs in terms of building relationships at the local level (Citizens Survey 2003, cited by Fjeldstad et al. 2004). Kilosa was also unusual in that citizens on balance felt taxes were being spent on service delivery, whereas some two-thirds of those interviewed in the urban councils did not. 2 CHARACTERISTICS OF SAMPLE A total of 270 households (3 x 3 x 30) were surveyed in three councils. Each council was zoned (either within or between wards) to ensure a balance between poor, average-income, and better-off households, based on indicators derived from key informants. The sample was purposively skewed toward urban areas. 11 Of the respondents, 55 percent were men and 45 percent women; 80 percent were household heads. Three-quarters owned their own houses, most of the remainder renting. Household size ranged from one to four individuals, averaging two persons. Eleven percent had no formal education, 13 percent were educated to Primary level 4, 48 to Primary 8, 17 percent to Secondary 4, and 3 percent had degrees or diplomas. A third lived in houses with four or more bedrooms, and two-thirds used electricity for lighting, although wood or charcoal was predominantly used (92 percent) for cooking. Fifty-three percent had access to piped water (including standpipes), but the majority (65 percent) used pit toilets. Asset ownership showed a wide spread: households with a computer (5 percent), a vehicle (9 percent), a telephone (16 percent), a bicycle (36 percent), a fridge (45 percent), a TV (50 percent), and radio (88 percent). Before summarizing the findings of the sample surveys of households and businesses some cautionary remarks about data accuracy would be pertinent. These apply particularly to more disaggregated levels of analysis, by councils, or social groups, or tax types. First, the reforms took place in two waves over a year apart: June 2003 (abolition of the flat rate development levy and a range of nuisance taxes) and July 2004 (abolition of business licenses below a threshold turnover, and a much reduced flat rate license fee above that level). The surveys were conducted in early 2005, implying that pre-reform recall for households and businesses refers to 2002, while post-reform experience refers to 2004 during which significant further changes were taking place. Second, there are overlaps in revenue jurisdictions over certain types of fees or dues (for example, plying and parking fees recorded by respondents in Ilala MC are collected by Dar es Salaam City Council, not by the Municipal Council). 3 IMPACT BY COUNCIL AND INCOME GROUP Total local government tax payments per household across the three sample councils showed a small, but statistically insignificant, decline of 0.8 percent after the reforms (table 7). The mean tax payment per household fell from Tsh 51,030 before the reforms to Tsh 50,622 per year (approximately US$45). When examined by tax category, the most marked change was the nearly 90 percent fall in development levy payments made by the sample households (Figure 1). Bicycle license fees, market fees, crop cesses, and business license fees also fell steeply in line with the tax categories that were abolished or drastically reduced in 2003 and 2004. Households in urban areas reported big rises in parking and plying fees (in Ilala MC, these are Dar es Salaam City Council charges), and in council rent payments. Table 7 also reveals that these newly intensified sources of revenue substantially compensated for the reduced revenue from taxes eliminated by the reforms. 12 Figure 1: Change in Taxes Paid by Socioeconomic Groups 15 10 14 5 0 -2 -5 -10 -15 -34 -20 -25 -30 -35 Better-Off Middle Income Poorer Table 7: Local Taxes Paid before and after the Tax Reforms by All Households N=270 Tax type Before 2003­04 After 2003­04 % change due to Tsh `000 Tsh `000 reform Development levy 548 65 -88 Property tax 2,326 2,742 18 Abattoir fees 3 3 0 Auction fees 0 0 0 Council rents 468 298 -36 Business license 4,440 2,403 -46 Advert fees 40 37 -8 Market fees 1,598 1,238 -22 Plying Fees 1,614 2,578 60 Parking fees 289 1,193 313 Livestock levy 0 0 0 Bicycle license 70 5 -94 Cattle market fee 2 0 -100 Local brew fee 31 182 485 Forest product fees 226 353 55.8 Crop cess 90 0 -100 Bus-stand fees 0 0 0 Land fee 312 314 0.8 Other 1,721 2,257 31.2 Total 13,778 13,668 -0.8 Average per household (Tsh) 51,030 50,622 13 When households are disaggregated into the three income groups (table 8) and by council (table 9) it becomes clear that there are significant differences in effects between urban and rural areas, and among socioeconomic groups. Generally, the reforms have been strongly progressive in impact. Local taxation (measured by mean taxes and other fees paid) increased by 14.3 percent for the better-off, decreased by 2.3 percent for the middle-income group, and decreased by 34.2 percent for the poorest. Table 8: Taxes Paid before and after the Tax Reforms: Households Stratified by Income N=270, Tsh High Medium Low Tax type Before After % Before After % Before After % change change change Development levy 1,985 88 -96 2,043 426 -79 2,027 22 -99 Property tax 15,414 19,606 +27 10,711 12,491 +17 3,007 3,232 +7 Abattoir fees 0 0 14 21 +50 11 0 -100 Auction fees 0 0 0 0 0 0 Council rents 706 1,735 +146 1,969 1,134 -42 1,752 830 -53 Business license 67,971 35,206 -49 14,319 8,106 -43 1,000 574 -43 Advert fees 1,088 1,088 0.0 21 0 -100 0 0 Market fees 10,588 10,588 0.0 2,609 2,621 +0.5 9,193 5,359 -42 Plying Fees 47,294 75,529 +60 0 0 0 0 Parking fees 8,471 30,024 +254 0 1,191 ++ 0 0 Livestock levy 0 0 0 0 0 0 Bicycle license 0 0 85 7 -91.8 612 37 -94 Cattle market fee 0 0 17 0 -100 0 0 Local brew cess 0 0 0 562 ++ 330 1,087 +230 Forest product fees 0 0 477 2,092 +339 1,686 603 -64 Crop cess 0 0 638 0 -100 0 0 Bus-stand fees 0 0 0 0 0 0 Land rents 794 1,088 37 1,490 1507 +1 782 676 -14 Other 7,059 9,559 35 8,338 11580 +39 3,179 3,094 -3 Total: 16,1370 184,511 14.3 42,731 41,738 -2.3 23,579 15,514 -34 Before the reforms, the better-off paid four-fifths of their taxes as business licenses, plying fees, and property taxes. A similar picture applied to the middle-income group, though with market fees standing in for plying fees. The main taxes levied on the lower-income group were market fees, property tax, and development levy. As a result of the reforms, the main increases for the better-off were higher council property rents (+146 percent), plying fees (+60 percent), and parking fees (+254 percent).7 Property tax (+27 percent) and land rents (+37 percent) also increased significantly (mainly in Moshi MC), while the biggest savings of the better-off have been on the business license (-49 percent). To the extent that councils have been able to substitute for the loss of some taxes by increasing others, these have impinged upon the better-off more than on middle-income or poor households. 7The last two dues are charges on commercial vehicle users and owners levied by Dar es Salaam City Council: no payments under these headings were recorded in Kilosa or Moshi urban councils. 14 Middle earners gained from a 79 percent reduction in development levy payments, lower business license payments (-43 percent), and the scrapping of bicycle licenses and the crop cess. However, they faced new parking fees and large increases in taxes on forest product sales (+339 percent) and local brewing. These latter two charges relate only to the taxpayers in Kilosa district. Property taxes also increased for middle- income households. The poor gained substantially from the abolition of the development levy (-99 percent), council rents (-53 percent), market fees (-42 percent), and business license payments (-43 percent). However, they incurred a substantial increase in the local brew cess (again, in Kilosa only), and a smaller increases in property taxes. By council, notable changes were the increases in parking and plying fees paid by residents of Ilala MC to Dar es Salaam City Council; big parking fee increases also in Moshi MC; rises in property taxes and land rents in Moshi MC; and rises in forest product and local brew fees in Kilosa district. Note that land rents are collected by councils on behalf of the Ministry of Lands and are partially retained or remitted to councils, except for the Dar es Salaam municipalities. Essentially from the household perspective, city councils and the rural poor were winners, but the urban better-off and rural councils were losers. Postreform tax payments by respondents in Ilala MC rose by 11 percent, those in Moshi fell by 7 percent, and those in Kilosa fell by 28 percent. These figures indicate that the major impacts in rural council areas were the scrapping of the development levy and the sharp reduction in market fees: in the case of Kilosa these represented some two-thirds of respondents' tax payments prior to reform. In Ilala, in contrast, citizens continued to pay diverse taxes (not all to Ilala MC) resulting in both higher payments and a progressive distributional impact. This differential impact between urban and rural councils has implications for the shifting social contract between state and citizen in different locations. PSIA survey returns based on detailed interviews with a sample of individual taxpayers can be compared, in the case of Moshi, with the district's figures on local revenue (table 10).8 For a variety of reasons-- sampling, stochastical, and the possibility of leakage--one would not expect to find exact correspondence between these.9 However, judging from the types of tax that have increased in the official figures, the two sources present a consistent picture of the burden of taxation being shifted onto those more likely to be able to pay. The main variations between the survey and official sources are for land rent (MC ­31.6 percent, PSIA +9.6 percent), market fees (MC ­23.7 percent, +9.8 percent), property tax (MC ­13.2 percent, PSIA +39.5 percent), and overall (MC ­0.7 percent, PSIA ­12.1 percent). 8Revenue figures are not available for Kilosa. Those for Ilala appeared to suffer from marked inconsistencies and compilation errors, and have therefore been omitted from this report. 9The PSIA analysis gives equal weight to the three income groups, which are likely not equally representative in the population. The Moshi MC revenue data refer to calendar years 2003 and 2004--there is a 9 percent difference between the MC's year 2003 figures when compared with the approved budget statement. The Government of Tanzania's fiscal review (2004c, box 1.2) considers a number of self-reported own revenue returns by LGAs to be of a "questionable nature." 15 Table 9: Taxes Paid by Sample HHs Before and After the Tax Reforms by Council N=270 Tsh `000 All councils Ilala Moshi Kilosa Type of Tax Before After % Before After % Before After % change change change Development levy 150 60 -60 188 3 -98.4 209 2 -99 Property tax 1,290 1,342 +4 1,052 1,468 39.5 1 0 -100 Abattoir fees 0 0 0 0 3 3 0 Auction fees 0 0 0 0 0 0 Council rents 285 174 -39 155 45 -71 0 0 Business license 2,187 1,308 -40 2,133 1,032 -52 62 54 -12 Advert fees 0 0 0 40 37 -7 0 0 0 Market fees 108 74 -31 394 432 +10 1,090 727 -33 Plying fees 1,608 2,568 +59 0 0 0 0 Parking fees 252 625 +148 36 564 +1,467 0 0 Livestock levy 0 0 0 0 0 0 Bicycle license 0 0 0 0 71 5 -93 Cattle market fee 2 0 -100 0 0 0 0 Local brew fee 0 0 0 0 31 181 +483 Forest product fees 0 0 0 0 226 352 +56 Crop cess 0 0 0 0 90 0 -100 Bus-stand fees 0 0 0 0 0 0 Land fee 214 217 +1 39 43 +10 1 0 -100 Other 3,063 3,785 +24 397 507 +27 147 75 -49 Total 9,159 10,153 +11 4,434 4,131 -7 1,931 1,399 -28 16 Table 10: Moshi Municipal Council Own Revenue Collection (Tsh) Tax type Jan.­Dec. 2003 Jan.­Dec. 2004 % change Development levy 15,343,295 0 -100 Service levy 145,721,560 196,745,218 35.0 Property tax 204,990,384 177,854,016 -13.2 Business license 138,606,027 16,244,734 -88.3 Local liquor license 1,193,000 891,000 -25.3 Intoxicating liquor license 12,620,000 10,051,000 -20.4 Road license 2,797,000 3,135,000 12.1 Hotel levy 3,293,106 3,974,362 20.7 Market dues and stalls rent 81,974,800 62,512,675 -23.7 Abattoir slaughter fees 14,962,520 21,649,000 44.7 Bus-stands 108,200,500 140,510,000 29.9 Application fee (intox. liquor) 926,000 714,000 -22.9 Application fee (local liquor) 166,000 63,000 -62.0 Application fee (business license) 2,312,000 1,306,000 -43.5 Advertising sign board 15,621,000 13,273,000 -15.0 Business inspection fee 2,078,000 805,000 -61.3 Refuse collection charges 15,062,800 19,391,000 28.7 Cesspit charges 28,162,500 22,763,000 -19.2 Vaccination charges 1,537,000 646,000 -58.0 Sales of plants and seeds 32,700 57,500 75.8 Building plan scrutiny fees 174,500 130,000 -25.5 House rent 6,133,036 3,315,646 -45.9 20 percent retention of land rent 8,967,000 6,129,000 -31.6 Miscellaneous collection 81,473,403 183,656,059 125.4 Total 892,348,131 885,816,210 -0.7 17 Chapter 4: Distributional Impacts on Businesses 1 SAMPLE CHARACTERISTICS Fifty-eight businesses were surveyed in the three council areas. Table 11 gives some of the basic characteristics of the sample. Most (69 percent) of the businesses were operated by their owners, most of whom (74 percent) were men (mean age 34). Petty retail was the most common type of enterprise, at 48 percent, with the next most common activity being hotels, restaurants, and bars at 19 percent. 2 IMPACTS In terms of the overall change as a result of the tax reform, the aggregate change for respondents was a 14 percent reduction in taxes (table 12). Common to respondents in all councils was the scrapping of the development levy, a large reduction in business license fees paid (-49 percent), and the effect of abolition of a range of minor taxes (with some variation as to what these were) of -27 percent. Offsetting these tax declines were rises overall in property taxes paid (+28 percent) and in advertising (billboard) fees (+24 percent). When taxation outcomes are disaggregated by level of turnover (figure 2 and table 13), micro-business (turnover less than 54,000/-) shows an 11 percent increase, small business (54,000/- to 156,600/-) a decrease of 36 percent, and medium business (>156,000/-) a decrease of 11 percent. Micro-businesses seem to have been caught by rises in previously minor taxes that took from them more revenue than the savings they enjoyed from the reduction in the business license. Small businesses in the turnover range of 54,000­156,600/- did best: in the PSIA sample they experienced an overall reduction in minor fees and taxes of 24 percent, and they made significant savings from the reduction in the business license fee (-64 percent) and market fees (­20 percent). The large savings of medium-sized businesses on the business license and market fees were offset by higher payments for parking and plying fees, and a 34 percent rise in property taxes (though the last still represented only a small proportion of the taxes paid). 18 Table 11: Business Survey: Key Characteristics of Respondents (N=58) Descriptive statistics for age of respondent No. of observations Minimum Maximum Mean Std. deviation 42 20 59 34.00 10.41 Position of the respondent in business Position Percent Deputy of the Group 2.4 Employee 19.0 Owner's wife 2.4 Business owner 69.1 Relative 7.1 Respondent's gender Gender Percent Male 73.8 Female 26.2 Respondent is the owner or one of the owners in this business Percent Yes 71.4 No 28.6 Type of business Response option Percent Wholesale 4.8 Retail 47.6 Skilled service provider 16.7 Restaurants, hotels, and bars 19.0 Other 11.9 Year started Quartiles Year started 25 1991 50 1997 75 2001 Average number of employees (family members not included) Average number of employees At present 2.43 Three years ago 2.33 Average number of employees who are family members Average number of employees At present 1.00 Three years ago 1.07 Education level of respondent Quartiles Education level 25 Primary 7/8 50 Primary 7/8 75 Secondary 4 19 Table 12: Summary View of Local Taxes Paid by Businesses, Sample Surveys Tax type Average taxes per business (n=58); (Tsh `000) Prior Mean post % change Development levy 2.2 0.0 -100.0 City service levy 1.2 1.2 0.0 Property tax 3.6 4.6 27.6 Property rents 4.4 0.3 -93.0 Business license 66.5 33.9 -49.0 Advert fees 2.0 2.4 23.5 Market fees 31.4 30.7 -2.1 Parking fees 0.0 3.1 ++ Plying fees 0.0 24.8 ++ Bicycle license 0.2 0.0 -100.0 Local brew fee 0.1 0.5 278.9 Forest prod. fees 1.2 0.0 -100.0 Crop cess 1.6 0.0 -100.0 Other 38.7 28.1 -27.3 Average 151.3 130.1 -14.0 Figure 2: Change in Taxes Paid by Scale of Business 15 10 11% 5 0 -5 -11% -10 -15 -36% -20 -25 -30 -35 -40 Middle Small Micro 20 Table 13: Business Survey: Average Local Taxes Paid before and after the Reforms, Stratified by Turnover Tax type < 54,000/- (micro) 54,000­156,600/- (small) > 156,600/- (medium) Before After % Before After % Before After % change change Change Development levy 1.0 0.0 -100.0 1.1 0 -100.0 2.1 0 -100.0 City Service levy 0.0 0.0 0 0 5.1 5.1 0 Property tax 0.0 0.0 5.9 6.9 16.9 9.1 12.2 34.1 Council rents 0.0 0.0 1.3 1.3 0 17.1 0 -100.0 Business license 23.3 11.8 -49.4 41.1 15 -63.5 197.9 110.3 -44.3 Advertisement fee 0.0 0.0 2.9 2.9 0 5.4 7.3 35.2 Market fees 16.0 15.9 -0.6 23.2 18.5 -20.3 36.7 31.2 -15.0 Parking fees 0.0 0.0 0 0 0 12.9 Plying fees 0.0 0.0 0 0 0 102.9 Other fees and taxes 1 3.0 18.3 510.0 32.3 24.6 -23.8 59.9 60.4 0.8 Other fees and taxes 2 0.4 2.6 550.0 0 0 64.6 10.6 -83.6 Average 43.8 48.6 11.0 107.7 69.1 -35.8 398 353 -11.3 There was some variation among councils (table 14), with respondents in Ilala showing a net increase in revenue paid of 6 percent, while Moshi respondents reported a decline of 36 percent and Kilosa respondents a decline of 21 percent. As already noted for households, plying and parking fees paid to Dar es Salaam City Council seem to have become a growing imposition on both residents and businesses in recent years, and these largely explain the increased overall tax payments recorded by business respondents in Ilala MC. In Moshi MC, the overall tax burden of business appears to have declined substantially post­tax reform, a 36 percent decline overall being recorded by business respondents despite a rise in property taxes in this period. In Kilosa likewise, a substantial drop in overall tax payments was recorded by businesses, offset for some of them by a steep rise in fees for local brewing. The policy objective of improving the business environment via the 2003 and 2004 tax reforms therefore appears from these surveys to have overall had a modestly beneficial effect, with the middle category of "small businesses" benefiting the most. A key informant in Kilosa council affirmed the beneficial impact of the tax reforms on the business environment (box 1). Box 1: Impact on Business, Kilosa (Key Informant, District Executive) Asked whether the burden of taxation for businesspeople had changed as a result of the tax reforms, the informant said that that it had gone down substantially, and that visibly more people were going into business since the license fee was abolished. Businesspeople were happy, he asserted, because they can reinvest savings and expand their businesses. Few businesses in Kilosa are large enough to qualify for the Tsh 20,000 flat rate business licenses; as a result, licenses are nowadays given out by the council free of charge. 21 Table 14: Business Survey: Total Local Taxes Paid before and after the Reforms, by Council (Tsh `000) Tax type Ilala Moshi Kilosa Before After % Before After % Before After % change Change change Development levy 21.0 0.0 -100.0 38.5 0.0 -100.0 68.0 0.0 -100.0 City service levy 72.0 72.0 0.0 0.0 0.0 0.0 0.0 Property tax 50.0 65.0 30.0 160.0 203.0 26.9 0.0 0.0 Abattoir fees 0.0 0.0 0.0 0.0 0.0 0.0 Auction fees 0.0 0.0 0.0 0.0 0.0 0.0 Property rents 240.0 0.0 -100.0 18.0 18.0 0.0 0.0 0.0 Business license 2,115.6 860.0 -59.3 1,557.0 1,059.0 -32.0 185.9 50.0 -73.1 Advert fees 40.0 40.0 0.0 75.0 102.0 36.0 0.0 0.0 Market fees 648.0 564.0 -13.0 415.0 355.0 -14.5 757.8 863.2 13.9 Parking fees 0.0 180.0 0.0 0.0 0.0 0.0 Plying fees 0.0 1,440.0 0.0 0.0 0.0 0.0 Livestock levy 0.0 0.0 0.0 0.0 0.0 0.0 Bicycle license 0.0 0.0 0.0 0.0 12.0 0.0 -100.0 Cattle market fee 0.0 0.0 0.0 0.0 0.0 0.0 Local brew fee 0.0 0.0 0.0 0.0 7.6 28.8 278.9 Forest prod. fees 0.0 0.0 0.0 0.0 72.0 0.0 -100.0 Crop cess 0.0 0.0 0.0 0.0 90.0 0.0 -100.0 Bus-stand fees 0.0 0.0 0.0 0.0 0.0 0.0 Land fee 0.0 0.0 0.0 0.0 0.0 0.0 Other 1,114.8 1,207.8 8.3 1,127.0 422.0 -62.6 0.0 0.0 TOTAL 4,193.4 4,425.2 5.5 3,390.5 2,178.2 -35.8 1,193.3 942.0 -21.1 22 Chapter 5: Stakeholder Perceptions and Suggestions: Households and Businesses Households and businesspeople were asked a range of questions concerning their experience of local government, their perceptions about local taxes and tax reform, and their attitudes toward service delivery by the public sector. 1 GENERAL VIEWS ON TAXATION REFORM Both households and businesses were strongly in favor of the reforms. While businessmen were less critical than households, these responses often revealed negative attitudes toward local government's service delivery, trustworthiness, and accountability. The following were among the comments recorded: · Citizens get very little or nothing from their councils. · The abolition of these taxes allowed increased investment; previously informants felt overtaxed. · At last the government was perceived to have done something reasonable. · More people were now free to start small businesses. · The social disturbance due to the collection of unpopular taxes had abated. · Collection of the development levy used to be accompanied by brutality by officials. · The reform was seen to have helped in the fight against poverty. 2 STANDARDS OF LOCAL GOVERNMENT The reservations about local government reflected in these replies coexisted with some positive perceptions, and indeed, the predominant view was one of improvement in standards. About 80 percent of both households and businesspeople perceived that taxes are collected fairly (table 15) and roughly 95 percent said that they are properly receipted (table 16). While 97 percent of household respondents stated that they received receipts for taxes paid, it is interesting that in Ilala and Moshi councils this was correctly interpreted as being about current taxes, while in Kilosa it was interpreted as being a question about past taxes (the development levy, bicycle tax, and market fees). The supplementary qualitative data on which taxes were stated as being properly receipted in the different survey sites are provided in table 17. Table 15: Responses to Whether Tax Collection Is Fair Yes No HH 80.3 19.7 Business 81.0 19.0 Table 16: Responses to Whether Taxes Collected Are Receipted Yes No HH 97.0 3.0 Business 92.9 7.1 23 Table 17: Taxes Stated by Household Respondents as Providing Receipts for Payment (No. of Affirmative Responses) Tax or fee Ilala Moshi Kilosa Garbage collection 43 0 0 Property tax 63 70 70 Land tax 0 11 11 Market fees 0 4 20 Development tax 0 0 63 Bicycle tax 0 0 51 No. of observations 88 85 134 Qualitative comments on these responses varied greatly across the three study sites. In Ilala council there were 33 such comments and 13 of these (40 percent) responded affirmatively that tax collection was conducted fairly. Other comments in Ilala were partly about confusion caused by variable property and garbage tax fees in the same place or for different places in the council (7 respondents); and about the poor quality of service delivery in relation to the tax paid (5 respondents). In Kilosa district, the preoccupation of almost all of those who provided comments (19 respondents) was with the negative experience of the former tax regime (development levy and bicycle tax), with numerous references to threats and harassment in the collection of the development levy. In Moshi council, many respondents who added qualitative comments about tax collection methods considered that property valuations for their ward had been set too high, resulting in an insupportable level of property tax (9 out of 22 respondents). Other comments were scattered across a broad range, and few referred to the methods of current tax collection. Table 18 provides data on the experience of households regarding unofficial payments. In general, the position seems fairly positive in this regard, with definite "no" responses to requests for unofficial payments varying between 82 and 98.5 percent, with one exception. Health services (13 percent "yes") and the police (6 percent "yes") were the worst-performing services from this point of view. In addition there was a catch-all "other officials" where somewhat more ambivalent views about unofficial payments could be inferred from the percentage of "other" replies. Table 18: Requests for Unofficial Payments and Amounts (Percent, Tsh) Households Yes No Other Amount (Tsh) Educational services 1.5 98.5 0.0 50 Health 13.0 82.0 4.8 600 Water 0.4 94.0 5.6 0 Electricity 3.7 90.7 5.6 3,700 Licenses and permits 1.5 92.9 5.6 200 Avoid local tax 2.6 91.8 5.6 200 Dealing with TRA officials 0.7 93.7 5.6 100 Local politicians 1.1 93.3 5.6 0 Police 6.3 89.2 4.5 700 Other officials 1.9 31.2 66.9 50 Few respondents added additional comments on whether or not corruption was prevalent in the delivery of local government services. A total of 22 respondents across all study sites stated that "informal" payments to get services had got worse, while 12 respondents gave reasons why they thought the position had stayed the same or improved in recent years. These are small numbers in relation to the overall sample, and reveal a reluctance perhaps to enter into too much detail concerning the issue. Qualitative commentary on levels of tax paid prior to the reforms was in general rather diverse and fragmentary. Nevertheless, 10 out of 19 supplementary remarks in the Ilala survey referred to the 24 innovation of the garbage fee (which is a contracted-out private service) that must have occurred about that time, with several positive comments on this fee as well as a few comments about late garbage collection. In the Kilosa household survey, 23 out of 30 supplementary remarks (77 percent) expressed delight at the abolition of the development levy and the bicycle tax, also referring to the punitive approach that had characterized the collection of the development levy, and the disruption to people's livelihoods as they sought avoidance strategies. Based on the Kilosa responses, abolition of the development levy has been certainly popular in rural areas. 3 DEMANDS FOR IMPROVEMENT IN LOCAL SERVICES While this has certainly marked considerable progress over the situation prior to 2003, a major difficulty remains concerning the perceived link between local taxation and the provision of services. Despite a recognition that services require revenue, the role of councils in this respect is not highly rated. Households and businesspeople had a broad range of perceptions (table 19), with only education generating a preponderance of favorable responses agreeing that this has improved (58 percent of households and 45 percent of businesses). More households believe that roads and marketed facilities have deteriorated than those who believe they have improved, while it is only for water services that the majority of businesspeople think matters have worsened. Respondents in the household survey were asked to state which of the public services were a priority for them (see table 20). Significantly the two most important services across all study sites were water supply and health services. However, there are interesting variations between sites concerning how these are prioritized. The residents of Ilala, Dar es Salaam clearly regard water supply as the priority above all other services; while in Moshi this was the fifth priority (presumably because water provision already exists or is less problematic there), while in Kilosa it was the second priority. Garbage removal was the second-highest priority, behind health, in the Moshi urban sample; while it was of little concern for Kilosa residents. Education shows a degree of consistency across sites, but it is interesting that this was not a top priority in the same way as water, health (and garbage collection) across the sites. In Kilosa, provision of grain mills was mentioned by some respondents as their priority service. This is, of course, normally a private rather than public service, providing a reminder that not all services of benefit to citizens need to be delivered by government, and sometimes facilitation of private provision is an appropriate way forward at no cost to local council budgets. It should be noted, however, that although user fees emerge as an effective means of recovering operating costs, it is not yet clear whether they are a sufficient mechanism to widen access to goods and services, especially for the poor. Table 19: Rating of Quality of Services Provided by Local Councils Households Good Average Bad Other Health 33.8 35.3 27.9 3.0 Education 58.4 29.4 7.4 4.8 Water 41.6 23.8 34.6 0 Roads 25.5 36.7 37.8 0 Waste disposal 33.8 33.5 30.3 2.5 Market facilities 18.6 37.5 30.3 13.6 Businesses Good Average Bad Other Health 35.7 33.3 21.4 9.5 Education 45.2 47.6 4.8 2.4 Water 38.1 16.7 45.2 0 Roads 33.3 33.3 33.3 0 Waste disposal 40.5 35.7 23.8 0 Market facilities 26.2 59.5 9.5 4.8 25 Table 20: Priority Ranking of Services by Survey Respondents Service Ilala Moshi Kilosa All Sites No. percent No. percent No. percent No. percent Water 69 77.5 10 11.5 41 46.1 120 45.2 Health 30 33.7 32 36.8 50 56.2 112 42.3 Education 17 19.1 19 21.8 16 18.0 52 19.6 Garbage 16 18.0 28 32.2 2 2.2 46 17.4 Roads 13 14.6 17 19.5 13 14.6 43 16.2 Sewerage 4 4.5 6 6.9 2 2.2 12 4.5 Mills 0 0.0 0 0.0 4 4.5 4 0.2 TOTAL: 89 100.0 87 100.0 89 100.0 265 100.0 4 OTHER FEES AND CHARGES Respondents' comments on current taxes revealed tax types and other charges that were not listed specifically in the questionnaire. Some 30 households made additional comments each in the Ilala and Moshi research sites, while over 70 households made additional comments in Kilosa district. In Ilala council, the main preoccupation was with the fee for the privatized garbage collection service, and whether an effective service is provided for it. In Moshi, the main concern was the steep rise in property taxes since the 2003 tax reform.10 Respondents in Kilosa provided a detailed and consistent list of the various fees and charges that they pay (table 21). The main one was referred to as the "secondary school contribution." This is the agreed community contribution to a construction project, the greater part of which will have been funded from other sources. However, there is clearly some confusion among respondents about this contribution's status as a temporary or future regular charge, and similar payments are mentioned by some respondents as being made to primary schools. Other charges mentioned by households include monthly payments for school security guards, water service fees (which vary from ward to ward), a contribution called the "independence torch," and various other payments mentioned by just a few respondents. Table 21: Other Fees and Taxes Mentioned in the Kilosa Household Survey 2005 Type of fee or tax Rate of fee or tax No of answers Outliers and other comments Secondary school 2,000/- single payment per Figures of 1000/- and 4000/- "contribution" household 55 mentioned. School security guard 100/- per month 9 Some said 400/- or 500/- Water fees 100/- per month 10 300/- per month 10 Different charges in different wards Independence torch 300/- per year 6 Some said 400/- or 500/- Business area tax 1,000/- per year 1 Health insurance 5,000/- per year 2 Not clear whether this was private or local council 5 WHO SHOULD PAY? Households were asked who should pay taxes for public services, selecting from "rich should pay more," "everyone pays equally," or "user-based charges" (table 22). Nearly 31 percent considered that the rich 10However, two respondents in Moshi said that their property tax had gone down. 26 should pay more.11 Few respondents opted for "user-based charges" although this may have been due to the framing of the question, since many who considered that "everybody should pay" qualified this by saying that the more use made of a service the more the user should pay. Overall, responses to this question demonstrated a strong preference for progressive taxation among ordinary citizens, together with a view that all citizens should make at least some contribution to the cost of services. Table 22: Responses on Relative Wealth and Payment for Services Who should pay? Ilala Moshi Kilosa All sites No. percent No. percent No. percent No. percent Rich more 29 32.2 27 30.3 26 29.2 82 30.6 Everybody equally 38 42.2 24 27.0 40 44.9 102 38.1 User-based charges 23 25.6 38 42.7 33 37.1 94 35.1 and other replies TOTAL: 90 100.0 89 100.0 89 100.0 268 100.0 6 ATTITUDES TOWARD PROPERTY TAX Two-thirds of the respondents were in favor of a residential property tax while less than one-third thought it was a bad idea, and about 5 percent stated that they did not have a view either way (table 23). These results provide quite a powerful mandate for residential property taxes. Table 23: Attitudes toward Residential Property Taxes Good or bad Ilala Moshi Kilosa All Sites No. percent No. percent No. percent No. percent Good 34 58.6 47 72.3 26 61.9 107 64.8 Bad 23 39.7 13 20.0 14 33.3 50 30.3 Don't know 1 0.1 5 7.7 2 4.8 8 4.8 No. of respondents 58 100.0 65 100.0 42 100.0 165 100.0 Sixty-two percent of businesses were in favor of property taxes. Significantly, several businesspeople said that they would be strongly in favor of all local taxes being consolidated into a single property tax. This is consistent with evidence from other studies showing that taxpayers are not in principle opposed to paying more in local taxes, but rather prefer a smaller number of taxes (Semboja and Franzen 2003, 12). Investigation of which taxes should be retained or introduced provides support to the preceding result. Respondents were asked which of the current taxes that they pay should be kept, and were also asked whether there were new taxes they would suggest. Not surprisingly, rather few respondents came up with new tax ideas. However, on the question of retention, a clear preference for property taxes was revealed (table 24). Indeed, if property and land tax preferences are combined, the proportion in favor (61 percent) is four times the next highest preference, business taxes (15 percent). The residents of Ilala council appeared to find the private garbage fee system acceptable. The comments from many households in Ilala indicate that this system more than any other tax seemed to achieve the elusive link between tax payment and service delivery advocated in the literature on taxation. 11This figure probably understates the preference for progressive taxation as respondents who opted for "everyone should pay" specified that that the better-off should pay more than the poor, or that the poor should be excused from paying taxes altogether. A proportion of the respondents whose replies fell into the "other" category also expressed approval of progressive taxation. 27 Table 24: Views on Preferred Taxes to Keep or Introduce Retain or Ilala Moshi Kilosa All sites introduce No. percent No. percent No. percent No. percent Property 43 46.7 35 50.0 2 9.1 80 43.5 Land 15 16.3 15 21.4 2 9.1 32 17.4 Business 4 4.3 8 11.4 16 72.7 28 15.2 Sewage 0 0.0 1 1.4 0 0.0 1 0.5 Garbage 21 22.8 2 2.9 0 0.0 22 12.0 All taxes 9 9.8 9 12.9 2 9.1 20 10.9 No. of 92 100.0 70 100.0 22 100.0 184 100.0 respondents 7 USER FEES Attitudes toward user fees were relatively favorable. For water supply and roads, there was wide support for user fees rather than taxes, whereas for other services opinions of both business and households were split (table 25). Table 25: Responses to How Services Should Be Paid For Households Taxes User fees Other Health 70.3 27.9 1.9 Education 81.8 17.1 1.1 Water 0.0 46.8 52.8 Roads 0.0 87.4 12.6 Waste disposal 53.2 46.8 0.0 Market facilities 65.1 34.3 0.0 Business Taxes User fees Other Health 42.9 52.4 4.8 Education 69.1 31.0 0.0 Water 0.0 31.0 69.0 Roads 0.0 76.2 23.8 Waste disposal 35.7 64.3 0.0 Market facilities 66.7 33.3 0.0 8 CONCLUSIONS The perceptions and suggestions of both households and business regarding local taxation are generally very encouraging. They demonstrate a rational interest in the use of taxes for improving the circumstances of their lives, and indicate that there have been improvements in the standards of local government. The responses on issues such as progressive taxation, user fees, and residential property taxes all point the way for further policy development in these areas. This is exemplified by the views of a businessman key informant in Moshi district (box 2). 28 Box 2: Moshi Businessman's Views on Tax Reform "Property taxes be should paid to enable the council to run its activities. Before 2003, I used to pay 15,000/- as property tax, but I now I pay 17,000/- [a 13.3 percent increase]. However, those who pay property tax should enjoy the benefit of doing so. For example, in the past the council used to fumigate the houses of those who paid property taxes, and also provided sewerage and sanitation services. These services are not adequately provided today. Charging better-off people higher taxes than less well-off people might be complicated to implement. The criteria used to distinguish better-off from less well-off people might induce corruption, since the better-off would seek to get reclassified. The use of a percentage of the value of the property would be enough. Finally, the valuation of houses should be done fairly and immediately." 29 Chapter 6: Stakeholder Perceptions and Suggestions: Local Government Officials 1 PERCEPTIONS OF OFFICIALS, KEY INFORMANTS, BUSINESS, AND CITIZENS Council Officials Council officials and councilors in Kilosa, Ilala, and Moshi were almost unanimous in taking strongly negative views on the 2003 and 2004 tax abolitions. While these individuals recognized the beneficial impact on the poor and acknowledged the problematic nature of nuisance taxes, they made the following points in opposition to the reforms. Kilosa: · The authority of the council has been eroded. · Council tax revenues have been reduced by the reforms. · Taxes not abolished have become more difficult to collect; as a widespread belief developed among citizens that all taxes had been abolished. · Taxation provided a control mechanisms over livestock movements (which formerly required a movement certificate and fee). · The incentives to undertake casual labor formerly provided by the need to generate cash to pay the development levy have been removed, resulting in a reduced supply of labor. · Citizens' rights to make demands on council has been reduced, resulting in their disempowerment. · The compensation by central government for revenue loss has been inadequate. · The uncertain flow of funds from central government inhibits planning. · Reduced revenue has had a negative effect on provision of services. Kilosa district council's own data provide support for a picture of declining revenue since the reforms. The council's forward estimates and actual own revenue collected in the period 2000 to 2004 are shown in table 26. While both forward estimates and actual collections grew steadily in the early 2000s (with 2003's forward estimates continuing the trend), actual collections fell by 37.3 percent and 25.8 percent in 2003 and 2004, respectively, so that in 2004 revenue collected was less than half that in 2002. 30 Table 26: Kilosa Council: Trends in Tax Revenue, 2000­04 Year Forward estimates Actual % change actuals c.f. previous collection year 2000 286.7 238.9 n/a 2001 339.3 308.4 29.1 2002 457.6 386.9 25.5 2003 513.6 242.4 -37.3 2004 163.2 179.9 -25.8 Source: Kilosa Council Revenue Department. Ilala: · Existing and planned development programs may be adversely affected by the reforms. · There is marked anxiety that deficits may not be made good and would deteriorate over time. · The role of the community as a stakeholder with voice (through the development levy) will be seriously reduced. Moshi: · The manner of implementation of the reform by the center was non-participatory and top-down, and perceived as politically favorable for central government but highly unfavorable for local government. · Local stakeholders were not consulted about the rise in property taxes and this caused deterioration in the relationship between the council and the community (especially non-full-time residents who have second homes in the town). Minority Official View in Favor One or two of the most articulate council officials did see benefits in abolition: · The reform allowed payers to reallocate former tax to savings, and to investment in new enterprises. · It represented welcome relief from citizens' complaints over alleged council incompetence and corruption. · User fees for schools and clinics seem to be preferred as a mechanism for achieving the link between taxation and delivery. · The reform has had a significant impact on increase in micro-business start-ups. · Many taxpayers, especially but not only the poorest, habitually used to go into hiding to avoid tax collectors, with negative impact on business and farming enterprise, as well as on child care and school attendance. · The reduction in the general tax burden has proved very popular. · Methods of local taxation had been crude, ineffective and inefficient; bearing in mind high collection costs saved and central transfer, the net effect of the reform is broadly neutral (see box 3). 31 · Fluctuations in weather and market conditions, as well as the prevalence of evasion, meant that, in the past, revenues from farmers and pastoralists were variable and unpredictable, rendering planning difficult. Box 3: Dissenting View on Tax Impact, Kilosa (Official, Kilosa Council) "The tax reform has had no impact on revenue or expenditure. Methods of collection were primitive, with excessive money being spent on collection. In the case of the development levy, for example, for every 4000/- collected, the cost was 10,000/-. The levy created a lot of disturbance and the cost of collection was too high. The revenue deficit that is seen in the figures is offset by the 20 percent collection cost savings and government compensations. Furthermore our [local government] revenue mostly depends on climatic conditions and movement of pastoralists, and therefore it was fluctuating. Pastoralists used to hide their cattle during the collection period, and few farmers were involved in farming activities for fear of being caught; for that matter the amount of council revenue was not reliable. Now we are sure of the compensation from government at zero cost." Summarizing these responses, in light of the findings of the surveys, it may be concluded that the tax reform had both benefits and disadvantages: 2 UNFAVORABLE OUTCOMES · The legitimacy of LGAs has been undermined. · There has been a high level of local revenue loss outside metropolitan areas. · The link between LGAs and service provision has been further weakened. · It has become more difficult to collect the taxes not abolished. · The dependency of LGAs on central government has increased. · The reform has created confusion at the LGA level in terms of what new taxes are permitted. · The reforms have eroded the basis for citizens' voice being heard locally. 3 FAVORABLE OUTCOMES · The popularity of central government has increased. · Local taxes for citizens and businesspeople have become significantly more progressive. · Urban areas with diversified tax base have been able to adjust rapidly. · There is evidence that tax savings have been invested, with a concomitant increase in micro- business start-ups. · The number of taxes levied has been (to a degree) reduced. · Time and resources spent on tax collection have been reduced. 4 STRATEGIES OF ADJUSTMENT TO TAX REFORM There has been considerable variation in the manner in which the three councils have responded to the tax reforms. It was not at all clear whether the officials of Kilosa DC and Moshi MC understood that any proposed new local taxes will require formal agreement by the Ministry of Finance at the central level before being legally enforceable. In discussions with key informants, the following views regarding adjustment were proposed: 32 4.1. Kilosa · Reconsider the use of private agents in the collection of taxes (for example, those on forest products) that have been both inappropriate (sometimes involving harassment) and uneconomic (in that costs often exceeded additional revenues obtained). · Introduce new fees on major national transit routes for buses at bus-stands at Gairo and Mikumi. · Introduce rents payable to the council for businesses with built structures on council land. · Extending the use of property taxes not being actively considered by officials, but considered feasible by businesspeople for select areas if this is the only tax measure. 4.2. Ilala · Improve tax strategies, including high penalties for late payment. · Educate tax payers. · Organize daily meetings for tax collectors. · Institute legal action against those failing to pay. 4.3. Moshi · Downsize services provided. · Target more effective collection of the remaining tax resources such as property tax and council service levy. · Introduce a new public transportation tax to be collected at Moshi's new bus-stand. · Build a conference room that will generate income. · Charge fees to agents wishing to operate in council markets. · Introduce new parking fees. · Build housing to increase the property tax take. · Build taxable recreational facilities. Ilala MC would appear to be adopting the most sensible course, namely that of maximizing revenue generation from existing taxes, combined with a communications strategy aimed at both collectors and taxpayers. The taxation of inter-regional buses (assuming this is approved) presumably will represent a mildly regressive increase in travelers' costs. (It is interesting to note that bus users in Morogoro staged a protest against Morogoro MC on these grounds, arguing that it constituted a double levy [The Guardian, March 16, 2005]12). Some of the activities being considered by Moshi MC might be more appropriate for private rather than public investment (conference, housing). Clearly there is room for more explicit guidance on the type of taxes that might be approved, the principles on which local taxation should be built, and the areas open to LGA discretion. 12The way in which Ghana has managed to tax the informal transport and bus sector without opposition provides an interesting model (compare to Joshi and Ayee 2002). 33 5 PROPERTY TAXES Moshi MC is notable for being the only surveyed council to increase the rate of property tax paid by respondents (although it should be noted that the council's own figures indicate that property tax income fell from 23 to 20 percent of revenue between 2003 and 2004 as shown in table 10. In Moshi, several key informants pointed to exclusions causing property tax to be a weaker source of revenue than could be the case (box 4). It is interesting to note the generally low levels in Tanzania of what is internationally the most common type of local council tax. Table 1 showed that in 2003 the share of property tax in average council own revenue was less than 7 percent. With property tax formally limited to the urban areas, this figure of course is not necessarily significant. It is worth noting here the comment of a key informant in Kilosa, a senior LGA official, who considered it extremely unfortunate that property taxes could not be more widespread in the council area (see box 5). Box 4: Property Tax Exclusions in Moshi Town One key informant commented that property tax is the leading source of revenue collection in Moshi; if people were well-educated and paid this tax, it could act as the best source of generating revenue for the council. However, he said that the source is partly affected by clauses in the 2003 and 2004 tax reforms. For example, incomplete houses and mud houses are exempted from the property tax. Regarding the idea that better-off people should pay more taxes than the poor, he said that taxes should be based on a percentage of the value of the buildings, and not otherwise. Supplementary information was provided by another key informant in the council. He said that in addition to mud houses and incomplete houses, tax could not be levied on empty houses, which are a significant proportion of the housing stock in Moshi owing to the purchase of holiday homes there by people living in Dar es Salaam. He added that many people have built nice houses, but these individuals just come during Christmas. He rejected the idea that better-off people should pay more tax than less well-off people, saying that the calculation of property tax should be based on certain percentages of the value of the property, regardless of whether the owner is poor or rich. He added that if property tax were paid only by the rich, it might discourage investment in housing. Box 5: Thoughts on Property Tax, Kilosa (Key Informant, District Executive) The key informant thought it was a good idea to introduce property tax as an alternative way of increasing local council revenue in Kilosa. However, she cautioned that in Kilosa district, collection of such a tax might be difficult because of the low value of most buildings. She supported the idea that better-off people should pay more than the disadvantaged, but she did not think other types of taxes should be scrapped. For a place such as Kilosa, it might change people's views about property tax if the tax could be used to formalize land and title deeds, which could then be used as collateral to get loans. For urban areas, property tax has typically ranged between 10 percent and 20 percent of all revenue, and has been second only to licenses in the three major tax categories (with the development levy), which among them accounted for some 80 percent of revenue in the past. Fjeldstad and Semboja (2000) studied 10 urban councils during 1995­98 with an average property tax of 20 percent; Steffensen, Tidemond, and 34 Mwaimpopo (2004, annex 4.6 and 4.7) show that six urban councils with this contribution at 10 percent for 2002, and Ilala for the years 2000­02, raised 12 percent of revenue on average from this source (while Kinondoni raised 19 percent). Because property tax is levied on an immobile asset, in theory it is hard to avoid payment. Property size, type, and location provide ready indicators of gradations of wealth, and constitute a significant part of household assets. In recommending guidelines for the rationalization of local taxes, a PO-RALG study (Dec. 2002) argued for the retention of property rates, noting that "this tax has enormous potential but valuation rolls are out of date and collection procedures are poor. This should be a major revenue source for all local authorities" (Government of Tanzania 2002). These views are echoed in a major recent study of Tanzania (McCluskey et al 2003) which argues that the taxation of real property is the superior route to creating a stable, predictable and sufficient revenue stream to provide the local governments with the bulk of their financial needs. Of all the options available, it is the most progressive and assured path. The property tax is typically levied at a rate of 0.3 percent of the value of residential buildings and 0.5 percent for commercial or institutional buildings. These values were established by the Ardhi Institute in 1991. Reviewing the unsatisfactory situation in Mtwara, Kobb raises the question of whether it might not be on the whole simpler and more desirable for all local taxes to be bundled into a property tax, utilizing only a few classifications (residential/commercial, density level, specified institutions), and ensuring that valuations are up to date and comprehensive (Kobb 2001c, 23). For Singida, he argues that the overall revenue base would be considerably enhanced if only the top 100 residences were rated for property tax, and the development levy reduced, thus ensuring that the regressive tax were balanced out by a highly progressive one (Kobb 2000). Further arguments have been put forward that in terms of fundamental tax principles, property taxes are appropriate for Tanzania (Short 2003): The taxation of land and property may be justified on the grounds of both benefit and ability to pay principles of taxation. Governments provide benefits in the form of law and order that enable the maintenance of property rights. In particular there is an argument that the construction of a road or other public infrastructure confers a benefit for which a tax may be charged. Ability to pay considerations suggest that the holding of property not only implies a capacity to receive property income but also implies an inherent form of potential consumption. Property taxes also can be used as a proxy for user charges for local services that are provided but for administrative purposes direct charges are not levied. Important issues relating to property taxes are the base for levying the tax, the valuation of the base, the tax rate that is applied to the valuation and the exemptions that are allowed by the authorities. Many difficulties have been noted in a range of studies of property and land taxes in Tanzania. These may be summarized as · administrative problems of collection and enforcement · valuation capacity and information systems · dispute and litigation levels in land and property ownership · political will Collection has proved the most difficult. Property and land taxes were rated the second most difficult (next to the development levy) to collect in Kibaha and Kilosa (Fjeldstad and Semboja, 2000, 16). 35 Compliance rates noted by Kobb for Mtwara were 27 percent, and Singida 18.5 percent (Kobb 2003, table 15). The capacity to carry out cadastral surveys and maintain updated records is generally lacking. A number of studies have pointed to the shortfall in numbers of skilled valuers with highly varied levels of capacity in different urban councils. Valuation rolls are noted as missing or badly dated for a number of wards. There has been a high propensity of property owners to take their objections to political leaders and the courts. The tax-paying cohort targeted by property taxes includes those members of the socioeconomic elite from whose ranks are drawn local councilors, politicians, and the vested interests clustered around political parties seeking to maintain or gain office. Thus harsh methods of enforcement (such as seizure and forced sale) are likely to encounter widespread resistance and evasion, and the swamping of courts of appeal are likely to induce political intervention. As noted above, attitudes by households and businesspeople in the PSIA surveys toward a property tax, especially if such a measure can be combined with tax consolidation, were distinctly favorable, and this was also the opinion of a number of key informants. The benefits of consolidating LGA taxes around the property tax are such that a continuing systematic effort should be mounted to reduce the institutional obstacles that this rational measure faces. 36 Chapter 7: Local Government Taxation in Tanzania: Conclusions, Policy Dialogue, and the Way Forward This chapter reviews the main findings of the study, outlines the foundations already established for policy dialogue, and proposes next steps for both policy development and analytic work. 1 SUMMARY OF MAIN FINDINGS The main elements of the reform were the abolition of the flat rate development levy in 2003 along with other "nuisance taxes," and the abolition, in 2004, of business license fees for enterprises below a certain size. Our findings are that the impact of these reforms was progressive--the tax burden on poor households fell by one third, that for median households remained about the same (­2 percent), and better-off households paid more (+14 percent). For businesses, there was an overall decrease of 14 percent in tax payments. Within this, medium-size businesses recorded 11 percent less tax, and small businesses 36 percent less tax. However, micro-businesses paid 11 percent more tax (probably owing to stricter enforcement of business license fees). There were suggestions that business start-ups and investment had benefited from the reforms, though this was not assessed directly. On the other hand, the reform had certain unintended consequences. Councils' perceived need to retain levels of local revenue led, within only a few months, to making more intensive use of those taxes for which they had not lost their legislative mandate. Examples are billboard and bus-stand taxes, which increased sharply during 2004. If this trend continues, the councils may see a return to the previous multiplicity of taxes. The distribution of impacts was spatially as well as socially uneven. The decline in tax payments was most marked in rural areas (­28 percent). In the provincial urban area (Moshi), payments fell by just 7 percent while in Dar es Salaam they increased by 11 percent. This pattern reflects both the former higher reliance of rural councils on the discontinued development levy and the greater range of taxable activities in urban areas (though urban councils did suffer from the discontinuation of the business levy). Overall, the trend reinforces the already lower levels of per capita revenue collected in rural areas, and leaves much of the rural population (by and large a relatively disadvantaged group) outside the tax net. Overall, the winners of the reform appear to have been rural households and city councils, while urban taxpayers and rural councils were losers. The PSIA found that taxpayers are not opposed to local taxation, provided that some return was seen in the form of locally delivered public services. Citizens were in favor of consolidating tax sources, and also expressed the view that the wealthy should pay more than the poor. Property taxes were widely perceived as a means of both consolidating taxation and of promoting an equitable tax system. User fees were also favorably regarded. 2 ESTABLISHING POLICY DIALOGUE The process of policy dialogue was initiated soon after the end of fieldwork for the study in two dissemination events. On April 6, 2005, preliminary findings of the study were presented to a joint meeting of the Tanzania Task Forces on Fiscal Decentralization and Tax Reform. Discussion at this meeting contributed greatly to refinement of policy recommendations in the first draft report. On April 20, a half-day seminar dedicated to the study was held for 32 participants--central government (including 37 the PO-RALG, President's Office ­ Planning and Privatization (PO-PP), Ministry of Finance, and sectoral ministries) officials, local government officials, as well as representatives from civil society, researchers, and international development partners. The seminar broadly endorsed the approach to the research and the findings, welcomed the progressive impact of the tax reforms the study revealed, and expressed approval for the policy directions outlined in the study. The workshop put forward a number of suggestions for the simplification and consolidation of local revenue generation (including the idea of amending the Land and Ratings Acts to permit the consolidation of land and property taxes, and the extension of this measure to cover district headquarters and rural trading centers). 3 NEXT STEPS Policy Dialogue This study has uncovered a lack of clarity and consensus about the objectives and means of fiscal decentralization in Tanzania. Debate on its findings presents an opportunity for improving dialogue in this critical area. This study has shown that the impact of the reforms has been, by and large, positive (that is, as far as equity and productivity are concerned). Nevertheless, the reforms removed key sources of local revenue without debate or discussion with local government, its representatives, or other key stakeholders. The suddenness, and indeed air of arbitrariness, with which the reforms were promulgated created an impression of recentralization which was at odds with the spirit of local government reform, and has left a legacy of ill will. Substantively, too, the opportunity to rethink systematically the haphazard and inefficient system of local revenue collection was lost. The need now is for the development of a consensus around a broad strategic framework defining the respective fiscal and developmental roles of local and central government within the context of decentralization. In particular, greater clarity would be beneficial with respect to (a) the objectives of local revenue generation (b) respective responsibilities of central and local government for service provision (c) levels of central government support to local governments (including equalization policies and criteria) and the amount that would need to be collected locally per capita to achieve reform objectives (d) the most efficient methods for realizing the target amount in terms of unit costs of collection (e) ensuring a progressive income distribution effect of tax collection so that the poor are taxed less than the better-off, and the extreme poor--those who would require a safety net in bad years--are excluded from the tax net altogether. In this way, a framework could be developed defining an overall set of principles that can be varied for councils of particular sizes and types in terms of the economic activities of their citizens, their infrastructure, and the mix of central and local government funding of public services that is appropriate to them. In addressing these wider questions, a number of more specific issues would need to be encompassed: · systems for administering a wider property tax, including cost-effective valuation and assessment methods, and the possibility of integrating land and property taxation systems · the role of compensation payments and the equalization grant system in addressing inequities between districts, and in particular the growing gap between rural and urban districts 38 · possible mechanisms for promoting downward accountability, transparency, and clear links between local revenue generation and local service delivery A dialogue on such a framework would need to bring together central and local government, representatives of businesses and citizens, of taxpayers and service users, as well as international development partners. Drawing on such wide participation, it would likely best be facilitated by the organizations charged with taking forward local government frameworks (that is, LGRP and LGSP). Since there is considerable overlap between the property/business-owning class and that of political decision makers, debate will need to broach frankly the divergent vested interests served by the present system--including the political and social barriers to greater emphasis on property taxation, and to a more legitimate business tax--as well as the transparency of the revenue collection system. Policy options will need to be championed from within relevant institutions, rather than parachuted into the policy discussion from the outside. Accordingly, the best way forward may be to work with a small number of councils that would provide detailed case studies in the first instance, ensuring a more general approach to creating an equitable and sustainable revenue system for local governments.. Analytic Work Policy dialogue will need to be informed by further analytic work, and the PSIA suggests several areas for investigation related to the strategic themes outlined above. 4 MODELING LOCAL TAXATION OPTIONS Based on previous experiences of revenue and expenditure for councils of different types, target amounts for local revenue generation--per capita and total--can be defined. A range of simple models of households and businesses may then be used to explore various scenarios, such as the imposition of residential and business property tax at different levels, and assess the impact of each. Possible models might include · property tax with no threshold · property tax with exemption below a certain value threshold (say, 10 million Tsh) · property tax, combined with business taxes (on turnover, inventory, or income) above a certain threshold that excludes small and micro-businesses · abolition of new "nuisance" taxes (such as billboard taxes and bus-stand taxes) in combination with each of the above options · expansion of a "user fee" approach to public services (compared to a "general revenue" approach) with implications for the management of services, local tax needs, and the revenues required The information generated by such simple models could inform central and local government decision makers regarding the implications of the options, as well as engage the public in the debate, illustrating what the implications for tax assessments of each option would be (which may, if spread widely and progressively, turn out to be relatively modest). Drawing on income and asset information, where available, would enable the income distribution implications in each model to be assessed. 5 OTHER ANALYTIC TASKS 39 As noted earlier, GSU researchers are reviewing the policy framework underpinning the structure of the financing of local government and local government revenues. There are a number of other data and analytic tasks that could inform policy dialogue as it progresses: · Distributional data on current local tax impacts at the top of the income scale and for the largest firms. Such data seem critical to productive dialogue about future local tax policy, especially with the Ministry of Finance and the Tax Reform Task Force. · Further comparison of the profile of receipts reported by LGAs with the reported profile of payments by taxpayers. This would indicate the degree and location of any leakage of funds. · Gathering data from additional LGAs to remedy the small sample size of the present study. · Research on the aggregate impact burden of taxes (and other revenue-raising measures), up and down the income and firm-size scale, including both local and national taxes and revenue-raising measures. This list is merely an indicative one; more important than the themes of the research themselves will be the emergence from a process of policy dialogue of demands for analytic work that can contribute to the consultative development of a simple, fair, transparent, equitable, and sustainable tax system. 40 Bibliography Adamolekun, L. 1999. Public Administration in Africa. Boulder, Colorado: Westview. Bah, M., et al, 2003. "Changing Rural-Urban Linkages in Mali, Nigeria and Tanzania." Environment and Urbanization 15(1). Brautigam, D. 2002. "Building Leviathan: Revenue, State Capacity and Governance." IDS Bulletin 33(3) (July): 10­20. Institute of Development Studies, Sussex. Bryceson, D. 1999. "African Rural Labour, Income Diversification and Livelihood Approaches: A Long- Term Developmental Perspective." Review of African Political Economy, vol. 80: 171­89. Cooksey, B. 2003. "Marketing Reform? The Rise and Fall of Agricultural Liberalization." Development Policy Review 21(January): 67­91. Cross, S., and M. Kutengule. 2005. "Decentralization and Rural Livelihoods in Malawi." In F. Ellis and H. A. Freeman, eds., Rural Livelihoods and Poverty Reduction Policies. Oxford: Routledge. Ellis, F., and N. Mdoe. 2002. "Livelihoods and Rural Poverty Reduction in Tanzania." LADDER Working Paper No. 11 (ODG, February). Also published in World Development [2003] 31(8): 1367­84. Evans, A., and E. Ngalewa. 2001. "PRSP Institutionalisation Study: Final Report--Chapter 9: Institutionalising the PRSP Approach in Tanzania." ODI Working Paper no. 165. April. Overseas Development Institute. Fjeldstad, O-H., 2001a. "Donors Turn Blind Eye to Extortion in Tax Collection in Africa." Development Today XI(8) (May). --------. 2001b. `Taxation, Coercion and Donors: Local Government Tax Enforcement in Tanzania." The Journal of Modern African Studies 39(2): 289­306. --------. 2002a. "Collectors, Councilors and Donors: Local Government Taxation and State-Society Relations in Tanzania." IDS Bulletin 33(3) (July): 21­29. --------. 2002b. "Fighting Fiscal Corruption: The Case of the Tanzania Revenue Authority." CMI Working Paper no.3. Bergen. --------. 2003. "Fiscal Decentralization in Tanzania: For Better or for Worse?" Journal of Public Administration 38(2) (June): 133­49. Available at http://www.cmi.no/staff/staff.cfm?staffid=121. --------. 2004. "Citizens' Access to Information on Local Government Finances." Project Brief No.7, REPOA/NIBR/CMI, April. Fjeldstad, O-H., and K. Nygaard. 2003. "Low Awareness amongst Citizens of Local Government Reforms." Project Brief No. 6, REPOA/NIBR/CMI. 41 Fjeldstad, O-H., and J. Semboja. 2000. "Dilemmas of Fiscal Decentralization: A Study of Local Government Taxation in Tanzania." Forum for Development Studies 27(1): 7­41. --------. 2001. "Why People Pay Taxes: The Case of the Development Levy in Tanzania." World Development 29(12): 2059­74. Fjeldstad, O-H., and B. Tungodden. 2003. "Fiscal Corruption: A Vice or Virtue?" World Development 31(8): 1459­67. Fjeldstad, O-H., with F. Henjewele, G. Mwambe, E. Ngalewa, and K. Nygaard. 2004. "Local Government Finances and Financial Management in Tanzania: Observations from Six Councils 2002­2003. REPOA Special Paper No. 16, Dar es Salaam. Fjeldstad, O-H., 2002. "Taxation, Aid and Democracy: The Evolution of Tax Systems in Namibia, Tanzania and Uganda--Progress Report for the Research Council of Norway." CMI, Bergen, August. Unpublished report. Fjeldstad, Odd-Helge. 2004. "To Pay or Not to Pay? Citizens' Views on Taxation in Local Authorities in Tanzania." Paper presented at REPOA's 9th Annual Research Workshop, March 25­26, 2004, Dar es Salaam. Available at http://www.repoa.or.tz/research_activities/formative/reports/ToPaynottoPay.pdf. Fozzard, A., and F. Naschold. 2002. "How, When and Why Does Poverty Get Budget Priority: Poverty Reduction Strategy and Public Expenditure in Tanzania." ODI Working Paper no. 165. April. Gemmell, N., with O. Morrissey. 2002. "The Poverty Impacts of Revenue Systems in Developing Countries." DFID, April. Unpublished report. Government of The Netherlands. 2004. "Poverty, Policy and Perceptions in Tanzania: An Evaluation of Dutch Aid to Two District Rural Development Programmes." Ministry of Foreign Affairs, Policy and Operations Evaluation Department (IOB), February. Unpublished report. Government of Tanzania. 2004d. "LGRP: Joint Government--Donor Review." Draft Report 1, November. Unpublished report. --------. 2001a. "Public Finance in Singida." Report prepared for Urban Authorities Partnership Programme (UAPP)/DFID, April. Unpublished report. --------. 2001b. "Public Finance in Mbeya" Report prepared for UAPP/DFID, April. Unpublished report. --------. 2001c. `Public Finance in Mtwara" Report prepared for UAPP/DFID, July. Unpublished report. --------. 2001d. "A Summary of UAPP's Experience of Raising Revenues in Three Local Authorities: A Background Report for Presentation to LGRP." August. Unpublished report. Government of Tanzania. 1998. "Policy Paper on Local Government Reform." Ministry of Regional Affairs and Local Government (MRALG), Dar es Salaam, October. 42 --------. 2002. "Guidelines for Rationalization and Harmonization of Local Government Sources of Revenue, Tanzania, Mainland." PO-RALG, Dodoma, December. --------. 2004a. "Key Documents on Fiscal Decentralization in Mainland Tanzania." PO-RALG, April. --------. 2004b. "Poverty Reduction Strategy: The Third Progress Report 2002/03." Dar es Salaam, April. --------. 2004c. "Local Government Fiscal Review." Ministry of Finance, December. Hyden, G., and B. Karlstrom. 1993. "Structural Adjustment As a Policy Process: The Case of Tanzania." World Development 21(9): 1395­1404. James, R., N. Mdoe, and F. Mishili. 2002. "Tanzanian Rural Livelihoods: Towards a More Enabling Institutional Environment." LADDER Working Paper No. 13 (ODG). Joshi, J., and J. Ayee. 2002. "Taxing for the State? Politics, Revenue and the Informal Sector in Ghana." In Moore and Rakner, op. cit. Kelly, R. 2000. "Property Tax Reform in East Africa: The Tale of Three Reforms." Development Discussion Paper No. 753, Harvard Institute for International Development, February. Kelly, R., and Z. Musunu. 2000. "Implementing Property Tax Reform in Tanzania." Lincoln Institute of Land Policy Working Paper 00RK1. Kelsall, T. 2000. "Governance, Local Politics and Districtization in Tanzania: The 1998 Arumeru Tax Revolt." African Affairs 99(397) (October): 533­51. Khan, M. H. 2001. "Agricultural Taxation in Developing Countries: A Survey of Issues and Policy." Agricultural Economics 24(3): 315­28. Kobb, D. 2000. "Corruption in Tanzania: Counting and Franchise Building." In J. McLaren, ed., "Institutional Elements of Tax Design and Reform," World Bank Technical Paper (ISSN 0253- 7494) no. 539. Kobb, D., and H. Clegg, ed. 2002. "LGRP Trainer's Manual." LGRP/UAPP/DFID Version II, January. Unpublished report. Kobb D., and C. Howard. 2002. "Reforming Local Government Finances in Tanzania: A Training Manual." Report commissioned by UAPP and DFID. Unpublished report. Luoga, F. 2002. "Taxpayers' Rights in the Context of Democratic Governance: Tanzania." IDS Bulletin 33(33). --------. 2003. "The Viability of Developing Democratic Legal Frameworks for Taxation in Developing Countries: Some Lessons from the Tanzanian Tax Reform Experiences." Law, Social Justice and Global Development Journal, vol. 2, Warwick University. Manyanga, E. 2003. "Challenges to the Implementation of Tax Reform in the Local Government in Tanzania: A Study of Dar es Salaam City Council and the Then City Commission." M. Phil. Thesis abstract, University of Bergen. 43 Mawhood, P. 1983. Local Government in the Third World: The Experience of Tropical Africa. Chichester, U.K.: Wiley. McCluskey, W., R. Franzsen, T. Johnstone, and D. Johnstone. 2003. "Property Tax Reform: The Experience of Tanzania." RICS Foundation (Royal Institute of Chartered Surveyors, London). Meena, R., and R. Mhamba. n.d. "A Gender Study on Tanzania's Government Revenue and Taxation." Tanzania Gender Networking Programme (TGNP) on behalf of the Gender Micro Policy Working Group. Moore, M. 1998. "Death Without Taxes: Democracy, State Capacity, and Aid Dependence in the Fourth World." In M. Robinson and G. White, The Democratic Developmental State: Politics and Institutional Design. Oxford University Press. Moore, M., and L. Rakner. 2002. "Introduction: The New Politics of Taxation and Accountability in Developing Countries." IDS Bulletin 33(3) (July): 1­9. Morrissey, O. 1995. "Political Commitment, Institutional Capacity and Tax Policy Reform in Tanzania." World Development 23(4): 637­49. Narayan, D. 1997. "Voices of the Poor: Poverty and Social Capital in Tanzania." Environmentally and Socially Sustainable Development Studies and Monographs Series no. 20. Environmentally and Socially Sustainable Development (ESSD) Work in Progress, World Bank Report no 17387. Olowu, D. 2003. "Local Democracy, Taxation and Multi-Level Governance in Africa." Draft paper prepared for Norwegian Association for Development Research (NFU) Annual Conference on Politics and Poverty, Oslo, October 23­24. Institute of Social Studies, The Hague, Netherlands. Rakner, L., and S. Golpen. 2003. "Tax Reform and Democratic Accountability in Sub-Saharan Africa." In N. van de Walle, N. Ball, and V. Ramachandran, eds., Beyond Structural Adjustment: The Institutional Context of African Development, pp. 77­100. New York: Palgrave MacMillan. Rønsholt, F., with R. Mushi, B. Shallanda, and P. Assey. 2003. "Results-Oriented Expenditure Management: Country Study--Tanzania. ODI Working Paper no. 204. Semboja, J., and R. C. D. Franzsen. 2003. "Local Revenue Enhancement Strategies for Dar es Salaam City Council and Three Municipal Councils." REPOA, unpublished study commissioned by the World Bank. Unpublished report. Short, J. 2003. "Tanzania: Review of Tax Policy and Tax Administration Increasing Revenue." Ministry of Finance, United Republic of Tanzania, January 10. Steffensen, J., P. Tidemond, and E. Mwaimpopo. 2004. "A Comparative Analysis of Decentralization in Kenya, Uganda and Tanzania: Final Report, Vols. I and II." Unpublished report Sundet, G. 2004a. "Public Expenditure and Service Delivery in Monitoring in Tanzania: Some International Best Practices and a Discussion of Present and Planned Tanzanian Activities." Report prepared for USAID, March 19, 2004. Available at http://www.usaid.gov/tz/PESDM.doc Sundet, G. 2004b. "Expenditure Tracking and Financial Accountability in Local Government in Tanzania." Draft report prepared for USAID, July 15. Unpublished report. 44 Stella, P., 1993. `Tax Farming: A Radical Solution for Developing Country Tax Problems?" IMF Staff Papers 40(1): 217­25. Therkildsen, O. 1988. Watering White Elephants: Lessons from Donor-Funded Planning and Implementation of Rural Water Supplies in Tanzania. Uppsala: Scandinavian Institute of African Studies. --------. 2000. "Public Service Reform in a Poor, Aid Dependent Country." Public Administration and Development 20(1) (February): 61­71. UAPP. 2003. "Auditing Poverty in Tanzania: A Joint Approach at Local Government Level." Prepared by the Urban Authorities Partnership Programme for LGRP, revised December. Unpublished report. United Republic of Tanzania Vice President's Office. 2004a. "Public Expenditure Review of Environment--Financial Year 2004: Final Report." May. Unpublished report. --------. 2004b. "National Strategy for Growth and Reduction of Poverty." Second draft, October 7. Unpublished report. World Bank. 2002a. "Upgrading Low Income Urban Settlements." Country Assessment Report: Tanzania. World Bank AFTU 1 and 2, Washington, DC. --------. 2002b. Tanzania at the Turn of the Century: Background Papers and Statistics. Washington, DC: World Bank with United Republic of Tanzania. World Bank. 2004. "Tanzania: Second Poverty Reduction Support Credit and Grant Project." Available at http://www- wds.worldbank.org/servlet/WDS_IBank_Servlet?pcont=details&eid=000012009_200407121129 38. 45