Report No: AUS10133 Republic of Malawi Malawi Urbanization Review Leveraging Urbanization for National Growth and Development April 15, 2016 GSU19 AFRICA Standard Disclaimer: 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. Copyright Statement: 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, e-mail pubrights@worldbank.org. Malawi Urbanization Review Leveraging Urbanization for National Growth and Development April 15, 2016 GSURR AFRICA Contents ACKNOWLEDGEMENT ..................................................................................................... IV ABBREVIATIONS ................................................................................................................V EXECUTIVE SUMMARY .................................................................................................... VI The Contribution of Urbanization to National Growth and Development ............................................................. vi The Institutional and Financing Capacity of Urban Local Governments ............................................................... viii Policy recommendations ...................................................................................................................................... ix 1 URBANIZATION AND ECONOMIC TRANSFORMATION IN MALAWI ................... 1 1.1 Introduction .............................................................................................................................................. 1 1.2 Urbanization Trends in Malawi ................................................................................................................. 1 1.3 Opportunities and Challenges of Urbanization .......................................................................................... 6 2 RURAL-URBAN LINKAGES FOR NATIONAL GROWTH AND DEVELOPMENT .... 12 2.1. Introduction ............................................................................................................................................ 12 2.2. Growth, Employment and the Rural and the Urban Economies .............................................................. 12 2.2.1. National growth and structural change .................................................................................................12 2.2.2. Rural and urban employment ................................................................................................................16 2.3. Malawi’s Rural and Urban Economies ..................................................................................................... 18 2.4. Urbanization and Investment Scenarios .................................................................................................. 22 2.4.1. Baseline or “business-as-usualâ€? scenario ..............................................................................................23 2.4.2. Faster urbanization scenario .................................................................................................................26 2.4.3. Urban investment scenario ....................................................................................................................29 2.4.4. Win-win investment scenario ................................................................................................................31 2.5. Conclusion ............................................................................................................................................... 31 3 LOCAL GOVERNMENT FUNCTIONS AND FINANCING ......................................... 33 3.1. Introduction ............................................................................................................................................ 33 3.2. Decentralization and Local Government System ..................................................................................... 33 3.2.1. Decentralization .....................................................................................................................................33 3.2.2. Local Government System .....................................................................................................................34 3.2.3. Urban Local Government .......................................................................................................................38 3.3. Functional Assignments, Performance and Expenditure Patterns ........................................................... 42 3.3.1. Local Government Functions .................................................................................................................42 3.3.2. City Councils and Urban Infrastructure and Service Delivery ................................................................44 3.3.3. Urban Local Government Expenditure Patterns ....................................................................................51 3.4. Local Government Revenues ................................................................................................................... 53 3.4.1. Own-Source Revenue Assignments, Patterns and Performance ...........................................................54 3.4.2. Inter-Governmental Fiscal Transfers .....................................................................................................59 3.5. Borrowing and Other Financing ............................................................................................................... 63 3.6. Conclusion ............................................................................................................................................... 64 REFERENCES ..................................................................................................................... 69 ANNEX I. ECONOMY-WIDE MODEL............................................................................... 73 Consumer and producer behavior........................................................................................................................ 75 National product markets and international trade .............................................................................................. 75 Government and investment demand ................................................................................................................. 76 Factor and product market equilibrium ............................................................................................................... 77 Capital accumulation ........................................................................................................................................... 77 Internal labor migration....................................................................................................................................... 78 Agglomeration, congestion and technical change ................................................................................................ 78 ANNEX II. STRUCTURE OF THE MALAWI MODEL ........................................................ 79 ANNEX III. LOCAL GOVERNMENT AREAS AND POPULATION SIZE (2008) .............. 80 i Boxes Box 3.1 Centralized and deconcentrated sector spending ......................................................................... 44 Box 3.2 Solid waste management in Lilongwe and Blantyre: snapshots .................................................... 47 Box 3.3 Mzuzu City Council and property rates .......................................................................................... 58 Box 3.4 The Constituency Development Fund ............................................................................................ 61 Figures Figure 1.1 Urbanization Rates in Sub-Saharan Africa, 2014 ......................................................................... 2 Figure 1.2 Size Distribution of Urban Centers, 2008..................................................................................... 3 Figure 1.3. Urban Population Expansion, 1998-2008 ................................................................................... 5 Figure 1.4 Migrant Share of Urban Population, 2011 ................................................................................... 6 Figure 1.5. Value added to GDP and employment by sector........................................................................ 9 Figure 1.6. Night-time lights compound growth 1996-2010 (% annual change) ........................................ 10 Figure 2.1 Economic Growth in Sub-Saharan Africa, 1998-2013 ................................................................ 13 Figure 2.2 Structural Change in Malawi, 1998-2013................................................................................... 15 Figure 2.3 Spatial Modeling Framework ..................................................................................................... 22 Figure 2.4 Structural Change in the Baseline Scenario, 2010-2030 ............................................................ 25 Figure 2.5 Growth and Welfare Outcomes in the Baseline Scenario, 2010-2030 ...................................... 25 Figure 2.6 Final Year Deviation in Growth and Welfare Outcomes in the Faster Migration Scenario ....... 27 Figure 2.7 Structural Change in the Faster Migration Scenario, 2010-2030 .............................................. 28 Figure 2.8 Final Year Deviation in Poor Households’ Welfare .................................................................... 30 Figure 3.1 map of Malawian Districts and Cities......................................................................................... 35 Figure 3.2 Local government - total expenditure ....................................................................................... 51 Figure 3.3 LG expenditure per capita.......................................................................................................... 51 Figure 3.4 Recurrent expenditure as % of total expenditure ..................................................................... 51 Figure 3.5 Recurrent expenditure as % of total expenditure (by City Council) .......................................... 51 Figure 3.6 Payroll spending as % of total expenditure ............................................................................... 52 Figure 3.7 Payroll spending as % of total expenditure (by City Council) .................................................... 52 Figure 3.8 Capital expenditure as % of total LG expenditure ..................................................................... 52 Figure 3.9 Cities: capital spending as % of total expenditure ..................................................................... 52 Figure 3.10 District Councils: revenue sources ........................................................................................... 54 Figure 3.11 All City Councils - main revenue sources ................................................................................. 54 Figure 3.12 OSRs as % of total LG revenues (by LG category) .................................................................... 55 Figure 3.13 City Councils - OSRs per capita (Kwacha) ................................................................................. 55 Figure 3.14 City Councils: breakdown of locally generated revenues ........................................................ 56 Figure 3.15 Central Government transfers as % of total revenues ............................................................ 60 Figure 3.16 All IGFTs per capita .................................................................................................................. 60 Figure 3.17 2012-2013: Central government grants to local governments (%) ......................................... 60 Figure 3.18 LG total revenues per capita (kwacha) .................................................................................... 65 ii Tables Table 1.1 National Population Growth, 1998-2008 ...................................................................................... 3 Table 1.2 Simple Population Projection, 2008-2040..................................................................................... 4 Table 1.3: Urban settlements and urban population (2008) ........................................................................ 4 Table 2.1 Growth and Employment, 1998-2013......................................................................................... 13 Table 2.2 Decomposing Gains in Worker Productivity, 1998-2013 ............................................................ 14 Table 2.3 Rural and Urban Job Creation, 1998-2013 .................................................................................. 17 Table 2.4 Characteristics of Cities, Towns and Rural Economies, 2010 ...................................................... 19 Table 2.5 Rural, Town and City Demand Self-Sufficiency Measures, 2010 ................................................ 20 Table 2.6 Model Simulations ...................................................................................................................... 23 Table 2.7 Baseline “Business-as-Usualâ€? Scenario, 2010-2030 .................................................................... 23 Table 2.8 Wage Differentials and Migration Flows, 2010-2030 ................................................................. 26 Table 2.9 Economic Growth Results, 2010-2030 ........................................................................................ 26 Table 2.10 Household Welfare Results, 2010-2030.................................................................................... 29 Table 3.1 local government revenues and expenditure (2011-2014) ........................................................ 37 Table 3.2 LG revenues & expenditure – international comparisons .......................................................... 38 Table 3.3 Urban population and urban local government (2008 Census) .................................................. 39 Table 3.4 Urban settlements in Malawi ...................................................................................................... 39 Table 3.5 Local government functional assignments ................................................................................. 42 Table 3.6 Local government own-source revenues .................................................................................... 54 Table 3.7 Local Development Fund Windows............................................................................................. 62 Note Malawi Fiscal Year: July 1 – June 30 Exchange rates: MWK 100 = US$ 0.6349 (2009 constant prices) iii Acknowledgement The Malawi Urbanization Review is part of the Africa Urban team’s multi-year, programmatic analytical work to enhance and provide support to the process of urban development and management in Africa. A series of Urbanization Review has been conducted in a selected number of African countries such as Ghana, Côte d'Ivoire, Senegal, Nigeria, Kenya, Ethiopia, Tanzania, Rwanda, and Malawi, while a few more Urbanization Reviews are currently under preparation in the Africa region. The current report was completed under the coordination by Narae Choi (Urban Specialist, GSURR) and with the guidance from Somik Lall (Lead Urban Economist, GSURR) and Roland White (Lead Urban Specialist, GSURR). The work was initially led by Michael Webster (Senior Water and Sanitation Specialist, GWADR) and Hyoung Gun Wang (Senior Economist, GSURR), and background papers prepared by consultants were incorporated wherever relevant. The final work was conducted by the International Food Policy Research Institute (IFPRI), led by James Thurlow and involving Paul Dorosh and Karl Pauw; and Michael Winter (Consultant). The team appreciates their timely delivery of high-quality work, which made a value addition to the series of analytical work on Malawi. Juliana Restrepo (GSURR) provided some preliminary analyses. The work also benefited from the technical input from Dean Cira (Lead Urban Specialist, GSURR) and insightful comments from three peer reviewers – Catherine D. Farvacque (Lead Urban Specialist, GSURR), Richard Record (Senior Economist, GMFDR), and Nancy Lozano Garcia (Senior Economist, GSURR). The team is grateful for the support received from the World Bank Management, including Bella Bird (Country Director, AFCE1), Laura Kullenberg (Country Manager, AFMW), Sameh Wahba (Practice Manager, GSURR), Idrissa Dia (Practice Manager, GSURR) and Andre Bald (Program Leader, AFCE1). The team also appreciates assistance from Sonia Wheeler (Program Assistant, GSURR) and Tamara Mwafongo (AFMMW). A generous support was received from Multi-Donor Trust Fund for Sustainable Urban Development and the Swiss State Secretariat of Economic Affairs. iv Abbreviations CDF Constituency Development Fund CEO Chief Executive Officer CGE Computable General Equilibrium DC District Council DDF District Development Fund DFLA Development Fund for Local Authorities GDP Gross Domestic Product GOM Government of Malawi GRF General Resource Fund ICR Implementation Completion Report IDF Infrastructure Development Fund IFMIS Integrated Financial Management Information System IGFT Inter-Governmental Fiscal Transfer IHS Integrated Household Survey ISD Infrastructure & Service Delivery JICA Japan International Cooperation Agency KfW Kreditanstalt Für Wiederaufbau (German Development Bank) LDF Local Development Fund LG Local Government LGDP Local Government Development Project MASAF Malawi Social Action Fund MGDS Malawi Growth and Development Strategy MLG&RD Ministry of Local Government & Rural Development MoED Ministry of Education MP Member of Parliament MWK Malawi Kwacha NDP National Decentralization Policy NLGFC National Local Government Finance Committee NPDP National Physical Development Plan NSO National Statistics Office OSR Own Source Revenue PBGS Performance-Based Grant System PPP Public-Private Partnership RGCP Rural Growth Centres Project RFA Road Fund Administration SAM Social Accounting Matrix SCDP Secondary Centres Development Programme SWM Solid Waste Management TST Technical Support Team WB Water Board v Executive Summary The policy discussion in Malawi has been cautionary against rapid urbanization and its adverse effects such as the urbanization of poverty. Malawi being predominantly rural, national development policies have naturally focused on rural development as an end in itself but also as a means to reduce rural to urban migration by raising agricultural productivity and value-addition and creating markets and job opportunities in rural villages (GOM 2011). For instance, the Malawi Growth and Development Strategy (MGDS) states that “The long-term goal is to develop rural growth centers… and reduce rural-urban migrationâ€? (GOM 2006). Likewise, the current MGDS-II (soon to be replaced by MDGS-III) demonstrates a concern with urbanization by introducing the notion of a Green Belt with an aim to “reduce rural- urban migrationâ€? as well as population policies aimed at “addressing the vulnerabilities caused by… migration and rapid urbanizationâ€? (GOM 2011). Some of these concerns appear to have been propelled by the projection of high urbanization rate – however, Malawi is at an early stage of urbanization and is urbanizing at a moderate rate. Unlike some alarmist projections of rapid urbanization, Malawi is urbanizing at a rate of 3.7-3.9 percent per year for the period of 1998-2008, more slowly than other African countries. Around 16 percent of Malawi’s population (2.8 million) lived in urban areas for the same period. Rural to urban migration has been the main driver of urbanization in Malawi but the net migration inflow, including both rural-to-urban and urban-to-rural migrants, stands only at 14,000 working-age migrants moving to cities and towns annually (2006-2010), which explains the moderate rate of urban population growth. At the current rate and pace of urbanization, Malawi is indeed well-positioned to plan ahead to maximize the benefits of urban agglomeration, while addressing the congestion effects. The Malawi Urbanization Review examines this proposition by analyzing the past and future contribution of cities and towns to national development and by measuring this up against the current institutional and financial capacity of local governments to manage urbanization. Analyses presented in this report are particularly timely as Malawi is planning for the coming half decade through the Malawi Growth and Development Strategy (MGDS) III (2016-2020). The Contribution of Urbanization to National Growth and Development The GDP and employment patterns in the past 15 years (1998-2013) indicate that Malawi’s economy is undergoing a positive structural change. In the past decade and a half (1998-2013), Malawi experienced an annual GDP growth rate of 3.9 percent on average. Industry was the major driver of economic growth, mainly due to expansion in manufacturing and construction at an annual growth rate of 5.4 percent, as compared to three percent growth for agriculture. Sectoral patterns of employment demonstrates a clear indication of “positive structural changeâ€? – a gradual shift out of agriculture and the movement of labor from low to high productivity sectors. Employment in the service sectors grew rapidly at almost eight percent per year for 1998-2013, increasing by 17 percentage points. In comparison, employment in agriculture declined in absolute terms by 20 percentage points, despite rural population growth. However, structural transformation in Malawi is not driven by urbanization, as conventionally known, since most of the expansion of Malawi’s non-farm sector occurred in rural areas. In Malawi, rural and urban economies do not neatly correspond to agriculture and non-agriculture. Almost a third of rural jobs are in non-farm activities, while one in every six urban jobs are in agriculture. The composition of vi non-agricultural sector differs between rural and urban areas, demonstrating production and consumption linkages between the two. Among services that dominate non-farm employment in Malawi, trade activities (retail and wholesale trade as well as restaurants) are important in rural areas, where they account for about half of all non-farm jobs. In contrast, a greater proportion of urban non- farm jobs are in finance, business services and public administration. Accordingly, urban centers become net importers of agricultural and food products from rural areas, while exporting transport, finance and business services to rural areas. At the same time, the urban economy demonstrated a potential of making a greater contribution to the national development, by growing faster than the overall economy. Although data constraints make it hard to gauge the precise contribution of cities and towns to national growth, Malawi’s four cities form the economic core of the national economy, with their contribution to national GDP (33%) far larger than their population share (13%). In comparison, rural areas contain 85 percent of Malawi’s population, but account for only 62 percent of national GDP. Secondary towns contain about three percent of the population, but contribute six percent to national GDP. There also exist complex but strong linkages between rural and urban economies through production, consumption and migration channels, which can be further reinforced in support of national growth and development. Agriculture is heavily concentrated in rural areas, while industrial and service sectors that are most important in cities and towns. This production pattern defines rural-urban consumption linkages whereby urban centers import most of their agricultural and food products from rural areas, while exporting services to rural areas, such as transport, finance and business services. These strong rural-urban linkages need to be carefully factored in the analysis of future urbanization trends and their implication for public investment options. The core issue for policy debate is not about rural versus urban but about the constraints and opportunities of urbanization for national development, including both rural and urban areas. The report presents four scenarios of urbanization with different impact on national growth for the period of 2010-2030, depending on modalities of public investments: (i) Baseline “business-as-usualâ€? scenario (with urbanization rate of 16.14% in 2030); (ii) Faster urbanization scenario (over 21% urbanization rate by 2030) – with no change in the spatial allocation of public investments; (iii) Urban investment scenario – faster urbanization with increased urban investment, but at the expense of rural investment; and (iv) Win-win investment scenario – faster urbanization and increased urban investment paid for by raising taxes on urban households (therefore the baseline level of rural investment). If urbanization continues at the current rate, it is anticipated that major cities will grow further, generating national growth gains (Baseline scenario). At an accelerated rate, both benefits and costs of urbanization would increase, with structural transformation taking place at the cost of urbanizing poverty, unless further capital investment is made to reduce congestion effects (Faster urbanization scenario). In order to address congestion effects associated with faster urbanization, investment needs to increase in line with the growing demands for infrastructure and services in urban areas. With a fixed amount of public resources, this means that more investment in urban areas would reduce the level of investment in rural areas, thereby lowering agricultural productivity and hurting the rural poor who are mostly smallholder farmers (Urban investment scenario). If the food price increases due to low vii agricultural output, the urban poor will also be negatively affected as net consumer of agricultural products. A win-win solution will be found when urbanization can be financed by increasing the urban tax base and the level of public investment in rural areas is maintained (Win-win Investment Scenario). Unlike the prevailing concern with the prospect of rapid urbanization and its adverse effects, urbanization can play a positive role in Malawi’s national growth and development. Urbanization does pose significant challenges for housing, sanitation and other services – challenges that were well recognized in the previous MGDS. However, a careful analysis of past growth patterns demonstrates that even small increases in the pace of urbanization and level of urban investment could greatly enhance Malawi’s long-term economic prospects by accelerating growth and bringing more meaningful structural change. Increased investment in urban areas, which comes at the cost of rural agricultural investment, increased food prices and urban poverty, can be self-financing if urban households/firms are taxed to pay for urban development. The Institutional and Financing Capacity of Urban Local Governments To materialize this desired scenario, it is imperative to understand the capacity of local governments to function as the principal mechanisms for the decentralized delivery of public goods, since already existing challenges of infrastructure and service delivery will not resolve themselves and are instead likely to intensify with faster urbanization. Urban local governments have a formal mandate to meet a wide range of urban needs but are de facto engaged in providing a more limited set of public goods and services. The Local Government Act (1998) and National Decentralization Policy (1998) assign a wide range and large number of functions to both urban and rural local governments, which are not differentiated and apply uniformly. They include functions related to security, regulation of public areas and general public goods, vital registration, infrastructure provision, social and economic services, and other orthodox local government activities. In practice, urban local governments are “crampedâ€? by parastatals and line ministries (which are heavily involved in roads, water and electricity, as well as education and health services) and seem to be limited to residual functions (e.g. solid waste and local market management) or not very developmental functions (e.g. cemetery or park management). Management of urban functions are less efficient and also suffer from limited resources. In the case of some services, notably solid waste management, City Council performance is sub-optimal as they resort to a force account modus operandi, relying directly on City Council equipment and labor, rather than through contracting with the private sector. In other cases, notably roads, City Councils face severe funding constraints, some of which can be attributed to the expenditure patterns of City Councils, characterized by high proportion of payroll-related costs in recurrent expenditures (40-50%) and the total expenditures (32-47%), in return for a very low rate of capital expenditure (less than 10% of all city spending). Urban local governments in Malawi rely heavily on own-source revenues (OSRs), which account for 65-80 percent of the total revenues of urban local governments. This is an unusually high level of dependency on OSRs and is largely due to the very limited transfers from the central government in Malawi, well below the levels where they can meet current public infrastructure and service delivery needs. Inter-governmental fiscal transfers (IGFTs) to local governments in Malawi are very limited, and particularly inadequate for urban local governments, for whom IGFTs only account for a little over 20 viii percent of their revenues. OSRs per capita for all City Councils are not large, either as they only amount to MWK 1,500-2,000 or US$ 5 per capita annually. Other methods of revenue generation such as borrowing and private-public partnerships are currently not utilized to raise finance. Property tax is the most important revenue item for City Councils, accounting for between 50-65 percent of total OSRs and offers a large potential. According to the Local Government Act, property taxes can be levied on all properties within a local government jurisdiction, with a few exceptions (such as roads, parks and cemeteries) – that is, in principle, City Councils have a fairly wide property tax base. In practice, most City Councils limit themselves to levying property taxes on properties within “ratableâ€? areas. The criteria by which areas are deemed to be “ratableâ€? are not clearly defined but the observed tendency is for commercial/industrial, administrative and formal low density (and high value) residential areas to be designated as such, while informal and high density residential areas tend to excluded. Local Governments also enjoy very wide discretion in terms of setting property tax rates, which can vary depending on the type of property (commercial, residential, etc.) and in terms of the actual percentage of assessed property value. Despite having a large property tax base and considerable latitude in setting property tax rates, City Councils have difficulties in mobilizing property tax revenues, due to a number of problems. Property tax rolls are generally gross underestimates of real properties in urban jurisdictions. The institutional and legal framework regulating property taxation is cumbersome, inappropriate, unrealistic and outmoded to effectively update the tax rolls. In general, direct incentives for encouraging better City Council staff performance are absent, while enforcement and compliance are major challenges in mobilizing property taxes. On the tax collectors’ side, there is a high risk of property taxes to be misplaced or misappropriated, while tax-payers are reluctant to pay, given the poor services provided by City Councils. These challenges underline that increase in revenues goes hand in hand with improvement in the service delivery capacity and institutional reform of the local governments. Policy recommendations To generate more investment in urban development without hurting rural areas, Malawi needs to strengthen the capacity of urban local governments to expand their financial resources and grow their roles in infrastructure and service delivery. For urbanization to contribute to national growth, the scope of urban policy needs to be broadened to include not only residential investments, but also investments that directly support urban enterprises and job creation. This report finds that urban centers account for a much larger share of national economic growth than their small populations would suggest. Better employment opportunities in urban centers can then accelerate economic growth faster than rural non-farm development can. Urbanization will also increase the demand of urban businesses and consumers for agricultural products, which in turn can contribute to reducing poverty in rural areas. Urbanization is therefore an opportunity, not a threat, to broad-based economic development. Increased investment in urban development can be financed through a systematic, focused effort to improve the own source revenue (OSR) system in urban areas, particularly for property tax. The Local Government Act allows a formally large tax base and considerable latitude in setting property tax rates. To mobilize property tax revenues, property valuation methods should be simplified to more effectively update property tax rolls and payment systems modernized to facilitate the collection process while ix increasing transparency in the overall property tax system. If local governments are to make a genuine contribution to urban development, ways of increasing the size of inter-governmental fiscal transfers to City, Municipal and Town Councils should also be explored. It will also be necessary to change the ways in which local government, in general, and City Councils, in particular, manage service provision and their budgets. Out-sourcing of some services to the private sector as compared to the current in-house management arrangement, is likely to be more cost- effective (as has been the case in waste collection and road works and which can also be applied to other activities such as solid waste disposal) and help the urban local governments to reduce payroll and other recurrent costs, which currently accounts for the majority of their expenses and thus reduces the level of capital investment. Increase in IGFTs should also be made available only on the basis of their performance and if they meet certain conditions such as increasing their fiscal effort (to raise OSRs), making agreed changes to their management, and gradually altering the structure of their budgets. In the long run, the gap between policy intentions of decentralization and reality requires reconciliation. Constitutional provisions point towards empowered local governments in cities (and towns), playing a major role in urban infrastructure and service delivery, as well as providing a forum for democratic governance at the local level. This policy vision is shared with many other countries, in which strong and downwardly accountable municipal governments are expected to provide (or facilitate the provision of) urban public goods and services. However, urban local governments in Malawi are marginalized and weak in reality. This policy disconnect will become more acute in the context of faster urbanization. Although it is likely to contribute greatly to facilitating national economic development, faster urbanization will also increase the need to address key urban challenges, such as providing basic services to an expanding urban population and improving transport infrastructure within cities and across regions. x 1 Urbanization and Economic Transformation in Malawi 1.1 Introduction The Malawi Urbanization Review aims to provide fresh perspectives on urbanization in Malawi, by analyzing the current and potential contribution of urbanization to long-term national development and the current institutional and financial capacity of local governments to manage the process. Analyses presented in this report are particularly timely as Malawi is planning for the coming half decade through the Malawi Growth and Development Strategy (MGDS) III (2016-2020). Malawi is urbanizing at a moderate rate and has a good chance of proactively managing the urbanization process. Opportunities may arise from a positive structural change that Malawi’s economy is undergoing, whereby the driver of growth and job creation moves from agriculture to non-agricultural sectors. Faster urbanization, with strong linkages with rural areas, can contribute further to deepening such structural change. To unlock the potential of urbanization as a catalyst for long-term economic development, it is necessary to strengthen the capacity of urban local governments to generate revenues and meet the key infrastructure and service needs in urban areas, which remain challenging even at the current rate of urbanization. 1.2 Urbanization Trends in Malawi Malawi’s past national development strategy has viewed urbanization as a constraint to development rather than as an opportunity. Malawi being predominantly rural, national development policies have naturally focused on rural development as an end in itself but also as a means to reduce rural to urban migration by raising agricultural productivity and value-addition and creating markets and job opportunities in rural villages (GOM 2011). For instance, the Malawi Growth and Development Strategy (MGDS) states that “The long-term goal is to develop rural growth centers… and reduce rural-urban migrationâ€? (GOM 2006). Likewise, the current MGDS-II (soon to be replaced by MDGS-III) demonstrates a concern with urbanization by introducing the notion of a Green Belt with an aim to “reduce rural- urban migrationâ€? as well as population policies aimed at “addressing the vulnerabilities caused by… migration and rapid urbanizationâ€? (GOM 2011). This position is premised on a scenario of urban expansion, which often revolves around the projection of rapid urbanization. Estimates of Malawi’s urbanization rate have indeed been on the high side. Some present Malawi as “one of the fastest urbanizing countriesâ€?, projecting that its urban population grows at the rate of 4.7 percent, with a potential warning of the urbanization of poverty and growth of slums (UN-HABITAT 2012). The Situation of Urbanisation in Malawi Report (2013), prepared for the Ministry of Lands, Housing and Urban Development, also notes an annual urbanization rate of 5.2 percent (i.e. the increase in the share of urban residents in the total population), citing the 2008 Population and Housing Census, which then leads to the projection of 50 percent urbanization by 2050.1 Unlike the common perception of rapid urbanization, Malawi is at an early stage of urbanization and its urban population is growing at a moderate pace. Around 16 percent of Malawi’s population (2.8 million) lived in urban areas for the period of 1998-2008 and the urban population grew on average at 3.7-3.9 percent per year during the same period (as compared to the total population growth rate of 1 The 2008 Population Census (NSO 2008, pp. 26) reports that Malawi’s urban population growth rate was 5.3 percent per year during 1987-1998 but only 3.7 percent per year during 1998-2008. However, summing the reported populations of individual urban centers yields an urban population growth rate of 3.9 percent per year. 1 2.8%).2 At this rate, by 2030, one in every five Malawians (or around 20% of the population) will be a city or town dweller and by 2050, the share will reach 30 percent of the population.3 Malawi’s urbanization rate is in fact slower than that of other countries at a comparable level of urbanization in Sub-Saharan Africa (Figure 1.1). The horizontal axis of Figure 1.1 shows the share of the total population in urban areas, and the vertical axis shows the percentage point gap between urban and rural population growth rates in 2014 (a higher value implies more rapid urbanization). Malawi (MWI) lies towards the lower left-hand side of the figure, because it has a low urban population share and its urban population growth rate is about one percentage point higher than its rural population growth rate. This is half the urbanization rate of Sub-Saharan Africa (SSA) as a whole. Urbanization in Malawi is slower than in other agrarian countries with low urban population shares, such as Ethiopia (ETH), Rwanda (RWA) and Uganda (UGA).4 Figure 1.1 Urbanization Rates in Sub-Saharan Africa, 2014 5 rwa Urban minus rural annual population 4 growth rate (% point) tza 3 eth uga ssa 2 zam mwi moz 1 0 0 10 20 30 40 50 60 70 Urban population share (%) Source: Own calculations using World Bank (2015a). Notes: MWI is Malawi; RWA is Rwanda; ZAM is Zambia; MOZ is Mozambique; UGA is Uganda; ETH is Ethiopia; TZA is Tanzania; and SSA is Sub-Saharan Africa. The above figure uses the National Statistics Office (NSO)’s definitions of “urbanâ€? areas, which varies across countries. Malawi’s 2008 Population Census (NSO 2010a) separates urban centers into three categories: (i) primary centers (e.g., Lilongwe); (ii) secondary centers (e.g., Mangochi); and (iii) other urban centers. This classification is not based solely on population size, but rather on an urban center’s 2 Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010). 3 Projections for Malawi’s urban population made by the National Statistical Office (see http://www.nsomalawi.mw/images/stories/data_on_line/demography/census_2008/Main%20Report/ThematicRe ports/Population%20Projections%20Malawi.pdf). 4 From a global perspective, the rate of urbanization in Rwanda, which is the fastest urbanizing country in Sub- Saharan Africa, is similar to that of China. Malawi’s urbanization rate is closer to that of India. 2 orientation towards non-agricultural activities, its population density, and the availability of services.5 As such, the ranking of centers by population size does not neatly divide into secondary and other towns. This is evident from Figure 1.2, which shows the size distribution of urban centers in 2008. Luchenza, for example, is a secondary town with a population of 11,000 people, whereas Nkhotokota has a population of 25,000 but is not classified as a secondary town. The figure also shows the large population gap between primary and secondary centers. Considering this, urban areas in this report are grouped into “primary citiesâ€? and “secondary townsâ€?, where the latter includes all of the “other centers.â€? Figure 1.2 Size Distribution of Urban Centers, 2008 674 661 Primary centers Secondary centers Population (1000s log scale) 134 Other centers 88 51 40 40 28 25 23 23 21 20 20 19 18 18 15 15 15 14 14 12 11 11 8 8 7 7 2 1 Zomba Salima Ntchisi Ngabu Kasungu Mulanje Dedza Mchinji Chiradzulu Mzuzu Mangochi Karonga Nkhotakota Mzimba Ntcheu Dowa Blantyre Lilongwe Mwanza Balaka Nsanje Thyolo Liwonde Mponela Luchenza Rumphi Chitipa Machinga Nkhatabay Chikwawa Monkeybay Source: Own calculations using the 2008 Population Census (NSO 2010a). The urban population grew at 3.7 percent over the decade (1998-2008). Malawi’s total population grew at 2.8 percent per year, which, given the large size of the rural population, was close to the rural population growth rate of 2.6 percent. Secondary towns grew more rapidly at 4.2 percent per year, but their small size meant that, in absolute terms, primary cities still dominated urban expansion. Table 1.1 National Population Growth, 1998-2008 Population (1000s) Annual 1998 2008 growth (%) National 9,934 13,077 2.8 5 It should be noted, however, non-agricultural activities including trading, manufacturing, transportation, social services, construction, financial services, mining and quarrying, exist in both urban and rural areas. The classification of urban areas by the NSO also differs from that of the Ministry of Lands, Housing and Urban Development. The 1987 National Physical Development Plan (NPDP) defines urban centers according to levels of service provision such as administration, commerce and business, health, education and infrastructure; and classifies them into the hierarchy of two national centers (Blantyre and Lilongwe), one regional center (Mzuzu) and six sub-regional centers (Karonga, Liwonde, Mangochi, Salima, Dedza and Bangula). 3 Urban 1,427 2,044 3.7 Primary 1,095 1,558 3.6 Secondary 157 236 4.2 Other 175 250 3.6 Rural 8,506 11,033 2.6 Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010). At current population growth rates, Malawi’s urban population share will remain below 20 percent until 2040. A simple projection of population levels for 2040, assuming that current annual growth rates for cities, towns and rural areas continue, suggests that Malawi’s total population could reach 32 million people by 2040 (Table 1.2).6 The urban population share increases from 15.6 percent in 2008 to 19.9 percent in 2040. Together, population growth and urbanization imply that Malawi’s towns and cities will need to accommodate an additional three million people by 2040, i.e., a threefold increase over 2008 population levels. Even though secondary towns are growing faster than major cities, almost three quarters of the increase in the urban population will take place in Lilongwe and Blantyre. Table 1.2 Simple Population Projection, 2008-2040 Population (1000s) 2015 2040 Change National 13,817 31,779 17,962 Urban 2,152 6,327 4,175 Primary 1,672 4,810 3,138 Secondary 235 805 570 Other 246 713 467 Rural 11,665 25,452 13,787 Urban share (%) 15.6 19.9 - Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010). Malawi’s urbanization pattern is defined primarily by its four main cities, which account for the majority of the urban population. Malawi’s urban population lives in 31 urban settlements, broken down into three categories of Primary, Secondary and Other settlements (Table 1.3). Malawi’s primary urban settlements are relatively small in terms of population by comparison with their peers in other countries; they are also low density settlements, spread out over substantial urban jurisdictions. Nonetheless, over 75 percent of the urban population lives in these cities, although their growth rate is not necessarily the highest, with some of the secondary towns (and other urban areas) growing faster. Table 1.3: Urban settlements and urban population (2008) Urban settlement Sub-category No of Total population % of total urban category settlements population Primary All 4 1,557,986 76.1 Cities (Lilongwe & Blantyre) 2 1,335,704 65.3 Emerging cities (Mzuzu) 1 133,968 6.5 Large towns (Zomba) 1 88,314 4.3 6 This is close to the United Nation’s population projection of 34 million for 2040 (UNDESA 2015). 4 Secondary 9 237,471 11.6 Other 18 250,754 12.3 Totals 31 2,046,211 100.0 Source: NSO (2008) There is considerable variation in population growth rates across individual urban centers. Population growth rates for urban centers ranked by their population size in 2008 show that average annual population growth rates among secondary centers ranged from 2.8 percent in Luchenza to 6.7 percent in Mangochi (Figure 1.3). There is even wider variation in population growth rates across “otherâ€? urban centers. In comparison, the four cities, particularly Lilongwe and Blantyre, with markedly larger populations, will continue to be dominant in terms of their absolute population size in the coming years. Malawi’s urban policies and investments will need to accommodate the needs of small towns like Mangochi, whose population is set to double over the next decade, as well as those of the major cities like Lilongwe, whose population is growing more slowly but whose size dominates overall urban expansion. Figure 1.3. Urban Population Expansion, 1998-2008 8 Annual population growth rate (%) 7 6 5 mzu lil 4 zom bla 3 Primary centers 2 Secondary centers 1 Other centers 0 1 10 100 1000 Population in 2008 (1000s log scale) Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010). Notes: LIL is Lilongwe; BLA is Blantyre; MZU is Mzuzu; and ZOM is Zomba. Internal migration is the driver of urbanization in Malawi, accounting for more than half of the annual population growth in cities and towns, although at a moderate scale. Figure 1.4 shows the cumulative share of rural-to-urban and urban-to-rural migrants in Malawi’s total urban population as of 2011, counting the migrants moved to or left the urban center over the 12 years since 1998. 7 In 2011, 7.9 percent of the urban population were migrants who had moved from rural to urban areas. This inflow was offset by a 2.4 percent outflow of urban residents moving to rural areas. The net effect is that 5.5 7 The Third Integrated Household Survey (IHS3) conducted in 2010/11 includes questions such as if households had always resided in their current location and, if not, where they had previously resided and which year they had migrated. This provides information on internal migration flows between and within rural and urban areas. 5 percent of the total urban population in 2011 were migrants who had migrated to towns or cities between 1998 and 2010. Figure 1.4 Migrant Share of Urban Population, 2011 10 Urban-to-rural migrants (outflow) Share of urban population in 2011 (%) Rural-to-urban migrants (inflow) 8 Net rural-to-urban (inflow) 7.9 6 5.5 4 2.4 2 0 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 Year after which migrant moved to or left the urban center Source: Own calculations using the Third Integrated Household Survey (NSO 2012). Rural to urban migration results from both push and pull factors – limited economic prospects in rural areas, where small landholdings severely constrain agrarian growth and alternative income generation opportunities are lacking; as compared to expanding economic opportunities and greater access to public services in towns and cities, whether actual or perceived (Manda 2013). Among 14,000 new working age migrants arriving in Malawi’s towns and cities each year between 2006 and 2010,8 11,000 migrants moved to major cities and the remaining 3,000 moved to secondary towns. With survey and census data combined, then this migrant inflow accounted for more than half (i.e., two percentage points) of average annual population growth in cities and towns. This means the underlying or natural rate of population growth in urban centers is lower than that of rural areas (i.e., about 1.7 percent versus 2.9 percent per year). 1.3 Opportunities and Challenges of Urbanization Despite cautionary measures, urbanization has continued and will continue in Malawi, albeit at a moderate rate. Attempts to manage urbanization either by focusing on rural development or by curtailing rural-urban migration are not unique to Malawi. According to UN Habitat (2013:25), 67 percent of the countries in the world are reported to have implemented policies aimed at reducing or reversing migration flows from rural to urban areas, even though rural-urban migration was no longer a major cause of urban growth. Globally, natural increase has accounted for 60 percent of urban population growth while 20 percent has been due to reclassification of urban areas. The trend is slightly different in Malawi where rural-urban migration remains a prime driver of urbanization (Manda 2013). 8 Note that is the net inflow of migrants (i.e., rural-to-urban migrants minus urban-to-rural migrants). 6 The challenge arises when urbanization is taking place in the absence of industrialization, job creating investments, or adequate service provision, not from urbanization per se. Given this, a key question is whether cities and towns in Malawi are ready to accommodate in-migrants with adequate housing, infrastructure and services, and by creating more jobs. Even at the current rate of urbanization and as their populations increase, Malawi’s cities and towns will face clear and growing challenges – not only so as to make a greater contribution to local (and thus national) economic growth, but also to provide local populations with better basic living conditions. With the exception of the water supply sector, there is little reliable information to allow a precise assessment of urban investment needs. Some sense of the scale of the urban development challenge in different investment and service delivery sectors is provided below by way of describing the magnitude of current shortages. ï‚· Electricity supply and distribution remains erratic and insufficient to meet growing industrial, commercial and residential demand in urban areas. In 2012, only 37 percent of the urban population (and only 10% of the total population) has access to electricity, which underlies deficiencies in the distribution network.9 Moreover, despite some improvements in recent years, electricity supply in Malawi (and its urban areas) is unreliable and insufficient to meet demand. In 2010, almost half of formal sector enterprises (which are usually located in cities) included in a survey reported themselves as having backup generators, which was twice the share found in other low-income African countries. Moreover, the rate of over 22 percent business turnover lost to power outages was more than three times as high as other low-income African countries. Based on a lost load of US$ 0.50 per kilowatt-hour (kWh), power outages are estimated to be costing the Malawian economy around two percent of its GDP.10 ï‚· Transport infrastructure.11 Although Malawi’s main road network is in much better shape than that of many other developing countries, urban and secondary road networks are often limited in coverage and feeder roads are in poor condition. Compounded by road maintenance spending that is well below the norm in Africa, this discourages public transport and increases the cost of intra- urban travel. ï‚· Water supply and sanitation.12 Malawi has a better track record in the urban water supply sector than many other developing countries, with some 96 percent of the urban population estimated as having access to improved water sources. However, only 33 percent of the urban population has access to water that is piped into their homes; the majority of those who access improved water supplies do so from either shared or multi-user sources. To ensure universal access to improved water supplies and also to secure urban water supplies until 2030, it is estimated that over US$ 1 billion will need to be invested. Sanitation in Malawi’s cities and towns is also in need of improvement: although 47 percent of the urban population has access to improved sanitation, the 9 Data from the World Bank’s Malawi country website (http://www.worldbank.org/en/country/malawi and http://data.worldbank.org/country/malawi), 10 All 2010 information on Malawi’s electricity supply sector is drawn from Foster, V. & Shkaratan, M. (2010): Malawi’s Infrastructure: a continental perspective, World Bank. 11 For more detailed information on Malawi’s transport infrastructure, see Foster, V. & Shkar atan, M. (2010): Malawi’s Infrastructure: a continental perspective, World Bank; and Foster, V. & Briceño-Garmendia, C., eds (2010): Africa’s Infrastructure: A Time for Transformation , World Bank. 12 See UN/JMP data; and World Bank (2012): Malawi Water Sector Investment Plan (Volumes I and II) 7 remaining 53 percent either share improved sanitation facilities with others, use unimproved sanitation or defecate in the open.13 Also, access to improved sanitation in urban areas has fallen from already low levels. It is estimated that around US$ 200 million will need to be invested to make a major improvement in urban sanitation. ï‚· Solid waste management. Currently, solid waste management in Malawi’s towns and cities is a major challenge. Regular waste collection – whether by public authorities or by the private sector – is generally limited to industrial, commercial and middle income housing areas. Low income housing areas, in which the majority of the urban population lives, are poorly served, if at all. Waste disposal is carried out at municipal landfill sites, where – for the most part – waste is simply dumped (rather than disposed of), and which are rapidly approaching saturation. ï‚· Natural hazard risk. Malawi is classified as one of the 15 countries with more susceptibility14 worldwide. In 2015 alone, weather shocks of flooding and drought contributed to the decline in the GDP growth rate (World Bank 2015c). In particular, urban areas are highly vulnerable15 (UNU-EHS, 2014). For instance, the Global Urban Risk Index classifies Blantyre, the second most populated city, as the city with highest mortality risk and the risk of economic loss due to earthquakes (Brecht, Deichmann & Wang. 2013). Substantial investments are needed to meet these urban challenges and modest growth in the past decade and a half (1998-2013) seems to offer some opportunities. During this period, Malawi experienced an annual GDP growth rate of 3.9 percent on average, recovering from a long spell of recession and stagnation in the 1980s and the 1990s. The continuation of positive performance is yet to be seen, however. Concerns of sluggish growth are looming again, as weak fiscal discipline results in a large deficit and macroeconomic instability, characterized by high inflation and volatile exchange rates (World Bank 2015b; 2015c). More importantly, Malawi’s economy is undergoing a “positive structural changeâ€? – a gradual shift out of agriculture and the movement of labor from low to high productivity sectors. Over the period of 1998-2013, industry was the major driver of economic growth, mainly due to expansion in manufacturing and construction at an annual growth rate of 5.4 percent, as compared to three percent growth for agriculture. In 2013, services accounted for 51.2 percent of GDP, while agriculture only accounted for less than one third share of value added. An indication of positive structural change becomes more apparent when sectoral patterns of employment is considered. For the same period, employment in the service sectors grew rapidly at almost eight percent per year, increasing by 17 percentage points, largely accounted for by trade and community services. In comparison, employment in agriculture declined in absolute terms by 20 percentage points, despite rural population growth. 13 WHO /UNICEF Joint Monitoring Program (JMP) for Water Supply and Sanitation. Water Sources and Sanitation Facilities. June 2015, Malawi. Source: www.wssoinfo.org 14 “Susceptibility refers to the likelihood of suffering from and experiencing harm, loss and disruption in an extreme event or natural hazard. Thus, susceptibility describes structural characteristics and framework conditions of a societyâ€? (UNU-EHS, 2014). 15 Vulnerability “relates to social, physical, economic and environmental factors which make people or systems susceptible to the impacts of natural hazards, the adverse effects of climate change or other transformation processes. Moreover, the term vulnerability covers factors which comprise the abilities and capacities of people or systems in order to cope with and adapt to the negative impacts of natural hazards. â€? (UNU -EHS, 2014). 8 Figure 1.5. Value added to GDP and employment by sector 100% 80% 60% 40% 20% 0% Agriculture Industry Services Agriculture Industry Services GDP Employment 1998 2013 Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010) and 2013 Labor Force Survey (NSO 2014) Non-agricultural growth and job creation, which is often associated with urbanization, cannot be simply equated with an expansion of urban economy in case of Malawi. While urban areas contribute more to job creation (12% employment in urban areas accounts for 17% of job creation), most of the expansion of Malawi’s non-farm sector over the past 15 years has occurred in rural areas. In Malawi, rural and urban economies do not necessarily correspond to agriculture and non-agriculture, respectively. Almost a third of rural jobs are in non-farm activities, while one in every six urban jobs are in agriculture. The nature of jobs within each sector also differs across the regions. Notably, among services that dominate non-farm employment in Malawi, trade activities (retail and wholesale trade as well as restaurants) are particularly important in rural areas where they account for about half of all non-farm jobs. In contrast, a greater proportion of urban non-farm jobs are in finance, business services and public administration. At the same time, the urban economy demonstrated a potential of making a greater contribution to the national development, by growing faster than the overall economy. Although data constraints make it hard to gauge the precise contribution of cities and towns to national growth, Malawi’s four cities form the economic core of the national economy, with their contribution to national GDP (33%) far larger than their population share (13%). In comparison, rural areas contain 85 percent of Malawi’s population, but account for only 62 percent of national GDP. Secondary towns contain about three percent of the population, but contribute six percent to national GDP. Capital accumulation is also being concentrated in the four main urban centers. Compound growth in lights from 1996 to 2010 shows urban areas have expanded across the country, especially in Lilongwe, Blantyre and Mzuzu. Night-time lights are a good proxy for capital accumulation in urban areas as infrastructure and industries tend to work at night and represent activity in non-agricultural sectors. 9 Figure 1.6. Night-time lights compound growth 1996-2010 (% annual change) Source: ITSQS GIS Group. There also exist complex but strong linkages between rural and urban economies through production, consumption and migration channels, which can be further reinforced in support of national growth and development. Agriculture is heavily concentrated in rural areas, while industrial and service sectors that are most important in cities and towns. This production pattern defines rural-urban consumption linkages whereby urban centers import most of their agricultural and food products from rural areas, while exporting services to rural areas, such as transport, finance and business services. These strong rural-urban linkages need to be carefully factored in the analysis of future urbanization trends and their implication for public investment options. Chapter 2 provides detailed analyses of the past growth trajectories, focusing on the rural-urban linkages, and proposes three scenarios of urbanization with different public investment options for the period of 2010-2030. The baseline is that even at the current rate of urbanization, it is anticipated that major cities will grow further, generating national growth gains. (i) At an accelerated rate, both benefits and costs of urbanization would increase, with structural transformation taking place at the cost of urbanizing poverty. (ii) In order to address congestion effects associated with faster urbanization, investment needs to increase in line with the growing demands for infrastructure and services in urban areas. With a fixed amount of public resources, this means that more investment in urban areas would come at the cost of investment in rural areas, thereby lowering agricultural productivity and hurting the 10 rural poor who are mostly smallholder farmers. (iii) A win-win solution may be found when urbanization can be financed by increasing the urban tax base and the level of public investment is maintained. The findings of the report suggests that even slightly faster urbanization would go a long way towards meeting Malawi’s need to accelerate growth and to undergo more meaningful structural change. Increased investment in urban areas, which comes at the cost of rural agricultural investment, increased food prices and urban poverty, can be self-financing if urban households/firms are taxed to pay for urban development. To materialize this desired scenario, it is imperative to understand the capacity of local governments to function as the principal mechanisms for the decentralized delivery of public goods, since already existing challenges of infrastructure and service delivery will not resolve themselves and are instead more likely to intensify with faster urbanization. Chapter 3 examines whether and to what extent current institutional and financing capacities of local governments are “fit for purposeâ€?, particularly within the context of recently revived decentralization, following the tripartite election in May 2014. The delivery of public goods and services in Malawi’s urban settlements is currently undertaken by a number of institutional actors. In normative terms, the role of local government is considerable as the Local Government Act provides City, Municipal and Town Councils with substantial statutory authority over a wide range of planning, infrastructure and service delivery functions. At the same time, other institutions such as central government line ministries and departments have similar functions and mandates and act on them with varying degrees of success. In practice, local governments play a relatively small part in infrastructure and service delivery in urban areas. “Crampedâ€? by parastatals and line ministries (which are heavily involved in roads, water and electricity, as well as education and health services), urban local governments seem to be limited to: residual functions (such as residential streets, lighting and pavements); functions that are not taken on by other agencies (such as solid waste management and running local markets); or very specific and not very “developmentalâ€? urban functions (such as cemetery management or parks and recreation). This can be partly attributed to the financial and capacity limitations of urban local governments, which needs to be addressed in order to manage urban as well as national development more effectively. Substantial inter-governmental fiscal transfers to urban local governments account only for 20 percent of their total revenues. Thus, Malawi’s cities and towns rely heavily on own-source revenues, which accounts for 65-80 percent of the total revenues, although its per capital share comes down to a meagre US$ 5 in four major cities. With recourse to borrowing being extremely limited, urban local governments are then poorly placed: (i) to take on all of their statutory functions; (ii) to “muscle inâ€? on activities already being undertaken by other actors; or (iii) to do a good job of meeting the growing needs for the municipal services that they currently try to provide (such as solid waste management, city road construction and maintenance). 11 2 Rural-Urban Linkages for National Growth and Development 2.1. Introduction This chapter reviews Malawi’s growth and employment trends between 1998 and 2013, in order to provide an empirical basis for projecting national development scenarios in the future. Economic growth in Malawi over the last decade and a half, which has increased average per capita incomes in the country, was largely driven by non-agricultural sectors. This movement of workers out of agriculture into more productive non-farm jobs has led to gradual positive structural change in the economy. Notably, most of the expansion of Malawi’s nonagricultural sector occurred in rural areas rather than towns and cities as conventionally observed. This is consistent with the country’s small urban population share and slow urbanization process relative to other developing countries.16 At the same time, the urban economy demonstrated a potential of making a greater contribution to the national development, by growing faster than the overall economy over the examined period. Based on the analysis of the economic structure of rural and urban areas and the linkages between the two, a dynamic economy- wide model is constructed to discern policy implications of faster urbanization for Malawi’s growth, employment and poverty (at national, sectoral and spatial levels). The chapter concludes by discussing different options of allocating public investment between Malawi’s major cities, secondary towns, and rural areas. 2.2. Growth, Employment and the Rural and the Urban Economies 2.2.1. National growth and structural change The last decade and a half was a period of modest but positive economic growth in Malawi. Although its economic growth is slower than in Sub-Saharan Africa as a whole, the country still performed well relative to its own historical record. Growth in national GDP averaged 3.9 percent per year between 1998 and 2013 (NSO 2015). This exceeded population growth of three percent per year (World Bank 2015a). As a result, GDP per capita increased from US$ 416 in 1998 to US$ 472 in 2013 (i.e., by about one percent per year). This is still lower than the growth rates of other African countries with per capita GDP below US$ 1000 in 2005 or the total GDP growth rate for all Sub-Saharan African countries (marked as SSA in Figure 2.1). Figure 2.1 also compares Malawi’s growth performance to that of other agrarian economies who are at a similar stage of development, such as Ethiopia, Rwanda and Uganda. Most of these countries have national development plans aimed at accelerating economic growth by promoting industrial expansion, particularly in manufacturing. 16 According to the World Bank (2015a), Malawi’s urban population share inc reased by 0.11 percentage points per year over the last five years (2010-2014). This is below the Sub-Saharan African average of 0.48 percentage points and is the eighth slowest urbanization rate in the region (see Section 2.3 for further discussion). 12 Figure 2.1 Economic Growth in Sub-Saharan Africa, 1998-2013 6 eth Average annual GDP per capita growth 5 rwa 4 moz zmb uga tza 3 2 ssa rate (%) mwi 1 0 -1 eri -2 caf -3 zwe -4 0 200 400 600 800 1,000 Average GDP per capita in 2005 ($) Source: Own calculations using World Bank (2015a). Notes: Per capita GDP is not adjusted for purchasing power parity. MWI is Malawi; RWA is Rwanda; ZAM is Zambia; MOZ is Mozambique; UGA is Uganda; ETH is Ethiopia; TZA is Tanzania; ZWE is Zimbabwe; ERI is Eritrea; CAF is Central African Republic; and SSA is Sub-Saharan Africa. Recent economic growth patterns indicate that Malawi’s economy is gradually shifting away from agriculture. As shown in Table 2.1, industry was the major driver of economic growth in Malawi, mainly due to expanded manufacturing and construction. Although services did not grow as quickly, it remains the country’s single largest source of value-added. Agriculture, by contrast, grew at a slower pace than the overall economy (3% as compared to 3.9% in terms of annual GDP growth rate), causing its share of total GDP to fall from 36.1 to 31.8 percent between 1998 and 2013. Table 2.1 Growth and Employment, 1998-2013 1998 2013 Annual growth, 1998- 2013 GDP ($ million) 4,133 7,287 3.9 Agriculture 1,491 2,315 3.0 Industry 512 1,135 5.4 Services 2,130 3,837 4.0 Employment (1000s) 4,446 5,547 1.5 Agriculture 3,736 3,556 -0.3 Industry 193 411 5.2 Services 517 1,581 7.7 GDP per worker ($) 929 (a) 1,314 (b) 3.2 Agriculture 399 651 4.5 Industry 2,658 2,764 0.4 Services 4,116 2,427 -4.7 13 Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010), 2013 Labor Force Survey (NSO 2014b), national accounts data (NSO 2015), and integrated household surveys (NSO 2000b, 2014a). Notes: Initial poverty rate is for 1997. Modest changes in the sectoral composition of GDP hide substantial shifts in sectoral patterns of employment. Employment in the service sectors grew rapidly at almost eight percent per year during 1998-2013, whereas employment in agriculture declined in absolute terms despite rural population growth (see Table 2.1).17 Rising agricultural GDP together with falling agricultural employment implies rising agricultural value-added per worker. Moreover, labor productivity is already low in agriculture and so the migration of workers to more productive sectors further increases national average GDP per worker. This movement from low to high productivity sectors is called “positive structural changeâ€? and is strongly associated with long-term economic development, which is typically characterized by a declining importance of agriculture (and with migration to urban centers). It is also assumed to be synonymous with urbanization, although, as will be seen, this has not been the case in Malawi. Using the methodology described in McMillan et al. (2014), economy-wide labor productivity is decomposed into two components: (i) within-sector gains in worker productivity; and (ii) gains arising from workers moving between sectors. The first component is the sum of sectoral productivity changes weighted by initial employment shares, assuming that changes in national employment are distributed proportionally across sectors. The second component is the additional effect of reallocating labor across sectors after accounting for changes in sectoral productivity. When workers move, in aggregate, from low to high productivity sectors or when job creation is faster in higher productivity sectors, then structural change is said to have contributed positively to national labor productivity. Table 2.2 reports the results from this growth decomposition analysis. Table 2.2 Decomposing Gains in Worker Productivity, 1998-2013 Change in value-added per worker (US$) Within- Between- Total sectors sectors change Total for all sectors 26.5 357.6 (c) 384.1 (d) Agriculture 211.7 (e) -129.7 82.0 Industry -1.3 73.7 72.5 Services -63.7 222.8 (f) 159.0 Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010), 2013 Labor Force Survey (NSO 2014b), and national accounts data (NSO 2015). Virtually all of the increase in national GDP per worker between 1998 and 2013 was due to “positive structural changeâ€? caused by a shift in employment out of agriculture, primarily into services. The total increase of US$ 384 in value-added per worker (i.e., (d) in Table 2.2. or (b) minus (a) in Table 2.1) is largely accounted for by the gains of US$ 358 arising from workers moving between sectors (i.e., (c) in the Table 2.2). Agriculture still contributed to the increase in national labor productivity, because 17 Agriculture’s falling employment share in Malawi is supported by various surveys and censuses. The share was estimated at 87 percent in 1987 (NSO 1989); 84 percent in 1998 (NSO 2000a); 73 percent in 2008 (NSO 2010a); and 64 percent in 2013 (NSO 2014b). The downward trend is consistent with the reanalysis of historical data conducted by the Groningen Growth and Development Centre (see de Vries et al. 2015). 14 average productivity rose for those workers who remained in the sector. In fact, the productivity gains within agriculture were almost equal to the productivity gains caused by farm workers moving into services (i.e., US$ 212 versus US$ 223 or (e) and (f) in Table 2.2 respectively). This suggests that agricultural growth may have been associated with labor-shedding technological change. One possible explanation is the Farm Input Subsidy Program (FISP), which provided chemical fertilizer and improved seeds to half of all farmers from 2006 onwards (see Arndt et al. 2014). This is important since it indicates that at least some of the recent changes in employment patterns (and possibly urbanization) may be due to “push factorsâ€? caused by productivity-enhancing agricultural policies.18 Of course, large productivity differentials across sectors are also a “pull factorâ€? encouraging workers to leave agriculture. The analysis suggests that workers’ exit from farming reduced average productivity levels within industry and services and narrowed sectoral productivity gaps in the country. Figure 2.2 helps visualize Malawi’s process of structural change. The vertical axis shows sectoral productivity relative to economy-wide productivity. A positive value means that a sector generated above-average value-added per worker in 1998. The horizontal axis shows the percentage point change in employment shares from 1998 to 2013. A negative value means that a sector’s share of total employment fell, even though employment in the sector may have grown in absolute terms. Finally, the size of the circles represents a sector’s initial contribution to total employment. Agriculture has the largest circle because 84 percent of Malawian workers identified themselves as being farmers in 1998. Figure 2.2 Structural Change in Malawi, 1998-2013 fbs Relative worker productivity 1.2 utl 0.8 min trn csv man 0.4 trd con 0.0 -0.4 agr -0.8 -25 -20 -15 -10 -5 0 5 10 15 Change in employment share (%-point) Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010), 2013 Labor Force Survey (NSO 2014b), and national accounts data (NSO 2015). Notes: Size of circle equals initial employment share. AGR is agriculture; MIN is mining; MAN is manufacturing; UTL is utilities (electricity and water); CON is construction; TRD is trade services; TRN is transport and communication; FBS is financial and business services; and CSV is public and community services. Overall, trade and community services generated half of the new jobs and three quarters of the positive structural change generated in Malawi over the last 15 years. Agriculture’s share of total 18 This is supported by decomposition analysis for the recent 2008-2013 period, which indicates an accelerated decline in agriculture’s employment share relative to the earlier 1998 -2008 period. 15 employment fell by almost 20 percentage points over the same period (see AGR in the figure). All of the other sectors shown in the figure experienced an increase in their employment share. The largest expansion in employment took place within trade services (TRD), followed by community services (CSV). Community services include high productivity services like health and education, as well as low productivity services like paid domestic work. Most of the observed increase in employment was within the lower productivity community services. 2.2.2. Rural and urban employment In Malawi, it is difficult to equate non-agricultural growth and job creation with an expansion of the urban economy. Malawi has a large rural non-farm economy and so sectoral patterns of growth can at best provide only a rough indication of spatial patterns of growth. As a result, GDP data alone is not sufficient to infer the contribution of urban centers to national economic development. Instead, household survey and population census data are used to compare employment levels in rural and urban areas. Table 2.3 shows rural and urban employment shares drawn from the 2013 Labor Force Survey. The table also reports each sector and region’s contribution to total job creation between 1998 and 2013 (based on the 1998 Population Census). Agricultural and non-agricultural sectors exist across rural and urban economies, although with different characteristics. The first three columns of the table indicate that the rural and urban economies do not correspond to agriculture or non-agriculture, respectively. Almost a third of rural jobs are in non-farm activities and one in every six urban jobs are in agriculture. The nature of farming varies between rural and urban areas. Urban agriculture mainly involves the production of food crops, such as maize and horticulture, which are non-tradable goods and tend to be locally consumed, whereas rural agriculture is more diverse and includes export-oriented, tradable crops like tobacco and cotton. Most agricultural jobs in Malawi involve smallholder farmers operating on small plots of land. In rural areas, however, there are also large-scale estate farms, mainly producing maize and sugarcane. Most rural manufacturing jobs are related to the agricultural sector, such as maize milling. On the other hand, urban manufacturing includes activities like the production of clothing, chemicals and furniture. Finally, construction is an important source of employment, especially in urban areas. Overall, it is services that dominate non-farm employment in Malawi, accounting for more than a quarter of total national employment. Most service jobs are in the locally traded sector, which includes wholesale and retail trade as well as restaurants. These activities are particularly important in rural areas, where they account for about half of all non-farm jobs. In contrast, a greater proportion of urban non-farm jobs are in sectors like finance, business services and public administration (see the FBS and CSV sector in the table). The urban economy is an important source of job creation in Malawi, with urban employment growing faster than the national average over the last decade and a half. The final three columns of Table 2.3 report contributions to national job creation. Urban employment growth accounted for 17.3 percent of all new jobs created in Malawi between 1998 and 2013 (i.e., (b) in Table 2.3). This is higher than its 11.6 percent share of national employment in 2013 (i.e., (a) in Table 2.3). 16 Table 2.3 Rural and Urban Job Creation, 1998-2013 Share of total regional Share of total sector Share of total national employment, 2013 employment, 2013 change in employment, 1998-2013 (%) National Urban Rural National Urban Rural National Urban Rural All sectors 100 100 100 100 11.6 (a) 88.4 100 17.3 (b) 82.7 Agriculture (AGR) 64.1 16.4 70.4 100 3.0 97.0 -16.4 1.6 -18.0 Industry 7.4 16.7 6.2 100 26.2 73.8 19.8 2.3 17.5 Mining (MIN) 0.3 1.1 0.2 100 41.9 58.1 1.3 0.6 0.7 Manufacturing (MAN) 4.1 7.7 3.6 100 21.7 78.3 10.4 0.8 9.7 Utilities (UTL) 0.4 0.8 0.4 100 22.3 77.7 1.4 0.0 1.4 Construction (CON) 2.6 7.2 2.0 100 32.0 68.0 6.7 1.0 5.7 Services 28.5 66.9 23.5 100 27.2 72.8 96.6 13.4 83.2 Trade services (TRD) 16.9 34.8 14.6 100 23.2 76.8 62.1 8.8 53.3 Transport, communication (TRN) 2.0 5.4 1.6 100 31.4 68.6 7.4 1.2 6.2 Financial, business services (FBS) 0.9 3.9 0.5 100 49.9 50.1 3.9 1.7 2.2 Community, public services (CSV) 8.7 22.8 6.8 100 31.0 69.0 23.3 1.8 21.5 Source: Own calculations using the 1998 and 2008 Population Censuses (NSO 2000a, 2010) and 2013 Labor Force Survey (NSO 2014). 17 That being said, most of the expansion in non-farm employment in Malawi occurred in rural areas rather than urban centers. Nearly 83 percent of the new jobs were created in rural area and moreover, all rural employment growth took place within non-farm sectors as rural agricultural employment declined over this period. As was shown in Figure 2.2, most of the shift in employment out of agriculture was into trade and community services. In summary, economic growth in Malawi over the last decade and a half was associated with substantial positive structural change, despite modest economic growth. Workers moved out of low productivity agriculture into higher productivity non-farm jobs, particularly into trade and community services. It is not possible, given data constraints, to accurately estimate to what extent urban centers were responsible for driving economic growth and structural change in the country. Employment data cautions against simply equating urban development with movements out of agriculture. Most new jobs in trade and community services were actually created in rural areas and so it is here where most of Malawi’s positive structural change is likely to have occurred. Nevertheless, the analysis indicates that the urban economy has more than likely played a disproportionately large role in promoting economic development in Malawi over the last 15 years. 2.3. Malawi’s Rural and Urban Economies Evidence from developing countries, including Malawi, suggest that there are strong linkages between agricultural and non-agricultural sectors (see Diao et al. 2010). As agricultural production expands, it generates demand for non-agricultural inputs, such as fertilizers and transport services. Similarly, manufacturing growth generates demand for agricultural products. Finally, agricultural incomes may be used to purchase non-agricultural products, and vice versa. These production and consumption linkages mean that growth in one sector may have “positive spillover effectsâ€? on other sectors. These linkages may also imply trade-offs between sectors, since resources are limited. For example, the expansion of employment in Malawi’s service sector after 1998 was matched by a decline in agricultural employment. Rural-urban linkages in Malawi are more complex than sectoral linkages, however, because there is no clear distinction between activities occurring in rural and urban areas. As discussed in the previous section, similar activities exist across a “continuumâ€? of areas that are more or less urban. For example, while all urban centers may generate backward linkages to rural areas, these linkages may be stronger for smaller towns than major cities given the former’s proximity to rural areas. Given this, Malawi’s urban centers are separated into towns and cities in order to better reflect the country’s rural-urban continuum and to estimate how economic structures and linkages vary across and between cities, towns and rural areas. Subsequently, national sectors and households are disaggregated across primary cities, secondary towns, and rural areas, using a range of data is utilized, including: Malawi’s 2010 social accounting matrix (SAM),19 which was built on national accounts and agricultural and economic surveys from the National Statistical Office; revenue and expenditure data from the Ministry of Finance; and balance of payments data from the Reserve Bank of Malawi.20 Information on labor and households 19 A SAM is a consistent accounting framework that captures all income and expenditures flows in Malawi during a given year. It is an economy-wide database because it includes all sectors and households as well as the government and the economy’s interactions with the rest of the world. 20 The SAM is documented in Pauw et al. (2015). Data reconciliation was done using cross-entropy techniques that attempt to preserve, as much as possible, the information contained in original data sources. 18 (incomes and expenditures) came from the IHS3 (NSO 2012) and the 2010 and 2011 Economic Surveys (NSO 2014c). Table 2.4 report key statistics describing the city, town and rural economies in 2010. Table 2.4 Characteristics of Cities, Towns and Rural Economies, 2010 Rural Towns Cities All Population (millions) 11.9 0.4 1.8 14.1 Share (%) 84.6 2.7 12.7 100 Poor population (millions) 5.4 0.1 0.2 5.6 Share (%) 95.8 1.2 3.0 100 Consumption per capita ($) 341 940 1,136 458 Poverty headcount rate (%) 45.3 18.4 9.4 40.0 Workers (1000s) 4,604 144 559 5,307 Finished secondary school 277 18 90 385 Finished primary school 1,417 55 229 1,700 Not finished primary school 2,910 72 241 3,222 Sector GDP shares (%) 100 100 100 100 Agriculture 49.2 9.5 4.5 32.3 Industry 15.5 20.0 17.7 16.5 Services 35.3 70.5 77.8 51.2 Regional GDP shares (%) 61.6 5.9 32.5 100 Agriculture 93.8 1.7 4.5 100 Industry 57.9 7.2 34.9 100 Services 42.5 8.1 49.4 100 Total consumption shares (%) 100 100 100 100 Agriculture 49.3 28.7 22.1 39.6 Processed foods 14.3 13.7 12.8 13.8 Industrial goods 9.2 18.8 12.9 10.9 Services 27.3 38.8 52.1 35.7 Product consumption shares (%) 63.0 5.5 31.5 100 Agriculture 78.4 4.0 17.6 100 Processed foods 65.2 5.5 29.3 100 Industrial goods 53.1 9.5 37.4 100 Services 48.1 6.0 45.9 100 Source: Own calculations using Malawi SAM and CGE model. Notes: The poverty line is set at the upper threshold of the second per capita consumption quintile. Malawi’s four cities form the economic core of the national economy, with their contribution to national GDP far larger than their population share. Even though cities contain only 13 percent of the population, they generate 33 percent of national GDP. In comparison, rural areas contain 85 percent of Malawi’s population, but account for only 62 percent of national GDP. Secondary towns contain about 19 three percent of the population, but contribute six percent to national GDP. Accordingly, average per capita consumption is highest in cities and is many times larger than the rural average, which is below the national average. Rural-urban production linkages are characterized by heavy concentration of agriculture in rural areas and that of services in cities. Agricultural GDP is heavily concentrated in rural areas, with an underrepresentation of industry and services compared to the overall economy. Rural households also spend almost two-thirds of their income on processed foods and agricultural products (including meals purchased away from the home). Although some agricultural production occurs within town boundaries, it is the industrial and service sectors that are most important. Households in towns are thus far more likely than rural households to consume processed foods than unprocessed agricultural products. Formal services are overwhelmingly concentrated in cities. This is reflected in city households’ consumption patterns, which consist of more industrial products and services. Rural-urban consumption linkages exist, notably in the consumption of agricultural products by urban areas and that of banking and financial services by rural areas. For example, city households are responsible for 18 percent of all agricultural products consumed in Malawi, but cities only produce five percent of agricultural output. Cities are therefore “net importersâ€? of agricultural products from rural areas, although they are also more likely to consume foods imported from abroad. Similar rural-to- urban trade flows exist for industrial goods and for services. Electricity, for example, is generated in rural areas but consumed by urban producers and households. Rural producers and households, on the other hand, are net importers of banking and financial services from urban centers, including small towns. Rural-urban consumption linkages are also identified by the gap between a region’s national supply and demand shares. If a region accounts for a larger share of demand than it does for supply (in Table 2.5), then it means that it is a net importer. For example, rural areas account for 93.5 percent of national food crop supply, but only 71.8 percent of food crop demand. Rural areas are therefore net suppliers of food crops to the Malawian economy and this is shown by 21.7 percentage point supply-demand gap reported in the table. In contrast, cities demand more food crops than they produce and so their supply- demand gap is negative 9.2 percentage points. At the national level, total supply equals total demand and so the gap for each product group is zero. Overall, urban centers import most of their agricultural and food products from rural areas, while exporting services to rural areas, such as transport, finance and business services. Urban centers’ largest rural trade deficits are for unprocessed food products – they are much closer to self-sufficiency for processed foods. Both rural and urban areas are net importers of most manufactured goods, which reflects the country’s overall dependence on foreign imports. This is particularly the case for heavier industrial goods, such as chemicals and machinery. Table 2.5 Rural, Town and City Demand Self-Sufficiency Measures, 2010 Gap between national supply and demand shares (%-point) Allocated Un-allocated Total gap Rural Towns Cities Total -0.9 0.2 0.6 0.1 0.0 Agriculture 20.5 -2.1 -11.7 -6.7 0.0 20 Food crops 21.7 -2.3 -9.2 -10.2 0.0 Other crops 26.6 -0.1 -15.8 -10.8 0.0 Livestock 23.6 -3.1 -21.0 0.5 0.0 Forestry 8.9 -1.3 -7.4 -0.2 0.0 Fishing 27.0 -3.8 -23.6 0.4 0.0 Industry -14.9 -1.7 -8.2 24.8 0.0 Mining products 69.7 11.3 -4.7 -76.3 0.0 Manufacturing -22.7 -2.9 -17.4 43.1 0.0 Processed foods 20.7 1.5 -3.6 -18.5 0.0 Textiles -12.9 -3.6 -25.7 42.3 0.0 Wood and paper -9.7 -3.8 -29.8 43.3 0.0 Chemicals -47.8 -3.6 -17.8 69.2 0.0 Non-metallic minerals -38.7 -3.7 -10.7 53.1 0.0 Metals and machinery -38.0 -6.1 -24.1 68.3 0.0 Other manufactures 1.0 -5.0 -70.5 74.5 0.0 Construction 27.2 2.5 50.3 -80.0 0.0 Utilities -23.0 4.9 18.1 0.0 0.0 Services 1.2 3.3 14.6 -19.1 0.0 Trade services -9.9 2.6 7.3 0.0 0.0 Hotels and catering -3.3 -4.1 -2.7 10.0 0.0 Transport -34.9 10.0 16.6 8.3 0.0 Communications -28.9 -1.8 30.3 0.4 0.0 Finance and business -40.7 -1.2 37.0 5.0 0.0 Real estate 5.9 -2.7 -3.1 0.0 0.0 Community services 25.1 9.4 25.3 -59.8 0.0 Source: Own calculations using Malawi SAM and CGE model. Notes: Allocated supply-demand gaps include domestic production, intermediate demand, and household consumption. Unallocated includes gross capital formation, government consumption, imports and exports. Unfortunately, existing data sources does not allow spatial allocation of investment, government consumption, and exports on the demand-side, or imports on the supply-side. Apart from separating production and household consumption spending across rural areas, towns and cities, it is difficult to infer the direction of rural-urban trade for certain goods and services. For example, the table reports that rural and urban areas are both net importers of community services. Although the government is the major consumer of these services, it cannot be discerned where government demand is located and thus it is not possible to determine which regions are net importers or exporters of community services. A similar problem exists for construction, because investment demand cannot be spatially identified. For most other products, however, the unallocated gap mainly reflects net imports. For example, the large unallocated gap in chemicals (mainly petroleum and fertilizer) indicates that Malawi satisfies the excess demand for chemicals in both rural and urban areas through imports. In summary, complex migration, production and consumption linkages exist between rural areas, towns and cities in Malawi. Malawi’s modest urbanization is mainly driven by internal rural-to-urban migration, which has accounted for more than half of urban population growth over the last decade and a half. The long-term implications of urbanization for Malawi’s economic development prospects is yet 21 difficult to gauge. Survey and census data reveal strong production and consumption linkages between rural areas, towns and cities. The existing structural differences and economic linkages will help determine the national benefits and trade-offs from urbanization, with different implications for Malawi’s development strategies. 2.4. Urbanization and Investment Scenarios Based on the past growth trajectory, an economy-wide model of Malawi is developed in order to examine whether faster urbanization would benefit Malawi in the long-run. The type of model used for the projection of future growth pattern, called a “Computable General Equilibriumâ€? (CGE) model, captures the structural characteristics of and linkages between urban and rural areas.21 The CGE model includes the complex workings of the economy, including all interactions between producers, households, government and the rest of the world. Figure 2.3 offers a stylized representation of the model.22 Drawing on the spatial SAM discussed in the previous section, Malawi’s economy is separated into three spatial units: primary cities, secondary towns, and rural areas. Each spatial unit contains up to 58 different sectors depending on the location of producers in Malawi. International or foreign trade, as well as internal migration between rural areas, towns and cities, are considered in the model. Figure 2.3 Spatial Modeling Framework Internal migration Small towns Major cities Agriculture Agricultur e Industry Industry Services Services Domestic trade Rural areas Agriculture National product market Industry Foreign trade Service s Rural-urban linkages in the model include trade flows in product markets, and labor migration and capital allocations in factor markets. Urbanization affects rural areas through production and consumption channels. Macroeconomic and resource constraints also generate spillovers between urban and rural economies. For example, a rapid expansion of rural agricultural exports generates the 21 This model is widely used to evaluate national policies and investment plans. In Malawi, CGE models have been used to inform national growth and investment strategies (World Bank 2007b; Benin et al. 2012); to estimate economic costs from climate risks (Pauw et al. 2010; Arndt et al. 2014) and foreign exchange shortages (Pauw et al. 2013); and to evaluate the Farm Input Subsidy Program (Arndt et al. 2015). 22 Detailed information on the model can be found in the appendix and in Diao and Thurlow (2012). 22 foreign exchange needed to pay off imported goods used by urban industries and households. Conversely, increased migration to urban centers means less labor is available in rural areas and can cause rural wages to rise. Finally, urban agglomeration and congestion effects are also included. The model is used to examine three scenarios of urbanization and public investment for the period of 2010-2030. First, in the “Faster Urbanizationâ€? scenario, the economic implications of more rapid rural- to-urban migration is explored. Secondly, in the “Urban Investmentâ€? scenario, faster urbanization is matched by larger public investments in urban areas. This scenario assumes that government resources are fixed and so higher urban investment reduces rural investment. Finally, in the “Rural Investmentâ€? scenario, there is faster urbanization and higher urban investment, but rural investment levels are maintained by increasing urban taxation. Table 2.6 summarizes the key differences between the scenarios – more detailed information for each scenario is provided below. Table 2.6 Model Simulations Model scenarios Relative to historical trend: maintain (→), increase (↑) or decrease (↓) Rural-to-urban Urban investment Rural investment Urban migration rate shares shares tax rates Baseline → → → → 1 Faster Urbanization ↑ → → → 2 Urban Investment ↑ ↑ ↓ → 3 Win-win Investment ↑ ↑ → ↑ Source: Malawi CGE model simulations. In order to measure the impacts of the three urbanization and investment scenarios, a baseline or reference scenario is developed first, on the assumption of a “business-as-usualâ€? growth path. 2.4.1. Baseline or “business-as-usualâ€? scenario The baseline scenario assumes that economic growth in Malawi during 2010-2030 replicates historical trends observed during 1998-2013. The national population grows at 2.8 percent per year with faster growth in cities and towns (see Table 2.7). Total labor supply expands more slowly than population growth at 1.4 percent, which is close to the 1.5 percent growth rate during 1988-2013 (see Table 2.1). This means that the national dependency ratio increases over time, as well as population density. The model captures this by assuming slow agricultural land expansion in rural areas (0.5 percent per year) and no land expansion in urban areas. Private capital accumulation rates are determined inside the model based on past levels of investment. Public investment is assumed to grow at four percent each year, which is faster than population growth.23 This means that public capital per capita is rising in the baseline and this contributes positively to urban economic growth by reducing congestion effects. Table 2.7 Baseline “Business-as-Usualâ€? Scenario, 2010-2030 Rural Towns Cities All Annual GDP growth (%) 3.04 4.91 5.22 3.96 Labor 1.10 2.90 2.90 1.38 23 In 2013-2014 capital expenditure for all local governments was just above five percent of the total expenditure (see Chapter 3 for further details). 23 Crop Land / Livestock 0.50 0.00 0.00 0.48 Private capital 4.09 6.17 6.37 6.22 Public capital 3.67 3.93 2.95 4.00 TFP 1.28 1.43 1.64 1.30 Annual migrant flow (1000s) -16.98 3.35 13.63 0.00 Inflow 0.00 3.80 13.63 17.43 Outflow -16.98 -0.45 0.00 -17.43 Share of workforce (%) -0.31 1.62 1.71 0.00 Population growth rate (%) 2.70 3.27 2.98 2.75 Source: Malawi CGE model results. Labor supply growth rates in rural and urban areas diverge over time due to migration. The model endogenously reallocates labor and populations across cities, towns and rural areas based on wage differentials. It is assumed that the wage differentials in the model’s base year (2010) reproduce the migration flows observed in the 2010 IHS3 (NSO 2012). Due to national population growth, a net annual inflow of 14,000 migrants into urban areas during 2005-2010 rises to an average 17,000 migrants per year during 2010-2030, with 3,800 moving to towns and 13,200 moving to cities (Table 2.7). There is also migration from towns to cities, which then receive a total of 13,600 migrants annually. Migration causes urban population growth to exceed that of rural areas, which is consistent with the historical trends. Economic growth in the models is determined by changes in factor supply and total factor productivity (TFP). As described in the previous section, productivity growth is determined by endogenous migration and agglomeration effects and by exogenous trends. Migration generates positive agglomeration effects, which are only partially offset by congestion effects. In order to replicate historical GDP growth rates (in Table 2.1), exogenous productivity growth is set at 1.5 percent per year for rural agriculture and industry and at 0.5 percent per year for rural services. Urban productivity grows faster at one and two percent per year for services and industry, respectively. Overall, total GDP in the baseline scenario grows at four percent per year (see Table 2.4) or 1.2 percent in per capita terms. This matches average GDP growth rates during 1998-2013. The nature of structural change in the baseline scenario (Figure 2.4) resembles the pattern of structural change that occurred between 1998 and 2013 (see Figure 2.2), with a falling share of the workforce employed in agriculture and a rising share in trade and community services. Poverty declines in the baseline scenario, but at a fairly modest rate. The lowest section in Figure 2.5 reports welfare changes for poor and non-poor household groups. Welfare is measured using “equivalent variationâ€?, which is a consumption-based measure that controls for changing prices. “Poor householdsâ€? are those in the bottom two national per capita consumption quintiles. Poor households’ welfare increases by 0.1 percent per year in the baseline causing poverty to fall over time. Non-poor households’ welfare grows much faster at 1.2 percent per year, causing the gap between poor and non-poor households to widen. 24 Figure 2.4 Structural Change in the Baseline Scenario, 2010-2030 Relative worker productivity 1.2 fbs 0.8 utl min trn 0.4 csv man trd con 0.0 -0.4 agr -0.8 -7.5 -5.0 -2.5 0.0 2.5 5.0 Change in employment share (%-point) Source: Malawi CGE model results. Notes: Size of circle equals initial employment share. AGR is agriculture; MIN is mining; MAN is manufacturing; UTL is utilities (electricity and water); CON is construction; TRD is trade services; TRN is transport and communication; FBS is financial and business services; and CSV is public and community services. Figure 2.5 Growth and Welfare Outcomes in the Baseline Scenario, 2010-2030 Average annual growth rate (%) 0 2 4 6 National 2.8 Population Cities 3.0 Towns 3.3 Rural 2.7 National 4.0 Cities 5.2 GDP Towns 4.9 Rural 3.0 National 1.0 Welfare Poor 0.1 Non-poor 1.2 Source: Malawi CGE model results. Note: Welfare is measured using equivalent variation, which is a consumption-based measure that controls for price changes. The baseline scenario is meant to provide a counterfactual against which the three alternative growth scenarios can be compared, and thus is broadly consistent with Malawi’s growth trajectory from the last 15 years. Based on observed migration and population growth, Malawi continues the gradual pace 25 of urbanization whereby its urban population share rises from 15.4 percent in 2010 to 16.2 percent in 2030. Agriculture’s slower growth relative to industry and services reduces its share of national GDP and reflects ongoing structural change in Malawi. 2.4.2. Faster urbanization scenario The first simulation explores the implications of faster urbanization (21.2% in 2030 as compared to the baseline urbanization of 16.2%). The urban population share rose to 16.2 percent in 2030 in the baseline scenario. By increasing rural-to-urban migration above historical rates, in the Faster Urbanization scenario this share is increased by five percentage points, i.e., to 21.2 percent in 2030. The larger flow of migrants increases job competition in urban areas and reduces labor availability in rural areas. This causes the ratio of average urban wages to average rural wages to fall (see the second column in Table 2.8). For example, by the end of the baseline scenario, the average urban wage was about 4.5 times higher than the average rural wage. However, the flow of migrants arriving in urban centers rises from 17,000 in the baseline to 47,500 in the Faster Urbanization scenario. This narrows the urban-rural wage gap to about 3.3 times the rural wage in 2030.24 Table 2.8 Wage Differentials and Migration Flows, 2010-2030 Baseline Urbanization scenarios scenario Faster Urban Win-win migration investment investment Average wage ratios, 2030 Towns / Rural areas 3.46 2.75 2.84 2.80 Cities / Rural areas 4.85 3.66 3.72 3.73 Cities / Towns 1.40 1.33 1.31 1.33 Annual net migrant flows (1000s) Rural areas -16.98 -47.45 -48.06 -49.10 Towns 3.35 7.95 7.73 8.37 Cities 13.63 39.49 40.33 40.74 Urban population share, 2030 (%) 16.24 21.18 21.27 21.42 Source: Malawi CGE model results. Agricultural GDP increases in the Faster Urbanization due to strong urban-to-rural demand linkages. More rapid urbanization reduces the total supply of labor in rural areas, which places downward pressure on agricultural GDP growth. However, urban producers benefit from higher labor supply and lower wages. Faster urban growth generates backward linkages to agriculture, mainly because rising urban incomes generate demand for agricultural products, almost all of which are produced in rural areas (see Table 2.5). This positive linkage-effect outweighs the loss of rural labor, causing agricultural GDP growth to accelerate in response to faster urbanization. The agricultural GDP growth rate in the Faster Urbanization scenario is 0.17 percentage points higher than the baseline (as highlighted in Table 2.9) – i.e., it is now 2.98 instead of 2.81 percent per year. Table 2.9 Economic Growth Results, 2010-2030 24 The falling urban-to-rural wage ratio acts as a disincentive to migrate. Thus, while historical migration rates are initially increased by fivefold, there is less than a threefold increase over the entire simulation period. 26 Total Baseline Urbanization scenarios GDP annual (%-point deviation from baseline) share, growth Faster Urban Win-win 2010 (%) rate (%) migration investment investment Annual GDP growth 100 3.96 0.66 0.68 0.82 Agriculture 32.3 2.81 0.17 -0.03 0.14 Industry 16.5 4.69 0.83 0.95 1.10 Manufacturing 10.6 4.60 0.89 1.01 1.10 Agro-processing 6.0 4.10 0.79 0.81 0.93 Other manufacturing 4.6 5.19 0.99 1.22 1.29 Other industry 4.7 4.90 0.90 1.00 1.32 Services 51.2 4.35 0.83 0.90 1.03 Rural areas 61.6 3.04 -0.29 -0.52 -0.33 Towns 5.9 4.91 1.23 1.22 1.54 Cities 32.5 5.22 1.60 1.83 1.91 Source: Malawi CGE model results. Faster urbanization leads to faster national economic growth, with the Malawi’s economy growing 14.1 percent larger than the baseline by 2030. Increasing the urban population share by five percentage points raises the total GDP growth rate by 0.7 percentage points per year, leading to the cumulative effect whereby the economy is 14.1 percent larger than it would have been without faster urbanization (Figure 2.6). It should be emphasized that this positive growth-effect is entirely due to more people leaving rural areas for urban centers, i.e., the total population and workforce force remains unchanged from the baseline. All additional growth resulting from faster urbanization occurs in cities and towns, whose economies are 37.4 and 27.7 percent larger in 2030 than in the baseline. Despite increased agricultural production, the rural economy contracts (relative to the baseline) due to a slowdown in the growth of rural non- farm activities. Changes in the composition of rural employment patterns is driven by two factors. First, greater demand from urban consumers causes food prices to rise faster than the consumer price index. Higher real food prices encourages rural workers to remain in agriculture. Second, and more importantly, urbanization and faster urban growth generates increased demand for imported goods. Most of Malawi’s foreign exchange is generated by agricultural exports, such as tobacco and cotton. As such, labor allocations within agriculture shift towards export crops and those food crops that are difficult to substitute with imports, such as livestock and fish. Urbanization therefore encourage a positive transformation within the agricultural sector towards higher-value activities. This is consistent with recent studies suggesting that urbanization in Sub-Saharan Africa will transform national food systems (see Tschirley et al. 2015). Figure 2.6 Final Year Deviation in Growth and Welfare Outcomes in the Faster Migration Scenario 27 Deviation from baseline in 2030 (%) -20 0 20 40 60 National 14.1 Cities 37.4 Towns 27.7 GDP Rural -5.7 Agriculture 3.5 Industry 18.0 Services 17.9 National 10.8 Welfare Cities -3.9 Towns -5.7 Rural 5.3 Source: Malawi CGE model results. Note: Welfare is measured using equivalent variation, which is a consumption-based measure that controls for price changes. Figure 2.7 Structural Change in the Faster Migration Scenario, 2010-2030 Relative worker productivity 1.2 fbs utl 0.8 trn min 0.4 man csv trd 0.0 con -0.4 agr -0.8 -3.0 -2.0 -1.0 0.0 1.0 2.0 Change in employment share (%-point) Source: Malawi CGE model results. Notes: Size of circle equals initial employment share. AGR is agriculture; MIN is mining; MAN is manufacturing; UTL is utilities (electricity and water); CON is construction; TRD is trade services; TRN is transport and communication; FBS is financial and business services; and CSV is public and community services. 28 Faster urbanization accelerates the pace of structural change in Malawi. Faster agricultural GDP growth is caused by farm labor moving away from food crops towards higher value agricultural activities. Migration to urban centers causes a decline in agriculture’s employment share and non- agricultural employment to increase, particularly trade and community services (Figure 2.7). However, unlike in the baseline scenario (Figure 2.4), more new jobs are created in manufacturing and construction, in part because new urban residents increase demand for manufactured goods, while urban GDP growth generates greater demand for investment goods and services, like construction. Finally, the increase in urban population exceeds the increase in urban GDP, leading to lower urban per capita GDP and welfare relative to the baseline (see Figure 2.5 and Table 2.7). Conversely, the decline in the rural population is larger than the decline in rural GDP, causing rural welfare to improve. Faster urbanization leads to higher national welfare and falling poverty, but most of the benefits accrue to non-poor households. Table 2.10 Household Welfare Results, 2010-2030 Average per Baseline Urbanization scenarios capita con- annual (Annual growth rate, %) sumption, growth Faster Urban Win-win 2010 ($) rate (%) migration investment investment National welfare 458 1.01 1.52 1.48 1.36 Poor 147 0.13 0.25 0.14 0.25 Non-poor 665 1.16 1.73 1.70 1.56 Urban welfare 1,102 1.74 1.54 1.58 1.29 Poor 174 0.95 -0.07 -0.16 0.00 Non-poor 1,217 1.90 1.88 1.92 1.63 Rural welfare 341 0.35 0.61 0.49 0.50 Poor 146 0.06 0.23 0.12 0.22 Non-poor 502 0.45 0.82 0.71 0.70 Source: Malawi CGE model results. Note: Welfare is measured using equivalent variation, which is a consumption-based measure that controls for price changes. In summary, faster urbanization accelerates the pace of economic growth and structural change and raises national welfare. Increasing urbanization from its current modest pace could help set Malawi’s on a more sustainable long-term development trajectory. Urbanization might also catalyze agricultural transformation. However, the results from the Faster Urbanization scenario suggest that urbanization, without increased public investment in urban infrastructure, might not generate sufficient economic growth to absorb new migrants, leading to slower welfare improvements for the urban poor. In the next scenario, it is examined whether allocating more public resources to urban areas can prevent an “urbanization of povertyâ€?. 2.4.3. Urban investment scenario The Urban Investment scenario replicates the Faster Urbanization scenario but now allocates more public investment to cities and towns. Urban investment levels are increased until urban public capital 29 per capita remains unchanged from baseline levels. This means that there are no negative congestion effects to offset positive urban agglomeration effects. This will lead to faster urban economic growth, which will increase the absorptive capacity of the urban labor market. Urban infrastructure per capita is higher than in rural areas and so maintaining per capita levels with an inflow of new migrants requires a substantial reallocation of public resources. Without raising the overall level of government spending in this scenario, increasing urban investment reduces investment in rural areas. The growth rate of rural public spending falls by 2.4 percent, i.e., from 3.4 percent per year in the baseline to 1.0 percent per year in the Urban Investment scenario.25 Subsequently, rural agriculture’s annual TFP growth rate26 falls by 0.4 percentage points (i.e., 0.15×2.4=0.4). This is the trade-off from increasing urban investment. Agricultural GDP growth in the urban investment scenario is then even slower than it was in the baseline scenario. In comparison, industrial and service sector GDP growth is higher than it was in the Faster Urbanization scenario (see Table 2.9 above). Faster urban GDP growth widens the gap between average urban and rural wages and encourages more migration to urban centers (Table 2.8). Combined together, urban welfare is higher and rural welfare is lower when investments are reallocated towards urban centers (Table 2.10). However, reallocating investment away from rural agriculture does not achieve the goal of preventing urban poverty from rising. Instead, reducing rural investment leads to worse outcomes for poor urban households (see Table 2.10), because slower agricultural growth leads to higher real food prices. Food purchases are a major share of poor urban households’ consumption baskets and so higher food prices lowers their real incomes, despite the increase in urban wages and job creation caused by faster urban economic growth. This finding is consistent with the reverse argument of the traditional development model that raising agricultural productivity benefits the urban poor by reducing food prices (see Diao et al. 2010). In short, increasing urban investment, while necessary to prevent an urbanization of poverty, should not come at the expense of rural investments. Figure 2.8 Final Year Deviation in Poor Households’ Welfare Deviation from baseline in 2030 (%) -6 -4 -2 0 2 4 6 Faster urbanization 5.1 Urban investment National 2.8 5.0 Win-win investment -1.3 Urban -3.1 0.0 4.8 Rural 2.5 4.5 Source: Malawi CGE model results. Note: Welfare is measured using equivalent variation, which is a consumption-based measure that controls for price changes. 25 Urban investment growth increases from 5.8 percent to 9.3 percent per year. 26 An agricultural spending to TFP growth elasticity is assumed to be 0.15, an estimation by Benin et al. (2009) using data from 18 African countries. 30 2.4.4. Win-win investment scenario In the final scenario, rural investment is maintained at baseline levels, while urban investments are also increased (thus named Win-win Investment scenario). This implies that total government spending increases (rather than rural investment is reduced to finance the increase in urban investment). Additional government spending comes from raising direct tax rates on urban enterprises and households. The rationale for this financing mechanism is that urban producers and non-poor urban households were the main beneficiaries from faster urbanization and increased investment in urban areas. They are also the main tax payers in the country. Maintaining rural investment while also increasing urban investment leads faster agricultural growth relative to the baseline by 0.14 percentage points per year (Figure 2.7). This is similar to the agricultural growth outcome in the Faster Urbanization scenario when rural public investment was also unchanged from baseline levels. By preventing a deceleration in agricultural GDP growth, the Rural Investment scenario leads to higher overall GDP growth, partly due to positive growth linkages between rural and urban economies. Both rural and urban GDP growth is higher when rural agricultural productivity does not fall. In fact, secondary towns are the main beneficiaries from maintaining rural investment levels, since the economy of towns is more closely linked to rural agriculture. Town households also tend to be poorer than city households and so food purchases and real food prices are more important for the town economy. Preventing a fall in agricultural productivity leads to faster rural-to-urban migration, and most of which occurs within secondary towns (Table 2.8). In the Win-win Investment scenario there is no decline in the welfare of poor urban households relative to the baseline, despite the even larger increase in urban migration. Faster urbanization on its own reduced poor urban household welfare by 1.3 percent, and increasing urban investment by reducing rural investments (Urban Investment scenario) caused this to fall even further by 3.1 percent (see Figure 2.8 above). When rural investment is maintained at baseline levels, poor rural households are also better off. Only non-poor urban households are worse off when taxes are raised to finance urban investments. However, even with higher taxes, non-poor urban households are still the main beneficiaries from faster urbanization – their per capita welfare is 38.1 percent higher in 2030 than it would have been without faster urbanization, and this is only slightly below the 46.4 percent welfare improvement they would have enjoyed had they not had to finance urban investments. The results from the final scenario indicate that raising taxes on urban households in order to finance urbanization and urban investment leads to win-win outcomes. There is faster economic growth, positive structural change, and improved national welfare. A tax-financed investment strategy prevents an urbanization of poverty, while also promoting agricultural transformation and reducing rural poverty. These positive outcomes are achieved with urban economies financing their own development. 2.5. Conclusion Economic growth over the last 15 years was mainly caused by workers moving out of agriculture into higher productivity jobs – a process known as “positive structural changeâ€?. Structural change is typically associated with urbanization and long-term economic development but most non-farm job creation in Malawi occurred in rural areas, which reflects Malawi’s small urban population and slow 31 urbanization process. This does not, however, mean that urban centers and urbanization could not contribute significantly to future economic development. A dynamic economy-wide analysis suggests that faster urbanization can be a potential catalyst for long-term economic development in Malawi. Three alternative implications for Malawi’s national development strategy can be drawn from the urbanization and public investment scenarios: First, slow agricultural growth might suggest that public investments should be directed towards agriculture in order to benefit the majority of Malawi’s population, who are smallholder farmers living in rural areas (Diao et al. 2010). Although rural poverty is falling, it is still more pronounced than in urban areas. Raising farm incomes should directly benefit the poor and allow farmers to diversify into rural non-farm activities (Benin et al. 2012; World Bank 2014). This agriculture-oriented perspective lies at the heart of Malawi’s first two Growth and Development Strategies (MGDS), which identified the country’s policy priorities for 2006-2016 (GOM 2006; 2011). The MGDS gives particular emphasis to agriculture, food security, and investments in rural villages. A stated goal in both MGDS’s is to reduce the pace of urbanization, which is viewed as a constraint to future economic development in Malawi. Secondly, faster nonagricultural growth over the last decade and a half might suggest that there is greater growth potential within the urban economy. An alternative strategy would direct more investments towards major cities in order to accelerate the pace of nonagricultural growth, create more non-farm jobs, and reinvigorate urban poverty reduction. Urban centers can also generate positive agglomeration economies that arise when workers and economic activities are spatially concentrated (Dorosh and Thurlow 2014). An urban-oriented strategy might encourage more people to migrate to cities where poverty and presumably the per capita cost of providing improved public services are lower. An urban-oriented strategy might also be better at promoting industrialization and economic transformation and may be more consistent with the aspirations of Malawi’s large youth population. Contrary to the thinking behind the MGDS, urbanization might promote rather than constrain future economic development. Finally, an intermediate strategy might be to invest in small towns, such as through road infrastructure and information systems. Small town development could strengthen the linkages between agriculture and other sectors (via product markets) and provide opportunities for farmers to diversify incomes (via temporary migration and local labor markets). Developing small towns might lead to incremental improvements along agricultural value-chains, while also providing a more gradual employment transition for rural-to-urban migrants. The MGDS prioritizes investment in “rural marketsâ€? rather than small towns, although there is often little distinction between the two in Malawi. Each of the three strategies has its strengths and limitations, and it is important to understand the relative merits and potential trade-offs between investment options for a national development plan. The comparative analysis of urbanization and public investment scenarios cautions against reducing investments in rural agriculture in order to finance urban development because this leads to an “urbanization of povertyâ€? that underpins many of the concerns about urban migration in Malawi. Instead, it is recommended that urbanization finances itself by increasing the urban tax base at the level where higher-income urban households still benefit from urban development. In discussing the ways to materialize the desired scenario, it is imperative to examine the current institutional and financing arrangements of local governments, which will be the focus of the next chapter. 32 3 Local Government Functions and Financing 3.1. Introduction This chapter aims to provide a descriptive and analytical overview of the local and intergovernmental institutional and fiscal system in Malawi, particularly in relation to urban management and service-delivery. In addition to the duplication of functions across the levels of governments or between institutions, limited capacity of local governments to efficiently deliver infrastructure and services remains a prime challenge for meeting current and future challenges of urbanization. This is largely related to the low level of resources available at the local government level and weak institutions for revenue generation. In view of this, the chapter concludes by making proposals to strengthen the system in order to better manage Malawi’s urbanization and urban development. One caveat to what follows is the limited availability of information and documentation on local government, inter-governmental relations, and urban development issues.27 While official documents – such as laws and regulations – are hard (or sometimes impossible) to locate, a good deal of the operational framework for local government in Malawi is less clearly defined by formal procedures than by “inheritedâ€? practice and unwritten rules. 3.2. Decentralization and Local Government System 3.2.1. Decentralization Following the establishment of multi-party politics in 1993 and the promulgation of a new constitution in 1994, the principle of elected local government was enshrined. Decentralization returned to the policy agenda after a long period of one-party rule and centralized political authority during Hastings Banda’s regime (1964-93).28 Post-1993 decentralization and representative local government in Malawi were clearly seen (by many Malawians and development partners) as part and parcel of a wider democratization process, as well as an important step towards improving infrastructure and service delivery through greater public participation and increased accountability. The constitutional commitment to local government was followed by the adoption of a National Decentralization Policy (1998) and, shortly afterwards, by the passing of the Local Government (LG) Act (1998). Local government elections were held in 2000; elected Assemblies assumed responsibility for the management and oversight of local government. As articulated in the NDP, Malawi’s decentralization policy is far-reaching and ambitious. In its own words, the NDP: 27 Examples of this problem include: the absence of published, readily available or official documentation on local government finance in general; the almost complete absence of any documentation on Local Development Fund (LDF) operations and activities; and the general paucity of official regulations and procedures. This chapter therefore relies on data and information that has been obtained through a large number of meetings with central and local government officials. Wherever possible, copies of legal and regulatory documents were used, but in some cases the chapter has had to rely on verbal hearsay with respect to statutory procedures. The National Local Government Finance Committee (NLGFC) does not publish reports on local government finance but kindly provided unofficial (and sometimes incomplete) spreadsheet files on Local Government finance issues. Finally, the chapter relies heavily on the hands-on experience and knowledge of one of its authors, who was (for many years) a senior member and then CEO of Blantyre City Council, as well as an active participant in recent local government policy discussions in Malawi. 28 For further information and analysis of local governance in the colonial and Banda periods, see the chapter by Chiweza in Tambulasi, R. (ed) (2010): Reforming the Malawian Public Sector, Council for the Development of Social Science Research in Africa) and Cammack, D., Kanyongolo, E. & O’Neil, T. (2009): ‘Town Chiefs’ in Malawi, APPP Working Paper No.3, ODI.; and Joint Malawi Government/Donor Review Team (May 2004): Review of the National Decentralization Program of Malawi. 33 ï‚· Devolves administration and political authority to the district level; ï‚· Integrates governmental agencies at the district and local levels into one administrative unit, through the process of institutional integration, manpower absorption, composite budgeting and provision of funds for the decentralized services; ï‚· Diverts the center of implementation responsibilities and transfers these to the districts; ï‚· Assigns functions and responsibilities to the various levels of government; and ï‚· Promotes popular participation in the governance and development of districts. However, in its early years (1998-2004), progress on implementing this ambitious decentralization policy was modest and effectively came to a hold, when the Government called off scheduled local elections in 2005. A combination of institutional inertia, vested and competing interests, and a lack of incentives slowed down the initial process of decentralization. Furthermore, during the period 2005-2013, locally elected councils were replaced by un-elected consultative committees appointed by the Central Government and recentralization took place. Underlying this were a number of factors: political consolidation under President Bingu wa Mutharika, the ruling party’s concerns that local elections might provide opposition parties with a power base, and the continued reluctance of central government agencies to devolve authority, functions and resources to local government. Unsurprisingly, this decade-long hiatus has further stalled any progress made on decentralization after the mid and late 1990s. After a series of postponements, new local elections were held in May 2014 (at the same time as presidential and parliamentary elections), restoring formal electoral accountability to the local government system. The Government is also currently in the process of reviewing the 1998 LG Act. As far as can be understood, however, most revisions will be related to “politicalâ€?, council and human resource issues – rather than to any rethinking of urban/rural distinctions, local government functions and finance, which will be discussed further in the following sections. 3.2.2. Local Government System The current system of local government in Malawi is structured and functions in accordance with constitutional provisions, the Local Government Act of 1998, a significant 2010 amendment to the Act, and a host of other sectoral and thematic laws. Both the Constitution and the LG Act provide local governments with broad and wide-ranging political, administrative and developmental responsibilities. In terms of development, local governments are responsible for local planning, as well as for general infrastructure and service delivery. Local government policy, oversight and support are in the hands of the MLG&RD. Local government finance, on the other hand, is the responsibility of the National Local Government Finance Committee (NLGFC), a constitutionally mandated body, which decides on local revenue sources, determines the scope and scale of inter- governmental fiscal transfers, exercises oversight with respect to local government budgets and submits a consolidated annual local government budget29 to the Ministry of Finance. There is only one tier of local government, currently consisting of 35 Councils: 28 District Councils, 4 City Councils, 2 Municipal Councils and 1 Town Council. The three administrative regions (central, northern and southern) that divide up Malawi’s national territory are not sub-national governments, do not have elected councils (or administrative heads), and largely serve as geographical units for the purposes of deconcentrated line ministries (Figure 3.1). Each local government area is under the 29 Based on draft LG budgets submitted by Councils, themselves partly based on indicative budget allocations calculated by the NLGFC. 34 authority of a local Council.30 Annex II presents local government areas with their population size and number of elected councilors. Figure 3.1 map of Malawian Districts and Cities Source: NSO: Statistical Year Book 2012 Local government councils consist of elected and ex-officio members, and are headed by Chairpersons (in the case of Districts) and Mayors (in the case of urban local governments). Elected councilors represent individual wards on the basis of first-past-the-post elections, either as party or as individual candidates, and serves five-year terms. Voting ex-officio members of councils include 30 Prior to its 2010 amendment, the LG Act referred to Councils as Assemblies. 35 the Members of Parliament (MPs) whose constituencies lie within the local government jurisdiction in question. There are five ex-officio non-voting representatives of special interest groups (e.g. women, youth), selected by other locally elected council members. Locally elected councilors are typically the largest group in a City Council.31 Council heads, elected by their respective councils from amongst their membership, are "mayors-in-councilâ€?, with no executive functions (which are the responsibility of Chief Executive Officers in urban LGs and District Commissioners in rural LGs). Both rural and urban Councils are assisted by Secretariats, headed by District Commissioners (DCs) and Chief Executive Officers (CEOs), respectively. Both DCs and CEOs are appointed by the Ministry of Local Government and Rural Development (MLG&RD). Other City Council Secretariat staff (with the exception of the Education Department Manager, who is on the Ministry of Education payroll) are recruited locally and paid directly out of the City Council budget. District Council Secretariats, on the other hand, rely on line department staff, paid by their respective line ministries, and on a small number of directly hired and paid core staff. City Council Secretariats operate as local executives and consist of a number of technical departments for planning, administration, public works (or engineering services), finance, education, health, commerce and other functions (e.g. parks, recreation and environment). They work alongside ministry line departments and other public sector agencies to provide infrastructure and services. In accordance with the LG Act, LG Councils operate on the basis of a number of statutory committees (at least six including finance, development, education, health, public works and administrative), in which both elected councilors and full-time Secretariat staff work together. Council committees are responsible for taking day-to-day decisions, overseeing activities and preparing submissions to full Councils for approval. In cities, Council Secretariats and their departments are responsible for day-to-day management. In principle, Councils (as local legislatures) direct and oversee the activities of their respective Secretariats. The designation and number of Secretariat departments are similar across City Councils. On the other hand, secretariat technical departments, in principle, are responsible for implementing decisions reached and policies made by the Council (and its service committees). When measured in terms of total expenditures and revenues, local government is a small player in the overall Malawian public sector and, in recent years, appears to have declined in importance, as can be seen from Table 3.1. Local government own-source revenues account for only about 0.3 percent of GDP and less than two percent of central government revenues. Over the period 2011- 2014, local government revenues have become even less important, declining from 1.7 percent of central government revenues in 2011-2012 to about 1.2 percent in 2013-2014. In terms of expenditure, local governments account for considerably less than two percent of GDP (declining from 1.68% of GDP in 2011-2012 to 1.22% in 2013-2014); and amount to less than six percent of central government spending (declining from 5.5% in 2011-2012 to 3.3% in 2013-2014). 31 In Mzuzu, for example, the voting membership of the City Council consists of 15 locally elected councilors and one MP; in Blantyre, the voting membership consists of 23 locally elected councilors and eight MPs; and in Lilongwe, of 27 local councilors and four MPs. 36 Table 3.1 local government revenues and expenditure (2011-2014) Revenue & expenditure FY 2011/12 2012/13 2013/14 Kwacha % of LG as % Kwacha % of LG as % Kwacha % of LG as % (billions) GDP of CG (billions) GDP of CG (billions) GDP of CG finance32 finance finance Public revenues Central Govt. - tax & non-tax revenues* 214.0 22.10 297.0 24.50 441.0 28.00 Local Govt. - tax & non-tax revenues** 3.6 0.34 1.7 4.0 0.28 1.3 5.5 0.32 1.2 Public expenditure Central Govt: - total expenditure* 324.0 33.40 489.0 40.50 647.0 41.00 - current expenditure (including grants to LGs)* 246.0 25.40 385.0 31.90 547.0 34.70 - development expenditure* 78.0 8.00 104.0 8.60 100.0 6.40 Local Govt: - total expenditure (including CG grants)** 17.7 1.68 5.5 22.5 1.59 4.6 21.3 1.22 3.3 - total expenditure (less grants from CG)** 2.1 0.20 3.6 0.25 4.4 0.25 * = IMF and World Bank data ** = NLGFC and IMF/WB data 32 CG finance refers to all central government revenues or expenditure, as voted by Parliament and executed by central government ministries and agencies. 37 Malawian local government revenues and expenditure are significantly smaller than is the case in other developing countries (Table 3.2). Clearly, caution is required in comparing local government revenues and expenditure patterns across countries – since LG revenues (as a percentage of GDP and total public revenues) will vary depending on revenue assignments, while the size of LG expenditures will depend on functional assignments. Nonetheless, the comparison does indicate that Malawian local governments spend or raise proportionately less than LGs in other countries, reflecting their relative unimportance when compared to central government. The very small fiscal footprint of local government in Malawi seems to belie the intent of national policy statements and frameworks. Table 3.2 LG revenues & expenditure – international comparisons Measure LGs in Malawi LGs in developing (2011-2014) countries (in the 2000s) Total expenditure % of total public expenditure 3.3-5.5 18.8 % of GDP 0.2-0.3 5.1 Total % of total taxes/revenues 1.2-1.7 11.4 taxes/revenues % of GDP 0.3 2.3 Source: Bahl & Bird (2013); Bahl & Bird use data for 16-20 (unspecified) developing countries, based on IMF Government Finance Statistics. However, since the elections in May 2014, the policy environment for decentralization has become considerably more favorable – in response to less authoritarian rule, widened democratization and donor lobbying. Very importantly, policy-makers appear increasingly willing to provide local government with more “teethâ€? in the form of greater fiscal resources. The 2015-16 budget presented to parliament allocated additional resources to local governments – notably, by including (for the first time) a development component (of MWK 5 billion) in the local government General Resources Fund (GRF) and by allocating MWK 6.5 billion to urban councils to finance road rehabilitation and upgrading. In addition, as part of a wider and ongoing public sector reform, moves are underway to provide local government with greater control over local human resources. 3.2.3. Urban Local Government Urban local government in Malawi currently consists of four City Councils, two Municipal Councils and one Town Council. Prior to 2010, the hierarchy of urban local governments included 12 units: the four City Councils and eight Town Councils. Two of the Town Councils (Luchenza and Kasungu) have since been “elevatedâ€? to the status of municipalities, while the remaining six have been subsumed into District Councils33. In 2012, after local lobbying, Mangochi was re-instated as an urban LG, becoming the only Town Council in the LG system.34 The initial “de-municipalizationâ€? of smaller towns (and their subsequent incorporation into District Councils) appears to have been justified in terms of efficiency, since establishing a recommended standard structure of directorates for Town Councils (formerly called 33 It should be noted that despite their incorporation into District Councils, these former Towns retained their classification as “rateable areasâ€? and thus remain subject to property taxation. 34 It is not entirely clear why Mangochi residents lobbied for town council status. Media reports imply that local citizens perceive town councils as being more likely to address issues such as solid waste management, local roads and urban amenities, in turn seen as important to tourism (for which Mangochi is perceived as a service center). See: http://allafrica.com/stories/201407120496.html and http://allafrica.com/stories/201404240085.html. 38 Assemblies) has incurred disproportionately high overhead costs as compared to relatively small areas and populations that they cover.35 Table 3.3 Urban population and urban local government (2008 Census) Urban LG area Population (2008) % of total Population density (per km2) Lilongwe City Council 669,021 40.8 1,479 Blantyre City Council 661,444 40.4 3,006 Mzuzu City Council 128,432 7.8 2,791 Zomba City Council 87,366 5.3 2,264 Kasungu Municipal Council 42,351 2.6 n/a Luchenza Municipal Council 10,751 0.7 n/a Mangochi Town Council 39,369 2.4 n/a Total population in urban LG areas 1,638,734 100.0 Total urban population in Malawi 2,003,309 The seven urban local governments account for just over 80 percent the total urban population of Malawi. The four City Councils are by far the largest in terms of population; and, of these, Lilongwe and Blantyre account for over 80 percent of the total urban LG population. Notably, almost 400,000 Malawians, then, live in urban settlements that are subsumed within district (or rural) jurisdictions. Indeed, there are 18 “non-municipalizedâ€? settlements that are bigger than Luchenza Municipal Council (see the table below). Table 3.4 Urban settlements in Malawi No. Urban Urban settlement Population (2008) Rank in 2008 ULG status settlement classification 1 Lilongwe Primary 674,448 1 ULG 2 Blantyre Primary 661,256 2 ULG 3 Mzuzu Primary 133,968 3 ULG 4 Zomba Primary 8,314 4 ULG 5 Mangochi Secondary 50,821 5 ULG 6 Karonga Secondary 40,334 6 7 Kasungu Secondary 39,640 7 ULG 8 Salima Secondary 27,852 8 9 Nkhotakota Other 24,726 9 10 Liwonde Secondary 23,463 10 11 Balaka Secondary 22,733 11 12 Mzimba Other 20,994 12 13 Dedza Secondary 20,241 13 14 Nsanje Other 20,179 14 15 Thyolo Other 18,589 15 16 Mchinji Other 17,881 16 35 Joint Malawi Government/Donor Review Team (May 2004): Review of the National Decentralization Program of Malawi 39 17 Rumphi Other 17,845 17 18 Mponela Other 15,399 18 19 Chitipa Other 14,753 19 20 Ntcheu Other 14,642 20 21 Mulanje Other 14,487 21 22 Mwanza Other 14,226 22 23 Monkey Bay Other 11,576 23 24 Nkhata Bay Other 11,265 24 25 Luchenza Secondary 11,207 25 ULG 26 Ntchisi Other 7,918 26 27 Ngabu Other 7,523 27 28 Dowa Other 7,408 28 29 Chikwawa Other 6,987 29 30 Chiradzulu Other 2,348 30 31 Machinga Secondary 1,180 31 Total population36 2,044,203 Lilongwe has been the national capital since 1975 and is now the largest city in the country in terms of population size, having “overtakenâ€? Blantyre in 2008. Lilongwe is an administrative center, the headquarters for the majority of national ministries and institutions, and also a services and agro- processing hub. Its growth has been strongly driven by its status as the national capital, in which almost all central government head offices are now located. Lilongwe’s growth has also been due to its role in servicing a large agricultural area. Blantyre, established by missionaries in the 1870s, is the oldest city in the country and is traditionally portrayed as Malawi’s manufacturing and commercial capital, serving a distinct regional/international hinterland. Until recently, Blantyre was the largest city in Malawi; although its population continues to grow, it currently does so at a slower pace than Lilongwe. Mzuzu serves as an administrative and commercial service center for northern Malawi. Alongside Lilongwe, Mzuzu is the fastest-growing city in Malawi with its population increasing by 4.4 percent annually between 1998 and 2008 and its surface area expanded six fold since the late 1940s. It is the hub of government administration, business, industry, commerce, and services for the northern region of Malawi. Zomba, the original national capital, is best-described as a large university town37 and old center of administration, serving a relatively small hinterland. Malawi’s secondary and other urban settlements are a combination of administrative centers, minor agricultural processing nodes, and market and transport hubs. Some of these smaller or emerging towns are growing rapidly and include Mangochi, Kasungu, Thyolo and Mchimji (NSO 2008). These towns are at the frontline of the rural-urban interface in Malawi – retaining much that is agrarian alongside an increasingly visible urban appearance and economy. There are no hard and fast or consistent definitions of what kinds of urban settlement “qualifyâ€? as urban local governments, thus leaving plenty of room for arbitrary decisions about LG status. There is 36 This figure for the total population of urban settlements is greater than the figure for the total urban population. Both figures appear in official 2008 census reports. 37 Chancellor College, located in Zomba, is the largest of the five constituent colleges of the University of Malawi. Five university faculties (humanities, science, education, law and social sciences) make up Chancellor College. 40 clearly a high degree of inconsistency in having a hierarchy of urban local governments that includes Luchenza Municipality (classified as a “secondaryâ€? settlement although considerably smaller than most “otherâ€? urban settlements) but excludes much larger towns (such as Karonga, Liwonde and Salima). It is also unclear that there are particular advantages for local citizens of towns like Liwonde if they were to become municipalities; conversely, there are no obvious ways in which the particular urban needs of a town like Liwonde are (or can be) met through being a part of a larger District. The existing legal and regulatory framework also makes little substantive distinction between urban and rural local government. Most distinctions are terminological or minor: e.g. “City, Municipal or Town Councilsâ€? as opposed to “District Councilsâ€?, CEOs as opposed to DCs, and urban mayors having longer terms than their rural counterparts. There are few formal differences in the responsibilities and powers ascribed to urban as opposed to rural local governments or in the procedural framework defining their activities or operations. In addition, there are no clear criteria for distinguishing between urban and rural jurisdictions. Classifying local government areas as urban (rather than rural) is at the discretion of the Minister of Local Government and Rural Development (in the case of municipalities and towns) or the President (in the case of cities).38 The lack of any substantive and formal distinction between urban and rural local governments can largely be attributed to an overwhelming policy focus on rural areas. Decentralization policy in Malawi is largely formulated in terms of Districts and pays relatively little attention to City, Municipal, or Town governments. This is understandable, given that over 80 percent of Malawians live and work in rural areas. However, the rural “biasâ€? that underpins decentralization policy does not mean that urban local governments operate as if they are more or less the same as district councils.39 In practice, urban local governments face rather different challenges to their rural counterparts – ensuring the provision of networked services (such as water and electricity supply) in high density residential areas, dealing with complex land tenure arrangements, providing transport infrastructure for home-to-work travel by city residents, managing environmental health issues in high density urban neighborhoods, facilitating industrial and commercial development. City and town governments tend to be neglected in Malawi’s institutional landscape and are thus subject to a “one size fits allâ€? approach. In practice, there are a number of significant differences in the ways that urban and rural local governments are structured and operate, inter alia: ï‚· Urban Councils do not include traditional chiefs as ex-officio members; rural councils include them. Whereas the provisions of the LG Act (1998 and 2010 amendment) do not clearly articulate on the membership of traditional chiefs in City Councils, the provisions of the Chiefs Act (1967) specify that traditional or customary authorities have no jurisdiction in urban areas. ï‚· The Mayors of Cities (and, to a lesser extent, Towns) tend to have a much higher political profile and more (but still limited) authority than their counterpart (Chairs) in Districts. This distinction has no basis in formal regulations – and would appear to be largely attributable to the “customaryâ€? or 38 According to officials in MLG&RD, the Government is currently in the process of establishing criteria for deciding on whether local government areas should be categorized as urban or rural. Subject to discussions, these will be incorporated into a revised LG Act. However, no agreement has yet been reached on the criteria (population density, population size, economic characteristics) that would be used to demarcate urban jurisdictions. 39 Indeed, in its only reference to urban local governments, the National Decentralization Policy (1998) states that Cities, Municipalities and Towns will be “districts in their own rightâ€?. 41 ceremonial status of mayors40, the greater prestige attached to urban councils, and the common perception that mayors are better educated and better connected than their District counterparts. ï‚· District Commissioners, the most senior civil servants in any District, appear to be administratively more influential and important than their CEO counterparts in urban local government areas. ï‚· Urban local governments directly employ many more staff than do District Councils. In “devolvedâ€? sectors (other than education), urban local governments employ their own staff and manage sectors in parallel to or separate from line departments. In rural areas, districts do not run parallel services, but instead rely on cooperation and support from line departments and line ministry staff. ï‚· Urban local governments also seem to be much less concerned by sector devolution policies and indeed receive very few conditional transfers for any “devolvedâ€? sectors, with the exception of education sector grants. Rural councils, on the other hand, receive conditional (or sector) grants for managing devolved functions (such as health, agricultural extension, water supply). A degree of ambiguity exists in urban areas as to who is responsible for the provision of services in devolved sectors: without any funding for specific sectors, urban local governments may not feel obliged to provide services, while line departments may assume that urban governments are responsible. ï‚· Finally (and as will be discussed in more detail), urban local governments rely far more on own- source revenues to finance their activities than do Districts. 3.3. Functional Assignments, Performance and Expenditure Patterns 3.3.1. Local Government Functions The Local Government Act (1998) and National Decentralization Policy (1998) assign a wide range and large number of functions to both urban and rural local governments, which are not differentiated and apply uniformly to both types of local government. They include functions related to security, regulation of public areas and general public goods, vital registration, infrastructure provision, social and economic services, and other orthodox local government activities. The table below provides a summarized listing of local government functional assignments (drawn up on the basis of Schedule 2 of the LG Act and the National Decentralization Policy), broadly linked to planning, infrastructure and service delivery (LG Act, Article 6). Table 3.5 Local government functional assignments FUNCTION PRINCIPAL SUB-FUNCTIONS LG Act (Second Schedule) National Decentralization Policy Planning ï‚· Development planning ï‚· Local Assembly designated as a ï‚· Physical and land use planning planning authority Health ï‚· Management of clinics, maternity ï‚· Health centers, dispensaries, clinics, health centers and maternity clinics and health posts dispensaries ï‚· Control of communicable diseases ï‚· Health education ï‚· Environmental sanitation Education ï‚· Management of schools ï‚· Nurseries and kindergartens ï‚· Primary Schools 40 Mayors of Malawian Cities (but not District Council Chairs) are, on formal occasions, referred to as “Your Worshipâ€?, reflecting the country’s colonial legacy. 42 ï‚· Distance Education Centers Waste ï‚· Collection, removal and treatment ï‚· Refuse disposal management of solid and liquid waste ï‚· Sewerage removal and disposal Water & ï‚· Sanitation ï‚· Provision and maintenance of sanitation ï‚· Provide for the supply of potable water supplies including: water - Boreholes - Piped water projects - Protected wells - Distribution of water - Gravity-fed piped water schemes Roads and ï‚· Construction and maintenance of ï‚· Construction, rehabilitation and transport roads (subject to provisions of the maintenance of roads not under Public Roads Act and other sector Central Government including: laws) - District roads ï‚· Street lighting - Township and City roads ï‚· Public transport - Estate roads Agriculture & ï‚· Management of agricultural ï‚· Agricultural and livestock extension natural resources extension services services ï‚· Veterinary & animal health services ï‚· Irrigation dams Market and retail ï‚· Management of markets and facilities market buildings and premises ï‚· Charge market fees Trade & ï‚· Trade development services ï‚· Licensing of small and medium commerce ï‚· Business licensing business Emergency ï‚· Ambulance and fire service ï‚· Ambulance and fire services services ï‚· Disasters (until they are declared national disasters) Public Amenities ï‚· Management of parks and public ï‚· Sports stadiums spaces (including sports & ï‚· Community halls recreational facilities, cemeteries, ï‚· Recreational parks and playgrounds etc.) ï‚· Public conveniences ï‚· Burial services Buildings and ï‚· Regulation & control of building ï‚· Control of building standards structures and construction Land & housing ï‚· Acquisition of land, lay out of ï‚· Housing estates housing areas, provision of residential services, construction & management of housing ï‚· Land management and surveying Other ï‚· Environmental reclamation ï‚· Environmental education 43 It is unclear from the LG Act or NDP how far this exhaustive assignment of functions is mandatory or exclusive or whether the intent is to spell out “permissiveâ€? functions, responsibilities that might be taken on by local governments. If the former, the LG Act or NDP would appear to be over-ambitious and unrealistic. If the latter, the law and policy lead to a high degree of ambiguity. Formally speaking, the functions assigned to local governments overlap extensively with sector/other laws and regulations. The Ministry of Education, for example, has much the same responsibilities as do local governments in the education sector. The same applies to other service delivery sectors. This formal functional overlap between line departments or other national institutions and local government (which is by no means unique to Malawi) is problematic insofar as: (i) accountability becomes blurred and; (ii) inefficient duplication can take place. 3.3.2. City Councils and Urban Infrastructure and Service Delivery In practice, it is clear that local governments do not provide all the public goods and services that they are mandated to by the LG Act or NDP. Indeed, urban local governments are actually only directly involved in a sub-set of their assigned functions and, in many cases, do so along with other service providers or agencies. For major infrastructure and service sectors, notably, education and health, the share of spending accounted for by local government is relatively small as most line ministry spending remains centralized. Box 3.1 Centralized and deconcentrated sector spending Line ministries: weak levels of decentralized spending “Most ministries centralize their spending and do not have an important share of their budget deconcentrated to districts, with the exception of the Ministry of Health … only one ministry, the Ministry of Health, which has given some spending authority to districts, has deconcentrated a large share (about 40%) of its budgetary resources [although none to City Councils]. The Ministry of Education has deconcentrated 4%, the Ministry of Agriculture less than 2%, and all the other ministries together 0.4%. The total amounts to about 6% of total ministerial expenditures.â€? Source: World Bank (2012): Malawi - A Snapshot of Decentralization in Malawi: An Overview of Progress, Challenges and Policy Options in 2010 and 2011, Public Sector Reform and Capacity Building Unit (AFTPR) Africa Region. Education and health sectors: budget devolution “Part of the education sector budget is devolved to local councils. In 2012-13, 6% of the education budget was devolved to the district level for [Other Recurrent Transactions, or non-payroll related] ORT of primary education. Local [government] councils [LGCs] are only responsible for the ORT part of recurrent expenditures for primary schools. Salaries for primary school teachers and key teaching and learning materials for primary education are still centrally procured. […] Most of the public [health sector] funds – tax financed and external pooled financing – flow through the Ministry of Health (75%), followed by the local government councils (25%). […] There is clear separation on what budget items are covered between the MOH headquarters and the other agencies and levels. MOH takes on the stewardship and management role. MOH headquarters finances PE [payroll and emoluments] for all of public health sector, and also finances some public health goods, such as vaccines. Additionally, MOH headquarters manages the larger capital investment projects, both externally and domestically financed. The responsibility for ORT is shared equally between MOH and its agencies and the LGC. Over time, there is a shift of responsibility for operations to LGC, which use about 35% of this ORT for drugs and supplies, and 65% for operations costs.â€? Source: World Bank (November 2013). Malawi Public Expenditure Review, Report No. 79865 44 Roads Formal institutional arrangements in the Malawian roads sector appear clear. As in most countries, responsibilities for road construction, rehabilitation and maintenance are divided up according to the classification of roads. Consistent with the Public Roads Act (1962), the Roads Fund Administration Act (2006), the Roads Authority Act (2006), the Local Government Act (1998), and the Urban Public and Private Streets Act (1956), the designated roads network consists of five categories of road: ï‚· Main roads are inter-territorial roads outside cities or towns, unilaterally designated by the Government and providing a high degree of mobility connecting major settlements and/or serving as international corridors. These are the responsibility of the national Roads Authority. ï‚· Secondary roads consist of roads outside cities or towns linking the main population and economy, to the main road network. These are the responsibility of the national Roads Authority. ï‚· Tertiary roads are outside cities or towns, acting as collector roads to arterial roads, accommodating shorter trips, and feeding the arterial road network. These are also the responsibility of the national Roads Authority. ï‚· District roads are outside cities or towns and provide intermediate levels of service, connecting districts, local centers of population and developed areas with the principal arterial system. These roads are the responsibility of District Councils. ï‚· Urban roads, consisting of other roads in urban area, provide accessibility over relatively short trip lengths at low speeds and services to smaller communities. Urban local governments are responsible for these roads. In addition to the designated roads network, the public roads network also includes a large number of community roads, in both rural and urban areas. Responsibility for community roads lies with local communities and local governments. In total, the designated network consists of about 15,000 km of roads (of which about 25% is paved, the rest being earth or gravel roads); and community roads (all of which are earth or, at best, gravel) make up a further 10,000 km of undesignated roads. In principle, City Councils are exclusively mandated to build, rehabilitate, upgrade and maintain urban roads and share responsibility for community roads in their jurisdictions. In practice, this clarity of assignment is muddied by finance and “capacityâ€? issues. Financially, City Councils do not have enough funds to ensure much in the way of either road construction or maintenance. As will be seen below, their capital budgets are very limited – partly because overall revenues are relatively small but also because most resources are spent on recurrent expenditure items – which explains the limited involvement of City Councils in major road construction or rehabilitation activities. As far as is known, City Councils only take on construction or rehabilitation works when development partners chose to provide financing.41 This lack of investment has clearly resulted in underdeveloped urban road networks and poor access to low-income areas. Most City Councils are involved in some kind of urban road maintenance but a very small proportion of roads are regularly and properly maintained. In Blantyre, for example, the public works department undertakes sporadic maintenance of minor urban roads – usually grading and occasional repairs – and 41 As is currently the case in Blantyre, where JICA is providing the finance for upgrading a major city road. However, this is off-budget funding and is being largely managed by international technical assistance, rather than by the City Council’s public works department. 45 estimates that only about 10 percent of urban roads are maintained every year. The same applies to Lilongwe and Mzuzu. In general, the maintenance of urban roads is undertaken on a force account basis, relying directly on City Council equipment and labor, rather than through contracting with the private sector. In all cases, City Councils are poorly equipped (Blantyre, Lilongwe and Mzuzu each have one functioning grader/bulldozer), but do have plenty of more or less full-time road workers available (the public works departments of Blantyre, Lilongwe and Mzuzu employ – respectively – about 200, 350 and 110 staff). Whatever maintenance does take place, then, is limited to intermittent labor-based work.42 As a result of all this, urban roads are in poor shape and now require costly rehabilitation. Given the financial and other constraints faced by City Councils in road construction and maintenance, the national Roads Authority has regularly and frequently taken on urban road activities.43 The Roads Authority, with external and Government financing, has been de facto responsible for the construction and/or rehabilitation of several main urban roads. It has also carried out maintenance of some urban roads, using finance provided by the Roads Fund Administration (RFA).44 In all such cases, the Roads Authority outsources actual works to private sector contractors. Although the RFA is mandated to provide local governments with funding for road maintenance, this has not been the case to date. Local governments, including City Councils, can qualify as “road agenciesâ€?, thus becoming eligible for RFA funding. On the basis of capacity assessments by the RFA, however, City Councils have not been rated as “road agenciesâ€? and are not able to access RFA funds for the maintenance of their urban roads. Electricity In Malawi, local governments have no legal mandate to provide electricity generation or supply services. Electricity supply and distribution are the exclusive function of the Electricity Supply Corporation of Malawi (ESCOM). This is consistent with utility management in many (but not all) countries and is the case in practice. Although City Councils play no direct role in providing electricity, they may need to get more involved in planning for electricity distribution (and thus coordinate network extensions or upgrades with ESCOM). The 2010 Lilongwe Urban Development Master Plan, for example, makes no reference to electricity, despite the fact that electricity is obviously a key urban utility. Water Supply Although the LG Act does assign water supply and sanitation functions to urban local governments, the preponderant responsibility for water supply management in Malawi’s cities now lies with parastatal Water Boards (WB): Blantyre City WB, Lilongwe City WB, the Northern Region WB (for Mzuzu and other towns) and the Southern Region WB (for Zomba and other towns). Practice on the ground reflects this – no City Councils are involved in the direct supply of water within their jurisdictions apart from being represented (by their CEOs) on the Board of Directors of each WB. City governments are expected to play a planning, coordination and monitoring role in the water sector. In addition, they are also directly involved in the management of a limited number of water kiosks in low income areas. Water Boards are responsible for all water supply networks and are expected to cover their costs. City Councils do not provide any capital funding for extensions or upgrading of the WB water supply 42 And probably associated with high levels of under-employment of staff or long periods of inactivity. 43 Much the same applies to District Councils. 44 The Roads Fund Administration revenues derive from a fuel levy, the proceeds of which are intended to finance maintenance of the designated road network. 46 networks. Although commercial and institutional clients account for a relatively small proportion of the client base of WBs (by connections)45, payments from them typically account for over 50 percent of all WB revenues; their residential clients (who account for the overwhelming majority of water connections) provide the remaining proportion of their revenues. WBs are – by local government standards – significant financial operations. Lilongwe and Blantyre WBs generated revenues of MWK 9.3 billion and MWK 7.7 billion in 2014-15, respectively, roughly three to four times greater than the total revenues of their City Council counterparts. Solid Waste Management Solid waste management (SWM) is one of the main urban services for which local governments have a near exclusive mandate. However, as bleakly illustrated in the box below, SWM in Malawi’s main cities is woefully inadequate – in terms of coverage, efficiency, and actual waste disposal – with low-income areas and poor communities being particularly underserved. Box 3.2 Solid waste management in Lilongwe and Blantyre: snapshots Lilongwe “Of the total waste generated within the City, only about 30% is collected. Collection from … high income areas is undertaken regularly from individual households. In low income areas, households dispose of waste either in waste pits on their plots or at skip sites which are collected by the City on an infrequent basis. No waste is collected in the squatter areas and limited accessibility is cited as the main problem. It would appear that in 15 areas within the City, residents are required to dispose of their own refuse. The use of skips is the common service delivery level in 16 areas with household collection occurring in 25 areas. Illegal dumping, especially building rubble, is common and difficult to police. There are no service delivery levels which are determined for the city as a whole and or different settlement types. The [only landfill] site [in Area 38] is declared as a prohibited area. Currently, there is no separation of waste into waste for recycling and non-recycling. Scavenging is not allowed at the landfill site but due to a lack of fencing, the practice is difficult to prevent and is therefore common. The poor waste management within the City contributes to soil and water contamination. The existing by-laws are outdated making enforcement difficult. The City does not manage any public awareness campaigns for improving the management of solid waste within Lilongwe.â€? (See: Lilongwe City Development Strategy 2010-2015). Blantyre “A recent [2010] document estimates that Blantyre residents produce an average of 0.9 kg of waste per capita per day (equivalent to 647 metric tons per day)—81 percent of which is organic and biodegradable (Berman, 2010). According to UN-HABITAT, only a third of the solid waste generated is actually collected (UNHABITAT, 2011). Solid waste collection services are available in high-income areas, but low-income areas do not have access to the same services … informal settlements and Traditional Housing Areas are severely underserved. In these areas, solid waste is either not collected, or skips are emptied irregularly, leading to the accumulation of waste and the indiscriminate disposal of solid waste. [The City Council] has placed skips/containers in many residential areas and markets, but collection is irregular, 45 Commercial and institutional water supply connections accounted for about 7 percent of all (55,000) connections served by Lilongwe Water Board (LWB) in 2014 (personal communication, LWB officials, 22 July 2015). 47 particularly when there are fuel shortages or when vehicles break down. In peri-urban areas there is no regular solid waste collection system in place, partly because there are few access roads. As a result, household solid waste is often dumped into pits, drains or indiscriminately discarded in the streets. … a number of private entities providing waste collection services, …but they focus on commercial/industrial companies and houses in high-income areas, leaving poor communities underserved. The [City Council] charges private waste collectors fees to dispose of waste at the dumpsite, and as a result, many of these collectors dump waste in undesignated areas to avoid the fees. The official site for solid waste disposal is located along the eastern boundary of the city … and is nearing full capacity. The site is not a proper landfill, but rather an uncontrolled open dump without leachate or gas management systems and without a fence, so it is accessible to residents, including children. This is problematic because hazardous waste is not separated from other waste. Hazardous waste includes paints, solvents, consumer batteries, construction and demolition debris, chemical and pharmaceutical waste, medical and infectious waste, tires and sewage sludge.â€? (see Maoulidi, M. (2012): Water and sanitation needs assessment for Blantyre City, Malawi, Millennium Cities Initiative Social Sector Working Paper, Series N° 27/2012). City Councils continue to rely on their own in-house health departments (akin to using force accounts in the public works sector) for the direct production of SWM services. Lilongwe City Council’s SWM operations in 2015 are currently undertaken by several hundred sweepers and garbage collectors, who are responsible for cleaning public areas and actual waste collection from residential zones, and commercial and industrial clients (the latter two commonly pay additional waste collection fees for their on-site skips). The City has eight garbage trucks (all obtained through private or international donations), of which only four are currently in operation (the other four are either undergoing repairs or out of action). The City Council has no official SWM sub-contracts with private sector operators and undertakes no systematic regulation of non-public SWM operations – although there are several in Lilongwe, which function on a commercial (and apparently profitable) basis. The situation is very similar in Blantyre where City Council’s SWM operations in 2015 employ about 250 sweepers and cleaners. The City Council has a fleet of 11 trucks for SWM, of which only seven are currently operational. Although there are private sector garbage collectors that operate in Blantyre, none of them are either contracted or regulated by the City Council. Seeking ways of out-sourcing to the private sector and then regulating privatized waste collection are likely to be more efficient way of ensuring waste collection and disposal. In Lilongwe, a small private sector operator is known to collect seven percent of the total amount of municipal solid waste with one truck and four employees; whereas the City Council has 600 employees in the cleaning department (although not all working at collecting) and four trucks, and collects 30 percent of the total waste.46 Clearly, there is a huge difference in efficiency. Education and Health City Councils, like all other local governments, are nominally responsible for the provision of primary education services within their jurisdictions. This is in line with the LG Act, the NDP and the Ministry of Education’s own devolution policy. Indeed, the education sector grants (see below) allocated to City Councils make up the largest proportion (just over 40%) of all their inter-governmental fiscal transfers. 46 Barré, J. (2014): Waste Market in Urban Malawi – A way out of poverty?, Master’s Thesis, Department of Urban and Rural Development Swedish University of Agricultural Sciences, Uppsala. 48 Education departments within City Councils, however, are relatively small: in Lilongwe, less than 3 percent of all directly hired Secretariat staff are employed by the education department. In practice, City Councils play a largely delegated and passive role in primary education and have limited expenditure responsibilities. Recurrent expenditure in the primary education sector is dominated by payroll spending on teachers’ salaries, all of which is managed by the Ministry of Education. In Blantyre, for example, all 1,700 primary school teachers are paid by the Ministry; the City Council only pays for a few administrative personnel in its Education Department.47 Through their education sector grants, City Councils finance operating expenditures, such as the purchase of some teaching and learning materials, school maintenance, and the costs of utilities for schools. Most such expenditures are routine, offer little room for budgetary discretion and are planned/budgeted by City Education Departments under close Ministry supervision and in collaboration with school management committees. Capital spending (on school classrooms and other facilities) is funded either directly by the Ministry of Education or through the LDF; in the case of LDF funding, the Ministry of Education appears to decide on the number of construction or rehabilitation projects, with City Councils playing a minor role in making decisions about which schools to build or rehabilitate. In Blantyre, for example, the City’s Education Department was not aware of which school construction projects had been selected for 2015- 16. City Councils only allocate very limited funds from their own-source revenues to finance primary education expenditure; in Blantyre, the Director of the Education Department insists that apart from paying for a small number of administrative staff salaries, the City Council allocated no own-source revenues to the primary education sector budget. Primary education is not seen by most Secretariat officials as something that is actively managed by local government – indeed, education is often not even thought of or presented as a City Council function or service48 by Secretariat staff. This is unsurprising given that: (a) the Ministry of Education (MoEd) continues to exercise near complete control over most decisions related to primary education; (b) City Education Departments are headed by directors employed and remunerated by MoEd, whereas all other Council departments are headed by officials paid out of the Council budget; (c) all primary school teachers are paid out of central government’s payroll; and (d) many school management decisions and operations are devolved to school management committees, made up of teachers and community representatives. Given the considerable extent to which they are earmarked, grants for education only appear to enter and exit City Council accounts as a “post boxâ€? or “pass-throughâ€? arrangement. In short, City Councils cannot genuinely be described as providing primary education services within their jurisdictions. It would be more accurate to describe City Councils as a delegated or passive conduit through which the Ministry of Education provides access to primary schools. In the primary health sector, City Councils (in contrast to District Councils) play a very limited role in direct service delivery. Much of this is probably due to the lack of budgetary resources (City Councils do not receive health sector grants unlike Districts), the presence of Ministry of Health funded major referral hospitals and clinics in urban areas, and the availability of not-for-profit and for-profit private sector health care services in cities. In Blantyre, the City Council’s Health Department does employ and 47 For further details on education sector spending in Blantyre, see Torre, K. (March 2010): An Education Needs Assessment of Blantyre City, Malawi, MCI Social Sector Working Paper series. 48 Strikingly, the section on Secretariat staffing in Mzuzu City Council’s Urban Profile (2013-2017) does not even include an enumeration of the staff employed by the Council’s Education Department. 49 pay about 50 staff in a few primary health centers but city residents very clearly depend on other facilities (not managed by the City Council) in order to access primary health care. The same is true of Lilongwe, where the City Council operates only one permanent public health care clinic (out of a total of 17). Both Lilongwe and Blantyre City Council health departments employ medical staff, but only nurses and no doctors; in addition, they fund limited operational expenditures. Other Municipal Services City Councils do provide a range of classic “municipalâ€? services: fire services, the management of local market places, burial services, and the like. Of these, the management of market places is clearly the most significant and important – in the eyes of most city residents. In return for daily fees charged to sellers, City Councils ensure that market places are cleaned and regulated. They are managed by market masters, employed by City Councils, and serviced by cleaners and other staff. As with roads and SWM, however, there have been no attempts made to out-source market management to the private sector. Planning and Coordination Urban local governments, especially City Councils, engage actively in planning and produce a range of plans – but with limited coordination and little implementation. Amidst active planning (land use or physical planning, investment planning, strategic planning) and the plethora of plans (master plans that project spatial and infrastructure intentions, strategic plans that outline the broad ways in which thematic issues are to be addressed, investment plans that list priority infrastructure), some of the lacunae characterize the ways in which local governments currently engage in urban planning. Firstly, local government planning in urban jurisdictions lacks “teethâ€? for implementation or execution. Because City Councils have limited resources (as will be discussed below), the implementation of any plans necessarily requires: (i) the active support of other stakeholders in the form of finance and the provision of other inputs; and (ii) a capacity on the part of local governments to ensure that a multitude of urban stakeholders comply with plans. Both of these implementation conditions appear to be absent in Malawi. There is, for example, little point in a City Council designating and planning future or upgraded low-income residential zones if key stakeholders cannot commit or are not committed to providing water connections, new or rehabilitated roads, electricity distribution, and the like. Nor do such plans make much sense if City Councils are unable to enforce (however liberally) compliance with zoning or land use regulations. In short, there is little in the way of meaningful city- wide and cross-sectoral planning – City Councils neither exercise control over key sectors nor have the authority to make other agencies align their plans to city-wide coordination. Secondly, the urban plans produced by City Councils tend to be too ambitious or rather unrealistic, with unattainable objectives or projections. There is little in the way of robust prioritization within the framework of hard-headed estimates of the levels of resources that are likely to be available or which accounts changing implementation realities of the 21st century, factoring in issues such as informal settlements, private sector initiatives, growing economic inequalities and technological change.49 For instance, Lilongwe’s 2010 Urban Development Master Plan, provides a wide-ranging overview of land use and physical planning and envisions a long list of actions to be taken. However, in its own evaluation 49 An interview with the UN-HABITAT during the fieldwork 50 of the process of drafting City Development Strategy (CDS), supported by Cities Alliance, the Lilongwe City Council recognizes the lack of funding as the prime constraint in implementing the Strategy.50 3.3.3. Urban Local Government Expenditure Patterns51 Urban local governments account for approximately 20 percent of total LG expenditure, while Districts account for about 80 percent (Figures 3.2 and 3.3). In per capita terms, urban local governments spend considerably more than do rural local governments. Figure 3.2 Local government - total expenditure Figure 3.3 LG expenditure per capita 25,000,000,000 2,500 20,000,000,000 2,000 Kwacha 1,500 15,000,000,000 1,000 All LGs All LGs 10,000,000,000 500 Rural LGs - Rural LGs 5,000,000,000 Urban LGs Urban LGs - Year City Council52 as well as District Council expenditures consist very largely of spending on recurrent items, as shown in Figure 3.4 below. Indeed, recurrent expenditure as a proportion of total appears to be slowly rising, from about 87 percent in 2011-2012 to just over 92 percent in 2013-2014. In all four cities, recurrent spending has amounted to between 70 percent and 100 percent of total expenditure over the last three years (see Figure 3.5). Figure 3.4 Recurrent expenditure as % of total expenditure Figure 3.5 Recurrent expenditure as % of total expenditure (by City Council) 100.00 100.00 80.00 80.00 60.00 All local 60.00 All Cities % % 40.00 governments 40.00 Lilongwe 20.00 District 20.00 - - Blantyre Councils Mzuzu City Councils Zomba Year Year 50 Lilongwe City Council (2015). Grant Completion Report: Grant for Lilongwe City Development Strategy Phase 3. 51 All the financial data for local governments used in this chapter is based on information provided by the National Local Government Finance Committee. 52 Among the seven urban local governments, the four City Councils account for around 95 percent of all expenditures – the following discussion therefore concentrates on the spending patterns of City Councils and does not consider municipal or township expenditure. 51 For City Councils, payroll-related costs have consistently accounted for between 40 percent and 50 percent of recurrent expenditures and between 32 percent and 47 percent of total expenditures (Figures 3.6 and 3.7). This is a considerably larger proportion than in District Councils, for which payroll costs account for less than 7 percent of total spending. This can be partly explained by the extent to which many payroll-related costs (sector department staff salaries) in District Councils are paid for out of the Central Government’s budget – in contrast, City Councils are directly responsible for paying many (but by no means all) such staff. In addition, payroll spending (like recurrent expenditure in general) among City Councils also appears to be steadily increasing over time. Figure 3.6 Payroll spending as % of total expenditure Figure 3.7 Payroll spending as % of total expenditure (by City Council) 100.00 100.00 80.00 80.00 60.00 All local 60.00 All Cities % % 40.00 governments 40.00 Lilongwe 20.00 District 20.00 - - Blantyre Councils City Councils Mzuzu Zomba Year Year Not surprisingly, City Council capital expenditure has been less than 10 percent of all city spending and appears to be declining. Nonetheless, it accounts for a greater proportion than in Districts, for which capital spending hovers around only 5 percent of total expenditure (Figures 3.8 and 3.9).53 In per capita terms, City Council spending on capital expenditure has also been low: for all four Cities, per capita capital expenditure was about MWK 230 (or US$ 1.46) in 2011-12, MWK 240 in 2012-13 (or US$ 1.29) and MWK 180 in 2013-14 (or US$ 0.76). It is not possible, on the basis of available data on City Council finances, to break expenditure down into service delivery or functional categories – although it can be said that most capital spending by cities appears to be on roads (construction, rehabilitation or maintenance) and on equipment (vehicles). Figure 3.8 Capital expenditure as % of total LG expenditure Figure 3.9 Cities: capital spending as % of total expenditure 50.0 30.0 40.0 25.0 30.0 20.0 All local 15.0 % % 20.0 governments Lilongwe 10.0 10.0 5.0 Blantyre - All Districts - Mzuzu All Cities Zomba Year Year 53 This may be because investments financed by the Local Development Fund (LDF – see below) are off-budget. 52 Levels of capital expenditure by Malawi’s local governments as a whole (rural and urban combined) are low compared to local governments in other developing countries54. Local government capital expenditure in Kenya, Tanzania, and Rwanda, for example, accounts for 15 percent, 26 percent and 43 percent of total spending – all considerably more than in Malawi. Malawian City Council capital expenditure patterns also tend to be outliers: out of nine sample East African cities, Lilongwe and Blantyre have the smallest shares of capital expenditure55. However, in comparison to other urban local governments in southern/eastern Africa, the share of payroll expenditure in Malawian city budgets appears to be around the norm, but with relatively greater spending on (non-payroll) operational expenditure items than most other cities.56 To sum up, although the LG Act formally assigns a wide range of functions to urban local governments, in practice, City, Municipal and Town Councils actually provide a circumscribed set of services, frequently in parallel to other institutions (which have the same functional mandates). In the case of some services, notably solid waste management, City Council performance is sub-optimal and could be improved through better management. In other cases, notably roads, City Councils face severe funding constraints (the revenue side of which will be examined in the next section). Some of the fiscal “squeezeâ€? can probably be attributed to relatively high payroll costs and a “force accountâ€? modus operandi. Although mandated to provide social services, in practice, City Councils either do so on the basis of limited delegation by line ministries (e.g. primary education) or do little (e.g. primary health care). As players in the arena of urban development, City Councils are small and not very effective. 3.4. Local Government Revenues Overall local government spending in Malawi is financed almost entirely from inter-governmental fiscal transfers (IGFTs) and own-source revenues (OSRs). Al local governments rely on IGFTs for over 75 percent of their total revenues, with the remaining balance (of 25% or less) accounted for by OSRs. Districts and Cities, however, are near mirror opposites in terms of their reliance on IGFTs/OSRs: while District Councils derive over 90 percent of their total revenues from IGFTs and less than 10 percent from OSRs (see figure 3.10), City Councils rely on OSRs for around 65-80 percent of all revenues, but on IGFTs for only about 20 percent (see figure 3.11). Among City Councils, property taxes account for between 40-50 percent of all revenues and other OSRs for 25-35 percent of total revenues. Sector or conditional grants account for the lion’s share of IGFTs, while General Resource Fund (GRF) block grants account for less than 4 percent of total revenues. 54 This may be because investments financed by the Local Development Fund (LDF – see below) are off-budget in Malawi. For cross-country comparison, different classifications of costs, different functional assignments and other factors can make it difficult to make meaningful comparisons. 55 See South African Cities Network (2008). 56 See South African Cities Network (2011). 53 Figure 3.10 District Councils: revenue sources Figure 3.11 All City Councils - main revenue sources 100.00 100.00 % of total revenues % of total revenues 80.00 80.00 60.00 60.00 CDF 40.00 40.00 20.00 20.00 Sector grants IGFTs - - GRF OSRs Other OSRs Property rates Year Year 3.4.1. Own-Source Revenue Assignments, Patterns and Performance The LG Act assigns a number of own-source revenues (OSRs), known as locally generated revenues in Malawi, to local governments. OSR assignments57 prescribed in Schedule 3 of the Act, in a cursory way, include: (i) Property rates (or taxes); (ii) Ground rent; (iii) Fees and licenses; (iv) Commercial undertakings; and (v) Service charges. Part VII of the LG Act is entirely devoted to spelling out detailed provisions for assessing, managing and collecting property rates (or taxes). Finally, Schedule 2 of the Act, which details the “additional functionsâ€? of the local government, includes additional guidance on user fees, licenses, and service charges. Other regulations that govern these OSRs, if any, are not readily available.58 The following table provides a summary of LG own-source revenue items.59 As in other developing countries, local government OSRs are typically of limited fiscal potential since the main sources of public revenue (e.g. VAT, customs and excise, income tax and natural resource royalties) are generally assigned to central government. Table 3.6 Local government own-source revenues Locally Generated Revenues Collected or levied by: Property rates Urban Councils or in areas designated as “rateableâ€?* Market fees All Councils Business licenses/permits All Councils Rental charges All Councils Advertising charges Urban Councils Car parking fees Urban Councils 57 In addition to these OSRs, the LG Act provides for “ceded revenuesâ€? (or shared revenues) as a source of revenue for local governments, to be passed on by central government. In practice, local governments do not receive any such shared revenues and these are therefore not discussed in this chapter. 58 In Malawi, nobody interviewed during the course of fieldwork was able to provide detailed and official regulations for LG own-source revenues. Nor are such regulations available on any official or other websites. Indeed, a good deal of the institutional and regulatory framework for local government own-source revenues would appear to be an “oral traditionâ€?, understood through past and established practice (rather than through documentation). 59 The table is adapted from World Bank (2015d): Malawi Urbanization Review: Challenges and Opportunities in Local Government Financing (Background Paper). 54 Bus departure fees Urban Councils Health fees and charges All Councils Refuse collection fees Urban Councils Hire of property – rooms, chamber and vehicles All Councils Interest on investments All Councils Fines and penalties for breaking by-laws All Councils Cemetery fees Urban Councils Plot charges/sale of immovable property Urban Councils Sale of water in markets All Councils Sale of forest produce All Councils Commercial undertakings - rest houses, bottle All Councils stores, bars and restaurants * Designated by the Ministry of Local Government & Rural Development (see below). City Councils rely on OSRs for 65-80 percent of their total revenues – This level of dependency on OSRs is unusual. Although larger urban local governments in developing countries typically can and do depend on OSRs more than rural LGs (as is the case in Malawi), few of them are as reliant on OSRs as are Malawian City Councils. In southern Africa, only Zambian and Namibian cities demonstrate comparable rates of dependence on OSRs.60 In contrast, municipalities in Nepal61 rely on OSRs for a little over 50 percent of their total revenues and larger Brazilian municipalities for between 38-55 percent.62 The relative importance of OSRs for City Councils in Malawi is largely due to the very limited transfers from central government. Although there are inter-annual variations and some inter-city variations (see figure 3.12), all City Councils rely on OSRs for most of their overall revenues. Figure 3.12 OSRs as % of total LG revenues (by LG category) Figure 3.13 City Councils - OSRs per capita (Kwacha) 100.00 5,000 80.00 4,000 Kwacha 60.00 All local 3,000 All Cities % 40.00 governments 2,000 Lilongwe 20.00 District 1,000 - - Blantyre Councils 2011-12 2012-13 2013-14 Mzuzu City Councils Zomba Year Year There is some difference between the four City Councils in terms of their OSRs per capita. OSRs per capita for all City Councils are not large and only amount to MWK 1,500-2,000 or US$ 5 per capita annually (Figure 3.13). Lilongwe’s per capita OSRs are generally the lowest and (at the same time) most stable (in nominal terms) – relatively low rates of OSR mobilization, according to Lilongwe City Council officials, are due to poor collection rates from Government agencies, which are concentrated in the national capital. Blantyre has been able to slowly increase OSRs per capita, while Mzuzu (a much smaller 60 See South African Cities Network (2011). 61 See LBFC Nepal (2011). 62 See World Bank (2014a): Municipal Finances – A Handbook for Local Governments. 55 city) has recently significantly increased its OSRs per capita.63 For unknown reasons, Zomba’s per capita OSRs have fluctuated considerably over the last three years. Figure 3.14 City Councils: breakdown of locally generated revenues 100.0 Other OSRs 80.0 60.0 Market fees % 40.0 20.0 - Fees and Service Charges Property Year rates Property tax is the most important revenue item for City Councils, accounting for between 50-65 percent of total OSRs, as shown in Figure 3.14. A wide range of fees and service charges make up the rest – market fees, fees for billboards and advertising, parking fees, land allocation and development fees, building permits, etc. Individually, these fees and charges are relatively minor, but together amount up. They are generally intended to cover the costs of specific services. Given their importance as OSRs (and thus as a key determinant of overall revenues), property rates are the primary focus of fiscal effort among City Councils. According to the LG Act, property rates/taxes can be levied on all properties within an LG jurisdiction, with a few exceptions (such as roads, parks and cemeteries). In practice, most City Councils limit themselves to levying property taxes on properties within “ratableâ€? areas. Although the criteria by which areas are deemed to be “ratableâ€? are not clearly defined, the observed tendency is for commercial/industrial, administrative and formal low density (and high value) residential areas to be designated as such, while informal and high density residential areas tend not to be designated as “ratableâ€?. The LG Act64 nonetheless allows for LGs to levy simple property taxes in non-ratable areas, allowing, in principle, City Councils to have a fairly wide property tax base. LGs also enjoy very wide discretion in terms of setting property tax rates, which can vary depending on the type of property (commercial, residential, etc.) and in terms of the actual percentage of assessed property value.65 63 Mzuzu’s improvement in performance can perhaps be attributed to recent changes to the way th at the City Council assesses and administers its property rates (see below). 64 LG Act 1998, Article 64: “In respect of any area which has not been designated by the Minister as a rateable area under this Act or which for any reason has not been assessed or is not assessable, the Assembly may levy (a) a fixed sum upon the owners of buildings which sum may be different in respect of buildings used for different purposes; and (b) a fixed sum per unit of area of land or a fixed sum per unit of superficial area or both such fixed sums.â€? 65 Indeed, it has been argued that the very wide statutory discretion that LGs enjoy with respect to setting the property tax rate should be capped. “Tax rates should be initially capped to protect the interests of the taxpayers. Currently Assemblies have unlimited authority to set tax rates in order to generate sufficient funds to meet all locally approved expenditures. Possible misuse of this freedom can place undue burden on taxpayers that can lead to mis-administration and possibly to tax revolts.â€? (see Kelly etalia. (UNDP/UNCDF) 2001). 56 Despite having a large property tax base and considerable latitude in setting property tax rates, City Councils have difficulties in mobilizing property tax revenues, due to the following problems: ï‚· Property tax rolls are generally gross underestimates of real properties in urban jurisdictions. For example, while Lilongwe City Council’s property valuation roll listed about 35,000 properties in 2011, it was estimated that there were perhaps a further 45,000 properties that should be on the list but were not (SACN 2011: 44). Blantyre City Council officials echo this: they estimate (in 2015) that their current property tax roll (drawn up in 2005) includes only a third of the guesstimated number of properties in the city. Part of this can be attributed to City Councils’ apparent reluctance to assess property values outside of officially designated ratable areas, even though some non- ratable areas have, over time, become indistinguishable from ratable areas. ï‚· The institutional and legal framework regulating property taxation is cumbersome, inappropriate, unrealistic and outmoded. Thus, the LG Act prescribes that the preparation and updating of valuation rolls can only be done by registered valuers66, of whom there are very few in Malawi. As a result, preparing and updating valuation rolls becomes very costly: Lilongwe City Council, for example, has estimated that updating its current property valuation roll will cost about MWK 400 million.67 Moreover, registered property valuers expect to be paid for their services as a function of the estimated values of the property they assess, which presumably provides them with incentives to over-estimate property values – and which almost certainly is cause for complaint and refusal to pay taxes on the part of property owners. Property valuation methods are also inappropriate and cumbersome – relying on individual, rather than mass, valuations – limiting City Councils to assessing high value properties (for which high taxes can be charged to justify the assessment costs). ï‚· In general, direct incentives for encouraging better City Council staff performance are absent. In Blantyre, for example, a dynamic and young senior official has recently tried to increase revenue collection rates by upgrading payments systems and setting revenue targets for his staff. Although there has been some improvement in revenue collection, the official insists that this would have been much better had it been possible to provide revenue collection staff with direct monetary incentives as a reward for improved performance.68 ï‚· Enforcement and compliance are major challenges in mobilizing property taxes – partly because procedures for enforcement are cumbersome, costly and counter-productive (interest charges on and penalties for non-payment add significantly to the tax bill itself and probably deter owners from settling their property tax bills), and partly because property owners (especially public sector agencies) are reluctant and slow to pay. Arrears on property rates are thus high: Lilongwe City Council estimates that it is owed almost MWK 8 billion in property taxes and is currently actively trying to recover some MWK 2.9 billion; although Blantyre City Council has fewer arrears, they still amount to about MWK 2 billion. ï‚· An unknown amount of collected taxes is misplaced or misappropriated by local government revenue collectors, often in collusion with tax-payers. 66 Registered under the Land Economy Surveyors, Valuers, Estate Agents and Auctioneers Act. 67 Roughly US$ 800,000 (as of July 2015). 68 The City Council and its Secretariat were unable to approve such an incentive-based scheme, for reasons that are not clear. 57 ï‚· Tax-payers are reluctant to pay, given the poor services provided by City Councils. Paying property tax is (rightly) perceived by Malawian city-dwellers as something for which they should receive services in return. Overall, property tax revenues are probably far lower than they should be, especially – but not only – in the larger urban centers as officially “ratableâ€? areas are out of step with urban development and do not include all properties, and tax collection rates are low. There clearly is potential to increase City Council property tax collection and thus its OSRs. The use of simplified mass valuation methods would allow City Councils to tax many more properties while legal and administrative reforms of the property tax are needed, along with a modernization and more effective revenue administration. Minor improvements in the tax payment system in Blantyre (where property taxes are now paid directly to a commercial bank), for example, have led to better collection rates, which have risen from less than 50 percent of the amounts billed in 2011-12 to almost 60 percent in 2013-14. The case of Mzuzu City Council, where a major reform and re-dynamization of property tax administration and collection has recently taken place, is further demonstration of the potential for improvement (see the box below). The reforms and new approach piloted in Mzuzu show that a good deal can be done in other cities to increase the number of taxable properties, to improve tax collection and administration and to raise more property tax revenues. Scaling up on what has been achieved in Mzuzu would require some regulatory changes (such as simplifying valuation procedures, rethinking the definition of “ratableâ€? areas), basic improvements in tax collection processes, and greater LG engagement with residents and property owners. Box 3.3 Mzuzu City Council and property rates Since 2013, with technical and other assistance from a German agency (RDF – Revenue Development Foundation) and funding support provided by Deutsche Gesellschaft für Internationale Zusammenarbeit GIZ, Mzuzu City Council has embarked on a major reform of its property tax administration system. Prior to doing so, the City Council had been collecting about MWK 120 million in property taxes; after a year of implementing new administrative methods and measures, property tax revenues had increased to a little over MWK 220 million, a significant increase.69 Several key improvements were made to Mzuzu’s property tax system: ï‚· Use of a mass valuation method, relying on GIS and local staff follow-up in the field, and based on a “points-basedâ€? assessment of property values – all carried out at reasonable cost, making use of existing staff and avoiding the costly direct use of registered valuers. This resulted in a dramatic increase in the number of assessed properties (from 8,000 to almost 40,000). ï‚· Establishment of an effective and transparent process for lodging complaints about property value assessments. ï‚· Modernizing tax payments through the banking system. ï‚· Undertaking neighborhood sensitization programs aimed at explaining property tax and its links to local service delivery. While technical improvements were instrumental in increasing property tax revenues, it is equally likely that the return of an elected City Council has played a role. 69 Although it should be pointed out that the rate of collection (taxes paid/taxes billed) did not improve radically. Instead, Mzuzu was able to dramatically increase the number of properties subject to propery tax. 58 Business licenses make up a small but important proportion of City Council OSRs. In Blantyre, business licenses account for a little over 10 percent of all OSRs. Collection rates in Blantyre (as measured by revenue estimates versus actual revenues) have apparently improved through incremental changes in revenue administration (with staff being assigned clear revenue targets and payments being made through the banking system, rather than to collectors). A study of LG revenue collection, carried out in 2001,70 made a number of recommendations for improving business licensing, although none have been taken up. The most important recommendation was to expand the scope of LG business licensing and to establish a “Single Business Permitâ€?, similar to what has successfully been implemented in Kenya. This would allow City Councils to widen the scope of business licensing and to increase their revenues. Daily fees paid by sellers in locally-managed markets are one of the major sources of City Council OSRs consists of market fees. About 10 percent of Blantyre City Council’s OSRs are derived from market fees. Market fees are closely linked to the provision of basic market services – cleaning, waste collection, maintenance, security, etc. – all of which require management and staff. The problem is that individual markets may often cost more to run than they generate in terms of fees: in Blantyre, where the city manages 24 local markets, City Council officials estimate that costs are barely covered out of fees, and that only four markets generate a surplus (used – in effect – to subsidize the remaining 20 markets). It is also unclear as to how far all market users actually pay daily fees and exactly what percentage of fees collected are actually registered. In short, while there may be room to increase revenues from market fees, this will probably require some degree of out-sourcing and semi-privatization (with City Councils divesting themselves of running costs and “leasing outâ€? markets to private sector management). 3.4.2. Inter-Governmental Fiscal Transfers Inter-governmental fiscal transfers (IGFTs) account for between 75-80 percent of all local government revenues in Malawi. District Governments are more dependent on IGFTs, which account for over 90 percent of all district revenues, than City Councils for whom IGFTs only account for a little over 20 percent of their revenues (Figure 3.15). For both rural and urban local governments, per capita IGFTs are low and amount to well below US$5. For City Councils, per capita IGFTs in Mzuzu and Zomba are about twice or three times greater than in Lilongwe and Blantyre, as they receive higher per capita education sector grants in reflection of higher poverty rates (Figure 3.16). These IGFT per capita differences also imply that relatively little weight is given to population size in determining grants to City Councils. 70 See Kelly, R. etalia. (2001): Improving revenue mobilization in Malawi: study on business licensing and property rates, UNDP/UNCDF. 59 Figure 3.15 Central Government transfers as % of total Figure 3.16 All IGFTs per capita revenues 100.00 1,400 80.00 1,200 1,000 Kwacha 60.00 800 % 40.00 600 Lilongwe 400 20.00 All LGs 200 Blantyre - - Mzuzu Districts Zomba Cities Year Year Transfers to local government consist of both unconditional block grants and earmarked or conditional grants. In both cases, the NLGFC provides coordination and overall management. The largest grants provided to local governments as a whole are for the health and education sectors. There are important differences between District and City Councils in terms of IGFTs (Figure 3.17). The only significant sector conditional grant allocated to City Councils is for education. City Councils are also provided with earmarked grants for roads rehabilitation and maintenance, unlike rural LGs. District Councils, on the other hand, receive up to 14 different types of sector grants, the largest of which is for health (only districts benefit from health sector grants), followed by education. Figure 3.17 2012-2013: Central government grants to local governments (%) 100.00 80.00 % of all grannts Education 60.00 Health 40.00 Other sectors 20.00 - GRF All LGs Districts Cities The largest category of conditional grants is earmarked for financing “devolvedâ€? sector expenditures (health, education, agriculture, etc.). Each sector allocation is based on a specific formula; for example, education sector allocations to local governments are calculated on the basis of the number of school- aged children in each LG jurisdiction. From the point of view of local governments, discretion over the use of such sector conditional grants is very limited – as they are earmarked not only for specific sectors, but also for specific spending items, determined by respective line ministries. In all cases, sector conditional grants are earmarked for financing non-payroll operating costs and are not used to finance any sector investments (such as classroom construction in the education sector). In practice, sector conditional grants are tightly controlled and to finance delegated functions and their use is closely monitored (on a monthly basis) by line ministries and the NLGFC and in-year disbursements are 60 contingent upon LG reporting on their correct use. Local governments act as “post boxesâ€? through which funds are channeled to de-concentrated line departments. Another type of conditional grant derives from the Constituency Development Fund, allocated on the basis of the number of parliamentary constituencies in each local government area. CDF grants are earmarked for spending in specific constituencies and to finance a range of small-scale “community- basedâ€? projects or investments (e.g. school classrooms). Currently (FY 2015-16), each local government receives an annual CDF grant of MWK 9 million71 for each of its parliamentary constituencies, which is intended to be a part of the local government’s regular budget. CDF Guidelines issued by the NLGFC in 2015 specify that while each local MP is the ultimate arbiter of how CDF funds are to be used, budgeting and execution are expected to be fully consistent with regular local government planning processes and to be managed in line with regular financial management and procurement procedures. In practice, however, CDF funds are rarely considered by local governments to be a part of their mainstream development budget as noted by a critical discussion of CDF grants in the box below. Box 3.4 The Constituency Development Fund “The Constituency Development Fund (CDF) is a result of political manoeuvring in the mid-2000s when President Mutharika, needing MPs’ support, agreed to their initiative. It is a fund voted by parliament annually for all its members, which is formally managed by DCs. In practice, MPs treat the funds as their own monies, and shun oversight by local government. The amount now [2014] stands at MWK 7million per annum for each MP. The projects on which the CDF are spent may not be those identified in area development plans, since most MPs aim to use the funds to gain votes and reward loyalists, not to fulfil a district development vision. That said, some parliamentarians involve their local party leaders in deciding how to spend the money … but across the country ordinary citizens report that their MPs do not involve them in CDF planning ... Similarly, some CDF projects stand as a testament to an MP, while others can be seen as a waste of public funds …. Overall, the CDF programme runs counter to the government’s professed objective of strengthening local government and its capacity to meet local needs, and to more transparent and accountable processes, as it involves the use of public funds at local levels that are outside the formal planning, budgeting and spending framework. MPs are in practice not held to account for how they use CDF funds, and the fund legitimises MPs’ role as deliverers of local development – which is how they are seen by most voters and which creates another reason for tension between them and councilors, local authorities and [traditional authorities].â€? Source: O’Neil, T. & Cammack, D. (2014): Fragmented governance and local service delivery in Malawi, ODI. City Councils (but not District Councils in rural areas) are also provided with earmarked grants for the construction, rehabilitation or maintenance of urban roads. These grants are sometimes itemized as being made from the Infrastructure Development Fund (IDF)72. Budget allocations of these grants appear to be made on an arbitrary basis, with each City Council receiving the same (large) allocation and 71 Approximately US$ 20,000 per MP (using 2015 MWK/US$ exchange rates). 72 The exact status of the IDF is not known. None of the Ministry of Finance or NLGFC officials met during the course of fieldwork were certain of what the IDF is or how it is determined. 61 each Municipal or Town Council receiving smaller (but identical) allocations.73 Actual amounts disbursed as IDF grants to urban local governments have been less than the initial budget allocations, due to a combination of funding shortfalls and under-spending. Block grants are allocated from the General Resource Fund (GRF), a central government funding pool . GRF block grants account for less than five percent of all LG grants, a little less than the proportion accounted for by grants from the Constituency Development Fund (CDF). Although the total size of the GRF funding pool appears to have been fairly steady over time, there is no statutory rule fixing the size of the GRF as a given proportion of central government’s revenues or budget. Annual transfers to individual local governments from the GRF are calculated (by the NLGFC) on the basis of a simple formula that uses population size (80%) and relative poverty (20%) as allocation criteria. Local governments can use their GRF block grants in any way that they wish, provided that this is consistent with the provisions of the LG Act and other regulations. In addition to “regularâ€? IGFTs, local governments are also provided with development finance from the Local Development Fund (LDF), a hybrid blend of the Malawi Social Action Fund (MASAF) and the District Development Fund (DDF).74 In principle, the LDF was established as a “basketâ€? for Government and development partner funding aimed at a variety of local-level projects; in practice, the LDF is made up of four funding “windowsâ€?, each of which is earmarked for specific types of investments or projects and is funded by different development partners (Table 3.7 ). City Councils can receive LDF funding either as part of their education sector grants or through the LDF’s Urban Window.75 Table 3.7 Local Development Fund Windows LDF Windows Target Funding source 1. Community Small-scale community-based projects World Bank (MASAF), supplemented Window (≤ MWK 15 million) by financing from the Education 2. Local Authority Larger inter-community-based projects (Sector Wide Approach) SWAp Window (> MWK 15 million, ≤ MWK 40 million) 3. Urban Window Socio-economic investments in urban Demand-driven component by settlements Kreditanstalt Für Wiederaufbau (KfW, or German Development Bank); supply-driven component by the African Development Bank (AfDB) 4. Performance Local government capacity-building Window activities 73 For 2014-15, each of the four City Councils was provided with a MWK 127.65 million IDF allocation and each of the three Municipal/Town Councils with MWK 20 million; for 2015-16, City Councils have been allocated MWK 150 million from the IDF, Municipal/Town Councils with MWK 20 million, and each of the five former Town Councils with MWK 10 million. 74 The LDF does not have a formal or distinct legal or institutional identity, although it is a separate vote (no. 272) within the Government’s overall budget. LDF funding decisions are made by a Steering Committee and day -to-day LDF activities are managed by a Technical Support Team (TST). The LDF’s primary “supervisorsâ€? are the Min istry of Finance (which chairs the LDF Steering Committee) and the Ministry of Local Government & Rural Development. 75 It is unclear as to whether LDF education sector funding for classrooms and teacher housing is on- or off-budget, although it is clear that the LDF TST is responsible for managing these types of investments. 62 The LDF Urban Window isa type of “challenge fundâ€? that can be tapped into so as to finance either planning exercises or infrastructure investments.76 Both urban and rural local governments submit project proposals for urban socio-economic investments or planning activities; these are then reviewed and appraised by the LDF and, if judged to be robust, will be approved for funding. To date, there have been two rounds of submissions, with a third round underway; up until the end of FY 2014/15, a little over MWK 700 million has been disbursed through the LDF Urban Window.77 LDF funds allocated for urban investments and projects are not on-budget (from the point of view of local governments) and are directly disbursed to contractors or consultants by the LDF itself. To date, an underwhelming number of proposals were submitted to the LDF Urban Window by urban local governments. Out of 128 urban investment project proposals submitted in the first two rounds of calls for funding by the LDF, 19 were approved and of these, only four were proposals submitted by City Councils and one by a Town Council. Only Zomba City Council and Mangochi Town Council have received LDF funding for infrastructure projects (e.g. sports stadiums, markets, and bus depots); Blantyre, Lilongwe and Mzuzu City Councils have used funds from the Urban Window to finance urban planning activities.78 LDF officials attribute the underrepresentation of City Councils to their lack of motivation and interest in seeking funding for urban investments, as well as the poor technical quality of the disproportionately small number of proposals submitted by City Councils. 3.5. Borrowing and Other Financing The LG Act allows for local governments to borrow, with approval by the Ministry of Local Government and Rural Development and the Ministry of Finance. It also allows local governments to operate short term overdrafts, provided that they do not exceed one-sixth of the previous year’s current revenues. Borrowing spreads the cost of such investments across generations and is thus more equitable than relying on current revenues as a source of finance, which are also unlikely to be sufficient to finance large infrastructure investments. Thus, many local governments – especially those in urban areas – often access to and rely on borrowing in order to finance larger infrastructure investments. In practice, City Councils (and other local governments) in Malawi borrow very rarely for the purposes of financing infrastructure investments and do not appear to make excessive use of overdraft facilities to manage any cash flow issues. Debt servicing costs79 are generally very low, amounting to less than one percent of City Council OSRs. Most debt servicing costs are associated with bank charges and interest on overdrafts, taken out to smooth over short term cash illiquidity. Low levels of borrowing on the part of urban local governments80 are not surprising, given their revenue/expenditure patterns, the paucity of municipal lending institutions and the absence of any 76 As such, the LDF’s Urban Window appears to function as an embryonic municipal development fund. 77 For the two years (2013-14 and 2014-15), the planned budget for the LDF’s Urban Window was a little over MWK 3.5 billion. Actual spending has been less due to delays and the need for multi-year investments. 78 To the LDF Urban Window, 101 proposals were submitted by 32 LGs in the first round, 27 proposals by 12 LGs in the second round, and 32 proposals by 17 LGs in the third round. Of this total of 160 submitted proposals, 19 infrastructure projects and nine planning exercises have been approved to date. 79 In the absence of audited local government balance sheets, debt servicing expenditure is the only systematic evidence of any borrowing or credit. 80 In the absence of recent Annual Financial Statements for any of the City Councils, it is not possible to specify exact debt stocks. The Directors of Finance of Lilongwe and Blantyre City Councils, however, insist that neither City Council has any substantial medium or long term debt. 63 form of municipal bond market. City Councils do not generate enough revenue to finance meaningful borrowing and what revenue that they do collect is very largely spent on recurrent costs. And, even if they did, there are few lenders. The Development Fund for Local Authorities (DFLA) is the only institution in Malawi that lends investment finance to local governments. DFLA was established (as a Trust Fund) in 1993 with a start- up capital of about US$ 8.5 million, as part of the World Bank funded Local Government Development Project (LGDP).81 DFLA, as intended, has provided grant and loan finance for local government investments: since 1993, DFLA has lent or granted funds to 31 local governments for a range of investments (construction of roads, market facilities, and other infrastructure, purchase of vehicles, etc.). Most DFLA grants/loans were approved in the period between 1994 and 1998; after the late 1990s DFLA activities appear to have ground to a standstill, as a result of capital depletion – itself due to poor rates of repayment, debt cancellation, attrition through the conversion of loans to grants (which became an ever larger proportion of DFLA funding), and management costs.82 Since 2010, DFLA has re- commenced operations on a very modest basis of a greatly diminished capital of about US$ 1.4 million and is seeking to replenish its revolving fund. Although it is still – in legalistic terms83 – a functioning institution, DFLA is effectively non-operational and no longer a significant source of investment finance for local governments, few of which are even aware of the Fund’s continued existence. Urban local governments in Malawi make very little use of Public-Private Partnerships (PPPs) to finance public investments or services. This is partly a reflection of the Malawian public sector as a whole, which appears to be generally averse to PPP-type arrangements (and, indeed, to out-sourcing), as well as a result of the limited opportunities for private sector involvement in the provision of public infrastructure and services. Nonetheless, City Councils – particularly Blantyre and Mzuzu – are well aware of the private sector, have sought to encourage firms to take on minor functions,84 and are open to PPP options that make sense. Given the limited commercial opportunities, the private sector would appear to have few incentives to engage in PPPs. 3.6. Conclusion Unlike rural Districts, City Councils depend very largely on own-source revenues (OSRs) alone and receive very little in the way of inter-governmental fiscal transfers (IGFTs), which account for 20 percent or less of their total revenues. In Malawi, central government transfers to local governments, in general, and especially to urban local governments are currently well below the levels that would normally be expected in a developing country. Even though District Councils are already highly 81 There is very little available documentation on the DFLA, either at the World Bank or in Malawi, apart from the Implementation Completion Report (ICR) by both parties. 82 The World Bank’s ICR (2001) for LGDP notes that: “The financing mechanism (DFLA) set up by the project was a failureâ€? (p. 17). While the DFLA, as a form of revolving fund, was expected to reconstitute itself through LG repayments, this did not happen “ … due to the chronic failure of local authorities to service their debt obligationsâ€?. The ICR further notes that: “The DFLA apparently never incorporated any systems to assess the creditworthiness of the local authorities to which it "lent". In practice, the costs of new sub-projects financed under the LGDP would automatically appear as "loans" to the local authorities with no ex ante analysis, discussion, or agreementsâ€?. Although infrastructure investments were completed (behind schedule), the DF LA was evaluated as being un-sustainable. 83 Its Board of Trustees, made up of representatives from central and local government, remains in existence. 84 Such as the maintenance and upkeep of roundabouts in Blantyre, in exchange for free advertising. 64 dependent on IGFTs for most of their revenues, the absolute amounts thus transferred are small to meet the needs. In the case of urban local governments, IGFTs are woefully inadequate. Such low levels of IGFTs to City Councils are not based on any official policy prescriptions as no explicit and formal policy statements rationalize differing levels of central government financing for urban/rural local governments. Instead, there appears to be an informal agreement within central government policy- making circles that rural LGs are unable to mobilize significant OSRs and therefore require more support (in terms of transfers) than do urban LGs. In all, the Government is a long way from meeting its stated fiscal commitment to decentralization by transferring at least five percent of national revenues (excluding development partner grants).85 Figure 3.18 LG total revenues per capita (kwacha) 3,000 2,500 Districts 2,000 Kwacha 1,500 Cities 1,000 500 Municipalities & towns - 2011-2012 2012-2013 2013-2014 All LGs Year Staffing levels in City Councils are relatively high, which is reflected in the high share of recurrent costs in the total expenditures. Lilongwe City Council, for example, currently employs about 2,000 staff (roughly one employee for every 350 city residents); Mzuzu employs about 520 staff (one employee for every 300 city residents); and Blantyre about 1,400 staff (one employee for every 500 city residents). Many such staff are unskilled (employed as laborers for public works activities, street sweepers, gardeners) and semi-skilled (employed in clerical/administrative positions, for market management, for revenue collection). All City Councils insist that there are many unfilled positions, especially for senior and technical jobs, where they have a hard time offering competitive salaries and career opportunities. The modus operandi of City Council management can only be described as staid and “time-warpedâ€?. Secretariat departments tend to see themselves as being directly responsible for the production (rather than the provision) of city services. Engineering sections, for example, tend to operate on the assumption that city public works activities will be undertaken on the basis of force account modalities, drawing on city-owned equipment and city-employed labor – rather than through out-sourcing to private sector contractors. For instance, solid waste collection is also seen as a task that should be undertaken by city garbage trucks and the city’s workforce, rather than private sector service providers contracted by the city. Although there are no regulatory limitations to out-sourcing service delivery activities to the private sector, City Councils seem averse to do so.86 85 As spelt out in the National Decentralization Policy (1998:13). 86 It is beyond the scope of this chapter to explore the rationale underlying a continued insistence on both the provision and production of public goods and services by local governments. However, there are clearly grounds for arguing that the private sector in Malawi is under-developed, has weak capacity to engage formally with the 65 The effectiveness of City Council financial management is hard to evaluate given the unreliability of audits. All City Councils use and have access to the Government’s Integrated Financial Management Information System (IFMIS). However, external audits by the National Audit Office have been consistently well behind schedule and, even when they have been carried out, audit reports have not been submitted to Councils or made publicly available.87 Moreover, there are some grounds for supposing that external audits by the Central Government have – in the past – been less than exhaustive and rigorous; they are also largely limited to compliance auditing, rather than full financial auditing. Although internal audit sections are in place, they are not seen as being effective.88 Some of the inertia and immobilization that characterizes city management can probably be attributed to the decade-long period (2005-2014) during which Councils did not have locally elected councilors. Indeed, there is much anecdotal evidence that the holding of local elections in 2014 has re- introduced a greater sense of purpose, urgency and dynamism to city management. Elected councilors appeared to be accountable to their ward voters and, in turn, put pressure on Secretariat officials to deliver services. On both sides, however, this has resulted in a degree of resentment. City mayors in Lilongwe, Blantyre and Mzuzu all complain about their Secretariats’ lack of responsiveness, reluctance to “do their jobsâ€? and obfuscation. Secretariat officials, on the other hand, complain about council micro- management, interference and inability to understand budgetary constraints. Malawian cities need more resources and strengthened public infrastructure and service delivery institutions. First, the role of City Councils in urban development needs to be streamlined to focus on infrastructure and service delivery responsibilities that they have comparative advantages. The current set of functional assignments provided in the LG Act is too extensive and needs to be revised to reflect how services are actually managed. Some functions, notably primary education and health services, are probably best left to deconcentrated line departments, given that they are already provided on such basis in urban areas. Other institutions (such as City Water Boards) are better-suited to the task of delivering certain urban public goods and services. The revision of the LG Act, in the long run, can provide a formal focus on fewer and more specific functions for urban local governments to meet the needs that are different to rural development needs.89 public sector and may itself be reluctant to take on responsibilities for the production of publicly-financed goods and services (see Tambulasi, R. (ed) (2010): Reforming the Malawian Public Sector, Council for the Development of Social Science Research in Africa). At the same time, keeping production “in -houseâ€? has its logic: providing regular and salaried employment, extending patronage to employees, avoiding conflict with organized labor, maintaining bureaucratic “fiefdomsâ€?, and the like. 87 The recent evaluation of NDP II, however, notes that the National Audit Office (with financial support from Irish Aid and KfW) has now audited Council accounts up to 2011-12. 88 “Based on the current research evidence, local-level financial control appears frequently to operate without the full range of expected ex-post and ex-ante controls.â€? p. 35, O’Neil, T. & Cammack, D. (2014): Fragmented governance and local service delivery in Malawi, ODI. 89 In addition, a major revision of the LG Act might include provisions for different types of urban local government, recognizing that large cities (such as Lilongwe or Blantyre) may need different municipal arrangements to those appropriate to emerging towns or smaller cities. Some countries (e.g. Bangladesh and India) distinguish between, for example, large City or Metropolitan authorities and smaller municipalities. This finds strong echoes in earlier recommendations in a World Bank (2012) study of decentralization in Malawi and in ODI-sponsored work on local governance and service delivery in Malawi (O’Neil, T. & Cammack, D. 2014). 66 City Councils can focus on planning and coordination, by strengthening an institutional architecture that improves downward accountability. Electoral representation does lead to some degree of downward accountability, and local governments can often facilitate citizen participation and engagement and – given the right circumstances – prove responsive to citizen priorities and needs. These are important attributes, which are not characteristic of other types of public institution. As things stand, City Councils have limited authority to implement any urban development plans; they can do little to make other players actively comply with spatial (or physical) plans, even though the involvement of agencies such as Water Boards or the Roads Authority is critical to planning initiatives aimed at, for example, upgrading informal residential settlements. This is partly a question of political will but may also require a formal provision of greater authority to City Councils to coordinate infrastructure and service delivery within their jurisdictions, including water supply, provision of roads, solid waste management, essential municipal services (fire services, for example) and planning for more integrated, more accessible and better services. It will also be necessary to change the ways in which local government, in general, and City Councils, in particular, manage service provision and their budgets. City Councils do very little in the way of out- sourcing the production of services, preferring instead to operate on the basis of force account modalities. Given that the private sector is probably more cost-effective in producing some services (e.g. waste collection, road works) than are in-house management arrangements, out-sourcing would not only enable savings (which could be invested in other activities, such as waste disposal or regulatory supervision) but might also allow City Councils to reduce payroll and other recurrent costs. Likewise, increases in IGFTs will also need to be carefully thought through in terms of the extent to which they either allow discretion at the local level or are earmarked (for, say, capital expenditure or specific sectors). As things stand, increases in fully discretionary block grants (such as those allocated from the GRF) might be counter-productive – City Councils might well use full discretion to simply increase payroll and recurrent spending. It will clearly not be easy, in political terms, to cut back on bloated City Council payrolls or staff numbers but there is a major imbalance in the current structure of expenditures which needs to be addressed before additional resources are raised or accessed by City Councils. Third, in order to meet the significant and growing needs of cities, urban local governments need more funds and greater capacities. There are a number of options that could be explored to actually increase fiscal resources in urban local governments. Increasing own-source revenues through improved property taxation. A range of measures and actions can be taken to increase property tax revenues, inter alia: ï‚· Amending legal and regulatory provisions to make prescribed property valuation methods and practices more up-to-date. Better ways of updating property tax rolls exist (and have been implemented in Mzuzu) and need to be provided with legal backing; ï‚· Modernizing payment systems to reduce leakage and encourage tax-payers to settle property rates; ï‚· Undertaking more sensitization of tax-payers; and ï‚· Providing City Council revenue officials with training and modern taxation tools (e.g. GIS). Needless to say, improvements in local infrastructure and service delivery would go a very long way towards making taxpayers more willing to meet their fiscal obligations. Increasing City Council OSRs will 67 at least provide them with a basis to make immediate improvements in infrastructure and services delivery and, in the longer term, will also enhance their creditworthiness and ability to borrow. Inter-governmental fiscal transfers. If local government is to make a genuine contribution to urban development, the Central Government and its development partners should explore ways of increasing the size of IGTFs to City, Municipal and Town Councils, while using those transfers as leverage for improvements in the way that local governments spend and manage their resources. In other words, additional transfers should only be made available to urban local governments on the basis of their performance and if they meet certain conditions such as increasing their fiscal effort (to raise OSRs), making agreed changes to their management, and gradually altering the structure of their budgets. Two basic models for this have been applied in other developing countries: ï‚· Performance-based grant systems (PBGSs) have been widely used in Sub-Saharan Africa and elsewhere to provide incentives for better local government management.90 The basic premise is that grants are provided to local governments as a function of measurable and sustained improvements in their performance such as full compliance with statutory requirements and enhanced fiscal effort. For Malawi, a PBGS for City Councils might, for example, be designed so as to incentivize reduced payroll and recurrent expenditure or efficient out-sourcing of service delivery. ï‚· Municipal contracts have also been used in many countries as instruments for driving reforms in urban local government.91 Municipal contracts between central and local governments spell out a range of reform steps to be taken at the local level; in return for embarking on reforms, urban local governments are provided with additional grants and other assistance by central government. Investment grants. City Councils may also need encouragement to provide financing for certain public services – but not get directly involved in service production. A case in point here is water supply: there is no reason why City Councils should not provide their respective Water Boards with investment grants in order to extend or upgrade water supply networks in low-income housing areas. This would also add teeth to the spatial planning and coordination functions of City Councils by financing key services. Borrowing. In the medium to long term, city governments in Lilongwe, Blantyre and Mzuzu will need to be able to access credit to finance the growing infrastructure requirements associated with urban growth. Developing appropriate lending arrangements will take time and a good deal of expertise. The task will probably be more difficult in Malawi than in other countries, if only because the potential in- country municipal market for loans is likely to be relatively small, perhaps too small to justify setting up a fully-blown set of domestic arrangements. Development partners could assist the Government in medium term strategizing here and be ready to share financial risks as any such system starts up. 90 see Steffensen (2006): Performance-Based Grant Systems: Concept and International Experience , UNCDF 91 See World Bank (2014a): Municipal Finances – A Handbook for Local Governments. 68 References Arndt, C., A. Schlosser, K. Strzepek and J. Thurlow. 2014. Climate Change and Economic Growth Prospects for Malawi: An Uncertainty Approach. Journal of African Economies 23(2): 83-107. Arndt, C., K. Pauw, and J. Thurlow. 2015. The Economywide Impacts and Risks of Malawi’s Farm Input Subsidy Program. American Journal of Agricultural Economics. Bahl, R. & R. Bird. 2013. Decentralization and Infrastructure in Developing Countries: Reconciling Principles and Practice. IMFG Papers on Municipal Finance and Governance, University of Toronto. Barré, J. 2014. Waste Market in Urban Malawi – A way out of poverty? Master’s Thesis, Department of Urban and Rural Development Swedish University of Agricultural Sciences, Uppsala. Benin, S., X. Diao, C. McCool and F. Simtowe. 2012. Malawi. In Diao X., J. Thurlow, S. Benin and S. Fan (eds.). Strategies and Priorities for African Agriculture: Economywide Perspectives from Country Studies. Washington DC, USA: IFPRI. Cammack, D., E. Kanyongolo, & T. O’Neil. 2009. ‘Town Chiefs’ in Malawi. APPP Working Paper No.3, ODI. Commonwealth Local Government Forum. 2013. The local government system in Malawi. Dervis K., J. De Melo, and S. Robinson. 1982. General equilibrium models for development policy. New York, USA: Cambridge University Press. de Vries, G.J., M.P. Timmer and K. de Vries. 2015. Structural Transformation in Africa: Static Gains, Dynamic Losses. Journal of Development Studies 51(6): 674-688. Diao, X., P. Hazell, and J. Thurlow. 2010. The Role of Agriculture in African Development. World Development 38(10): 1375-1383. Diao, X. and J. Thurlow. 2012. A Recursive Dynamic Computable General Equilibrium Model, in X. Diao, J. Thurlow, S. Benin and S. Fan (eds.) Strategies and Priorities for African Agriculture: Economywide Perspectives from Country Studies. Washington DC, USA: International Food Policy Research Institute. Dorosh, P., and J. Thurlow. 2014. Can Cities or Towns Drive African Development? Economywide Analysis for Ethiopia and Uganda. World Development 63(10): 113-123. Foster, V. & Shkaratan, M. 2010. Malawi’s Infrastructure: A Continental Perspective. AICD Country Diagnostic. Foster, V. & Briceño-Garmendia, C., (eds). 2010. Africa’s Infrastructure: A Time for Transformation, World Bank. GOM (Government of Malawi). 1998. National Decentralization Policy. GOM. 2004. Joint Malawi Government/Donor Review Team (May 2004): Review of the National Decentralization Program of Malawi. GOM. 2006. Malawi Growth and Development Strategy: From Poverty to Prosperity, 2006-2011. Lilongwe, Malawi: Government of Malawi. GOM. 2011. Malawi Growth and Development Strategy II, 2011-2016. Lilongwe, Malawi: Government of Malawi. Henderson, J.V. and H.G. Wang. 2005. Aspects of the rural-urban transformation of countries. Journal of Economic Geography 5: 23–42. Hove, T. 2011. Waste management in developing countries: A case study of Blantyre City, Malawi. EWB- UK National Research & Education Conference. 69 Jere, P. 2012. Improving land sector governance in Malawi – Implementation of the Land Governance Assessment Framework. Ministry of Lands, Housing and Urban Development. Kelly, R. etalia. 2001: Improving revenue mobilization in Malawi: study on business licensing and property rates. UNDP/UNCDF. Lea, Nicholas, and Lucia Hanmer. 2009. Constraints to Growth in Malawi. Policy Research Working Paper 5097. The World Bank. Washington DC, USA. Lilongwe City Assembly. 2009. Lilongwe City Development Strategy. Lilongwe City Council. 2015. Grant Completion Report: (Cities Alliance) Grant for Lilongwe City Development Strategy (Phase 3). Unpublished. Local Bodies Fiscal Commission (Nepal). 2011. Municipal finance aggregates (2008-09). Malawi Roads Authority. 2013. Annual Report 2012. Malawi Roads Authority. 2012. Annual Report 2011. Manda, M. 2013. Situation of Urbanization in Malawi (Consultancy services to prepare a National Urban Policy Framework), Ministry of Lands and Housing. Maoulidi, M. 2012: Water and sanitation needs assessment for Blantyre City, Malawi. MCI Social Sector Working Paper Series n° 27/2012. McMillan, M., D. Rodrik, and I. Verduzco-Gallo. 2014. Globalization, Structural Change, and Productivity growth with an Update on Africa. World Development 63(11): 11-32. Minister of Finance. 2015. Budget Speech 2015-16. Mzuzu City Council. 2013. Urban profile 2013-17. National Local Government Finance Committee. 2014. Constituency Development Fund Guidelines. NSO (National Statistics Office). 1989. Population and Housing Census 1987 (IPUMS Subset). Zomba, Malawi: National Statistics Office. NSO. 2000a. Population and Housing Census 1998 (IPUMS Subset). Zomba, Malawi: National Statistics Office. NSO. 2000b. First Integrated Household Survey 1997-1998. Zomba, Malawi: National Statistics Office. NSO. 2006. Second Integrated Household Survey 2004-2005. Zomba, Malawi: National Statistics Office. NSO. 2008. Malawi Housing and Population Census (PHC): Spatial analysis. Zomba, Malawi: National Statistics Office. NSO. 2010a. Population and Housing Census 2008 (IPUMS Subset). Zomba, Malawi: National Statistics Office. NSO. 2010b. Analytical Report Volume 9: Population & Housing Census Spatial Distribution and Urbanisation. Zomba, Malawi: National Statistics Office. NSO. 2012. Third Integrated Household Survey 2010-2011. Zomba, Malawi: National Statistics Office. NSO. 2014a. Integrated Household Panel Survey 2013. Zomba, Malawi: National Statistics Office. NSO. 2014b. Labor Force Survey 2013. Zomba, Malawi: National Statistics Office. NSO. 2014c. Annual Economic Survey Report 2010-2011. Zomba, Malawi: National Statistics Office. NSO. 2015. National Accounts Online Database. Zomba, Malawi: National Statistics Office. < http://www.nsomalawi.mw/national-accounts.html - last accessed July 2015> O’Neil, T. & D. Cammack. 2014. Fragmented governance and local service delivery in Malawi. ODI. Patel, N. & M. Wahman. 2015. The presidential, parliamentary and local elections in Malawi, May 2014. Africa Spectrum, 50 (1): 79-92. Pauw, K., P. Dorosh and J. Mazunda. 2013. Exchange Rate Policy and Devaluation in Malawi. Discussion Paper 1253. Washington DC, USA: International Food Policy Research Institute. 70 Pauw, K., F. Schuenemann and J. Thurlow. 2015. A 2010 Social Accounting Matrix for Malawi. Washington DC, USA: International Food Policy Research Institute. Rosenthal, S.S., and W.C. Strange. 2004. Evidence on the nature and sources of agglomeration economies. In J.V. Henderson and J.-F. Thisse (eds.) Handbook of urban and regional economics: Volume 4. Amsterdam: Elsevier. South African Cities Network. 2011. The Financing of City Services in Southern Africa. Tambulasi, R. (ed). 2010. Reforming the Malawian Public Sector. Council for the Development of Social Science Research in Africa. Torre, K. 2010. An Education Needs Assessment of Blantyre City, Malawi. MCI Social Sector Working Paper series. Trémolet, S. & G. Mansour. 2014. Improving performance through service contracts: can it be done? Independent Assessment of the Private Sector Service Contract between Vitens Evides International and the Blantyre and Lilongwe Water Boards in Malawi. Tschirley, D., T. Reardon, M. Dolislager and J. Snyder. 2015. The Rise of a Middle Class in East and Southern Africa: Implications for Food System Transformation. Journal of International Development 27(5): 628–646. UNDESA. 2015. 2015 Revision of World Population Prospects. New York NY, USA: United Nations Department for Economic and Social Affairs. UN-Habitat. 2011. Malawi: Blantyre urban profile. UN-Habitat. 2011. Malawi: Lilongwe urban profile. UN-Habitat.2011. Malawi: Mzuzu urban profile. UN-Habitat. 2012. Malawi national urban profile. UN-Habitat. 2011. Malawi: urban housing sector profile. UN-Habitat. 2013. State of the World’s Cities 2012/2013: Prosperity of Cities. Wiseman Chijere Chirwa. 2014. Malawi – Democracy and Political Participation, A review by AfriMAP and the Open Society for Southern Africa. World Bank. 2001. Malawi Local Government Development Project – Implementation Completion Report. Washington DC, USA. World Bank. 2007a. Malawi Poverty and Vulnerability Assessment: Investing in Our Future. Report 36546-MW. Washington DC, USA. World Bank. 2007b. Malawi Public Expenditures Review. Report 40145-MW. Washington DC, USA. World Bank. 2012a. A Snapshot of Decentralization in Malawi: An Overview of Progress, Challenges and Policy Options in 2010 and 2011. Washington DC, USA. World Bank. 2012b. Malawi Water Sector Investment Plan (Volumes I & II). Washington DC, USA. World Bank. 2013. Malawi – Public Expenditure Review. Report No. 79865. Washington DC, USA. World Bank. 2014a. Municipal Finances – A Handbook for Local Governments. Washington DC, USA. World Bank. 2014b. Trends and Drivers of Household Welfare in Malawi 2005-2011: Implications for Policy. Washington DC, USA. World Bank. 2015a. World Development Indicators Online Database. Washington DC, USA. World Bank. 2015b. Malawi Economic Monitor: Managing Fiscal Pressures (March 2015). Lilongwe, Malawi. World Bank. 2015c. Malawi Economic Monitor: Adjusting in Turbulent Times (October 2015). Lilongwe, Malawi. 71 World Bank. 2015d. Malawi Urbanization Review: Challenges and Opportunities in Local Government Financing (Background Paper). 72 Annex I. Economy-Wide Model This Annex details the characteristics of Malawi’s economy and associated assumptions that are taken into account in the functioning of the model. The overview of the model is narrated first, followed by the elaboration of each sub-component. In the model, Malawi’s economy is separated into three spatial units: primary cities, secondary towns, and rural areas. Each spatial unit contains up to 58 different sectors depending on the location of producers in Malawi. Representative producers in each sector and spatial unit produce output by combining land, labor, capital and other inputs according to their production needs or “technologiesâ€?. Farmers, for example, mainly combine crop land with seeds, fertilizers and family labor, whereas the transport sector mainly uses vehicles, petroleum and hired labor. Information on production technologies comes from the NSO’s supply-use table and annual economic surveys, and from the Ministry of Agriculture’s farm budget analysis. These production technologies (or “input-output coefficientsâ€?) capture the rural-urban production linkages discussed in the previous section by indicating to what extent urban sectors use goods and services produced by rural sectors, and vice versa. Using employment data from the household survey (IHS3), rural and urban workers are separated into three education categories – primary, secondary and tertiary – to examine their employability in sectors with different skill requirements. In order to capture seasonal labor shortages and skills scarcity in Malawi, it is assumed that workers are fully employed. This is consistent with the fact that almost all working-age adults in Malawi have at least one job (i.e., working at least four hours per week), even though they could become more productive by working longer hours per week or more weeks per year. Separating workers into education categories is important because it captures how different sectors differ in skill requirements. Urban workers tend to be better educated than rural workers (Section 2.3.2). As such, new migrants from rural areas may find it difficult to compete with better-educated urban workers and so have limited job opportunities in sectors that are less skill-intensive, such as informal trade and manual labor in construction. The model’s detailed technology and labor market specification allows it to capture these important issues. As in most countries, data on internal trade is not available in Malawi and it is assumed that rural and urban producers supply their output to national product markets, where a single national price equates demand and supply. In other words, while rural and urban areas do trade with each other in the model, the exact flow of products moving between each spatial unit cannot be tracked. This could cause the model to exaggerate rural-urban linkages because, for example, increases in rural consumer demand can be supplied by both rural and urban producers. This assumption is less problematic when rural and urban areas have some degree of specialization, because it can be said with greater certainty whether it is rural or urban areas that are supplying particular goods or services. As discussed, there is some specialization in Malawi, with rural areas producing most agricultural products and urban areas producing most of the higher-value services. Non-food-related manufactured goods tend to be imported or produced in urban areas. However, the two large sectors with less specialization are construction and trade services, and this may cause the model to overestimate rural-urban growth linkages. The model captures international or foreign trade. As a result, increased domestic demand can be satisfied by either greater imports or domestic production. Likewise, increased domestic production may lead to more exports rather than increased supply to domestic markets. The decision to import or 73 export is based on changes in the relative price of domestic and foreign goods. There are rigidities to this decision-making due to differences in product qualities and the difficulties in interacting with foreign markets. To capture this rigidity, the model treats domestic and foreign goods as imperfect substitutes. This means that even when prices in foreign markets rise, producers will continue supply domestic markets, but the level of supply may fall and cause domestic prices to rise. Conversely, when import prices fall, consumers in Malawi continue to purchase domestic goods, but the price of these goods fall due to import competition. Finally, from a global perspective, Malawi is a small economy and so its foreign trade decisions are unlikely to influence global commodity prices. To capture this, the model fixes world prices, which is the standard assumption for most developing countries. The model allows for internal migration between rural areas, towns and cities. During the year, workers can only migrate across sectors within their spatial unit. Between years, however, they can migrate between rural and urban areas in response to spatial wage differentials. Migration rates are initially calibrated so as to replicate observed flows of 14,000 new migrants arriving in urban centers each year (Section 2.3.1). Over time, population growth and widening wage gaps can lead to larger absolute migration flows. Ultimately, total labor supply in rural areas, towns and cities depends on previous period labor supply; exogenous population growth; and net internal migration. Importantly, it is assumed that workers migrate with their family based on observed worker dependency ratios. However, it is not assumed that rural-to-urban migration simply leads to higher rural dependency ratios; or that urban centers benefit from new working-age migrants without also having to accommodate and feed their families. Finally, new urban migrants adopt the consumption patterns of existing urban households at similar levels of income, for example, consuming more import-intensive consumption baskets and so urbanization leads to greater demand for imported goods and services in Malawi. Standard accounting principles state that the level of investment should equal the level of private savings less the fiscal deficit. The latter is the difference between the government’s tax revenues and recurrent spending. New capital stocks depend on previous period investment, and are allocated to rural and urban sectors according to profit rate differentials. Once invested, new capital cannot move between rural and urban sectors. It is worth noting that migration tends to increase returns to capital in the receiving region, and so urbanization can attract larger allocations of new capital and cause faster economic growth in urban centers. Urbanization also affects the rate of technical change or productivity growth in each rural and urban areas. Following Henderson and Wang (2005), an assumption is made that agglomeration spillovers are a positive function of population density. As workers move to a city or town they raise the population density and hence productivity in all sectors.92 However, urbanization reduces the amount of public capital available to each urban resident, and, without additional public investment, reduces productivity gains from urban agglomeration. In other words, the model captures what are called “congestion effectsâ€?, which arise when urbanization outpaces urban public investment, leading to overstretched transport systems, inadequate housing, etc. Congestion effects reduce worker productivity by, for example, increasing the time taken to get to work on crowded roads, or increasing morbidity (sick days) because of inadequate water and sanitation infrastructure. The model therefore allows for an “urbanization of povertyâ€? that can arise when the urban economy cannot absorb enough migrant 92 The elasticity linking population density to agglomeration effects is set at 0.08 based on estimates by Rosenthal and Strange (2004). It is assumed that sparsely populated rural areas do not experience agglomeration spillovers. 74 workers, and when urban infrastructure (capital) cannot accommodate enough migrant residents. The allocation of public investments across rural areas, towns and cities is determined outside of the model and is the main policy instrument in the simulations. Finally, given the low population density of rural areas, agglomeration and congestion effects are allowed to occur only in cities and towns. In summary, the model captures the detailed economic characteristics of Malawi’s cities, towns and rural areas, thereby enabling a simulation of alternative future urbanization and investment scenarios, with different implications for growth, employment and poverty (at national, sectoral and spatial levels). Consumer and producer behavior Representative consumers and producers in the model are treated as individual economic agents. It is assumed that households (consumers) make decisions so as to maximize welfare (utility) subject to a budget constraint. For this a linear expenditure system (LES) of demand is employed: (1 − 𝑠𝑎 − 𝑡𝑑𝑎 ) ∙ 𝑌𝑎 (1) 𝑃𝑖 ∙ 𝐶𝑖𝑎 = 𝑃𝑖 ∙ 𝛾𝑖𝑎 + 𝛽𝑖ℎ𝑎 ∙ ( − ∑ 𝑃𝑖′ ∙ 𝛾𝑖′𝑎 ) 𝐿𝑆𝑎 𝑖′ where C is per capita consumption of good i in area a (i.e., cities, towns or rural areas), γ is a minimum subsistence level, β is the marginal budget share, P is the market price of each good, Y is total household income, LS is total labor supply (a proxy for population), and s and td are savings and direct tax rates, respectively. The demand functions allow consumption patterns and income elasticities to vary across households in cities, towns and rural areas. We assume producers maximize profits subject to input and output prices. A constant elasticity of substitution (CES) function determines output quantity X from sector i in area a: −𝜌 −𝜌𝑖𝑎 −1â?„𝜌𝑖𝑎 (2) 𝑋𝑖𝑎 = 𝛼𝑖𝑎 ∙ (𝛿𝑖𝑎 ∙ 𝐿𝑖𝑎 𝑖𝑎 + (1 − 𝛿𝑖𝑎 ) ∙ 𝐾𝑖𝑎 ) where α reflects total factor productivity (TFP), L and K are labor and capital demands, and δ and Ï? are share and substitution parameters. Our production functions permit technologies to vary across producers and areas. Maximizing profits subject to Equation 2 gives the factor demand equations: 𝐿𝑖𝑎 𝑟 ∙ 𝐷𝑖𝑎 1 − 𝛿𝑖𝑎 1â?„(1+𝜌𝑖𝑎) (3) =( ∙ ) 𝐾𝑖𝑎 𝑊 𝑎 𝛿𝑖𝑎 where W is the labor wage in area a, and r is a fixed economy-wide capital rental rate adjusted by a sector/area-specific distortion term D. The factor substitution elasticity is a transformation of Ï?. Higher elasticities means producers can more readily substitute between labor and capital when relative prices change. Intermediate demand is not shown in the equations, although this is included in the model. The producer price PX is the sum of factor payments per unit of output: 𝑃𝑋𝑖𝑎 ∙ 𝑋𝑖𝑎 = 𝑊𝑎 ∙ 𝐿𝑖𝑎 + 𝑟 ∙ 𝐷𝑖𝑎 ∙ 𝐾𝑖𝑎 (4) National product markets and international trade Products are traded in national markets at a single market-clearing price P. The national market assumption is needed because internal trade data is unavailable. Output from each area is combined into a composite national good Q using a CES function: 75 −1â?„í µí¼?𝑖 (5) âˆ’í µí¼? 𝑄𝑖 = 𝜙𝑖 ∙ (∑ 𝜆𝑖𝑎 ∙ 𝑋𝑖𝑎 𝑖 ) 𝑎 Equation 5 permits imperfect substitution between goods from different areas. Relative producer prices are determined by the following first order condition, derived from minimizing the composite supply price of each good: −1 (6) âˆ’í µí¼? âˆ’í µí¼? −1 𝑃𝑋𝑖𝑎 = 𝑃𝑖 ∙ (1 − 𝑡𝑖𝑖 ) ∙ 𝑄𝑖 ∙ (∑ 𝜆𝑖𝑎′ ∙ 𝑋𝑖𝑎′ 𝑖 ) ∙ 𝜆𝑖𝑎 ∙ 𝑋𝑖𝑎 𝑖 𝑎′ where ti is the indirect tax rate applied to domestic sales. This function implies that demand for an area’s output rises when its supply price falls relative to those in other areas. The equations governing international trade are not presented. However, the model permits two-way trade assuming imperfect substitution between domestic and foreign goods. A constant elasticity of transformation (CET) function determines exports and a CES function determines imports. World commodity prices are fixed under a small country assumption. The current account balance is fixed in foreign currency units and the real exchange rate is flexible (i.e., a price index of tradable to non- tradable goods). Government and investment demand Assuming all factors in an area are owned by households in that area, then total income Y is 𝑌𝑎 = ∑ (𝑊 (7) 𝑎 ∙ 𝐿𝑖𝑎 + 𝑟 ∙ 𝐷𝑖𝑎 ∙ 𝐾𝑖𝑎 ) + ℎ𝑎 ∙ 𝐿𝑆𝑎 𝑖 where h is per capita transfer payments from the government. The government is treated as a separate agent. Total domestic revenue is the sum of direct and indirect taxes, as shown on the left-hand side of the following equation: ∑ 𝑡𝑑𝑎 ∙ 𝑌𝑎 + ∑ 𝑡𝑖𝑖 ∙ 𝑃𝑖 ∙ 𝑄𝑖 = ∑ 𝑃𝑖 ∙ 𝐴 ∙ 𝑔𝑖 + ∑ ℎ𝑎 ∙ 𝐿𝑆𝑎 + 𝐵 (8) 𝑎 𝑖 𝑖 𝑎 The government uses revenues to purchase goods and make transfers (i.e., recurrent spending) and to save (i.e., finance public capital investment). This is shown on the right-hand side of Equation 8. The macroeconomic closure for the government account assumes that public consumption spending is equal to base-year quantities g multiplied by an exogenous adjustment factor A. The fiscal balance B adjusts to equalize total revenues and expenditures. A savings-driven investment closure is assumed, i.e., total investment adjusts to the level of savings in the economy. As shown below, a national savings pool finances investment: ∑ 𝑠𝑎 ∙ 𝑌𝑎 + 𝐵 = ∑ (𝑃𝑖 ∙ 𝐼 ∙ í µí±–í µí±?𝑖 + 𝑃𝑖 ∙ 𝐺 ∙ 𝑖𝑔𝑖 ) (9) 𝑎 𝑖 where ip and ig are fixed base-year quantities of private and public investment, respectively, multiplied by adjustment factors I (endogenous) and G (exogenous). For a given level of savings, an increase in public investment G must be matched by a decline in private investment I, i.e., the government “crowds- outâ€? private investors. 76 Factor and product market equilibrium It is assumed that labor is fully employed. As such, total labor supply LS in each area is fixed and, in equilibrium, must equal the sum of all sector labor demands: 𝐿𝑆𝑎 = ∑ 𝐿𝑖𝑎 (10) 𝑖 Unlike labor, which is mobile across sectors, capital is sector/area-specific. Both factor demand K and the economywide rental rate r are therefore fixed (see Equation 3) and the rental rate distortion term D adjusts so that sectoral profit rate equate capital demand and supply. Finally, product market equilibrium requires that the composite supply of each good Q equals total private and public consumption and investment demand: 𝑄𝑖 = ∑ 𝐶𝑖𝑎 ∙ 𝐿𝑆𝑎 + 𝐴 ∙ 𝑔𝑖 + 𝐼 ∙ í µí±–í µí±?𝑖 + 𝐺 ∙ 𝑖𝑔𝑖 (11) 𝑎 Market prices P adjust to ensure equilibrium is achieved. Together, the above 11 equations simultaneously solve for the values of 11 endogenous variables (i.e., C, X, L, D, Q, PX, Y, B, I, W and P). The national consumer price index (CPI) is the numéraire. Capital accumulation The model is recursive dynamic; i.e., it consists of distinct within- and between-period components. The above equations specify the within-period component. Between-periods, exogenous variables and parameters are updated based on externally-determined trends and previous period results. The processes of capital accumulation, labor migration, and agglomeration and technical change are described. While not shown in Equations 1-11, each variable has a time subscript associated with it. Sector-level capital accumulation is determined endogenously based on previous period investment. As shown in Equations 12-14, the quantity of new capital N is based on the value of private investment and the capital price PK (i.e., a composite price derived from investment demand shares ip). New capital is allocated to sectors/areas after applying a national depreciation rate Ï… and according to a capital allocation factor SK (0