Report No. 112433-PE Peru: Building a More Efficient and Equitable Fiscal Decentralization System Macroeconomics and Fiscal Management Global Practice Latin America and the Caribbean Region MARCH 2017 Contents Abbreviations and Acronyms . . . . . . . . . . . . . . . . . . . . . . v Enhancing the equalizing impact of FONCOMUN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Reducing the volatility of resource-revenue Chapter 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Chapter 2: An Overview of Fiscal Decentralization Appendix 1: Methodology for New Ordinary in Peru . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Resources and Simulation . . . . . . . . . . . . . . . . . . . . . . 63 Chapter 3: Strengthening the Institutional Appendix 2: Current Methodology Arrangements for Decentralization . . . . . . . . . . . . . . . 15 of the FONCOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Chapter 4: Clarifying Expenditure Appendix 3: Methodology of the New Responsibilities and Reducing the Investment FONCOR Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Bias of Local Spending . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Appendix 4: The Municipal Compensation Fund Chapter 5: Improving Subnational Taxation . . . . . . 27 (FONCOMUN) and the Proposed Reform . . . . . . . . . . 73 Chapter 6: Toward More Efficient, Equitable, References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 and Stable Intergovernmental Transfers . . . . . . . . . .35 List of Boxes Ordinary Resources and FONCOR . . . . . . . . . . . . . . . 38 Box 1: Partial decentralization in the health FONCOMUN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 A Stabilization Fund for Resource-Revenue Box 2: The Brazilian basic education fund Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 (FUNDEB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Chapter 7: Conclusions and Policy Box 3: International experience in multilevel indirect Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 taxation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 The Institutional Framework . . . . . . . . . . . . . . . . . . . . 57 Box 4: Prioritization criteria of regions . . . . . . . . . . . . . 67 Consolidating regional governments at the List of Figures permanent intermediate government level . . . . . 57 Figure 1: Regional disparities in per capita GDP Reducing municipal fragmentation . . . . . . . . . . . .58 and access to basic services . . . . . . . . . . . . . . . . . . . . . . . . 4 Expenditure Assignments . . . . . . . . . . . . . . . . . . . . . . 58 Figure 2: The fiscal decentralization diamond . . . . . . . . 5 Clarifying expenditure responsibilities . . . . . . . . . 58 Figure 3: Subnational spending as a share of total Taxation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 government spending, 2012 . . . . . . . . . . . . . . . . . . . . . . . . 6 Assigning tax bases to regional governments . . 58 Figure 4: Public investment by local governments as a share of GDP, Peru and selected comparators, Improving the efficiency of revenue collection 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 at the local level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Figure 5: The vertical fiscal gap, Peru and OECD Intergovernmental Transfers . . . . . . . . . . . . . . . . . . . . 59 countries, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Establishing a more stable, less discretionary, Figure 6: The evolution of regional government and more equalizing system of transfers revenues (in constant 2007 PEN billions) . . . . . . . . . . . . 10 to regional governments . . . . . . . . . . . . . . . . . . . . . 60 i ii    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 7: Revenue structure of regional governments, Figure 25: Relative gains and losses from the 2007–14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 proposed changes in FONCOMUN transfers (in PEN per capita) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Figure 8: The evolution of local government revenues (in constant 2007 PEN billions) . . . . . . . . . . . . 11 Figure 26: Annual resource-revenue transfers, actual and stabilization-fund scenario, 2004–2014 Figure 9: The revenue structure of local governments, (in constant 2007 PEN millions) . . . . . . . . . . . . . . . . . . . 55 2007–14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure 27: New methodology to distribute Figure 10: Per capita resource-revenue transfers FONCOMUN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 to regional and local governments, 2004–14 . . . . . . . . 12 Figure 11: Average municipal population size, Peru List of Tables and selected comparators . . . . . . . . . . . . . . . . . . . . . . . . . 16 Table 1: Intergovernmental transfers in Peru . . . . . . . . . 9 Figure 12: Property-tax-to-GDP ratio, Peru Table 2: Per capita spending among provincial and selected comparators . . . . . . . . . . . . . . . . . . . . . . . . 28 and district municipalities . . . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 13: Distribution of resource-revenue transfers Table 3: Municipalities by population, 2012 . . . . . . . . . . 16 among the top 1, 5, and 10 percent of recipient municipalities, 2011 (in PEN thousands) . . . . . . . . . . . . 36 Table 4: Subnational government revenues as a share of GDP, 2004–14 . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 14: Subnational per capita spending in constant 2014 PEN and regional poverty Table 5: Potential regional government revenue rates in 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 derived from a 1–3 percent flat rate income-tax surcharge, 2015 (in PEN thousands) . . . . . . . . . . . . . . . . 30 Figure 15: Per capita spending by regional governments, 2014 (in constant 2007 PEN) . . . . . . . . 38 Table 6: Share of ordinary resources represented by the potential tax revenue from the income-tax Figure 16: Per capita spending by municipalities, surcharge, 2015 (in PEN thousands) . . . . . . . . . . . . . . . . . 31 2014 (in constant 2007 PEN) . . . . . . . . . . . . . . . . . . . . . . 38 Table 7: Subnational indirect taxes . . . . . . . . . . . . . . . . . 32 Figure 17: Relation between per capita ordinary resources and per capita GDP . . . . . . . . . . . . . . . . . . . . . 39 Table 8: Simulations transforming the system for ordinary resource transfers into a revenue- Figure 18: The current distribution of ordinary sharing-plus-FONCOR equalization grant, 2016 revenue transfers, 2016 (in PEN millions) . . . . . . . . . . . 42 (in PEN millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Figure 19: The distribution of ordinary resources Table 9: Simple regressions of current ordinary under the new revenue-sharing mechanism resource transfers and transfers under scenarios and the original FONCOR distribution formula one and two on GDP per capita . . . . . . . . . . . . . . . . . . . . 47 (in PEN millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Table 10: Simple regressions of FONCOMUN Figure 20: The distribution of ordinary resources under transfers on household expenditure per capita . . . . . . 53 the new revenue-sharing mechanism and the new FONCOR distribution formula (in PEN millions) . . . . . 44 Table 11: Annual resource-revenue transfers, actual and stabilization-fund scenario, 2004–14 Figure 21: Relative gains and losses from the (in constant 2007 PEN millions) . . . . . . . . . . . . . . . . . . . 56 introduction of the proposed revenue-sharing mechanism and the original FONCOR distribution Table 12: New ordinary resources distribution formula (in PEN millions) . . . . . . . . . . . . . . . . . . . . . . . . . . 45 index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Figure 22: Relative gains and losses from the Table 13: FONCOR distribution index . . . . . . . . . . . . . . . 67 introduction of the proposed revenue-sharing Table 14: Probabilistic—Logit regression . . . . . . . . . . . . 68 mechanism and the new FONCOR distribution formula (in PEN millions) . . . . . . . . . . . . . . . . . . . . . . . . . 46 Table 15: Calculation of the FONCOR index for 2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Figure 23: Geographical distribution of existing FONCOMUN transfers (in PEN per capita) . . . . . . . . . . 50 Table 16: Financing sources of FONCOMUN . . . . . . . . . 73 Figure 24: Geographical distribution of the proposed Table 17: FONCOMUN: Distribution criteria changes in FONCOMUN transfers (in PEN (weights) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 per capita) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 This report was prepared by Fernando Blanco (Lead Economist, GMFDR), Jorge Martinez-Vazquez (Regents Professor, Department of Economics, International Center for Public Economics, Georgia State University), and Janet Porras Mendoza (Georgia State University). Maryam Ali-Lothrop and Sean Lothrop edited the report. The authors would like to thank Pablo Saavedra (Practice Manager, GMFDR), Pedro Luis Rodriguez (Program Leader, LCC6C), and Marc Tobias Schiffbauer (Senior Economist, GMFDR) for their comments and support. Special thanks is due to Gonzalo Neyra from Peru’s Ministry of Economy for his valuable advice and assistance. iii Abbreviations and Acronyms CND Decentralization National Council CTAR Transitory Council of Regional Administration DGAES General Directorate of Economic and Social Affairs ENAHO National Household Survey (Encuesta Nacional de Hogares) FOCAM Socioeconomic Development Fund of the Camisea Project (Fondo de Desarrollo Socioeconómico de Camisea) FONCOMUN Municipal Compensation Fund (Fondo de Compensación Municipal) FONCOR Regional Compensation Fund (Fondo de Compensación Regional) FONIPREL Fund for the Promotion of Regional and Local Public Investment (Fondo de Promoción a la Inversión Pública Regional y Local) FUNDEB Brazilian Basic Education Fund (Fundo de Manutenção e Desenvolvimento do Ensino Basico e Valorização do Magistério) GDP Gross Domestic Product GFS Government Finance Statistics (IMF) IGM Municipal Management IMF International Monetary Fund INADE National Institute of Development INEI National Statistical System (Instituto Nacional de Estadística e Informática) IPIM Municipal Promotion Tax MEF Ministry of Economy and Finance (Ministerio de Economía y Finanzas) MINSA Ministry of Health (Ministerio de Salud) MML Provincial Municipality of Lima OECD Organisation for Economic Co-operation and Development PCM Secretary of Decentralization PEN Peruvian Nuevos Soles PIMGMM. Municipal Improvement Incentive Program (Programa de Incentivos a la Mejora de la Gestión y Modernización Municipal) PpR Result-based-Budgeting (Presupuesto por Resultados) SIS Main health-insurance program (Seguro Integral de Salud) SNGs Subnational Governments SNIP National System of Public Investment (Sistema Nacional de Inversion) SUNAT National Tax Administration (Superintendencia Nacional de Administración Tributaria) VAT Value-added Tax v Executive Summary Over the past two decades, Peru has achieved held in November of the same year. The 2004 remarkable economic success. Average annual Fiscal Decentralization Law established the Gross Domestic Product (GDP) growth has exceeded policy actions and fiscal resources necessary to 5 percent since 2001. Poverty has been consistently complete the decentralization process and defined reduced, and sustained improvements have been the implementation schedule for transferring observed in social and human development. The expenditure responsibilities to municipal and poverty incidence rate fell from 58 to 23 percent regional governments. between 2004 and 2014, and households’ incomes at the bottom 40 percent grew 50 percent faster than The main objectives of this process were to the national average. The structural transformation increase the overall efficiency of the public sector of Peru’s economy striking fast and widely shared and strengthen the democratic decision making growth transformed Peru into an upper-middle- by enhancing regional and local governments. income and diversified economy. The rationale for fiscal decentralization was that proximity between citizens and government will Peru is now embarking on a new stage in its enable public services to be better tailored to the development, and as an emerging upper- specific needs of local communities. Also, enhanced middle-income country with aspirations to accountability should improve efficiency and quality become a high income country, it will face new of public spending. Finally, an important motivation challenges at the same time that it will need to in Peru was to achieve a more regionally balanced address long standing development challenges. economic growth favoring increased economic To continue growth in a context of unpredictable dynamism and improving socioeconomic conditions external markets, Peru’s development model must in the less developed regions of the country. transition from one based on factor accumulation to one based on productivity and human capital However, the implementation of Peru’s fiscal enhancements. At the same time, to attenuate decentralization stalled in 2005, when the the large regional disparities that have branded creation of macro-regions was defeated in a the country’s socioeconomic development, a more referendum. Under the original decentralization equitable, efficient, and tailored provision of public plan, the 26 regional governments were to be goods and services across the country will need consolidated into 12 macro-regions, which would to be supported by enhancing the current fiscal serve as an intermediate level of government, decentralization arrangements. but without clear spending responsibilities and lacking own revenue sources and tax assignments. Peru began a process of political and fiscal Moreover the establishment of a new revenue- decentralization in the early 2000s. The 2002 sharing mechanism for the income tax and value- Constitutional Reform transferred political power added tax (VAT) could not be implemented, as it to subnational jurisdictions, and elections for was conditional on the formation of the macro- newly established regional governments were regions, which made regional governments fully vii viii    Peru: Building a More Efficient and Equitable Fiscal Decentralization System dependent on discretionary transfers from the recognize that enhancing internal and external national government. controls is a necessary condition for improving accountability, transparency, and efficiency A second important variable affecting the in decentralized service delivery. In the same country’s decentralization process was the direction, recent evidence has shown that a heavy unexpected (and significant) revenue windfall reliance on transfers may also be associated with that extractive industries generated for the excessive subnational indebtedness. Effective fiscal central government and a few municipalities rules supported by sound accounting systems over the past decade. As a result of the revenue- and meaningful enforcement are critical to keep sharing mechanism between the central, regional, subnational finances on track.1 Finally, it is worth and local governments for the corporate income tax mentioning that this report is not intended as a generated by extractive industries and the boom in substitute to the preparation of a detailed blueprint commodity prices, an unintended decentralization for enhancing the current fiscal decentralization of fiscal revenues was spurred in the second part of framework, but it aims at providing policy options the 2000s. Moreover, the adoption of the derivation on specific areas and tools to simulate impacts of principle (origin of the resources revenues) as a proposed reforms. distribution criteria resulted in few municipal governments receiving vast amounts of resources, especially in gas and in mining areas, even though Institutional Arrangements the economies of scale of the public infrastructure Peru’s fiscal decentralization framework remains they provide and their size, institutional capacity, uneven and incomplete. Regional governments and spending needs were not commensurate to the have assumed some of the functions originally resources availability. assigned to the macro-regions and are the de This report analyzes recent trends of the fiscal facto intermediate level of government. In theory, decentralization process in Peru and presents a set they are responsible for providing or coordinating of reform options designed to harvest the envisaged the delivery of public goods and services with efficiency and equity gains in service delivery spatial spillovers across local jurisdictions, and for that the fiscal decentralization was expected to coordinating policy priorities and service delivery bring. The analysis and policy options are presented between the local and national authorities. They in a conceptually logical order: (i) departing from also prepare the formal development program for institutional arrangements in the vertical structure their region. However, unlike what was envisaged of subnational governments passing to (ii) the need for the macro-regions, the regional governments of a clearer definition of spending responsibilities lack own-revenue sources or a transparent and among levels of government that needs to be stable revenue-sharing mechanism. Moreover, followed by (iii) a commensurate redefinition of regional governments have limited autonomy revenue assignments and (iv) enhancing equalization to allocate the transfers they receive from the role of the transfer system. central government, adjust public services to suit the local context, or coordinate the provision of The report does not cover other important topics public goods and services between municipalities of the fiscal decentralization agenda such as in their jurisdictions. This is particularly important subnational accountability and transparency for capital spending, where regional governments and subnational borrowing restrictions. The could bring scale to individual investments report looks at expenditure responsibilities, by municipalities, but can also apply to social however it does not provide detailed analysis at programs. the sector specific level. While accountability and transparency can be fostered by expanding the ability of subnational governments to collect 1  At the time of completion of this report (December 2016), their own revenues and by improving the design the government enacted a new regulation establishing a fiscal of intergovernmental transfers, it is important to rule for subnational governments. Executive Summary    ix The failure to establish macro-regions service delivery platforms or infrastructure prompted the central government to shift focus projects with economies of scale that serve from decentralization to administrative de- several jurisdictions at the same time to improve concentration. Today, two types of situations efficiency. A case in point is investment spending. prevail. First, central government line ministries The number of investment projects at the local might retain direct control over the provision of level increased from 2,100 in 2004 to more than public goods and services and over budget allocation 15,000 in 2014, and the average costs per project decisions, but their regional directorates or other shrank; currently the average cost per project is administrative arrangements have been gradually less than 1 million Peruvian Nuevos Soles (PEN). incorporated into regional governments. However, The large number of small projects suggests that the regional governments have very limited control investment decisions are uncoordinated, limiting over the directorates, which typically remain under positive spatial spillover effects. Unfortunately, the effective authority of the line ministries and are incentives for consolidating projects within the financed through annually determined allocations public investment system (SNIP), e.g., the extra of Ordinary Resources (Recursos Ordinarios) from points given in project evaluation to regionally- the central government budget based on historical bundled projects, do not compensate for the strong norms—although a small proportion is allocated via incentives provided by the fiscal decentralization incentives for certain actions undertaken by regional framework. governments. Second, a central government line ministry holds all budget resources, but transfer Perhaps most important is the unclear them, generally under opaque rules, throughout the distribution of responsibilities across different fiscal year to subnational governments for them to levels of government, which has further implement specific projects. weakened the effectiveness of decentralization. A large number of functions are either shared or Meanwhile, a high and increasing degree of delegated between different government levels fragmentation and lack of scale at the local level rather than assigned to a specific level. While some is undermining the efficiency of decentralized degree of shared responsibilities is normal, formal service delivery and productive infrastructure. mechanisms for intergovernmental coordination Many municipalities are too small to have the must be established, such as binding master plans, scale for service delivery, and their institutional cofinancing arrangements, conflict-resolution and technical capacities are inadequate. Also, their forums, and joint implementation mechanisms. scope for local revenue collection is limited, leaving Yet despite the extensive legal framework, the them heavily dependent on intergovernmental decentralization framework does not yet include fiscal transfers. Most importantly, the otherwise formal institutions and procedures to support well-designed revenue-sharing transfer to local intergovernmental coordination. In principle, the governments, the Municipal Compensation Fund regional directorates of the line ministries should (Fondo de Compensación Municipal, FONCOMUN), has facilitate coordination between the national and two main shortcomings. First, it provides incentives regional governments, but in practice this is not the to subdivide municipal jurisdictions rather than case, and coordination with local governments is incentives for consolidation. As a result, the number even less well structured. of local government units (and the associated fixed costs of public administration) have been steadily Expenditure and revenue increasing over the past several years. Second, decentralization: Vertical it does not account for own-revenues and, thus, and horizontal gaps even rich municipalities or municipalities that are beneficiaries of large resource transfers in gas and Peru’s intergovernmental fiscal framework mining areas still receive FONCOMUN transfers. reveals a high degree of asymmetry between revenue and expenditure decentralization. The fragmentation and the current transfer The “fiscal decentralization diamond” presented system provide little or no incentives to create below provides a simple conceptual model for x    Peru: Building a More Efficient and Equitable Fiscal Decentralization System understanding trends in fiscal decentralization. also affects revenue incentives and influences While certain expenditure responsibilities have been the quality of subnational spending and tax- either deconcentrated or decentralized (particularly collection effort. with regard to recurrent spending programs), the central government continues to collect the • The horizontal fiscal gap defined as the imbalances between expenditure needs and overwhelming majority of public revenue. As a a regional or local government ability to result, regional and local governments are heavily raise revenues. The imbalances are reflected reliant on intergovernmental transfers. This is in disparities in per capita spending across especially true among subnational governments in jurisdictions, implying that service delivery can less developed regions of the country, and in some vary significantly across the territory depending cases transfers finance more than 95 percent of on where citizens actually reside. Equalizing subnational spending. Due to their limited access transfers are typically used to reduce horizontal to credit markets and tight restrictions, regional fiscal gaps, but again, unless they are carefully and local government borrowing is limited, albeit designed, they can affect the efficiency of has been increasing. the recipient governments’ spending and own revenue collection. In particular, asymmetries between revenue decentralization and expenditure decentralization Vertical fiscal gaps are common in fiscal generate two types of gaps that are today present decentralization arrangements; the problem is in Peru to various degrees: that the vertical fiscal gap in Peru has become • The vertical fiscal gap defined as the asymmetry excessively large and places a number of serious challenges. Peru has one of the largest vertical between subnational revenue generation and fiscal gaps among comparable countries. At less spending responsibilities. If in deficit, this gap than 1 percent of GDP, subnational own-source increases subnational governments’ reliance revenues—which include taxes, user fees, and other on intergovernmental transfers (possibly minor revenue streams—represent 5 percent of motivating excessive transaction costs and total public revenue, yet subnational governments “lobbying” in the system). This gap can be were responsible for a full 40 percent of total public bridged by tax revenue sharing and other spending in 2014. As noted above, large vertical intergovernmental transfers or by borrowing. fiscal gaps tend to reduce expenditure quality, At the same time, the availability of transfers The fiscal decentralization diamond Transfers Transfers Spending Taxes Spending Taxes Debt Debt Peru Peru Brazil Mexico Source: World Bank staff. Executive Summary    xi discourage revenue mobilization by recipient to at least at two of four basic services—education, governments, and undermine fiscal discipline. health care, sanitation, and housing—compared to fewer than 2 percent of the population in Lima, The similarly large horizontal fiscal gap is Ancash, and Moquegua. reflected in the uneven coverage and quality of basic public goods and services across regions. Resource-revenue sharing transfers (canon, The concentration of economic activity in Lima, sobre canon, and royalties) exacerbate horizontal Callao, and a few resource-rich provinces results fiscal imbalances. These revenues are distributed in a highly uneven distribution of local revenue according to their point of origin, and as bases. Fiscal transfers do little to reduce regional municipalities receive the bulk of these transfers, imbalances in revenue capacity or equalize the fiscal inequality is especially pronounced at the supply of public goods and services. More than municipal level. For example, over 25 percent of 15 percent of the population living in the Andean canon mining-revenue transfers go to just 1 percent and Amazonian regions, which together cover of municipalities. Moreover, district municipalities almost 70 percent of Peru’s territory, lack access receive more than 50 percent of resource-revenue Regional disparities in per capita GDP and access to basic services 7.3 19.9 17.8 7.9 2.1 4.9 10.1 1.3 2.3 5.8 15.5 7.2 5 8.6 Share of the 1.1 population with at 4.1 2.7 least 2 out of 4 unmet basic needs 1.2 7.9 2.8 according to the 5.1 2014 Unmet Basic Needs Index. 1.7 GDP per capita 2014 (nuevos soles) 2014 1.7 5,935–8,737 16%–20% 1 8,738–11,551 8%–15% 11,552–19,770 6%–7% 19,771–48,616 1%–5% Sources: INEI and MEF. Note: The Unmet Basic Needs Index estimates the share of the population with access to adequate housing, sanitation, basic education, and a subsistence-level income. xii    Peru: Building a More Efficient and Equitable Fiscal Decentralization System transfers in each region where an extractive part of its provincial competences, the MML has industry is located. As a result, the wealthiest responsibilities over the 43 districts that are part provincial municipality spends 80 times more, per of the province of Lima. Beyond the issue of greater capita, than the poorest provincial municipality, coordination, which could involve a “metropolitan while the wealthiest district municipality spends area authority,” Lima also raises the question of 250 times more than its poorest counterpart. the need for a special fiscal regime—regarding functional responsibilities and also revenue Expenditure responsibilities sources—a need that can be extended to other large urban centers in the country. The administrative and expenditure respon­ sibilities of subnational governments are still Limited institutional and technical support not clearly defined. The 2002 Constitutional and capacity prevents regional governments Reform, the organic laws for regional and local from expanding their expenditure capacity. governments, and sector-level regulations all Under the original decentralization strategy, assign administrative functions across government institutional capacity-building plans were designed levels. However, an excessive number of shared to boost the ability of regional governments to functions and a lack of precise assignments spend resources efficiently. Once a subnational create overlapping responsibilities rather than government was deemed capable of assuming complementary roles or concurrent competencies. its expanded role, annual plans would be drafted This ambiguity undermines public accountability by detailing the specific functions, sub-functions, making it difficult to determine which government programs, and actions to be undertaken, as well as level is ultimately responsible for the delivery of the fiscal resources, human resources, and capital specific services. assets to be transferred, along with a timetable for completing the process. However, this framework The situation is worsened by the atypical feature has largely been abandoned, and subnational of Peru’s decentralization system of having governments have been receiving resources and two categories of local government (provincial assuming responsibilities regardless of their ability municipalities and district municipalities) with to manage them effectively. often overlapping responsibilities. A conspicuous case of the need for greater coordination is Tax assignments the case of “Metropolitan Lima,” which refers to the geographic area that includes the 43 The decentralization system has not significantly districts of Lima Province and the six districts enhanced the own-source revenue capacity of of the Constitutional Province of Callao. This subnational governments. The 2002 and 2004 all involves a complex governance system. The fiscal decentralization laws, and all subsequent provincial municipality of Lima (MML) itself has a legislation, did not alter the original allocation of tax special regime that combines district, provincial, ownership between different levels of government. and regional competences.2 The Lima district As a result, the municipal tax bases have not (Cercado de Lima) is run directly by the MML. The changed since 1993 and remain limited to the urban Lima Metropolitan Council exercises powers and property tax (predial), the real estate transaction functions equivalent to the Regional Council, and tax (alcabala), and other minor taxes on lotteries, the mayor of the MML exercises the powers and entertainment, and gambling. In particular, functions equivalent to the Regional President. As district municipalities collect property taxes and taxes on the transfer of real estate. Provincial municipalities collect taxes on motor vehicles and 2  In the region (department) of Lima, the regional government on public entertainment, lotteries, and other forms does not have authority over the province of Lima. The election of gambling. Regional governments have no tax of regional authorities in the region of Lima does not include the province of Lima, but rather only the other nine remaining authority at all, as they were established to be provinces of the Lima region. The regional seat or capital city transitory entities that would be subsumed into of the Lima region is Huacho. macro-regions. Regional governments collect all of Executive Summary    xiii Property-tax-to-GDP ratio, Peru and selected comparators 1.2 1.0 0.8 0.6 0.4 0.2 0 a il le a a o y ru y C e az ua ua in bi al ic ag LA hi Pe ex nt om m Br C g g er ge te ra ru M av ol ua Pa U Ar C D G EC O 2008 2010 2013 Source: OECD. their own-source revenue from user fees and other Nevertheless, local tax collection in Peru has small revenue sources. improved in recent years. Property-tax revenue has risen from 0.17 percent of GDP in 2002 to In Peru, subnational taxes represent around 0.24 percent in 2015. This likely reflects a variety 0.45 percent of GDP, a lower share than in of factors. In the first place there is the role of the comparable countries in Latin America. This recent real estate property boom, which has also results from a combination of limited subnational been reflected in the tax revenue collections from tax powers and low collection efficiency. The the property transfer tax (alcabala). The technical largest tax bases—income and consumption— and institutional support from the national are leveraged by the central government, and government have also played a role. However, subnational governments are restricted to improvements in revenue collection have been exploiting more marginal revenue sources, such offset by reductions in other revenue streams, as property taxes, various administrative charges, and total subnational own-source revenues have and user fees for public services. Moreover, remained broadly constant over time. insufficient administrative capacity prevents the subnational government from fully exploiting their Intergovernmental transfers limited revenue base, and a heavy reliance on intergovernmental transfers weakens incentives While the intergovernmental transfer system to strengthen tax collection. Even though local closes the vertical fiscal gap, it does little governments have larger tax powers than regional to generate adequate incentives to improve governments, local tax revenues are very low. fiscal (revenue) effort and more expenditure Most local governments do not have cadasters of efficiency at the subnational level. The use of properties or if they have they are not updated. ordinary resource transfers as the main revenue Problems with the registration of vehicles have source for regional governments presents several also constrained the ability to collect revenues important drawbacks. First, since these transfers from this tax source. The share of local property- are based on historical or inertial criteria, such as tax revenues in Chile and Colombia are both far existing expenditure needs, regional governments higher than in Peru at about 0.8 and 0.7 percent have no incentives to generate efficiency gains of GDP, respectively. Peru’s property-tax-to-GDP in their use (i.e., less inputs, less money). Second, ratio is comparable to those of Paraguay, Mexico, the lack of a clear distribution criteria for ordinary and Guatemala. resource transfers and also the lack of a rule to xiv    Peru: Building a More Efficient and Equitable Fiscal Decentralization System determine the amount of funds to be distributed, within each province according to their rurality, creates unpredictability of regional governments’ territory, and municipal management capacity. budgets, fosters excessive negotiations between Due to this two-stage distribution, it is possible subnational governments and MEF, and increases that two identical municipalities (same population, administrative procedures and management unmet needs, rurality, territory, and capacity) may costs. Third, and most importantly, a substantial receive different transfer amounts just because share of ordinary resource transfers is determined they are located in two different provinces (with throughout negotiations during a given fiscal different socioeconomic conditions). In the third year, generating month-to-month uncertainty, stage two adjustments to the amounts defined in negatively affecting annual budget execution, the two stages are undertaken: a minimum transfer and further undermining medium-term budget level (eight tax units) and the application of the planning and spending for results. They also harmless clause that ensures that municipalities introduce an element of unfairness, as the relative should receive at least the same amount (in real bargaining power of different regional governments terms) that they had received in 2009. After these likely affects their transfer allocations. Finally, the two adjustments the FONCOMUN equalizing lack of a clear rule for determining the total pool of formula distribution may be further dissipated. funds to be transferred annually creates additional Moreover, the minimum transfer received by all uncertainty. municipalities also encourages local administrative fragmentation. Due to the minimum FONCOMUN The central government provides two equalizing transfer level, two municipalities can significantly transfers to regional and local governments, increase the transfers they receive by splitting, as yet their relatively small size and the specifics compared to when they were one. of their distribution formulas reduce their equalizing impact. The Regional Compensation The increasing size of resource-revenue transfers Fund (Fondo de Compensación Regional, FONCOR) is has shifted public investment responsibilities a capital-budget transfer to regional governments to subnational governments, yet municipalities that promotes equalization by basing allocations often have limited capacity to execute investment on regional investment needs and revenue capacity. projects and face weak incentives to build It is well defined and its distribution criteria is interjurisdictional infrastructure. Canon proceeds clearly equalizing, but its small size (currently at increased from 0.4 percent of GDP in 2004 to 690 million of PEN) renders it useless—creating 2 percent of GDP in 2012, though the recent decline in scope for discretionary “ordinary” transfers, rather global commodity prices reduced them to 1.4 percent than rule-based allocations. FONCOMUN is the in 2015. By law, resource-revenue transfers are largest transfer for municipalities to cover recurrent earmarked for infrastructure investment. As the costs. It is also a well-designed equalizing transfer, amount of these transfers increased over the last except that its distribution formula does not decade, subnational governments eclipsed the include revenue generation capacity, which enables national government as the primary source of wealthy municipalities to receive FONCOMUN public investment. Local governments now account resources. for 47 percent of total public investment, while regional governments account for 23 percent. Certain aspects of FONCOMUN distribution However, the limited administrative capacity of criteria create additional distortions that subnational governments diminishes the quality weaken its equalizing effects. FONCOMUN has a of capital spending. While execution rates for complex three-stage distribution procedure. In the investment projects have improved, regional and first stage, resources are allocated at the provincial local governments still execute an average of less level according to equalizing criteria (socio- than 80 percent of their investment budgets. Low economic conditions reflected by population and execution rates and poor implementation quality unmet needs for public services). The second stage negatively affect the efficiency of investment distributes the resources allocated in the first phase spending. Moreover, individual municipalities at the provincial level to the district municipalities have limited incentives to invest these resources Executive Summary    xv Distribution of resource-revenue transfers Subnational per capita spending in constant among the top 1, 5 and 10 percent of recipient 2014 PEN (left) and regional poverty rates municipalities, 2011 (in PEN ten thousands) in 2010 (right) 450 6,000 70 400 60 5,000 350 50 300 4,000 Nuevos soles 40 Percentage 250 x 10,000 3,000 200 30 2,000 150 20 100 1,000 10 50 0 0 0 Ta s eq a U ipa La M ali be n C P d am ra C ca ac co ua á o o ov ica s* io Ar cn Li artí rta H Hu ch m nca uc aj iu ce n y ar Ay us D pr vel u Sa ca 03 04 05 06 07 08 09 10 11 u n in de 20 20 20 20 20 20 20 20 20 re ad Top 1% recipients a M Li Top 5% recipients Subnational per capita spending Top 10% recipients Poverty rate (2010) Canon minero (total GL) Sources: MEF and INEI. Source: Erman, 2015. *The poverty rate of metropolitan Lima, Lima provinces and Callao is not disaggregated; INEI reports a rate of 13.5 percent for Lima and Callao regions. in interjurisdictional infrastructure, which Peru urgently needs. The existing legal framework for pooling resources is weak, and most investment In addition, the allocation criteria for ordinary projects are small and generate limited economic resource transfers to regional governments do returns. The failure of the establishment of not account for regional expenditure needs or macro-regional governments and the absence revenue capacity. While FONCOR does account of a coordination entity at the national level for resource-revenue transfers, its small size limits coordinating subnational governments’ investments its equalization effect. In addition, the inherent have prevented the seizing of economies of scale volatility of resource revenues has greatly increased and spillovers associated to larger infrastructure the unpredictability of subnational government projects. budgets, further undermining the efficiency of both capital and current expenditures. Resource-revenue transfers have exacerbated budgetary volatility and increased horizontal inequalities. While FONCOMUN and other A Summary of Reform Options intergovernmental transfer mechanisms are based on equalization criteria, the enormous size and Peru has several policy options to improve its vastly uneven distribution of resource revenues fiscal decentralization system. Below we present effectively eliminates the transfer system’s these options as follows: (i) the institutional capacity to attenuate interregional disparities. arrangements determining the relationship xvi    Peru: Building a More Efficient and Equitable Fiscal Decentralization System between different levels of government; (ii) the d. Fully incorporate the regional directorates importance of clarifying expenditure responsibilities of line ministries into the institutional at all government levels; (iii) the commensurate structure of regional governments; and redefinition of the tax authority to boost e. Eliminate the current system of ordinary subnational own-source revenue capacity and resource transfers and replace it by enhance collection efficiency; and (iv) proposals to (i) assigning tax bases to regional increase the transparency, stability, and equalizing governments and (ii) establishing an impact of the intergovernmental transfer system. unconditional revenue-sharing transfer (see options on taxation and intergovernmental The institutional framework transfers presented below). Reforms should acknowledge that the macro- 2. Reducing municipal fragmentation regions are unlikely to be established, and focus on strengthening the existing regional a. Declare a moratorium on the creation of governments. While important progress has been new municipalities; made, the decentralization process has been slow Complete the legal demarcation of b. and uneven. In this context, future efforts should municipal boundaries to end territorial focus on the consolidation of regional governments disputes, which are the most important as the intermediate unit of government that loophole allowing for the creation of new articulates policies and interventions for a region municipalities; and serves as the interphase between the central and local governments. A second stream of reforms c. Tighten regulatory requirements to further under this area is to stop the further creation of discourage the establishment of new new municipalities and develop mechanisms to municipalities; foster consolidation. Significantly d. reduce the minimum FONCOMUN transfer level, diminishing the 1. Consolidating regional governments at the incentive to create new municipalities; and permanent intermediate government level e. Design stronger incentives for municipal Review the legislative framework for a. consolidation and cooperation in service regional governments to formalize their delivery. role as a full-fledged intermediate-level government; Expenditure assignments b. In parallel establish and enhance institutions for national-regional coordination in decision- A clear definition of expenditure responsibilities making processes (e.g., establishment of an will be necessary to increase the autonomy and entity at the national government responsible accountability of subnational governments for the coordination with regional and local and improve the efficiency of service delivery. governments, supporting the recently created In particular, further clarification is needed of the Association of Regional Governments); shared or concurrent functions, by disentangling sub-functions to define what level of government is c. Prioritize the resumption of institutional ultimately responsible for the sub-function without capacity-building efforts at the regional ambiguities. A clear differentiation between level, identifying a suitable mechanism at deconcentrated and delegated functions and the the central level to deliver such programs definition of financing sources (own-revenue, block (the entity at the national government grants, or conditional grants) in each case would responsible for the coordination with other also be needed. As this work tends to be inherently levels of government may be responsible dependent on the nature of the sector (e.g., health, for capacity building and accreditation of roads, etc.), a commission comprised of sectoral subnational governments); Executive Summary    xvii and fiscal experts will probably need to be mobilized consideration is also important because Peru has for each sector. a relatively small tax revenue base as a whole and, thus, only sharing or devolving existing taxes might 3. Clarifying expenditure responsibilities be at the expense of central government’s ability a. Review the organic laws for regional and to address national-level priorities if the overall tax local governments to identify sectors with base is not expanded. duplicative or overlapping responsibilities; 4. Assigning tax bases to regional governments Unbundle shared responsibilities into b. a. Establish a regional surcharge on the sub-functions (regulation, financing, and personal income tax collected by the service delivery) and allocate responsibility national government to be levied at a flat for each sub-function to the appropriate rate to minimize interregional labor factor government level following the subsidiarity movements; principle; b. Enable regional governments to define a Eliminate c. overlapping responsibilities local rate for the surcharge of 1–3 percent; between regional and local governments, and assigning them to the appropriate level, and balancing economies of scale and the c. Explore other tax-policy options, including subsidiarity principle; a presumptive income business tax, or surcharges on existing national excise d. Streamline the framework for delegated taxes or on potentially new ones.3 responsibilities; and 5. Improving the efficiency of revenue collection Establish e. permanent formal inter­ at the local level governmental coordination mechanisms to harmonize activities, resolve conflicts, Give municipalities discretion to set a. and reach agreements on specific issues. property-tax rates within a limited range Again, a national government entity and determined at the national level; representative associations of regional/local b. Establish a national framework for coop­ governments may play a facilitating role. eration between regional and municipal revenue agencies, or for developing Taxation administrative cooperation agreements for shifting tax administration responsibilities Narrowing the vertical fiscal gap will require (e.g., collection) between government levels, increasing the tax revenue generated by regional or between district and provincial tax and municipal governments. This can be achieved agencies; through the combination of broadening tax bases c. Streamline user charges and fees collected assigned to them, enabling them the ability to by local governments and simplify licensing define rates for their own taxes, and improvements and payment procedures; and in tax-collection efficiency. Regional governments completely lack tax bases while local governments d. Create an Office of the National Cadaster have small bases but without autonomy to define to enhance property-tax collection by rates. Proposed are the following policy options managing the registration and valuation of to increase subnational taxation. It is, however, properties at the national level. important to mention that tax and revenue sharing reform options for subnational governments need to be discussed based on a proper reassessment of 3  To expand their tax bases, an increasing number of countries the overall tax system and their compound effects and most recently Mexico (2014) and Chile (2015) have adopted excise taxes on tobacco, alcoholic, and sugar-containing non- on labor, investment, and consumption decisions alcoholic beverages and junk food with corrective effects on that affect growth and income inequality. This consumption and beneficial impacts on health. xviii    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Intergovernmental transfers and resources are withdrawn in events of lower than average commodity prices. The proposed reforms in this area would trans­ form the ordinary resource transfers into a more 6. Establishing a more stable, predictable, and transparent and predictable financing system equalizing system of transfers for regional based on revenue sharing and equalization governments criteria. The size of the revenue-sharing mechanism a. Reform ordinary resource transfers by: and an enhanced FONCOR equalization pool could be adjusted to enhance the equalizing properties of Defining a share of personal and i. the entire system. The new formula for FONCOR corporate income-tax (excluding the used in the second scenario could also be adjusted corporate income tax on extractive by changing its weights to increase the equalization industries that is already transferred via effect of the transfer. A fiscal simulation of the the canon) and VAT revenue equivalent impact of these reforms on the finances of regional to the current level of ordinary resource governments are presented in detail in Chapter 6. transfers; ii. Dividing that pool of funds into a formula- A number of reforms to FONCOMUN will enhance driven revenue-sharing component its equalizing properties and remove the (two-thirds) and an equalization- incentives to create new municipalities. These transfer component (one-third); reforms would also require the earmarking of resource-revenue transfers to capital investment iii. Basing the revenue-sharing distribution projects, enabling a broadened use. These reforms formula on two equally weighted include the inclusion of fiscal capacity estimates criteria—regional GDP and population— in the distribution criteria and the streamline of in order to incorporate both the point- the three-stage distribution. A simulation of the of-origin principle for resource transfers impact of these reforms on the finance of each and the equalization principle; and municipality is presented in detail in Chapter 6. iv. Developing a strategy for phasing in the new system that minimizes the shock Establishing a Stabilization Fund for Resource- to regional government budgets. Note Revenue Transfers and defining clear rules of the reform can be designed so as to accumulation and withdrawal will strengthen be neutral with respect to the central the current revenue stabilizing mechanism government level of revenues. based only on withdrawals and transfers. Compensating regional and local governments Enhance the equalization impact of b. for the fall in resource-revenue transfers by transfers to regional government by increasing other transfers has functioned as a type incorporating part of ordinary resource of revenue-stabilizing mechanism. In particular, transfers to the pool of FONCOR funds; as local governments were the most affected by Either preserve the current FONCOR c. the fall in resource-revenue transfers, the central formula, or alter it to better account for government enabled them to withdraw the unused regional differences in expenditure needs balances of resource-revenue transfer proceeds and revenue capacity, thereby enhancing accumulated in previous years. However, more its equalization impact; and permanent mechanisms to stabilize subnational revenues with transparent rules of accumulation d. No longer earmark FONCOR transfers for and use are needed. Chapter 6 simulates a scenario infrastructure investment, but instead in which canon resources accumulate in periods of make them unconditional transfers higher than historical average commodity prices, Executive Summary    xix designed to finance the decentralization temporary compensation mechanisms to of functions currently funded by ordinary minimize budgetary shocks; and resource transfers. f. Explore options to increase the pool of Enhancing the 7. equalizing impact of resources transferred under the reformed FONCOMUN FONCOMUN mechanism by, for example, increasing the national VAT surcharge or a. Alter the distribution formula to include funding it with excise taxes. revenue-capacity criteria—including the size of resource-revenue transfers and 8. Reducing the volatility of resource-revenue other transfers (except transfers from transfers FONCOMUN itself)—in order to enhance its a. Establish a stabilization fund for resource- equalization effect; revenue transfers based on well-defined b. Ease the earmarking rules for resource- rules for accumulation and withdrawal; revenue transfers, and enable provincial and b. Integrate the stabilization fund into the district governments to use them to finance macro-fiscal framework and the fiscal recurrent expenditures, helping to offset rules for subnational governments; the decrease in FONCOMUN transfers that may result from the proposed reforms; c. Create separate accounts for each recipient government; c. Eliminate minimum transfers, at least for newly created jurisdictions, as they incen­ Ensure that the management of the d. tivize the proliferation of municipalities; fund is fully transparent and backed by appropriate oversight and accountability Reform FONCOMUN’s three-stage dis­ d. mechanisms; and tribution rule by creating separate criteria for transfers to provincial and district Strengthen the system for pooling e. municipalities; resources among municipalities to finance interjurisdictional infrastructure projects. e. Pursue a gradual and properly sequenced reform process, which may include 1 Introduction Following a period of strong centralization in more accurately tailor public services to suit the the 1990s, Peru launched an ambitious fiscal specific needs of citizens in their jurisdictions, decentralization process in 2001–02. As opposed while also strengthening the accountability of to the centralism imposed in President Fujimori’s government officials. Attenuating the country’s tenure (1990–2000), the new democratic vast socioeconomic disparities and meeting the government elected in 2001 initiated a rapid pro­ rising demand for public services in its most cess of political and fiscal decentralization. Shortly underserved regions was also a major objective after the election, the authorities formalized the of the decentralization agenda. The drive to decentralization process as a permanent state strengthen democratic institutions to enable policy. The March 2002 constitutional reform4 decisions to be taken at the local level also meant established autonomous regional governments an accelerated wave of political decentralization. based on the existing administrative jurisdictions (departamentos), and elections for the newly While its initial actions were swiftly established governments were held in November implemented, Peru’s decentralization plan was of the same year. The decentralization program expected to be phased in over a number of years. envisaged the consolidation of the 26 regional The 2002 Decentralization Law5 and the 2004 governments into 12 macro-regions, which were Fiscal Decentralization Law6 defined the policy to serve as an intermediate level of government actions and fiscal resources necessary to complete between the national and local governments. the decentralization agenda and specified the implementation schedule for transferring The decentralization process had three main administrative and expenditure responsibilities to objectives: to improve the overall efficiency of local and regional governments. First, a system public spending, to reduce Peru’s large regional was established to identify functions that could disparities in the provision of public goods and be transferred to subnational governments services, and to strengthen local democratic without overwhelming their institutional capacity. decision processes and institutions. As in other Second, complementary legislation was developed countries, the rationale for fiscal decentralization to execute an orderly transfer of functions from relied on the idea that their relative proximity to the national government to regional and local their constituents enables local governments to governments. Third, institutional capacity- building plans for regional governments were formulated to ensure that regional authorities 4  Law 27680 (March 2002) defined the basis for the decentralization process, including the election of new regional would be adequately prepared to assume their new government officials. The fiscal decentralization process had responsibilities. Finally, a plan to gradually transfer already begun with the adoption of new legislation in July 2001 mandating that local governments and departments receive 80 and 20 percent, respectively, of public revenues generated 5  Law 27783 (June 2002). by extractive industries. 6  Legislative Decree 955 (February 2004). 1 2    Peru: Building a More Efficient and Equitable Fiscal Decentralization System fiscal resources to regional governments and then them heavily dependent on intergovernmental to the macro-regions was established. In the first fiscal transfers. Moreover, the distribution criteria stage of the fiscal transfer, Ordinary Resources for intergovernmental transfers creates incentives (Recursos Ordinarios) from the national government to further subdivide municipal jurisdictions, would be reallocated to finance the staff and increasing both the number of local government facilities being transferred from the line ministries units and the fixed costs of public administration. to the regional authorities. In the second stage, more permanent revenue-sharing mechanisms Increasing distortions in the intergovernmental would be created to finance the operations of the fiscal relations underscores the pressing need macro-regional governments. to refocus attention on the core objectives of the fiscal decentralization effort. While However, the failure to establish macro-regional expenditure responsibilities increasingly devolve governments derailed the decentralization upon subnational governments, revenue process. A 2005 referendum rejected the proposed collection remains highly centralized, and the merger of 16 regional governments into five macro- reliance of regional and local governments on regions. A second referendum on the consolidation intergovernmental transfers creates perverse of other macro-regions had been scheduled for incentives that negatively impact the quality of December 2009, but was postponed indefinitely. the public spending and the efficiency of service The capacity-building plans for regional delivery. The dependence of regional and local governments were designed, but have not yet been governments’ budgets on intergovernmental fiscal implemented. As the decentralization of service transfers diminishes incentives for expenditure delivery responsibilities was delayed, emphasis efficiency at the subnational level while further shifted to administrative deconcentration, with discouraging own-source revenue collection.7 line ministries assuming a greater role in sectoral policy making. Rising revenue from the extractive industries has driven the fiscal decentralization process, but As a result, an intermediate level of government, the increasing importance of resource revenues which was expected to balance economies of scale has also inadvertently shifted responsibilities for with local autonomy, has not yet been created, public investment to subnational governments and regional governments have attempted to fill and exacerbated their budgetary volatility. By the gap. Regional governments have assumed the law, resource revenues (called canon, sobrecanon functions originally assigned to the macro-regions, and royalties8 in Peru) are shared and transferred to becoming the de facto intermediate level between regional and especially to local governments, and the local and national authorities. However, regional they are earmarked for infrastructure spending. The governments have weak technical and institutional share of fiscal revenues generated by the extractive capacity and lack either own-source revenues industries and transferred to regional and local or transparent and stable revenue-sharing governments increased from 0.4 percent of GDP in mechanisms. Moreover, regional governments have limited autonomy to allocate the ordinary resource transfers they receive, to tailor public services to 7 Recent research on fiscal decentralization also suggests local needs, or to coordinate the provision of public that a heavy reliance on transfers is associated with excessive goods and services by municipalities. subnational indebtedness, as financial implications of spending decisions are not necessarily fully internalized. 8 The canon is a resource-revenue transfer financed by Meanwhile, a high and increasing degree of 50 percent of the corporate income tax paid by extractive fragmentation at the local level is undermining industries to the central government. In the case of oil the efficiency of decentralized service delivery. resources there is the canon and sobrecanon that correspond Many small municipalities have inadequate to 10 percent ad valorem of the oil production (for the canon) institutional and technical capacities, and their and 2.5 percent (for the sobrecanon). Mining royalties are an limited scope for local revenue collection leaves ad-valorem levy applied to the gross value of the extracted minerals. Introduction    3 2004 to 2 percent of GDP in 2012, before falling to by the heavy concentration of resource revenues in 1.4 percent of GDP following the recent decline in certain provinces. commodity prices. As a result, local governments have eclipsed the national government as the Large fiscal disparities between regions continue primary source of infrastructure investment, but to be reflected in the uneven distribution of their limited administrative capacity has negatively basic public goods and services. While inter­ affected the quality and efficiency of investment governmental transfers are designed to shift fiscal spending. In addition, the inherent volatility of resources from wealthier regions to poorer ones, revenues from the canon, sobrecanon and royalties regional inequalities in education, health care, and (referred to hereafter as canon) has made the basic services such as water and sanitation are budgets of regional and local governments highly highly correlated with the regional distribution of unpredictable, further undermining the efficiency of income (Figure 1). both capital investment and current expenditures at the local level. The following report analyzes the status of fiscal decentralization in Peru and presents a set The growing importance of canon revenues has of reform options designed to promote greater also jeopardized recent progress in improving efficiency and equity in the fiscal decentralization the interregional equity of public service framework. Chapter 2 provides a brief overview of delivery. Historically, economic activity has been recent progress on the decentralization agenda heavily concentrated in the capital, Lima, and and the main challenges faced by policy makers. the coastal region of Callao.9 Together, Lima and Chapter 3 assesses the institutional arrangements Callao are home to a third of Peru’s population that define the vertical and horizontal structure and produce more than 50 percent of its GDP. The of subnational governments. Chapter 4 describes concentration of economic activity in Lima, Callao, the allocation of spending responsibilities between and a few resource-rich provinces results in highly different levels of government and explores uneven local revenue bases. Fiscal transfers were options for promoting greater accountability and designed to address these imbalances, but canon efficiency in subnational spending. Chapter 5 resource revenues have exacerbated disparities in assesses the distribution of revenue authority fiscal capacity across subnational governments. A and tax-collection rates at various government large share of resource revenues goes to regional levels and simulates the effectiveness of different and local governments in areas where extractive policy options designed to bolster the own- industries are located, more than offsetting the source revenue capacity of regional and local equalizing effect of intergovernmental fiscal governments. Chapter 6 analyzes the system of transfers, which have increased at a much slower intergovernmental fiscal transfers and evaluates rate. Consequently, fiscal decentralization is both prospective reforms to the distribution criteria excessive at the local level, as fiscal transfers exceed intended to enhance their equalization effect. local administrative capacity, and inadequate to Chapter 7 summarizes the report’s conclusions and compensate for the regional imbalances caused policy recommendations.10 9  Lima and Callao are specially designated as a Metropolitan Municipality and a Constitutional Province, respectively. Despite their relatively small geographic size, the special status of Lima and Callao puts them on equal footing with 10 Given Peru’s very low level of subnational indebtedness, regional authorities rather than other municipalities. this report does not include a chapter on borrowing. 4    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 1: Regional disparities in per capita GDP and access to basic services 7.3 19.9 17.8 7.9 2.1 4.9 10.1 1.3 2.3 5.8 15.5 7.2 5 8.6 Share of the 1.1 population with at 4.1 2.7 least 2 out of 4 unmet basic needs 1.2 7.9 2.8 according to the 5.1 2014 Unmet Basic Needs Index. 1.7 GDP per capita 2014 (nuevos soles) 2014 1.7 5,935–8,737 16%–20% 1 8,738–11,551 8%–15% 11,552–19,770 6%–7% 19,771–48,616 1%–5% Sources: INEI and MEF. Note: The Unmet Basic Needs Index estimates the share of the population with access to adequate housing, sanitation, basic education, and a subsistence-level income. 2 An Overview of Fiscal Decentralization in Peru Peru’s public sector consists of the national decentralization diamond” presented in Figure 2 government, 26 regional governments,11 provides a simple analytical framework for and 1,845 municipalities, of which 195 are understanding the present asymmetries in provincial municipalities and the rest are the fiscal decentralization framework. The district municipalities. The fiscal decentralization diamond reflects the four dimensions of public framework consists of the laws, rules, and financial management: taxation, spending, institutions that allocate revenue and expenditure intergovernmental transfers and debt. In Peru, responsibilities among the three levels of spending is far more decentralized than taxation. government. It also includes the intergovernmental As a result, intergovernmental transfers are the transfer system and the fiscal arrangements dominant financing source for regional and local regulating the fiscal performance of regional and governments. Due to their reduced access to credit local governments and borrowing by subnational markets and the tight borrowing restrictions authorities. imposed on them, regional and local governments have limited recourse to borrowing to finance their Peru’s fiscal decentralization trends highlight expenditure responsibilities, leaving them heavily the profound asymmetry between revenue dependent on intergovernmental transfers.12 and expenditure decentralization. The “fiscal Figure 2:The fiscal decentralization diamond Transfers Transfers Spending Taxes Spending Taxes Debt Debt Peru Peru Brazil Mexico Source: World Bank staff. 12  In 2014, debt to GDP ratio of regional governments was 11  This includes the Constitutional Province of Callao and the 0.45 percent. Metropolitan Municipality of Lima. 5 6    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 3: Subnational spending as a share of total government spending, 2012 90 80 70 60 50 40 30 20 10 0 Canada Australia Peru Poland Colombia Romania Thailand Chile Malaysia Peru Colombia Chile Canada United States Denmark Switzerland Sweden Germany Australia Belgium Finland Peru Austria Norway Poland Netherlands Iceland Italy Estonia United Kingdom Czech Republic France Slovenia Slovak Republic Hungary Chile Israel Portugal New Zealand Luxembourg Ireland Greece Structural peers Regional OECD peers Peru Peers Average Sources: IMF GFS database and OECD. Figure 4: Public investment by local governments as a share of GDP, Peru and selected comparators, 2012 3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 Peru Thailand Romania Poland Malaysia Colombia Australia Canada Chile Peru Colombia Chile Peru Korea, Rep. Czech Republic Estonia Slovenia United States Poland New Zealand Finland Iceland Sweden Japan France Israel Slovak Republic Norway United Kingdom Australia Greece Netherlands Canada Belgium Chile Italy Austria Switzerland Denmark Hungary Spain Portugal Germany Ireland Structural peers Regional OECD peers Peru Peers Average Source: IMF GFS database, 2014. Expenditure decentralization progressed rapidly out by local and regional governments, respectively. in recent years. Over the past decade responsibility Indeed, local governments in Peru invest significantly for the provision of key services, such as education more than their peers (Figure 4). Consequently, the and health, has been transferred to regional efficiency of subnational governments is especially governments, while municipalities now implement critical to the quality of public service delivery, and the bulk of public infrastructure investment. especially to the provision of public infrastructure. Regional and local governments account for close to 40 percent of all primary spending (Figure 3) and By contrast, taxation remains highly centralized. about 70 percent of all investment spending, with The national government collects more than 47 and 23 percent of investment spending carried 90 percent of total tax revenue, and the low An Overview of Fiscal Decentralization in Peru    7 tax-collection rate of subnational governments level, then transfer resources to subnational has been a persistent feature of Peru’s fiscal governments to finance spending at the local decentralization process. Tax and other own- level. However, increasing own-source revenue source revenues represent less than 1 percent collection by subnational governments can also of GDP and finance less than 10 percent of promote expenditure efficiency by strengthening subnational expenditures. Regional governments local control and taxpayer accountability. In light rely on user fees and other nontax revenue streams of this tradeoff, the balance between subnational to supplement their low tax revenues. Local taxation and intergovernmental transfers will vary governments are empowered to collect certain from country to country depending on the relative property taxes, but these revenues are modest and efficiency of centralized tax collection and the collection efficiency is very low. effectiveness of decentralized service delivery. This asymmetry between concentrated In Peru, however, the vertical fiscal gap has revenues and decentralized expenditure— become excessively large and poses a number known as the vertical fiscal gap—is common of serious challenges. Subnational governments in other fiscal decentralization arrangements (SNGs) in Peru execute about 40 percent of total worldwide and not necessarily very problematic. spending and collect about 10 percent of tax In principle it may be nothing more than a benign revenues—the largest vertical fiscal gap among effect of efforts to reduce economic distortions OECD countries (Figure 5). Large vertical fiscal gaps associated with taxation or economies of scale in can reduce expenditure quality and undermine tax administration, as a central revenue authority fiscal discipline. Although the optimal vertical fiscal may be better positioned to collect personal and gap for any given country is difficult to determine corporate income taxes, as well as certain indirect precisely, recent research indicates that very taxes. Meanwhile, public spending tends to be large fiscal gaps, or very low own-source revenue more effective when it reflects local conditions collection by subnational governments, diminishes and priorities. As a result, it may be most efficient incentives for expenditure efficiency. Given Peru’s to collect the bulk of tax revenues at the national very large vertical fiscal gap, increasing the share Figure 5: The vertical fiscal gap, Peru and OECD countries, 2012 80 80 PER NLD (%) of SNGs on general government spending KOR CAN DNK NOR 60 60 CHE DNK LUX SWE MEX ESP USA EST KOR 40 PERU FIN DEU 40 CZE BEL NOR SVK NLD POL ITA AUT FIN GBR CZE ISL EST SWE HUN 20 FRA SVN 20 SVK FRA PRT ISR LUX IRL CHE DEU 0 0 USA 0 20 40 60 80 0 20 40 60 80 (%) of SNGs on general government tax revenues Fiscal Gap = [(%) SNG spending – (%) SNG revenue]/(%) SNG spending Sources: OECD and World Bank. 8    Peru: Building a More Efficient and Equitable Fiscal Decentralization System of tax revenues raised by subnational governments Transfers to local governments have increased could boost the overall efficiency of public spending. substantially in recent years. Local governments receive fiscal resources from three main sources. Peru’s very wide vertical fiscal gap is The first is the Municipal Compensation Fund compounded by a similarly large horizontal fiscal (Fondo de Compensación Municipal, FONCOMUN), gap, as both revenue and expenditure levels a non-earmarked revenue-sharing transfer vary enormously between regions. These large financed by a 2 percent surtax rate on the central vertical and horizontal fiscal gaps have resulted in government’s value-added tax (VAT). FONCOMUN a complex set of intergovernmental transfers. This resources are distributed to all municipalities system includes non-earmarked transfers financed according to equalization criteria that reflect their through revenue-sharing mechanisms, earmarked fiscal needs and ensure that they have adequate transfers financed through the national budget, revenue to execute their core functions. The second and special fiscal arrangements for using resource source is resource-revenue transfers. Municipalities revenues to fund infrastructure investment. receive 80 percent of all income-tax revenue from the extractive industries, and as with transfers The size and composition of intergovernmental to regional governments these resources are transfers have both changed substantially in distributed according to the derivation principle recent years. As a result of the increase in fiscal and earmarked for infrastructure investment. revenues from extractive industries and the The third source is a conditional transfer for implementation of the decentralization agenda in the modernization of municipalities. While local the second part of the 2000s, per capita transfers in governments also receive transfers of ordinary 2014 were around three times larger than they were resources, the amounts are much smaller than in 2004. Moreover, the importance of earmarked those transferred to regional governments.13 transfers—both budgetary transfers to finance spending in specific sectors and resource revenues Regional and local governments are transferred to fund infrastructure investment— overwhelmingly dependent on intergovernmental has increased relative to non-earmarked transfers, transfers, and this dependence has increased reducing the autonomy of subnational governments over time as the spending needs of local to allocate their resources according to regional governments have intensified while their own- and local preferences. source revenue capacity has remained limited. The recent evolution of regional government Transfers to regional governments are entirely revenues (Figure 6). reveals a continuous increase earmarked, restraining their autonomy. These in the share of ordinary resources transferred include ordinary resources, which were previously from the central government, which rose from transferred from the central government budget 65 percent of regional government revenues in to the regional directorates of its line ministries, 2007 to 80 percent in 2014 (Figure 7). Meanwhile, but which are now transferred directly to falling global commodity prices caused resource- regional government budgets and earmarked revenue transfers to decline from 15.4 percent of for specific decentralized functions. They also include the Regional Compensation Fund (Fondo de Compensación Regional, FONCOR), a 13 Other intergovernmental transfers to regional and local capital-budget transfer distributed according to governments include the Fund for the Promotion of Regional investment needs and fiscal capacity criteria, and and Local Public Investment (Fondo de Promoción a la Inversión Pública Regional y Local, FONIPREL), which provides resource-revenue transfers (canon and sobrecanon), matching grants for investments in infrastructure and social which represent 20 percent of the income tax on service delivery, and the Socioeconomic Development Fund extractive industries. Resource-revenue transfers of the Camisea Project (Fondo de Desarrollo Socioeconómico are distributed according to the derivation principle de Camisea, FOCAM), which is also distributed on a and earmarked for infrastructure investment derivation basis and finances environmental and basic social projects (see Table 1). infrastructure investments in regions and municipalities affected by the country’s major hydrocarbon sector project. An Overview of Fiscal Decentralization in Peru    9 Table 1: Intergovernmental transfers in Peru Size Formula for Fund (% of GDP) Distribution criteria Use distribution Transfers to regional governments Ordinary resources (recursos ordinarios) 3.19 Finance decentralized Earmarked for No functions service delivery FONCOR 0.12 Equalizing based on Earmarked to Yes fiscal needs/fiscal investment capacity Resource revenues (canon and 0.35 Derivation principle Earmarked to Yes sobrecanon) investment Transfers to local governments FONCOMUN 0.85 Equalizing based on General Yes fiscal needs Resource revenues (canon and 1.07 Derivation principle Earmarked to Yes sobrecanon) investment Municipal improvement incentive 0.00 Result-based transfer Enhance public No program for (tax collection) administration Ordinary resources 0.07 Negotiations Earmarked to No specific uses Transfers to regional and local governments FONIPREL 0.007 Competitive selection Earmarked to No of investment projects investment FOCAM 0.08 Derivation principle Earmarked Yes investment Sources: MEF and World Bank Staff. regional government revenues in 2007 to 9 percent government revenue rose from 14 percent in 2007 in 2014, and stagnant funding for FONCOR capital to 35 percent in 2009, though the recent decline transfers progressively diminished their share in commodity prices caused this share to fall to in regional government revenues. Nevertheless, 31 percent in 2014 (Figure 8). The inherent volatility canon resource-revenue transfers are an important of resource revenues and their increasing fiscal budgetary component in resource-producing importance has increased the unpredictability of regions such as Ancash, Cuzco, Loreto, and Piura, local government budgets. FONCOMUN transfers while in non–resource-producing regions transfers have represented about a quarter of local of ordinary resources represent more than government revenue since 2004 (Figure 9), but 85 percent of regional government revenues. own-source revenues fell from 34 percent of total revenues in 2007 to 26 percent in 2014. However, While the revenue structure of local governments these figures mask significant regional disparities. is more balanced, they are also highly dependent Municipalities in resource-producing regions are on transfers from the central government, and heavily dependent on resource-revenue transfers, this dependence intensified during the 2000s as while municipalities in non–resource-producing resource-revenue transfers increased. The share regions are similarly dependent on FONCOMUN of natural resource-revenue transfers in total local transfers. However, across all regions local 10    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 6: The evolution of regional government Figure 7: Revenue structure of regional governments, revenues (in constant 2007 PEN billions) 2007–14 16 100% 1.7% 0.8% 5.9% 3.6% 4.2% 14 90% 5.4% In constant 2007 PEN (billions) 12 80% 10 70% 8 60% 70.7% 64.8% 80.3% 6 50% 4 40% Capital revenues 2 30% Other current transfers 20% 5.2% 4.8% 0 Ordinary resources 2007 2008 2009 2010 2011 2012 2013 2014 11.4% 2.9% FONCOR 10% 15.4% Canon/sobrecanon Ordinary resources 9.1% Canon/sobrecanon FONCOR Own revenues 7.7% Own revenues 0% 3.3% 2.6% Capital transfers Other transfers 2007 2009 2014 Sources: Ministry of Economy and Finance, MEF; World Bank staff calculations. Sources: Ministry of Economy and Finance, MEF; World Bank staff calculations. own-source tax revenues typically represent less Intergovernmental transfers have been than 5 percent of total revenues. Only municipalities partially successful in reducing horizontal in the greater Lima metropolitan area generate fiscal imbalances. Equalization transfers are substantially larger shares of own-source revenues. used to reduce differences in own-source revenue capacity between different regional and local The evolution of canon resource revenue has governments. Tax revenue finances both ordinary become the key determinant of both the vertical resource transfers and FONCOMUN transfers. and horizontal distribution of fiscal resources, These revenues are largely collected in the most as well as the revenue and spending patterns of economically developed regions of the country subnational governments. The rapid increase in and then distributed to less-developed regions resource-revenue transfers spurred an unintended according to their demographic and socioeconomic and likely temporary decentralization of fiscal characteristics. revenues. This process did not result from the reassignment of tax bases to lower levels of Nonetheless, the increasing importance of government, nor did it reflect increased own-source resource-revenue transfers has weakened the revenue capacity among subnational governments, equalization effect of the overall intergovernmen­ and it was not part of a strategic plan to finance tal transfer system. Resource-revenue transfers decentralized service provision. Moreover, the disproportionately benefit a small number of increasing reliance of subnational governments regional and local governments in resource-rich on resource-revenue transfers left their budgets areas (Figure 10). Moreover, the allocation criteria highly vulnerable to external shocks, and greater for ordinary resource transfers and FONCOMUN budgetary volatility may be undermining the transfers do not account for other revenue streams, quality of subnational spending. and these transfers are allocated to resource-rich An Overview of Fiscal Decentralization in Peru    11 Figure 8: The evolution of local government Figure 9: The revenue structure of local governments, revenues (in constant 2007 PEN billions) 2007–14 7 100% 6% 6% 9% 3% 90% 3% 6 12% 7% 2% 80% 6% 5 In constant 2007 PEN (billions) 24% 70% 25% 4 28% 60% 3 50% 35% 2 40% 14% 31% Capital revenues 1 30% Other current transfers Ordinary revenues 0 23% 20% 17% 13% FONCOMUN 04 05 06 07 08 09 10 11 12 13 14 Canon and 20 20 20 20 20 20 20 20 20 20 20 sobrecanon 10% Nontax revenues Canon/sobrecanon 11% 12% 13% Taxes and FONCOMUN contributions Ordinary resources 0% Other transfers 2004 2009 2014 Taxes and contributions Nontax revenues Sources: Ministry of Economy and Finance; World Bank staff calculations. Sources: Ministry of Economy and Finance; World Bank staff calculations. The distribution rules also disproportionately benefit district municipalities, which receive more areas without regard to their total income. FONCOR than 50 percent of resource-revenue transfers in is the only transfer that does account for resource each region where an extractive industry is located. revenues, but its relatively modest value reduces As a result, the wealthiest provincial municipality its equalization effect. Five of the country’s 26 spends 80 times more, per capita, than the poorest administrative regions receive almost 50 percent of one, while the wealthiest district municipality resource-revenue transfers, with a full 25 percent spends 250 times more than the poorest district going to the Cuzco region alone. municipality (Table 2). As municipalities receive the bulk of resource- Though subnational indebtedness has increased revenue transfers, fiscal inequality is even more significantly in recent years, the subnational pronounced at the municipal level. For example, debt stock remains modest and does not yet over 25 percent of canon mining–revenue transfers represent a source of macroeconomic risk. in 2011 went to the top 1 percent of municipalities.14 Faced with increasing spending obligations, lower resource-revenue transfers due to the 14  Erman, 2015. drop in international commodity prices, and a 12    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 10: Per capita resource-revenue transfers to regional and local governments, 2004–14 (annual averages in constant 2007 PEN) Average 2004–2014 (nuevos soles of 2007) 0–21 22–175 176–293 294–1,936 Sources: MEF and Bank staff calculations. Table 2: Per capita spending among provincial and district municipalities (in constant 2007 PEN)   2005 2008 2011 2014 Maximum 1,793 2,530 3,449 4,625 Province municipality Purus (Ucayali) Jorge Basadre– Jorge Basadre– Jorge Basadre– Locumba (Tacna) Locumba (Tacna) Locumba (Tacna) Minimum 20 34 40 62 Province municipality Ascope (La Libertad) Ascope (La Libertad) Ascope (La Libertad) Ascope (La Libertad) Variation coefficient 1.27 1.35 1.2 1.09 Number of provincial 188 195 195 195 municipalities Maximum 5,447 17,378 29,657 26,342 District municipality El Algarrobal Ilabaya (Tacna) Ilabaya (Tacna) Ilabaya (Tacna) (Moquegua) Minimum 43 87 87 102 District municipality San Pedro (Puno) Sauce (San Martin) San Juan de Comas (Lima Met.) Miraflores (Lima) Coefficient 1.05 1.31 1.31 1.18 Number of district 1,548 1,632 1,634 1,639 municipalities Source: Ministry of Economy and Finance. An Overview of Fiscal Decentralization in Peru    13 limited capacity to raise own-source revenues, Nacional de Administración Tributaria, SUNAT many subnational governments have turned to according to a “public works for taxes” financing borrowing. As a result, the total subnational debt scheme15). When those debts are added to the stock rose from 0.17 percent of GDP in 2007 to financial debt, the total subnational debt stock 0.42 percent in 2016. may amount to 4 percent of GDP, according to preliminary estimates from the MEF. While However, there are certain local governments’ subnational indebtedness does not yet appear practices of incurring debt—through arrears to to be a source of systemic risk overall, certain providers and to central government agencies regional and municipal governments have become that need to be addressed. Most regional highly indebted. In some cases, debt levels exceed government financial debt is held by Peru’s Ministry 100 percent of the annual revenues of regional and of Economy and Finance (Ministerio de Economía y municipal governments, and these governments Finanzas, MEF). Other debts are arrears to private may need to adopt fiscal adjustment programs to providers, to the national pension system, and to ensure long-term debt sustainability. the National Tax Administration (Superintendencia 15 This new borrowing modality is increasingly common among local governments. Under a “public works for taxes” arrangement, private providers construct local public infrastructure as a way of reducing their tax liability to the central government. That tax liability is then transferred to the local government, with SUNAT as the creditor. 3 Strengthening the Institutional Arrangements for Decentralization Peru’s institutional framework for fiscal initially viewed as transitory entities that lacked decentralization remains unfinished. Following the capacity to fully assume certain expenditure the establishment of regional governments and responsibilities, and central government agencies the gradual devolution of administrative functions were slow and sometimes reluctant to transfer supported by increased intergovernmental their functions to regional governments. transfers, the process of consolidating regional governments into macro-regions stalled. While As a result, an administrative de-concentration the fiscal decentralization law included a generous model was adopted in place of the planned revenue-sharing mechanism, earmarking decentralization model. Rather than directly 50 percent of national income-tax and VAT revenues increasing the authority of subnational for macro-regional governments, the newly elected governments, sector ministries were given direct regional authorities feared that the establishment control over all budgetary decisions. In recent years of macro-regional governments would entail a the regional directorates of line ministries have significant sacrifice in regional political authority. been incorporated into regional governments, and Meanwhile, the central government was resistant their operations have been financed by ordinary to the idea of transferring half of its two most resource transfers from the central government important revenue streams (income tax and VAT) budget.16 The result is a hybrid deconcentration/ to the macro-regional governments. Finally, local decentralization model, in which sector ministries authorities may not have viewed the creation and other central government agencies retain of a new intermediate level of government to be authority over key policy areas that were in beneficial, as it could reduce local control over principle decentralized, including education, public service delivery and would require municipalities health, and water management.17 In addition, to share a portion of the resources transferred to a majority of technically decentralized fiscal them by the national government. resources go to finance the public payroll, which regional governments have a limited ability to The failure to establish macro-regional influence. Consequently, subnational governments governments halted the fiscal decentralization rely on intergovernmental transfers, yet they have process on several fronts. First, it left the regional authorities as the sole intermediate level of government, a role that they were not 16 Law 28926 of 2006 integrated the sectoral regional directorates into the regional governments as regular line designed to assume alone. Second, it delayed the bodies. decentralization of fiscal revenues, as the income 17  The World Bank report “The Decentralization Process and its tax and VAT revenue-sharing mechanism was Links with Public Expenditure Efficiency” (2010) describes the conditional on the formation of macro-regions. ambiguous role of the regional directorates on the education Third, it delayed the process of administrative sector. The World Bank report “Hacia un sistema integrado de decentralization, as regional governments were ciudades” documents the coordination problems between levels of governments in sanitation and water management. 15 16    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 11: Average municipal population size, Peru and selected comparators 60,000 50,000 40,000 30,000 20,000 10,000 0 le o r a il ia ru e e y n ce do az l ic bi ag p ai Ita hi liv Pe an ro ex om Sp ua Br C er Bo Eu Fr M Eq av ol C D EC O Sources: World Bank, OECD. Table 3: Municipalities by population, 2012 District Provinces Number Cumulative (%) Number Cumulative (%) 0–1,000 195 11.9 0 0.0 1,000–5,000 739 56.8 33 16.9 5,000–10,000 337 77.4 23 28.7 10,000–50,000 298 95.5 92 75.9 >50,000 74 100.0 47 100.0 Total 1,643 Source: Cheasty and Pichihua (2015) based on INEI. limited autonomy over expenditure decisions, fewer than 10,000,18 while 4 percent have more which undermines accountability and expenditure than 50,000 (Table 3). Peru’s average municipal efficiency at the regional level. population is comparable to that of certain high- income and highly fragmented countries in Europe, The fragmentation of municipal governments such as France, Italy, and Spain.19 However, while further weakens the efficiency gains of municipalities in wealthier countries often have decentralized service delivery. The average relatively strong local revenue bases and efficient municipalities in Brazil, Bolivia, Colombia, Ecuador, local governments, the degree of municipal Mexico, and Chile have populations between two and fragmentation in Peru reflects inadequate three times the size of Peru’s average municipality (Figure 11). This is due to the large number of very small municipalities in Peru. Fifty-seven percent 18  A minimum population of 10,000 is generally regarded as of Peru’s district municipalities have fewer than the threshold for leveraging economies of scale in local public administration. 5,000 inhabitants, and around 80 percent have 19  Martinez-Vazquez, 2013. Strengthening the Institutional Arrangements for Decentralization    17 local fiscal resources, high fixed administrative complementarity investments, or encourage costs, and an inability to leverage economies of projects with positive spillover effects. As a scale in local service delivery. Moreover, small result, inadequate leadership at the regional level municipalities often have difficulty executing has further diminished the returns to municipal their budgets efficiently.20 In addition, the lack spending. of formal demarcation of boundaries among many municipalities remains a potential source The situation is exacerbated by the atypical of conflict, especially among local governments feature of Peru’s decentralization system of located in areas with natural resource deposits.21 having two categories of local government (provincial municipalities and district Efforts to consolidate municipalities have not municipalities) with often overlapping respon­ succeeded in reducing the fragmentation of sibilities. A conspicuous case of the need for local governments. The 2007 Local Government greater coordination is the case of “Metropolitan Association Law enables municipalities to form Lima,” which refers to the geographic area that voluntary associations in order to jointly provide includes the 43 districts of Lima Province and certain services and execute certain infrastructure the six districts of the Constitutional Province of investments, and it regulates service-delivery Callao.22 This all involves a complex governance agreements between municipal and regional system. The provincial municipality of Lima (MML) governments. The law established financial itself has a special regime that combines district, incentives for municipal associations by enabling provincial, and regional competences.23 The Lima them to access additional FONCOMUN resources. district (Cercado de Lima) is run directly by the However, the law has not led to widespread MML. The Lima Metropolitan Council exercises municipal consolidation. To date, 198 municipal powers and functions equivalent to the Regional associations have been formed, but they comprise Council, and the mayor of the MML exercises the just 0.05 percent of local spending. For political powers and functions equivalent to the Regional reasons many municipal governments prefer to President. As part of its provincial competences, provide services and implement infrastructure the MML has responsibilities over the 43 districts projects individually rather than share credit with that are part of the province of Lima. Beyond the other municipalities. The central government could issue of greater coordination, which could involve a attempt to overcome this barrier to consolidation “metropolitan area authority,” Lima also raises the by focusing on the revenue side, and indeed the question of the need for a special fiscal regime— MEF appears to have successfully incentivized regarding functional responsibilities and also joint property tax (impuesto predial) collection at revenue sources—a need that can be extended to the local level. other large urban centers in the country. Due to Peru’s weak intergovernmental coor­ More than a decade after the launch of the fiscal dination mechanisms, regional governments decentralization process, the establishment have been unable to foster greater cooperation of macro-regions appears unlikely. Meanwhile, among municipalities. Regional governments lack in spite of an unclear division of responsibilities, the capacity to proactively support collaborative efforts between municipalities, promote 22 Metropolitan Lima has a significant presence in the population and the economy of Peru. With about 0.2 percent 20  Loayza et al. (2011) found a strong correlation between the of the land area of the country, it represents over 40 percent of size of the population and the rate of budget execution, both Peru’s GDP and about one-third of its total population. for current and capital spending. 23  In the region (department) of Lima, the regional government 21  The National Directorate of Territorial Demarcation does not have authority over the province of Lima. The election indicated that about 80 percent of district municipalities of regional authorities in the region of Lima does not include and 92 percent of provincial municipalities lack permanently the province of Lima, but rather only the other nine remaining defined boundaries (World Bank, 2010). It appears that to date provinces of the Lima region. The regional seat or capital city no progress has been made in this dimension. of the Lima region is Huacho. 18    Peru: Building a More Efficient and Equitable Fiscal Decentralization System an incomplete devolution of tax authority despite the adoption of some restrictions, has and increasingly convoluted revenue-sharing continued in recent years. While recent legislation mechanisms, the regional decentralization process has sought to prohibit the establishment of new has significantly advanced. In this context, municipalities, regulatory loopholes have allowed future decentralization efforts should focus the process to continue. Completing the legal on clarifying the expenditure responsibilities of demarcation of boundaries between municipalities regional governments,24 ensuring that they have would end territorial disputes, thereby eliminating adequate resources25 and promoting coordinated the most important loophole that allows for action between them. While exceptions exist, the creation of new municipalities. Tightening most countries with a geographic and population regulatory requirements would also help discourage size similar to that of Peru have three levels of the establishment of new municipalities. Finally, government: local, regional, and national. This reducing the minimum FONCOMUN transfer level arrangement is typically the most efficient because would eliminate an important incentive for the it allows each public sector function to be executed creation of small municipalities. at its optimal level of centralization.26 Efforts to encourage voluntary consolidation The authorities should prioritize the resumption at the municipal level have been largely of institutional capacity-building efforts at the unsuccessful, both in Peru and elsewhere, as regional level. This should be accompanied by a local officials tend to strongly resist measures thorough review of the legislative framework for that might render their positions redundant. regional governments with a view to clarifying As a result, an increasing number of countries in their expenditure responsibilities and identifying Europe have embraced mandatory consolidation appropriate permanent revenue sources. Regional strategies, though these efforts also tend to face tax authority should be properly defined, and bureaucratic resistance. One promising alternative rules for revenue sharing should be specified. The is to foster greater coordination and cost-sharing first step should be to reevaluate the institutional between municipalities without reducing the arrangements that were planned for the macro- number of municipal governments. While the regions and reformulate them to reflect the status Local Government Association Law provides an of the regional authorities as the sole intermediate appropriate foundation for local-level coordination, level of government. greater fiscal incentives for cooperation will need to be provided in order to enhance its effectiveness. In order to address municipal fragmentation, In addition, the government should promote more the authorities should declare a moratorium effective local service provision through competitive on the creation of new municipalities, which, contracting at the municipal level. 24 New expenditure responsibilities at the regional level should be assigned according to the guidelines of the national government. Accomplishing this will require capacity-building efforts and the use of incentives; for example, the government could increase certain fiscal transfers on the condition that regional governments recruit well-trained public managers, as it currently does under the SERVIR program. 25 Recent proposals by the General Directorate of Public Revenues and SUNAT require further evaluation. 26  For a more detailed discussion of the role of intermediate levels of governments see Lago-Peñas and Martinez-Vazquez (2013). 4 Clarifying Expenditure Responsibilities and Reducing the Investment Bias of Local Spending Despite extensive legislation on the issue, spending. According to the decentralization law, expenditure responsibilities are not clearly the transfer of sectoral functions must be based on defined, and intergovernmental coordination an evaluation system designed to ensure that the mechanisms for public service provision are institutional capacities of subnational governments largely ineffective. The 2002 Constitutional are adequate to execute the transferred functions. Reform, the organic laws of regional and local Institutional capacity-building plans were prepared governments, and sector-level regulations assign to support subnational governments in expanding administrative functions across government their administrative capabilities. Once a subnational levels. This framework distinguishes between three government was deemed capable of assuming types of functions: (i) those undertaken exclusively their expanded role, annual plans would be drafted by the central government; (ii) those shared by detailing the specific functions, sub-functions, multiple levels of government; and (iii) those that programs, and actions to be undertaken, as well can be delegated by the national government to as the fiscal resources, human resources, and lower government levels. The excessive number of capital assets to be transferred, and a timetable shared functions included in the organic laws—and for completing the process. the lack of a clear and precise description of the functions and sub-functions (regulation, financing, However, the institutional capacity-building and implementation) exclusively assigned to each plans were not satisfactorily implemented, government level or shared between levels—creates and the process for certifying the capabilities overlapping responsibilities rather than clearly of subnational governments required more shared or concurrent competencies. In addition, time than initially planned, which derailed the lack of clarity regarding which functions can be the decentralization process. In 2006 the delegated, and how they should be financed, has government attempted to accelerate the transfer undermined accountability by making it difficult of functions by eliminating the requirements to determine which government level is ultimately to implement capacity-building plans, verify responsible for the delivery of specific services.27 the capabilities of subnational governments, or prepare detailed annual plans for the transfer The absence of clearly defined expenditure of resources. The accelerated decentralization responsibilities has also delayed the process further obscured the distribution of decentralization process and has negatively expenditure responsibilities. Some functions were impacted the overall efficiency of public transferred to subnational governments that lacked the institutional capacity to execute them, and confusion regarding the transfer of staff and 27 World Bank (2010) provides an extensive description of the ill-defined and frequently overlapping expenditure budgetary resources required the continuous responsibilities in the education sector, while the Peru Ministry involvement of the central government. The of Health presented a similar analysis for the health sector. disarray of the expenditure decentralization 19 20    Peru: Building a More Efficient and Equitable Fiscal Decentralization System process weakened governmental accountability directorates are financially decentralized yet and undermined the efficiency of public spending. administratively deconcentrated. The lack of formal mechanisms for coordination Expenditure arrangements at the regional level between levels of government has also weakened are still in transition. Expenditures financed the effectiveness of decentralization.28 The lack through the national government budget within the of coordination starts with some line ministries framework of shared functions could be considered which do not internalize in their decisions the decentralized spending, since these funds effectively assignment of functional responsibilities to are transferred to subnational government budgets subnational governments. This issue is even more and earmarked for specific purposes. However, pronounced for the new ministries created after as decentralized service delivery financed via the the launch of the decentralization process (e.g., shared-functions framework becomes permanent, Ministry for Development and Social Inclusion earmarked transfers to regional governments or Ministry of Tourism and Foreign Trade) (OECD, could become regular revenue flows and could be 2016). Due to the large number of functions that included in the budget of regional governments in are either shared or that can be delegated, different the same way that “ordinary resources” are. For government levels must coordinate closely. example, payments from the central government Yet despite its expansive legal framework, the to the regional government to finance the country’s decentralization process does not yet encompass main health-insurance program (Seguro Integral formal intergovernmental coordination or conflict- de Salud—SIS) could be completely decentralized, resolution mechanisms. In principle, the regional meaning that resources would only be budgeted directorates of the line ministries should facilitate and executed at the regional level, and would coordination between the national and regional appear as transfers—not executed expenditures— governments, but in practice this mechanism has at the central level. However, these funds are still proven ineffective. registered in the national government budget, and they continue to be recorded as deconcentrated Precisely establishing the role of deconcentrated spending rather than decentralized spending.29 This and decentralized units at the regional level has underscores the need to identify items currently also proven challenging. As described above, a included in the national budget that should be number of sectoral functions were deconcentrated formally decentralized. This process could represent through the establishment of regional directorates, a second phase of fiscal decentralization (Box 1). which were incorporated into regional governments yet remained under the control of the sector Well-defined expenditure responsibilities are ministries. The complexity of this arrangement is necessary to increase the autonomy and reflected in the classification of spending by these accountability of subnational governments and decentralized units. As the regional directorates are improve the efficiency of public service delivery. formally part of the regional governments, they are This will require further clarification of shared or financed through “ordinary resources” transferred concurrent functions. Shared assignments can from the central government to the budgets of be unbundled into their respective sub-functions the regional governments. As a result, the regional (regulation, financing, and service delivery) and responsibilities for each sub-function can be allocated to the appropriate government level. This 28 Note that as the (OECD, 2016) reports the government approved a law in 2007 to establish intergovernmental commissions (ICC), which were to include national and 29  Mexico’s health insurance program (Seguro Popular) has subnational government representatives. Recently, a similar structure: the federal government reimburses state the Decentralization Bureau has considered only three governments for the use of health care facilities and staff, but commissions to be active (health, labor, and education). A main these expenditures are registered in the state governments’ difficulty has been the selection of subnational government budgets with the revenue source classified as an earmarked representatives to participate in the commissions. transfer from the federal government. Clarifying Expenditure Responsibilities and Reducing the Investment Bias of Local Spending    21 Box 1: Partial decentralization in the health sector Decentralization of health services took place salaries are normally paid by regional governments, between 2005 and 2009 with regional governments inputs are provided by MINSA and incidentally may officially becoming owners of the public health be financed by SIS—however, the lack of coordination facilities located in their territory. However, financing among these entities may result in incomplete inputs flows and budgetary arrangements have not enabled that prevent the regular full delivery of vaccination regional governments’ autonomy in planning, services. allocating, and executing budgets toward their health goals. Indeed, regional governments have not been This organizational disarray undermines efforts to able to manage and coordinate the budgets used improve planning, budget allocation, and execution. In to finance the various components of their health fact, while budget allocation in the health sector has services provision. In general, goods are acquired been trying to move away from inertial allocations and assigned by the Ministry of Health (Ministerio de based in historical budgets toward mechanisms that Salud–MINSA), bought by Budget Units directly from provide incentives for more effective allocation of funds assigned by the central government treasury, funds, such as the Result-based-Budgeting or they are assigned by Seguro Integral de Salud (Presupuesto por Resultados—PpR) and the (SIS). Human resources and services are contracted Comprehensive Health Insurance Scheme, or the at the regional level with resources assigned by the SIS, these attempts have had only a limited impact Treasury or by SIS. Investments are carried out either on budget management efficiency. About half of by MINSA or by regional governments.* the health budget is still assigned using historical budgets, and the different financing mechanisms As a result, the different inputs used for the delivery (historical budgets, PpR, SIS) “compete” with each of a particular health service are the responsibility other and provide contradictory incentives preventing of different entities and levels of government, and the efficient assignment and execution of funds. there is no guarantee that all inputs are available to provide those services in a timely and adequate Source: World Bank Staff. *According to Carlos Ricse, former Vice Ministry of Benefits and quality manner. For example, for vaccination services, Insurance of the Ministry of Health (Minsa). process may require several rounds of unbundling, and even to certain local governments with as some functions are more complex than others. sufficient administrative capacity. In addition, the A clear definition of responsibilities will also possibility of reverse delegation from lower to higher require the establishment of permanent formal government levels has not been fully explored. coordination mechanisms to harmonize activities, Given the weak administrative capacity of many avoid conflicts, and reach agreements on specific local governments, some district municipalities issues. may wish to delegate more sophisticated functions with clear economies of scale—such as Successful decentralization will also require the administration of the property tax (predial)— eliminating the overlapping responsibilities upward to provincial municipalities and regional between regional and local governments and government. streamlining the framework for delegated responsibilities. Reviewing the organic laws for Fully incorporating the regional directorates regional and local governments would enable into the institutional structure of regional policy makers to identify sectors with duplicative governments and providing them with adequate or overlapping responsibilities. The delegated- authority and resources to perform their functions framework remains incomplete and functions will be crucial to the decentralization underutilized. The central government could process. This will require severing the administra­ delegate more functions to regional governments tive links between the regional directorates and 22    Peru: Building a More Efficient and Equitable Fiscal Decentralization System the line ministries. The central government could the effectiveness of all other reforms in the fiscal continue to pursue sectoral policy objectives by decentralization system. Lessons can be learned creating incentives for regional directorates, such from many other countries that have successfully as conditional matching grants and national confronted similar difficulties by, for example, regulations, but it would no longer involve itself introducing national programs to train and certify in the day-to-day operations of the regional key local government functionaries in charge of directorates. budget planning and implementation (Box 2). Resuming the use of institutional capacity- Increased coordination and collaboration among building plans for subnational governments would different levels of government should become complement the clarification of expenditure priorities. The lack of coordination among the responsibilities. Most capacity-building plans different levels of government remains a key were suspended several years ago, and a lack of obstacle to the effectiveness and efficiency of the institutional capacity has left a large number of entire public sector. Several attempts have been municipalities unable to execute their decentralized made in the past to increase vertical and horizontal functions. Precisely defining the responsibilities of coordination, but so far they have largely failed these governments would enable them to develop (OECD, 2016). Institutional reforms leading to new, more accurate capacity-building plans. adequate coordination and dialogue among the Capacity building is a critical area that conditions different levels of government should be a reform Box 2: The Brazilian basic education fund (FUNDEB) The financing arrangements, budget allocation contribute 15 percent of their current revenues to a process, and service delivery responsibilities of the common pool in each state, which is then distributed three levels of government in the provision of education according to the number of students in the state services in Brazil represents a good example of a sound and municipalities. If the resources in a state’s pool assignment of responsibilities, predictable financing, are not sufficient to cover the minimum expenditure and accountability among central and subnational needs per student, the federal government makes governments. The system has consistently yielded up the difference. In this sense, the transfer from adequate service delivery. Brazil’s Fund for Primary the federal government can be considered a regional School Maintenance and Development and Teacher equalization transfer. The distribution formula for Training (Fundo de Manutenção e Desenvolvimento do FUNDEB also incentivizes good performance through Ensino Basico e Valorização do Magistério, FUNDEB) is a the Index of Development of Basic Education, which multi-governmental matching transfer that finances is included as an indicator of the quality of education, basic education services provided by state and combining information on students’ performance municipal governments. Municipalities administer from national assessments. primary education (levels 1–8), while states are responsible for secondary education (levels 9–12). FUNDEB has promoted regional equalization in Both states and municipalities are responsible for the per capita expenditures and has been successful management of human resources, including teachers’ in improving the coverage and quality of education payroll and school construction and operation, while services. However, because the minimum level of the federal government provides regulatory oversight expenditures per student determined by the federal for basic education. government is low, regional disparities in expenditures per capita persist, with the more developed south All three levels of government contribute to FUNDEB. and southeastern regions (which have greater tax The federal government defines regional minimum revenue) seeing higher education expenditures per levels of expenditures per student, which can vary student. according to region, grade, and location (urban/ rural). State governments and municipalities Source: World Bank staff. Clarifying Expenditure Responsibilities and Reducing the Investment Bias of Local Spending    23 priority. At the central-regional level, there is a need While the process of administrative decen­ to clearly delineate the roles of deconcentrated tralization has been slow and uneven, the units of the central government and to make them system has made considerable advances in the a separate permanent establishment from the decentralization of expenditures. Over the past “regional directorates.” Beyond that, there is a need decade Peru has achieved an intermediate-to-high to create permanent coordinating committees by level of spending decentralization by international sectoral area between the central line ministries standards (Figure 3). The share of total primary and agencies and the regional governments, spending executed by subnational governments which should meet on a periodic basis to discuss rose from 30 percent in 2004 to around 40 percent issues of coordination and potential conflict in 2014.30 Municipal governments drove this trend, between the two levels of government. Bilateral as the share of municipal government spending negotiations between the central authorities and rose from 12 percent in 2004 to 22 percent in regional governments should be avoided unless 2014. Meanwhile, the share of regional government they are issue specific, because they tend to lack spending remained broadly unchanged at around transparency and can give rise to perceptions of 20 percent, reflecting the modest progress favoritism. They are also more expensive for the in devolving additional responsibilities to the central authorities to hold. Similar permanent intermediate level of government. coordinating committees should be established at the regional level for the dialogue and coordination Rising natural resource revenues drove the rapid between regional and local authorities in the main increase in local government spending, and strict areas where there are concurrent responsibilities. rules for using resource revenues have created a powerful bias in favor of capital investment. There is a need for introducing mechanisms The bulk of resource revenues is transferred addressing the resolution of conflicts in to local governments, and these resources are expenditure assignments among different earmarked for infrastructure investments. As a levels of government. Even when mechanisms for result, the share of investment spending executed coordination and collaboration are in place, conflict by subnational governments rose from 44 percent is likely to arise because of different interpretations in 2004 to 68 percent in 2014, surpassing the of the legal system regarding expenditure OECD average of 64 percent, and investment now assignments. Conflicts among government levels accounts for about 60 percent of local government should be first addressed by mixed administrative budgets. On average, resource-revenue transfers sectoral committees with representation of the financed about two-thirds of local investment different levels of government. At a second stage, spending between 2004 and 2014. unsolved issues could be addressed formally (as is now done informally) in the meetings of the The recent decline in resource-revenue transfers national president and regional governors. If the has not reversed the fiscal decentralization conflict is among municipalities in the same region, process or significantly reduced investment this second stage could consist of the mediation of spending by local and regional governments. the regional authorities (for example, the regional Between 2012 and 2015 the central government president). If the conflict is between regional and compensated for the decrease in natural resource- municipal governments, the mediation could be revenue and FOCAM transfers by boosting ordinary entrusted to the prime minister. At a third stage, resource transfers to regional governments and this issue could be considered by administrative enabling local government to use their accumulated courts or even by the Administrative Bench (Sala resource-revenue balances. This enabled local de lo Contencioso-Administrativo) of the Supreme government investment to rise by 15 percent Court. The Constitutional Court should work as between 2010–2011 and 2012–2014. the last instance for conflict resolution after other avenues have been exhausted. 30  Cheasty and Pichihua, 2015. 24    Peru: Building a More Efficient and Equitable Fiscal Decentralization System The rationale for allocating natural resource The poor quality of many local investment revenues to capital investment is widely accepted projects and the limited public investment in the international literature. Investing resource management capacity of local governments revenues creates a sustainable long-term income undermine the effectiveness of resource-revenue stream that compensates future generations transfers. Inadequate strategic planning prevents for the loss of nonrenewable resources. As such, the efficient prioritization of public investment resource revenues can be used to finance physical projects, while low technical capacity for project assets, human capital formation, or environmental appraisal results in the implementation of projects protection, or they can be accumulated as cash with low rates of return. Capacity limitations reserves for future use. In Peru, the law mandates at the local level can delay implementation and that resource revenues be used exclusively for increase costs, and weak budget management infrastructure investment and the maintenance of may fail to ensure that assets will be maintained capital assets. over time and that resources will be set aside to finance future recurrent costs. Indeed, projections The distribution of resource revenues based for future operation and maintenance spending are on point of origin, their exclusive use for normally not included in local government budgets, infrastructure investment, and weaknesses in resulting in dilapidated, idle, or underutilized capital. public investment management at the local level Finally, exclusively earmarking resource revenues all limit the contribution of resource revenues to for infrastructure investment, rather than for economic growth and welfare objectives. Under recurrent spending on education and public health, the current legal framework, a small group of local has slowed human-capital formation and created governments in resource-producing regions receive a barrier to sustainable economic growth. large amounts of resources regardless of their spending needs. The influx of resource revenues has Shifting the composition of subnational overwhelmed the capacity of local governments to spending, especially local government spending, properly design and execute investment projects, in favor of human-capital formation could and as a result many local investment projects have enhance its effectiveness in advancing key social relatively low rates of return. These resources would and economic development objectives. This will likely be spent more effectively at the regional level require relaxing the earmarking rules for resource- or by local governments in other areas. Rather than revenue transfers, along with other changes in how identifying spending needs and allocating resource subnational governments are financed. For example, revenues accordingly, revenues are allocated first the establishment of a dedicated stabilization and then spending needs are determined. fund for regional and local governments could help smooth expenditures over time, allowing them to The fragmentation of investment decisions increase their recurrent expenditures. Meanwhile, limits opportunities for coordination and enabling regional and municipal governments to complementarity. The number of investment allocate a share of resource-revenue transfers to projects at the local level increased from 2,100 the education and health sectors would greatly in 2004 to more than 15,000 in 2014, while the increase their budgetary flexibility. These measures average cost per project is less than 1 million would be complementary, as education and health Peruvian Nuevos Soles (PEN).31 The large number of services typically entail a larger share of recurrent small projects suggests that investment decisions expenditures than capital investment projects. are uncoordinated and that few projects produce However, strict oversight will be necessary to ensure spillover effects. The distribution rules for resource- that budgeted expenditures are appropriately revenue transfers do not allow for the implementa­ registered and accounted for, as relaxing spending tion of large strategic investment projects that rules could create incentives to misuse resource would benefit more than one municipality. revenues. The government could also recentralize some 31  Erman, 2015. investment decisions in order to ensure that Clarifying Expenditure Responsibilities and Reducing the Investment Bias of Local Spending    25 public investment projects are aligned with Reforming Peru’s National System of Public regional and national objectives and that Investment (Sistema Nacional de Inversión they fully leverage complementarities and Pública, SNIP) and adapting its procedures for maximize positive spillover effects. This partial use by regional and local governments could recentralization could reduce the fragmentation further enhance investment efficiency. SNIP’s of municipal investment decisions and increase strict procedural rules include the preparation the overall efficiency of investment spending, as of pre-feasibility studies for investment projects regional and national governments are better financed by resource revenues. Though it was suited to designing and implementing certain originally highly centralized, SNIP’s authority has types of investment projects. Policy makers should gradually devolved upon investment units in local also assess the possibility of using accumulated governments. Yet even as its regulations were subnational resources to finance projects in other eased, SNIP came to be viewed as a bottleneck parts of the country through grants from the to public investment at the local level. Meanwhile, national government. with less stringent standards in place the quality of local public investment deteriorated. Rather Capacity constraints are also present at the than further relax its selection criteria and regional level. Capacity limitations can be as acute implementation guidelines, SNIP could develop for regional governments as for local governments. different procedures and methodologies for project Given the limited technical capabilities of many design and implementation at the regional and local regional governments, capacity building in levels. Moreover, using canon resource revenues to strategic planning and project appraisal could improve project appraisal, implementation, and ensure sound project screening and the efficient maintenance capacity could enhance the impact implementation of regional projects of greater of local public investment on economic growth and scope and complexity. welfare. 5 Improving Subnational Taxation The decentralization process launched in the have a cadaster all contribute to low tax-collection early 2000s has not significantly enhanced the efficiency. own-source revenue capacity of subnational governments. Fiscal decentralization laws did Tax-collection rates vary significantly by region. not alter the allocation of tax authority between In Lima per capita local tax revenue is 100 times the different levels of government. The tax authority national average, while Lima’s income per capita of municipalities has not changed since 1993, is only around 40 percent above average. About and regional governments have no tax authority, 95 percent of property-tax revenue is collected by as they were established to be transitory entities about 10 percent of all large municipalities, though that would be subsumed into the macro-regions. together these municipalities represent about two- Regional governments collect all of their revenue thirds of Peru’s total population. Overall, property- from user fees and other small revenue sources. tax revenue in 2014 represented just 0.215 percent The tax structure remains highly centralized, and of Peru’s GDP, well below the Latin American around 95 percent of total tax revenue is collected average of 0.42 percent, the global developing- at the national level. country average of 0.6 percent and the OECD average of 2.1 percent. However, total property-tax Municipal governments’ limited tax authority revenue in Peru has increased in recent years, likely is compounded by low collection efficiency. due to the MEF’s Municipal Improvement Incentive Income and consumption are taxed by the national Program (Programa de Incentivos a la Mejora de la government through the income tax, the VAT, Gestión y Modernización Municipal, PIMGMM). and excise taxes. District municipalities collect property taxes and taxes on the transfer of real Subnational taxes represent around 0.45 percent estate.32 Provincial municipalities collect taxes of Peru’s GDP, a lower share than in comparable on motor vehicles and on public entertainment, countries in Latin America. Local property-tax lotteries, and other forms of gambling. All tax rates revenue represents about 0.8 percent of GDP in and tax authorities are defined by the national Chile and about 0.7 percent in Colombia (Figure 12). government. Generous exemptions, weak tax The ratio of total local revenues to GDP in Peru is administration in rural areas, the lack of a property also much lower than in Argentina (5 percent of registry (or “cadaster”) in many municipalities, and GDP) and Brazil (11 percent), though comparable to outdated real estate values in municipalities that Mexico and Guatemala. Nevertheless, local tax collection in Peru has 32  Provincial municipalities also receive property taxes, but improved in recent years. MEF data for 2014– only for properties located within the capital district of the 15 indicate that property-tax revenue rose to province. The capital district of the province does not have its own district government, and the provincial government 0.24 percent of GDP. This likely reflects a variety governs both the province and the capital district. of factors. In the first place there is the role of 27 28    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 12: Property-tax-to-GDP ratio, Peru and selected comparators 1.2 1.0 0.8 0.6 0.4 0.2 0 a il le a a o y ru y C e az ua ua in bi al ic ag LA hi Pe ex nt om m Br C g g er ge te ra ru M av ol ua Pa U Ar C D G EC O 2008 2010 2013 Source: OECD. Table 4: Subnational government revenues as a share of GDP, 2004–14 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Local government taxes 0.28 0.27 0.25 0.34 0.36 0.37 0.39 0.41 0.43 0.46 0.44 Local government user 0.59 0.57 0.48 0.50 0.52 0.55 0.49 0.49 0.51 0.46 0.43 fees and other revenues Total local government 0.87 0.84 0.73 0.84 0.88 0.92 0.88 0.90 0.94 0.92 0.87 revenues Regional government user 0.11 0.10 0.12 0.14 0.14 0.28 0.16 0.20 0.17 0.13 0.11 fees and other revenues Total subnational own- 0.98 0.94 0.85 0.98 1.02 1.20 1.04 1.10 1.11 1.05 0.98 source revenues Source: MEF. the recent real estate property boom, which has Boosting subnational own-source revenue also been reflected in the tax revenue collections capacity will require both expanding subnational from the property transfer tax (alcabala). The tax authority and improving tax-collection technical and institutional support from the efficiency. As it now appears that the macro-regional national government—the MEF’s PIMGMM—may governments will not be formed, tax bases should also have played a role.33 However, improvements be assigned to the existing regional governments. in tax collection were offset by reductions in other However, the decentralization of tax authority does revenue streams, and total subnational own-source not imply the decentralization of tax administration revenues remained broadly constant (Table 4). or enforcement. Indeed, it may be both desirable and feasible to give regional governments some autonomy to set regional rates while keeping tax 33 Note that the increase in property-tax collections has collection under the purview of SUNAT. been mostly in urban areas while rural governments continue to raise almost nothing from the tax. But, it could also be the case that the MEF’s PIMGMM has been less effective in rural Allocating tax authority to the intermediate areas. level of government in a decentralized context Improving Subnational Taxation    29 requires a careful assessment of the trade-offs a presumptive-income business tax,35 separate and potential distortions inherent in multi- subnational corporate income taxes,36 or indirect level taxation. Regional governments require an taxes. New indirect taxes could include surcharges adequately large tax base to finance a substantial on existing national excise taxes, such as taxes on share of their expenditures. A thorough analysis will telephone or electricity services. Subnational taxes be required to determine the optimal tax authority on merchandise could include a regional tax on to assign to regional governments, but the most final sales or a centrally coordinated VAT; however, promising option could be to establish a surcharge subnational VATs entail serious design, efficiency, on the national personal income tax. and tax administration issues and are generally a less-preferred option. Whatever the mechanism, Ideally, a regional surcharge on the personal merchandise tax rates should be uniform in order to income tax would be levied at a flat rate and avoid creating incentives for interregional smuggling, collected concurrently with the national tax. This and the establishment of any new taxes always surcharge could be between 1 and 3 percent and entails a risk of distortive effects. Box 3, summarizes levied on the same basis as the national income the international experience with indirect taxation tax. The rate for each region could be decided by at the intermediate level of government. the regional government, but collection would be administered by SUNAT. Even at 3 percent, the There are several policy options for enhancing additional revenues generated by this surcharge local tax-revenue capacity. One possibility would would be relatively modest (Table 5). be to give municipalities some discretion over local property-tax rates within a limited range established To illustrate this proposed reform’s modest at the national level.37 While inconsistent tax rates effect on regional government revenues, create economic distortions and encourage the Table 6 presents the estimates from Table 5 as migration of tax bases, some modest latitude over percentages of the ordinary resource transfers the definition of tax rates (such as the definition of received by regional governments in 2015. This a rate interval) would not only enhance the revenue is the appropriate perspective from which to capacity of local governments, but would also consider the relative importance of the revenues tighten the link between local taxes and local public that could be generated by a flat-rate surcharge services, thereby strengthening the accountability on the personal income tax. While the amounts of municipal governments. Another potential source involved are small, they are significantly biased of municipal tax revenue is the “municipal license” downward by the estimation methodology. After (patente municipal) for business and commercial applying a correction factor, which could be as high activities, which is currently used in Chile. as fivefold or more, these amounts would remain small, but they would not be insignificant. Despite its effective status as a regional government, important spending responsibilities were not 35  For example, Colombia’s industry and commerce tax, which yields significant revenues, is designed to tax presumptive net transferred to the Metropolitan Municipality income. Gross income can be transformed into net income of Lima, and as a result it receives a marginal either by applying differentiated tax rates to gross income, amount of ordinary resources. Given the heavy which already incorporate the presumptive rate of profitability concentration of income-tax collection in Lima, a for each type of business, or by using standard lump-sum surcharge on the income tax would substantially deductions for the cost of doing business for each type of boost its own-source revenue capacity. business category and then applying a uniform tax rate. 36 For example, Germany and the United States, among other countries, use subnational corporate income taxes. Other policy options include assigning regional Determining the right apportionment of net income across governments the authority to levy a payroll tax,34 jurisdictions is the most serious difficulty with this type of tax. 37  As mentioned above, some rate discretion existed prior to the reforms implemented by the Fujimori administration in the 34  For example, Mexico levies a state payroll tax on employers. early 1990s. Having a minimum rate also means that the use Rates range from 1 to 3 percent and exemptions vary by state. of the tax is compulsory for all municipalities. 30    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Table 5: Potential regional government revenue derived from a 1–3 percent flat rate income-tax surcharge, 2015 (in PEN thousands) National personal income Alternate rates Department tax collection 1% 2% 3% Amazonas 25,598 256 512 768 Ancash 127,726 1,277 2,555 3,832 Apurimac 41,530 415 831 1,246 Arequipa 932,956 9,330 18,659 27,989 Ayacucho 53,685 537 1,074 1,611 Cajamarca 136,609 1,366 2,732 4,098 Callao 1,225,783 12,258 24,516 36,774 Cusco 404,670 4,047 8,093 12,140 Huancavelica 15,611 156 312 468 Huánuco 65,404 654 1,308 1,962 Ica 339,588 3,396 6,792 10,188 Junín 229,329 2,293 4,587 6,880 La Libertad 775,178 7,752 15,504 23,255 Lambayeque 217,457 2,175 4,349 6,524 Lima provinces 114,608 1,146 2,292 3,438 Loreto 195,346 1,953 3,907 5,860 Madre de Dios 35,619 356 712 1,069 Moquegua 38,339 383 767 1,150 Pasco 28,772 288 575 863 Piura 413,414 4,134 8,268 12,402 Puno 167,412 1,674 3,348 5,022 San Martín 115,583 1,156 2,312 3,467 Tacna 91,388 914 1,828 2,742 Tumbes 42,618 426 852 1,279 Ucayali 164,413 1,644 3,288 4,932 Total 5,998,636 59,986 119,973 179,959 Metropolitan Lima 28,746,804 287,468 574,936 862,404 Source: Authors’ calculations based on SUNAT data for 2015. Note: These data are for actual collections, not taxable income.38 38  This imparts a significant downward bias to the estimates, Beyond traditional capacity-building programs, which is larger the smaller the effective average rate on the government has several options for taxable income for the national tax in the region. A corrective enhancing subnational tax-collection efficiency. factor should be applied reflective the inverse of the effective These include establishing a national framework rate on taxable income in the region. It should also be noted for cooperation between regional and municipal that these data are based on the place of withholding and payment of the tax, which may not necessarily coincide with revenue agencies, or developing administrative place of residence of the taxpayer. For example, if the effective cooperation agreements for outsourcing tax average tax rate at the national level is 25 percent, collections administration from lower to higher government for Amazonas a flat rate of 3 percent should be PEN 3,072, levels, or between district and provincial tax but if the average national effective rate is 15 percent, then agencies. Administrative cooperation agreements collections for Amazonas would be PEN 5,120. Improving Subnational Taxation    31 Table 6: Share of ordinary resources represented by the potential tax revenue from the income-tax surcharge, 2015 (in PEN thousands) Ordinary resources Personal income tax potentially collected (in PEN thousands) as a share of ordinary resources Regional government 1% 2% 3% RG Amazonas 575,323 0.04% 0.09% 0.13% RG Ancash 923,039 0.14% 0.28% 0.42% RG Apurimac 696,738 0.06% 0.12% 0.18% RG Arequipa 1,124,108 0.83% 1.66% 2.49% RG Ayacucho 1,056,010 0.05% 0.10% 0.15% RG Cajamarca 1,188,583 0.11% 0.23% 0.34% RG Callao 610,352 2.01% 4.02% 6.02% RG Cusco 1,083,458 0.37% 0.75% 1.12% RG Huancavelica 701,745 0.02% 0.04% 0.07% RG Huánuco 820,606 0.08% 0.16% 0.24% RG Ica 689,200 0.49% 0.99% 1.48% RG Junín 1,155,639 0.20% 0.40% 0.60% RG La Libertad 1,159,319 0.67% 1.34% 2.01% RG Lambayeque 972,208 0.22% 0.45% 0.67% RG Lima 907,607 0.13% 0.25% 0.38% RG Loreto 1,031,540 0.19% 0.38% 0.57% RG Madre De Dios 271,837 0.13% 0.26% 0.39% RG Moquegua 301,949 0.13% 0.25% 0.38% RG Pasco 442,395 0.07% 0.13% 0.20% RG Piura 1,205,200 0.34% 0.69% 1.03% RG Puno 1,211,842 0.14% 0.28% 0.41% RG San Martin 1,080,027 0.11% 0.21% 0.32% RG Tacna 309,374 0.30% 0.59% 0.89% RG Tumbes 285,822 0.15% 0.30% 0.45% RG Ucayali 519,666 0.32% 0.63% 0.95% Total 20,323,587 0.30% 0.59% 0.89% Municipality of Metropolitan Lima 38,555 745.60% 1,491.21% 2,236.81% Source: Authors’ calculations based on SUNAT data. are currently being used in Mexico with very Uruguay, and Spain. While the national agency positive results. would maintain the cadaster, property taxes would still be collected locally. As some large urban One especially promising option for enhancing municipalities in Peru can maintain accurate and administrative cooperation would be the comprehensive cadasters and collect property creation of an Office of the National Cadaster. taxes efficiently, qualifying municipalities could This agency would enhance property-tax collection be allowed to maintain their own cadaster, while by managing the registration and valuation of the cadasters of other municipalities would be properties at the national level. Similar agencies managed by the national office. have been successfully established in Colombia, 32    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Box 3: International experience in multilevel indirect taxation There are a wide variety of arrangements for multilevel regardless of their budgetary impact is sometimes taxation, though taxes at the intermediate level are referred to as “the fiscal war between the states.” much more common in federal countries than in In addition to reducing state revenue, tax base and unitary states such as Peru (Table 7). Although some rate differentials encourage inefficient resource countries, notably the United States, impose state- allocation, as firms attempt to relocate based on level retail sales taxes, several studies suggest that more favorable tax treatment rather than economic the VAT is the least distortive consumption tax. efficiency. Firms can contribute to widening base and rate differentials by pressuring state authorities for However, the VAT suffers from a number of theoretical more favorable treatment. Ultimately, all firms may and practical limitations. These include: suffer as tax-related uncertainty deters investment and inadequate revenue mobilization causes state- (a) Intergovernmental coordination. The design and level fiscal crises. Coordination between states can operation of a multilevel tax system poses serious end this type of fiscal competition. challenges in terms of federal-state coordination. Due to a desire to assert jurisdiction, federal entities (c) Administrative capacity and costs. Complex tax may be unwilling to coordinate policy, which can lead systems can strain the ability of public agencies to to negative outcomes. administer the tax system and enforce compliance with tax laws. The challenge of administering a (b) Tax base and tax rate differentials. A lack of multilevel tax system can divert attention away from coordination can lead to large differentials between audit and oversight functions, as well as transparency the tax base and the tax rate. While such differentials and accountability mechanisms such as taxpayer are an inevitable result of state autonomy, education and data publication. uncoordinated policies can have a perverse impact, as states may compete to offer tax incentives that (d) Compliance capacity and costs. Complex tax serve their individual interests but undermine their systems also increase the cost of compliance. collective interests. In Brazil, competition between Compelling taxpayers to parse tax laws and states to offer ever more generous tax incentives regulations for different jurisdictions involving Table 7: Subnational indirect taxes Federal Subnational Cross-border Country VAT sales tax Type of subnational tax treatment Argentina Yes Yes Gross-receipts taxes Origin Australia Yes No All VAT revenue goes to states Destination Austria Yes No States receive a share of VAT revenue Destination Belgium Yes No None Destination Brazil Yes Yes VAT Origin (limited) Canada Yes Yes Some provinces use VATs, some use retail Destination sales taxes, and some do not use a provincial consumption tax Germany Yes No States receive a share of VAT revenue Destination India Yes Yes States are currently moving from producer taxes Origin (limited) to VAT Switzerland Yes No None Destination United States No Yes Most have retail state taxes Destination Improving Subnational Taxation    33 different bases, rates, filing requirements, and other subject to the rate of the importing jurisdiction, and features can greatly increase the administrative cost that exports to other domestic jurisdictions or abroad of paying taxes. An excessively complicated tax code be zero-rated. Thus, interstate trade should be can cause especially acute problems in a developing subject to the VAT rate of the importing jurisdiction. country environment, marked by low levels of general As few countries regulate interjurisdictional trade, education, high transportation, and communications unregistered importers have no incentive to self- costs, and limited taxpayer education programs. impose the VAT, and registered exporters have an When tax compliance represents a substantial incentive to inflate VAT credits to take advantage of burden, it creates an incentive for informality and the zero-rating of domestic exports. tax evasion, leading to both revenue losses and economic inefficiencies. Ease of compliance is (f) Tax evasion in cross-border trade: The most especially important for VATs, as registered traders common forms of fraud on cross-border trade are essentially act as tax collectors on behalf of the tax “carousel fraud,” false credit and refund claims, and administration. If they are not able to fulfil this role, import non-compliance with the VAT. The most serious the tax administration will not be able to fulfil its own. of these, particularly in the EU, is carousel fraud, also known as “missing trader intra-community fraud.” (e) Tax evasion in interjurisdictional trade. Many of Carousel fraud works by exploiting the zero-rating the VAT’s advantages depend on it being an efficient of exports under the deferred payment mechanism and effective indirect tax on consumption under the used to collect VAT on EU imports. destination principle. This requires that all imports be 6 Toward More Efficient, Equitable, and Stable Intergovernmental Transfers Due to Peru’s vertical and horizontal fiscal of resource-revenue transfers, undermine the imbalances, decentralized spending is over­ effectiveness of the transfer system in reducing whelmingly financed through intergovern­mental interregional disparities. transfers. The vertical fiscal gap results from the fact that the largest and most dynamic tax bases— The distribution of resource-revenue transfers, income and consumption—are assigned to the and their increasing fiscal importance, is a national government, while regional governments serious obstacle to achieving the objectives of have no tax authority, and municipalities are fiscal decentralization. The main rationale for restricted to collecting property taxes and a few distributing resource revenues to areas where other minor taxes. Horizontal fiscal imbalances extractive industries are located is to compensate are the effect of large interregional socioeconomic local communities for the negative externalities disparities and the spatial concentration of certain generated by the exploitation of natural resources.40 economic activities, especially natural resource However, the distribution criteria would appear to extraction. The concentration of economic activity be excessively generous to affected communities, in Lima and a few resource-rich areas greatly as their impact eliminates the equalization effect increases their own-source revenue capacity, and of FONCOR and FONCOMUN transfers. Resource- as a result, the ability of subnational governments revenue transfers are extremely uneven at the to self-finance public services varies widely. While municipal level (Figure 13), and the distribution rules the growth of the extractive industries created disproportionately benefit district municipalities, new economic centers outside of Lima, it has also which receive more than 50 percent of resource- intensified regional differences in per capita income revenue transfers in each region where an and subnational fiscal revenue, particularly as extractive industry is located. Because the location many resource-rich areas are sparsely populated. of extractive industries determines resource- revenue transfers, there is no correlation between The intergovernmental transfer system helps regional governments’ per capita spending and to close the vertical fiscal gap, but it does very local poverty rates (Figure 14). Indeed, the less- little to attenuate the horizontal gap. Most of the populated regions where most extractive industries existing transfers to regional and local governments are located tend to have relatively modest poverty are based on equalization criteria. However, the rates, yet their per capita expenditures are among absence of fiscal-capacity criteria in the formulas the highest in the country. for practically all other intergovernmental transfers,39 and the increasing importance 40  The externality argument is not officially recognized in the current legislation of Peru. The Constitution, Articles 193 and 196, state that natural resource royalties (canon) are regional 39 The exception is FONCOR, which explicitly accounts for rights, while Article 66 establishes that natural resources, fiscal capacity. However, FONCOR’s resource pool is small and renewable and nonrenewable, are patrimony of the nation, and remains constant over time. the state is sovereign in their utilization. 35 36    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 13: Distribution of resource-revenue Figure 14: Subnational per capita spending transfers among the top 1, 5, and 10 percent of in constant 2014 PEN (left) and regional poverty recipient municipalities, 2011 (in PEN thousands) rates in 2010 (right) 450 6,000 70 400 60 5,000 350 50 300 4,000 Nuevos soles 40 Percentage 250 x 10,000 3,000 200 30 2,000 150 20 100 1,000 10 50 0 0 0 Ta s eq a U ipa La M ali be n C P d am ra C ca ac co Li an ánu o o ov ica s* io Ar cn Li artí rta H Hu ch m ca c aj iu ce n y ar Ay us D pr vel u Sa ca 03 04 05 06 07 08 09 10 11 u in de 20 20 20 20 20 20 20 20 20 re ad u Top 1% recipients a M Top 5% recipients Subnational per capita spending Top 10 % recipients Poverty rate (2010) Canon minero (total GL) Sources: MEF and INEI. Source: Erman, 2015. *The poverty rate for metropolitan Lima, the Lima provinces and Callao is not disaggregated; the INEI reports a rate of 13.5 percent for the Lima and Callao region. In decentralized fiscal systems, equalization transfers are used to bridge the gap between expenditure needs and tax-collection capacity and given the size of the government’s assigned at different levels of government. In most cases tax bases. The optimal equalization transfer to a the majority of tax revenue is collected by the subnational government increases in line with its central government in the most developed regions expenditure needs and decreases in line with its of the country, and a portion of this revenue is then revenue capacity. redistributed according to a formula that reflects fiscal capacity and expenditure needs. In principle, Several aspects undermine the fiscal equaliza­ this should result in the systematic redistribution tion role of Peru’s intergovernmental transfer of fiscal resources from richer areas to poorer ones. system. While transfers are distributed according Expenditure needs are defined as the amount a to fiscal needs criteria, most do not consider subnational government would need to spend to revenue capacity. As a result, governments that provide a standard level of public services based are able to raise substantial own-source revenues on the size of the population, local socioeconomic or other revenue streams continue to receive large conditions, and the costs of providing those equalization transfers. For example, while ordinary public services. Revenue capacity is the ability resource transfers and FONCOMUN transfers of a government to raise own-source revenues use needs-based distribution criteria, they do not based on an average level of administrative effort, include revenue capacity. As a result, municipalities Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    37 with substantial own-source revenues and those received by a regional government was more than that receive resource-revenue transfers also receive 20 times higher than the largest FONCOMUN ordinary resource transfers and FONCOMUN transfer and more than five times higher than the transfers proportionate to their expenditure largest FONCOR transfer. As noted above, resource needs.41 As noted above, recent changes to the revenues are transferred based on their point FONCOR allocation criteria may enhance its of origin and are not correlated with recipients’ equalization impact, as capital expenditure needs expenditure needs or revenue capacity. are now adjusted to reflect revenue capacity, including resource-revenue transfers. However, Finally, specific aspects of the FONCOMUN in practice the equalization capacity of FONCOR distribution formula may require reform. In remains limited due to its very small size.42 particular, the minimum transfer level encourages fragmentation, and the allocation of resources in The distribution criteria for ordinary resource two stages is opaque and may produce undesirable transfers to regional governments primarily results. For example, two identical districts may reflect supply-side considerations rather than receive different transfer amounts just because demand for public services or local expenditure they are in regions with different overall fiscal needs. This mechanism was established in order needs. to finance the transfer of institutional and human resources necessary for the decentralized provision As a result of their weak equalizing criteria, of public services. However, the distribution criteria intergovernmental transfers do not reduce fiscal do not explicitly account for either expenditure disparities. Indeed, per capita spending at the needs or fiscal capacity criteria at the regional regional and municipal levels reinforces existing level. Although regional governments currently regional disparities in socioeconomic development have no tax authority, some regional governments (Figure 15 and Figure 16). receive resource-revenue transfers, which greatly increase their revenue capacity. Another disruptive aspect of canon resource- revenue transfers is the volatility that it adds to Equalization transfers are relative small, subnational and especially local public finances. especially compared with resource-revenue Resource-revenue transfers to regional and local transfers. With the strong increase in revenue governments increased from 0.4 percent of GDP from the extractive industries, resource-revenue in 2002 to almost 2 percent in 2012, then fell transfers became the country’s most important to 1.5 percent in 2014. While this substantially fiscal transfer system. For example, in 2012 the improved subnational government finances, it largest per capita resource-revenue transfer also increased fiscal risks. Resource-revenue transfers represent almost 10 percent of total regional revenues and more than 30 percent 41  There is also some question as to whether ordinary resource of local government revenues. The increasing transfers are based on an accurate assessment of expenditure dependence of subnational governments on an needs that fully reflects local demand for public services. 42  FONCOR’s limitations stem from three sources. First, its inherently volatile revenue source, their limited equalization effectiveness is limited because the pool of funds capacity to mitigate revenue shocks, and the distributed is relatively small and fixed at PEN 690.33 million. absence of national-level stabilization mechanisms In 2015 FONCOR represented less than 3 percent of the total leaves subnational governments highly exposed to revenue of regional governments. Second, its distribution commodity-price volatility. formula has a strong fixed component (23.6 percent) and a variable component that partially reflects equalization criteria (76.4 percent). However, 50 percent of the variable component There are a number of policy options for enhanc­ is inertial and fixed, leaving only 38.2 percent of the PEN ing the efficiency, predictability, and equalization 690.33 million truly variable. Third, funds are earmarked impact of intergovernmental transfers. The most exclusively for public investment projects, which limits the direct way, albeit the most politically difficult and ability of regional governments to improve the provision of perhaps least viable one, would be to modify the decentralized services and equalize access to them. 38    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 15: Per capita spending by regional Figure 16: Per capita spending by municipalities, governments, 2014 (in constant 2007 PEN) 2014 (in constant 2007 PEN) Tumbes Ayacucho Apurimac Huancavelica Amazonas Pasco Cusco Tacna San Martin Loreto Ucayali Lima Arequipa Callao Huánuco Ica Ancash Junín Puno Cajamarca 62–400 Lambayeque 401–810 Piura 811–4,625 La Libertad Regions 0 500 1,000 1,500 Source: Ministry of Economy and Finance. In constant 2007 PEN Source: Ministry of Economy and Finance. origin basis of transfers by allocating a share to alter the FONCOMUN distribution formula to of resource revenues according to equalization include revenue- and expenditure-capacity criteria. principles. While the majority of resource-revenue The third would be to establish a stabilization fund transfers would be distributed according to their using natural resource revenues. point of origin as compensation for the negative externalities generated by the extractive industries, the use of equalization criteria would mitigate the Ordinary Resources and FONCOR disproportionality of their impact. Since the process of establishing macro- regions stalled, regional governments have This chapter simulates several prospective played an increasingly important role in reforms to the intergovernmental transfer the decentralization process. While regional system aimed at improving its efficiency, as governments were supposed to be temporary well as its equalization and stabilization effects. entities, they have become the established The first would be to reform ordinary resource intermediate level of government. Their main transfers into a revenue-sharing mechanism financing source continues to be ordinary resource for regional governments based on two equally transfers from the central government budget. weighted distribution criteria—regional GDP and regional population. Under this scenario, the size The use of ordinary resources for inter­ of FONCOR transfers would also be increased, and governmental transfers is atypical, and presents steps would be taken to enhance its equalization several important drawbacks. As ordinary role. The second prospective reform would be Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    39 Figure 17: Relation between per capita ordinary resources and per capita GDP y = 0.00000178*(gdppc^2) – 0.0727216*(gdppc) + 1541.883 1,400 1,300 1,200 rroo pc (in nuevos soles) 1,100 1,000 900 800 20,427 700 600 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 gdp pc (in nuevos soles) Sources: MEF and World Bank calculations. resource allocations are based on historical or execution and medium-term budget planning. In inertial criteria, such as existing expenditure needs, addition, it introduces an element of unfairness, regional governments face weak incentives to use as regional government bargaining power likely these resources efficiently. As a result, ordinary affects transfer allocations. Indeed, the distribution resource transfers tend to undermine the quality of of ordinary resources favors very poor and very public financial management and the efficiency of rich regions and has an ambiguous effect on public spending at the regional level. This system equalization (Figure 17). The lack of a clear rule for also introduces perverse incentives to expand public determining the total pool of funds to be transferred staff and facilities, as rising expenditure needs also generates uncertainty and weakens budget are likely to entail increased revenue transfers. execution and medium-term planning. While the increase in ordinary resource transfers designed to compensate for the recent decline in Ordinary resource transfers could be reformed by resource-revenue transfers has proven effective dividing them into two components: a formula- in stabilizing regional budgets, it is unclear how driven revenue-sharing component and an this policy response has impacted the quality and equalization-transfer component. The revenue- efficiency of subnational public spending. sharing component could be funded through a revenue-sharing rule similar to the one planned for The lack of clarity in the distribution criteria financing the macro-regions, which would have set for ordinary resource transfers and the rules for aside 50 percent of income-tax and VAT revenue. determining the pool of funds to be distributed That share would likely need to be recalibrated increases the unpredictability of regional to reflect the new administrative context. In the government budgets. A substantial share of simulations presented below, funding is equivalent ordinary resource transfers is determined through to two-thirds of the ordinary resource transfers in-year negotiations between the regional and for 2016, which would represent approximately central governments. This process generates 10 percent of current income tax and VAT revenue. uncertainty that negatively affects annual budget The revenue-sharing distribution formula is based 40    Peru: Building a More Efficient and Equitable Fiscal Decentralization System on two equally weighted criteria—population and where they are generated and redistributing them regional GDP. These criteria are commonly used to other regions. in subnational revenue-sharing systems, as the inclusion of regional GDP reflects the point-of-origin The first scenario, which simulates the impact of principle in resource transfers, while population size the reformed ordinary resource transfer system, reflects the equalization principle.43 uses two-thirds of the ordinary resources transferred in 2016. The allocation is based on two The equalization impact of regional transfers equally weighted criteria, the regional population could be further enhanced by reforming as a share of the total population and regional GDP FONCOR. Under the first reform scenario, the pool as a share of total GDP in 2014. The computation of FONCOR funds would be substantially increased of the formula is presented in Appendix 1. In this by adding the remaining one-third of the pool of scenario the enhanced FONCOR transfer system ordinary resources for 2016, which is equal to uses a pool of funds that is equal to one-third of approximately 5 percent of income-tax and VAT ordinary resource transfers in 2016. revenue. The equalization criteria of the current FONCOR formula would be preserved in the first Two alternative formulas are used to allocate scenario, and it would account for expenditure these resources. In the first scenario, the new needs and revenue capacity, including resource- pool of resources is distributed using the existing revenue transfers. In the second scenario, the FONCOR formula (see Appendix 2). The “fixed” and FONCOR formula would be modified to enhance the “base” amounts of the FONCOR remain the same its equalizing role. Finally, under both scenarios as they were in 2016, while the additional resources transfers would no longer be earmarked for (one-third of ordinary resource transfers in 2016) capital investment and would instead represent are distributed using the existing FONCOR index. an unconditional transfer designed to finance the In the second scenario, the new pool of resources decentralization of functions currently financed by is also distributed based on the “fixed” and the ordinary resource transfers. “base” amounts of the “old FONCOR” formula, but the additional FONCOR funds are distributed using These reforms would result in a more transparent a “new FONCOR” formula. This new formula keeps and predictable financing system for ordinary the existing criterion of “relative poverty,” but resource transfers based on revenue-sharing the second criterion, the “relative transfer” index, and equalization criteria. The size of the revenue- is adjusted to reflect fiscal capacity. The fiscal- sharing mechanism and the enhanced FONCOR capacity criterion includes both resource-revenue equalization pool could be altered to make the transfers and the new revenue-sharing funds. entire system more equalizing. The more funds This methodology is described in Appendix 3. The that are allocated to the equalization component proposed change in the FONCOR formula also and the larger the weight of the population includes a more straightforward and transparent criterion in the revenue-sharing formula, the distribution approach that eliminates the logit greater the equalization effect. The new formula regressions used in the current distribution formula. for FONCOR used in the second scenario could also be altered by changing its weights to increase the The results of the simulations are summarized in equalization effect of the transfer. The extent of the Table 8. Column 1 shows the current distribution equalization effect is ultimately a political economy of ordinary resource transfers, as well as funds decision, and different countries will reach different transferred under the current FONCOR formula. balances between allocating resources to the areas Figure 18 illustrates the geographical distribution of these resources. Column 2 in Table 8 shows the regional distribution of funds based on the 43  It is important to note that this distribution would differ proposed revenue-sharing mechanism plus significantly from the previously envisioned revenue-sharing FONCOR transfers using the current formula but mechanism for the macro-regions, which was intended to the revised resource pool. Figure 19 presents the reflect only the point-of-origin principle. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    41 Table 8: Simulations transforming the system for ordinary resource transfers into a revenue-sharing- plus-FONCOR equalization grant, 2016 (in PEN millions) (1) (2) (3) Old New revenue New revenue (4) = (2) – (1) (5) = (3) – (1) ordinary sharing 2016 sharing 2016 plus Winners Winners Regional resources plus FONCOR reformed FONCOR and losers and losers government 2016 (1st scenario) (2nd scenario) (1st scenario) (2nd scenario) Amazonas 264 461 260 197.4 –3.6 Ancash 451 370 416 –80.9 –34.5 Apurimac 299 481 269 182.4 –29.2 Arequipa 497 489 514 –7.9 16.7 Ayacucho 449 423 320 –25.8 –128.4 Cajamarca 593 426 563 –166.7 –29.6 Callao 276 373 397 97.9 121.2 Cusco 480 450 504 –30.0 24.2 Huancavelica 309 458 238 148.7 –71.0 Huánuco 343 573 432 230.5 88.7 Ica 327 303 330 –24.3 2.4 Junín 492 502 557 9.9 65.0 La Libertad 553 524 643 –28.6 89.5 Lambayeque 505 453 538 –52.1 32.8 Loreto 521 256 398 –265.1 –122.9 Madre De Dios 113 145 112 32.6 –0.5 Moquegua 133 137 159 4.1 25.7 Pasco 154 354 192 199.9 37.9 Piura 558 517 641 –41.5 82.4 Puno 542 605 595 62.9 52.2 San Martin 415 387 403 –27.5 –11.8 Tacna 150 133 156 –16.8 6.1 Tumbes 156 73 98 –82.4 –58.0 Ucayali 255 127 215 –128.3 –39.5 Lima region 421 233 305 –188.4 –116.0 Sources: MEF and World Bank Calculations. corresponding geographical distribution of funds. formula presented in the first scenario and the Column 3 in Table 8 shows the regional distribution current distribution of ordinary resource transfers of funds using the new revenue-sharing mechanism shown in column 4 of Table 8. The corresponding plus the new FONCOR formula. Figure 20 presents geographical distribution of gains and losses is the corresponding geographical distribution. depicted in Figure 21. Finally, the gains and losses for the second scenario are shown in column 5 of The proposed reforms generate regional Table 8, and illustrated in Figure 22. gains and losses. These are computed as the difference between the sum of the new revenue- Avoiding drastic changes in the resources sharing mechanism plus the original FONCOR available to regional governments will require 42    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 18: The current distribution of ordinary revenue transfers, 2016 (in PEN millions) 155.9 521.1 263.6 558.3 505 592.7 414.6 553 450.7 309.2 254.9 153.8 492 275.5 112.8 421.3 342.9 479.6 448.5 298.6 327.2 542.5 Ordinary resources for 2016 497.2 (in millions of nuevos soles) 112.8–155.9 133.3 156.0–342.9 149.5 343.0– 505.0 505.1–592.7 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Note: This figure corresponds to column 1 in Table 8. Metropolitan Lima (in yellow) is not part of the simulation. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    43 Figure 19: The distribution of ordinary resources under the new revenue-sharing mechanism and the original FONCOR distribution formula (in PEN millions) 73.5 256 516.8 461 453 425.9 387.1 524.4 369.8 457.9 126.6 353.7 501.8 373.4 145.4 232.9 573.4 449.6 422.8 481 302.9 605.4 Reformed ordinary resources—1st scenario 489.4 (in millions of nuevos soles) 137.3 73.5–145.4 145.5–387.1 132.7 387.2–501.8 501.9–605.4 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Note: This figure corresponds to column 2 in Table 8. Metropolitan Lima (in yellow) is not part of the simulation. 44    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 20: The distribution of ordinary resources under the new revenue-sharing mechanism and the new FONCOR distribution formula (in PEN millions) 97.9 398.2 640.8 260 537.8 563.1 402.8 642.6 416.2 238.2 215.4 191.7 556.9 396.7 112.3 305.4 431.6 503.8 320.1 269.4 329.5 594.7 Reformed ordinary resources—2nd scenario 513.9 (in millions of nuevos soles) 159 98–215 216–330 155.6 331–432 433–643 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Note: This figure corresponds to column 3 in Table 8. Metropolitan Lima (in yellow) is not part of the simulation. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    45 Figure 21: Relative gains and losses from the introduction of the proposed revenue-sharing mechanism and the original FONCOR distribution formula (in PEN millions) –82.4 –265.1 197.4 –41.5 –52.1 –166.7 –27.5 –28.6 –80.9 148.7 –128.3 199.9 9.9 97.9 32.6 –188.4 230.5 –30 –25.8 182.4 –24.3 62.9 Winners & losers—1st scenario –7.9 (in millions of nuevos soles) –265.1– –128.3 4.1 –128.2–9.9 –16.8 10.0–97.9 98.0–230.5 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Note: This figure corresponds to column 4 in Table 8. Metropolitan Lima (in yellow) is not part of the simulation. 46    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 22: Relative gains and losses from the introduction of the proposed revenue-sharing mechanism and the new FONCOR distribution formula (in PEN millions) –58 –122.9 82.4 –3.6 32.8 –29.6 –11.8 89.5 –34.5 –71 –39.5 37.9 –116 65 121.2 –0.5 88.7 24.2 –128.4 –29.2 2.4 52.2 Winners & losers—2nd scenario 16.7 (in millions of nuevos soles) 25.7 –128.4– –116.0 –115.9– –29.2 6.1 –29.1–37.9 38.0–121.2 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Note: This figure corresponds to column 5 in Table 8. Metropolitan Lima (in yellow) is not part of the simulation. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    47 Table 9: Simple regressions of current ordinary resource transfers and transfers under scenarios one and two on GDP per capita (1) (2) (3) (4) Current ordinary Reformed new Reformed FONCOR— Reformed FONCOR— resources revenue sharing 1st scenario 2nd scenario Variables Model 1 Model 2 Model 3 Model 4 GDP per capita 0.00324 0.0122 –0.00899 –0.00378 (0.00345) (0) (0.00726) (0.00289) Constant 444.3*** 139.8 345.6*** 226.7*** (53.41) (0) (112.3) (44.74) Observations 25 25 25 25 R-squared 0.037 1.000 0.063 0.069 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 a transition strategy for phasing in the new FONCOR formula used in the second scenario may system. The most gradual approach would entail enhance the equalizing impact of the proposed the use of a “hold harmless” strategy, in which reforms. all regions would receive the same amount they received the year prior to the reform and only the annual increase in funds would be applied in each FONCOMUN subsequent year. The transition period can be The set of proposed FONCOMUN reforms open-ended or limited to a specific timeframe. explored here would both increase its size and alter its distribution formula. With respect to the The equalization impact of the proposed reforms latter, the new formula would include a revenue- on ordinary resource transfers and FONCOR is capacity criterion, reflecting both the capacity shown by regressing the regional distribution to raise own-source revenues and other revenue of per capita ordinary resource transfers before resources received from transfers (excluding and after reforms with regional per capita GDP. those for equalization purposes). FONCOMUN is a Table 9 shows the results from regressing the traditional equalization fund financed through the current distribution of ordinary resource transfers municipal promotion tax, which is a surtax of up and the distribution under scenarios one and two to 2 percent applied to the VAT collected by the with regional per capita GDP. Column 1 shows national government. One option to increase the that the current distribution of ordinary resource size of FONCOMUN would be to increase the surtax transfers is largely neutral with respect to the rate and reduce the national VAT rate to keep the regional distribution of per capita GDP. This is also tax burden from VAT constant. FONCOMUN’s pool the case for scenario one, as the revised allocation of funds could also be increased by allocating for is based on regional populations and regional per that purpose a share of the overall pool being used capita GDP, and thus it is closely linked with GDP for all other transfers. per capita. FONCOR transfers under scenario one (column 3) and scenario two (column 4) appear to be Adjustments in FONCOMUN’s distribution mildly redistributive and progressive, as indicated formula are also needed. In line with FONCOR’s by the negative regression coefficient, though recent reforms, the most obvious adjustment in neither case is that coefficient statistically would be to add revenue capacity as a distribution significant. The apparently more progressive criterion. Currently, the distribution formula only 48    Peru: Building a More Efficient and Equitable Fiscal Decentralization System takes into account differences in expenditure which ensures that the FONCOMUN transfer to needs. As revenue capacity would include income each district is at least equal to the amount they from resource-revenue transfers (other than received in 2009 (in real terms). equalization), its inclusion would help correct the inequalities generated by the unevenly distributed The complexity of the three-stage distribution revenues from extractive industries. This would rule undermines the intended equalizing effect also require adjustments in the current legal use of of FONCOMUN. The application of the first the resource-revenue transfers. Because resource- and second stages may result in cases in which revenue transfers are currently earmarked almost two districts with identical populations; unmet exclusively for capital investment purposes, for needs for public services; and rurality, territory many governments that receive resource-revenue and municipal management receive different transfers, the funds from FONCOMUN are actually FONCOMUN transfers just because they are the only revenue source they have with which located in provinces that at the aggregated to finance recurrent expenditures. Relaxing the level have different socioeconomic conditions rules for using these resources would help offset (population and unmet needs). Moreover, the the potential decrease in FONCOMUN funds that application of the minimum level and hold harmless subnational governments would likely experience adjustments in the third stage may distort the as a consequence of including revenue capacity as supposed equalizing effect of the first and second an additional criterion in the distribution formula. stages. In fact, after performing the third stage allocation adjustment, it is hard to guarantee A second adjustment would be the simplification that the allocations corresponding to the first and of the FONCOMUN’s three-stage distribution second phase actually are maintained. rule. FONCOMUN’s distribution rule consists of three stages. In the first, the FONCOMUN national The following scenario simulates the effect of pool is distributed among provinces (which include reforming the FONCOMUN formula to additionally provincial municipalities and district municipalities) reflect the fiscal capacity of local governments, according to the population of the province and especially by accounting for resource-revenue an equalizing criterion represented by the index transfer funds. The proposed reform would only of unmet needs for public services. FONCOMUN alter the first stage of the distribution methodology transfers to each province are divided between the by including the fiscal capacity aggregated at the provincial municipality (20 percent) and the district provincial level as an additional criterion in the municipalities (80 percent). In a second stage, the initial allocation of the pool of resources. It must 80 percent of FONCUMUN transfers to the province be recalled as outlined above that FONCOMUN’s that correspond to the district municipalities distribution methodology is based on three stages: are allocated among them according to three (i) a geographical index combining population and factors: rural status, territory size, and municipal unmet needs for public services is used to allocate management. Rurality and territory are intended the pool of funds across all the provinces; (ii) several to reflect expenditure needs while the third factor, other criteria are used to allocate the funds for municipal management, is intended to stimulate each province across districts within each province; improvements in public financial management. The and (iii) the minimum amounts that each district three factors are applied with the following weights: should receive are established. The new proposed 85 percent for rurality, 5 percent for territory, methodology is explained in detail in Appendix 5. and 10 percent for municipal management. In the third stage two types of adjustments are adopted: The new criterion is based on an index of fiscal the first is to ensure that district municipalities capacity measured at the provincial level. Fiscal receive a minimum level of FONCOMUN transfer capacity is measured by examining the sum of all corresponding to eight tax units, and the second the transfers received by the municipalities in the that corresponds to the hold harmless clause province, including resource-revenue transfers— (established when FONCOMUN was last reformed) but excluding the FONCOMUN itself—and potential Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    49 own-source revenues. Potential own-source metropolitan Lima, 35 would see losses under the revenues are calculated through a regression new formula. This includes the district municipality analysis of per capita local governments’ own- of Santa Maria del Mar and the district municipality source revenues, such as taxes and user fees on of Comas (PEN –3). Among the eight municipalities per capita household spending (as a proxy for per in metropolitan Lima that would gain under the capita income). reform, seven municipalities would see gains of less than PEN 3.09, although the provincial municipality The inclusion of local government fiscal capacity of Lima would see a gain of PEN 242. as a distribution criterion significantly alters the allocation of FONCOMUN funds. Figure 23 It is important to note, that the equalization shows the geographical distribution of actual impact of the proposed reform on FONCOMUN is FONCOMUN transfers at the district-municipality weaker than the one of the current FONCOMUN. and provincial-municipality levels for 2014. This is shown by regressing the regional distribution Figure 24 depicts the results of the simulation of per capita FONCOMUN transfers before and using the new distribution of FONCOMUN funds in after the proposed reform of FONCOMUN with 2014. The geographical distribution of gains and the variable municipal household expenditure losses is shown in Figure 25. per capita. Table 10 shows that the implied re- distribution of current FONCOMUN transfers is The gains and losses resulting from the slightly higher than the one that would result from FONCOMUN reform are approximately the proposed FONCOMUN reform. bounded within +/– PEN 250 per capita, which represents about 16.6 percent of average local The lack of a redistributional impact from the expenditures per capita. Gains and losses were introduction of a fiscal capacity indicator in defined based on a comparison of the per capita the FONCOMUN formula signals the need to allocation of FONCOMUN funds for 2014 under complement this reform by streamlining the the new formula and the existing formula. Out current three-stage distribution rule. In fact, the of the 1,845 local governments included in the equalizing effect of FONCOMUN is undermined simulation, 669 were classified as losers under the not only by the lack of fiscal capacity criteria new formula, with losses varying from PEN 284 but mainly by the complexity of the three- (the district municipality of Santa Maria del Mar stage rule which treats districts with the same located in metropolitan Lima) to PEN 0.005 (the characteristics differently, defines minimum district municipality of Sachaca in the Arequipa FONCOMUN transfers, and eventually works to region). Among the losers there were 118 provincial preserve the distribution criteria and results of municipalities, with losses ranging from PEN 280 previous arrangements. In particular, without (the provincial municipality of Jumbilla in Amazonas relaxing over time the minimum transfer clauses region) to PEN 0.6 (the provincial municipality of in the third stage of the formula, little could be Trujillo in La Libertad region). On the other hand, changed in terms of redistributing resources. 1,176 local governments would gain under the new formula, with amounts ranging from PEN 0.06 (the But redistribution should be gradual without district municipality of Surquillo in metropolitan imposing drastic changes in municipal budgets. Lima) to PEN 314 (the district municipality of Rosa A comprehensive reform of FONCOMUN should Panduro in Loreto region). Among the provincial avoid drastic changes in the resources available to municipalities that would benefit from the new provincial and district municipalities. This can be formula, the gains range from PEN 0.9 (the accomplished by adopting a transition strategy for provincial municipality of Contamana in Loreto phasing in the new system over a period of three region) to PEN 242 (the provincial municipality to five years. However, a hold harmless clause is of Lima). Among the 43 municipalities located in neither required nor desirable. 50    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 23: Geographical distribution of existing FONCOMUN transfers (in PEN per capita) Old FONCOMUN 2014 (in nuevos soles per capita) 15.4–221.0 221.1–437.0 437.1–890.9 891.0–2087.7 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    51 Figure 24: Geographical distribution of the proposed changes in FONCOMUN transfers (in PEN per capita) New FONCOMUN 2014 (in nuevos soles per capita) 10.6–250.9 251.0–516.9 517.0–1075.6 1075.7–2314.4 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. 52    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Figure 25: Relative gains and losses from the proposed changes in FONCOMUN transfers (in PEN per capita) Winners & losers (in nuevos soles per capita) –284.3– –63.8 –63.7–8.4 8.5–68.1 68.2–247.6 Sources: Ministry of Finance and Ministry of Agriculture; World Bank calculations. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    53 Table 10: Simple regressions of FONCOMUN transfers on household expenditure per capita Reformed FONCOMUN Variables Current FONCOMUN (using adjusted formula) Household expenditure pc_2013 –0.282*** –0.261*** (0.0277) (0.0279) Constant 421.1*** 398.6*** (12.27) (12.35) Observations 1,845 1,845 R-squared 0.053 0.045 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Source: World Bank staff calculations. A Stabilization Fund for address the effects of revenue volatility given the larger number of policy tools at its disposal Resource-Revenue Transfers to increase fiscal revenues. Moreover, the The increasing share of resource-revenue central government’s access to credit markets transfers in the budgets of regional and local would enable it to attenuate exogenous shocks. governments, coupled with fluctuations in Subnational governments, which have substantially commodity prices, have made subnational less revenue sources and limited access to credit revenues less predictable, possibly affecting the markets, would benefit from more stable revenue continuity of service delivery at the local level. sources. A reduction in the share of resource- Faced with increasing expenditure obligations revenue transfers to subnational governments, and a very limited capacity to boost own-source compensated by expanded tax revenue-sharing revenues, the finances of regional and local mechanisms, would help rebalance risks between governments have been affected by the fall in the central government and subnational entities, resource-revenue transfer proceeds, triggering a with the former receiving a greater share of higher compensatory increase in transfers from the cen­ risk revenues. tral government. The current decline in commodity To attenuate the recent fall in resource- prices is highlighting the strong exposure of revenue transfers, the national authorities have subnational governments and may open a window been taking actions to stabilize subnational of opportunity to reconsider the role of resource- governments’ fiscal revenues. The national revenue transfers in subnational public finances. authorities’ policy response of compensating The fiscal impact of the fall in commodity prices on regional and local governments for the fall in subnational government budgets may, moreover, resource-revenue transfers by increasing other reduce anticipated resistance to reforms in the transfers has functioned as a type of revenue- distribution of resource-revenue transfer proceeds. stabilizing mechanism. In 2012–14, the central From a risk-management perspective, the government significantly increased ordinary authorities should rebalance risks among the resource transfers to regional governments, as well various levels of government by increasing the as other current transfers to local governments.44 share of volatile revenues going to the central Furthermore, as local governments were the government and ensuring a more stable stream of fiscal revenues to subnational governments. 44  Ordinary resources to regional governments increased by The central government is better prepared to 23 percent in real terms from 2012 to 2014. 54    Peru: Building a More Efficient and Equitable Fiscal Decentralization System most affected by the fall in resource-revenue The Stabilization Fund for Resource-Revenue transfers, the central government enabled them Transfers could be based on the following to withdraw the unused balances of resource- accumulation and withdrawal rules. Thirty revenue transfer proceeds accumulated in previous percent of the resource-revenue transfer proceeds years.45 Therefore, the accumulated balances of are accumulated in the funds46 whenever the resource-revenue transfers have been serving as actual price of the corresponding commodity a de facto revenue-stabilization fund, as proceeds exceeds its average level observed in 2004–14, accumulated during the period of rising commodity and 10 percent of the accumulated balance of the prices are being used to offset the strong decline in fund is withdrawn whenever the actual price of the extractive industry revenue. corresponding fund is lower than its average level observed in 2004–14. Given the measures already adopted to stabilize subnational revenues, establishing Comparing the actual flow of resource-revenue a Stabilization Fund for Resource-Revenue transfer funds for the period from 2004–14 Transfers could face less political resistance with a simulation based on the above-mentioned than the centralization of resource-revenue rules of accumulation and withdrawal show the transfers. As mentioned above, the authorities stabilizing effects of the proposed instrument. have been using accumulated resource-revenue Figure 26 and Table 11 show how the stabilization transfer proceeds as a sort of revenue-stabilization fund, based on differences between actual and fund. However, the functioning of this stabilization average prices, attenuates the volatility of the mechanism needs to be regulated in at least resource-revenue transfers received by regional five ways. First, the rules for accumulation and and local governments. Volatility is reduced for the withdrawal need to be clearly defined and four most important sources of resource revenue simplified. Second, the subnational stabilization (gas, oil, mining, and mining royalties). For the total fund would need to be integrated into the macro- resource-revenue proceeds, volatility is reduced by fiscal framework and reflected in the fiscal rule for approximately 20 percent. subnational governments. Third, the stabilization role of the proposed fund could be strengthened Automatic rules for accumulating and with­ if the constraint on the use of resource-revenue drawing resources would ensure a stable flow transfer funds to exclusively finance investment of resources, which would in turn support the expenditures and infrastructure maintenance was continuity of service delivery by regional and relaxed. This would enable disbursed resources local governments. While rules governing the from the fund to be assigned to a broader set of inflow of resources could reduce their availability expenditures. Fourth, as the distribution rule for when prices are high, rules for withdrawing resource-revenue transfers among regional and resources could help avoid expenditure cuts that local governments is based on separate ownership would be necessary to balance the budget in the of resources, the stabilization fund’s proceeds absence of stabilization funds. should involve individual separate accounts. Fifth, a transparent and professional investment policy for the resource-revenue transfer balances would be needed. 45 Classified as the sale of assets (unused balances of resource-revenue transfer proceeds), grew by 750 million nuevos soles, equivalent to an increase of more than 100 percent in real terms between 2012 and 2014, partially compensating the fall of 1,500 million nuevos soles in canon 46  There would be seven stabilization funds, one for each type proceeds (or 25 percent decrease in real terms) in the same of resource revenue: mining, gas, oil, hydropower, fisheries, period. forestry, and mining royalties. Toward More Efficient, Equitable, and Stable Intergovernmental Transfers    55 Figure 26: Annual resource-revenue transfers, actual and stabilization-fund scenario, 2004–2014 (in constant 2007 PEN millions) Gas canon Mining canon 2,000 4,500 1,800 4,000 1,600 3,500 1,400 3,000 1,200 Millions Millions 2,500 1,000 2,000 800 1,500 600 400 1,000 200 500 0 0 04 05 06 07 08 09 10 11 12 13 14 04 05 06 07 08 09 10 11 12 13 14 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Actual Fund based on price Actual Fund based on price Oil canon-sobrecanon Mining royalities 1,000 800 900 700 800 700 600 600 500 Millions Millions 500 400 400 300 300 200 200 100 100 0 0 04 05 06 07 08 09 10 11 12 13 14 04 05 06 07 08 09 10 11 12 13 14 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Actual Fund based on price Actual Fund based on price Hydropower canon Forestry canon 180 60,000 160 50,000 140 Thousands Hundreds 120 40,000 100 30,000 80 60 20,000 40 10,000 20 0 0 04 05 06 07 08 09 10 11 12 13 14 04 05 06 07 08 09 10 11 12 13 14 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Actual Fund based on price Actual Fund based on price Sources: MEF; World Bank calculations. 56    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Table 11: Annual resource-revenue transfers, actual and stabilization-fund scenario, 2004–14 (in constant 2007 PEN millions) Actual flows Simulated flows Standard Coefficient Standard Coefficient Average deviation of variation Average deviation of variation Gas 802 675 0.84 490 308 0.63 Mining canon 2,460 1,210 0.50 1,707 775 0.45 Oil 467 244 0.52 316 94 0.30 Mining royalties 359 180 0.50 242 95 0.40 Hydropower 104 28 0.27 80 27 0.39 Forestry 2.1 1.9 0.98 1.8 1.7 0.98 Total 4,242 2,056 0.48 2,879 1,135 0.39 Sources: MEF and World Bank calculations. 7 Conclusions and Policy Recommendations This report presents policy options designed Consolidating regional governments to improve specific aspects of Peru’s fiscal at the permanent intermediate decentralization system. These options are presented in logical order, beginning with (i) the government level vertical institutional arrangements that determine a. Review the legislative framework for regional the relationship between different government governments to formalize their role as a full- levels, then moving on to (ii) the importance fledged intermediate-level government; of clarifying expenditure responsibilities at all government levels, and then (iii) the commensurate b. In parallel establish and enhance institutions redefinition of tax authority to boost subnational for national-regional coordination in decision- own-source revenue capacity and enhance making processes (e.g., establishment of an collection efficiency, and concluding with for tax entity at the national government responsible collection, before concluding with (iv) proposals to for the coordination with regional and local increase the transparency, stability, and equalizing governments, supporting the recently created impact of the intergovernmental transfer system. Association of Regional Governments); Prioritize the resumption of institutional c. capacity-building efforts at the regional level, The Institutional Framework identifying a suitable mechanism at the central More than a decade after the launch of the fiscal level to deliver such programs (the entity at decentralization process, it appears highly the national government responsible for the unlikely that macro-regions will be established, coordination with other levels of government and reform options must focus on strengthening may be responsible for capacity building and the existing regional governments. While accreditation of subnational governments); important progress has been made, the d. Fully incorporate the regional directorates of decentralization process has been slow and uneven. line ministries into the institutional structure In this context, future efforts should focus on of regional governments; and the consolidation of regional governments as the intermediate unit of government. Meanwhile, at the e. Eliminate the current system of ordinary municipal level, policy makers should take steps to resource transfers and replace it by prevent the creation of new municipalities, which (i) assigning tax bases to regional governments has significantly increased the fragmentation of and (ii) establishing an unconditional revenue- local government in recent years. sharing transfer (see options on taxation and intergovernmental transfers presented below). 57 58    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Reducing municipal fragmentation c. Eliminate overlapping responsibilities between regional and local governments, assigning a. Declare a moratorium on the creation of new them to the appropriate level, and balancing municipalities; economies of scale and the subsidiarity b. Complete the legal demarcation of municipal principle; boundaries to end territorial disputes, which Streamline the framework for delegated d. are the most important loophole allowing for responsibilities; and the creation of new municipalities; e. Establish permanent formal intergovernmen­ c. Tighten regulatory requirements to further tal coordination mechanisms to harmonize discourage the establishment of new activities, resolve conflicts, and reach municipalities; agreements on specific issues. Again, a d. Significantly reduce the minimum FONCOMUN national government entity and representative transfer level, diminishing the incentive to associations of regional/local governments create new municipalities; and may play a facilitating role. Design stronger incentives for municipal e. consolidation and cooperation in service delivery. Taxation Narrowing the vertical fiscal gap will require increasing the tax revenue generated by regional Expenditure Assignments and municipal governments. This can be achieved through the combination of broadening tax bases A clear definition of expenditure responsibilities assigned to them, enabling them the ability to will be necessary to increase the autonomy and define rates for their own taxes, and improvements accountability of subnational governments in tax-collection efficiency. Regional governments and improve the efficiency of service delivery. completely lack tax bases while local governments In particular, further clarification is needed of the have small bases but without autonomy to define shared or concurrent functions, by disentangling rates. Proposed are the following policy options sub-functions to define what level of government is to increase subnational taxation. It is, however, ultimately responsible for the sub-function without important to mention that tax and revenue sharing ambiguities. A clear differentiation between reform options for subnational governments need deconcentrated and delegated functions and the to be discussed based on a proper reassessment of definition of financing sources (own-revenue, block the overall tax system and their compound effects grants, or conditional grants) in each case would on labor, investment, and consumption decisions also be needed. As this work tends to be inherently that affect growth and income inequality. This dependent on the nature of the sector (e.g., health, consideration is also important because Peru has roads, etc.), a commission comprised of sectoral a relatively small tax revenue base as a whole and, and fiscal experts will probably need to be mobilized thus, only sharing or devolving existing taxes might for each sector. be at the expense of central government’s ability to address national-level priorities if the overall tax Clarifying expenditure responsibilities base is not expanded. a. Review the organic laws for regional and local governments to identify sectors with Assigning tax bases to regional duplicative or overlapping responsibilities; governments b. Unbundle shared responsibilities into sub- Establish a regional surcharge on the a. functions (regulation, financing, and service personal income tax collected by the delivery) and allocate responsibility for each national government to be levied at a flat sub-function to the appropriate government level following the subsidiarity principle; Conclusions and Policy Recommendations    59 rate to minimize interregional labor factor into a more transparent and predictable movements; financing system based on revenue sharing and equalization criteria. The size of the revenue- b. Enable regional governments to define a local sharing mechanism and an enhanced FONCOR rate for the surcharge of 1-3 percent; and equalization pool could be adjusted to enhance the c. Explore other tax-policy options, including equalizing properties of the entire system. The new the establishment of a regional payroll formula for FONCOR used in the second scenario tax, a presumptive income business tax, or could also be adjusted by changing its weights to surcharges on existing national excise taxes or increase the equalization effect of the transfer. A on potentially new ones.47 fiscal simulation of the impact of these reforms on the finances of regional governments was Improving the efficiency of revenue presented in detail in Chapter 6. collection at the local level A number of reforms to FONCOMUN will enhance a. Give municipalities discretion to set property- its equalizing properties and remove the tax rates within a limited range determined at incentives to create new municipalities. These the national level; reforms would also require the earmarking of resource-revenue transfers to capital investment b. Establish a national framework for cooperation projects, enabling a broadened use. These reforms between regional and municipal revenue include the inclusion of fiscal capacity estimates agencies, or for developing administrative in the distribution criteria and the streamline of cooperation agreements for shifting tax the three-stage distribution. A simulation of the administration responsibilities (e.g., collection) impact of these reforms on the finance of each between government levels, or between district municipality was presented in detail in Chapter 6. and provincial tax agencies; c. Streamline user charges and fees collected by To offset the recent decline in resource-revenue local governments and simplify licensing and transfers, the central government has increased payment procedures; and the size of other transfers. This has stabilized subnational budgets, especially at the municipal d. Create an Office of the National Cadaster to level, where the drop in resource-revenue transfers enhance property-tax collection by managing was most acute. In addition to increasing other the registration and valuation of properties at transfers, the central government enabled the national level. municipalities to withdraw and spend the unused balances from resource-revenue transfers in previous years. Intergovernmental Transfers Establishing a Stabilization Fund for Resource- The proposed reforms in this area would Revenue Transfers backed by clear rules for transform the ordinary resource transfers accumulation and withdrawal could further reinforce the stability of subnational budgets. Such a mechanism would enable resource revenues 47 To expand their tax bases, an increasing number of accumulated during periods of higher-than- countries and most recently Mexico (2014) and Chile (2015) have adopted excise taxes on tobacco, alcoholic, and sugar- average commodity prices to be withdrawn when containing non-alcoholic beverages and junk food with commodity prices fell, reducing the volatility of the corrective effects on consumption and beneficial impacts on intergovernmental transfer system. health. 60    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Establishing a more stable, less of resource-revenue transfers and other discretionary, and more equalizing transfers (except transfers from FONCOMUN itself)—in order to enhance its equalization system of transfers to regional effect; governments Ease the earmarking rules for resource- b. a. Reform ordinary resource transfers by: revenue transfers, and enable provincial and district governments to use them to finance i. Defining a share of personal and corporate recurrent expenditures, helping to offset the income-tax (excluding the corporate decrease in FONCOMUN transfers that may income tax on extractive industries that is result from the proposed reforms; already transferred via the canon) and VAT revenue equivalent to the current level of c. Eliminate minimum transfers, at least for ordinary resource transfers; newly created jurisdictions, as they incentivize the proliferation of municipalities; ii. Dividing that pool of funds into a formula- driven revenue-sharing component (two- d. Reform FONCOMUN’s three-stage distribution thirds) and an equalization-transfer rule by creating separate criteria for transfers component (one-third); to provincial and district municipalities; iii. Basing the revenue-sharing distribution e. Pursue a gradual and properly sequenced formula on two equally weighted criteria— reform process, which may include temporary regional GDP and population—in order compensation mechanisms to minimize to incorporate both the point-of-origin budgetary shocks; and principle for resource transfers and the f. Explore options to increase the pool of resources equalization principle; and transferred under the reformed FONCOMUN iv. Developing a strategy for phasing in the mechanism by, for example, increasing the new system that minimizes the shock to national VAT surcharge or diverting ordinary regional government budgets. resource transfers to FONCOMUN. b. Enhance the equalization impact of regional transfers by adding a share (one-third) of Reducing the volatility of resource- ordinary resource transfers to the pool of revenue transfers FONCOR funds; a. Establish a stabilization fund for resource- c. Either preserve the current FONCOR formula, revenue transfers based on well-defined rules or alter it to better account for regional for accumulation and withdrawal; differences in expenditure needs and revenue capacity, thereby enhancing its equalization Integrate the stabilization fund into the b. impact; and macro-fiscal framework and the fiscal rules for subnational governments; d. No longer earmark FONCOR transfers for infrastructure investment, but instead make c. Create separate accounts for each recipient them unconditional transfers designed to government; finance the decentralization of functions d. Ensure that the management of the fund is currently funded by ordinary resource transfers. fully transparent and backed by appropriate oversight and accountability mechanisms; Enhancing the equalizing impact and of FONCOMUN e. Strengthen the system for pooling resources among municipalities to finance a. Alter the distribution formula to include interjurisdictional infrastructure projects. revenue-capacity criteria—including the size Conclusions and Policy Recommendations    61 These proposed policy options reflect four set corresponding distribution of revenue authority. conclusions regarding the relationship between The third is the crucial role that regional and local Peru’s institutional framework and three of own-source revenue capacity plays in the long- the dimensions of fiscal decentralization. term sustainability of fiscal decentralization. And The first is the importance of consolidating the fourth is the significance of the framework for vertical institutional arrangements to reflect intergovernmental transfers in fostering equity of the presumptive permanence of the regional service delivery nationwide while minimizing their governments. The second is the need for a potential negative effects on public expenditure clearer, more precise definition of expenditure efficiency and tax collection efforts. responsibilities between government levels and a 1 Appendix Methodology for New Ordinary Resources and Simulation The pool of resources to be distributed in this new formula is based on an index that considers the revenue-sharing mechanism will be equivalent regional population and regional GDP both with to two-thirds of the current value of Ordinary equal weights of 50 percent (see Table 12). Resources (budgeted in 2016). The distribution Table 12: New ordinary resources distribution index Regional GDP 2014 Population 2014 (thousands of nuevos soles) Regional government Total % Total % Distribution index Amazonas 421,122 0.019 3,235,930 0.013 0.016 Ancash 1,142,409 0.052 16,558,359 0.065 0.059 Apurimac 456,652 0.021 2,710,392 0.011 0.016 Arequipa 1,273,180 0.058 23,985,309 0.095 0.076 Ayacucho 681,149 0.031 5,461,714 0.022 0.026 Cajamarca 1,525,064 0.069 11,817,954 0.047 0.058 Callao 996,455 0.045 19,249,771 0.076 0.061 Cusco 1,308,806 0.059 21,390,383 0.084 0.072 Huancavelica 491,278 0.022 3,535,902 0.014 0.018 Huánuco 854,234 0.039 5,128,940 0.020 0.029 Ica 779,372 0.035 15,373,558 0.061 0.048 Junín 1,341,064 0.061 13,108,553 0.052 0.056 La Libertad 1,836,960 0.083 20,498,411 0.081 0.082 Lambayeque 1,250,349 0.057 10,924,336 0.043 0.050 Loreto 1,028,968 0.047 8,858,322 0.035 0.041 Madre de Dios 134,105 0.006 1,940,575 0.008 0.007 Moquegua 178,612 0.008 8,683,376 0.034 0.021 Pasco 301,988 0.014 5,291,774 0.021 0.017 Piura 1,829,496 0.083 18,875,817 0.074 0.079 Puno 1,402,496 0.064 10,016,943 0.039 0.052 San Martin 829,520 0.038 5,720,598 0.023 0.030 Tacna 337,583 0.015 6,327,816 0.025 0.020 Tumbes 234,638 0.011 2,710,404 0.011 0.011 Ucayali 489,664 0.022 4,092,239 0.016 0.019 Lima provinces 933,749 0.042 8,249,610 0.033 0.037 Total 22,058,913 1.000 253,746,986 1.000 1.000 Sources: Ministry of Finance and National Institute of Informatics and Statistics. Note: Metropolitan Lima is not included in the simulations. 63 2 Appendix Current Methodology of the FONCOR48 The Regional Compensation Fund (FONCOR) was the Ministry of Finance (MEF) through a Ministerial established in article 37 of the Law No. 27783, the Decree, supported by the Decentralization Framework Decentralization Law. It is an equalization National Council (CND) report (now Secretary of transfer seeking to distribute additional resources Decentralization (PCM) on the basis of the proposal for regional governments, under the criteria of issued by the General Directorate of Economic and equity and compensation. These resources should Social Affairs (DGAES) of the Ministry of Finance. be fully used in projects of regional investment that complied with the standards set by the National System of Public Investment (SNIP). Fixed Amount When FONCOR was created in 2004, according to The FONCOR is currently made of the resources the law on decentralization, this would be financed comingfromthe base programs and public investment with the financial resources of the programs and projects of the former Transitory Council of Regional regional projects of the then National Institute of Administration (CTAR) (regional investment), and a Development (INADE) and resources product of the variable part constituted 30 percent of the resources privatization of regional companies. generated by privatizations and concessions, as well as an additional amount of regular resources from Currently the government still transfers to the the Treasury, established by the general direction of regional governments that amount as stipulated in the public budget. the Framework Decentralization Law.49 The use of such expenses is mainly operating expenses as the According to article 39 paragraph 2 of the payment of the project workers, insurance works, Law No. 27783—Decentralization Base Law— etc.50 FONCOR is distributed proportionally among all regional governments with criteria of equity and compensation, considering factors of poverty, 49 Article 39—Regional compensation fund 39.1. The Fund of unmet needs, border location, population, tax Regional compensation (FONCOR) was initially constituted with: contribution to the Treasury, and performance (a) financial resources corresponding to all investment projects of regional scope in charge of respective Regional indicators in the execution of public investment. Administration transitory Council, and to all the public investment projects of regional scope in the field of Article 39 of the Law No. 27783 and article 15 of agriculture, fishery, industry, agro-industry, trade, tourism, the Law No. 28411, General Law of the National energy, mining, roads, communications, education, health, and Budget System, established that the Regional environment, present in his constituency, in accordance with Compensation Fund (FONCOR) index is approved by the principle of neutrality and fiscal responsibility with equity and compensation criteria considering factors of poverty. (b) the proceeds from the process of privatization and 48  The information in this chapter has been extracted concessions, as established in the third complementary from the website of the Ministry of Finance (MEF). https:// provision of this law. www.mef.gob.pe/index.php?option = com_content&view = 50  Two known cases are the Sihuas Majes project in Arequipa article&id=284&Itemid=100847&lang=en and the Chavimochic project in the North. 65 66    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Nineteen regional governments receive this fixed Variables or criteria used for the amount; regions that do not get this fixed amount construction of the index to allocate never had such projects. The total fixed amount is the amount to be distributed S/.162,841,356. Following the criteria established in the Law No. 27783, the index is composed of the following Variable Amount indicators and sources of information: FONCOR also features a variable amount that Allocation methodology sums S/.527,488,644, which comes from the The methodology to be used for the construction of Ordinary Resources of the public Treasury. the FONCOR index is based on the prioritization of regions through a probabilistic approach. The allocation of this amount comes in two parts: a base amount (50 percent) and an amount to be To define and identify as priority regions, the distributed (50 percent). The aim of this structure methodology is based on Box 4, where “type A” is that the implementation of the methodology regions have higher priority since they have fewer does not generate distortions in the allocation of resources for investment projects and have higher resources or leave investment projects unfinished. rates of poverty or unmet needs. On the contrary, The variable amount is structured into two parts “type D” regions are lower priority since they have to establish a gradual and permanent allocation more resources and lower poverty rates. rule so that regional governments receive annually not less than 50 percent of what they received the In that sense, the probabilistic approach defines previous year. what is the probability that a region is type A given its poverty and transfers indicators. In this sense, Base amount it will be assigned greater amounts of FONCOR to those regions that are more likely to be type A. The base amount corresponds to 50 percent of Two relative indicators will be used (poverty and what regional governments received the previous transfers), which will be explained. year as a variable amount. In that sense, the distribution of the variable amount of FONCOR a. Equity criteria for 2017 will assign as base amount, 50 percent of The equity criterion is based on the construction of what was assigned as a variable amount in 2016. a relative poverty index, in which the variables type population, unmet needs, and extreme poverty are The aim of this annual fixed rate is to not affect the considered. The construction is as follows: planning of regional governments in relation to this source of funding. (a.1) Poverty index (p): The total base amount is S/. 263,744,322.51 p = POB20.1 × (0.5 Car + 0.5 Pobex) Amount to be distributed Where: The remaining 50 percent of the variable amount • POB252: regional population + regional population in frontier districts. is assigned according to a methodology based on relative probabilities estimated through a • Car: unmet needs index in the region. logit model. The methodology incorporates two dimensions: one of deficiencies or relative poverty • Pobex: extreme poverty rate in the region. and the other on transfers received. 52  The 0.1 exponent is justified to reduce the urban bias that may have the population variable (like Metropolitan Lima and other major cities). On the other hand, it assumes a heterogeneous production scale where the unit costs of the investment project are higher in more dispersed 51  The total amount is frozen since 2009. areas, which are associated with poorer localities. Current Methodology of the FONCOR    67 Table 13: FONCOR distribution index Relation to the Criteria (Law Variables Description Source FONCOR index 27783, Art. 39) 1 Population Regional population INEI-census (latest Positive Population data available) 2 Population Population in frontier districts INEI-census (latest Positive Border location in frontier data available) areas 3 Extreme Percentage of the population in INEI-ENAHO (latest Positive Poverty poverty rate extreme poverty data available) 4 Unmet needs Average of the rates of Census 2007 Positive Unmet needs index population with a deficit in access to water and sewage (to public network) and electricity 5 Transfers Transfers to regional MEF-SIAF Negative Transfers governments by concept of canon, sobrecanon, mining royalties, income from customs, and FOCAM 6 Balance Balance of “certain resources” MEF-DNTP Negative Execution of transfers (recursos determinados) (does not investments including FONCOR) to December 31 of the year prior to the calculation Sources: Ministry of Economy and Finance; World Bank. Box 4: Prioritization criteria of regions Poverty/unmet needs More Less I. Priority regions: Less A B Type A (less resources and poorest) Transfers II. Intermediate regions: Type B (less resources and less poor) Type C (more resources and poorest) III. Low priority regions: More C D Type D (more resources and less poor) 68    Peru: Building a More Efficient and Equitable Fiscal Decentralization System (a.2)  Relative poverty index (P): the balance of “certain resources” (recursos determinados) (does not include FONCOR) to pi P= December 31 of the year prior to the calculation. ∑ 26 p i =1 i It is the regional government disposable income for investment projects in a given year. b. Compensation criteria • POB: regional population The compensation criterion is based on the construction of a relative transfer index, in which (b.2)  Relative transfer index (T): the variables of regional governments’ available t income of certain resources for public investment T= ∑ 26 projects are considered: t i =1 i (b.1)  Transfer index (t): c. Calculation of the FONCOR index (Disposable Income ) For the calculation of the FONCOR index, it was t= estimated as a probabilistic logit model based on a POB0 .1 pool of data on indicators of poverty and transfers. • Disposable income: transfers to regional The results in Table 14 show a degree of considerable governments by concept of canon, sobrecanon, mining royalties, income from customs, and significance (all under a 0.01) in the constant and FOCAM in the year of the calculation plus independent variables of the model. Table 14: Probabilistic—Logit regression . logit prioridad ipobreza itransferencias Iteration 0:   log likelihood = –127.56404 Iteration 1:   log likelihood = –81.996729 Iteration 2:   log likelihood = –65.872284 Iteration 3:   log likelihood = –61.320306 Iteration 4:   log likelihood = –61.15241 Iteration 5:   log likelihood = –61.152096 Iteration 6:   log likelihood = –61.152096 Logistic regression Number of obs = 208 LR chi2(2) = 132.82 Prob > chi2 = 0.0000 Log likelihood = –61.152096 Pseudo R2 = 0.5206 prioridad Coef. Std. Err. z P>|z| [95% Conf. Interval] ipobreza 138.995 25.05356 5.55 0.000 89.89087 188.099 itransfere~s –111.6345 22.51362 –4.96 0.000 –155.7603 –67.50858 _cons –4.74878 1.016194 –4.67 0.000 –6.740484 –2.757076 Note: 10 failures and 0 successes completely determined. Current Methodology of the FONCOR    69 Finally, for the calculation of the odds chances Finally, the index is defined as the probability for “prediction” is taken into account associated with each regional government expressed as a proportion the logit model: of the total sum of the estimated probabilities. ey The simplicity of the methodology is that its annual PROBABILITY (TYPE A = 1) = 1 + ey update will only correspond to the update of the y = −4 .74878 − (111 .6345 × T ) + (138 .995 × P ) indicators “T” and “P” not of the coefficients, which are updated every four years. Table 15 shows the Where: index for the year 2017. • T is the relative transfer index The total amount to be distributed is S/. • P is relative poverty index 263,744,322. Table 15: Calculation of the FONCOR index for 2017 Poverty Transfers FONCOR index Regional government indicator (P) indicator (T) Y Score 2017 Amazonas 0.056189236 0.005554794 2.441136144 0.919910833 0.1353871152 Ancash 0.034302468 0.112737246 –12.5662745 0.000003488 0.0000005133 Apurimac 0.051228587 0.004162577 1.907050173 0.870687387 0.1281426953 Arequipa 0.018422326 0.038607924 –6.49814515 0.001503965 0.0002213448 Ayacucho 0.051884298 0.022685711 –0.06963001 0.482599528 0.0710261860 Cajamarca 0.067522485 0.055464513 –1.55524539 0.174329961 0.0256568677 Callao 0.017204102 0.099564800 –13.4723626 0.000001409 0.0000002074 Cusco 0.040015598 0.175447100 –18.7727612 0.000000007 0.0000000010 Huancavelica 0.058499017 0.033530093 –0.36082433 0.410760034 0.0604532679 Huánuco 0.056048505 0.002795861 2.729567473 0.938748972 0.1381596027 Ica 0.019853106 0.022566488 –4.50849622 0.010895003 0.0016034631 Junín 0.035968465 0.008561616 –0.70511493 0.330679158 0.0486674313 La Libertad 0.037740897 0.069659087 –7.2793414 0.000689164 0.0001014272 Lambayeque 0.030996653 0.003696581 –0.85306615 0.298790059 0.0439741795 Loreto 0.052909655 0.039189025 –1.76944974 0.145610773 0.0214301449 Madre de Dios 0.029504551 0.004794557 –1.1830329 0.234507311 0.0345134194 Moquegua 0.017877687 0.034954281 –6.16597463 0.002095272 0.0003083699 Pasco 0.048988337 0.012923625 0.617631494 0.649679677 0.0956160686 Piura 0.043588613 0.077559486 –7.34849515 0.000643146 0.0000946545 Puno 0.056225213 0.020345404 0.794994475 0.688902734 0.1013886895 San Martin 0.046603108 0.021256936 –0.64418849 0.344300339 0.0506721174 Tacna 0.018703863 0.029998654 –5.49792138 0.004078572 0.0006002605 Tumbes 0.023285593 0.030294889 –4.89415379 0.007434558 0.0010941749 Ucayali 0.038718141 0.016920108 –1.25601978 0.221659827 0.0326226014 Lima 0.029304927 0.050761228 –6.34224604 0.001757251 0.0002586219 Metropolitan Lima 0.018414571 0.005967413 –2.85541592 0.054402035 0.0080065745 (special case) Total 1.000000000 1.000000000 –96.10778000 6.794670465 1.000000000 Sources: Ministry of Economy and Finance and World Bank staff calculations. 3 Appendix Methodology of the New FONCOR Index The new FONCOR index keeps the relative poverty average, the fiscal capacity gap is zero and criterion in the old FONCOR index but changes the regional government is not eligible for this the relative transfer criterion by adjusting its component of the FONCOR index. computation and its content, and thus this criterion • If the per capita fiscal capacity of the regional is renamed the “Fiscal Capacity” criterion government is less than the national average, the fiscal capacity gap is: The “Fiscal Capacity” criterion is now defined as: fcg i = (fc − cfi ) × POB if fc > fci (Disposable Income ) fci = POB Where: Where: • fc: per capita fiscal capacity – national average • fci: per capita fiscal capacity of regional • fci: per capita fiscal capacity of regional government i government i • Disposable income: “revenue sharing” (or • POB: regional population reformed ordinary resources) plus transfers to regional governments by concept of Note: A variation to this rule is to use the maximum canon, sobrecanon, mining royalties, income tax capacity instead of using the average as the from customs, and FOCAM in the year of base; this criterion would only keep ineligible the the calculation plus the balance of “certain richest region in per capita terms. resources” (recursos determinados) (does not include FONCOR) to December 31 of Subsequently, the fiscal capacity of each regional the year prior to the calculation. Note that government as gap index is defined: since regional governments do not have tax fcg i sources assigned to them, none are included FCG = ∑ 26 here. However, if new taxes were assigned to i =1 fcg i regional governments, the potential revenue from those taxes should also be included in The final new FONCOR index is: this definition of disposable income. IDFi = (0 .5 × P ) + (0 .5 × FCG ) • POB: regional population • Then, it is calculated the national average of Where: P: relative poverty index the Fiscal Capacity per capita and the Fiscal Capacity gap per capita relative to this national FCG: relative fiscal capacity gap index average as follows: • If the per capita fiscal capacity of the regional government is higher than the national 71 4 Appendix The Municipal Compensation Fund (FONCOMUN) and the Proposed Reform Definition compensatory allocation within the FONCOMUN to the municipalities located in rural areas. The Municipal Compensation Fund (FONCOMUN) is established in the Peruvian Constitution with • Article 4 of the Law No. 29332, that creates the Plan of incentives to improve municipal the objective of promoting investment in the management, sets that the distribution criteria municipalities of the country, with a redistributive of FONCOMUN are applied progressively from criterion in favor of the most remote and deprived the budget for Fiscal year 2010. areas, giving priority in the allocation to rural and marginal urban areas of the country. • The Supreme Decree No. 060-2010-EF approved the criteria, procedures and methodology for the distribution of the FONCOMUN, setting Legal Background the allocation for each provincial and district • Paragraph 5 of article 196 of the Peruvian municipality according to the criteria described in the Supreme Decree No. 156-2004-EF. Constitution states that the resources allocated by the Municipal Compensation Fund (FONCOMUN) concept are revenues of municipalities. Funding Sources • Article 87 of the law of Municipal taxation, Article 86 of the Legislative Decree No. 776 (the Law approved by Supreme Decree 156-2004-EF, of Municipal Taxation) (amended by the article 31 stipulates that the FONCOMUN is distributed of the Legislative Decree No. 952), determines the under the criteria of equity and compensation, source of the funds that make the FONCOMUN and that it aims to ensure the functioning of all (see Table 16). municipalities. Table 16: Financing sources of FONCOMUN • Article 88 of the law of Municipal taxation establishes that the monthly resources   Percentage municipalities perceive by the concept of the a. Municipal promotion tax (IPM) 95.6% FONCOMUN may not be lower than the amount equivalent to eight tax units (UIT) in force at the Tax on vehicles that use gasoline b.  4.3% date of approval of the budget law of the Public (impuesto al rodaje) Sector of each year. c. Tax on recreational crafts 0.1% • Article 146 of the Law No. 27972, the Organic Municipalities Law, establishes a priority and 73 74    Peru: Building a More Efficient and Equitable Fiscal Decentralization System Distribution Procedure obtained in the second phase are adjusted so that no district is transferred monthly less than eight The procedure for the distribution of the Fund tax units (UIT) by concept of FONCOMUN. It is also comprises three (3) phases:53 guaranteed that no transfer will be less than the amount transferred by the concept of FONCOMUN First phase: Geographical allocation to the provinces. in 2009. The total national FONCOMUN “pool of funds” is divided into 196 parts, which are the 196 geographical provinces that make up the country. Existing and Proposed Distribution Criteria Second phase: Inter-district allocation. The amount assigned to each geographical province in the first Essentially, the existing distribution criteria are phase, is distributed among all the districts that kept the same except that we propose to change make up each of the 196 provinces across the the methodology in the first phase in order to country. include the fiscal capacity of the provinces as an additional criterion in this phase. Third phase: The setting of eight UIT and 2009 district allowance as a “minimum amount.” The amounts Figure 27: New methodology to distribute FONCOMUN FIRST PHASE SECOND PHASE THIRD PHASE Calculate the amount by province District distribution Adjustment of the amount Provincial distribution criteria District distribution criteria Restrictions (minimum amount) Population • 8 UITs Provincial (# habitants) 20% • FONCOMUN 2009 municipality Amount by province Unmet needs (% of people who 1. Population in poverty: 85% do not have water, • (Unmet basic needs—for Amount per district electricity, and sewage) Metropolitan Lima and Callao) • (Rurality—for the rest of the country) Fiscal capacity 80% (transfers and 2. Municipal management: 10% predicted own • Execution of investment revenues) • Efficiency in tax collection 53  These three phases are described in the law of Municipal taxation (approved by Supreme Decree No. 156-2004-EF). Article 87 states: “. . . The procedure of distribution of the Fund comprises, first, a geographic allocation by province and, on this basis, a distribution among all district municipalities of the province. . . .” Article 88 mentions: “. . . Resources received in the municipalities by concept of Municipal compensation fund may not be less than the amount equivalent to eight (8) tax units . . .” The Municipal Compensation Fund (FONCOMUN) and the Proposed Reform    75 First phase: Allocation by province (b.1)  Transfers received during the year analysis The allocation by province currently uses an index We add all components of the item “transfers” combining population and the index of unmet (except for FONCOMUN) received by the province. needs for public services. n Thus the original index is calculated as: TRj = ∑ TR i =1 i (a)  Geographic index for provinces (IGPj) Where: POPULATIONj * ICj TRi = CANON _ SOBRECANON + ROYALTIES + CD + OTHER IGPj = 196 TR i = CANON _ SOBRECANON + ROYALTIES + CD + OTHERS ∑ i (POPULATIONj * ICj ) Where: TRj: total transfers received by province j Where: CANON_SOBRECANON: Canon Forestal, Canon IGPj: geographic index for province j Gasífero, Canon Hidroenergético, Canon Minero, POPULATIONj: population in province j Canon Pesquero, Canon and Sobrecanon Petrolero ICj: unmet needs index of province j of district i ROYALTIES: mining royalties of district i The unmet needs index is defined as: CD: custom duties of district i OTHERS: Fideicomiso Regional, FOCAM, FONIE,  NOWATERj + NOSEWAGEj + NOELECTRICITYj  ICj =  participants and balance of transfers of district i  3   n: number of districts that are part of province j Where: (b.2) Estimating potential own-revenues ICj: unmet needs index for province j NOWATERj: percentage of people that have no access We first calculate actual own-revenues54 of each to water in province j district, then we estimate the per capita own- NOSEWAGEj: percentage of people that have no revenues as follows: access to sewage in province j NOELECTRICITYj: percentage of people that have no OWN _REVi IPpci = electricity in province j POPULATIONi Proposed scenario Where: We propose to add an index of fiscal capacity to IPpci = own-revenues per capita of district i the criteria used in the first phase. OWN_REVi = own-revenues of district i POPULATIONi = population of district i (b)  Fiscal capacity index (ICFj) Then in order to estimate potential own-revenues Fiscal capacity is measured by all the transfer per capita, we run a regression for all districts having funds received by the province, including canon, but as a dependent variable the own-revenues per excluding the FONCOMUN itself plus potential own capita calculated and as an independent variable revenues—which are calculated using regression analysis. These components of fiscal capacity are 54 For the calculation of own-revenues it was considered further explained in what follows. the revenues coming from the “Municipal Tax” and “resources directly collected.” From both sources of financing, we do not take into account the generic entries “Indebtedness,” “Balance,” nor “non-Financial Assets Sale.” Other sources not considered: “tax on casinos” “taxes on slot machines,” transfers by canon, sobrecanon, concession rights, FOCAM royalties, custom duties, and FONCOMUN. 76    Peru: Building a More Efficient and Equitable Fiscal Decentralization System the average household private expenses in each FONCOMUN) plus potential own-revenues district expressed in per capita terms. We run this (calculated as predicted values based on regression regression using the information for 2013, the year analysis): for which we have the most updated information. FC j = TR j + IPj 7 Subsequently, we used the regression parameter to predict the values of the district own-revenues Where: per capita for the year of the analysis. In the case TRj: total transfers received by province j IPJ: potential own-revenues for province j 7 that the predicted value is negative, the district is assigned a value of zero. FCj: fiscal capacity of province j FC j 7 IPpci = 0 .5625352 × Gtoavgpci − 129 .5626 FCpc j = POPULATION j 7 IPpci = 0 if IPpci < 0 7 Where: Where: IPpci = predicted value of own-revenues per capita of FCpcj: fiscal capacity per capita of province j 7 district i POPULATIONj: population of province j Gtoavgpci = average household private expenses in per capita terms of district i (b.3)  Calculating the fiscal capacity gap We multiply the predicted value of own-revenues In this step we first calculate the national average per capita calculated by the district population for fiscal capacity per capita and then we calculate the year of analysis the fiscal capacity gap per capita for each province relative to this national average as follows: IPi = IPpci × POPULATIONi 7 7 • If the per capita fiscal capacity of the province Where: is higher than the national average, the fiscal IPi = predicted value of own-revenues of district i 7 capacity gap is zero and the province is not POPULATIONi = population of district i for the years eligible for this component of the index. of analysis • If the per capita fiscal capacity of the province In order to arrive at the potential own-revenues is less than the national average, the fiscal at the province level, we sum the predicted value capacity gap is: of own-revenues for all districts including the fcg j = (FCpc − FCpc j ) × POPULATION j (if FCpc > FCpc j ) provincial municipality, n fcg j = (FCpc − FCpc j ) × POPULATION j (if FCpc > FCpc j ) ∑ IP 7 7 IPj = i i =1 Where: fcgj: fiscal capacity gap of province j Where: FCpc: fiscal capacity per capita – national average 7 7 IPi: predicted value of own-revenues of district i FCpcj: fiscal capacity per capita of province j IPj: predicted value of own-revenues of province j POPULATIONj: population of province j n: number of districts that are part of province j (it includes the provincial municipality in each province) The fiscal capacity index of each province is defined as its share in the total fiscal capacity gap of all The fiscal capacity of the province is thus defined qualifying provinces: as the revenues coming from all transfers (except fcg j FCG j = 196 ∑ (fcg ) 1 j The Municipal Compensation Fund (FONCOMUN) and the Proposed Reform    77 Where: The relative weights for the three criteria are shown FCGj: fiscal capcity index of province j in Table 17: (c) The information on the fiscal capacity index is Table 17: FONCOMUN: Distribution criteria combined with the original provincial allocation (weights) index (IGPj) to estimate the proposed provincial allocation index. Indicator Weights Rurality 0.85 We give the original index a weight of 70 percent and the added fiscal capacity index a weight of Territorial extension 0.05 30 percent. Both criteria are pro-poor and other Municipal management 0.10 combinations of weights could be simulated to Total 1.00 arrive to more (or less) redistributive outcomes: Source: Ministry of Economy and Finance. API j = 0 .7 × IGPj + 0 .3 × FCG j (a)  Rurality index (IRj) Where: APIj: adjusted allocation index for province j Article 146 of the Organic Law of Municipalities, IGPj: geographic index for province j approved by Law No. 27972, states that rural areas FCGj: fiscal capacity index for province j have priority for receiving compensatory allocation of FONCOMUN resources. The index is calculated Second phase: The allocation across as a weighted summation of the rural and urban population of the district where a double weight districts within each province is provided to the rural population. The index is After determining the allocation by province, applied as follows: the next step is the distribution between the district and provincial municipalities in each of   the 196 provinces, taking into consideration the  1 * URBANi + 2 * RURALi  IRi =  n  following:  [1 * URBANi + 2 * RURALi ]  ∑  i =1   20 percent of the first step allocation is assigned to the provincial municipality (of each province). Where: IRi: rurality index of district i 80 percent of the geographical allocation by URBANi: urban population in district i province is distributed among all the district RURALi: rural population in district i municipalities of the province, taking into account n = number of districts that are part of the province three criteria: (i) the rurality (prioritization where district i is located in districts with higher rural population); (ii) the territorial extension; and (iii) municipal In the case of the districts located in Metropolitan management (the generation of income and the Lima and Callao province, the unmet needs prioritization of spending in investment). 78    Peru: Building a More Efficient and Equitable Fiscal Decentralization System indicator (NBI) is used in substitution of the rurality Where: index. The NBI’s indicators taken into account are: TERRITORYi: territory index of district i KM2: territorial extension of district i (in square • Population kilometer—KM2). • Households with inadequate housing (NBI1) n: number of districts that are part of the province where district i is located • Households without access to water service. (NBI2) (c)  Municipal Management (IGM) • Households with no access to sewage service. The municipal management indicators incor­ (NBI3) porated into the interdistrict allocation are based • Households with children between 6 to 12 years on the Article 32 of Legislative Decree No. 952 old who do not attend school. (NBI4) that modifies the Legislative Decree No. 776— • Households with very low economic capacity (NBI5) Municipal Taxation Law. Therefore the index applicable to the districts of The municipal management index is calculated on the provinces of Lima and Callao is: the basis of two criteria: (i) the generation of own- revenues; and (ii) the prioritization of spending in   investment.56  POPULATIONi * NBIi  IPOi =  n   [POPULATIONi * NBIi  • The own-revenues index 57 (IPI) is defined as:  ∑  i =1   0 .1   ING _PROPt − 1   NBIi = NBI 1i + NBI2i + NBI3i + NBI 4i + NBI5i    POPULATIONt − 1   IPi = 1 +     1 + ING PROPt −2   Where:    IPOi: poverty index for district i   POPULATIONt −2     POPULATIONi: population of district i NBIi: unmet needs index for district i Where: n: number of districts that are part of the province IPi: own-revenues index of district i where district i is located ING_PROPi: own-revenues of district i POPULATIONi: population of district i (b)  Territorial extension index (TERRITORY) t: year for which the index is calculated Since 2010, the FONCOMUN index incorporates the territorial extension variable.55 This variable aims to compensate districts that due to their territorial extension have difficulties providing basic services to their more remote populations. The index is calculated as: 56 In the case of the municipalities whose value of the indicator is zero (0) for the years of assessment, they are   assigned 50 percent of the minimum value of the indicator of  KM2i  municipal management district registered nationwide. TERRITORYi =  n  57 For the calculation of own-revenues it was considered  [KM2i ]  ∑  i =1   the revenues coming from the “Municipal Tax” and “resources directly collected.” From both sources of financing, we do not take into account the generic entries “Indebtedness,” “Balance,” 55  The Legislative Decree No. 952 (February 2004) modifies nor “non-Financial Assets Sale.” Other sources not considered: the Legislative Decree No. 776—Municipal Taxation Law. Its “tax on casinos” “taxes on slot machines,” transfers by canon, Article 32 sets that within the new distribution criteria for sobrecanon, concession rights, FOCAM royalties, custom FONCOMUN, the territorial extension should be included. duties, and FONCOMUN. The Municipal Compensation Fund (FONCOMUN) and the Proposed Reform    79 The expenditures prioritization index (IPGIi) Where: considers the acquisition of nonfinancial assets58 IND_DISTRi: allocation index for district i (ADQ_ACT_NO_FINAN) and total spending59 IRi: rurality index of district i (GASTO_TOTAL). The index is defined as: TERRITORYi: territorial extension index for district i IGMi: municipal management index for district i  ADQ _ ACT _NO _FINANFONCOMUNi  IPGIi =  1 +  In the case of Metropolitan Lima and Callao, the  GASTO _TOTALFONCOMUNi  allocation is: Where: IND _DISTRi = [IPOi ] * 0 .85 + [TERRITORYi ] * 0 .05 + [IGMi ] * 0 .1 IPGIi: expenditures prioritization index of district i IND _DISTRi = [IPOi ] * 0 .85 + [TERRITORYi ] * 0 .05 + [IGMi ] * 0 .10 ADQ_ACT_NO_FINANFONCOMUNi: acquisition of nonfinancial assets financed with FONCOMUN of Where: district i IND_DISTRi: allocation index for district i GASTO_TOTALFONCOMUNi: total expenditure financed IPOi: poverty index based on unmet needs of district i with FONCOMUN of district i TERRITORYi: territorial extension index for district i FONCOMUN: expenditures financed with “FONCOMUN” IGMi: municipal management index for district i Therefore, the municipal management index for district i is set by the following formula:60 Third phase: District allocation adjustment for 8 UIT and 2009     FONCOMUN allocation as “floor value”  IPi   IPGIi   n + n  The district allocation obtained in the second      ∑ IPi   ∑ IPGIi  phase should be corrected so that no district IGMi =  i = 1   i = 1  receives less than 8 tax units (UIT) per month. To 2 do this, the municipalities that have a preliminary monthly allowance higher than 8 UIT, would yield Where: their allocation marginally (prorated) to ensure this IGMi: municipal management index for district i minimum level FONCOMUN. This legal “minimum” is IPi: own-revenues index for district i established by the Article 33 of Legislative Decree IPGI: prioritization of expenditures index for district i No. 952 which modifies the Legislative Decree n: number of districts that are part of the province No. 776—Municipal Taxation Law. where district i is located The formula used for the adjustment is: Based on the previously developed indexes, the allocation index for district i is set by the following n m formula: ∑ Surplus + ∑ Deficit i i RA8UIT = i =1 n i =1 IND _DISTRi = [IRi ] * 0 .85 + [TERRITORYi ] * 0 .05 + [IGMi ] * 0 .10 ∑ Surplus i = [IRi ] * 0 .85 + [TERRITORYi ] * 0 .05 + [IGMi ] * 0 .10 i =1 Where: RA: adjustment ratio for 8 tax units. Surplusi: positive difference between the preliminary 58  It is considered the expenditures under the generic “Acquisition of non-financial assets” financed with FONCOMUN monthly allocation of FONCOMUN and 8 tax units in and are intended for investment projects. district i 59 It consists of total spending financed with FONCOMUN. Deficiti: negative difference between the preliminary 60 In the case of the own revenues index (IP), before the data monthly allocation of FONCOMUN and 8 tax units in is converted into an index, a previous numeric transformation district i is done through which these numbers are converted to a scale n: number of districts in surplus between 1 and 2. m: number of districts in deficit 80    Peru: Building a More Efficient and Equitable Fiscal Decentralization System With the adjustment ratio, we calculate the deficit will be given by the difference between the adjusted monthly allocation of FONCOMUN to be adjusted monthly allocation of FONCOMUN to no less than 8 UIT as follows: 8 tax units and the monthly amount of FONCOMUN transferred in 2009. (a) if the district is in surplus: FONCOMUNi = 8UIT + RA8UIT * Surplusi Fluctuations in the “pool of funds” used in the allocation (b) if the district is in deficit: FONCOMUNi = 8UIT The resources distributed by FONCOMUN are mainly determined by the collection of the Where: Municipal Promotion Tax (IPM), which is closely FONCOMUNi: adjusted monthly allocation of linked to the performance of the general sales tax FONCOMUN in district i (VAT). Article 76 of the Legislative Decree No 776 UIT: Tax units sets that the IPM is levied at a rate of 2 percent over the operations affected by the VAT. In addition of making the adjustment of 8 tax units, the methodology also sets a “floor amount,” Therefore, the monthly allocation of FONCOMUN meaning that it guarantees that no transfer is is directly related to the performance of the tax less than the amount of FONCOMUN transferred revenue of the previous month: when the revenue in 2009. The methodology for the calculation is increases, the allocation of the following month similar to the one established for the adjustment increases and when the levy is reduced, the to 8 tax units, the difference is that in order to municipalities receive a lower allocation. obtain the adjustment ratio (RA), the surplus and References Cheasty, A. and Pichihua, J. 2015. “Fiscal Martinez-Vazquez, J. 2013. “Fiscal Decentralization decentralization: Progress and challenges for the in Peru: A Perspective on recent Developments future,” Chapter 10 in: Santos, A. and Werner, A. and Future Challenges,” Working Paper 13–24, (eds), Peru: Staying the Course of Economic Success, International Center for Public Policy, Andrew International Monetary Fund, Washington, D.C. Young School of Public Studies, Georgia State University, November. Erman, Alvina. 2015. ‘The state of the canon minero’ [PowerPoint slides]. World Bank. OECD. 2016. OECD Territorial Reviews: Peru 2016, OECD Territorial Reviews, OECD Publishing, Lago-Peñas, Santiago, and Martinez-Vazquez, Paris, https://doi.org/10.1787/9789264262904- Jorge. 2013. The Challenge of Local Government en. Size Theoretical Perspectives, International Experience and Policy Reform. World Bank. 2010. “Peru: The Decentralization Process and Its Links with Public Expenditure Loayza, Norman V.; Rigolini, Jamele; and Calvo- Efficiency.” Report No. 52885-PE, World Bank, Gonzalez, Oscar. 2011. More than you can handle: Washington, DC. decentralization and spending ability of Peruvian municipalities. Policy Research working paper; no. WPS 5763. Washington, DC: World Bank. https://hubs.worldbank.org/docs/ImageBank/ Pages/DocProfile.aspx?nodeid=14826727 81