Report No. 52885-PE PERU The Decentralization Process and Its Links with Public Expenditure Efficiency June 7, 2010 Poverty Reduction and Economic Management Bolivia, Chile, Ecuador, Peru, and Venezuela Country Management Unit Latin America and the Caribbean Region Document of the World Bank REPUBLIC OF PERU FISCAL YEAR January 1 to December 31 CURRENCY EQUIVALENTS Currency Unit = Soles 1 US Dollar = S/. 2.84 WEIGHTS AND MEASURES Metric System ABBREVIATIONS AND ACRONYMS LBD Decentralization Law (Ley de Bases de la AAA Analytical and Advisory Activities Descentralización) ANGR National Assembly of Regional LEMU Local Education Management Unit Governments (Asamblea Nacional de (UGEL, Unidad de Gestión Educativa Gobiernos Regionales) Local) AMPE LGs Local Governments APRA American Revolutionary Popular Alliance LOGR Organic Law of Regional Governments (Alianza Popular Revolucionaria (Ley Orgánica de Gobiernos Regionales) Americana) LOM Organic Law of Municipalities (Ley CND National Decentralization Council Orgánica de Municipalidades) (Consejo Nacional de Descentralización) LOPE Organic Law of the Executive Power (Ley CTAR Transitory Council for Regional Orgánica del Poder Ejecutivo) Administration (Consejo Transitorio de MEF Ministry of Economy and Finance Administración Territorial) OLS Ordinary Least Squares DEA Data Envelope Analysis PCM Office of the Prime Minister (Presidencia DNPP National Budget Directorate (Dirección del Consejo de Ministros) Nacional de Presupuesto Público) PNUD Programa de Desarrollo de las Naciones ENAHO National Household Survey (Encuesta Unidas Nacional de Hogares) PSE Public Sector Efficiency FONCOMUN Municipal Compensation Fund (Fondo de PSP Public Sector Performance Compensación Municipal) RBB Results-Based Budgeting FONCOR Regional Compensation Fund (Fondo de REMURPE Rural Municipalities Network (Red de Compensación Regional) Municipalidades Rurales del Perú) GDP Gross Domestic Product REO Regional Education Office (DRE, GOP Government of Peru Direcciones Regionales de Educación) GNP Gross National Product RG Regional Government IDB Inter-American Development Bank SD Secretariat of Decentralization IFC International Finance Corporation SIAF Public financial system (Sistema Integrado IGV General sales tax (Impuesto General a las de Administración Financiera) Ventas) SIAF-GL Public financial system for local IMF International Monetary Fund governments (Sistema Integrado de INEI National Statistics Institute (Instituto Administración Financiera-Gobiernos Nacional de Estadística e Informática) locales) KfW German development bank, Reconstruction SNIP National public investment system Credit Institute (Kreditanstalt für (Sistema Nacional de Inversión Pública) Wiederaufbau) UNDP United Nations Development Program LAC Latin America and the Caribbean USAID US Agency for International Development Vice President, LCR: Pamela Cox Country Director, LCC6C: C. Felipe Jaramillo Sector Director, LCSPR: Marcelo Giugale Sector Manager, LCSPE: Rodrigo Chaves Sector Leader, LCSPR: Carlos Silva-Jauregui Task Team Leader: Oscar Calvo-Gonzalez ACKNOWLEDGMENTS This report was prepared by a team led by Oscar Calvo-Gonzalez (LCSPE) under the overall supervision and guidance of Carlos Silva-Jauregui (Lead Economist and Sector Leader, LCSPR), Rodrigo Chaves (Sector Manager, LCSPE), and C. Felipe Jaramillo (Country Director, LCC6C). The team also included Apostolos Apostolou (LCSPE), Ana Maria Aviles (LCSPE), Carlos Casas (Consultant, Universidad del Pacifico), Patricia Chacon Holt (LCSPE), Barbara Cunha (LCSPE), Kent Eaton (Consultant, University of California Santa Cruz), Javier Illescas (LCSPE), Kai Kaiser (Senior Economist, PRMPS), Milagro Mayuri (LCC6C), Diana Ortiz (LCSPE), and Emmy Yokoyama (LCC6C). Comments and data were also received from Jonas Frank (Senior Public Sector Management Specialist, LCSPS), Victor Grande (IFC), Nelson Gutierrez (LCSHS), Anna Okola (LCSTR), Claudia Sanchez Lanning (LCSHE), Fernando Rojas (Lead Public Management Specialist, LCSPS), Geeta Sethi (Lead Economist, LCSPE) and Steve Webb (Consultant). Jorge Martinez-Vazquez (Georgia State University) kindly shared insights from an ongoing technical assistance program on decentralization in Peru. The report was prepared in parallel to complementary studies on decentralization by the HD and SD sector units and the team benefited from joint missions and interactions with the teams led by Ines Kudo and Willy Reuben, as well as from the overall guidance of Livia Benavides in the CMU. The peer reviewers for the report were Wolfgang Fengler (Lead Economist, AFTP2), Lili Liu (Lead Economist, PRMED), and Martin Valdivia (GRADE). The team received valuable insights through meetings with a variety of individuals, including officials of the Peruvian Government as well as representatives of civil society organizations. The team would like to thank the Peruvian authorities for their continued cooperation, especially the support of the Ministry of Economy and Finance (MEF), the Secretariat of Decentralization, the Defensoría del Pueblo, and the Regional Government of Cajamarca for their hospitality during a field visit. Our appreciation also goes to the IMF, KfW, and USAID. 4 The Decentralization Process and Its Links with Public Expenditure Efficiency in Peru Table of contents Executive summary............................................................................................................. 7 1. Introduction .............................................................................................................. 15 A. The goals and principles underlying the decentralization process ...................... 15 B. How the process unfolded ................................................................................... 20 C. Explaining the route travelled so far: the role of incentives ............................... 25 2. Assignment of responsibilities for service delivery ................................................. 33 A. Assignment of functions ..................................................................................... 33 B. Evolution of the transfer process ......................................................................... 40 C. Coordination among levels of government ......................................................... 42 3. Expenditure patterns ................................................................................................ 45 A. Spending by different levels of government ....................................................... 45 B. Deconcentration vs. decentralization of expenditures......................................... 50 C. Public spending and regional disparities ............................................................. 53 4. Assignment of revenue ............................................................................................ 59 5. Efficiency of public expenditure across departments .............................................. 67 A. Introduction ......................................................................................................... 67 B. Analysis ............................................................................................................... 69 C. Results ................................................................................................................. 73 D. Conclusions ......................................................................................................... 83 List of references............................................................................................................... 84 Annex A. Basic indicators by department ................................................................... 89 Annex B. Transfer of functions in health and social programs ................................... 93 Annex C. Public expenditure per capita by level of government and by function across departments .......................................................................................................... 103 Annex D. Efficiency analysis .................................................................................... 107 5 List of boxes Box 1: The institutional framework for decentralization since 2002 ............................... 16 Box 2: Levels of government ............................................................................................ 19 Box 3: A brief history of de- and re-centralization in Peru .............................................. 23 List of tables Table 2.1: Allocation of responsibilities by level of government ..................................... 34 Table 2.2: Comparison of functions assigned to regional and local governments ........... 35 Table 2.3: Allocation of education responsibilities by level of government .................... 37 Table 3.1: Public spending by level of government, 2004-2008 ...................................... 46 Table 3.2: Public spending by function and level of government, 2004 and 2008........... 49 Table 3.3: Deconcentrated and decentralized spending, 2004 and 2008 .......................... 51 Table 3.4: Public spending by level of government and by function ............................... 55 Table 3.5: Comparison of public spending by level of government and by function, Apurimac and Huancavelica ............................................................................................. 56 Table 4.1: Revenue sources of sub-national governments................................................ 59 Table 4.2: Local tax revenues ........................................................................................... 61 Table 4.3: Rules by which canon is distributed ................................................................ 63 List of figures Figure 1.1: Results from the referendum for the formation of macro-regions in 2005 .... 30 Figure 2.1: Regional government organization chart ....................................................... 43 Figure 3.1: The importance of public investment in different levels of government ....... 47 Figure 3.2: Disbursement rates for public investment by level of government ................ 49 Figure 3.3: Breakdown of regional governments’ deconcentrated and decentralized spending over time ............................................................................................................ 50 Figure 3.4: Deconcentrated and decentralized spending by regional government ........... 52 Figure 3.5: Public expenditure per capita and GDP per capita across departments ......... 53 Figure 3.6: Total public spending per capita by department and level of government .... 54 Figure 3.7: Local public expenditures by function ........................................................... 57 Figure 3.8: Local public capital and current expenditures and main revenue sources ..... 57 Figure 4.1: Sources of revenue of sub-national governments........................................... 60 Figure 4.2: Local public spending by revenue source ...................................................... 62 Figure 4.3: Earmarked sources of revenue for sub-national governments ....................... 63 Figure 4.4: Public spending per capita by department and by revenue source ................. 64 Figure 4.5: Share of resources by regional and local governments accounted for by canon and other natural resource-related revenues ..................................................................... 65 6 Executive summary ES. 1. Decentralization in Peru is an ongoing process that is in its infancy and faces key challenges. This study aims to take stock of the process and identify these challenges. The main goal of the report is to help think through these issues with a view of identifying priorities and opportunities for fruitful engagement in this area. Much of the stock-taking serves as an introduction to the topic for the general reader and draws on what has become by now extensive secondary literature. The report distils the key challenges of the decentralization process and adds value in three specific areas, as detailed below. ES. 2. A first distinct contribution of this report is to highlight the importance of political economy factors in explaining the ebbs and flows of the decentralization process. Electoral, institutional, and bureaucratic incentives cannot be ignored if we are to understand the path that the decentralization process has taken so far. A clear example – discussed further below – is provided by the failed referendum on the creation of macro-regions in 2005 (the last of many previous attempts at creating a middle level of government in Peru). The decision by legislators to hold the referendum on the issue after having established regional governments in all departments created strong incentives against the amalgamation into a smaller number of regional governments. Unsurprisingly, the individuals who were elected as regional presidents and as regional assembly members in 2002 faced strong incentives to campaign against the macro-regions since the possible fusion of regional governments could have resulted in the elimination of their own positions. All regional integration proposals were rejected despite the fact that fiscal incentives were on the cards for amalgamated regions, which would have received 50 percent of all tax revenues collected in the macro-region instead of the current system of transfers. This episode provides a cautionary tale of how institutional incentives can trump even relatively large fiscal incentives. ES. 3. Second, the report provides new data on truly decentralized spending – as opposed to deconcentrated – as well as on public spending per capita across districts. The distinction between decentralized spending (where the sub-national government has discretion over how the money is spent) and deconcentrated spending (where the sub-national government has no autonomy and does not control how the money is spent) is an important one in the literature. In this report we provide a first systematic approach at differentiating between these two types of sub-national spending in Peru. Finally, the analysis on a per capita basis of local public spending is now possible due to the combination of a number of new data sources that have become available. The integrated public financial system was rolled out to all local governments in 2008, and therefore now includes spending by each of the three levels of government. In addition, the new population census of 2007 provides the basis for all per capita calculations. To the best of our knowledge, this is the first report that combines these data sources to provide a picture of how much different localities spend on a per capita basis. ES. 4. Third, this report provides a novel analysis of the efficiency of public spending across departments in Peru. The report compiles measures of public sector 7 performance and efficiency for education, health, and transport across all departments. The analysis draws mainly on the newly available results indicators collected as part of the Results-Based Budgeting while the public expenditure data is from the integrated public financial system (SIAF - Sistema Integrado de Administración Financiera). The combination of these data sources allows us to compute how the expenditure of the three levels of government in a given department is correlated or not with results on the ground. Challenge #1: An institutional design overwhelmed by messy realities ES. 5. A flurry of legislation in 2002-2004 delineated a detailed path towards decentralization. Between 2002 and 2004 a total of 11 major laws were passed by Congress laying out the institutional framework for decentralization. The laws passed included a constitutional reform making decentralization a “permanent policy of the state�; a framework law for decentralization; organic laws for regional governments and municipalities; and as a fiscal decentralization law. The framework was cautiously delineated to reflect a number of lessons learned from previous episodes of decentralization in Latin America. In particular, the principle of fiscal neutrality was paramount since it requires that increases in expenditures by sub-national governments need to be compensated for by decreases at the central level. Fiscal rules were put in place to limit the ability of sub-national governments to borrow and thus prevent unsustainable debt positions that could eventually lead to bail outs by the central government. ES. 6. The gradualist approach envisaged fiscal decentralization to take place once macro regions had been formed. Gradualism was another key principle of the decentralization framework. Political decentralization, i.e., the holding of regional elections, took place first. A regional government was set up in every department in Peru, although departments are expected to be merged into fewer macro-regions via referenda at a later date. Administrative decentralization, by which we mean the assignment of functions over which the sub-national governments would have autonomy, was envisaged to take place over a number of years depending on the capacity of sub-national governments to take on responsibilities. To that end, an accreditation system was set up and annual transfer plans were put in place to govern the transfer of responsibilities. Fiscal decentralization, and specifically the assignment of resources to sub-national governments to fulfill their new responsibilities, was also to take place in a phased approach. In a first stage the regional governments would receive the transfers that “correspond� to the competencies already transferred in strict compliance with the fiscal responsibility framework. In a second stage, regional governments would receive half of the national taxes (sales tax, corporate and income tax) effectively collected in their territory. This second stage, however, was made contingent on the formation of macro- regions through the amalgamation of one-department regional governments. ES. 7. The process was thrown off track rather quickly as the referendum for the formation of regions failed. In 2005 a referendum on the formation of macro- regions was held but all of the proposed regions were rejected by voters. Sequencing is key to explaining these results: as regional authorities had already been elected in late 8 2002 in each department they had little to gain by merging and a lot to lose –they would risk losing their positions as a result of amalgamation. Given these incentives, it is unsurprising that regional authorities campaigned against the formation of macro-regions. ES. 8. The timetable for new referenda for the creation of regions set in the law has now been postponed sine die. The framework decentralization law set out that referenda for the formation of regions would take place in 2005, 2009, and 2013. After the failed referendum in 2005, however, the authorities decided not to call any referenda in 2009. Instead, a series of inter-regional coordination committees have been set up and a so-called “pilot region� has been discussed but this has not led to any reduction in the number of regional governments and runs the risk of creating one more intermediate institutional level between the current regional governments and the central government. ES. 9. The sequencing envisaged in the law was disrupted and some stages planned in the gradual approach are in limbo – notably fiscal decentralization. The referendum results posed a challenge to the process as the legal framework in place did not envisage a situation where macro-regions would not be created. Both fiscal decentralization, in the form of revenue sharing, and the transfer of functions were designed to take place after the formation of regions. As a result of the failed referendum fiscal decentralization was put on hold. However, the transfer of functions took place, reversing the steps initially envisaged. ES. 10. Accreditation of regional governments did not remedy the institutional capacity constraints of sub-national governments. The National Decentralization Council mapped the institutional capacity weaknesses of regional governments and developed a capacity building plan as early as 2004. Functions were to be transferred upon the accreditation of regional governments capable of undertaking those functions – a politically important step to strengthen support for the gradualist approach that was adopted. However, this and subsequent plans have made little difference, in part because capacity constraints are large and difficult to address, but also because many of the capacity building activities have not been undertaken. Moreover, 39 percent of the functions accredited to regional governments have been carried out on the basis of agreements whereby the parties acknowledge the insufficient levels of capacity but undertake to make efforts to improve it. This is the so-called ‘por potenciar’ modality of accreditation introduced in 2007 following the ‘decentralization shock’ announced by the central government. A final important point regarding capacity constraints is that it is common to gauge institutional capacity of sub-national governments in public discussion by their ability to spend their investment budget. This is a highly inadequate proxy for true institutional capacity as it ignores both the efficiency of spending and also the distortions introduced by the boom in rents from natural resources which have in some cases flooded the coffers of sub-national governments in natural resource-rich areas. ES. 11. Reducing regional disparities remains a key challenge. One of the stated goals of decentralization in Peru was to improve living conditions and public services across the territory and also reducing inequalities. However, there was no instrument envisaged to help address inequalities in the country. The fiscal decentralization planned in the framework could not have contributed to a reduction in inequities in the level of 9 public services that citizens receive across the territory since it merely would have transferred half of national taxes collected in a region to the regional government (a prospect that, in any case, posed significant methodological difficulties) The central government has modified the allocation of the regional compensation fund (Fondo de Compensación Regional, FONCOR) to benefit poorer departments but the pro-poor bias is only slight. The government has also set up a fund to promote regional investment (Fondo de Inversión Pública Regional, FONIPREL) but this fund is relatively small compared to the resources that are accruing to the regional governments through other sources. At the local level, the municipal compensation fund (Fondo de Compensación Municipal, FONCOMUN) is also not designed as a true compensatory fund but rather depends on collections from the sales tax in a given area. ES. 12. There is no institutionalized forum for intergovernmental consensus- building following the elimination of the National Council of Decentralization. As in many unitary countries undergoing decentralization, the inter-governmental institutional arrangements have been in flux. With all its shortcomings, the end of the National Council of Decentralization eliminated the only official governmental space that allowed for the collective expression of sub-national governments. After its elimination, relationships between the central authorities and the sub-national ones have been ad hoc and largely bilateral. However, regional presidents have responded by creating a new collective body: the National Association of Regional Governments (Asociación Nacional de Gobiernos Regionales or ANGR). Under a rotating leadership, this association has emerged as an important voice in articulating the demands for more resources. ES. 13. Despite repeated changes to the legislative framework some key issues still need to be addressed. A paradox is that in a country with an abundance of laws on decentralization, the space that is emerging for dialogue between the regional governments and the central government has no legal entity and is not envisaged in any law. The National Association of Regional Governments has now established itself as the counterpart of the President of the Republic and holds meetings every six months. This is not to say that the policy implication is to quickly pass a law formalizing the National Association of Regional Governments. The point is that attempting to fine-tune the legislative framework for decentralization is unlikely to solve the problems when there is no underlying consensus. Yet the tendency to tinker is entrenched: notice the case of the recent minor modifications to the fiscal rules for sub-national governments and the more relevant introduction of performance-based grants to sub-national governments. ES. 14. The overall institutional framework provides enough reasons for reconsideration – an area where the Bank could add value. To be clear, this report does not call for new rushed legislative initiatives to address the perceived shortcomings in the institutional set up. What is called for is the creation of spaces for inter-institutional and inclusive dialogue to help establish a consensus on some basic principles. Only then would changes to the current institutional framework be appropriate. 10 Challenge #2: The unintended consequences of the canon ES. 15. The mechanism to share natural resource revenues among levels of government has contributed to obfuscate the decentralization process. The canon and other mechanisms by which revenues from natural resources are divided between the central, regional and local governments have had three negative consequences that relate to the decentralization process and its goals: (i) they have exacerbated regional disparities, which the decentralization process aimed to reduce; (ii) they have generated vertical asymmetries between levels of government; and, (iii) they have not contributed to improving the efficiency of public expenditure, which was also a stated goal of the process. To be sure, these are not challenges faced only by Peru but are common challenges faced by most countries with abundant natural resources. ES. 16. Horizontal inequities have increased, this calls into question the ability of the decentralization process to help to reduce regional disparities. This report shows that public expenditure on a per capita basis is highest precisely in those areas of the country that are richest. For example, the citizens in Moquegua, which enjoy a per capita income more than twice the average for the country as a whole, benefit from twice as much public spending as the average for the country. In neighboring Puno, where the per capita income is half the countrywide average, its citizens benefit from only half the countrywide average for public spending per capita. The unequal sharing of natural resource revenues account for part of that result. All of this is well known and, to some extent, only loosely related to the decentralization process. What is not so well-known is that today the citizen in Moquegua still benefits from more public spending than her neighbor in Puno even after we deduct expenditures that are financed with the canon. The inequalities stemming from the uneven sharing from natural resources have been superimposed on the decentralization process which, despite having being launched with the intention of reducing regional disparities as one of its goals, has precious few mechanisms to achieve that goal. The boom in the canon obfuscates this fact as well as it complicates the chances that a political consensus could be reached over basic issues such as the extent of horizontal equity that is to be sought – an issue which was not addressed when the decentralization process was launched. ES. 17. Vertical asymmetries have increased as the canon has disproportionately benefited one level of government – the local one. Much has been said about how the rules by which the canon is distributed across the territory benefit some (the citizen in Moquegua) more than others (the citizen in Puno). But much less has been said about how the canon favors the local level of government versus the regional government at a ratio of approximately 3 to 1. Together with the VAT-based municipal compensation fund (FONCOMUN), transfers linked with natural resources to local governments have been the single most important factor affecting the landscape of intergovernmental fiscal relations in Peru since 2004. The share of local governments in total public spending has indeed increased from 12 percent to 18 percent over the five years through 2008. This increase has not been associated with any realignment of competencies that would have seen an increase in the functions of local governments. In contrast, regional governments, for all the talk about transfers of functions to that level of government, have increased their share in total public spending from 16 to 18 percent: a figure which, estimates in the 11 report suggest, would be closer to one-third of that, 6 percent. This figure excludes the spending run through the deconcentrated offices of line ministries which is formally accounted for as spending by regional governments, but over which the regional government has generally no control. Thus, the canon obfuscates the decentralization process by favoring the local level of government, when the goal had been to develop a stronger middle level in a few regional governments, and by effectively removing the link between the transfer of functions to that of resources. ES. 18. The mechanism to share natural resource revenues is also affecting the efficient use of public funds. To begin with, the canon imposes the restriction that amounts be spent on capital expenditures. While this is a positive control mechanism that has helped to keep current expenditures at bay, it has led to a rapid increase in investment at the sub-national level: in 2008 the central government undertook approximately only one-third of the total public investment. At the local level things are even more extreme as more than half of the budget of local governments was spent on investment. The boom in canon resources has created a de facto division of labor whereby the central government runs current expenditures and sub-national governments do the investment. This is unlikely to be conducive to efficiency if only because of the interrelation between capital and current expenditures that is required for the provision of most public services. In addition, the fact that revenues from natural resources are all incorporated in the annual budget and none has to be saved for the future creates a number of electoral and institutional incentives to try to spend the money as quickly as possible. There is actually a cadre of civil society organizations ready to “name and shame� those local politicians who display low disbursement rates, which are unambiguously equated to be a proxy for success. But all of this ignores the highly volatile nature of commodity prices and the fact that most of the natural resources exploited are non-renewable. The canon therefore obfuscates the decentralization process by promoting a division of labor between levels of government that may be inefficient and by promoting investment horizons focused on the short-term. ES. 19. While modifying the canon has so far proved to be politically unfeasible, there is an increasing recognition that the status quo is inadequate. Most noteworthy is the latest proposal on fiscal decentralization put forward by a congressional working group calls for “delinking the calculation of the canon as a share of the corporate income tax� (Congreso de la República 2009). While it is beyond the scope of this report to consider what may become politically feasible or not, it is important to recognize that addressing the shortcomings identified in the decentralization process so far, including the horizontal and vertical asymmetries, will likely necessitate some degree of reform to the canon and to other existing mechanisms in order to share revenues from natural resources. 12 Challenge #3: Increasing public spending efficiency ES. 20. One of the goals of the decentralization process is better public services. The goal of improving public sector performance through decentralization is not only laid out in the Peruvian legal framework but is also often emphasized by the many civil society organizations that have been created to monitor and opine on the decentralization process. So far, the decentralization process has not been able to produce the expected results. Based on the national household surveys of 2004-2005, 70 percent of respondents thought that decentralization had done nothing or little to improve public services. Drawing on newly compiled data on result indicators and gathered as part of the Results- Based Budgeting initiative of the Ministry of Economy and Finance, an analysis of the extent to which public spending contributes to better outcomes across departments in Peru helps to shed some light on those issues. ES. 21. Our analysis points to the importance of underlying economic and geographic factors – such as income levels and population density – in explaining public sector performance. As expected, higher levels of per capita income tend to be associated with larger amounts of public spending and better government outcomes (though not necessarily with higher efficiency, an important issue to which we will return below). Population density is positively correlated with government efficiency. This result is consistent with the findings of the World Development Report 2009 which highlights how densities and agglomeration effects help to reduce production costs and increase productivity. Of course, factors like population density can only slowly be influenced by policy – if at all - and are instructive about the challenges that increasing performance will pose. ES. 22. Early analyses suggest that decentralization of expenditure responsibilities is not yet associated with better quality or with more efficient service delivery. The analysis suggests that the share of local and regional spending in total spending is not related to measures of either performance or efficiency of public spending. The only exception to this is in the case of transport services, where a higher share of sub-national spending in total spending is weakly associated with better government outcomes. The lack of a positive relationship between more sub-national spending and better outcomes is in line with the popular perception on the initial effects of the program. ES. 23. At the same time, there is some evidence to suggest that decentralization in revenues collection seems to play a role in improving public sector performance and efficiency. Notably, the share of public revenues collected sub-nationally is directly associated with efficiency and performance in Health and Education. The correlation is even stronger when specific programs are considered, such as Primary Education and Nutrition, which can be closely monitored by the population. These results indicate that the ongoing decentralization process has been only partially successful in improving service delivery, further analysis is suggested before a final verdict. ES. 24. Efficiency is negatively correlated with levels of public spending, and they are significantly lower for natural resource-rich regions. The relationship between 13 public sector efficiency and public spending is negative for all sectors (and sub-sectors) analyzed. This result is partially driven by natural resource-rich regions. These regions combine good performance with very low levels of efficiency at the expense of large amounts of public resources. Whether driven by technological factors or by incentives schemes, this fact suggests the existence of diminishing returns in the production of public goods and services. In this case, significant gains could be made by redistributing public resources away from big spenders to more efficient regions. 14 1. Introduction This chapter reviews the origins and development of the recent decentralization process and provides an overview of the key milestones along the way. The chapter also provides an interpretation of these events from a political economy perspective, emphasizing the importance of incentives, particularly electoral, institutional, and bureaucratic incentives, to help explain how the process has unfolded. A. The goals and principles underlying the decentralization process 1.1 What is decentralization? Decentralization is an ambiguous term because it refers both to a system and a process (Prud’homme 2003). As a system it refers to a situation where a substantial share of power is granted to local or regional governments. But it can also mean the process by which one moves from a centralized to a decentralized system of government. Moreover, decentralization is often defined in terms of its goals, which can be particularly confusing since decentralization can have many goals, as elaborated further below. In the case of Peru the ambiguity between decentralization as a system and a process, and the recourse to defining it in terms of its goals is enshrined in the constitutional amendment of 2002 setting out decentralization as a “form of democratic organization� that is a “permanent policy of the state� and that has as its “fundamental objective the comprehensive development of the country� while speaking also of the “decentralization process.�1 In this report, following on from World Bank (1999), we will focus on decentralization as a process. In any case, one must recognize at the outset that the term decentralization often means different things to different people and, to some extent, its use as a catch-all phrase complicates our object of analysis. 1.2 Decentralization can have many goals. Borrowing from the literature we can distinguish at least four common goals that authorities may have when promoting decentralization (Smoke 2003): (i) improved governance by empowering local populations; (ii) improved efficiency with the belief that sub-national service providers would better understand the realities on the ground; (iii) improved equity as sub-national governments may be in a better position to address poverty within their jurisdictions (though of course the redistribution from richer to poorer areas must be the responsibility of central governments); and (iv) improved development, again on the basis of the better understanding of sub-national governments on the ground. The institutional framework in place in Peru identifies the “comprehensive, harmonious, and sustainable development of the country� as the ultimate goal of decentralization and sets out that the division of power between the different levels of government is a means to the end of benefiting the population.2 At the same time, the same framework law also lays out a much broader 1 Article 188 of the Constitution as amended by Law No. 27680, Law of Constitutional Reform of Chapter XIV of Title IV on Decentralization, enacted March 7, 2002. 2 Article 3, Chapter I of the Decentralization Framework Law (Ley de Bases de la Descentralización), Law No. 27783, enacted July 20, 2002. 15 range of objectives, a total of 19, classified under five headings: political, economic, administrative, social, and environmental. The objectives cover a wide range of issues from increasing participation and equity in the distribution of resources of the state to modernizing administration processes. 1.3 Among its many goals, decentralization in Peru aimed to improve living conditions and public services and to reduce inequalities across the territory. While political goals were part of the decentralization process, public discussion of decentralization in Peru was also couched in terms of improving public service delivery (Casas 2009). For example, an early and influential assessment of the decentralization process emphasized how the process of transferring functions had as its “principal goal, to improve the services provided to the citizenry� (Prodes 2004, 35). A well-known stock-taking of the decentralization process in Peru conducted in 2006 reflects this view of decentralization as the engine for shared and inclusive growth: “The objective of decentralization is to create a model for territorial development that satisfies the legitimate expectations of inclusion, prosperity, and access to opportunity to exercise the rights [of the population]� (Azpur et al., 2006, p. 7). This is also the assessment which the World Bank has been using to support the process since its Decentralization Structural Adjustment Loan series.3 Box 1: The institutional framework for decentralization since 2002 Law (No.) Year Main characteristics Constitutional 2002 This Constitutional Reform established the decentralization process in reform Peru, whereby citizens’ participation was emphasized in the design of (27680) regional and local plans, budgets, and accountability mechanisms. Three main aspects were introduced in this reform: (i) regional governments will be created based on departments; (ii) regions can be created from one or more contiguous departments, after a positive result in a referendum; (iii) citizens will participate in the drafting of regional and local budgets and plans, as well as hold decentralized governments accountable. Framework 2002 Set the fundamentals for the gradual implementation of decentralization Decentralization in Peru. It identified key actions and complementary legislation needed Law for the process, such as the establishment of regions; the design of (27783, Ley de capacity building plans for regional and local governments; and the Bases de la transference of social programs and sectoral functions from the national Descentralización) government to regional and local governments. Ley de 2002 Established the importance of the geographical delimitation and Demarcación prohibited the creation of new districts and provinces until 2006 y Organización (extended to 2011 later). Territorial (27795) 3 The program document for the first Decentralization and Competitiveness Structural Adjustment Loan in 2003 notes that “decentralization is seen as an integral part of modernization and democratization, as well as a way to reduce inequalities and to increase responsiveness to local needs� (World Bank 2003, p. 13). 16 Organic Law of 2002 Set responsibilities for regional governments on economic development Regional promotion in their jurisdictions. It established the creation of the Governments Councils of Regional Coordination (Consejos de Coordinación Regional (27867) or CCR), with participation of mayors and representatives of the civil society, and indicated the need for checks and balances twice a year. Organic Law of 2003 Set responsibilities for local governments on economic development Municipalities promotion in their jurisdictions. It also established the creation of the (27972) Councils of Local Coordination. Framework Law 2003 Set the legal framework for the participatory budgeting at sub-national for Participatory governments. Budgeting (28056) Accreditation 2004 Established the framework for the transfer of sectoral functions from the System for national government to the regional governments. Regional and Local Governments (28273) Fiscal 2004 Established that the transfer of resources and fiscal competencies should Decentralization be implemented gradually. The transfer of resources should be aligned Law to the transfer of responsibilities to sub-national governments. Regional (DL 955) governments may participate in some aspects of national tax collection. Incentives for the 2004 Promoted the configuration of larger territorial governments (macro- Integration and regions). Formation of Regions (28274) Municipal 2007 Allowed municipalities to be associated with others to provide services Associations and execute public investment jointly. (29029, Ley de la Mancomunidad Municipal) Organic Law of 2007 Indicated the need of new legislation that clarifies the organization and the Executive functions of the Executive under the ongoing decentralization process Power (e.g., shared competencies between ministries at national level and (29158) regional and local governments). Source: Congreso de la República del Perú, Comisión de Descentralización, Regionalización, Gobiernos Locales y Modernización de la Gestión del Estado (2008). 1.4 The key principles to guide the process were fiscal neutrality and gradualism – the first stage of the process being the creation of regional governments. According to the 2002 framework law, decentralization would begin in the political sphere with the creation of regional governments using of Peru’s twenty-five existing administrative departments as a base. The law stipulated that the first elections for regional presidents were to be held before the end of the year (2002), and that regional officials were to assume their duties in January 2003. 17 1.5 In a second stage of the process macro-regions were to be formed on the basis of referenda. During the second stage of decentralization, the newly constituted regional governments were to be replaced by larger “macro-regions� based on the fusion of two or more regional governments. In contrast to the unilateral demarcation of macro-regions in the 1980s (see Box 3 in the next section for an overview of previous decentralization efforts), the regions themselves would formulate “technical plans� (expedientes técnicos) to amalgamate, and voters would then decide in a special referendum whether or not to approve these new macro-regions. Despite the passage of an ancillary Law of Incentives for the Integration of Regions in 2004, when the formation of macro-regions was put to a referendum in 2005, voters rejected all five proposals (affecting 16 departments). 1.6 During the third stage of decentralization, the national government would transfer functions to local and regional governments. Transfers of functions were to be governed by a 2004 law that established a system of accreditation according to which functions could only be transferred to sub-national governments whose sufficient capacity had been verified in advance by the national government. Transfers took place in all sectors except education and health, which would only be transferred in the fourth and final stage of decentralization (Conterno 2007). 1.7 The intended path to fiscal decentralization also reflected these principles of fiscal neutrality and of gradualism. The 2004 Law of Fiscal Decentralization adopted a two-phased approach to fiscal transfers. In the first phase, fiscal revenues would be transferred to sub-national governments as a function of the specific administrative responsibilities assumed by these governments. In the second phase, revenue sharing (coparticipación) would be introduced once macro-regions had been created via referendum. In particular, the regional governments of macro-regions would receive 50 percent of the national taxes collected in their territory – including the sales tax; the selective consumption tax on petrol and other items; and the corporate and personal income tax (art. 18 of the Fiscal Decentralization Law). Besides this, to induce the formation of macro-regions other fiscal incentives were put in place. Restrictions on borrowing placed on macro-regions would be somewhat alleviated, which would allow the regional governments to borrow domestically without prior authorization by the central government and would give priority to their requests for international borrowing. In addition, the macro-regions would also benefit from an investment fund for decentralization (Fondo de Inversión Descentralizada) to be set up by the central government. 18 Box 2: Levels of government In Peru there are three levels of government: national, regional, and local. Throughout this report we will use this terminology for consistency and to avoid any confusion between the levels of government and the territorial divisions. In fact, the territory is formally divided into four groups: districts, provinces, departments, and regions. There are a total of 1,833 districts grouped into 195 provinces. All of these districts have a local government. However, 1,638 of them are known as district municipalities while 195 are known as provincial municipalities. These provincial municipalities are in effect the municipalities of the capital of each province. The provinces do not constitute a level of government per se, although the 195 local governments that are known as provincial municipalities cover the entire jurisdiction of the province where the district is in principle and for some functions. Therefore, there are a total of 1,833 local governments. The territory is also formally divided in 24 departments and the Constitutional Province of Callao. Despite its name of “province,� Callao has the same status as a department. Therefore, it is common to refer to Peru as having 25 departments. There are, however, a total of 26 regional governments since, in addition to the regional governments in each of the 24 departments and the Constitutional Province of Callao, the Metropolitan Municipality of Lima also has the standing of a regional government. The large number of local and regional governments in Peru has been repeatedly commented upon in the literature. Of the 1,834 local governments only 109 have more than 50,000 inhabitants and 60 percent are rural. A total of 237 local governments (around 13 percent) have populations that are lower than 1,000 inhabitants. Institutional capacity and economic bases are therefore unevenly distributed. Finally, demarcation of boundaries remains an unresolved issue. Approximately 80 percent of districts and 92 percent of provinces are still not correctly demarcated, according to the National Directorate for Territorial Demarcation. The lack of clear boundaries between provinces and districts can lead to flare ups especially in situations where natural resources have begin to be exploited in a certain jurisdiction: in the current system of allocation of revenues from natural resources a high share of rents accrue to the jurisdiction where the resources are exploited (see Chapter 4). Source: Prodes (2007a), Gómez, Martínez-Vázquez, and Sepúlveda (2008), and Dirección Nacional Técnica de Demarcación Territorial de la Presidencia del Consejo de Ministros (2009). 19 B. How the process unfolded4 1.8 The main milestones in the decentralization process since 2002 have been the following: (i) Elections to regional governments were first held in November 2002 and subsequently in November 2006; (ii) A referendum for the creation of macro-regions was held in October 2005. Two-thirds of the voters rejected the proposed macro-regions. No further referenda have been called on the subject. All departments remain functioning as individual regional governments; (iii) The National Council for Decentralization was formed but was dismantled in 2007 and converted into a Secretariat of Decentralization under the Presidency of the Council of Ministers without representation of sub- national authorities; (iv) The accreditation system was put in place, and a national plan for capacity building was prepared, but it has not ensured the effective capacity of sub- national governments to perform the transferred functions; (v) In 2006 a “decentralization shock� was announced by the new administration aimed largely at ensuring that functions were transferred to the regional governments in an expedited way; (vi) As no macro-regions were formed there has been no revenue sharing. The sources of revenue for sub-national governments continue to be transfers from the central government, which in 2008 accounted for 97 percent of revenues for regional governments and 77 percent of revenues for local governments; (vii) The resources transferred to sub-national governments have increased notably but this has been the unintended consequence of the boom in natural resources, not the result of a matching of resources to transferred functions (see Chapter 2); (viii) The presidents of regional governments have created their own assembly, which has proposed further resources to be transferred and has managed to establish regular dialogue with the President of the Republic. 4 This section provides a brief overview of the process and is not intended as an exhaustive literature review. The number of documents that take stock of the decentralization process has reached the point in which a comprehensive summary of the literature would require a report in itself. In fact such a literature review is already available in Rodríguez Doig (2008). 20 1.9 Most of these issues – regarding transfers of functions and resources, as well as issues regarding inequality – will be dealt with at length in chapters 2 through 4. In the remainder of this section we will underline a few traits of the process that will not feature prominently in the rest of this report but that have significant importance in explaining the process. 1.10 The decentralization process faced a number of constraints from the outset, including an accelerated timetable. In 2002 there was a determined push to move decentralization forward. Less than a year into President Toledo’s administration, Congress approved a constitutional reform in March 2002 that mandated decentralization as a permanent and irreversible policy of the state. From March to November 2002 the push manifested itself in a total of six major pieces of legislation laying down the institutional framework for decentralization (see Box 1). However, the preparation and passing of those laws was not always smooth sailing and revealed some degree of improvisation. Reflecting the rushed nature of decentralization, it was only the day before regional elections were held on November 20, 2002 that Congress actually passed the Organic Law of Regional Governments, thereby specifying the legal prerogatives and responsibilities of the governments that were constituted the following day. The organic laws of municipalities and regional governments did not establish a great deal of clarity about the exact division of responsibilities between different levels of government (national, regional, local) – an issue to which we will return in Chapter 2. 1.11 More importantly, there was no substantive national debate on the issues related to decentralization policy. The lack of a consensus was highlighted early on in the process by observers. An influential early stocktaking of the process back in 2005 notes that “a pending issue is the implementation of a consultative and consensus- building policy over the most relevant aspects of the process� and called for the need to “renew the political agreement in favor of decentralization� (Prodes 2005, p. 8). The lack of a vision about what decentralization intends to achieve has been a frequent key shortcoming of the process (Prodes 2007, p. 7). The need for such a national debate is an aspect which has been identified in the literature as the basic first step for the sequencing of decentralization (Bahl and Martínez-Vázquez 2006). In a sense, Peru provides a case study where the detailed subsequent steps in the process have been compromised by the lack of clarity about the objectives of decentralization and by the lack of a sufficiently shared vision about what it would entail. The many recent attempts at decentralizing and recentralizing in Peru’s recent history underline the importance of a meaningful first step of national consensus-building (see Box 3 below). 1.12 Some of the accountability instruments adopted emphasized participation of civil society at the expense of the autonomy of elected sub-national officials. The law on participatory budgeting was passed as part of the decentralization agenda. Participatory budgeting has been resisted by many elected sub-national officials, who see it as diluting their electoral mandate. In addition to regional elections, the framework law also introduced participatory reforms by stipulating the mandatory creation of Regional Coordination Councils (Consejos de Coordinación Regional or CCRs) through which civil society groups would help shape regional budgets. 21 1.13 There is no clear forum for inter-governmental dialogue and dispute resolution. Resolving disputes about competencies is particularly difficult. The decentralization framework law intends for such conflicts to be resolved through recourse to the Constitutional Tribunal (article 16 of the Framework Decentralization Law). Moreover, following the dismantling of the Consejo Nacional de Descentralización (CND), which was established by the framework law for decentralization, there is now no formal institutional forum where consensus about the way forward for the decentralization process can take place. The new Secretariat of Decentralization is largely an administrative body charged with monitoring the transfer process and capacity building efforts, and it lacks political clout.5 1.14 Regional government presidents have nevertheless found new spaces for communicating their demands for greater resources. The central authorities have since attempted to launch working groups with sub-national authorities but issues of representation have made this particularly cumbersome. The national assembly of regional governments (ANGR) has gained standing as a partner for the central government but no major initiative has yet come out of their regular dialogue with the President of the Republic and the Prime Minister. In practice, the main inter- governmental mechanisms for dialogue are ad hoc bilateral negotiations between individual sub-national governments and the Ministry of Economy and Finance over budgetary issues. In isolated cases of conflict over entitlements from natural resources, street violence has taken place – as in the case of the incidents in Moquegua and Tacna in June 2008 or those of November 2008, which left three people dead.6 1.15 The creation of macro-regions remains an issue since the current division of the territory into departments does not fit well with natural economic corridors. It is admittedly difficult to establish precisely the number of “natural� regions in Peru. This is reflected in the fact that throughout different past attempts at decentralization, the number of proposed regions has varied widely: 12 in the 1987 regionalization; and seven in the 1979 Constitution, while earlier proposals in the 1930s identified 11 or 12 regions (CND 2006, pp. 12-16). What is common in all of those designs is that the number of regions was far fewer than the 26 regional governments currently in place. The division of the territory into too many regional governments can have a negative impact on the efficiency of public service delivery. Besides the economies of scale that can be reached with larger populations, a main issue in Peru is the fact that due to the difficult topography of the territory many districts and provinces naturally have easier access to 5 The very organizational structure of the Decentralization Secretariat makes this clear: the secretariat is divided into three offices: the Office for Transfers, Monitoring and Evaluation of Competencies, the Office of Capacity Building at Regional and Municipal level and Intergovernmental Coordination, and the Office of Investment Management (Presidencia del Consejo de Ministros 2008, p. 15). 6 The conflict between Moquegua and Tacna stems largely from the fact that the same company, Southern Peru Copper Company has two operations close by in Cuajone (Moquegua) and Toquepala (Tacna) which for the purposes of calculating the canon are reported separately and which, according to the company’s figures have seen rather different performances in recent years. Thus, the amount of canon received by Tacna has increased from 151 million nuevos soles in 2005 to 712 million in 2008, while in Moquegua the increase has been much more modest, from 149 million in 2005 to 212 million in 2008. The two mines are 35 kms apart and from an economic standpoint constitute a single complex that shares a number of processing facilities. 22 the capital of a neighboring department than to the capital of its own. This is the case for the provinces in the South of the Ayacucho department; the provinces in the rainforest in Pasco; the provinces of Jaén and San Ignacio in the department of Cajamarca – where the easiest route to the city of Cajamarca goes first through the city of Chiclayo in the Lambayeque department; the province of Sánchez Cerro in Moquegua; or the districts of Inkawasi and Cañaris in Lambayeque (Prodes 2007a, pp. 19-22). One source of inefficiency stems from the fact that regional governments have an interest in spending on intra-departmental network links (e.g., a direct road from the city of Cajamarca to the province of Jaén, which would help integrate the department) when, from an economic geography point of view, it may be more efficient to consider inter-departmental links with a neighboring department (e.g., for the province of Jaén it would be more natural to improve communications with the city of Chiclayo in Lambayeque). Box 3: A brief history of de- and re-centralization in Peru Direct elections for municipal offices took place for the first time in 1963 but were interrupted after only five years when a military coup in 1968 ended this incipient experience with local democracy. While the generals inherited a quite centralized political system, the highly statist and nationalist policies they pursued while in office brought the country to a still more centralized political system. Unlike other countries in Latin America, in Peru the transition to democracy in 1980 did not provide an opening for decentralization, despite the return to municipal elections and despite the fact that the new 1979 Constitution called for the creation of regional governments. For most of the 1980s, two factors kept decentralization off the table: an internal armed conflict between the state and insurgent groups, and an economic crisis that culminated in major macroeconomic instability and hyperinflation. The armed conflict worsened the environment for decentralization both because the Shining Path explicitly targeted mayors and other local leaders with acts of violence, and because the national government responded by giving the military free rein over much of the Peruvian interior. On the economic front, President Alan García responded to the international debt crisis by adopting expansionary fiscal and monetary policies upon the historic election of his APRA party in 1985 (Alianza Popular Revolucionaria Americana or American Revolutionary Popular Alliance). The expansionary economic policies put in place, however, only served to worsen the crisis. But the disastrous economic performance indirectly set the stage for political decentralization. As the president’s popularity continued to plummet along with the economy, it became clear that the APRA party would fare poorly in the next presidential elections scheduled for 1990. In this context, new electoral spaces at the regional level were created – a move that has been interpreted as driven by a hope that APRA candidates could win some of these regional elections and in this way retain a foothold in government with which to recover and rebuild the party (Schmidt 1989, Tanaka 2002, O’Neill 2005). The administration’s gambit worked in the sense that the APRA won control over half of the new regional governments in elections held in 1989 and 1990, even as its candidate for president failed to make it beyond the first round of 23 the 1990 presidential elections. While this “pre-emptive� creation of regional governments enabled the APRA to retain political relevance despite losing the presidency, this attempt at decentralization proved to be short-lived. Elections in 1990 brought political outsider Alberto Fujimori to the presidency at a moment of tremendous political and economic uncertainty in Peru. After only two years in power and with the support of the military, Fujimori enacted the so-called auto golpe (self-coup), which he justified as necessary in order to tackle the twin problems of insurgency and hyperinflation. Though his closing of the legislative branch and his intervention in the judiciary received more attention in the press and from scholars, Fujimori also abolished the regional governments and dismissed the regional authorities who had been elected for the first time only a few years before. Through the creation of the CTARs (Concejos Transitorios de Administración Regional or Transitory Regional Administrative Councils), Fujimori downgraded the regions from governments back to mere administrative units and placed them firmly under the political authority of the national government (Zas Friz 2004). As part of his attempt to suppress the more politically threatening regions, Fujimori privileged the municipal sphere of government and allowed elections to continue at the municipal level throughout his decade in power (Dickovick 2007 and forthcoming). On the fiscal side, Fujimori retained centralized control over public revenues, heavily manipulating spending in the successful attempt to manufacture support for the passage of a new constitution in 1993 and his own re-election in 1995 (Roberts 1996). In a variety of dimensions, Fujimori brought centralism in Peru to new heights, which provides the backdrop against which decentralization emerged as a salient reform proposal in the decade since his fall from power. This box draws from Eaton (2009). For a history of decentralization efforts back to Peru’s first Constitution in 1823 see Consejo Nacional de Descentralización (2006). 24 C. Explaining the route travelled so far: the role of incentives7 1.16 The decentralization process reflects the choices made by politicians and policy-makers. To date, analysts have typically focused on a series of technical and structural factors in the attempt to identify the factors that have shaped the outcome of decentralization in Peru. Some, for example, focus on the lack of clarity in the assignment of expenditures and on the absence of an effective system of public financial management (Ahmad and García-Escribano 2008). Others underscore the excessive fragmentation of local government in Peru, which denies individual sub-national units the economies of scale that would enable them to provide goods and services more effectively (Gómez, Martínez-Vázquez and Sepúlveda 2008). There is also a widely-held view that insufficient levels of institutional capacity at the sub-national level are critical for explaining the process and its delays. While all of the above help explain the process, and we will return to them in this report, as a complement to these mostly technical approaches, this section seeks to shift the focus toward the political economy aspects that have shaped the decentralization process in Peru. 1.17 This report draws on a recent diagnostic framework that emphasizes the electoral, institutional, and bureaucratic incentives faced by decision-makers as a key factor when explaining decentralization outcomes. In order to understand why decentralization was adopted in the first place, how it was designed, and what has prevented it from achieving greater traction, we need to pay greater attention to the incentive structures through which politicians and bureaucrats alike viewed decentralization. In doing so, this section applies a more general diagnostic framework developed by Eaton, Kaiser and Smoke (2009), which provides an organizing structure for looking systematically at the electoral, institutional, and bureaucratic incentives. According to the argument developed here, the timing and content of decentralization in Peru cannot be explained without reference to the underlying partisan, electoral, institutional and coalitional incentives that face national officials. Electoral incentives 1.18 Electoral and partisan incentives have long been recognized as helping to explain how the previous decentralization process unfolded in the 1980s and 1990s. As described in Box 2 above, the timing of decentralization in the late 1980s, at the end rather than at the beginning of the first administration of President Alan García, has been interpreted as reflecting the APRA’s electoral fears as a more important factor than the normative commitment to decentralization that APRA had long espoused. Recentralization in the 1990s likewise conformed to President Fujimori’s electoral calculations. Canceling regional elections eliminated spaces of contestation in which Fujimori could have faced strong opposition – a calculation that may have also led him to insist on a shift, in the election of legislators, away from smaller regional districts and toward the single nation-wide district that was introduced in the 1993 constitution. 7 This section draws from Eaton (2009). 25 1.19 Decentralization resurged as a theme in the 2000 and 2001 presidential campaigns. As a presidential contender, Alejandro Toledo championed decentralization on the campaign trail in the 2000 and 2001 elections. As a candidate, Toledo’s embrace of decentralization in 2000 as a policy also bolstered his support in the departmental capitals and secondary cities. In fact, many of these regional economies had turned strongly against Fujimori and railed at his centralizing practices (Gonzales de Olarte 2000). Though Fujimori won the 2000 election when Toledo withdrew from the second round, his regime collapsed shortly thereafter in a devastating corruption scandal (McClintock 2006). In the subsequent 2001 elections, Toledo reiterated and strengthened his proposals for decentralization in an attempt to neutralize the pro-decentralization credentials of his chief rival, former President Alan García. On the campaign trail, Toledo promised to hold regional elections immediately as a means of undoing Fujimori’s centralizing legacy. 1.20 The holding of regional elections early in President Toledo’s term responded to his electoral campaign promise. Once in office, President Toledo held his promise to call for regional elections early in his term. He certainly had cause to worry that APRA would do well in these elections given that it was the only party that had survived the collapse of the party system under Fujimori (Cameron and Levitsky 2003, Grompone 2005). However, by scheduling the elections early in his administration, Toledo may have hoped that his popularity would help produce regional victories for candidates of the Perú Posible party that had served as his electoral vehicle (though by the time regional elections were held, the president’s popularity had already begun to seriously slip). Bold action vis-à-vis decentralization was also appealing because it helped Toledo distinguish himself from Fujimori, whose market-oriented economic policies he intended mostly to sustain. A heterogeneous electoral coalition had brought Toledo to the presidency, and decentralization was one of the few issues upon which virtually all of his main supporters were agreed – unlike, for example, the direction of economic policy. 1.21 The outcome of the first regional elections may have contributed to slowing down the decentralization momentum. If electoral dynamics account for the sudden introduction of regional elections in November 2002, the outcome of inter-party competition in those elections explains the equally sudden decline in the president’s enthusiasm for decentralization thereafter. In part reflecting its status as the only party in Peru with any real organizational capacity in the regions, the opposition APRA party emerged as the big winner of the 2002 elections, taking the presidency in 12 of the 25 regions. In contrast, Toledo’s governing party was only able to win the presidency in one region. Thus for Toledo political decentralization was, as Tanaka argues (2002), a “major miscalculation.� In the aftermath of his 2002 electoral debacle, for Toledo to push forward the administrative and fiscal measures that were supposed to follow political decentralization would have been to strengthen governments that were controlled by the opposition. Worse, as Toledo’s public approval ratings remained very low for most of his administration, and as many came to question whether he would be able to finish his mandate at all, the president could ill afford to devolve resources away from the center as this could have further weakened his position and authority. 26 1.22 The results of the second regional elections in 2006 may have further contributed to limit the interest of the central authorities in decentralization. Partisan incentives may also help explain García’s lack of interest in decentralization subsequent to his reelection as president in July 2006 – despite the reality that he regularly excoriated Toledo for his lack of interest in promoting decentralization. Only six months after García’s presidential victory, all but one of APRA’s regional presidents lost their re-election bids in the December 2006 regional elections (one other APRA candidate was elected regional president, and as a result APRA currently governs in 2 of the 25 regions). Like Toledo before him, García has faced a set of regional presidents who are not his political supporters, generating in turn strong partisan reasons to move slowly on decentralization. While the introduction of sub-national elections always increases the possibility of greater political friction between sub-national and national officials, this effect has been particularly pronounced in Peru given the inchoate nature of its party system and the significant fragmentation of politics along territorial lines. 1.23 The limited influence of regional power brokers in national politics is also a key factor shaping the decentralization process. While inter-party dynamics are critical for understanding the rhythms of decentralization in Peru, intra-party dynamics in contrast are less relevant. In other countries in Latin America, including Argentina, Brazil, Colombia and Mexico, sub-national party leaders wield a great deal of influence within their parties, which they can and do use to prevent national officials from re- centralizing and/or dragging their feet on decentralization (Willis, Haggard and Garman 1999). In Peru, national party leaders have controlled their parties from the top, monopolizing decisions about campaign finance and even the selection of candidates in local races. While the regular holding of sub-national elections could in principle enable sub-national party leaders to begin to exercise greater authority within Peruvian national parties, recent changes in electoral laws make this unlikely. Peru’s 2003 electoral law, for example, stipulates that candidates no longer need to be affiliated with a political party in order to run for municipal office, a rule that discourages party building and that has exacerbated the fragmentation of political authority among geographically-specific regional movements (Vergara 2007). 1.24 Overall, electoral and partisan incentives help to explain the punctuated nature of decentralization in Peru with the successive opening and closing of rather brief reform windows. Institutional incentives 1.25 While electoral and partisan incentives may help to explain the timing of reforms, institutional incentives appear to explain the content of those reforms. Like their counterparts everywhere, national politicians in Peru have resisted endorsing changes that reduce their policy making authority and that transfer power to potential rivals at the sub-national level. Unlike their counterparts in many other countries, however, national politicians in Peru have been relatively free to design decentralization to their liking. 27 1.26 National politicians in the 2000s faced a number of institutional incentives that made devolving substantial power to sub-national governments less likely. First, the legislative power saw a recovery of influence after the Fujimori period in which its influence was minimal. Having just regained this power at the central level, Congress was unlikely to be prone to divesting itself of much power too quickly. Second, the election of legislators in a single nation-wide district rather than in sub-national electoral districts, did not contribute to the support for more generous decentralizing measures in Peru’s unicameral legislature. 1.27 The emergence of nearly chronic protests in the regions has paradoxically strengthened the incentives of central authorities not to decentralize. Regional politics has yet to become intertwined with national political parties. As a result, regional leaders play a limited role within the apparatus of national political parties and therefore wield limited influence. In many regions national political parties are actually of limited importance – as reflected in the fact that a majority of regions are now governed by “regional movements� (movimientos regionales) rather than political parties. But these often ephemeral, grievance-oriented and locally-specific movements have had a difficult time scaling up their mobilizing efforts (Panfichi 2007). In many countries, national politicians support decentralization because it appeals to important constituent groups in the coalitions that keep them in office. In the power vacuum that was created by the sudden collapse of the Fujimori regime in 2000, however, national politicians in Peru have been relatively unconstrained by the coalitional pressures that stronger societal actors might have exerted. Therefore, in the presence of widespread but atomized protest, and in the absence of more coherent and larger-scale societal groupings, coalitional pressures have been far less important in Peru than the electoral, partisan and institutional incentives that have directly shaped national politicians’ behavior. As discussed below, the cautious and ambivalent attitudes of national politicians toward decentralization are reflected in a variety of important design choices. 1.28 The gradualist approach and the reliance on central systems of control can be seen as signs of the caution and suspicion with which many central authorities perceived the decision to decentralize. Legislators endorsed an ad hoc approach to administrative decentralization according to which services would be transferred as the result of negotiations between the national government and individual sub-national units. The requirement that sub-national governments had to be accredited before they could receive services guaranteed the continued influence of national politicians. Furthermore, the national government has also maintained tight administrative control over sub- national governments through the National System of Public Revenues (Sistema Nacional de Ingresos Públicos or SNIP), which requires national-level approval for sub-national expenditures above a low threshold. Overall, at almost every step of the way, national politicians opted for a highly discretionary approach that maximized national control. 1.29 The limited fiscal decentralization envisaged in the 2004 framework law contrasts with the changes typically seen in other Latin American countries. There is very little in the 2004 fiscal decentralization law that could be construed as in any way politically threatening to national politicians – in contrast to more sweeping fiscal changes in other Latin American countries that have given sub-national officials a more 28 significant share of public revenues. Rather than transfer tax bases to sub-national governments, or legislate a set of criteria that would be used to distribute revenues from nationally-collected taxes, national politicians endorsed a much more discretionary approach. Even the rates for the local property taxes were still set by national authorities, the local authorities being able only to affect the collection through enforcement efforts, as discussed in Chapter 4. Fiscal transfers were made contingent on administrative transfers and on the creation of macro-regions in a national referendum (a politically unlikely scenario, as discussed below). The insignificance of fiscal decentralization in the Peruvian case cannot be understood as a result of the country’s unitary design; it is not just federal countries that have adopted revenue sharing with sub-national governments but unitary systems like Bolivia, Colombia, and Ecuador. 1.30 Institutional incentives trumped the fiscal incentives for the formation of regions. The sequencing decision by legislators to hold the referendum on the creation of macro-regions after elections for regional officials reflects a serious design flaw if the intention was truly to create a smaller number of regional governments. Unsurprisingly, the individuals who were elected as regional presidents and as regional assembly members in 2002 faced strong incentives to campaign against the macro-regions since the possible fusion of regional governments could have resulted in the elimination of their own elected positions. As an analyst put it on the eve of the 2005 referendum, the proposed formation of regions was a technical exercise that ignored existing political interests and doubted that the departmental political elites “would go for an uncertain process that would not necessarily favor them� (Muñoz 2005, p. 54). At the same time, we must recognize that there are also actors at the sub-national level that recognize the benefits of regional integration. Although the referendum failed, ongoing negotiations have continued among three groups of regions (the juntas de coordinación inter-regional that were created by the 2004 Law of Incentives to Integrate Regions) in the North, center and Amazonian parts of the country. 1.31 The failed formation of regions provides a cautionary tale of how institutional incentives can trump even relatively large fiscal incentives. The election results show a negative correlation between per capita income and the share of votes in a given department in favor of the formation of macro-regions (right panel in Figure 1.1). This suggests that citizens in richer departments were wary of joining macro-regions with poorer departments that could lead to subsidies to the poorer areas. However, it is remarkable that even in the poorest areas support for macro-regions was lackluster. These results may be explained by the campaigning by regional officials against macro-regions and suggest that even in the presence of potentially large fiscal incentives – amalgamated regions would have received 50 percent of all tax revenues collected in the macro-region instead of the current system of transfers – the role of institutional incentives can easily trump them. In the words of Eaton, Kaiser, and Smoke (2009) officials from the regional governments displayed a case of “where you sit is where you stand.� The conclusion drawn from this analysis is therefore subtly but significantly different from the standard assessment of the reasons for the failed referendum, which point to the “weak incentives� in place (Prodes 2006, p. 7). In our analysis, it is not so much that the incentives were weak but rather that they were of the ‘wrong’ type: they were fiscal in nature when powerful institutional incentives were going in the opposite direction. 29 Figure 1.1: Results from the referendum for the formation of macro-regions in 2005 Source: World Bank staff calculations based on Oficina Nacional de Procesos Electorales and PNUD (2008). Bureaucratic incentives 1.32 A fuller account of the process must also incorporate analysis of bureaucratic incentives and bureaucratic struggles. As a reflection of his commitment to decentralization in the early phase of his administration, Toledo pushed for the creation of a new bureaucratic agency: the National Council for Decentralization (Consejo Nacional de Decentralización or CND) that would be responsible for directing and implementing decentralization. According to the 2002 framework law, the CND would be composed of representatives of both national and sub-national governments, which was important given the extent to which the debate among politicians in the legislative sphere was dominated by national rather than sub-national voices. The four members of sub-national governments broke down accordingly: two members represented regional governments, one member represented the provinces and one member represented municipal districts. Representing the national government, two members of the CND were officials from the Ministry of Economics and Finance (MEF) and two members were officials from the Presidency of the Council of Ministers (PCM, the body that coordinates the policies and programs of the executive branch in Peru). The ninth member of the CND (Luis Thais) served as a representative of the President of the Republic. 1.33 The National Council for Decentralization engaged in regular confrontations with other central government bodies. In general, the CND favored a deeper, more extensive and more rapid approach to decentralization than the MEF, which favored gradualism and which consistently encouraged caution. Within the bureaucracy, MEF was the agency charged with defending fiscal stability and macroeconomic balance. MEF favored the type of responsible decentralization that would not put the country’s hard won economic stability in jeopardy. For example, in debating the 2004 Law of Incentives for the Integration of Regions, representatives of sub-national governments on the CND 30 wanted to pledge as an incentive that macro-regions would receive not just revenue transfers but their own tax bases as well. The MEF, in contrast, successfully argued that macro-regions could receive either tax bases or revenue transfers but not both. Likewise, the MEF insisted on the imposition of locks (candados) on sub-national governments to prevent them from borrowing without the prior approval of the national government. It has also been argued that the support of the MEF for the participatory dimensions of Peru’s decentralization process responded in part to a belief that greater civil society participation and oversight would help to ensure greater efficiency in public spending by acting as a check on sub-national authorities from below, as well as a check on legislators that could introduce pork-barrel projects in the budget during their discussion in Congress (McNulty 2008). In this sense, decentralization in Peru was shaped not only by MEF’s conflict with CND, but by its longstanding struggle with Congress as well. 1.34 Different bureaucratic incentives were also reproduced within the National Council for Decentralization. Due to the membership structure of the CND, these inter- bureaucratic conflicts were reproduced within the CND itself thanks to the participation of its two MEF representatives. As a newly created agency, the CND was simply no match for the MEF in the conflict that developed between these two bureaucracies. In addition to losing specific policy arguments with the MEF, it must be said that the CND also failed to offer much leadership vis-à-vis the decentralization agenda. Most problematic was its lackluster and sporadic public relations campaign to inform voters about the 2005 referendum on macro-regions, the failure of which led to paralysis in the decentralization process. Beyond the referendum for the formation of macro-regions, the CND was also widely criticized for the slowness of administrative transfers and for generalized confusion in the process by which sub-national governments were accredited to receive devolved functions (Conterno 2007). 1.35 After five years of pushing decentralization – sometimes consistently, sometimes ineffectively – the CND was eliminated in 2007. While the CND received numerous and well-justified criticisms, its elimination marked an important setback for sub-national governments and a further sign that the decentralization agenda that had been so meticulously delineated on paper could simply not be implemented as planned. The responsibilities of the CND were transferred to a new Secretariat for Decentralization under the direct control of the Presidency of the Council of Ministers (PCM). This new Secretariat was assigned only one-third the staff that the CND had employed and included no regularized role for any representatives of sub-national governments (Conterno 2007). The end of the CND thus eliminated the only official governmental space that allowed for the collective expression of sub-national governments. After its elimination, García was free to pursue direct relationships with individual regional presidents outside of and unmediated by the bureaucracy. Regional presidents have responded by creating a new collective body: the National Association of Regional Governments (Asociación Nacional de Gobiernos Regionales or ANGR). Under a rotating leadership, ANGR has emerged as an important voice on such issues as fiscal decentralization and articulating the demands to relax SNIP controls on sub-national spending. 31 Shifting incentives 1.36 The mining boom in recent years has altered some of the incentives faced by different stakeholders. The most important source of dynamism in recent years is the system of royalty payments, known as canon, according to which the national government transfers to sub-national governments 50 percent of the corporate income tax that it receives from mining and other companies (including in the gas, forestry and hydroelectric sectors). Royalty payments disproportionately benefit the municipal district where the natural resource is extracted and to a lesser extent the surrounding jurisdictions within the province and the department. Departments where no natural resources are extracted receive no revenues from this scheme. Importantly, these natural resource transfers are not related to any functions or services performed by the recipient sub- national governments and are effectively super-imposed on the decentralization framework. The generous terms of the royalty system, which stand in stark contrast to the 2004 Fiscal Decentralization Law, were established in the 1993 constitution and by the interim administration of President Valentín Paniagua (2000-2001), before the onset of the global mining boom. Subsequently, the spike in mineral prices in global markets as well as increased output have generated significant revenue flows to those sub-national governments that happen to have mineral wealth. 1.37 The growing importance of the canon system has complicated the decentralization process in Peru. For example, the structure of royalty payments, which seems to depart from the view that has traditionally identified subsoil wealth in Latin America as the patrimony of all the nation’s citizens, may have impacted the politics of the 2005 referendum on macro-regions. Opposition in several regions (e.g. Cajamarca, Moquegua) to the formation of macro-regions was particularly high due to concerns that they would have to share their mineral wealth with others. Gross inequality in the levels of canon payments sent to different municipalities, provinces and regions prompted an attempt by President García to introduce a compensation fund for “have not� regions, a proposal that provoked violent responses from producing regions and that forced the president to back down. Finally, canon payments have contributed to the confusion over the true extent of the transfer of functions that has taken place. Thanks to the canon system, representatives of the national government can say that they are transferring substantial revenues to sub-national governments, despite the failure to endorse a coherent fiscal decentralization law that would benefit all sub-national governments. 1.38 The remainder of the report reviews, in turn, the allocation of roles (Chapter 2) and spending (Chapter 3), the sources of finance (Chapter 4) and the efficiency of public spending in different departments (Chapter 5). 32 2. Assignment of responsibilities for service delivery In this chapter we review the functions that are assigned to different levels of government and the accreditation process by which those functions were devolved. We also discuss the main ‘contact points’ between different levels of government which are all the more important in ensuring intergovernmental coordination given the lack of clearly defined functions. The chapter uses the example of the education sector to illustrate some of the issues regarding the overlap in functions across levels of government, while details regarding the health and social policy sectors are provided in the annex to this report. A. Assignment of functions 2.1 The organic laws of the executive power, regional governments, and municipalities set out the assignment of functions for each level of government. The assignment of functions across different levels of government is set out in the Constitutional Reform, and in the organic laws. The framework laid out three types of responsibilities: (i) those that are the exclusive charge of the central government, (ii) those that are shared by different levels of government; and (iii) those that may be delegated by the central government to other levels of government. In practice, however, the organic laws set out a large number of shared competences. While this may provide for some flexibility in the implementation of the framework it also implies that there is limited clarity as to who is responsible for what. As Table 2.1 below shows, functions such as education, public health, or security are shared across all levels of government. 2.2 Even though subsidiarity is endorsed in principle, the design and supervision of sectoral policies is an exclusive charge of the central government – thus limiting autonomy of sub-national governments. The principle of subsidiarity is endorsed in the Organic Law of Regional Governments of 2002 (art. 8) and, as noted above, many sectoral functions such as education or health are shared obligations. However, the more recent organic law of the executive power from 2007 states in art. 4 that the design and supervision of national “and sectoral policies, which are compulsory for all entities of the State at all levels of government� is an exclusive competence of the central government. This article goes further in undermining the autonomy of sub-national governments by explicitly stating that “the national and sectoral policies take into account the general interests of the State and the diversity of the local and regional realities.� Thus, it clearly puts the central government – as the only level of government charged with designing sectoral policies – in charge of identifying regional and local peculiarities. Additionally, art. 23 of the Organic Law of the Executive Power sets out that each Ministry is in charge of formulating, planning, directing, coordinating, implementing, supervising, and evaluating national and sectoral policy under its direction, “applicable to all levels of government.� All of this would suggest that there is no space to transfer legal authority to sub-national governments for the elaboration of sectoral policies. This peculiar situation adds more complexity to the functioning of the government. 33 Table 2.1: Allocation of responsibilities by level of government Central government Regional governments Local governments Exclusive  Foreign relations  Regional development  Urban and rural  Defense, national security, plans and executing municipal development and armed forces corresponding socio-  Management and  Justice, with the exception of economic programs regulation of local public Justice administration  Internal organization of services  Internal order, national and the regional government  Internal organization of border police  Promote and implement the local government  Tax administration of public investment of  Local development plan national scope and national regional scope in roads,  Execution and public borrowing communications, and monitoring of local  Foreign trade and tariff policy basic services public infrastructure  Regulation of merchant  Development of tourism marine and commercial air circuits transport  Administer state land  Regulation of public services within their jurisdiction  Regulation of public (except municipal land) infrastructure of national  Demarcation of territorial scope limits within the region  Any others set by law in  Modernization of small accordance with the and medium enterprises Constitution  Promote sustainable use  Design and supervision of of forestry and national and sectoral policies, biodiversity resources which are compulsory for all levels of government Shared  All other responsibilities  Education: management  Education: take part in of education services for management of pre-school, primary, education services as secondary, and higher would be determined in education (except the sectoral law university)  Public health  Public health  Culture, tourism,  Regulation of economic recreation, and sports activities in their sphere  Security (seguridad  Sustainable management ciudadana) of natural resources and  Monument conservation improving the  Public transport and environment traffic  Preserving and  Housing and urban administering regional rehabilitation natural reserves  Service and management  Culture and arts of social programs  Regional competitiveness  Management of social and job promotion programs  Citizens’ participation  Waste management Source: Compilation by World Bank staff based on Articles 4 and 5 of the Organic Law of the Executive Power; articles 35, 36, 42, and 43 of the Decentralization Law; and the Organic Law of Regional Governments. 34 2.3 The functions assigned to regional and local governments overlap with each other and with the functions assigned to the central government. Despite the fact that the design of sectoral policies are exclusive responsibility of the central government, the range of functions assigned to the regional and local governments is so wide that they seem to overlap with the design of sectoral policies. This problem becomes evident simply by looking at the long list of functions that the organic laws of regional governments and municipalities assign to each level of government. As an example, Table 2.2 shows the assignment of functions in education by level of government. We can see that the regional governments are tasked, for example, with designing “the regional education project� and with “diversifying national curricula� – tasks which would appear to benefit, and require, a substantial degree of autonomy. Table 2.2: Comparison of functions assigned to regional and local governments Regional governments Local governments 1. Formulate, approve, execute, evaluate and 1. Promote sustainable human development at administer the education, culture, science and the local level, favoring the development technology, sports and recreation regional of educational communities. policies for the region. 2. Design, execute and evaluate the 2. Design, execute and evaluate the regional educational project of their jurisdiction, in educational project, the culture development coordination with the Regional Education programs, science and technology, and the Office (REO) and the Local Education sports and recreation development program. Management Unit (LEMU) and 3. Diversify the national curriculums, contributing to the national and regional incorporating significant contents of their educational policy with an intersectoral sociocultural, economic, productive and action focus. ecological reality. 3. Promote curriculum diversification adding 4. Promote a culture of rights, peace and equality significant content to their sociocultural, of opportunities for everyone. economic, productive and ecological 5. Promote, regulate, incentivize and supervise reality. services related to elementary, primary, 4. Supervise the pedagogic and administrative secondary and non-university superior management of the educational institutions education; in coordination with the local under their jurisdiction, in coordination government (LG) and in accordance with the with the REO and the LEMU and policies and regulation of the sector. strengthening their institutional autonomy. 6. Modernize the decentralized educational 5. Construct, equip and maintain the management systems and favor the formation of infrastructure of the educational educational networks, in coordination with establishments in their jurisdiction in MINEDU. accordance with the agreed Regional 7. Execute and evaluate, jointly with the LG, the Development Plan and the assigned literacy programs. budget. 8. Integrate the different educational regional 6. Support the creation of educational programs in an integral policy oriented to the networks as an expression of participation improvement of the region’s productivity; the and cooperation between the centers and equality of opportunities, integration and the educational programs of their inclusion at the regional level; the strengthening jurisdiction. of the citizen participation; and to the 7. Promote and organize the Local Education development of a culture of respect and Participative Council in order to generate recognition for diversity. agreements and to foster the citizens’ 9. Permanently promote intercultural education. surveillance and control. 10. Promote and incentivize investigation and 8. Support the incorporation and development 35 Regional governments Local governments extension in the universities and other superior of new technologies for the improvement educational institutions. of the educational system. This process 11. Promote and disseminate cultural expressions enhances relationships with other sectors. and foster artistic and cultural institutions of the 9. Promote, coordinate, execute and evaluate region, in coordination with the LG. with the RG the literacy programs within 12. Protect and preserve, in coordination with the the frame of national policies and LG, the national cultural heritage of the region, programs. as well as the non-official cultural property of 10. Strengthen the supportive spirit and the the region. collective work oriented to development of 13. Design and implement infrastructure and social, harmonious and productive equipment policies with the LG. coexistence; the prevention of natural 14. Identify, implement and promote the use of disasters and the security of the citizens. effective and efficient new technologies in order 11. Organize and support cultural centers, to improve the quality of different levels of libraries, theaters and art workshops in education. provinces, districts and populated centers, 15. Develop and implement information systems 12. Foster the protection and promotion of the and make them available to the population. cultural heritage of their jurisdiction and 16. Periodically and systematically evaluate the the defense and preservation of the achievements of the region in educational archeological, historical and artistic matters and support the evaluation and monuments, contributing with the relevant measurement of the Ministry of Education national and regional organizations. (MINEDU), as well as contribute to the 13. Promote the culture of prevention through development of an accreditation and the education for the preservation of the certification policy for educational quality. environment. 17. Encourage and participate in the design, 14. Promote and administrate, directly or execution and evaluation of research projects through agreement or concession, zoos, and educational experimentation and innovation botanic gardens and natural forests. that will contribute to the regional development 15. Promote sustainable tourism and regulate and the improvement in the quality of the the services related to that purpose in education service. coordination with the relevant entities. 18. Develop the professionalization, capacity- 16. Foster a civic culture of respect for public building and actualization processes for teachers goods and maintenance, cleanliness, and administrative personnel, in accordance with conservation and improvement of the local the continuous formation national plan. property. 19. Strengthen, in accordance with the LG, 17. Promote participative, educational and educational institutions by promoting their recreational spaces aimed at local elder autonomy, innovation capacity and democratic adults. functioning, as well as the intersectoral 18. Regulate, coordinate and promote sports articulation with the participation of the society. and child recreation through the 20. Articulate, advise and supervise local construction of sport and recreational management unities in the administrative and centers or the temporary use of appropriate pedagogical fields. urban zones for this purpose. 21. Promote and articulate the participation of 19. Promote diverse cultural activities. universities, enterprises and civil society 20. Promote the consolidation of a democratic institutions in the execution of regional culture and strengthen the cultural identity development plans. of the rural, native and afro-Peruvian population. Source: Compilation by World Bank staff based on art. 47 of the Organic Law of Regional Governments and the Organic Law of Municipal Governments. 36 2.4 Deconcentrated entities play, however, a substantial and increasingly important role in day-to-day management issues. Keeping with the education sector as our example, we can see that a large number of functions were already performed at lower levels of government before the decentralization process was launched in 2003. This is because a number of local units and Regional Offices (or Regional Directions) for educational management were already ascribed to the Regional Administration Transitory Councils and were subsumed into the regional governments upon their formation. This solution was applied in the case of other line ministries too. In other words, the functional and regulatory structure of the education sector was already deconcentrated in three tiers: (i) the Education Ministry; (ii) the Regional Education Office (REO or Regional Education Direction); (iii) the Local Education Management Unit (LEMU, Unidad de Gestión Educativa Local or UGEL). These were, and to a certain extent remain, non-concentrated entities from the Central Government within the Education Ministry (MINEDU). The Local Education Management Units follow the guidelines set out by the Regional Education Office and provide the pedagogical, institutional and administrative technical support to the educational centers, working as the executive organizations of the Regional Offices. In turn, the Regional Offices follow the guidelines and policies set out by the Ministry. 2.5 The decentralization process was to take place on top of this deconcentrated structure. The decentralization process assigned a number of education competences to the regional and local governments. These competencies, or specific functions, are in many cases so loosely defined that the risks of duplication or overlapping are large, as shown in Table 2.3. Hence, there are three regulations aimed at defining the role of each government level, which triples the responsibility of management supervision within one jurisdiction and the language is imprecise enough so as not to give a clarity as to who is supposed to do what. A final layer of complexity is added by the inclusion of community participation instances through the Participative Educational Councils at the regional (COPARE), local (COPALE) and school (CONEI) levels. Table 2.3: Allocation of education responsibilities by level of government Central government (Regional Education Offices and Local Regional governments Local governments Education Management Units)  “The Regional Education Offices  Promote, regulate, incentivize  Supervise the pedagogic and […] are responsible for the and supervise the services administrative management of promotion, coordination and related to elementary, the education institutions evaluation of educational primary, secondary and non- under their jurisdictions, in development, science and university superior education; coordination with the REO and technology, culture, recreation in coordination with the LG the LEMU and strengthen their and sports on their field.� and in accordance with the institutional autonomy. policies and regulation of the sector. Source: Compilation by World Bank staff based on article 2 of the Regional Education Offices and Local Education Management Units Organization and Functions Rules (approved by Supreme Decree Nº 015-2002-ED), article 47 of the Organic Law of Regional Governments and the Organic Law of Municipalities. 37 2.6 Faced with this degree of complexity the transfer plans did not initially envisage the transfer of the education sector’s functions. Even though the organic laws of regional governments and municipal governments gave shared competencies to sub-national governments over education, the gradualist approach envisaged that the issue of education, like that of health, would be tackled in the last stage of the decentralization process. While this appears to be a cautious approach in principle, it effectively voids of much significance the decentralization process as it involved no autonomy of the lower levels of government. In practice, the functions were already being fulfilled by the deconcentrated entities of the Ministry of Education (the regional directions and local management units). The regional directions were nominally transferred to the Regional Government but for all intents and purposes continued to act in the same way as they had been doing in the past. 2.7 The initial plan to decentralize education last was changed as part of the so- called “decentralization shock� launched in October 2006. Among the reforms that were part of the decentralization shock, a primary education municipalization process was started, formalized by the Supreme Decree 078-2006-PCM where the MINEDU was authorized to implement a pilot plan for elementary and primary education that would start in 2007 in 56 district municipalities. 2.8 Under the new Education Management Municipalization Plan, local authorities would be significantly empowered. With this new plan, the scheme described above ceased to function and the Local Government became the main player in the management of educational institutions through the Local Educational Council (LEC) led by the mayor and was responsible for the main decisions in the pedagogical, institutional and administrative fields. The LEC, in this case, would be formed by members from the Institutional Education Council (CONEI) and by local government members. Additionally, an important management role is given to the society through the Institutional Education Councils that would become entities with the legal authority to make decisions in the pedagogical, institutional and administrative fields. 2.9 The new approach to decentralize education is still in its early stages and has yet to be rolled out across the territory. The new plan to decentralize education has three planned stages: a first pilot stage (2007-2008), a second stage of expansion where there would be an increase in the number of districts during 2009-2010; and a final generalization stage during 2011-2015 where the total transfer of educational management functions to all the districts and provinces is expected. 2.10 The education municipalization plan has the potential to clarify the roles of the different levels of government but given capacity limitation, it is unclear that municipal management would be the most effective way forward across the territory. According to this new approach to decentralizing education, the central government entities, REO and LEMU, will be in charge of the transfer of capacities, faculties and resources. With that new scheme, and among other functions, the municipalities will have to: (i) undertake the hiring of teachers and administrative personnel; (ii) design and execute the Local Educational Project; (iii) conduct curriculum diversification; (iv) build, equip and maintain the infrastructure of the scholar 38 establishments within their jurisdiction; (v) execute the functions established in Article 82 of the Municipalities Organic Law; (vi) link the education objectives with the national, regional and local development plans; (vii) improve educational coverage, diminish drop- out and failing ratios and progressively increase the efficiency, inclusion, retention and quality ratios in their jurisdiction; and (viii) provide annual statements of educational management. 2.11 Overall, clarifications across the functions of the public sector are needed. The review of the assignment of responsibilities by level of government shows that many areas are not clearly defined. This can lead to potentially large losses in efficiency due to duplicity of functions as well as intergovernmental conflict. The only formal avenue for resolving conflicts regarding competences is recourse to the Constitutional Tribunal – which does not facilitate a speedy resolution of these issues. In practice, and in the absence of clear assignments, different sub-national governments have adopted different approaches regarding how much autonomy they want to exercise in individual functions. As we will see below, this has put strain on the regional directions, which are the points that connect the regional government and the central government. The regional directions are notionally part of the regional government but have traditionally been deconcentrated entities of the central government. The example of the education sector, as well as those of health and social programs reviewed in the Annex, suggest that considerable work still needs to be undertaken to more clearly define responsibilities though there is also promise that new approaches have finally been launched to resolve some of these issues. 39 B. Evolution of the transfer process 2.12 The Decentralization Law established a highly structured transfer process based on the capacity of regional governments and a sequenced approach. The law determined that the process will be based on a preparatory stage and four implementation stages: a first stage of decentralized governments’ installation, then the consolidation of the decentralization process and the formation of macro regions, after that the transfer and reception of sectoral competences (except for health and education), and finally the last stage is intended to be the transfer of functions in health and education. The accreditation system aimed to ensure that sub-national governments would be in a position to undertake the functions that were being transferred. This ambitious aim carried with it the development of transfers plans, underpinned by the criteria, tools, procedures and regulation necessary for the determination of whether sub-national governments were in a position to undertake the functions. In addition, a set of training modules and plans to build capacity were devised. Reflecting the complexity of the undertaking, the Accreditation Law was only passed by Congress in July 2004 and the accreditation system was therefore slow to start. 2.13 Transfers of functions were supposed to take place as part of Annual Plans once sub-national governments accredited their capacity to perform those functions. The process was to be launched with Annual Plans issued by the Presidency of the Council of Ministers based on inputs from different sectors. The annual plans were to detail the functions that would be transferred; the resources linked to those functions, including human resources, physical assets, as well as budgetary resources required; the requirements that each sub-national government would need to meet to be accredited for the transfer of the function; and a detailed timeline of the process. In addition to the annual transfer plans, the legal framework formulated an obligation for each sector to propose medium term transfer plans and also multi-annual plans to set out the transfer for five-year periods. 2.14 While awaiting the Accreditation Law, a temporary verification system was implemented in 2004. Even before the accreditation system was formally in place, a temporary “verification� mechanism was set up to purportedly ensure that the functions were transferred to entities that were in a condition to take over those functions. With the approval and promulgation of the Local and Regional Governments Accreditation System Law in July 2004, a generalized system was created for the transfer of functions based on defined requirements aimed at ensuring the compatibility of the sub-national government’s internal structure with the required resources. However, because functions that were transferred through the verification system gave rise to “verification compromises,� in practice the two parallel systems have remained in existence, adding to the complexity of the issue. 2.15 Throughout 2005 and early 2006 there were no major advances in the transfer process. In late 2006 the authorities announced a “Decentralization Shock.� The so-called “Decentralization Shock� (formally via the Supreme Decree 068-2006- PCM) consists of 20 decentralizing measures, including the acceleration of the transfer of 185 sectoral functions established in the Organic Law of Regional Governments. In addition, and as discussed above, the measures also include the launch of pilot programs 40 for the Municipalization of the Education Management and Primary Health Assistance (programs that will be described in the annex to this document). Initially, this proposal established the end of the transfers on December 31, 2007 but then it expanded the term twice, through end-2009. 2.16 The transfer of functions has only rarely implied transfers of personnel and budgetary resources. The Decentralization Framework Law stated that any transfer of competences must be accompanied by the financial, technical, tangible assets and human resources directly linked to the transferred services to ensure their continuation and efficiency. In practice, this transfer of resources has simply not happened (Defensoría del Pueblo 2009, p. 166). Some line ministries argued that the transfer of assets from the CTARs to the regional governments in 2002 already fulfilled the required transfer of resources. This might be true for tangible assets and data archives, but it has not been the case for human resources or the budgetary resources to fund current expenditures. Only in exceptional cases, such as in the projects from the National Development Institute (INADE), have personnel been transferred. Line ministries have identified a total of 80,000 workers that would need to be transferred but, due to constraints, such as the legal requirement that the worker would agree on the transfer,– it has not been possible to complete any personnel transfers. As for budgetary resources, the “linking� of functions being transferred to budget line-items has proven to be particularly difficult and has complicated the process, as discussed in Chapter 3 below. 2.17 The accreditation did not ensure the capacity of sub-national governments to effectively perform the transferred functions. The accreditation system was limited to verifying that certain pre-requisites were met for the transfers and did not verify whether sub-national governments were effectively able to carry out those functions. Moreover, since 2007, transfers of functions have taken place even in the absence of evidence that the sub-national authorities meet the expected requirements. This new modality of accreditation is known as “[capacity needs] to be further developed� (por potenciar) whereby the function is still accredited for being devolved provided that the central and regional governments subscribe to some cooperation agreement that will help build the institutional capacity that is said to be still lacking at the sub-national level. Peru’s ombudsman office, Defensoría del Pueblo, produced a comprehensive assessment of transferred functions until February 2009 which suggests that of the total of 4,500 functions to be accredited (180 functions times 25 regional governments), 1,734 functions or 39 percent, had been accredited on a ‘por potenciar’ basis. 2.18 Overall, the administrative costs of compliance with the complex transfer and accreditation system have been high while its benefits have been limited. 41 C. Coordination among levels of government 2.19 No institutional framework can hope to assign every possible function of the state to the smallest detail. Therefore, coordination between levels of government is important. The international experience suggests that no country has ever devised the perfect assignment of functions across levels of government at all times. In reality, the assignment of functions will be inevitably messy to some extent. Therefore, of critical importance is how the different levels of government coordinate with each other on a day-to-day basis to solve those problems. It is to these coordination mechanisms that we now turn our attention. 2.20 The main ‘contact point’ between the central and regional governments daily activities are the so-called regional directions. The regional directions of each sectoral ministry were originally the de-concentrated agencies of the corresponding sector. However, the decentralization process set out to convert these into dependent units of the Regional Management Offices of the regional governments. 2.21 The regional directions have become somewhat more integrated into the regional governments since regional presidents have been able to designate the regional directors. One of the measures of the October 2006 Decentralization Shock was to give regional presidents the authority to name the regional directors. Still, frictions are almost inevitable for a number of reasons. First, designating the regional director does not change the fact that the staffers of the regional direction have traditionally seen themselves as staff of the respective line ministries. Second, the regional government also has a management office (gerencia regional) for a given sector which has its own set of appointees that would be jockeying for power and competing with the regional director. For example, the regional manager for social development would be directly competing with the regional directors for health and education. All in all, the designation of regional directors has not changed the level of autonomy of the regional governments in any significant way. 2.22 However, the regional directions still act largely as deconcentrated bodies of the central government, which sets their budget. It is most significant that the budget assigned to individual regional directions is set and approved by the sector and by the National Direction of Public Budget of the Ministry of Economy and Finance. The Regional Government has no competence whatsoever to modify the budget allocation to the regional direction in any way. Thus, the regional directions are administratively dependent on the regional government but functionally dependent on the central government. 2.23 There are also difficulties stemming from differences in how the central and regional governments are organized by sectors. Another factor that complicates the day-to-day coordination between the central and regional governments is the lack of a counterpart between the management offices of the regional government and the line ministries and their regional directions. 42 Figure 2.1: Regional government organization chart Source: Compilation by World Bank staff. 2.24 The mismatch in the organizational structures of the central and regional government does not facilitate coordination. As shown in Figure 2.1 the regional governments’ organizational structure, set by law, includes a total of five regional management offices: economic development; social development; infrastructure; natural resources and environment; and planning, budget and territorial conditions. However, the sectoral structure from the line ministries does not match this, which complicates coordination as one single office in the regional government is supposed to be closely coordinated with a large number of central agencies. The economic development regional management office, for example, would need to coordinate with at least four ministries: agriculture, tourism and trade, production, and mining and energy, which appears to be a heavy burden and is unlikely to create effective relationships. 2.25 There are almost no mechanisms of coordination between the regional and local governments. In many instances these government levels not only lack coordination but they actively compete with each other. A common situation is for regional governments to carry out infrastructure works that correspond to local governments. In principle the Regional Coordination Councils may be a good space to coordinate the activities and policies, but in practice they have been little more than a formality in many cases. 43 2.26 The efficiency losses from this lack of coordination between levels of government can be high. Limited coordination can result in duplication of public services being offered. As an example, both the Callao municipality and the regional government of Callao provide nightly security patrols in the streets (serenazgo). The case of the multiple serenazgo services in Callao is particularly noteworthy since both entities share the exact same territory over which they have jurisdiction (Vega 2008). Without entering into the specifics of any given public service being provided in Peru, it is important to recognize that budgetary resources may be diverted to redundant tasks or to activities that may not be meeting the needs of the population. 44 3. Expenditure patterns This chapter analyzes how spending by different levels of government has evolved in recent years. One contribution of the chapter is to undertake an estimation of the extent to which spending by regional governments is truly decentralized or deconcentrated. Finally, since decentralization in Peru has been in part motivated by the goal of reducing regional disparities, the chapter also reviews the uneven distribution of total public spending across the territory. A particular contribution of this chapter is to provide local spending on a per capita basis across all local governments – an analysis that has not been presented in the literature before. A. Spending by different levels of government 3.1 A data gathering effort was launched to analyze public spending by level of government. The analysis below is the result of combining two data sources. The first source of information is the Integrated Financial Administration System (SIAF, Sistema Integrado de Administración Financiera) which provides detailed information on expenditures by the central government (since 1999), the regional governments (since 2004), and local governments (since 2008). For information on local governments prior to 2008 we obtained data from the National Direction of Public Accounting (DNCP, Dirección Nacional de Contabilidad Pública), to which all of the state’s entities (including municipalities) must report their accounts. This allows us to have data on budget expenditure, including the classification of expenditures by function, for more than 1,830 municipalities since 2004.8 The choice of 2004 as the starting point of our analysis responds to the availability in SIAF of data on regional governments’ expenditure since that year. 3.2 All levels of government have increased their spending sharply. Expenditures have increased sharply and total public spending including all levels of government rose by 49 percent from 2004 to 2008. Of the three levels of government, it is the central government that has seen the lowest increase: 32 percent, while regional governments’ expenditures increased by 64 percent and that of local governments more than doubled, increasing by 129 percent. Notably, as the period 2004-2008 was one of fast economic growth, averaging at an annual rate of 7.6 percent, the overall share of public spending on output remained roughly constant dropping slightly from 20 to 19 percent of GDP. 3.3 Sub-national governments have increased their share in total spending but mainly due to the rise in local government spending, while the share of regional governments has remained stable. Table 3.1 below provides a summary of the patterns of actual expenditure for the three levels of government on the basis of accrued amounts. 8 The analysis based on the DNCP data for local governments back to 2004 draws from Casas (2009). We would also like to thank the staff of MEF for facilitating the actual SIAF database with all records from 2004 to 2008 and for their help in addressing our queries about the contents of the database. 45 The share of the central government decreased from 72 to 64 percent from 2004 to 2008. The corresponding increase in the share of sub-national spending has been explained largely by the increase in the role of local governments, and not by regional governments. The limited increase in the share of regional governments showed by these figures contrasts with a relatively widespread perception to the contrary, a perception that is rooted in the fact that regional governments have a higher role in so far as public investment is concerned, an issue that attracts constant media attention. As for why has the share of local governments increased substantially, we will see below that the system by which revenues are earmarked for local governments helps to explain most of the difference. In fact, very little of the increased spending by either local or regional governments has been the result of resources being transferred in line with devolved functions. Table 3.1: Public spending by level of government, 2004-2008 2004 2005 2006 2007 2008 In millions of nuevos soles unless otherwise noted Central 34,421 36,070 39,938 39,650 45,537 Regional 7,849 8,987 9,963 11,571 12,879 Local 5,732 6,055 7,858 6,858 13,162 Total 48,002 51,113 57,758 58,079 71,578 As percent of GDP 20.2 19.5 19.1 17.3 19.0 Percent of total Central 72 71 69 68 64 Regional 16 18 17 20 18 Local 12 12 14 12 18 Total 100 100 100 100 100 Note: Reflects amounts actually accrued (devengado). These figures are in current terms since, given the low inflation of the period, the analysis based on nominal figures is deemed to be more transparent and still informative about trends in real terms. Source: Financial administration system (SIAF), and National Directorate of Public Accounts (Dirección Nacional de Contabilidad Pública), Ministry of Economy and Finance. 3.4 The limited increase in the share of spending by regional governments is explained by the fact that the transfer of functions implied limited transfers of budgetary resources. In principle the legal framework in place envisages that functions transferred would be done with all the resources linked to those functions, including human resources, physical assets, as well as budgetary resources. In practice, however, the transfer of resources has been limited for a variety of reasons. First, in many cases the sectors pointed out that physical assets and documentary archives were already transferred to the regional governments by virtue of the fact that the deconcentrated offices of the sectoral ministries, the regional directions, were formally absorbed by the newly formed regional governments 46 (Defensoría del Pueblo 2009). Second, many of the transferred functions were not well enough defined to be able to identify the actual budgetary resources necessary to perform them. The Ministry of Economy and Finance, through its National Direction of the Public Budget, established certain criteria to match functions to budget lines but in practice these have not helped to resolve the issues involved. The lack of clear rules has caused the transfer of budgetary resources linked to functions to have been determined by the outcome of bargaining between the decision-makers and by the political will of the senior management in each ministry. This situation has aggravated the grievances of regional government and central authorities alike. As part of the follow up work from the “Decentralization Shock� it was estimated that the transfer of the 180 sectoral functions to all of the 25 Regional Governments (Metropolitan Lima not included), would imply a transfer of resources of around 2.6 billion nuevos soles. While this is no small amount, it is important to remember that in 2008 the overall expenditure of regional governments (excluding Metropolitan Lima) was 12.8 billion nuevos soles (out of a planned budget of 16.3 billion nuevos soles). 3.5 Around two-thirds of public investment in 2008 was undertaken by sub- national governments. As shown in Figure 3.1 below, in 2008 local governments undertook 7.4 billion out of 16.3 billion nuevos soles of total public investment spending, while regional governments accounted for a further 3.3 billion nuevos soles. The budgets of sub-national governments are in fact heavily tilted towards capital expenditures: 67 percent of local governments’ budgets are programmed for public investment expenditures while the equivalent figure for regional governments is 39 percent (although the share of investment expenditures is somewhat lower when we consider actual expenditures as accrued instead). Figure 3.1: The importance of public investment by different levels of government Breakdown of public investment Share of investment in the budget (2008, billions of nuevos soles) (2008, percent of total budget) Local Regional Central Central Regional Local 35 70 67 30 60 56 25 50 13.2 39 20 40 15 30 26 6.3 7.4 10 18 20 12 3.3 5 9.8 10 5.6 0 0 Budgeted Spent Budgeted Spent 47 Notes: Investment expenditures are taken as the sum of “investments� and “other capital expenditures� but excludes “financial investments.� Budgeted amounts are those of the “Modified Institutional Budget� (Presupuesto Institucional Modificado, PIM) while spent amounts refer to accruals (devengado). Source: MEF (SIAF). 3.6 The heavy emphasis on capital expenditures at the sub-national level can limit their control over the quality of public service delivery. Current expenditures are key components in the provision of social services; for example in health and education the quality of the services provided are influenced by elements such as the salaries of teachers and medical doctors, inter alia. The heavy focus on capital expenditures by sub- national governments could therefore have a detrimental effect on public sector efficiency. While this issue falls outside the scope of this report, the question arises as to what would be an appropriate mix of central vs. sub-national capital expenditures. 3.7 The focus on investment also helps to explain the lower budgetary disbursement rates of sub-national governments. A common topic of discussion in the public debate is the poor performance of sub-national governments in disbursing their budgets. This is typically taken as a sure indicator that those sub-national governments lack the capacity to perform their assigned functions. Some would consider this an argument for re-centralizing. It is true that the overall disbursement rate of sub-national governments is lower than that of the central government (see left panel in Figure 3.2 below). However, once we limit our analysis to public investment projects, there is no such “disbursement gap� between levels of government (right panel in Figure 3.2). The overall differences in disbursement rates can be fully explained by the fact that capital expenditures, which are more difficult to execute and are more prone to accumulating delays, account for a higher share of the budget of sub-national governments. Thus, leaving aside the point of whether disbursement rate is a good proxy for institutional capacity or not, this evidence could hardly be interpreted as supporting a recentralization effort on the grounds of limited ability to spend by sub-national governments. 3.8 Sub-national governments formally account for more than half of education and health public expenditure. The functional classification of expenditure by level of government, shown in Table 3.2., reveals that the sub-national levels of government account for relatively high shares of spending in key functions such as education or health. In particular, regional governments account for more than half of education spending and more than a quarter of health expenditures. However, important caveats apply when interpreting these data. In fact, as Table 3.2 shows, regional governments already had a high share of spending in those sectors back in 2004, well before any attempts at decentralizing those two sectors. The reason why regional governments appear to implement most of those expenditures is because the deconcentrated offices of the line ministries: the regional directions, are counted in the column of regional governments. However, as we saw in Chapter 2, the regional governments have limited control over those regional directions and no autonomy over the use of their budgets. As a result, it is more accurate to classify expenditures by the regional directions as deconcentrated instead of decentralized. A detailed exercise to sort out what is truly decentralized versus what is deconcentrated is provided in the next section. 48 Figure 3.2: Disbursement rates for public investment by level of government Budget execution rate in 2008 Public investment by regional by level of government governments, 100 Central Regional Local billions of nuevos soles 83 7 79 Budgeted 80 6 66 Disbursed 57 5 60 52 56 4 40 3 20 2 1 0 0 Overall budget Public investment 2005 2006 2007 2008 projects Source: MEF (SIAF). Table 3.2: Public spending by function and level of government, 2004 and 2008 Percent of spending in a given function accounted for by different levels of government 2004 2008 Central Regional Local Central Regional Local Admin. and planning 81 2 17 73 4 23 Agriculture 63 30 7 36 35 29 Communications 99 1 – 99 1 – Defense and nat. security 100 – – 99 1 Education and culture 41 55 4 37 53 11 Energy and mining 70 12 18 87 6 7 Fisheries 74 23 3 61 27 13 Foreign affairs 100 – – 100 – – Health and sanitation 56 27 17 47 29 24 Industry, com. & services 51 9 40 37 8 55 Justice 100 – – 100 – – Labor 95 5 – 93 7 – Legislative 100 – – 100 – – Social protection 77 12 11 74 13 13 Transportation 54 13 32 39 15 46 Urban dev. & housing 31 6 63 47 3 50 Total 72 16 12 64 18 18 Note: Reflects amounts as per the modified budget (Presupuesto Institucional Modificado). Source: World Bank staff calculations based on MEF (SIAF and National Directorate of Public Accounts, Dirección Nacional de Contabilidad Pública). 49 B. Deconcentration vs. decentralization of expenditures 3.9 The share of deconcentrated spending (by regional directions of line ministries) has decreased over time but it remains above two-thirds of the spending formally classified as being undertaken by the regional government. Information on the spending of regional directions is available both from SIAF and from the Dirección Nacional de Contabilidad Pública. Figure 3.3 shows the results of splitting the spending among the regional directions in all sectors, leaving only the purely decentralized spending over which the regional governments have autonomy. A word of caution about the exercise is also called for as it relies on the classifications available within SIAF. Stating that regional governments have absolutely no autonomy over deconcentrated spending would be going a step too far as it is possible, and does happen, that certain regional directions may work very closely together with the regional government authorities and the latter may be effective at exercising de facto control over decisions on the ground and on spending. Still, the fact that the budget of regional directions is determined centrally by the Ministry of Economy and Finance suggests that these types of expenditures ought to be considered as deconcentrated. Figure 3.3: Breakdown of regional governments’ deconcentrated and decentralized spending over time Percent and billions of nuevos soles Deconcentrated Decentralized 100% S/.1.2 S/.1.5 S/.1.9 S/.2.4 S/.3.5 S/.4.3 80% 60% S/.5.5 S/.6.4 S/.7.1 S/.7.6 S/.8.1 40% S/.8.6 20% 0% 2003 2004 2005 2006 2007 2008 Source: World Bank staff calculations based on MEF (SIAF), INEI. 3.10 At the regional level, there is limited room to modify or increase the amount of the budget that is deconcentrated. To a large extent, the share of deconcentrated spending remains high since current expenditures like teachers’ salaries are channeled through the regional directions. As a result, regional authorities have taken to lobbying the executive directly in order to increase their say over public spending in their departments. The consequence has been relatively lengthy and ad hoc trilateral negotiations between a given regional authority, the Ministry of Economy and Finance, and the line ministry involved over transfers of functions and resources without clear rules of the game. The different willingness of line ministries to give up control helps to 50 explain why different sectors may have a higher or lower degree of deconcentrated spending (Table 3.3), and may also explain some of the differences across regional governments (see Figure 3.4 below). Table 3.3: Deconcentrated and decentralized spending, 2004 and 2008 Millions of nuevos soles 2004 2008 Regional governments Deconcentrated Decentralized Deconcentrated Decentralized by function Administration and planning – 293.2 2.0 606.1 Agriculture 71.9 219.9 150.3 549.2 Communications 0.3 0.4 0.7 0.0 Defense and national security 0.6 6.9 2.0 84.5 Education and culture 4,141.3 130.4 5,396.7 761.1 Energy and mining – 37.0 – 169.0 Fisheries – 22.5 – 47.4 Health and sanitation 898.2 336.4 1,515.8 1,052.5 Industry, commerce & services – 24.9 0.3 48.7 Justice – 0.1 – 2.6 Labor – 14.9 – 20.2 Social protection and pensions 1,207.9 103.5 1,314.4 178.7 Transportation 60.9 249.7 132.9 724.8 Urban development & housing 1.4 16.8 0.9 23.3 Total (regional govts.) 6,382.4 1,456.5 8,515.9 4,268.1 Percent of total 81 19 67 33 Total public spending (all levels 48,002 71,578 of government) Of which Decentralized 7,189 17,430 spending* (as percent of total) 15 24 Notes: Deconcentrated spending reflects the expenditure accounted for by the regional directions (direcciones regionales) of the sectoral ministries, while decentralized spending reflects the amounts spent by the main office of the regional governments (sede central). The sum of these functions is slightly lower than the numbers reported in Table 3.1 since not all functions are reported here (functions where the central government undertakes the vast majority of spending were not analyzed). * Decentralized spending in this calculation includes spending by local governments plus the spending undertaken by the regional governments’ main offices. Source: World Bank staff calculations based on MEF (SIAF and National Directorate of Public Accounts, Dirección Nacional de Contabilidad Pública). 51 3.11 Overall, excluding deconcentrated expenditures, the share of sub-national spending is estimated at 24 percent of total public expenditure (36 percent if we include deconcentrated spending, see Tables 3.3 and 3.1). Figure 3.4: Deconcentrated and decentralized spending by regional government Percent of regional budget, 2008 Deconcentrated Decentralized 100 15 18 20 27 27 28 29 29 31 31 36 37 38 38 40 41 80 41 43 43 46 47 51 57 60 40 20 0 Madre De… Tumbes Puno Tacna Pasco Ancash Cusco Cajamarca Junin Piura Ica Ayacucho La Libertad Arequipa Lambayeque Huancavelica Ucayali Huanuco Moquegua Loreto San Martin Apurimac Source: World Bank staff calculations based on MEF (SIAF). Amazonas 3.12 There is a strong link between the canon system and the extent to which regional spending is de-concentrated or decentralized. Some of the larger recipients of natural resource-related revenues (canon, royalties and alike) have the lowest degree of deconcentrated spending. Tacna and Moquegua, which are large recipients of canon from their mining activities, have the largest share of regional expenditure under the responsibility of their regional governments (54 percent on average). A similar case is that of Cusco, a larger recipient of canon from the gas sector. In these cases, the level of autonomy has increased considerably since 2004, when deconcentrated expenditures exceeded 70 percent in all cases. This suggests that much of the observed reduction in the share of deconcentrated expenditures has been the result of the swelling of natural resource-related revenues earmarked to sub-national governments. 3.13 Moreover, sub-national governments face additional restrictions on the natural resource-related funds that are directly channeled to them. Natural resource- related revenues are earmarked for capital expenditure (details are provided in Chapter 4 below). As a result, sub-national governments have a limited and indirect influence over public service delivery. Regional governments claim that they have little impact on the improvement of the quality provision of the services for their citizens. Critical services like health and education, which are intensive in human resources and other current expenditures, remain therefore in the realm of regional directions: in 2008 the regional directions controlled 88 percent of all the regional expenditure in education and in social protection, and 60 percent in health; these three functions concentrated 97 percent of the total spending by the regional directions. 52 C. Public spending and regional disparities 3.14 The reduction of regional disparities is one of the overall goals of the decentralization process in Peru. For this reason this section pays closer attention to how public spending is distributed across the territory. We first discuss the combined expenditure of all levels of government by department and present the data on a per capita basis. Data for the expenditure of the central government at the departmental level is sufficiently reliable, though some questions arise as to whether expenditure in Lima may reflect some purchases of goods subsequently delivered elsewhere. The analysis then moves on to the district level, where we present expenditure only for the local government (aggregating all levels of government expenditure by district is not possible since matching central government data to districts is not entirely accurate). Expenditure by all levels of government across departments 3.15 The distribution of public expenditure reinforces the inequality gap across departments. The top five departments with the largest expenditure per capita (Moquegua, Lima, Tacna, Madre de Dios, and Pasco), are among the ones with the highest GDP per capita and, excluding Pasco, have poverty rates below the national average. These departments receive on average three times more than the five departments with the lowest per capita expenditure (Piura, Cajamarca, Lambayeque, La Libertad and Puno). Notably, besides Pasco no other departments with higher than average poverty rates receive more than the country average in expenditure (i.e., the top right quadrant in Figure 3.5 is mostly empty). Figure 3.5: Public expenditure per capita and poverty across departments In current soles 6,000 Country average Moquegua 5,000 Lima Public expenditure 4,000 per capita in 2008 Tacna (all levels of 3,000 Madre de Dios Pasco Country average government) in Tumbes current soles, by 2,000 Arequipa Ayacucho Huancavelica department Ica Apurimac 1,000 Cajamarca Huanuco 0 0 20 40 60 80 100 Poverty rates in 2008 by department, percent Source: World Bank staff calculations based on MEF (SIAF), INEI. 53 3.16 The three levels of government spending reinforce the unequal distribution of public expenditure. All three levels of government exhibit some bias towards spending more on a per capita basis in richer departments. For example, Moquegua and Tumbes are among the top ten departments with the largest per capita expenditure at every level of government. Five other departments of the top ten recipients of total per capita public funds are among the largest recipients by at least two levels of government. For example, Tacna and Pasco are both among the top ten departments with the largest per capita expenditure by the local and the regional governments. Figure 3.6: Total public spending per capita by department and level of government Nuevos soles, 2008 MOQUEGUA 3,126 1,419 659 = 5,204 LIMA 362 69 4,065 = 4,495 TACNA 2,015 1,192 467 = 3,674 MADRE DE DIOS 311 1,155 1,347 = 2,814 PASCO 1,262 923 592 = 2,778 TUMBES 594 1,011 759 = 2,364 ANCASH 1,159 720 482 = 2,361 CUSCO 1,017 729 612 = 2,358 AMAZONAS 338 815 1,077 = 2,230 UCAYALI 430 758 738 = 1,926 HUANCAVELICA 638 855 430 = 1,923 AYACUCHO 489 794 611 = 1,894 APURIMAC 427 890 493 = 1,809 AREQUIPA 430 692 680 = 1,802 ICA 338 594 781 = 1,712 SAN MARTIN 231 800 629 = 1,660 LORETO 425 805 401 = 1,631 P.CONS.DEL… 442 458 616 = 1,516 JUNIN 395 565 552 = 1,512 HUANUCO 318 536 601 = 1,455 PUNO 425 555 448 = 1,427 LA LIBERTAD 348 535 488 = 1,371 LAMBAYEQUE 256 467 646 = 1,369 CAJAMARCA 414 542 359 = 1,315 PIURA 390 464 413 = 1,267 - 1,000 2,000 3,000 4,000 5,000 Local government* Regional government National government Notes: * Local government reflects the sum of the expenditure of all local governments situated in a given department. Source: World Bank staff calculations based on MEF (SIAF), INEI. 54 3.17 The level of government that accounts for the largest share of expenditure varies by function. Of the three sectors that account for the largest per capita expenditure: education, social protection, and health, it is important to note that the regional government dominates spending in education, while the central government does so in social protection and health and sanitation. The local level of government accounts for the largest share of spending in transportation, urban development and housing, and industry, commerce, and services (Table 3.4). Table 3.4: Public spending by level of government and by function Data refer to 2008 Total expenditure by government level (percent) Expenditure Regional governments per capita Local Regional Central (nuevos soles) by function Administration and planning 23 4 73 630 Education and culture 11 53 37 430 Social protection and pensions 13 13 74 423 Health and sanitation 24 29 47 323 Defense and national security – 1 99 246 Transportation 46 15 39 211 Energy and mining 7 6 87 107 Agriculture 29 35 36 74 Justice – – 100 63 Urban development & housing 50 3 47 31 Industry, commerce & services 55 8 37 22 Foreign relations – – 100 19 Legislative – – 100 9 Fisheries 13 27 61 6 Labor – 17 83 4 Communications – 1 99 3 Source: World Bank staff calculations based on MEF (SIAF), INEI. 3.18 The varied importance of different sectors across functions may reflect sector-specific characteristics but may also be driven by bureaucratic struggles. It is possible, of course, that these patterns reflect that optimal allocations between levels of government vary by sector. But the process discussed in Chapter 2 about how negotiations for the transfer of functions take place in an environment of relatively unclear rules suggests otherwise. In any case, this appears to be a significant finding that may merit further investigation. While a thorough take on this question is outside the scope of this report, there is some further evidence to support the view that bureaucratic struggles drive these variations in the importance of levels of government in different functions. In particular, a look across pairs of similar departments suggests relatively large differences in the role played by one or another level of government in different 55 departments for the same function. Table 3.5 below shows these results for one such pair of similar departments (Apurímac and Huancavelica, the two poorest departments in the country, both in the Southern Andean Sierra, and with limited revenues from natural resources). The data for all other departments and by some expenditure functions is given in Annex C, basic data as provided on the top panel of Table 3.5 for all departments is shown in Annex A. The results of this exercise suggest that in Apurímac the regional government plays a relatively greater role in both education and health than it does in Huancavelica (where local governments play a relatively bigger role). Table 3.5: Comparison of public spending by level of government and by function, Apurimac and Huancavelica Basic data Population Population GDP as GDP per Average Poverty rate (2007) density share in capita real GDP (2008, percent) (2007) national (2008, growth in GDP thousands 2004-08 (per cent) of current (percent) nuevos soles) Apurímac 404,190 19.3 0.4 3.3 5.5 69 Huancavelica 454,797 20.5 0.7 6.1 3.3 82 Per capita public spending by level of government, nuevos soles Education Health Local Regional Central Local Regional Central Apurímac 46 441 89 49 241 172 Huancavelica 104 415 96 73 184 164 Source: World Bank staff calculations based on MEF (SIAF), INEI. 56 Spending by local governments only – analysis of 1,834 municipalities 3.19 Compared to other levels of government, local governments spend relatively more on sectors like transportation. At the same time, social sectors still account for a substantial amount of spending. Notably, there is relatively limited dispersion on the share of expenditure devoted to social sectors: around half of the departments allocated between 25 percent and 45 percent of their local expenditure to social sectors. Figure 3.7: Local public expenditures by function Local expenditure by all districts according Social expenditure (Health, Social protection to functional classification, 2008 and Education) as a share of total expenditure Education & 100 Number of departments Culture 10% 80 Social Administrat. Protection & Planning 60 12% 31% 40 Health 16% 20 Agriculture 4% Energy & Transportati Mining 0 on 2% 1% 7% 13% 20% 26% 32% 39% 45% 51% 58% 64% 70% 77% 83% More 20% Industry, Housing & Trade & Urban Dev. Fishing 3% 0% Services Social expenditure as a share of total 2% 3.20 However, the allocation by local governments to capital or current expenditures varies substantially and is related to their main sources of revenues. Districts where a larger share of their resources came from natural resources spent a greater share of their resources in capital expenditure; likewise, districts where a larger share of their resources came from FONCOMUN spend a greater share of their resources in current expenditure. Figure 3.8: Local public capital and current expenditures and main revenue sources 100% 100% Current expenditure as share of Capital expenditures as share of 80% 80% 60% 60% total total 40% 40% 20% 20% 0% 0% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Canon, royalties, custom income and particip. as a share of total resources FONCOMUN as a share of total resources 57 58 4. Assignment of revenue This brief chapter reviews both the assignment of tax bases as well as the main sources of revenue and the mechanisms by which revenues from natural resources are channeled directly to sub-national governments. 4.1 Only the central and local levels of government are owners of a particular tax base. By design, sub-national governments are almost exclusively dependent on transfers from the central government. Even in the case of local governments, which,have local taxes (see Table 4.1), all decisions regarding rates or bases are set by Congress at the center. For example, local governments lack the discretion to modify the rates on property taxes. In that sense, such revenues are closer in concept to shared revenues (with a 100 percent share for the local government) than own-source taxes (Ahmad and García- Escribano 2009). Local governments receive the so-called Municipal Promotion Tax (Impuesto de Promoción Municipal), which is collected through 2 percentage points out of the 19 percent of VAT. It is considered a central government transfer to the municipalities in the form of the Municipal Compensation Fund (FONCOMUN). Table 4.1: Revenue sources of sub-national governments Resource Local government Regional government Taxes and fees Mainly property tax No taxes. (see table 4.2 below) Can collect fees Borrowing With restrictions With restrictions Transfers FONCOMUN (Municipal Natural resource-related (pre-determined and compensation fund) equivalent to 2 transfers (cánon, see table earmarked) percentage points of the VAT collected 4.3 for rules of in the department, distributed using a distribution) formula that considers poverty and unmet needs Natural resource-related transfers (cánon, see table 4.3 for rules of distribution) Transfers Yes, but small FONCOR (Regional (discretionary by the compensation fund), central govt) others Source: Ahmed and García-Escribano (2008). 3.21 For local governments the formula-driven transfers (FONCOMUN and natural resources-related) account for more than half of their revenues. Three sources of financing covered 75% of the total local expenditure for all districts in 2008. 59 These were: ‘Canon, royalties, custom income, and participations’ (37%); the ‘Municipal compensation fund, FONCOMUN’ (24%); and, ‘Revenues directly collected’ (14%). In nearly one-half of the departments, more than 40% of their resources came from ‘Canon, royalties, custom income, and participations’. Figure 4.1: Sources of revenue of sub-national governments 14,000 244 81 12,000 384 Borrowing 3,000 1,975 10,000 Local taxes and fees 8,000 6,000 Transfers from CG 8,024 10,432 (formula driven) 4,000 Transfers from CG 2,000 (discretionary) 1,894 - Local Government Regional Governemnt Source: SIAF - MEF. Source: MEF. 4.2 Even though local governments have taxes, collections are small. The main local tax is the property tax (predial) which, as shown in Table 4.1 below, accounts for around 51 percent of all local tax revenues. 9 This is followed by the real estate transfer tax (alcabala), which accounts for 32 percent. However, the importance of these property taxes is only high in a small number of municipalities. According to the latest National Registry of Municipalities (RENAMU, 2008), only 364 local governments reported having a cadastre. In the aggregate, all local taxes account for around 13 percent of the current revenues of local governments. Among the three levels of government in Peru, only the central and the local levels are owners of a particular tax base. It is important to mention that in 2007, the Foncomun transfer was about S/. 2.8 billion, more than twice the S/. 1.1 billion raised through local taxes. 4.3 In contrast, natural resource-related revenues are highly important for local governments. As shown in Figure 4.2, on aggregate natural resource-related revenues account for more than a third of local government revenues. In fact there are more than 330 local governments for which natural resource-related revenues account for 70 percent or more of their total revenues. The importance of natural resource-related revenues is still significant for regional governments, though to a lesser degree. For regional governments it is the flow of so-called ‘ordinary resources’ from the central government that accounts for the greatest share of revenues. 9 For a discussion on the local property tax see Alfaro and Rühling (2007). 60 Table 4.2: Local tax revenues 2007 Share of Share of Share of local local Thousands local tax govt. govt. of nuevos soles revenues current total (percent) revenues revenues (percent) (percent) Production and consumption 85,173 7.9 1 0.6 Gambling 715 0.1 0.0 0.0 Public spectacles (excl. sports) 16,611 1.5 0.2 0.1 Games 6,626 0.6 0.1 0.0 Slot machines 55,503 5.2 0.7 0.4 Casinos 5,717 0.5 0.1 0.0 Property 965,759 89.6 11.8 7.2 Vehicles 66,986 6.2 0.8 0.5 Real estate transfer (alcabala) 346,572 32.2 4.2 2.6 Real estate tax (predial) 552,201 51.2 6.8 4.1 Other 26,712 2.5 0.3 0.2 Total local tax revenues 1,077,644 100 13.2 8.0 Source: MEF. 4.4 In fact, the increase in natural resource revenues goes a long way to explain the rise in the share of local government in total expenditure. As we saw in Chapter 3, one of the stylized facts of the recent period has been the rise in the share of local governments in total public expenditure, up from 12 percent in 2004 to 18 percent in 2008. This sharp increase was in contrast with the less prominent increase by the regional governments, which went from 16 percent to 18 percent of total public expenditure during the same period. As shown in Figure 4.3, the stronger increase displayed by local governments can be explained by the behavior of two formula-driven transfers that local governments receive: the FONCOMUN (2 percentage points of VAT and hence increasing fast in a period of very fast economic growth) and the natural resource-related revenues. 4.5 The reason why the natural resource boom has favored the local level of government is due to the particulars of the way the canon is allocated. The rules of dividing the canon benefit the local level of government versus the regional government in a ratio of approximately 3 to 1 (see Table 4.3 below). In addition, the main compensation fund for regional governments, the FONCOR, is not based on the same formula as the FONCOMUN. Thus, the increased municipalization of public expenditure has been the unintended consequence of a transfer system designed under quite different conditions of commodity prices and tax collections from VAT. Thesystem has also contributed to the much discussed issue of rising inequities among municipalities. 61 Figure 4.2: Local and regional public spending by revenue source (a) Local governments Source of financing of local public Canon, royalties, custom income and participations expenditure in 2008 as share of total total resources received in 2008 300 249 Ordinary Official 250 237 243 235 Sources credits Nuber of departments 2% 203 Grants and 6% 200 174 transfers 161 8% 150 128 Canon, 103 Municipal royalties, 100 taxes customs, & 66 9% participat. 34 37% 50 Revenues directly Municipal 0 collected compensati 14% on fund 1 10 20 30 40 50 60 70 80 90 100 24% Share of resources received (b) Regional governments Public spending of regional governments by revenue source, 2008 Source of financing of regional public Item Amount (Nuevos soles) Percent of total expenditure in 2008 Ordinary resources (transfers Revenues from central government for directly decentralized functions) Official 9,053,861,258 70.3 collected Grants and credits Ordinary resources (FONCOR) 673,725,000 5.2 3% transfers 1% Canon - forestry 1,236,244 0.0 5% Canon - gas 168,563,885 1.3 Canon & royalties -mining 1,145,895,175 8.9 Canon, Canon - fisheries 17,534,193 0.1 royalties, Canon - hydroenergy 35,577,867 0.3 customs, & Canon - oil 321,949,851 2.5 participat. Ordinary Camisea gas rent 37,018,730 0.3 15% Sources COFIDE rents 130,049,853 1.0 76% FONIPREL 14,348,573 0.1 Customs rent 152,860,112 1.2 SIS (Seguro Integral de Salud) 301581796 2.3 Other grants and transfers from central agencies 361,679,434 2.8 Revenues directly collected 381,649,653 3.0 Borrowing 80,586,040 0.6 62 Figure 4.3: Earmarked sources of revenue for sub-national governments Millions of nuevos soles Local governments Regional governments Foncomun revenues FONCOR revenues Natural resource-linked revenues Natural resource-linked revenues 14,000 14,000 Total expenditure Total expenditure 12,000 12,000 10,000 10,000 8,000 8,000 6,000 6,000 4,000 4,000 2,000 2,000 - - 2004 2005 2006 2007 2008 2004 2005 2006 2007 2008 Source: MEF (Portal de Transferencias and Dirección Nacional de Contabilidad Pública). Table 4.3: Rules by which canon is distributed Amount of canon is determined as 50 percent of corporate income tax paid by mining company Which is shared, in percentage terms, as follows District municipality where the resource is extracted 10 Municipalities of the province where the resource is extracted 25 Municipalities of the department where the resource is extracted 40 Regional government where the resource is extracted 20 Universities in the department where the resource is extracted 5 Note: In 1992 the canon was determined as 20 percent of corporate income tax paid by mining companies. The percentages shown in the table reflect those in place since December 2004.Source: Arellano-Yanguas (2008). 4.6 The mechanism to share natural resources has contributed to an increase regional disparities. The exercise shown in Figure 4.4 consists of identifying how much of the total spending of the three levels of government in a given department is financed through revenues linked to natural resources. The data is shown on a per capita basis. Figure 4.4 shows how most of the top ten departments in terms of public expenditure per capita are also the largest recipients of natural resource-related revenues. This result stems from a lack of compensation mechanisms. In other words, the rest of the public expenditure system is not compensating the large natural resource-related revenues with much lower revenue allocations under other mechanisms. 63 Figure 4.4: Public spending per capita by department and by revenue source All levels of government combined, in nuevos soles 6,000 Revenues not related to natural resources 5,000 Natural resource-related revenues 4,000 3,000 2,000 1,000 - Hunanuco Tacna Lima Pasco Puno Ancash Cusco Ica San Martin Cajamarca Piura Madre de Dios Ayacucho Moquegua Tumbes Ucayali Hunacavelica Arequipa Junin La Libertad Lambayeque Loreto Apurimac Amazonas 4.7 The call for greater redistributive efforts is common among civil society organizations but is a highly sensitive political issue. It is not uncommon to read proposals that emphasize the need to consider all sources of revenue at once before deciding how monies are to be allocated. For example calls for “substantive adjustments in the origin and equitable distribution of monies from the canon, sobre-canon, mining royalties, municipal and regional compensation funds, voluntary contributions from the mining sector, and the fund to promote regional public investment� (RED-Peru 2008, p. 13). Yet, so far any attempt at modifying the rules by which natural resource revenues are shared have been fiercely resisted, and, as discussed in Chapter 1, violent protests have not been uncommon. 4.8 Besides the issue of exacerbating regional disparities, the rules for spending natural resource-related revenues do not necessarily contribute to the efficiency of public spending. One issue that arises relates to the requirement that those resources be devoted to capital expenditure. While this may be a well-intentioned and viable option under scenarios in which these resources account for a smaller share of the budget, it appears impractical and not necessarily desirable when hundreds of local governments have in excess of 70 or 80 percent of revenues stemming from the canon or other natural resources (as shown in Figure 3.8 in Chapter 3 above). But the issue is not only one affecting local governments. In Figure 4.5 we have aggregated all local governments located in a given department and calculated how much of their resources are linked to natural resources. We then add the regional government of that particular department and how much of its revenues are linked to natural resources. Combining the two we have a measure of how much of the entire expenditure of sub-national levels of government (local or regional) in a given department is accounted for by natural resource-related 64 revenues. The results show that there are five departments in which natural resources contribute more than half of the entire sub-national spending in that department. Figure 4.5: Share of resources by regional and local governments accounted for by canon and other natural resource-related revenues 2008, percent 90 Regional and local governments 80 Regional governments 70 Local governments 60 50 40 30 26 20 10 0 Arequipa Ica Ayacucho Lima Ancash Tacna Pasco Piura Loreto Huancavelica La Libertad Puno Moquegua Cusco Cajamarca Junin Huanuco Tumbes Lambayeque Apurimac Total Ucayali Madre de Dios Amazonas San Martin Note: Includes all cánons, regalías, participaciones and renta de aduanas. Source: MEF (SIAF). 4.9 Finally, the volatility of commodity prices and the non-renewable nature of most of the natural resources being exploited call for moving toward long-term horizons in investment planning. It is true that, unlike regular budget resources, any unspent balances from natural resource-related revenues at the end of the year can be carried over to the following year’s budget. As such, there is no real binding constraint that would force a given local or regional authority to spend the resources in a given year. Yet, the fact that all revenues from natural resources to which a sub-national government is entitled in a given year are in fact incorporated in the annual budget creates a number of electoral and institutional incentives to try to spend the money as quickly as possible. There is of course the old-age temptation of politicians to build patronage networks and secure electoral support. In addition, the public discussion of these issues has focused so much on disbursement rates, which are implicitly equated with success, that there are many civil society efforts to ‘name and shame’ those sub-national authorities that are unable to spend. There are a number of voices from civil society organizations that have been calling for such mechanisms to ensure inter-generational equity and improved efficiency in spending (Grupo Propuesta Ciudadana 2007, 2009). Still, so far the discussion of this issue has been relegated to other issues perceived to be of greater short- term relevance, many related to measures on how to speed up disbursements now such as simplifying the national system of public investment (SNIP) or procurement rules. 65 66 5. Efficiency of public expenditure across departments One of the key goals of the decentralization process is better quality and efficiency of public service delivery. This section assesses the performance of public spending at province level and takes an initial step in the direction of identifying the determinants of government efficiency in Peru. It does so by using two different concepts to measure performance and efficiency, the Public Sector Performance index (PSP) and Public Sector Efficiency index (PSE) (originally proposed by Afonso, Schuknecht and Tanzi, 2005). The indices are confronted with different socio-economic indicators including indicators of fiscal decentralization. Results show that: (i) there are significant differences in government performance and efficiency across regions; (ii) efficiency seems to partially compensate for lower public spending, mitigating instead of accentuating disparities in performance; (iii) governments in wealthier regions perform better, but they are not necessarily more efficient. On the other hand, governments in densely populated areas are more efficient, but have no significant advantages in performance; (iv) while decentralization in spending has no apparent relationship with government performance, a larger share of locally collected revenues is associated with higher PSP and PSE scores; (v) finally, governments in natural resource-rich regions tend to outperform, but they have much lower efficiency scores. A. Introduction 5.1 The ongoing decentralization process strengthened the debate on quality and efficiency of public services delivery at sub-national level. Theory suggests that the transferring of well-defined responsibilities and resources to sub-national governments can increase the efficiency of public spending and improve the quality of public services. Decentralization, the argument goes, gives more autonomy to those levels of government that are closest to the citizens, and hence more informed of their needs and more accountable to them. It is of economic and policy interest to analyze the extent to which the theoretical predictions apply to the Peruvian decentralization experience. 5.2 In line with the current debate, this section assesses the performance of public spending at province level and takes an initial step in the direction of identifying the determinants of government efficiency in Peru. It does so by using two different concepts to measure performance and efficiency, the Public Sector Performance and Public Sector Expenditure indices (proposed by Afonso, Schuknecht and Tanzi, 2005). Public sector performance (PSP) is measured as a weighted average of a variety of outcome indicators related to public service delivery, while public sector efficiency (PSE) is measured as the total outcome generated by unit of public resource spend. The calculated indices are then confronted with different socio-economic indicators that relate to the government’s ability to generate results, including indicators of fiscal decentralization. The analysis focuses on three key sectors for public service delivery: Education, Health and Sanitation, and Transport. 67 5.3 The efficiency of public spending is receiving increasing attention from both policy and economic literature. The first round of works on public sector performance focuses on the analysis of national government. Works such Tanzi and Schuknecht (1997), Afonso, Schuknecht and Tanzi (2005, 2006), and Herrera and Pang (2005) compare public sector efficiency, measured by relating government expenditure to socio- economic indicators targeted by this spending, across a large set of countries. Motivated by the methodological shortcomings of cross-country comparisons, Bardhan and Mookherjee (2006) and Hauner (2008) develop the first analyses of government performance at sub-national level. The first study targets a specific local government program in India, while the second study compares government efficiency among Russian regions. To our knowledge, Herrera and Francke (2007) is the only work in the literature to assess public efficiency at sub-national level in Peru. The authors analyze the relative efficiency of Peruvian municipalities in 2003. 5.4 This chapter builds on the existing literature for Peru by constructing objective measures of government performance and by linking government outcomes to the ongoing process of decentralization. Preliminary evidence suggests the decentralization process has not been able to produce the expected results. Based on the national household surveys of 2004-2005, 70 percent of respondents thought that decentralization had done nothing or little to improve public services, while 72 percent expressed similar negative views about the ability of decentralization to pay more attention to citizens’ interests (Gonzáles and Shiva 2007). By formally measuring government’s performance and efficiency at sub-national level, this chapter will help to assess the decentralization program and whether the negative perception is justifiable. Unlike Herrera and Francke (2007), this analysis compares public sector efficiency across region and uses recent data (2004 to 2008), the latest observations date almost six years from the beginning of the decentralization process. 5.5 Five main findings stand out from the analysis: i) Government performance and efficiency differ significantly across regions. While average performance scores for the Education Sector corresponds to 75 percent of the top performance, the averages for Health and Sanitation and Transport are 59 percent and 25 percent of the top score, respectively. Given the large regional dispersion in public spending, differences become even larger for efficiency scores, average scores range from 66 percent of the top performer in Education to 15 percent in Transport. ii) Efficiency seems to partially compensate for lower public spending, mitigating instead of accentuating disparities in government outcomes. Public sector performance is determined by the amount of public resources and by how efficiently these resources are allocated. For all sectors public sector efficiency scores are negatively correlated to public spending, i.e., regions that spend less tend to be more efficient in allocating public resources. iii) Governments in wealthier regions perform better, while governments in densely populated area tend to be more efficient. For all sectors analyzed, per capita income is positively correlated with public sector performance, but it has no significant relationship with public sector efficiency. On the other hand, 68 population density is clearly linked to the ability to provide services efficiently, but it shows no direct effect on government performance. iv) While decentralization in spending has no apparent relationship with government performance, larger shares of locally collected revenues are associated with higher PSP and PSE scores. Local and regional levels of spending in Education and Health and Sanitation have zero correlation with PSP and PSE scores. This relationship is positive but small in the Transport sector. The opposite is true for revenues. Locally collected revenues are positively associated with high quality and efficiency of services in Education and Health. v) Finally, governments in natural resource-rich regions tend to outperform in all sectors, but they have much lower efficiency scores. 5.6 This chapter is organized as follows: section B discusses the methodology and data used in the analysis; section C presents performance and efficiency scores and discusses the links between government performance/efficiency and key socio-economic indicators; conclusion and policy implications are debated in section D. B. Analysis Methodology 5.7 The measurement of the public sector performance (PSP), the outcomes generated by public sector activities, and public sector efficiency (PSE), the outcomes relative to the resources employed, follow the strategy proposed by Afonso, Schuknecht and Tanzi (2005). The authors assume PSP in a given sector or activity can be linearly approximated by a weighted average of outcome indicators (Ik) related with this activity. For simplicity, all indicators are assumed to carry equal weights. Therefore, for a region j and an activity i : n PSP   I ijk ij k 1 PSE is defined as the performance per unit of invested resources. PSE in a given activity can be expressed as the ratio between the PSP for this activity and the relevant public expenditure (PEX). For a region j and activity j: PSP PSEij  ij PEXij 5.8 This methodology has the advantage of being simple and intuitive, but it is subject to a few shortcomings. Index analysis, together with frontier analysis, is the approach most commonly used to assess public sector performance and efficiency. Although straightforward, this approach has a few disadvantages relative to more sophisticated methods. PSP and PSE indices depend on the arbitrary choice of outcome indicators and relative weights used to aggregate these outcomes. Indices are not flexible enough to account for external and technological factors that could affect government results. 69 5.9 These limitations are partially addressed in the analysis. The debate on the choice of outcomes is minimized as we use the outcome indicators adopted by the government as part of the results based budgeting program. These indicators are a good proxy of measure of performance considered as relevant by Peruvian authorities. Finally, a second part of the analysis compares performance and efficiency scores to key regional characteristics. This exercise can be seen as a first step in the direction of identifying the determinants of government efficiency in Peru. 5.10 The analysis is concentrated in three key sectors: Education, Health and Sanitation, and Transport. These are the sectors most frequently studied in the public spending efficiency literature10. They are not only strategic for public services delivery, but they are also convenient from the point of view of data. They are also the sectors covered by a database on relevant outcomes from public spending recently gathered by the Peruvian government. The data is part of the Results-Based Budgeting (RBB) program currently under being implemented. In addition to the analysis presented in the main text, we considered alternative specifications including specific programs in health (Nutrition and Maternal/neonatal health) and education (Initial and Primary Education). The results for the specific programs are presented in the appendix. Data 5.11 Outcome indicators are chosen, taking the government performance indicators analyzed as part of the Results-based Budgeting Program as a baseline. As part of the implementation of the RBB Program, the government of Peru collected a baseline results database in 2008. The database includes indicators of outcomes (and some outputs) related to childhood nutrition; maternal and neonatal health; initial and primary education; and access to basic services and market opportunities. Although strategic for the government and the World Bank11, these activities account for only a minor fraction of government expenditures in Health and Sanitation, Education and Transport. 5.12 In order to develop a broader sector analysis the exercise considers a few additional indicators. For Education, in addition to the four outcomes for initial and primary education proposed by the RBB database, the analysis will include two outcome indicators for secondary education (based on household surveys for 2008). For Health and Sanitation, the exercise combines nutrition, and maternal and neonatal outcome indicators to form a general health indicator, the share of the population covered by health insurance. Finally for Transport, in addition to three RBB indicators for access, we include: share of rehabilitated roads, share of paved roads, number of car accidents by 10,000 inhabitants. All outcome indicators are listed in Table D.1 in the appendix and are associated with the respective sector of analysis. 5.13 The information on public expenditure in the different sectors comes from the integrated system of information on financial management of the public sector 10 A few examples include: Hauner (2008), Afonso and Aubyn (2006) for Education; Herrera and Pang (2005) and Afonso, Schuknecht and Tanzi (2006) for Health; Herrera and Francke (2007) and Worthington (2000) for Transport. 11 The Bank supports several project in these areas. 70 (SIAF). SIAF records expenditure and revenues from the public sector by source, and by destination (including location, sector, and nature of activity). SIAF is a useful policy tool, but is subject to a few shortcomings. As established in a recent Bank-MEF study, municipalities consistently under-register commitments and accruals with suppliers and contractors, in addition, there is no correspondence between planned investment expenditures (as reflected primarily in the plans of approved investment projects) and budget formulation. Another possible drawback of using current expenditure information is the possible gap between investing resources and collecting outcomes. In order to mitigate some of these problems, whenever possible (for all sectors and sub-sectors except nutrition and maternal health), public expenditure is measured as the average of accrued expenditure for the period 2004-2008. Spending by region is defined as the total amount of public resources (spent by all levels of government) directed to the region. All expenditure indicators are listed in Table D.1 in the appendix and are associated with the respective sector of analysis. Figure 5.1: Different measures for government spending PUBLIC SECTOR PERFORMANCE (PSP) IN EDUCATION & PUBLIC SECTOR PERFORMANCE (PSP) IN EDUCATION & SPENDING IN EDUCATION AS A % GDP SPENDING IN EDUCATION PER CAPITA 1.20 1.20 1.00 1.00 0.80 0.80 0.60 0.60 0.40 0.40 0.20 0.20 0.00 0.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 0.00 100.00 200.00 300.00 400.00 500.00 600.00 700.00 800.00 SPENDING % GDP EDUCATION SPENDING PER CAPITA PSP Linear (PSP) PSP Linear (PSP) Source: World Bank staff estimates. 5.14 Expenditures are measured as public spending per capita. Afonso, Schuknecht and Tanzi (2005) and most of the literature that follow this work measure expenditures as the ratio of public spending to GDP. In this case, public expenditure is not an absolute measure, but it is relative to the total spending capacity of the economy. This indicator tends to punish wealthy regions and benefit poor ones. Figure 5.1 compares the two measures of public expenditure against PSP scores in education. When the PEX variables are measured as spending in education as a share of regional GDP, it is negatively related to performance. This happens because relatively wealthy regions tend to dedicate a smaller share of GDP to public spending, even if they spend more in absolute terms. Conversely, absolute measures, such as public spending on education per capita, tend to be positively correlated with performance. These measures are more intuitive and can be interpreted as traditional inputs in the production of public services and good. 71 5.15 Our measures of public sector performance and efficiency are compared to a set of key variables possibly linked with public sector outcomes. These variables include: regional GDP per capita and population density, constructed taking the information available from the national statistics institute (INEI) as a baseline; share of sub-national (local and regional) spending in total public spending, and share of sub- national (local and regional) revenues in total spending, constructed taking the SIAF database as a baseline; geographic region and natural resource endowments; these variables are used to group regions (see groups in Table D.3 in the appendix). The natural resource-rich group is formed by the five largest recipients of natural resources revenues.12 12 We also investigate the relationship between PSP and PSE scores and transparency at sub-national level as measured by the level of compliance with transparency and public information requirements by regional and municipal governments. We found no significant relationship. 72 C. Results Public Performance and Efficiency Scores 5.16 Based on the methodology discussed above, 12 scores were calculated for each region: PSP and PSE for Education, Health and Sanitation, Transport, Initial and Primary Education, Nutrition and Maternal and Neonatal Health (Figures D.2 to D.7 in the appendix). This section summarizes the results for Education, Health and Transport; the remaining results are presented in the appendix. Arequipa and Lima, the largest regions in Peru, are constantly among the top performers in both public spending outcomes and efficiency. Moquegua and Tacna are also relatively successful in producing the desired results. These regions however tend to spend a large amount of resources, so that performance per unit of resource invested remains low. 5.17 One important pattern stands out from plotting the performance scores, a wide range of expenditures leads to similar levels of performance. Figure 5.2 plots performance scores against the level of spending per capita in Education, Health and Sanitation, and Transport. While PSP scores in education are concentrated between 0.6 and 0.8, resources invested vary significantly across regions. The top spender, Moquegua, invests more than two times the amount invested by other large regions such as Lima or Lambayeque. PSP normalized scores in Health and Sanitation and Transport tend to be lower on average and more disperse than those in education (most regions score between 0.3 and 0.6 in Health and Sanitation and between 0 and 0.3 in Transports). Dispersions in per capita spending are even higher for these sectors. The comparison of the actual performance scores with a reference piecewise linear technological frontier confirm that there are large efficiency differences among regions in Peru. 5.18 The top performing regions tend to be among the most efficient in allocating public resources. A second group of graphs in Figure 5.2 compares scores from both PSP and PSE indexes. Further intuition on this relationship can be drawn from the Spearman rank correlation test results in Table 5.1. The relationship between performance and efficiency is significantly positive in all sectors but health. This means that regions delivering better quality services also tend to be the ones spending public resources more efficiently. In some regions such as Arequipa and Lambayeque, efficiency compensates for lower levels of per capita spending and allows for high standards of performance. Moquegua and Tacna are among the exceptions, for these regions high performance is mainly the result of very large public investments, with low levels of outcome per unit invested. As we will discuss in more detail later in this section, evidence suggests that efficiency contributes to reducing regional disparities in service delivery instead of fostering them. 73 Figure 5.2: Performance and Efficiency scores Education 1.0 1.0 LAMBAYEQ LA AREQUIPA MOQUEGUA U LIMA AREQUIPA LIBERTAD 0.9 0.9 TACNA PIURA JUNIN LIMA 0.8 0.8 LAMBAYEQU ICA CAJAMARC ICA PSE PSP JUNIN AMAZONAS E APURIMAC 0.7 HUANUCO A LA ANCASH CUSCO TUMBES AMAZONAS CAJAMARCA HUANCAVELPASCO SAN PUNO 0.7 LIBERTAD ICA AYACUCHO HUANUCO PUNO MADRE DE 0.6 MARTIN CUSCO APURIMAC PIURA DIOS HUANCAVE SAN ANCASH AYACUCHO 0.6 0.5 MADRE DE LI TACNA MOQUEGU MARTIN LORETO UCAYALI DIOS A LORETO UCAYALI PASCO TUMBES 0.5 0.4 200.0 300.0 400.0 500.0 600.0 700.0 0.50 0.70 0.90 SPENDING PER CAPITA PSP Health and Sanitation 1.0 1.0 LAMBAYEQ TACNA U 0.9 0.9 LA 0.8 0.8 LIBERTAD LIMA CAJAMARC AREQUIPA MOQUEGUA PUNO AREQUIPA TUMBES A 0.7 ICA 0.7 PSE TACNA PSP LAMBAYEQ MADRE DE SAN ICA PIURA CUSCO UE DIOS 0.6 MARTIN TUMBES 0.6 LA CUSCO JUNIN APURIMAC HUANUCO LIMA LIBERTAD ANCASH PUNO AYACUCHO SAN PIURA UCAYALI 0.5 HUANUCO AMAZONAS MARTINUCAYALI 0.5 AYACUCHO JUNIN HUANCAVEL ANCASH HUANCAVE CAJAMARCA PASCO ICA APURIMAC LICA 0.4 0.4 AMAZONAS MADRE DE LORETO PASCO LORETO DIOS 0.3 MOQUEGU 0.3 A 0.3 0.5 0.7 0.9 100.0 200.0 300.0 400.0 500.0 SPENDING PER CAPITA PSP Transport LIMA 1.0 1.0 LIMA 0.8 0.8 0.6 0.6 PSP PSE AREQUIPA 0.4 PUNO MOQUEGUA LA ANCASH AMAZONAS 0.4 PIURA CUSCO SAN LA AREQUIPA HUANUCO LIBERTAD UCAYALI ICA AYACUCHO MARTIN PIURA LIBERTAD PUNO 0.2 LAMBAYEQU LAMBAYEQ ICA E APURIMAC HUANCAVELTUMBES 0.2 HUANUCO CAJAMARCA TACNA UE JUNIN PASCO MADRE DE AYACUCHO CUSCO SAN LORETO ICA CAJAMARC ANCASH 0.0 DIOS APURIMAC DE MARTIN MOQUEGU JUNIN A UCAYALI HUANCAVE AMAZONAS LORETO TACNA MADRETUMBES 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 PASCO A (0.1) 0.0 DIOSLICA 0.2 0.4 0.6 0.8 1.0 SPENDING PER CAPITA PSP Source: World Bank staff estimates 74 Table 5.1: Spearman rank order correlation test Initial and Primary Health and Education Nutrition Maternal Health Transport Education Sanitation PSP PSE PSP PSE PSP PSE PSP PSE PSP PSE PSP PSE Education PSP 1.00 PSE 0.38* 1.00 Initial and Primary Education PSP 0.92*** 0.39* 1.00 PSE 0.55*** 0.84*** 0.64*** 1.00 Health and Sanitation PSP 0.57*** 0.13 0.62*** 0.33 1.00 PSE 0.19 0.66*** 0.34* 0.67*** 0.24 1.00 Nutrition PSP 0.58*** 0.21 0.71*** 0.42** 0.89*** 0.38* 1.00 PSE 0.39* 0.11 0.59*** 0.35** 0.64*** 0.26 0.75*** 1.00 Maternal Health PSP 0.49** 0.03 0.51** 0.26 0.83*** 0.12 0.63*** 0.67*** 1.00 PSE 0.23 0.18 0.36* 0.31 0.05 0.29 0.20 0.41** 0.15 1.00 Transport PSP 0.26 0.45** 0.29 0.44** 0.33 0.26 0.45** 0.15 -0.01 0.16 1.00 PSE 0.12 0.75*** 0.19 0.64*** 0.20 0.59*** 0.25 0.12 -0.01 0.20 0.75*** 1.00 Note: i) *, ** and *** indicates: significantly different from zero at 10%, 5% and 1% respectively ii) Efficiency scores for Nutrition and Maternal Health are base in 2008 expenditure figures. Source: World Bank staff estimates. 5.20 Public sector performance in health and education tend to be correlated, but they seem unrelated to service delivery in transport. Figure 5.3 compares the relative scores in performance and efficiency across sectors. Regions are ordered according to their ranks in the Education sector. Further intuition on cross-sector rank correlation is presented in Table 5.1. A few messages arise from the cross-sector comparison. There is a clear intersection among the top performers for the different sectors. Moquegua, Arequipa and Lima are constantly among the top five PSP scores for all sectors. The rank comparison, taking all regions into account, reveals strong complementarities between the quality of education and health services. Regions with high scores in education tend to perform well in health as well. Similar patterns arise from the analysis of specific activities within a sector. PSP Ranks in Nutrition, Maternal Health, and Primary Education are significantly correlated with each other. Transport service provision however seems unrelated to the performance in other sectors. 5.21 Relative efficiency in public spending is consistent across all sectors. Top regions in terms of efficiency are similar across the different sectors. Lambayeque, Arequipa and La Libertad appear among the top five PSE scores in Education, Health and Sanitation, and Transport. When taking all regions into account, results on government efficiency became even stronger than the ones on performance. Rank correlation statistics 75 for PSE scores between sectors (0.75 for Transport and Education; 0.59 for Transport and Health and Sanitation; and 0.66 for Education and Health and Sanitation) are significantly different from zero with 99 percent of confidence. Correlations are weak only for Nutrition and Maternal and Neonatal Health, but this result is probably driven by the differences in the definition of PEX. Figure 5.3: Performance and Efficiency across Sectors Public Sector Performance (PSP) Scores 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 PSP EDUCATION PSP HEALTH PSP TRANSPORTATION Public Sector Efficiency (PSE) Scores 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 PSE EDUCATION PSE HEALTH PSE TRANSPORTATION Source: World Bank staff estimates 5.22 In addition to documenting differences in public sector performance and efficiency, one should inquire about factors that can explain such differences. Indentifying the determinants of public sector performance and establishing causal relationships are methodologically challenging exercises that go beyond the scope of this chapter. Instead we focus on a simple but still valuable task, the search for regional characteristics that correlate to performance and efficiency. 76 Correlates of public sector performance and efficiency 5.23 This sub-section compares the PSP and PSE scores to six key variables. Figures 5.5, 5.6 and 5.7 present the plots of PSP and PSE scores in Education, Health and Sanitation, and Transport sectors respectively, against the following indicators:  Per capita income: Higher levels of per capita income tend to be associated with larger amounts of public spending and, ceteris paribus, better outcomes. In addition, richer individuals are likely to use private services to complement public services. Finally, richer regions could have stronger capacity to manage and invest government resources. Therefore, high income regions are expected ex-ante to have higher PSP and PSE scores.  Population density: This variable is included in most efficient analyses as a technologic parameter. As pointed out by many works in the literature, including the World Development Report 2009, densities and agglomeration effects help to reduce production costs and increase productivity. These results possibly extend to the provision of public services.  Share of sub-national spending in total spending: This variable seeks to link government performance and the decentralization process. It helps to investigate whether the transfer of expenditure responsibilities to sub-nationals improves performance and efficiency.  Share of sub-national revenues in total spending: The theory suggests that the decentralization of tax collection promotes accountability of governments. When individuals are directly taxed by local governments they are more inclined to monitor the use of public resources. This indicator will help us to test whether the theory applies to the regions in Peru.  Per capita spending: The previous section established a link between spending and performance. Now, we seek to understand whether government efficiency is related to the volume of resources invested.  Endowments: Revenues from natural resources-related activities play an important role in the political debate. Some argue for further redistribution, while others defend the status quo. In line with this debate we compare performance and efficiency of public spending between resource-rich regions and the remaining group. 5.24 Higher per capita income is associated with higher PSP scores, but there is no statistically significant relationship between per capita income and PSE. The first graph on the top left corner of Figures 5.5, 5.6 and 5.7 describes the relationship between PSP/PSE scores and per capita income across regions. For the three sectors, public sector performance is positively correlated with the level of income. The correlation between efficiency and income is clearly zero for Education. Although, the same coefficients are negative for Health and positive for Transport they are not statistically significant at 5 percent. This result is consistent with the idea that government outcomes in high income regions are driven by large amounts of public spending. There is no strong evidence that wealthy regions have better capacity in spending public resources. 5.25 Population density is directly related to public sector efficiency, but it has no apparent link with performance. The second graph on Figures 4.5, 4.6 and 4,7 describes the relationship between PSP/PSE scores and population density. Density is 77 positively correlated with efficiency in all sectors, confirming the existence of gains from density in service provision. Technological constraints are partially overcome by the use of public resources such that the correlation between public sector performance and population density is zero for all sectors. 78 Figure 5.5: Key correlates of Public Sector Performance and Efficiency in Education PSP,PSE & GDP PER CAPITA PSP,PSE & POPULATION DENSITY 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 0.0 20.0 40.0 60.0 80.0 100.0 GDP Per Capita POPULATION DENSITY PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSP,PSE & LOCAL SPENDING AS A SHARE OF TOTAL PSP,PSE & LOCAL REVENUES AS A SHARE OF TOTAL SPENDING SPENDING 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.6 0.7 0.8 0.9 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 share of local expenditure share of local revenue Ed PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSE & SPENDING PER CAPITAL AVERAGE SCORES 1.2 0.9 0.8 1.0 0.7 0.8 0.6 0.5 0.6 0.4 0.4 0.3 0.2 0.2 0.1 0.0 0.0 0.0 200.0 400.0 600.0 800.0 Costa Sierra Selva Natural Other TOTAL PUBLIC SPENDING resource PSE Linear (PSE) PSP PSE Rich Source: World Bank staff estimates. 79 Figure 5.6: Key correlates of Public Sector Performance and Efficiency in Health and Sanitation PSP,PSE & GDP PER CAPITA PSP,PSE & POPULATION DENSITY 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 0 20 40 60 80 100 GDP Per Capita POPULATION DENSITY PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSP,PSE & LOCAL SPENDING AS A SHARE OF TOTAL PSP,PSE & LOCAL REVENUES AS SHARE OF TOTAL 1.2 1.2 SPENDING SPENDING 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.5 0.6 0.7 0.8 0.9 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 share of local expenditure share of local revenue PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSE & SPENDING PER CAPITA AVERAGE SCORES 1.2 0.8 0.7 1.0 0.6 0.8 0.5 0.4 0.6 0.3 0.4 0.2 0.1 0.2 0.0 0.0 Costa Sierra Selva Natural Other 0.0 100.0 200.0 300.0 400.0 500.0 600.0 resource TOTAL PUBLIC SPENDING Rich PSE Linear (PSE) PSP PSE Source: World Bank staff estimates. 80 Figure 5.7: Key correlates of Public Sector Performance and Efficiency in Transport PSP,PSE & GDP PER CAPITA PSP,PSE & POPULATION DENSITY 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 0.0 20.0 40.0 60.0 80.0 100.0 GDP Per Capita POPULATION DENSITY PSP PSE PSP PSE Linear (PSP ) Linear (PSE) Linear (PSP ) Linear (PSE) PSP,PSE & LOCAL SPENDING AS A SHARE OF TOTAL PSP,PSE & LOCAL REVENUES AS A SHARE OF LOCAL SPENDING SPENDING 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 share of local expenditure share of local revenue PSP PSE PSP PSE Linear (PSP ) Linear (PSE) Linear (PSP ) Linear (PSE) PSE TRANSPORTE TERRESTRE & TOTAL PUBLIC AVERAGE SCORE 1.2 SPENDING 0.4 1.0 0.3 0.3 0.8 0.2 0.6 0.2 0.4 0.1 0.2 0.1 0.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 0.0 (0.2) Costa Sierra Selva Natural Other TOTAL PUBLIC SPENDING resource Rich PSP PSE PSE Linear (PSE) Source: World Bank staff estimates. 81 5.26 Decentralization in expenditure responsibilities, measured by the share of local and regional spending in total spending in the sector, seems unrelated to performance and efficiency in most sectors. The third graph on Figures 5.5, 5.6 and 5.7 describes the relationship between PSP/PSE scores and decentralization in spending. Except for Transport, where sub-national spending is directly linked with government outcomes, the local/regional share is not related to PSP scores. Correlation between public sector efficiency and sub-national spending is zero for all sectors. These preliminary results suggest that the ongoing decentralization in expenditure responsibilities has not been successful in improving service delivery. 5.27 On the other hand, there is a positive relationship between decentralization in revenues collection, measured as the share of local and region revenues of total spending, and both PSP and PSE. The fourth graph on Figures 5.5, 5.6 and 5.7 describes the relationship between PSP/PSE scores and decentralization in revenue collection. While the share of revenues collected sub-nationally seems irrelevant for public developments in the Transport sector, it is directly linked to both efficiency and performance in Health and Sanitation, and Education. The correlation is even stronger when considering specific programs, such as Primary Education and Nutrition (Appendix D), which can be closely monitored by the population. These results are consistent with the theoretical predictions. 5.28 There seems to be a negative relationship between the amount of public resources invested and the levels of relative efficiency across regions. The fifth graph on Figures 5.5, 5.6 and 5.7 describes the relationship between PSE scores and per capita spending. This relationship is strongly negative for all sectors (and sub-sectors) analyzed, suggesting diminishing returns in the production of public goods and services. It also suggests that efficiency tends to compensate for lower public spending, helping to mitigate disparities in government performance. This result could be driven by technological factors, it is harder to produce better outcomes when you are already performing well, or by incentives: “necessity is the mother of invention� (by Plato). In both cases, gains could be made by redistributing public resources from big spenders to frugal regions. 5.29 Finally, public sector performance and efficiency vary significantly between natural resource-rich and other regions. The sixth graph on Figures 5.5, 5.6 and 5.7 presents the average scores group for regions by geographic characteristics (Costa, Sierra and Selva regions) and by natural resources’ endowment (Natural resource-rich and Others). Regions in the Costa perform on average significantly better with respect to both PSP and PSE. They are followed by regions in the Sierra; regions in the Selva on average have the lowest PSP and PSE scores. This result is consistent across all sectors. Natural resource rich-regions, on average, have higher (in Education and Health and Sanitation) or at least equivalent (in Transport) PSP scores with respect to other regions. In terms of efficiency, this group scores significantly less in all sectors. For natural resource rich- regions, public sector outcomes tend to be the result of abundant spending but low public sector productivity. Overall, the evidence suggests that gains could be made by redistributing resources across regions. 82 D. Conclusions 5.30 This chapter proposed a formal assessment of public sector performance and efficiency at regional level in Peru. Following Afonso, Schuknecht and Tanzi, 2005, we constructed Public Sector Performance and Public Sector Expenditure indices and confronted those indices with different socio-economic indicators related to the government’s ability for generating outcomes. Results confirm significant disparities in government performance and efficiency across Peruvian regions. While top performers tend to be more efficient in allocating public resource, some regions are able to compensate low levels of efficiency with a large amount of public resources. 5.31 Efficiency scores are negatively correlated with levels of public spending, and they are significantly lower for natural resource-rich regions. This evidence suggests that gains could be made by redistributing resources toward more efficient regions. The relationship between PSE scores and PEX is negative for all sectors (and sub-sectors) analyzed. This result is partly driven by natural resource-rich regions. These regions combine good performance with very low levels of efficiency at the expense of large amounts of public resources. Whether driven by technological factors or by incentives schemes, this fact suggests the existence of diminishing returns in the production of public goods and services. In this case, gains could be made by redistributing public resources away from big spenders to frugal regions. 5.32 Finally, decentralization of expenditure responsibilities cannot be associated with better or more efficient service delivery, especially for Health and Education. On the other hand, decentralization in revenue collection seems to play a role in public sector performance. Except for Transport, where sub-national spending is mildly linked to government’s outcomes, the share of local and regional spending in total spending is not related with PSP or PSE scores. In contrast, the share of public revenues collected sub-nationally is directly associated with efficiency and performance in Health and Sanitation and Education. The correlation is even stronger when considering specific programs, such as Primary Education and Nutrition, which can be closely monitored by the population. These results indicated that the ongoing decentralization process has been only partly successful in improving service delivery, but further analysis is suggested before a final verdict. 83 List of references Afonso, Antonio and Miguel St. Aubyn (2006), “Cross-country efficiency of secondary education provision: A semi-parametric analysis with non-discretionary inputs,� Economic Modeling, Elsevier, vol. 23(3), pages 476-491, May. Afonso, Antonio, Ludger Schuknecht, andVito Tanzi (2005), “Public sector efficiency: An international comparison,� Public Choice, Springer, vol. 123(3), pages 321-347, June. Afonso, Antonio, Ludger Schuknecht, and Vito Tanzi (2006), “Public Sector Efficiency: Evidence for New EU Member States and Emerging Markets,� ECB Working Paper No. 581, January. Ahmad, Ehtisham and Mercedes García-Escribano (2008), “Constraints to Effective Fiscal Decentralization in Peru.� Paper prepared for the Conference on Obstacles to Decentralization: Lessons from Selected Countries, Georgia State University. Previous version published as IMF Working Paper No. 06/120 available at http://www.imf.org/external/pubs/ft/wp/2006/wp06120.pdf Alfaro Díaz, Javier and Markus Rühling (2007), “La incidencia de los Gobiernos Locales en el Impuesto Predial en el Perú,� Instituto de Investigación y Capacitación Municipal (INICAM), Lincoln Institute of Land Policy, and Konrad Adenauer Stiftung. January. Aragón, Fernando and Carlos Casas Tragodara (2009), “Technical capacities and local spending: evidence from Peruvian municipalities� Artículo a ser publicado en la revista Perspectivas de la Corporación Andina de Fomento. Arellano-Yanguas, Javier (2008), “A Thoroughly Modern Resource Curse? The New Natural Resource Policy Agenda and the Mining Revival in Peru,� IDS Working Paper No. 300, Brighton, UK. Azpur, Javier, Eduardo Ballón, Luis Chirinos, Epifanio Baca, and Gerardo Távara (2006), Descentralización. Aportes para el gobierno peruano 2006-2011, Consorcio de Investigación Económica y Social – Propuesta Ciudadana, Lima Bahl, Roy and Jorge Martínez-Vázquez (2006), “Sequencing Fiscal Decentralization,� World Bank Policy Research Working Paper 3914, May. Bardhan, P., & Mookherjee, D. (2006), “Pro-poor targeting and accountability of local governments in West Bengal�, Journal of Development Economics, 79(2), 303–327. Bird, Richard (1999), "Rethinking Sub-national Taxes: A New Look at Tax Assignment", Washington, International Monetary Fund. Bird, Richard and Francois Vaillancourt (eds). (1998) “Fiscal Decentralization in Developing Countries� Cambridge University Press. Cameron, Maxwell (1994), Democracy and Authoritarianism in Peru: Political Coalitions and Social Change. New York: St. Martin’s Press. Cameron, Maxwell and Steven Levitsky (2003), “Democracy without Parties? Political Parties and Regime Change in Fujimori’s Peru,� Latin American Politics and Society 45 (3). Casas Tragodara, Carlos (1997). “Descentralización Fiscal: El Caso de Perú�. Proyecto Regional de Descentralización Fiscal. CEPAL-GTZ. 84 Casas Tragodara, Carlos (2001). “El Proceso de Descentralización en el Perú�. Universidad del Pacífico. Lima. Casas Tragodara, Carlos (2005) “Aspectos Económicos y Fiscales de la Descentralización en el Perú�. Programa de descentralización y buen gobierno. Defesoría del Pueblo. Primera Edición, Lima. Casas Tragodara, Carlos (2009), “Review of the Decentralization Process and its Links with Public Expenditure Efficiency in Peru.� Unpublished background paper prepared for this report. Chirinos, Luis (2004), “Participación ciudadana en gobiernos regionales: El caso de los consejos de coordinación regional,� in Grupo Propuesta Ciudadana, La participación ciudadana y la construcción de la democracia en América Latina. Lima: Propuesta Ciudadana. Conaghan, Catherine (2006), Fujimori’s Peru: Deception in the Public Sphere. Pittsburgh: University of Pittsburgh Press. Consejo Nacional de Descentralización (2006), Glosario de la Descentralización, Lima. Consejo Nacional de Descentralización (2006), La descentralización en Perú. Compendio Normativo, Lima. Conterno, Elena (2007), La agenda a cuatro años de iniciado el proceso. Lima: USAID. Contreras, Carlos (2002), El centralismo peruano en su perspective histórica. Lima: Instituto de Estudios Peruanos. Dafflon, Bernard (2006), “The assignment of functions to decentralized government from theory to practice.� In “Handbook of Fiscal Federalism�. Edward Elgar USA. Defensoría del Pueblo (2009), “Hacia una descentralización al servicio de las personas: recomendaciones en torno al proceso de transferencia de competencias a los gobiernos regionales,� Informe defensorial no. 141, March. Dickovick, Tyler (2007), “Municipalization as Central Government Strategy: Central-Regional- Local Politics in Peru, Brazil and South Africa,� Publius 37 (1). Dickovick, Tyler (Forthcoming), When the Center Holds: Decentralization and Recentralization in the Developing World. University Park: Penn State University Press. Eaton, Kent, Kai Kaiser and Paul Smoke (2009), “The Political Economy of Decentralization in Developing Countries: A Development Partner Perspective.� Paper prepared for the World Bank. Giugale, Marcelo y Steven Webb (eds.) (2000), “Achievements and Challenges of Fiscal Decentralization. Lessons from Mexico,� The World Bank. Goldfrank, Benjamin (2007), “Lessons from Latin America’s Experience with Participatory Budgeting,� in Anwar Shah, ed., Participatory Budgeting (Washington, DC: World Bank Institute, Public Sector Governance and Accountability Series). Gómez, Juan Luis, Jorge Martínez-Vázquez and Cristián Sepúlveda (2008), “Diagnóstico del Proceso de Descentralización Fiscal en el Perú: Informe Final.� Georgia State University. Gonzáles de Olarte, Efraín (2000), Neocentralismo y neoliberalismo en el Perú. Lima: Instituto de Estudios Peruanos. 85 Gonzáles, Alberto and Marco Shiva (2007), “Evaluando los determinants de la percepción de beneficios de la descentralización en Perú: un análisis a nivel departamental,� Informe final de investigación, Lima: CIES – Macroconsult. Grompone, Romeo (2005), La escisión inevitable: partidos y movimientos en el Perú actual. Lima: Instituto de Estudios Peruanos. Grupo Propuesta Ciudadana (2007), Optimización de los recursos del Cánon en el Perú: los fondos intergeneracionales, Lima. Grupo Propuesta Ciudadana (2009), Gasto public y Cánon en el Perú. Análisis y recomendaciones para el major aprovechamiento de las rentas del gas de Camisea, Lima. Hauner, D. (2008), "Explaining Differences in Public Sector Efficiency: Evidence from Russia's Regions," World Development, Elsevier, vol. 36(10), pages 1745-1765, October. Herrera, Pedro and Pedro Francke (2007), “Análisis de la eficiencia del gasto municipal y sus determinants,� September. Herrera, S. & Pang, Gaobo (2005), “Efficiency of public spending in developing countries : an efficiency frontier approach,� Policy Research Working Paper Series 3645, The World Bank. Litvack, Jennie y Jessica Sedon (Editoras) (2002) “Decentralization Briefing Notes� World Bank Institute. Lizárraga Bobbio, Raúl (2009), “Descentralización en el Perú. Viejos Mitos, Nuevos Retos,� RED-Perú Iniciativas de Concertación para el Desarrollo Social, Lima. Lopez Murphy, R (1994), “Descentralización Fiscal y Política Macroeconómica� Serie Política Fiscal 87. Comisión Económica para América Latina y el Caribe – CEPAL/GTZ. McLure, Charles (2001), “Tax Harmonization, Fiscal Federalism, and Regional Economic Integration: International Experience and Lessons for Brazil.� Seminario Internacional de Sistema Tributatio y Competitividad. Confederación Nacional de la Industria Brasilera. McClintock, Cynthia (2006), “Electoral Authoritarian Versus Partially Democratic Regimes: The Case of the Fujimori Government and the 2000 Elections,� in Julio Carrión, ed., The Fujimori Legacy: The Rise of Electoral Authoritarianism in Peru. University Park: Penn State University Press. McNulty, Stephanie (2008), Voice and Vote: Decentralization, Participation and the Crisis of Representative Democracy in Peru, unpublished book manuscript. Ministerio de Economía y Finanzas. “Implementación de Reglas fiscales para los Gobiernos Regionales y Locales�. Boletín de Transparencia Fiscal- Informe Especial. Lima. Muñoz, Paula (2005), “¿Quién piensa el país?� Cuestión de Estado, no. 37/38 (October), pp. 53- 58. Nickson, Andrew (1995), Local Government in Latin America. Boulder: Lynne Rienner Press. Netzer, Dick y Matthew P. Drennan (editores) (1997) “Readings in State and Public Local Public Finance�. Blackwell Publishers Ltd. Reino Unido Oates, Wallace (1972). “Fiscal federalism�. NY: Harcourt Brace Jovanovich. 86 PNUD (2008). Cifras para la Descentralización. Versión temática. Edición especial – Presidencia del Consejo de Ministros – Programa de las Naciones Unidas para el Desarrollo, Lima. Prud’homme, Remy (2003). “Fiscal decentralization in Afirca: A framework for considering reform,� Public Administration and Development, 23, 17-27. RED-Perú (2008), Propuesta de la Agenda Prioritaria para profundizar la Descentralización en el Perú, Lima, October. REMURPE (2006). “Manual de Gestión Moderna de Municipalidades Rurales�, Red de Municipalidades Rurales en el Perú. O’Neill, Kathleen (2005), Decentralizing the State: Elections, Parties and Local Power in the Andes. Cambridge: Cambridge University Press. Panfichi, Aldo (2007), “The Challenges of the New APRA Government: Social Conflicts, Political Economy and Governance.� Paper prepared for the Latin American Studies Association Conference, Montreal, Canada, September 1-4. Presidencia del Consejo de Ministros. Secretaría de Descentralización (2008), Informe Annual sobre el Proceso de Descentralización, Lima, May. PRODES (2005), Proceso de Descentralización 2004: Balance y Recomendaciones para una Agenda Pendiente, USAID, Lima. PRODES (2006), Proceso de Descentralización 2005 – Abril 2006: Balance y Desafíos, USAID, Lima, May. PRODES (2007), Proceso de Descentralización – La Agenda a Cuatro Años de Iniciado el Proceso, USAID, Lima, January. Roberts, Ken (1996), “Neoliberalism and the Transformation of Populism in Latin America: The Peruvian Case,� World Politics 48 (1): 82-116. Rodríguez Doig, Enrique (2008), “Descentralización. La investigación económica y social en el Perú: balance 2004-2007,� Consorcio de Investigación Económica y Social, Lima. Schmidt, Gregory (1989), “Political Variables and Governmental Decentralization in Peru: 1949- 1988,� Journal of Interamerican Studies and World Affairs 31: 193-234. Shah, Anwar (2004), “Fiscal Decentralization in Developing and Transition Economies�. World Bank Policy Research Working Paper 3282. Smoke, Paul (2003), “Decentralisation in Africa: Goals, Dimensions, Myths and Challenges,� Public Administration and Development, 23, 7-16. Tanaka, Martín (2002), “La dinámica de los actores regionales y el proceso de descentralización: ¿el despertar del letargo?� Lima: Instituto de Estudios Peruanos. Documento de Trabajo 125. Tanzi, Vito and Ludger Schuknecht (1997), "Reconsidering the Fiscal Role of Government: The International Perspective," American Economic Review, vol. 87(2), pages 164-68, May. Ter-Minassian, T. y Craig, Jon (1997), “Control of Sub-national Government Borrowing� en Fiscal Federalism in Theory and Practice. Cap 7. Publicación: Fondo Monetario Internacional FMI. Washington. 87 Tiebout, Ch. (1978), “A Pure Theory of Local Expenditure". Hacienda Pública Española. N# 5. Madrid. Tullock, G. (1969), “Federalism: problems of scale�, Public Choice, 4, 19–29 USAID (2006), Proceso de Descentralización 2005-abril 2006. Balance y desafíos, Lima Véliz, Claudio (1980), The Centralist Tradition in Latin America. Princeton: Princeton University Press. Vergara, Alberto (2007), “El Choque de los Ideales: Reformas Institucionales y Partidos Políticos en el Perú Post-Fujimorato,� Paper presented at the Latin American Studies Association Conference, Montreal, Canada, September 1-4. Willis, Eliza, Stephan Haggard and Christopher Garman (1999), “The Politics of Decentralization in Latin America,� Latin American Research Review 34 (1): 7-56. World Bank (1999), Beyond the Center. Decentralizing the State. World Bank (2003), “Program Document for a Proposed Programmatic Decentralization and Competitiveness Structural Adjustment Loan in the amount of US$ 150 million for the Republic of Peru,� Report No. 26837-PE, November 3, 2003. World Bank (2008), “World Development Report 2009: Reshaping Economic Geography,� The World Bank, Washington, DC. November. Worthington, A. (2000), “Cost efficiency in Australian local government: A comparative analysis of mathematical programming and economic approaches�, Financial Accountability & Management, 16(3), 0267-4424, August. Yashar, Deborah (2005), Contesting Citizenship in Latin America: The Rise of Indigenous Movements and the Post-Liberal Challenge. Cambridge: Cambridge University Press. Zas Friz Burga, Johnny (2004), “La insistencia de la voluntad. El actual proceso peruano de descentralización política y sus antecedentes inmediatos (1980-2004)�. Defensoría del Pueblo. Primera Edición. Perú. 88 Annex A. Basic indicators by department Table A.1: Selected indicators, by department Population Population GDP as GDP per Average Poverty rate (2007) density share in capita real GDP (2008, percent) (2007) national (2008, growth in GDP thousands 2004-08 (per cent) of current (percent) nuevos soles) Amazonas 375,993 9.6 0.6 4.8 7.3 59.7 Ancash 1,063,459 29.6 3.4 15.4 5.6 38.4 Apurímac 404,190 19.3 0.4 3.3 5.5 69.0 Arequipa 1,152,303 18.2 5.2 16.6 9.3 19.5 Ayacucho 612,489 14.0 0.9 5.1 10.0 64.8 Cajamarca 1,387,809 41.7 2.4 6.3 1.9 53.4 Cusco 1,171,403 16.3 2.3 8.8 9.4 58.4 Huancavelica 454,797 20.5 0.7 6.1 3.3 82.1 Huánuco 762,223 20.7 0.9 3.9 3.4 61.5 Ica 711,932 33.4 2.8 15.0 13.3 17.3 Junín 1,225,474 27.7 2.9 8.6 6.5 38.9 La Libertad 1,617,050 63.4 4.3 9.9 10.8 36.7 Lambayeque 1,112,868 78.2 2.4 7.6 8.1 31.6 Lima * 9,322,088 266.7 47.5 17.1 9.4 18.3 Loreto 891,732 2.4 1.7 7.8 4.8 49.8 Madre de Dios 109,555 1.3 0.4 15.0 8.2 17.4 Moquegua 161,533 10.3 1.2 37.8 2.5 30.2 Pasco 280,449 11.1 1.0 15.2 5.6 64.3 Piura 1,676,315 46.7 3.6 10.2 8.1 41.4 Puno 1,268,441 17.6 1.9 5.8 5.7 62.8 San Martín 728,808 14.2 1.1 5.2 8.3 33.2 Tacna 288,781 18.0 1.2 16.9 4.6 16.5 Tumbes 200,306 42.9 0.4 7.8 6.5 17.2 Ucayali 432,159 4.2 0.9 7.6 5.9 32.5 PERU 27,412,157 21.3 13.1 7.64 36.2 Notes: * Includes Lima provinces, metropolitan Lima, and Callao. Source: INEI. 89 Table A.2: Fiscal indicators, by department Public spending, all Share of public spending Public debt of regional levels of government (2008, percent) by level of governments (2007), as (2008, current million of government percent of nuevos soles) Total Per capita Central Regional Local Regional Regional tax GDP revenues Amazonas 839 2,230 48 15 37 0.8 164 Ancash 2,511 2,361 20 49 31 0.7 43 Apurímac 731 1,809 27 24 49 0.7 82 Arequipa 2,076 1,802 38 24 38 0.8 12 Ayacucho 1,160 1,894 32 26 42 1.0 106 Cajamarca 1,824 1,315 27 32 41 0.6 44 Cusco 2,762 2,358 26 43 31 0.4 4 Huancavelica 875 1,923 22 33 44 0.6 191 Huánuco 1,109 1,455 41 22 37 1.0 99 Ica 1,219 1,712 46 20 35 1.6 54 Junín 1,853 1,512 37 26 37 0.2 8 La Libertad 2,217 1,371 36 25 39 0.5 12 Lambayeque 1,524 1,369 47 19 34 3.0 121 Lima * 39,292 4,215 89 9 3 0.9 4 Loreto 1,455 1,631 25 26 49 0.8 33 Madre de 308 2,814 48 11 41 0.7 37 Dios Moquegua 841 5,204 13 60 27 0.0 3 Pasco 779 2,778 21 45 33 0.2 26 Piura 2,124 1,267 33 31 37 0.8 13 Puno 1,811 1,427 31 30 39 0.6 52 San Martín 1,210 1,660 38 14 48 0.8 55 Tacna 1,061 3,674 13 55 32 0.3 15 Tumbes 474 2,364 32 25 43 2.7 138 Ucayali 832 1,926 38 22 39 0.9 13 PERU 70,884 2,586 63 19 18 n.a. n.a. Notes: * Includes Lima provinces, metropolitan Lima, and Callao. Source: INEI, Congreso de la República (2009). 90 Table A.3: Potential fiscal viability of departments Public spending Tax collection (2006, millions of (2006, millions of current soles) Sub-national current soles) spending minus According to revenues collected in Sub- According to All levels of actual location department** national residency of government of economic level taxpayer activity* Amazonas 400 251 9 153 -248 Ancash 1,329 792 217 2,505 1,176 Apurímac 493 323 14 240 -252 Arequipa 1,343 827 1,105 2,514 1,171 Ayacucho 794 518 28 393 -402 Cajamarca 1,297 1,021 111 1,853 556 Cusco 1,464 1,018 841 1,456 -9 Huancavelica 574 339 5 693 120 Huánuco 626 412 28 527 -99 Ica 840 485 226 1,160 320 Junín 1,279 802 240 1,558 279 La Libertad 1,317 841 629 2,209 892 Lambayeque 1,028 534 226 1,350 322 Lima 34,763 2,239 41,088 21,288 -13,475 Loreto 1,107 871 157 1,123 16 Madre de 246 93 23 125 -121 Dios Moquegua 494 433 51 938 444 Pasco 485 237 23 1,063 578 Piura 1,548 1,101 652 2,075 527 Puno 1,122 800 97 1,072 -50 San Martín 712 489 55 661 -51 Tacna 661 555 125 684 23 Tumbes 355 237 28 213 -143 Ucayali 607 377 240 356 -251 Notes: * As per estimations conducted by INDE (2007) on behalf of the association of regional government presidents. ** Using the estimates of tax collections by department on the basis of actual location of economic activity. Negative figures indicate that the Source: INDE (2007), BCRP. 91 92 Annex B. Transfer of functions in health and social programs Health The health sector has maintained a highly centralized institutional and functional structure with non-concentrated management in recent years through different executive units such as Health Care Areas all over the country and in Regional Offices. These institutions were part of the Regional Administration Transitory Councils prior to the decentralization process, and are still part of the Regional Governments. At the same time, a shared and participative way through Local Communities of Health Management (LCHM) was created as a strategy of community participation for health management. In 2007, there were more than 760 associations that covered the management of more than 30% of the health establishments. In 2002, with the formulation and development of the decentralization process, the National System of Coordination and Decentralization of Health was created with 3 levels of organization: National Health Counsel (NHC), Regional Health Counsel (RHC) and Province Health Counsel (PHC). That represented the development of a coordinating mechanism among the different government levels related to health. However, last years’ registers show that the participation of RHC and PHC has diminished significantly in many regions for various reasons such as lack of resources, lack of attendance from its members, lack of projects to implement monitoring, among others. Since 2003, as the legal frame for decentralization came into force, the basis for the health management: the Regional Governments Organic Law (RGOL) and the Municipalities Organic Law (MOL), were developed. The Decentralization Basis Law mentions on its 36 and 43 articles, regarding the shared functions of Regional and Local Governments, the function of public health without specifying functions or competences for each institution. The MOL defines 16 specific functions related to health that need to be transferred to the Regional Governments, as detailed below. 93 Health Functions for the Regional Governments a. Formulate, approve, execute, evaluate, manage, control and administer the health policies of the regions in accordance with national policies and sectoral plans. b. Formulate and execute the Regional Health Development Plan. c. Coordinate the actions of integral health at the regional level. d. Participate in the Coordinated and Decentralized National Health System as mandated by the current regulation. e. Promote and execute the activities of health promotion and prevention. f. Organize the assistance and administration levels of the government health institutions that provide services in the region, in coordination with the Local Governments. g. Organize, implement and maintain health services for the prevention, protection, recovery and rehabilitation, in coordination with the Local Governments. h. Supervise and fiscalize public and private services. i. Manage and execute, in coordination with the respective institutions, the prevention and control of emergency and disasters risks and damages. j. Supervise and control the production, commercialization, distribution and consumption of pharmaceutical and other related products. k. Promote and preserve the environmental health of the region. l. Plan, fund and execute sanitary infrastructure and equipment projects, promoting the technological development of health in the region. m. Provide useful information about the management of the sector as well as health infrastructure and services to the population. n. Promote the formation, capacitation and development of human resources and articulate health services in the teaching, investigation and projection for the community. o. Periodically and systematically evaluate the sanitary achievements of the region. p. Execute, in coordination with the Local Governments, effective actions that contribute to improving the nutritional levels in the population of the region. Source: RGOL The MOL establishes 18 functions related to healthiness and health in general, separating the functions between Province Municipalities and District Municipalities. 94 Health Functions for the Local Governments 1. Specific functions of province municipalities: a. Regulate and control the process of solid and liquid wastes and industrial residuals at the province level. b. Supervise and control the emission of smoke, gases, noises and other elements that contaminate the atmosphere and the environment. 2. Specific functions shared by province municipalities: a. Administer and regulate the potable water and drain and sewage services; public cleaning and solid waste treatment Concession processes are executed by the province municipalities and coordinated with the national organizations for investment promotion. b. Provide rural drainage services when they are not provided by the district municipalities or rural population centers, and coordinate with them the execution of epidemic and animal health control campaigns. c. Disseminate environmental health programs in coordination with the relevant district municipalities and national and regional organizations. d. Manage primary health assistance, health center construction and equipment, first aid kits and health offices in populated centers; in coordination with district municipalities, populated centers and national and regional organizations. e. Execute preventive medicine, first aid, health education and local prophylaxis campaigns. 3. Specific functions of district municipalities: a. Provide public cleaning services determining waste accumulation areas and proper use of industrial residuals. b. Regulate and control cleanliness, hygiene and healthiness in commercial and industrial establishments, houses, schools, pools, beaches and other public places. c. Install and maintain public restrooms and hygienic services. d. Supervise and control the emission of smoke, gases, noises and other elements that contaminate the atmosphere and the environment. e. Issue health care cards. 4. Specific functions shared by district municipalities: a. Administer and regulate, when they have the capacity, the potable water and drain and sewage services, public cleaning and solid waste treatment. b. Provide rural drainage services and coordinate the execution of epidemic and animal health control campaigns with populated centers. c. Disseminate environmental health programs in coordination with the relevant province municipalities and national and regional organizations. d. anage primary health assistance;, health center construction and equipment; first aid kits and health offices in populated centers, in coordination with province municipalities, populated centers and national and regional organizations. e. Execute preventive medicine, first aid, health education and local prophylaxis campaigns. Source: MOL According to the Decentralization Basis Law, the transfer of the health and education functions mentioned above would have to be part of the last stage of the decentralization process. Nevertheless, in practice the transfer has not been rigid about what is stated on the Basis Law. Even though the first Quinquennial Plan of 2005 established the Transfer Plan for the Health Sector, those responsibilities were already being performed by the Regional Governments through their Regional Offices, and the transfer just represented the formalization of this 95 situation. Furthermore, it is important to recall that the Regional Offices are responsible for the execution of regional and national sectoral policies which are also under the responsibility of the respective sector; in this case the Health Ministry, MINSA and the Regional Management, which implies regulative dependence on the sector and administrative and budgetary dependence on the Regional Government. However, the same health legislation establishes that there are some functions shared among 3 levels of government, according to articles 26 and 43 of the Decentralization Basis Law: For the Central and Regional levels:  Regulation  Funding  Process Management  Sectoral Conduction  Social Participation  Investment For the Regional and Local levels:  Services Organization  Assistance  Sectoral Conduction  Social Participation  Investment Additionally, it is worth mentioning that the Basis Law suggests that the following are exclusive competences of the different governments (related to health): Central Level • Approve national and sectoral policies. • Regulate their public services. • Regulate and manage the national public infrastructure. Regional Level • Regional Planning • Execute Programs • Approve Agreed Development Plan • Approve Internal Organization and Budget. • Promote and Execute Public Investments. Local Level • Regional and District Planning. • Regulate and Manage Collective Public Services. • Approve Internal Organization and Budget. • Approve Local Development Plan. 96 That, in practice, generates a series of legal confusions about who has the responsibility for which function and, given that there are no coordinating mechanisms, it is complicated to continue the transfer process. For instance, one of the EXCLUSIVE functions of the central level is the “infrastructure management� and, at the same time, one of the EXCLUSIVE functions of the intermediate level is the execution of public investment. iven that the main part of investments on this sector come from the acquisition and construction of the infrastructure improvement, it is difficult to clearly understand who is responsible for that task. In 2005 a Sector Transfer Plan for 2005-2009, which was a proposal that encompassed the results of the negotiations about responsibilities being transferred13 between MINSA and the Regional Governments, was developed. That document contained 13 functions and 37 faculties contemplated in 3 blocks to be transferred on 3 different stages: Block 1: Basically those that were being executed regionally and in a non-concentrated way by the Regional Health Offices (DIRESA). Mainly functions in which their budgetary execution did not require additional regulatory or human capacities. Block 2: This type of function requires partly-developed capabilities at the regional level that could be complemented with additional programs of some government levels. Block 3: These functions require high capacities that are poorly developed at the decentralized levels and that needed to be complemented with intense capacitating processes or important regulatory developments. Finally, at the end of 2006, and as part of the “decentralization shock�, there was an announcement of 20 measures related to the transfer of functions both in general matters and specifically in health; also the beginning of a Primary Health Management Municipalities Transfer pilot program. Health Municipalization For the execution of these transfers MINSA approved the technical document “Health Function Development in Local Governments� on May 2007 and “Implementation Guide of Health Decentralization for Local Governments Pilot Projects� approved on August 2007. This project implied the transfer of primary health care management that involves the planning, programming, monitoring, supervision and evaluation of health care in 98% of the establishments (approximately 6,466 nationally). 13 It is worth mentioning that the Medium Term Sectoral Plan is a reference for the elaboration of the Quinquennial Plan. Even though there were 13 functions registered for transfer in the Sectoral Plan elaborated by the MINSA, the Quinquennial Plan only considered 7. The 13 were included in the Annual Plan. 97 Initially the elaboration of pilot plans were established for the end of 2006 in order to start the transfer during 2007 and 2008; also the local capacities development process was to start in 2008. Health Municipalization Chronogram Preparatory Beginning Primary Health Development of Stage of Pilot Assistance Local Capacities Projects Management Transfer October – December 2006 2007 2008 2008 – and forward Agreed Municipalization Proposal Design of Education Municipalization Pilot Projects Municipalización de la educación Source: MINSA MONITORING BASELINE - EVALUATION In 2008 there were 40 outstanding pilot projects in 16 regions as depicted in the following chart, the complete list can be found on Exhibit I: Pilot Projects for local decentralization 2008 Regions Nº of Pilot Projects Amazonas 3 Apurímac 3 Piura 3 Huánuco 2 Pasco 3 Ayacucho 2 Huancavelica 5 San Martín 3 Madre de Dios 2 Cajamarca 2 Ica 2 Arequipa 3 Lambayeque 3 Tumbes 1 La Libertad 1 Ucayali 2 TOTAL 40 Source: MINSA 2009 98 Social programs The transfer of Social Programs to Local and Regional Governments has been an active process since the beginning of the decentralization process. The second complementary disposition was established in the Decentralization Basis Law thus : “social programs and regional productive investment projects transfer […], since 2003 the transfer of the social programs that fight against poverty and provide regional investment and productive infrastructure projects to the local and regional governments starts […]�. This represented a strong willingness from the legislators to initiate the decentralization process with the transfer of this type of program and project. Specific functions related to different sectors at the national level would come later. At the same time, in the definition of the process stages in the Transitory Dispositions, the first stage adds the transfer and reception of the Social Programs described on the Second Complementary Disposition together with the RATC assets and the design of the respective Capacitating Plans. In the same regulation the “Attention and administration of social programs� was included within the functions of the Regional Governments (different from the health and education sectoral functions); it is important to mention that this function was not included as part Local Government functions. This regulation, like the other cases previously described, was complemented by the development of the specific functions in the RGOL: Article 61 – Social Development and Equality of Opportunities functions for the Regional Governments a) Formulate, approve and evaluate policies related to social development and equality of opportunities, in accordance with the general policy of the NG, the sectoral plans and the respective LG programs. b) Supervise and evaluate the accomplishment of the execution of sectoral policies by LG and the functioning of the programs of fighting against poverty. c) Formulate policies, regulate, manage, execute, promote, supervise and control the actions oriented towards the prevention of political, familiar and sexual violence. d) Promote citizenship participation in the planning, administration and surveillance of social investment development plans; giving the advice and support that the organizations involved need. e) Manage and facilitate the support from international cooperation and the private sector in programs for the fight against poverty and for social development. f) Promote a culture of peace and equality of opportunities. g) Formulate and execute particular actions and policies aimed at the inclusion, prioritization and promotion of the rural and native communities in their jurisdiction. h) Formulate and execute particular actions and policies that will enhance social assistance by the support and protection for children, teenagers, women, disabled people, elder adults and social sectors in a risky or vulnerable situation. Source: RGOL According to this regulation, the Regional Governments would be mainly in charge of supervising and evaluating the execution of the Local Governments’ social programs. Local 99 Governments received, through the MOL, the executive function for these programs and that function is being assumed exclusively or shared by the province and district municipalities, as follows: Article 84 – Social Development and Equality of Opportunities functions for the Local Governments 1. Specific functions of province municipalities: 1.1 Plan and promote the social development of their jurisdiction in accordance with the regional and national policies and plans. 1.2. Establish communication channels among the institutions that work for the defense of the rights of children, teenagers, women, disabled people and elder adults. 1.3. Regulate the actions of the Children and Teenagers Municipal Defense (DEMUNA). 1.4. Execute the Glass of Milk program and other programs of nutritional support with the participation of the population and in accordance with the respective regulation, when the district municipality is not able to assume that function. 1.5. Establish communication and cooperation channels among neighbors and social programs. 1.6. Have an updated register of province youth organizations, as well as their active participation in the political, social, cultural and economic life of the LG. 1.7. Create a protection, participation and organization office for disabled neighbors as a dependent program of the Social Services Office. 2. Specific functions of district municipalities: 2.1. Plan and arrange the social development of their jurisdiction in accordance with the regional and province plans, applying participative strategies. 2.2. Acknowledge and register the organizations and institutions that execute social action and promotion coordinating with the local governments. 2.3. Organize, administer and execute the local programs that fight against poverty and the social development programs. 2.4. Organize, administer and execute the local programs of assistance, protection and support for groups of people that are discriminated against. 2.5. Contribute to the design of national, regional and province policies and plans for social development and protection and support for the population in risk. 2.6. Facilitate and participate on the citizenship participation and agreement spaces. 2.7. Promote public and private cooperation in the various local social programs. 2.8. Organize and implement the Children and Teenagers Municipal Defense Service. 2.9. Promote the integral development of the youth in order to promote their physical, psychological, social, moral and spiritual welfare as well as their active participation in the LG. 2.10. Solve conflicts among neighbors and fiscalize the accomplishment of the agreements between the proprietaries of buildings and between local neighborhoods. 2.11. Execute the Glass of Milk program and other programs of nutritional support with the participation of the population and in accordance with the respective regulation. 2.12. Create the Disabled Neighbors Protection, Participation and Organization Office as a program that depends on the Social Services Office. 3. Specific functions shared by district municipalities: 3.1. Disseminate and promote children, teenagers, women, and elder adult rights; favoring spaces for their participation at the municipal stances level. 3.2. Promote, organize and support, in accordance with their possibilities, crèches, nurseries, protection establishments for children, disabled people and elder adults, as well as shelters. 3.3. Promote the equality of opportunities. Source: MOL These soon to be transferred functions are classified into four types: funds, projects, poverty fighting programs and regional scope-productive infrastructure programs. The first three used to 100 be performed by sections of the Ministry of Women and Social Development: MWPHD, and are grouped as follows: a. Assistance and Food Security - PRONAA b. Social and productive investment to assist in the reduction of poverty and capacity- building – FONCODES, COOPOP, PAR, INABIF. c. Support and promotion programs for people and families at risk and in discriminatory circumstances – INABIF, PAR, Wawa Wasi, Program against Family and sexual violence. The process started with PRONAA transfers in the province municipalities and FONCODES in the district municipalities. Initially transferred programs – PRONAA and FONCODES Contribute to the improvement of the nutritional level of boys and Homes and teenagers in moral and health risk that live in homes, shelters or Shelters rehabilitation centers. PRONAA Improve access to food for people with low income that live in Popular Dining urban-marginal and rural sectors, providing food to vulnerable Rooms groups with high nutritional risk through popular dining rooms. Promote community pro-bono work by family members of the Food for work rural and peri-urban families that participate in community work by giving them food as compensation for work. Improving Financing social and economic infrastructure projects oriented to FONCODES your life satisfying basic needs. Promote temporary employment through the rehabilitation, Rural “Let’s maintenance and building of social infrastructure for the benefit work� of the community. Source: PRODES PRONAA was initially created as a Public Decentralized Organism which depended on the Ministry of Women and Social Development (MWPHD) and was basically in charge of nutritional areas and food support. The decentralization of each took place independently because of the differences in design and specialization. On the other hand, FONCODES was created as an executing unit of the Ministry of Women and Social Development (MWPHD) and its functions were to provide minor infrastructure in health, education, hygiene, as well as the development of productive projects. The transfer, in this case, was controlled by local governments. In both cases the transfer process suffered somewhat with respect to the budget and administrative structure of MWPHD, which remained unaltered in terms of staff and administrative structure due to the fact that the Nutrition function accounted for almost 80% of the total functions of this program in the case of PRONAA. Furthermore, it lacks a cost estimation system that can estimate the total costs of the program for the purpose of resource allocation to the local governments. Although later, the Annual and Sectoral Plans were developed and established a sequence in the process of transferring the rest of programmed functions. The decentralization shock accelerated the medium run process and the deadline of 2010 was brought forward to December of 2007. 101 This deadline was then delayed until 2008 and at the moment it runs until the 31 of December of the 2009. Previously, the MWPHD had elaborated a strategy for the transference program for a 3-year period, from 2003 to 2005. This plan included a first stage of capacity strengthening. The second and third stages implied the transference of the competences with credited and enabled governments. Within this plan the MWPHD developed a strategy to transfer the Integral Program of Nutrition, Services of Social Protection; and charities and social participation meetings to Local Governments; whereas the transfer of sectoral functions of population, social development and equal opportunities was programmed for the Regional Governments. - As of December 31st 2008, there were 25 regional governments recognized for the transfer functions of Population, Social Development, Equal Opportunities, 25 of which had nominally subscribed to the functions transfer, 23 of which ended the transfer of sector functions and budget parts and 19 of which had subscribed to management agreements. - For the program, services and projects transfer, we have verified 100% of the local province governments for the PRONAA. 100% of municipal districts have been verified for the Social and Productive Infrastructure Projects, and 73% of the local province governments have been verified for the transfer of the Programs of Integral Nutrition and Programs of Service of Social Protection. - For the transfer of Benefit Societies and Social Participation board of the INABIF, procedures and verification rules were established. Finally, in December 2008, the Resolution of the Secretary of Decentralization Nº 072-2008- PCM/SD was published. In this Resolution, the Special Commission for the transfer of Social Protection Programs and Services was set up. This commission was to be integrated by representatives of MWPHD, the Ministry of Economics and Finance, MEF, the Rural Municipalities Network of Peru: AMPE, the Regional Government Association, ANGR, and the Decentralization Secretary: an institution assigned to the PCM which absorbed the deactivation functions of the National Advisory of Decentralization. This recently formed commission will be responsible for following up the functions transfer process and for elaborating management strategies for Social Protection Programs and Services within the framework of regional and local intervention. It will also define the procedures and tenors to implement the transfer of these programs. There are four main participants within the functions of the Social Programs and the rest of the projects associated with development and social protection:  The Central Government: has the responsibility for designing the line-ups of the different services.  The Regional Government: implements and promotes the implantation of the line-ups defined at the Central level.  The local province government: in charge of delivering services and identifying users in coordination with the local District Government and the Regional level.  The local District Government: facilitates and promotes social participation and contributes mainly to the work of provincial governments. 102 Annex C. Public expenditure per capita by level of government and by function across departments All data refers to accrued amounts spent, as per SIAF Agriculture: public expenditure per capita by department by government type, 2008 (NS) MOQUEGUA 353 142 34 = 529 TACNA 150 122 12 = 284 HUANCAVELICA 70 84 28 = 182 CUSCO 98 47 23 = 167 MADRE DE DIOS 42 119 = 163 ANCASH 81 65 17 = 162 TUMBES 12 20 121 = 153 AYACUCHO 55 51 34 = 139 PASCO 71 21 33 = 125 APURIMAC 42 40 35 = 116 CAJAMARCA 14 65 19 = 99 PUNO 43 12 41 = 96 AREQUIPA 12 46 38 = 96 JUNIN 16 17 58 = 91 LORETO = 81 AMAZONAS = 70 Local government* SAN MARTIN = 52 Regional government LA LIBERTAD = 50 UCAYALI = 48 National government LAMBAYEQUE = 42 PIURA = 42 LIMA = 36 ICA = 33 HUANUCO = 26 P.CONS.DEL CALLAO- - =1 (50) 50 150 250 350 450 550 * Local government reflects the sum of the expenditure of all local governments situated in a given department. Education & culture: public expenditure per capita by department by government type, 2008 (NS) MOQUEGUA 355 615 48 = 1,018 PASCO 281 502 154 = 937 TACNA 232 541 142 = 916 ANCASH 205 421 83 = 710 TUMBES 49 494 129 = 671 CUSCO 134 359 148 = 640 HUANCAVELICA 104 415 96 = 615 APURIMAC 46 441 89 = 576 AYACUCHO 57 396 107 = 560 MADRE DE DIOS 19 398 140 = 557 ICA 21 269 248 = 537 UCAYALI 33 378 102 = 513 LORETO 33 369 99 = 502 AMAZONAS 29 394 54 = 476 PUNO 59 319 81 = 459 CAJAMARCA 54 309 61 = 423 Local government* AREQUIPA 43 265 114 = 422 HUANUCO 42 271 109 = 422 Regional government SAN MARTIN 15 315 70 = 400 JUNIN National government 45 273 60 = 378 LA LIBERTAD 41 232 79 = 352 LIMA 10 34 296 = 340 LAMBAYEQUE 14 214 106 = 334 PIURA 32 214 60 = 307 P.CONS.DEL CALLAO 11 205 74 = 290 (100) 100 300 500 700 900 1,100 * Local government reflects the sum of the expenditure of all local governments situated in a given department. 103 Industry, trade, & services: public expenditure per capita by department by government type, 2008 (NS) MOQUEGUA 289 9 = 299 TACNA 188 5 = 193 ANCASH 39 1 = 40 PASCO 33 3 = 36 CUSCO 20 8 1 = 30 AMAZONAS 7 13 7 = 27 LIMA 3 18 = 21 PUNO 13 4 1 = 18 HUANCAVELICA - 13 2 = 15 JUNIN - 12 2 = 15 AYACUCHO - 9 4 = 13 AREQUIPA 9 0 3 = 12 MADRE DE DIOS 6 6 = 12 LORETO 7 4 = 11 ICA = 11 P.CONS.DEL CALLAO = 10 Local government* CAJAMARCA = 10 UCAYALI = 10 Regional government APURIMAC =8 SAN MARTIN National government =7 HUANUCO =7 PIURA =6 TUMBES =6 LAMBAYEQUE =5 LA LIBERTAD =5 - 50 100 150 200 250 300 * Local government reflects the sum of the expenditure of all local governments situated in a given department. Transportation: public expenditure per capita by department by government type, 2008 (NS) MADRE DE DIOS 40 202 792 = 1,034 MOQUEGUA 715 112 114 = 941 AMAZONAS 48 72 696 = 815 UCAYALI 102 75 361 = 538 PASCO 285 177 67 = 528 TACNA 413 90 7 = 510 TUMBES 138 138 233 = 508 SAN MARTIN 35 142 315 = 491 CUSCO 254 52 128 = 433 ANCASH 216 17 97 = 330 P.CONS.DEL CALLAO 89 18 177 = 284 AYACUCHO 82 70 60 = 212 PUNO 83 38 88 = 210 HUANCAVELICA 89 71 44 = 204 HUANUCO 48 29 124 = 201 AREQUIPA 86 57 36 = 178 Local government* APURIMAC 60 44 74 = 178 LA LIBERTAD 72 47 44 = 163 Regional government CAJAMARCA 80 20 61 = 160 JUNIN National government 84 22 50 = 156 LIMA 83 7 45 = 135 PIURA 63 9 37 = 109 LAMBAYEQUE = 81 ICA = 76 LORETO = 76 (150) 50 250 450 650 850 1,050 * Local government reflects the sum of the expenditure of all local governments situated in a given department. 104 Health: public expenditure per capita by department by government type, 2008 (NS) MOQUEGUA 412 327 134 = 873 MADRE DE DIOS 61 305 113 = 479 TACNA 188 237 52 = 477 APURIMAC 49 241 172 = 462 HUANCAVELICA 73 184 164 = 421 AMAZONAS 69 194 152 = 415 PASCO 209 132 60 = 402 TUMBES 102 181 112 = 394 ANCASH 177 109 86 = 371 HUANUCO 61 122 183 = 366 AYACUCHO 76 147 141 = 363 ICA 72 158 121 = 351 LORETO 69 139 138 = 347 CUSCO 116 142 85 = 343 LIMA 67 18 259 = 343 UCAYALI 74 141 120 = 336 Local government* P.CONS.DEL… 95 28 203 = 326 PIURA 75 101 113 = 289 Regional government JUNIN 54 150 80 = 285 AREQUIPA 75 168 27 = 270 National government LA LIBERTAD 53 110 90 = 252 LAMBAYEQUE 89 98 65 = 252 SAN MARTIN = 238 PUNO = 225 CAJAMARCA = 224 - 100 200 300 400 500 600 700 800 900 * Local government reflects the sum of the expenditure of all local governments situated in a given department. Housing & urban development: public expenditure per capita by department by government type, 2008 (NS) MOQUEGUA 133 1 = 133 TACNA 88 1 = 88 ANCASH 62 2 3 = 68 PASCO 53 1 = 54 LIMA 11 42 = 54 TUMBES 32 3 9 = 44 ICA 7 1 22 = 29 CUSCO 27 2 = 29 HUANCAVELICA 16 5 1 = 22 LORETO 12 6 = 19 AMAZONAS 6 2 9 = 17 P.CONS.DEL CALLAO 16 = 16 APURIMAC 9 5 0 = 15 PUNO 14 0 = 14 PIURA 11 1 = 13 JUNIN 12 = 13 Local government* AYACUCHO 10 1 2 = 12 UCAYALI 11 1 = 12 Regional government AREQUIPA 10 1 1 = 11 MADRE DE DIOS National government 6 1 3 = 11 LA LIBERTAD 9 1 = 10 SAN MARTIN 8 2 = 10 CAJAMARCA =9 LAMBAYEQUE =8 HUANUCO =8 - 20 40 60 80 100 120 140 * Local government reflects the sum of the expenditure of all local governments situated in a given department. 105 106 Annex D. Efficiency analysis Table D.1: List of outcome and expenditures indicators by sector Sector: Education Outcomes (and outputs) AVG SD MIN MAX SOURCE % of students completing primary edu with satisfactory reading comprehension 13.67 7.65 3.70 31.30 RBB % of students completing primary edu with satisfactory math skills 7.24 2.91 2.10 13.60 RBB Enrolment in first year of primary edu, with 7 years of age or less 96.44 1.79 90.85 98.89 RBB % of students enrolled in first year of primary edu that have attended preschool 85.07 11.70 55.28 100.00 RBB Net coverage for secondary school 73.27 9.11 57.49 89.87 INEI Rate of delay at secondary school 17.95 6.54 10.08 29.08 INEI Expenditure Per capital spending in education and culture 401.61 109.18 255.29 667.77 SIAF Sector: Health Outcomes (and outputs) Prevalence of chronic malnutrition in children under 5 years old (WHO standard) 0.05 0.03 0.02 0.16 RBB Infants under 6 months old exclusively breastfed 0.02 0.00 0.01 0.03 RBB Prevalence of anemia in infants under 36 months old 0.02 0.00 0.01 0.02 RBB Prevalence of diarrhea (acute) in infants under 36 months old 0.06 0.02 0.03 0.09 RBB Prevalence of acute respiratory infections in infants under 36 months old 0.05 0.02 0.03 0.09 RBB Prevalence of low weight at birth 0.14 0.03 0.08 0.23 RBB Maternal mortality rate at giving birth 0.08 0.07 0.04 0.37 RBB Prevalence of birth control use 72.39 4.11 60.50 80.60 RBB Percent of births at health care facilities, out of mothers from rural areas 56.23 20.46 11.90 93.10 RBB Percent of births at health care facilities, in last five years 75.05 16.18 43.90 97.20 RBB Health insurance coverage 56.28 8.70 43.79 75.14 INEI Expenditure Per capital spending in health 213.45 76.13 121.69 480.23 SIAF Sector: Transport Outcomes (and outputs) Time to reach a health center on foot 0.01 0.01 0.00 0.02 RBB Time to reach an education center on foot 0.02 0.02 0.00 0.06 RBB Time to reach a market on foot 0.01 0.03 0.00 0.15 RBB Share of Rehabilitated roads 0.05 0.05 0.00 0.18 RBB Share of paved roads 0.14 0.10 0.01 0.40 PNUD Number of car accidents per 10000 inhabitants 1.51 1.36 0.06 5.81 CIDATT Expenditure Per capital spending in ground transportation 184.51 142.69 43.78 554.21 SIAF 107 Table D.2: List of outcome and expenditures indicators by sector Sub-Sector: Initial and Primary Education Outcomes (and outputs) AVG SD MIN MAX SOURCE % of students completing primary edu with satisfactory reading comprehension 13.67 7.65 3.70 31.30 RBB % of students completing primary edu with satisfactory math skills 7.24 2.91 2.10 13.60 RBB Enrolment in first year of primary edu, with 7 years of age or less 96.44 1.79 90.85 98.89 RBB % of students enrolled in first year of primary edu that have attended preschool 85.07 11.70 55.28 100.00 RBB Expenditure Per capital spending in initial and primary education 29.98 14.18 13.97 80.26 SIAF Sub-Sector: Nutrition Outcomes (and outputs) Prevalence of chronic malnutrition in children under 5 years old (WHO standard) 0.05 0.03 0.02 0.16 RBB Infants under 6 months old exclusively breastfed 0.02 0.00 0.01 0.03 RBB Prevalence of anemia in infants under 36 months old 0.02 0.00 0.01 0.02 RBB Prevalence of diarrhea (acute) in infants under 36 months old 0.06 0.02 0.03 0.09 RBB Prevalence of acute respiratory infections in infants under 36 months old 0.05 0.02 0.03 0.09 RBB Prevalence of low weight at birth 0.14 0.03 0.08 0.23 RBB Expenditure Per capital spending in childhood nutrition 49.93 39.44 6.49 139.25 SIAF Sub-Sector: Maternal and Neonatal Health Outcomes (and outputs) Maternal mortality rate at giving birth 0.08 0.07 0.04 0.37 RBB Prevalence of birth control use 72.39 4.11 60.50 80.60 RBB Percent of births at health care facilities, out of mothers from rural areas 56.23 20.46 11.90 93.10 RBB Percent of births at health care facilities, in last five years 75.05 16.18 43.90 97.20 RBB Expenditure Per capital spending in maternal health 13.64 7.93 4.89 33.48 SIAF Table D.3: Grouping according to geographic regions and natural resource endowment Group Regions Costa Ica, La Libertad, Lambayeque, Lima, Moquegua, Piura, Tacna, Tumbes Sierra Ancash, Apurimac, Arequipa, Ayacucho, Cajamarca, Cusco, Huancavelica, Huanuco, Junin, Pasco, Puno Selva Amazonas, Loreto, Madre de Dios, San Martin, Ucayali Natural resource- Ancash, Cusco, Moquegua, Pasco, Tacna rich Amazonas, Apurimac, Arequipa, Ayacucho, Cajamarca, Huancavelica, Huanuco, Junin, Ica, La Libertad, Other Lambayeque, Lima, Loreto, Madre de Dios, Piura, Puno, San Martin, Tumbes, Ucayali 108 Figure D.1: Ranks and Scores Education DEPARTMENT PSP SCORE PSP RANK PSE SCORE PSE RANK SPENDING PER CAPITA SPENDING RANK AMAZONAS 0.77 8 0.70 10 370.36 13 ANCASH 0.75 10 0.55 16 463.51 5 APURIMAC 0.75 11 0.58 14 433.93 9 AREQUIPA 0.96 2 0.99 2 325.63 17 AYACUCHO 0.70 17 0.54 17 432.79 10 CAJAMARCA 0.76 9 0.75 8 342.15 16 CUSCO 0.73 14 0.58 13 417.48 11 HUANCAVELICA 0.73 13 0.56 15 436.79 8 HUANUCO 0.68 19 0.71 9 318.50 19 ICA 0.80 5 0.75 7 359.72 14 JUNIN 0.78 7 0.85 5 304.92 20 LA LIBERTAD 0.72 15 0.94 4 256.60 23 LAMBAYEQUE 0.79 6 1.00 1 265.18 22 LIMA 0.86 4 0.95 3 302.55 21 LORETO 0.59 23 0.45 22 438.14 7 MADRE DE DIOS 0.66 20 0.49 20 447.19 6 MOQUEGUA 1.00 1 0.50 19 667.77 1 PASCO 0.71 16 0.44 23 539.82 4 PIURA 0.64 21 0.84 6 255.29 24 PUNO 0.68 18 0.63 11 358.59 15 SAN MARTIN 0.60 22 0.63 12 322.44 18 TACNA 0.91 3 0.51 18 593.74 2 TUMBES 0.74 12 0.44 24 572.40 3 UCAYALI 0.58 24 0.47 21 413.25 12 Source: World Bank staff estimates Figure D.2: Ranks and Scores Health DEPARTMENT PSP SCORE PSP RANK PSE SCORE PSE RANK SPENDING PER CAPITA SPENDING RANK AMAZONAS 0.51 18 0.44 19 228.97 7 ANCASH 0.54 12 0.48 17 222.50 9 APURIMAC 0.57 9 0.43 20 257.15 5 AREQUIPA 0.71 6 0.77 4 182.10 15 AYACUCHO 0.53 14 0.48 16 215.11 11 CAJAMARCA 0.48 21 0.78 3 121.69 24 CUSCO 0.56 10 0.62 11 177.23 16 HUANCAVELICA 0.48 22 0.47 18 200.86 13 HUANUCO 0.51 19 0.57 13 174.59 17 ICA 0.72 5 0.69 6 204.11 12 JUNIN 0.50 20 0.58 12 167.67 18 LA LIBERTAD 0.55 11 0.79 2 136.42 22 LAMBAYEQUE 0.68 7 1.00 1 132.61 23 LIMA 0.76 3 0.56 14 265.17 4 LORETO 0.41 24 0.36 23 217.88 10 MADRE DE DIOS 0.64 8 0.39 21 325.74 2 MOQUEGUA 0.78 2 0.32 24 480.23 1 PASCO 0.47 23 0.38 22 242.91 6 PIURA 0.54 13 0.66 8 161.18 19 PUNO 0.53 15 0.74 5 139.93 21 SAN MARTIN 0.52 17 0.64 9 160.13 20 TACNA 1.00 1 0.69 7 285.26 3 TUMBES 0.72 4 0.62 10 228.24 8 UCAYALI 0.52 16 0.53 15 195.14 14 Source: World Bank staff estimates 109 Figure D.3: Ranks and Scores Transport DEPARTMENT PSP SCORE PSP RANK PSE SCORE PSE RANK SPENDING PER CAPITA SPENDING RANK AMAZONAS 0.35 6 0.05 14 346.65 3 ANCASH 0.36 5 0.08 10 212.78 8 APURIMAC 0.07 20 0.03 18 114.21 13 AREQUIPA 0.45 2 0.30 2 75.04 18 AYACUCHO 0.23 14 0.08 9 134.96 12 CAJAMARCA 0.16 16 0.07 11 112.46 15 CUSCO 0.27 9 0.06 12 208.92 9 HUANCAVELICA 0.09 18 0.02 21 186.84 11 HUANUCO 0.27 11 0.16 8 83.23 17 ICA 0.25 13 0.19 7 62.91 19 JUNIN 0.07 21 0.03 19 112.54 14 LA LIBERTAD 0.29 7 0.24 4 60.55 20 LAMBAYEQUE 0.17 15 0.20 6 43.78 24 LIMA 1.00 1 1.00 1 49.67 23 LORETO 0.04 23 0.03 17 56.91 21 MADRE DE DIOS 0.03 24 0.00 24 554.21 1 MOQUEGUA 0.43 3 0.04 16 505.63 2 PASCO 0.06 22 0.01 23 329.53 4 PIURA 0.29 8 0.26 3 55.82 22 PUNO 0.40 4 0.22 5 90.90 16 SAN MARTIN 0.26 12 0.06 13 205.43 10 TACNA 0.09 19 0.01 22 288.33 6 TUMBES 0.12 17 0.03 20 219.88 7 UCAYALI 0.27 10 0.04 15 317.04 5 Source: World Bank staff estimates Figure D.4: Ranks and Scores Initial and Primary Education DEPARTMENT PSP SCORE PSP RANK PSE SCORE PSE RANK SPENDING PER CAPITA SPENDING RANK AMAZONAS 0.64 11 0.39 16 97.80 6 ANCASH 0.62 13 0.43 13 81.77 10 APURIMAC 0.60 15 0.30 20 108.74 2 AREQUIPA 0.96 2 0.94 2 50.66 22 AYACUCHO 0.50 20 0.32 18 102.41 3 CAJAMARCA 0.66 9 0.47 11 85.88 8 CUSCO 0.58 17 0.52 9 68.68 16 HUANCAVELICA 0.55 18 0.31 19 99.85 4 HUANUCO 0.48 21 0.45 12 72.29 15 ICA 0.77 5 0.56 8 62.18 18 JUNIN 0.73 7 0.73 4 63.93 17 LA LIBERTAD 0.65 10 0.72 5 52.33 21 LAMBAYEQUE 0.75 6 1.00 1 47.47 24 LIMA 0.83 4 0.76 3 49.58 23 LORETO 0.40 24 0.18 24 98.08 5 MADRE DE DIOS 0.50 19 0.29 21 78.95 13 MOQUEGUA 1.00 1 0.42 14 91.32 7 PASCO 0.63 12 0.41 15 82.84 9 PIURA 0.59 16 0.64 6 55.36 20 PUNO 0.60 14 0.50 10 76.48 14 SAN MARTIN 0.47 22 0.33 17 81.28 11 TACNA 0.88 3 0.59 7 61.14 19 TUMBES 0.71 8 0.22 23 111.94 1 UCAYALI 0.44 23 0.25 22 79.65 12 Source: World Bank staff estimates 110 Figure D.5: Ranks and Scores Nutrition DEPARTMENT PSP SCORE PSP RANK PSE SCORE PSE RANK SPENDING PER CAPITA SPENDING RANK AMAZONAS 0.53 13 0.08 19 64.13 6 ANCASH 0.58 11 0.10 17 50.63 8 APURIMAC 0.49 19 0.03 23 139.25 1 AREQUIPA 0.78 3 0.67 2 10.55 23 AYACUCHO 0.48 20 0.04 22 117.08 3 CAJAMARCA 0.50 17 0.05 20 84.99 5 CUSCO 0.52 14 0.08 18 60.91 7 HUANCAVELICA 0.46 22 0.03 24 134.14 2 HUANUCO 0.48 21 0.04 21 114.21 4 ICA 0.72 6 1.00 1 6.49 24 JUNIN 0.50 18 0.14 11 33.00 13 LA LIBERTAD 0.59 10 0.11 16 47.19 10 LAMBAYEQUE 0.67 8 0.27 7 22.33 20 LIMA 0.75 5 0.49 4 13.82 21 LORETO 0.42 24 0.13 14 29.83 15 MADRE DE DIOS 0.67 7 0.20 8 30.23 14 MOQUEGUA 0.94 2 0.33 5 25.49 19 PASCO 0.42 23 0.13 12 29.06 16 PIURA 0.56 12 0.18 9 28.93 17 PUNO 0.62 9 0.12 15 48.17 9 SAN MARTIN 0.51 15 0.14 10 33.19 12 TACNA 1.00 1 0.33 6 27.35 18 TUMBES 0.76 4 0.58 3 11.88 22 UCAYALI 0.51 16 0.13 13 35.48 11 Source: World Bank staff estimates Figure D.6: Ranks and Scores Maternal and Neonatal Health DEPARTMENT PSP SCORE PSP RANK PSE SCORE PSE RANK SPENDING PER CAPITA SPENDING RANK AMAZONAS 0.41 20 0.36 11 8.14 20 ANCASH 0.45 17 0.53 3 6.17 22 APURIMAC 0.56 8 0.14 23 27.94 3 AREQUIPA 0.58 6 0.36 10 11.57 12 AYACUCHO 0.48 11 0.11 24 31.64 2 CAJAMARCA 0.38 22 0.31 12 8.89 17 CUSCO 0.53 10 0.27 13 14.22 9 HUANCAVELICA 0.40 21 0.23 15 12.32 11 HUANUCO 0.43 18 0.21 19 14.80 7 ICA 0.69 3 1.00 1 4.89 24 JUNIN 0.46 16 0.37 8 8.73 18 LA LIBERTAD 0.43 19 0.23 16 13.40 10 LAMBAYEQUE 0.63 4 0.39 5 11.44 13 LIMA 0.75 2 0.37 9 14.50 8 LORETO 0.28 24 0.22 17 9.07 15 MADRE DE DIOS 0.58 7 0.19 22 21.94 4 MOQUEGUA 0.54 9 0.21 20 18.64 5 PASCO 0.47 12 0.47 4 7.15 21 PIURA 0.46 15 0.38 6 8.51 19 PUNO 0.34 23 0.25 14 9.92 14 SAN MARTIN 0.46 14 0.21 21 16.04 6 TACNA 1.00 1 0.21 18 33.48 1 TUMBES 0.61 5 0.89 2 4.94 23 UCAYALI 0.47 13 0.37 7 8.97 16 Source: World Bank staff estimates 111 Figure D.7: Performance and Efficiency scores Primary and Initial Education 1.0 1.0 LAMBAYEQ AREQUIPA MOQUEGUA UE AREQUIPA 0.9 0.9 TACNA LIMA 0.8 LA 0.8 LAMBAYEQ ICA JUNIN LIMA LIBERTAD 0.7 UE JUNIN PIURA 0.7 PSP 0.6 PSE LA PASCO TACNA LIBERTAD AMAZONAS CUSCO ICA 0.6 CUSCO CAJAMARCA ANCASH APURIMAC 0.5 PUNO PIURA PUNO CAJAMARC HUANCAVEL HUANUCO MADRE DE ANCASHA MOQUEGU 0.5 ICA 0.4 PASCO HUANUCO SAN AYACUCHO SAN A DIOS AYACUCHO AMAZONAS MARTIN 0.3 HUANCAVE MARTIN DE APURIMAC MADRE 0.4 UCAYALI LORETO DIOS LICA UCAYALI 0.2 TUMBES LORETO 0.3 0.1 40.0 60.0 80.0 100.0 0.3 0.5 0.7 0.9 1.1 SPENDING PER CAPITA PSP Nutrition 1.0 ICA 1.0 TACNA 0.9 MOQUEGUA 0.8 AREQUIPA 0.8 0.6 AREQUIPA TUMBES PSE LIMA TUMBES PSP 0.7 ICA LIMA MADRE DE 0.4 DIOS LAMBAYEQ MOQUEGU PUNO LA TACNA 0.6 UE A ANCASH LIBERTAD LAMBAYEQ PIURA MADRE DE SAN CUSCO AMAZONAS 0.2 SANPIURA UE DIOS LAPUNO 0.5 JUNIN MARTIN UCAYALI APURIMAC CAJAMARCA AYACUCHO HUANUCO HUANCAVEL LORETOJUNIN ANCASH PASCO CUSCO MARTIN UCAYALILIBERTAD HUANUCO CAJAMARC AMAZONAS AYACUCHO APURIMAC HUANCAVE PASCO LORETO I 0.0 A 0.4 0.4 LICA0.6 0.8 1.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 SPENDING PER CAPITA PSP Maternal and neonatal Health 1.0 ICA 1.0 TACNA TUMBES 0.8 0.8 LIMA 0.6 ICA PSE PSP LAMBAYEQ MADRE DE ANCASH 0.6 TUMBES AREQUIPA UE CUSCO PASCO APURIMAC DIOS MOQUEGUA 0.4 PIURA LAMBAYEQ PASCO LA SAN JUNIN AREQUIPA LIMA UCAYALI PIURA AYACUCHO JUNIN UCAYALILIBERTAD ANCASH HUANUCO MARTIN CAJAMARC CUSCO UE AMAZONAS 0.4 AMAZONASHUANCAVEL PUNOLA HUANCAVE MADRE DE HUANUCOA SAN MOQUEGU TACNA PUNO ICA CAJAMARCA 0.2 LORETO LIBERTAD LICA DIOS A MARTIN APURIMAC LORETO AYACUCHO 0.2 0.0 3.0 8.0 13.0 18.0 23.0 28.0 33.0 0.2 0.4 0.6 0.8 1.0 SPENDING PER CAPITA PSP Note: Efficiency scores are base on public spending figures for 2008 only. SIAF do not present disaggregated data at program level for the previous year. Source: World Bank staff estimates. 112 Figure D.8: Key correlates of PSP and PSE in Initial and Primary Education PSP,PSE & GDP PER CAPITA PSP,PSE & POPULATION DENSITY 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.0 20.0 40.0 60.0 80.0 100.0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 GDP Per Capita POPULATION DENSITY PSP PSE PSP PSE Linear (PSP ) Linear (PSE) Linear (PSP ) Linear (PSE) PSP,PSE & LOCAL SPENDING AS A SHARE OF TOTAL PSP,PSE LOCAL REVENUES AS A SHARE OF TOTAL 1.2 SPENDING SPENDING 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.9 0.9 1.0 1.0 1.1 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 share of local expenditure share of local revenue PSP PSE PSP PSE Linear (PSP ) Linear (PSE) Linear (PSP) Linear (PSE) PSE & SPENDING PER CAPITA AVERAGE SCORES 1.2 0.9 0.8 1.0 0.7 0.8 0.6 0.5 0.6 0.4 0.4 0.3 0.2 0.2 0.1 0.0 0.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0 Costa Sierra Selva Natural Other resource SPENDING PER CAPITA Rich PSP PSE PSE Linear (PSE) Source: World Bank staff estimates 113 Figure D.9: Key correlates of PSP and PSE in Nutrition PSP,PSE & GDP PER CAPITA PSP,PSE & POPULATION DENSITY 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0 5,000 10,000 15,000 20,000 25,000 30,000 0.0 20.0 40.0 60.0 80.0 100.0 POPULATION DENSITY GDP Per Capita PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSP,PSE & LOCAL SPENDING AS A SHARE OF TOTAL PSP,PSE & LOCAL REVENUE AS A SHARE OF TOTAL 1.2 SPENDING SPENDING 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 share of local expenditure share of local revenue PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSE & SPENDING PER CAPITA AVERAGE SCORES 1.2 0.8 0.7 1.0 0.6 0.8 0.5 0.6 0.4 0.4 0.3 0.2 0.2 0.1 0.0 0.0 0.0 50.0 100.0 150.0 Costa Sierra Selva Natural Other (0.2) resource TOTAL PUBLIC SPENDING Rich PSP PSE PSE Linear (PSE) Note: Efficiency score are based on public spending figures for 2008 only. SIAF do not present disaggregated data at program level for the previous year. Source: World Bank staff estimates 114 Figure D.10: Key correlates of PSP and PSE Maternal and Neonatal Health PSP,PSE & GDP PER CAPITA PSP,PSE & POPULATION DENSITY 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0 5,000 10,000 15,000 20,000 25,000 30,000 0.0 20.0 40.0 60.0 80.0 100.0 GDP Per Capita POPULATION DENSITY PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSP,PSE & LOCAL SPENDING AS A SHARE OF TOTAL PSP,PSE & LOCAL REVENUES AS A SHARE OF TOTAL SPENDING SPENDING 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 share of local expenditure share of local revenue PSP PSE PSP PSE Linear (PSP) Linear (PSE) Linear (PSP) Linear (PSE) PSE & SPENDING PER CAPITA AVERAGE SCORES 1.2 0.7 1.0 0.6 0.5 0.8 0.4 0.6 0.3 0.4 0.2 0.2 0.1 0.0 0.0 0.0 10.0 20.0 30.0 40.0 Costa Sierra Selva Natural Other resource TOTAL PUBLIC SPENDING Rich PSP PSE PSE Linear (PSE) Note: Efficiency score are based on public spending figures for 2008 only. SIAF do not present disaggregated data at program level for the previous year. Source: World Bank staff estimates 115