ADEX Asociación de Exportadores (Exporters’ Association) AFP Administradoras Privadas de Fondos de Pensiones ANA Autoridad Nacional del Agua (National Water Authority) ASBANC Asociación de Bancos del Perú (Peruvian Association of Banks) Comisión de Eliminación de Barreras Burocráticas (Committee for the CEBB Elimination of Bureaucratic Barriers) Centro de Estudios Distributivos Laborales y Sociales (Center for CEDLAS Distributive and Social Studies) CIAS Comisión Interministerial de Asuntos Sociales Centro de Investigación Forestal (Center for International Forestry CIFOR Research) COP Paris 21st Community of Parties COPD Chronic Obstructive Pulmonary Disease CPF Country Partnership Framework DHS Demographic and Health Survey ECA Estándar de Calidad Ambiental (Environmental Quality Standard) ECE Evaluación Censal de Estudiantes (National Student Assessment) ENAHO Encuesta Nacional de Hogares (National Household Survey) Estrategia Nacional de Desarrollo Rural (National Rural Development ENDER Strategy) EPS Empresa Pública de Servicios (Water Provision Enterprise) FDI Foreign Direct Investment FED Fondo de Estímulo al Desempeño (Development Incentive Fund) Fondo de Desarrollo Socioeconómico de Camisea (Socioeconomic FOCAM Development Fund of the Camisea Project) Fondo de Cooperación para el Desarrollo Social (Social Development FONCODES Cooperation Fund) FONCOMUN Fondo de Compensación Municipal (Municipal Compesation Fund) FONCOR Fondo de Compensación Regional (Regional Compensation Fund) Fondo para la Inclusión Económica en Zonas Rurales (Rural Economic FONIE Inclusion Fund) Fondo de Promoción a la Inversión Pública Regional y Local (Fund for the FONIPREL Promotion of Regional and Local Public Investment) GDP Gross Domestic Product GNI Gross National Income HOI Human Opportunities Index HSAI Health Services Availability Index IDB Inter-American Development Bank IEP Instituto de Estudios Peruanos (Peruvian Studies Institute) IFC International Finance Corporation i ILO International Labor Organization IMF International Monetary Fund Instituto Nacional de Defensa de la Competencia y de la Protección de la Propiedad Intelectual (National Institute for the Defense of Competition and INDECOPI Intellectual Property Rights) Instituto Nacional de Estadistica e Infomatica (National Institute of Statistics INEI and Informatics) IT Information Technology IVQ Individual Vessel Quotas LAC Latin America and the Caribbean LAPOP Latin American Public Opinion Project LOPE Ley Orgánica del Poder Ejecutivo (Executive Power Organic Law) LPI Logistics Performance Index LRI Lower Respiratory Infections MDG Millennium Development Goal MEF Ministerio de Economía y Finanzas (Ministry of Economy and Finance) Ministerio de Desarrollo e Inclusión Social (Ministry of Development and MIDIS Social Inclusion) MINAM Ministerio del Ambiente (Ministry of Environment) MINEDU Ministerio de Educación (Ministry of Education) MINEM Ministerio de Energía y Minas (Ministry of Energy and Mining) MRTA Tupac Amaru Revolutionary Movement NRW Non-Revenue Water NTMs Non-Tariff Measures OECD Organization for Economic Co-operation and Development Organismo de Evaluación y Fiscalización Ambiental (Agency for OEFA environmental Assessment and Control) OPI Oficinas de Programación de Inversiones (Investment Programming Offices) ONP Oficina de Normalización Previsional PAN Programa Articulado Nutricional (Articulated Nutritional Program) PCL Public Consultations Law PCPSL Communist Party of Peru—Shining Path PEN Peruvian Nuevo Sol PISA Program for International Student Assessment PM Particulate Matter PPP Purchasing Power Parity Centro de Estudios y Prevención de Desastres (Center for the Study and PREDES Prevention of Disasters) PRODUCE Ministerio de la Producción (Ministry of Production) Programa Nacional de Asistencia Alimentaria (National Food Assistance PRONAA Program) ii Programa Nacional de Manejo de Cuencas Hidrográficas y Conservación de Suelos (National Water Source Management and Soil Conservation PRONAMACHCS Program) PROVIAS Programa Nacional de Vías (National Roads Program) RENAMU Registro Nacional de Municipalidades (National Municipal Registry) SCD Systematic Country Diagnostic Servicio de Agua Potable y Alcantarillado de La Libertad, Sociedad SEDALIB Anónima (La Libertad Water and Sanitation Service, INC.) SEDLAC Socio-Economic Data Base data base for Latin America and the Caribbean SERVIR Autoridad Nacional del Servicio Civil (National Authority of Civil Servants) SIS Sistema Integrado de Salud (Health Integrated System) SISFOH Sistema de Focalización de Hogares (Household Targeting System) SMEs Small and medium-sized enterprises SNIP Sistema Nacional De Inversión Pública (National Public Investment System) SOEs State-Owned Enterprises Superintendencia Nacional de Aduanas y de Administración Tributaria SUNAT (National Customs and Tax Administration Superintendence) SUSALUD Superintendencia Nacional de Salud (National Health Superintendence) TFP Total Factor Productivity TRC Truth and Reconciliation Commission U.S. United States of America UNFCCC Framework Convention of the United Nations on Climate Change UNODC United Nations Office on Drugs and Crime WDI World Development Indicators WEF World Economic Forum WHO World Health Organization iii We would like to thank the members of the Peru Country Team from all Global Practices and IFC, as well as all the stakeholders in Peru (government authorities, think-tanks, academia, and civil society organizations), who have contributed to the preparation of this document. We are very grateful for their inputs, knowledge and advice. The team was co-led by Ana Maria Oviedo (Senior Economist, GPVDR), Marc Schiffbauer (Senior Economist, GMFDR), and Luciana Harrington (Strategy Officer, CBCCE). In addition, the following people provided overall guidance and substantive inputs: Alberto Rodriguez (Country Director, Andean Countries), Eduardo Wallentin (Senior Manager, CBCCF), Marc Tristant (Country Head & Principal Investment Officer, IFC), Oscar Calvo-Gonzalez (Practice Manager, GPVDR), Pablo Saavedra (Practice Manager, GMFDR), Marcelo Selowsky (Special Advisor to the Peru SCD), Livia Benavides (Program Leader, LC6CC), Pedro L. Rodriguez (Program Leader, LC6CC), Oliver Braedt (Program Leader, LC6CC), and Jamele Rigolini (Lead Economist, GSP01). The following members of the SCD team contributed substantial inputs or support to this document: Antonio Skarica (CBCCF), Juan Jose Rossel (CNGMI), Daniel Barco, Melanie Laloum (GMFDR), Hugo Brousset (GSPDR), Gabriela Farfan, Gonzalo Rivera (GPVDR), Christel Vermeersch, Andre Medici (GHNDR), Ines Kudo, Luciana Velarde, Javier Botero (GEDDR), Fernando Blanco (GMFDR), Klas Sander, Raul Tolmos, Fernando Loayza (GENDR), Juan Jose Miranda (GENGE), Michael Morris, Griselle Vega, Luz Diaz (GFADR), Habab Taifour, Gustavo Saltiel, Caroline Van den Berg, Iris Marmanillo, Gustavo Perochena, Malva Baskovich (GWA04), Alvaro Quijandria, Tanja Goodwin, Marta Licetti (GTCDR), Irene Portables, Georges Darido (GTIDR), Zoe Trohanis, Fernando Ramirez Cortes (GSURB), Alberto Leyton, Jorge Luis Silva Mendez (GGOPS), Janina Franco, Javier Aguilar (GEEDR), Marcos Castro (GCCCF), Leyla Castillo (GFM04), German Freire, Sergi Perez (GSURR), Juan Pablo Checura, Frankz Kastner, Karina Olivas, Guillermo Romero (LCC6C), Carol Yagui, Luisa Yesquen, and Itami Okumura (LCC6C), Clifton Wiens and Suzana Abbott (Editors). The team would like to thank the peer reviewers, Omar Arias (GSPDR), Luc Dejonckheere (CNGP7), and Truman G. Packard (GSP04) for their comments, and Augusto de la Torre, Daniel Lederman (LCRCEO) and Jorge Araujo (LCRVP) for their advice and guidance. The team would also like to thank officials from the Ministry of Economy and Finance for the comments and feedback received throughout the process. However, the views of the SCD represent those of the World Bank Group alone. Any factual errors are their responsibility. iv Abbreviations and Acronyms ......................................................................................................................................................................i Acknowledgements ....................................................................................................................................................................................... iv 1. Overview .................................................................................................................................................................................................. 9 From a failed economy to a success story.....................................................................................................................................11 New exogenous conditions highlight Peru’s specific structural challenges ahead ....................................................14 Areas for priority action .......................................................................................................................................................................18 Process and structure of the SCD ......................................................................................................................................................25 2. Poverty and Inclusion .......................................................................................................................................................................27 A remarkable decade of poverty reduction and increased shared prosperity .............................................................27 Poverty in Peru is predominantly rural .........................................................................................................................................31 Basic services and human development have accompanied poverty reduction, but gaps remain .....................35 3. Economic Growth ...............................................................................................................................................................................50 A decade of fast economic growth and convergence ...............................................................................................................50 Past growth was driven by private capital accumulation ......................................................................................................51 Going forward, growth will have to rely on higher productivity ........................................................................................63 4. Sustainability ........................................................................................................................................................................................73 Peru faces important environmental sustainability risks .....................................................................................................73 Maintaining fiscal sustainability could become difficult ........................................................................................................83 Institutional weaknesses lead to a low level of government coordination in services delivery ..........................91 5. Defining Priority Areas for Action ........................................................................................................................................... 105 Applying prioritization criteria ...................................................................................................................................................... 107 Improving connecting infrastructure and public services ................................................................................................. 109 Raising human capital......................................................................................................................................................................... 121 Reducing factor and product market rigidities ....................................................................................................................... 130 Supporting and sustainability conditions .................................................................................................................................. 142 Conclusion ................................................................................................................................................................................................ 144 References ..................................................................................................................................................................................................... 146 Appendix A: Country Comparators .................................................................................................................................................... 155 Appendix B: Poverty Profiles ................................................................................................................................................................ 157 Annex C: IFC industry notes .................................................................................................................................................................. 160 Infrastructure ......................................................................................................................................................................................... 160 Agribusiness ............................................................................................................................................................................................ 164 Public-Private Partnerships ............................................................................................................................................................. 167 Financial Sector ..................................................................................................................................................................................... 170 Health and Education .......................................................................................................................................................................... 178 Mining ........................................................................................................................................................................................................ 180 List of Figures Figure 1: Per capita income started to converge to U.S. levels in the early 2000s...........................................................12 Figure 2: The combination of macro-structural reforms and favorable exogenous conditions created a virtuous cycle of growth and shared prosperity ....................................................................................................................................13 Figure 3: Micro-structural reforms are needed to continue on the path of shared prosperity .................................18 Figure 4: Evolution of poverty, Peru and LAC, 2000–14 .............................................................................................................27 Figure 5: Labor income contributed the most to poverty reduction, followed by transfers in rural areas ..........29 Figure 6: Income inequality measured by the Gini coefficient, by rural/urban, Peru 2004–15 ................................30 Figure 7: Income growth (%), Bottom 40 and the Entire Population – LAC, 2003–13..................................................30 Figure 8: Middle class and vulnerable population, Peru and LAC, 2004–14 ......................................................................31 v Figure 9: Poverty headcount (%) in urban and rural areas, 2004 and 2015 .....................................................................32 Figure 10: District-level poverty, 2009 and 2013 ..........................................................................................................................32 Figure 11: Progress in extreme poverty rates by region, 2004–15 ........................................................................................33 Figure 12: Distribution of districts, by changes in poverty rates between 2007 and 2012–13.................................34 Figure 13: Improved access to water and sanitation, 2015 .......................................................................................................37 Figure 14: Households with drinking water, 2015 ........................................................................................................................38 Figure 15: Households with safe water, 2015 ..................................................................................................................................38 Figure 16: Neonatal, infant and child mortality rates by area of residence, 2013–15. ..................................................39 Figure 17: Maternal mortality, 1990–2015. ......................................................................................................................................39 Figure 18: Stunting rates for children under 5 by area, 2005–15 ...........................................................................................40 Figure 19: Prevalence of anemia in children between 6 and 36 months by area, 2005–15 ........................................40 Figure 20: Share of students that reached a satisfactory level in ECE, 2007–15 ..............................................................41 Figure 21: Performance of students in reading and mathematics in PISA, 2015 .............................................................41 Figure 22: Employed distribution and informal work by sector in Peru, 2004–15 ........................................................42 Figure 23: Indigenous population in LAC ...........................................................................................................................................43 Figure 24: Indigenous and Non-indigenous poverty (%), 2005 and 2015 ..........................................................................43 Figure 25: Deprivation of services and poor housing conditions for urban populations (2010) .............................44 Figure 26: Stunting among indigenous children, 2013 ................................................................................................................45 Figure 27: Share of above-age students by ethnicity, 2014 .......................................................................................................45 Figure 28: Indigenous 4th graders’ performance in Spanish as a second language in ECE, 2014–15......................45 Figure 29: Gender gaps in education performance ........................................................................................................................48 Figure 30: Domestic violence by income quintile ...........................................................................................................................49 Figure 31: Peru’s GDP per capita has recovered in the past 15 years ...................................................................................51 Figure 32: Growth was driven by private sector capital accumulation ................................................................................52 Figure 33: More than half of the US61 billion Peru received in FDI from 2006-2016 was in mining .....................53 Figure 34: 10-70 percent of the output of several large domestic sectors is sold directly to mining firms .........54 Figure 35: Countercyclical fiscal policy stabilized the economy and contributed to high savings ...........................56 Figure 36: Investment converged to pre-crisis levels thanks to lower macroeconomic uncertainty .....................57 Figure 37: Labor moved from subsistent agriculture to basic services while the composition of GDP hardly changed since 1990 ..........................................................................................................................................................................59 Figure 38: Peru’s share of labor in high-, medium-, and low-skill occupations has not changed since 1995 .....59 Figure 39: Peru developed few new manufacturing export successes in its high growth period, 2000-10 ........61 Figure 40: The contribution of TFP to aggregate growth has been small ............................................................................64 Figure 41: Raising Peru’s productivity (TFP) to U.S. levels would bring Peru close to U.S. per capita income ..65 Figure 42: Misallocation in services point to severe restrictions in product and labor market competition .....66 Figure 43: Peru’s informal sector is very large by international standards .......................................................................68 Figure 44: Knowledge applicability in its embodied technologies .........................................................................................69 Figure 45: Equilibrium of weak labor demand for productive, well-paid jobs .................................................................72 Figure 46: Only one-third of wastewater is treated ......................................................................................................................75 Figure 47: Evolution of PM2.5 concentration in Lima, 2001–14 .............................................................................................77 Figure 48: Pollution cases by mining and extractives are rising .............................................................................................80 Figure 49: Illegal gold mining surpasses drug trafficking ...........................................................................................................81 Figure 50: Aerial view of illegal mining & coca fields in San Gaban, Puno ..........................................................................81 Figure 51: Peru’s tax revenues are low compared to its peers .................................................................................................84 Figure 52: Public spending varies substantially and has little impact on reducing interregional disparities ....87 Figure 53: Public Investment by local governments as a share of GDP ................................................................................88 vi Figure 54: Peru falls behind in perceived public spending efficiency ...................................................................................89 Figure 55: Asymmetries in decentralization of revenues and spending ..............................................................................91 Figure 56: Peru’s complex and fragmented system of health financing and spending responsibilities ................94 Figure 57: Higher control has detected more irregularities ......................................................................................................95 Figure 58: Peru’s enforcement of regulations and the rule of law ranks below the regional average ...................98 Figure 59: As incomes increase, households opt out of public services ...............................................................................99 Figure 60: Security, not unemployment, became the main concern ................................................................................... 100 Figure 61: Coca cultivated surface, 2000–14 ................................................................................................................................. 100 Figure 62: Coca cultivated surface by region, 2014 .................................................................................................................... 100 Figure 63: Crime control is not perceived to be effective, affordable, or accessible .................................................... 101 Figure 64: Confidence in judicial system is low ............................................................................................................................ 102 Figure 65: Judicial system is seen as ineffective ........................................................................................................................... 102 Figure 66: Peruvians have low trust in the state ......................................................................................................................... 103 Figure 67: Peruvians say the state is not transparent ............................................................................................................... 103 Figure 68: Trust in national Congress ............................................................................................................................................... 104 Figure 69: Trust in municipal government ..................................................................................................................................... 104 Figure 70: Few people think that tax evasion is never justified ............................................................................................ 104 Figure 71: Progress in one key outcome reinforces the other ............................................................................................... 106 Figure 72: Annual investment in water and sanitation by urban and rural .................................................................... 110 Figure 73: Households with no or precarious housing ............................................................................................................. 111 Figure 74: Households with property title, 2015 ........................................................................................................................ 111 Figure 75: Population distribution by socioeconomic group and distance to city center ......................................... 112 Figure 76: Climate vulnerability by region ..................................................................................................................................... 112 Figure 77: Road network in Peru, 1925-2011............................................................................................................................... 113 Figure 78: Road infrastructure spending, 1999-2011 ............................................................................................................... 113 Figure 79: Share of households with access to basic telecommunications services, 2004-2015 .......................... 113 Figure 80: Peru’s large gap in connecting infrastructure ......................................................................................................... 114 Figure 81: Peru is falling behind in accessibility, affordability, and duration of transportation ............................ 116 Figure 82: There is significant geographic disparity in the access to the nearest larger domestic market ...... 116 Figure 83: Shortcomings in public investment management undermine public investment efficiency ............. 117 Figure 84: Logistic costs are high especially for customs and infrastructure ................................................................. 119 Figure 85: Broadband internet penetration is low ..................................................................................................................... 120 Figure 86: Share of population with health insurance by area, 2004-2015 .................................................................... 122 Figure 87: Out Of Pocket Expenditure, 2000-2014 ..................................................................................................................... 122 Figure 88: Share of above-age students in Peru in secondary, 2015 .................................................................................. 124 Figure 89: Share of students that reached a satisfactory level in ECE, 2011-2015 ...................................................... 124 Figure 90: Teachers' income relative to other professionals', 2004-14 ............................................................................ 124 Figure 91: Budget allocation, 2012-2015 (% of GDP) ............................................................................................................... 126 Figure 92: Non-wage budget by purpose in basic education, 2012-2015 (% of GDP)................................................ 126 Figure 93: Social protection spending per capita, Peru vs. LAC ............................................................................................ 129 Figure 94: Labor rigidities, skills, and bureaucratic barriers to competition (licenses and permits) are among the top five constraints perceived by managers ............................................................................................................... 131 Figure 95: Non-wage labor costs across different regimes ..................................................................................................... 132 Figure 96: Regulatory constraints are the most important factor explaining the large difference in labor informality between Peru and Chile ...................................................................................................................................... 132 Figure 97: Government promotion of agriculture and export performance ................................................................... 134 vii Figure 98: Product market competition performs is stronger than in the average LAC country but significant regulatory barriers for licenses and permits and in service sectors remain ....................................................... 136 Figure 99: Costs of credit are high for SMEs and banking is highly concentrated and profitable ......................... 138 Figure 100: Many bureaucratic barriers to competition and entry are subnational ................................................... 139 Figure 101: Low public investments in innovation compared to peers ............................................................................ 140 Figure 102: Peru’s National Science, Technology and Innovation (STI) System has high coordination costs . 141 List of Tables Table 1: Priority areas and key bottlenecks to address the two structural challenges .................................................19 Table 2: There is little presence of national parties in local governments ..........................................................................96 Table 3: Priority areas and key bottlenecks according to prioritization criteria .......................................................... 108 Table 4: Regions with higher increase in road network (in km) .......................................................................................... 113 Table 5: Pension schemes by worker types ................................................................................................................................... 133 Table 6: Household poverty profile, 2015 ...................................................................................................................................... 157 Table 7: Household profile of indigenous and non-indigenous, 2015 ................................................................................ 158 Table 8: Household profile, by gender of household head, 2015 ......................................................................................... 159 List of Boxes Box 1: Peru's demographic and human capital changes..............................................................................................................10 Box 2: The effects of internal migration on demographics in Peru ........................................................................................17 Box 3: A coordinated approach to combat chronic malnutrition ............................................................................................39 Box 4: Sources of growth: a sectoral approach ................................................................................................................................70 Box 5: Mining, social conflict, and the Public Consultation Law ..............................................................................................82 Box 6: A decade of fiscal decentralization ..........................................................................................................................................90 Box 7: The rural electrification program experience ....................................................................................................................96 Box 8: Government promotion of agribusiness and sector performance ......................................................................... 134 Box 9: Knowledge gaps ............................................................................................................................................................................ 145 viii 1. Peru has been one of the most prominent performers in Latin America in the last 25 years. With GNI per capita of US$5,975 in 2015 (2011 PPP), its economy is one of the largest in Latin America and the Caribbean (LAC). Peru’s rapid economic growth, averaging 5.3 percent since 2001, was second only to Panama’s in LAC. Its population of about 31 million is relatively young, with more than half being under 30 years of age (Box 1). After a massive urbanization process over the last 60 years, Peru is today a mostly urban country, with about 80 percent of the population living in urban areas. Economic growth has been widely shared. The poverty incidence rate fell from 58 to 23 percent from 2004–14, and households’ incomes at the bottom 40 percent grew 50 percent faster than the national average. The fast and widely shared growth transformed Peru into an upper-middle income economy, with aspirations to become a high-income economy in the next 20 years. 2. Peru is characterized by a complex and diverse geography that holds wealth in natural resources and several spatial development challenges. Peru’s geography is incredibly diverse, with the massive Andean cordillera dividing its surface into three natural regions: the Costa (arid coastal plains, where about 55 percent of the population resides) to the west; the Sierra (highlands, with 32 percent of the population); and the Selva (the lush Amazon rainforest, with 13 percent of the population) to the east. The country’s latitude, its mountain ranges, sharp variations in topography, and the Humboldt (ocean) Current create dramatically different climatic zones. Its geographic diversity exposes Peru to natural hazards—it has seven of the nine possible characteristics that make a country vulnerable to natural disasters: earthquakes, flash flooding, landslides, and volcanic activity, among them. The country’s varied geography causes challenges for connectivity, raising the cost of service delivery. The vast Sierra and Selva regions are difficult to traverse, having historically isolated their remote communities from those living in the Costa region, thereby resulting in large development gaps among the country’s regions. 3. Its geographic diversity makes Peru a resource rich country. It has large reserves of ores, including copper, gold, silver, zinc, lead, iron, and tin. Copper accounts for about one-third of total exports and Peru has become the world’s second largest copper exporter after Chile. It also has oil and gas reserves, is the third largest producer of fish (the largest exporter of anchovy), and has become a leading exporter of fruits and vegetables. Commodity exports accounted for about 70 percent of total exports in the 2000s. 4. Peru has a remarkable cultural heritage and rich ethnic diversity. Home to the oldest civilization in the Americas (Caral; 3,000–1,800 BC) and to the largest empire in pre-Columbian America (the Inca empire, which spread over 2 million square kilometers), Peru was already a land of great cultural, economic, and scientific wealth well before the 16th century Spanish conquest. 9 The European, African, Arab, Chinese, and Japanese migration that gradually took place from the conquest through the 20th century formed a multi-ethnic society, with a unique blend of cultures and traditions. Peru’s indigenous population, also diverse, makes up about one-quarter of the country’s total population and, paradoxically, is disadvantaged in terms of poverty and access to services. This diverse history has made Peru one of the leading cultural heritage and gastronomic destinations in the world. 5. Peru’s geography, natural endowments, and diverse population have shaped its unbalanced economic development. The country’s plentiful natural and human endowments shaped its economic development based on capital-intensive growth that resulted in spatially unbalanced outcomes for the population. The cost of service delivery and connectivity in the vast Sierra and Selva regions is high, which has concentrated economic activity in the Costa region, especially in the area of Lima, the country’s capital. The Lima area now accounts for one-third of Peru’s population and one half of its GDP. As an example of its unbalanced development, the primary transport network connects Lima to other coastal cities, but not necessarily to medium and smaller cities in the Sierra or Selva, other than those important to the mining industry. The country’s abundant resources have attracted large foreign investments in mining and enabled growth based on fast capital accumulation, albeit with few gains in productivity and little export diversification. Mining activities are centered in few areas, which has disproportionately benefitted a small number of districts under current decentralization arrangements. 6. Geography and resource abundance have thus led to a spatial concentration of economic activities and opportunities, creating large disparities in development across the country’s territory and its population groups. Poverty is unevenly distributed—only 180 out of more than 1,800 district municipalities account for half of all poverty in the country, and districts with the highest poverty incidence are located mostly in the Sierra and Selva. Extreme poverty incidence in rural areas is 13 times higher than in urban centers. Furthermore, a historically low presence of the State in isolated regions fueled a lack of trust in the State that is still visible in the generally low compliance with rules and regulations, which in part explains Peru’s abnormally high economic informality. Peru is close to the apex of its demographic transition and its population is living longer. In 2015, out of the 31 million Peruvians, 65 percent were of working-age (15–64 years old), whereas 28 percent were below 15 years old, and only 7 percent were 65 and above. This is the result of a progressive decline in the total fertility rate, from 3.5 children per woman in 1995 to 2.5 in 2015, which has reduced the share of children in the population, and significantly increased the share of working-age adults. Meanwhile, the dependency ratio continues to decline, and it is expected to reach 0.49 in 2016 (that is, there are two working-age adults per dependent member of the household). As with many Asian countries in the 1990s, this demographic trend is a unique opportunity in growth potential, considering that Peru has high labor force participation rates (at 78 percent, it is the third in LAC, after Barbados and Bahamas); 7 percentage points higher than the LAC average, and nearly 10 percentage points higher 10 than the world average. Moreover, Peru has steadily improved its human development outcomes, implying an overall increase in human capital. For instance, life expectancy increased by 8.8 years in the last 25 years (to 75 years), infant and maternal mortality rates fell by more than 70 percent —the largest drop in the region—and stunting rates in children under five years old dropped by half in 10 years (from 28 percent in 2005 to 14 percent in 2015). 7. During the 1980s, Peru faced one of its worst economic crises and the start of an internal armed conflict. The Latin American debt crisis in the 1980s, combined with an inherited state-led growth model, drove the country into a cycle of GDP contraction and hyperinflation. By the end of the decade, inflation rates had hit four digits (reaching 7,500 percent) and output had fallen by 25 percent. GDP per capita, relative to that of the United States, fell from 25 percent in the 1970s to 15 percent in 1990. Moreover, in the early 1980s, the Communist Party of Peru— Shining Path (PCPSL) launched an armed struggle against the Peruvian state that was soon followed by the Tupac Amaru Revolutionary Movement (MRTA). This armed rebellion—similar in spirit, but much bloodier than other communist movements in the region—claimed to empower the peasantry against the State and the ruling classes. The armed conflict left behind close to 70,000 civilian casualties, most of whom were poor, rural Sierra peasants, mainly from Quechua and other indigenous communities.1 Victims also included local authorities and community leaders. The nature of the PCPSL terror and military tactics, and the lack of a proper response by the Peruvian State subsequently led to thousands of human rights abuse cases on both sides of the conflict, a legacy of violence with which the country is still struggling to come to terms. The human rights abuses and the economic turmoil of that period further eroded social trust. 8. During the 1990s, Peru undertook several deep macroeconomic reforms supporting monetary and fiscal policy discipline, but also reverted upon some institutional progress. These reforms included trade and financial liberalization, the privatization of state-owned enterprises, and implementation of a more flexible exchange rate regime. To ensure a credible monetary policy and a sustainable fiscal policy, greater autonomy was given to the Central Bank of Peru (Banco Central de Reserva del Perú; BCRP), and the National Customs and Tax Administration Superintendence (Superintendencia Nacional de Aduanas y de Administración Tributaria; SUNAT) was created. The reforms laid the basis for the subsequent output recovery (Figure 1). The volatility of output growth declined somewhatin the 1990s but investment uncertainty remained unusually high as high-level corruption scandals and an increasing abuse of executive power undermined the rule of law. During this period previous institutional progress, for instance, in decentralization, transparency, and accountability mechanisms to control executive was partially reverted. 1 The Truth and Reconciliation Commission Report points out that out of the estimated 69,280 victims, 79 percent lived in rural areas, 56 percent were farmers, and 75 percent were had an indigenous mother tongue. In addition, 85 percent of them lived in the departments of Ayacucho (40 percent), Apurimac, Huánuco, Huancavelica, Junín, and San Martin. See TRC (2003). 11 40% Commodity boom, Oil crisis Debt crisis Initial macro-trade reforms Structural reforms GDP per capita relative to the United States 35% 30% 25% 20% 15% 10% 5% 0% 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2014 Chile Malaysia Thailand Peru Source: Authors; based on World Bank data. Favorable exogenous conditions, macro-structural reforms, and equitable growth 9. Over the last 15 years, Peru experienced a period of remarkable growth led by factor accumulation, resulting in an output recovery. As the economy stabilized and began to recover from the crisis of the 1980s, high-return investment opportunities increased, and so did capital accumulation, which has accounted for more than two-thirds of aggregate growth since 2001. During this period, average growth reached 5.3 percent, among the highest in the LAC region. 10. A strong growth in internal demand generated a surge in service jobs, often in the urban informal sector. During this period, labor moved from farming to mostly unskilled service jobs, especially in urban areas. Indeed, the employment share of agriculture declined from 32 percent in 2001 to 24 percent in 2013, while the employment share of services such as retail trade, hotels and restaurants, and transport—all of which have high informality rates—expanded. At the same time, agricultural productivity and incomes increased, due to better connectivity in the Sierra and increased industrialization of the agro-food exports sector in the Costa. Higher agricultural productivity and the expansion of urban informal service jobs led to a surge in the labor income of the bottom 40 percent, which further fueled growth by raising domestic consumption. 11. Economic growth was inclusive and reduced poverty significantly. Poverty responded strongly to growth—for each percentage point increase in GDP growth, poverty fell by 1.4 percentage points. Thus, from 2004–15, 9.3 million Peruvians escaped poverty, moderate poverty fell by more than half, from 58 to 22 percent and extreme poverty fell from 16 to 4 percent. More broadly, households’ incomes in the bottom 40 percent increased by an average of 6.8 percent per year, against 4.4 percent for the average income. As a result, inequality declined by 12.6 percent over the period. The middle class grew by 18 percentage points from 2004–14, and outnumbered 12 the poor by 2013. Higher labor incomes explain the lion’s share of the reductions in poverty and inequality. However, non-labor incomes (mostly public transfers from social programs) have helped to reduce rural poverty, especially over the last five years: the non-labor income contribution to rural poverty reduction had increased to 20 percent in 2014. 12. The virtuous cycle of growth and shared prosperity can be explained in large part by a combination of two main forces: favorable exogenous conditions and successful macro- structural reforms (Figure 2). The commodity price boom that started in the early 2000s boosted investment in mining, which had important ramifications for Peru’s economy. Foreign direct investment (FDI) increased four-fold as a share of GDP. Mining investment, which peaked at US$10 billion in 2013, accounted for 20 percent of total private investment from 2001-13. The mining boom also spurred private investments in upstream sectors such as chemicals, metal products, electricity and gas, land transport, and financial services. And in mining regions, higher incomes boosted consumption, especially of services 13. A number of important macroeconomic reforms supported the acceleration in capital accumulation. The Government developed a sound macro-fiscal framework for managing public resources. It adopted an inflation-targeting regime, eliminated interest rate controls, and strengthened banking supervision and regulation. It also reduced tariffs and removed other barriers to trade and foreign investments. These structural reforms led to low inflation expectations and together with the adoption of a countercyclical fiscal policy, allowed the Government to accumulate savings from the windfall that came from the commodity boom. In contrast to other Latin American countries, Peru raised its savings during the boom years by almost 2 percent annually from 2000–13, which helped finance the higher investment rates. Over that period, the 13 public sector ran average fiscal surpluses contributing to the growth in aggregate savings. International reserves increased from 17 to 32 percent of GDP from 2000–15, twice the size of the share of international reserves in Chile, Colombia, or Mexico, while public debt remained low (at about 20 percent of GDP in recent years, with net public debt at only 3 percent of GDP in 2015). 14. Macroeconomic stability and higher fiscal revenues enabled public investment and social progress. The stable macroeconomic environment reduced investment uncertainty, which, concomitantly with higher revenues from the mining boom, enabled public investment in connecting infrastructure, education, health, and social programs, diversifying rural economies and supporting urbanization. From 2004-15, mobile phone use increased from 1 to 73 percent in rural areas, and from 23 to 90 percent in urban areas. The total kilometers of paved roads more than doubled in several regional departments, such as Huancavelica (243 percent) and Ayacucho (189 percent). Higher spending for education materials and an increase in the proportion of certified teachers helped increase net enrollment for pre-primary education (3–5 years) from 54 to 83 percent from 2001–15 and in secondary education from 69 to 83 percent. The expansion of public health insurance reduced the socioeconomic gap in enrollment, especially in rural areas. And, targeted, multi-sectoral social programs such as the Articulated Nutritional Program (Programa Articulado Nutricional, PAN) yielded great progress in reducing child stunting and improving other dimensions of human development. 15. Past sources of shared prosperity are drying out. A weak external environment, with declining commodity prices, together with the global economic slowdown have led to a deceleration of private investment. The prices of copper and gold, which together account for 35 percent of Peru’s total exports, declined by almost one-half and one-third, respectively, since the beginning of 2013. Lower commodity prices reduced the value of Peru’s exports—commodities accounted for about 70 percent of total exports in 2014. The deterioration in terms of trade has also had a significant impact on private investment, which declined by 6.1 percent in 2014, and then stabilizing somewhat in 2015. Further, the higher volatility of global financial markets will likely increase firms’ financing costs, especially for the large pool of dollar-denominated corporate bonds, making only those projects with high returns attractive. Moreover, Peru’s long recovery from the hardships of the 1980s and 1990s is ending, implying that the large pool of profitable investment projects is drying out. The investment share has converged to its pre-crisis plateau where it is expected stagnate in the absence of productivity improvements that raise the returns to investment. Finally, the wage differential across sectors has declined, removing an important channel for the poor to raise their labor income by migrating from rural agriculture to urban services. 14 16. The new headwinds indicate that the past virtuous cycle of growth and shared prosperity may have reached its limit. The dividends for shared prosperity from employment growth in unskilled, informal services are shrinking. The share of capital investment in GDP peaked at a historically high level on the back of high commodity prices and cannot be sustained without raising private sector productivity and, consequently, without increasing returns to capital in non-mining activities. Despite lower commodity prices, Peru’s mining exports and investments did not collapse as a result of the sectors’ high competitiveness, and the mining sector is projected to continue to support moderate GDP growth rates of about 3 percent per annum over the medium term. But it will not create sufficient new, better-paying jobs to sustain the labor income growth needed to reduce poverty further and raise the incomes of the middle class. 17. Still, labor income will have to continue to be the main mechanism for reducing poverty and raising the income of the bottom 40 percent. The scope of public transfers to reduce poverty is limited and labor income will continue to be the main engine of shared prosperity. But many Peruvians still do not have the skills and the access to better-paying income opportunities. As urban migration slows, Peru needs to go beyond the static gains from the shift from rural agriculture to informal urban services and begin to generate dynamic productivity gains within the urban economy. However, labor-intensive activities such as tourism, commerce, and transport still have low productivity levels, limiting the opportunities to create higher-paying jobs. 18. These new headwinds highlight two structural challenges that have emerged from Peru’s specific endowments, and that constrain the opportunities for income growth of the bottom 40 percent. Peru’s geography, natural endowments, and diverse population supported shared prosperity over the past 15 years, relying on previous spatially unbalanced economic opportunities, combined with fast recovery of capital accumulation and the expansion of low productivity informal services providing income opportunities for the bottom 40 percent. To unleash a new virtuous cycle of shared prosperity Peru needs not only to continue to utilize its endowments, it also has to address two main structural challenges emerging from these endowments to provide all Peruvians with access to better-paying income opportunities and to unleash the private sector’s potential to create those better-paying jobs. These structural challenges, described in detail below, are the country’s persistent spatial disparities in development and its low productivity. 19. First, the persistence of large spatial disparities in development consistently undermine the ability of certain population groups, particularly indigenous and Afro- Peruvians, from overcoming poverty. Despite the impressive progress in poverty reduction over the last decade, at least 40 percent of the population is still vulnerable to falling into poverty, and there remain large gaps in income and human development across the socioeconomic spectrum. Many segments of the population are still largely underserved. The rural population is almost a quarter of the total, but it accounts for half of the poor and 82 percent of the extreme poor. Fifty 15 four percent of Peru’s poor reside in only 181 of its more than 1,800 districts. Human development is lower in rural areas. For instance, child mortality and child malnutrition rates are about twice as high in rural areas, and education outcomes are worse than in urban areas. This situation hurts particular groups disproportionately—indigenous and Afro-Peruvians are still behind in human development and access to services relative to other Peruvians. But even though poverty is higher among the indigenous population, much of this difference is driven by the fact that they live in rural areas. 20. Moreover, the capital-centric development model contributes to imbalances within the urban sector. Today, Peru is one of the most capital-centric countries in the world—Lima accounts for 32 percent of the population (40 percent of the urban population) and 45 percent of the national GDP.2 The unplanned urbanization of Lima has increased disparities between the capital city and the rest of the country, and between Lima’s affluent neighborhoods and its informal urban dwellings. Disparities in access to basic services such as water and sanitation, housing, transport, and security reduce the scope of opportunities that people can access throughout their lifetime, affecting their prospects for a better livelihood for themselves and their children, and ultimately, hurt the growth and shared prosperity potential of the country as a whole. 21. Addressing Peru’s spatial disparities in a manner that would balance access to opportunities for all of its citizens will require paying close attention to the underlying features of poverty and vulnerability. The specificities of Peru—in particular geographic segmentation and high urbanization—pose important challenges to ensuring that growth continues to be inclusive and that it consolidates the process of social mobility, not just between poverty and vulnerability, but towards a solid middle class. These challenges include:  Chronic rural poverty. Despite the impressive reduction in poverty and extreme poverty, in half of Peru’s regions, 40 to 90 percent of districts still have a poverty incidence above 50 percent. In other words, there are a large number of districts—even if these are small— that have not benefited from income growth either from better labor market opportunities or from direct social interventions. In terms of individuals, there is evidence that most of the poor in Peru today are chronically poor: they were already poor 10 years ago.3  Low service quality. In general, access to services has improved tremendously in the last decade. But quality gaps are still very large. This affects not only the poor in rural areas, but also the large numbers of vulnerable individuals across the country (42 percent of the population in 2015). In many aspects, quality is a significant factor in lifting people out of poverty and vulnerability into the middle class. 2 The concentration of some economic activities can be inefficient. In Colombia or Mexico, for instance, manufacturing exporters cluster in less congested, well connected smaller cities. 3 Vakis, Rigolini and Lucchetti (2015). 16  Large urban populations. Recent demographic changes pose new challenges for the Government’s inclusion agenda. Urban areas have grown tremendously, reaching around 75 percent of the total population. Lima has grown 10-fold since the 1950s (Box 2). This urban growth puts pressure on the capacity of municipalities to provide good services for their residents that respond to their needs. Urban areas are characterized by having limited urban transport systems, unstable labor markets with high levels of informality, poor planning, and deficient land management, among others.4 Social services to the poor and vulnerable in these areas are generally non-existent. A process of urbanization driven by rural-urban migration has changed the demographic landscape of Peru. The share of the population living in urban areas increased rapidly from 47 percent in 1960 to about 76 percent in 2015, one of the highest in the region and typically higher than in peer countries. The rapid urbanization has not been limited to the growth of Lima—the share of the urban population living in the capital city remained almost constant at about 40 percent between 1960 and 2015. Still, given its size —it accounted for one-third of the total population in 2015—the capital city absorbed the largest number of migrants in absolute terms. Since 1950s, rural population from the Sierra began a massive migration to the Coast, especially to Lima, in search of better opportunities. During the next 44 years, Lima grew 10-fold, and rural population went from 65 percent of the total to less than 35 percent. In the early 1980s, close to half of the Limeños were migrants, and 80 percent of Lima’s population lived in newly established barriadas (shantytowns). A second wave of migrants came to Lima and other urban centers in the 1980s and early 1990s, pushed by the armed conflict in the Sierra. From 1988-93, 8.4 percent of the total population migrated internally. Migration continued to be strong in the 2000s as the urban population has grown steadily of about 2 percent annually. Undoubtedly, migration improved the living standards of millions who lived in extreme poverty in the Sierra with little or no State presence. Migration also changed urban economies, as migrants relied heavily on informal employment in services to generate income. Rural-urban migration is likely to slow down as Peru’s share of urban population is higher than in most Latin American countries and already accounts for close to 80 percent of the population. Source: Matos Mar (1984); INEI (2009); Yamada (2009). 22. Peru’s second structural challenge relates to the large productivity gap of its private sector relative to its peers, which is constraining the demand for better-paying jobs and income opportunities. Aggregate productivity (TFP) is low and has stagnated over the past 20 years—it contributed only 11 percent to economic growth from 2000-14, much lower than in peer countries such as Malaysia (23 percent) and Thailand (29 percent). Raising aggregate TFP in Peru has the highest growth payoff―the country would increase its relative GDP per worker from 25 to 65 percent of that of the United States if it had the same aggregate productivity5—while the potential gains from increasing capital or labor are relatively modest. Peru’s growth dividend from closing the large TFP gap with high-income countries is significantly larger than in other Latin American countries. 4 World Bank (2016ª). 5 Caselli (2015) and World Bank (2015a). Peru’s GDP per worker is more on par with its peer countries due to the high capital intensity and relatively large size of the mining sector. 17 23. Peru’s low aggregate productivity stems in part from substantial misallocation of capital and labor as its more productive firms do not necessarily hire more workers or invest more. The large number of informal firms likely contributes to this misallocation, limiting the size and growth of more productive (informal) firms, especially in service sectors.6 The lack of technology adoption and integration into global value chains further limit productivity growth. Low productivity, in turn, constrains firms’ export diversification, their ability to adopt new technologies, and their demand for better-paying jobs, likely contributing to Peru’s high labor informality. Low productivity, high informality, and little export diversification likely reinforce each other, describing an equilibrium of weak labor demand for more productive, better-paying jobs. Addressing the structural challenge of low productivity is thus imperative in order to sustain high returns to investment and continue the path of shared prosperity. 24. While past policies have been successful and sustaining them continues to be necessary, they are no longer sufficient to keep the country on the path towards high, inclusive, and sustainable growth. As highlighted in Figure 3, addressing Peru’s two key structural challenges described above—namely reducing its persistent spatial disparities in development and removing the barriers to productivity growth—would unleash a new virtuous cycle of shared prosperity. Overcoming these challenges requires a new generation of micro- structural reforms and reforms to address key long-term sustainability risks. If these risks are not addressed, the impact of social and public services reforms might not be sustained in the long term. 6 See a detailed analysis in Section 3. 18 25. The Systemic Country Diagnostic (SCD) prioritizes policy constraints that have the greatest impact on Peru’s structural challenges of reducing the large spatial disparities and boosting private sector productivity. The SCD uses the following selection criteria to identify the constraints with the largest impact on achieving shared prosperity going forward. First, it identifies constraints that significantly affect one or both of the two main structural challenges. Second, it identifies policy constraints that present synergies to overcome these structural challenges. Third, it identifies constraints that support the sustainability of addressing Peru’s structural challenges. Priority Areas Key Bottlenecks Prioritization criteria sustainability of outcomes Increases productivity P resent synergies to Reduces disparities achieve outcomes Support the Improving connecting Large gaps in access to water and sanitation     infrastructure and public Lack of urban planning and cadasters     services Large gaps in connecting infrastructure    Fragmented coverage and quality of services in health     Raising human capital Low quality of services in education     Inefficiencies across social services     Labor and tax regulations that hamper productivity and    Reducing factor and product formalization market rigidities Regulatory barriers to competition (at the subnational   level) that hamper productivity and formalization Low efficiency of public spending   Right-sizing the fiscal Low tax revenues   envelope Unfinished decentralization process   Weak enforcement of the rule of law in the justice   system Improving government Weak enforcement of regulations   coordination, capacity, and Low level of government coordination and red tape     law enforcement Low levels of trust and social capital (citizen-   government) High exposure to climatic risks and natural hazards     Reducing environmental Reactive and loosely coordinated natural resources risks   management 26. Applying the three criteria described above, the SCD identifies a set of constraints that are pivotal to address Peru’s two main structural challenges and should thus be the focus of policies in coming years. Macroeconomic stability and openness to trade and investment will remain pre-conditions for prosperity. But a new generation of micro-structural priority areas needs to be tackled, including (i) improving connecting infrastructure and public services; (ii) raising human capital; and (iii) reducing factor and product market rigidities. At the same time, 19 sustainability risks undermine the long-term impact of the micro-structural reforms. These include (i) an inadequate fiscal envelope; (ii) low capacity and inefficient government coordination and law enforcement; and (iii) environmental risks. Without policies that address these risks, any progress from sector-specific, micro-structural reforms on reducing spatial imbalances will be reversed. The list of critical constraints and the specific bottlenecks associated with them are summarized in Table 1. The need for micro-structural reforms 27. Micro-structural reforms are needed to raise the endowments of and incentives to people, firms, and communities to increase their productivity. The constraints that have the greatest potential effect on Peru’s structural challenges are discussed in more detail below.7 Improving connecting infrastructure and public services 28. Improving the provision of public services, especially in rural and marginal urban areas, will have a large impact on reducing spatial disparities in Peru. Throughout the country, there are important inefficiencies in the quality and reliability of public service provision. For example, in spite of high levels of public investment in water and sanitation, only 15 percent of rural households receive drinking water, compared to 90 percent of urban households. And access to water can be inequitable, as it costs up to 65 Soles (S/.) per 5 cubic meters in informal urban settlements, compared to only S/.12 in residential areas. Water provision enterprises (EPS) incur significant water losses of about 40 percent, much higher than in developed countries (estimated at 15–20 percent), and also higher than the level attained in the best-performing utilities in developing countries (estimated at 20–25 percent). Service continuity also is a problem, and there is increasing water pollution from untreated wastewater, dumping of industrial solid waste, and uncontrolled use of agrochemicals. In 2012, only about a third of wastewater was treated before being sent back into the environment. The situation in water and sanitation mirrors the quality of service provision in many other sectors, including electricity, waste collection, and public transportation. Typically, rural households have much lower access to good quality services. But even in urban areas, unplanned growth has led to large inequities within cities, as many low- income neighborhoods have inadequate waste collection, security, public spaces, and so forth. A considerable share of the urban population lacks adequate housing, which increases their exposure to earthquakes and landslides. And, despite the importance of access to land for investment, planning, and services, only 8 of more than 1,800 municipalities hold comprehensive and up-to- date cadasters. 7 The analysis for the SCD was conducted in 2016 and does not include the legislation passed by the new administration in the last 2 months of 2016 when it was allowed to pass new laws in several defined areas without the approval of congress. For instance, the government passed a law in December 2016 to further strengthen the power of the competition authority (INDECOPI) to make legally binding recommendations to align regulations with the competition law. 20 29. Investing in connecting infrastructure will help improve the competitiveness of Peruvian firms. Markets towards the interior, south, and north of the country have low accessibility. The lack of connectivity divides domestic markets, undermining efficiency gains from competition and economies of scale. Poor connectivity also reduces the export competitiveness of domestic firms and diminishes incentives for FDI, thus limiting access to foreign markets and technologies. In contrast, markets are better connected in Colombia and Ecuador, which also have a complex geography. Peru’s logistics costs—about 32 percent of product value—are among the highest in Latin America, while many larger Peruvian cities face transport cost markups for commercial and social exchanges of 50 percent or higher relative to Lima.8 The export competitiveness of Peruvian firms is further undermined by the high bureaucratic costs of Peru’s customs administration, in which Peru ranks below its peers.9 Raising human capital 30. Improving human capital will require efforts to increase access to quality health services and to scale up successful social assistance programs. While Peru has expanded health care coverage for the poor through the Comprehensive Health Insurance system (Seguro Integral de Salud, SIS), there are still severe disparities in the coverage and quality of health services. Enrollment in health insurance reached 69 percent of the population in 2014, but while SIS covers the rural poor and formal sector workers are covered by the contributory social security system (EsSalud), there is still a “missing middle” of non-poor informal workers who lack health coverage. Moreover, poorer regions tend to have fewer doctors per capita, and out-of-pocket health expenditures continue to be high, constraining the access to quality health service for the poor and vulnerable. Social assistance programs are often well targeted but have low coverage. The Juntos conditional cash transfer (CCT) program, for instance, is limited only to districts with more than 40 percent of poverty incidence; it is thus effective in reducing poverty in rural areas, but its aggregate poverty impact is limited. 31. Improving the quality of education and especially reducing disparities in quality is required to ensure that all Peruvians develop to their potential. From the start of their lives, all Peruvians should have access to the structures and assets that they need to learn and develop their skills and job readiness, to exploit their potential as productive adults, and finally, to have income protection in their later years when they are no longer able to sustain themselves. Yet, many Peruvians, especially among the poor and in rural areas, lack access to the necessary structures and support systems, and therefore cannot expect to achieve good and secure income opportunities over their lifetime. The low average human capital, for example reflected in Peru’s Programme for International Student Assessment (PISA) scores (improving, but still among the 8 Well above Colombia’s (23%), Chile’s (18%), Brazil (26%), and Argentina (27%). 9 Logistic performance index, 2015. 21 lowest worldwide), masks notable gaps between different groups. For example, net enrollment in secondary school for extremely poor students is 20 percentage points lower than for non-poor students. Likewise, the proportion of above-age students at the primary and secondary levels is higher among rural and indigenous population, and the performance of students in national and international tests is significantly lower for rural students. This is due in part to poorer teaching conditions in rural and remote areas (including infrastructure, teachers, and other inputs), but also to the lower nutritional levels among children in disadvantaged areas, which is an impediment to cognitive development. In tertiary education, there is a large heterogeneity in the quality of training and higher education institutions, which results in large mismatches in the labor market. Reducing factor and product market rigidities 32. Peru needs to reduce the rigidities of its labor market to encourage the allocation of human capital to more productive activities. Only three out of 15 Latin American countries have more rigid regimes for hiring and firing employees. For instance, dismissals for economic reasons are severely limited and require explicit authorization from the Ministry of Labor (Ministerio del Trabajo y Promoción del Empleo; MTPE). Nonwage formal sector labor costs under the general labor regime account for 68 percent of the basic wage, by far the highest in the region. These labor market rigidities undermine private sector competitiveness and also, potentially, formal job creation. While past initiatives to reduce labor informality have failed, a successful formalization process probably requires a coordinated approach that combines reducing the costs of formality (hiring and firing, labor costs and contributions, tax burden, and other regulations) while increasing the benefits of formality (public services, more consistent enforcement). 33. Improving the productivity of Peru’s firms will also require actions aimed at reducing regulatory barriers, especially at the subnational level, that constrain market entry and competition. Peru has removed most tariff and many non-tariff technical trade barriers, introduced a best-practice legal framework for competition, and established an independent competition authority. But according to the WBG-OECD product market regulation indicators, the complexity of regulatory procedures and protection of incumbents stifle competition, especially in backbone service sectors such as transport, telecom, retail, and professional services. Unlike in Chile, Mexico, and Colombia, professional service firms in Peru self-regulate the entry conditions for new providers. In some cases this leads to anticompetitive practices. According to the national competition authority, subnational government bodies imposed all of 76 percent of the bureaucratic barriers (licensing, permits, and inspections). Further, the Doing Business data show that entrepreneurs in Peru, unlike in peer countries, spend almost all of the time required to open a business (26 days versus 8.3 days in OECD) dealing with municipalities (15 days) and notaries (9.5 days). 22 Supporting and sustainability conditions 34. Failure to address important sustainability risks would undermine potential achievements from micro-structural reforms. As discussed earlier, shared prosperity in Peru relies on raising the quality human capital, improving the connectivity infrastructure and public service provision, and reducing factor and product market rigidities. But to make the impact of these reforms sustainable over time, Peru needs to address important sustainability risks. The country needs to strengthen environmental risk management, right size the fiscal envelope to afford investments and services with sustainability, and implement institutional reforms improving government coordination, capacity, and law enforcement. Right-sizing the fiscal envelope to afford investments and services with sustainability 35. The low efficiency of public spending undermines the quality of infrastructure and public services. Inefficiencies in public spending limit the quality of public and social services.10 The low efficiency of public infrastructure spending, for instance, is related to shortcomings in public investment management such as the absence of multiyear budgeting or insufficient national or sectoral planning. Peru does not publish projections of capital spending beyond the current budget year and does not have multiyear targets or ceilings on capital expenditure by ministry or program. Also, there is no official record regarding commitments in future years from signed public investment contracts. The high administrative costs associated with public investment management are to some extent the consequence of the decentralized system with small, potentially weak capacity municipalities accounting for a large share of public investments. As a response to the potential low local capacity, the central government put in place additional procedures and control mechanisms to enforce quality standards. For instance, it created 110 investment committees in all three levels of government to ensure prioritization of strategic projects. 36. Low tax revenues constrain the quality of public services. Peru’s government has kept fiscal spending at sustainable levels, but this has also constrained investments in public services and the Government’s ability to reduce the large spatial gaps. Peru has a relatively small government compared to other upper-middle income countries. For instance, Peru collects a lower share of GDP in taxes.11 On average, Peru spends only 9 percent on all social expenditures (including pensions and social assistance programs), less than in most peer countries. An important challenge ahead is to increase the efficiency of spending without sacrificing quality and to expand the revenue base to finance higher investments in sectors where spending is relatively low. The 10 An ongoing World Bank Public Expenditure Review for Peru analyzes the efficiency of public spending in Peru and benchmarks the size of the state in different sectors relative to peer countries. 11 According to OECD (2015), the tax revenue in Peru in 2013 was 18.3 percent of GDP, compared to 21.3 percent for LAC on average and 34.1 percent for the OECD. 23 SCD finds, for instance, that spending in social assistance programs is low despite the existence of well-targeted programs. Reducing environmental risks 37. Regulating the use of natural resources and better enforcing environmental laws will mitigate the risks facing crucial resources like air, water, and fish stocks. Increasing urbanization and incomes have raised Peru’s levels of congestion and air pollution. Further, the pollution of water and land—often related to illegal mining activities—negatively affects people’s health and thus the country’s human capital endowment. Key economic sectors (agriculture, extractives, and tourism) depend on natural resources; managing them sustainably is a precondition to increasing their growth potential. Moreover, Peru is highly exposed to climatic risks and natural hazards. In the last 35 years, Peruvian glacier surface area fell by 22 percent. It is estimated that by 2020 all glacier surfaces below 5,000 meters will disappear, endangering the sustainability of water resources and the production of hydroelectricity. While more stringent environmental regulations have been implemented in selected cases, the lack of proper regulations and enforcement appears to be the norm. For instance, while the Government has put into place an effective quota system for anchovies—a major export crop—similar regulatory frameworks for other species are absent. And, while the 2012 Public Consultation Law, enacted after the sharp increase of mining related conflicts since 2007, improved the legal framework through which to channel local community concerns about extractive industries, specific sector roles and monitoring obligations are still unclear. Further, there is limited funding for consultation processes, limiting the interpretation and implementation of the law. Improving government coordination, capacity, and law enforcement 38. Current institutional arrangements are not well suited to provide the level and quality of services demanded by Peru’s population while coping with the economy’s increasing complexity. Clusters of excellence in government coexist with other government bodies that lack capacity and appropriate resources. More importantly, there is often a substantial overlap in responsibilities across government tiers, while coordination mechanisms are typically absent. Further, the unfinished decentralization process has generated additional inefficiencies in public investment and service delivery. The transfer of public investment responsibilities to the smallest government tier discourages investments in larger, cross-jurisdiction infrastructure projects with higher social returns, such as roads connecting cities, and has led instead to a large number of small-scale public investment projects. Further, mining revenues are primarily distributed to the few municipalities hosting the mines—only four of the 1,842 municipalities receive more than 50 percent of the total canon transfers to local governments. This creates abundant resources in some municipalities and severe shortages in others, thus undermining the efficiency in public spending and raising spatial inequalities in public service delivery. The decentralization of administrative 24 functions to subnational governments, without a decision-making authority and in the absence of relevant own revenue sources has led to misaligned incentives and the reduced efficiency of service delivery. 39. Improving law enforcement and access to justice for all citizens would also increase public trust and enhance levels of compliance. The lack of compliance with the legal framework has many facets, including building permits, municipal licenses, taxes, and labor standards. The belief of many Peruvians that regulatory evasion is justified, combined with the Government’s low enforcement capacity, generates risks and debilitates the social contract which is reflected in the increasing number of social conflicts throughout the country (207 as of September 2016, Ombudsman Office). Citizens’ perception of transparency and efficiency of the state are among the most pessimistic in LAC. The lack of transparency and accountability mechanisms, especially at the local level, together with the abundance of corruption investigations has led to a generalized mistrust in public institutions and the justice system. Only 21 percent of all citizens trust the judicial system, and most consider it inequitable. The limited access to judiciary services also undermines the ability of large parts of the population, especially the poor and vulnerable, to protect their rights. A case in point is the pervasiveness of domestic abuse, which most often goes unreported and rarely ends in prosecution. An inefficient justice system also increases firms’ investment uncertainty and reduces the appropriability of returns from more risky and productive activities, such as innovation. Finally, it increases the exposure to crime and the private costs of security. 40. The list of priority areas for action reflects the fact that the quality of government policies and services has fallen behind the pace of economic growth and the aspirations of the country’s population. Macroeconomic stability and favorable exogenous conditions helped Peru to achieve fast economic growth in the past 15 years. But public services and institutions did not improve at a similar pace. Despite rapid growth in recent years, the quality of social, infrastructure, and regulatory services still needs to improve for Peru to become an inclusive society with a solid middle-class. Sectoral reforms will not yield the desired results in Peru unless institutional constraints to improve government efficiency are eliminated, including improving the Government’s ability to plan, coordinate, and implement policies, and to guarantee law enforcement and access to justice. 41. The SCD seeks to identify the critical constraints and priority interventions necessary for Peru to achieve the goals of reducing poverty and improving shared prosperity. The study steps back from the World Bank’s existing portfolio, and even the Peruvian Government’s strategy, to conduct a broad overview of socioeconomic development progress in Peru. It is 25 designed as an analytical input for stakeholders to debate development priorities, and also to inform the preparation of a new Country Partnership Framework (CPF). 42. In preparation of this SCD, extensive consultations were held with various stakeholders. Three meetings with the Country Team were held on January 20, June 8 and October 13, 2016. A first specific SCD mission took place February 1–5, 2016. A second consultation mission that met with private sector representatives took place in July 4-8, 2016. Moreover, many bilateral meetings with Global Practice (GP) teams have been held to obtain feedback from colleagues and build consensus on priorities. The knowledge base on Peru is substantial and the analysis draws heavily on World Bank studies recently prepared across various GPs, new specific analysis prepared by the SCD team, and external work. The challenge of an SCD is to prioritize the binding constraints in order to tackle the most pressing development challenges. 43. The guiding question of this SCD is: What is necessary for Peru to continue on the path of high growth, poverty reduction, and shared prosperity? The SCD is organized around the following set of questions: • What are the critical factors determining poverty and inclusion? (Section 2) • What are the critical factors determining aggregate growth? (Section 3) • How sustainable (environmentally, socially, and fiscally) is the current pattern of growth, distribution, and poverty reduction? (Section 4) Section 2 reviews the recent trends and determinants of poverty, inequality, and inclusion—the analysis identifies Peru’s large spatial disparities as the main structural challenge to sustain poverty reduction and inclusion. Section 3 analyzes the trends and determinants of aggregate growth—the analysis identifies Peru’s low private sector productivity as the main structural challenge to sustain high growth. Section 4 focuses on the environmental, fiscal, and socio-institutional risk factors that will undermine the impact of microstructural reforms in reducing spatial disparities and boosting productivity, if not addressed. Finally, Section 5 identifies crucial policy areas that must be addressed to overcome the binding constraints to shared prosperity identified in Sections 2, 3, and 4. The SCD concludes with a set of priority areas and key bottlenecks for Peru that emerge from the analysis. 26 Over the last decade, poverty and inequality dropped dramatically in Peru, thanks to economic growth and increasing labor incomes, especially at the bottom of the income distribution. Human development and access to services also improved. Rural areas, however, still have high poverty rates, and their populations have lower access to services and lower human capital than their urban counterparts. Women face disparities in the labor market, and they are highly exposed to violence, in particular in the domestic realm. And, Indigenous and Afro-Peruvians face further disadvantages in living conditions. However, location plays a much larger role than ethnicity in explaining poverty. Peru’s large spatial disparities in development are thus a key impediment hindering many Peruvians from acquiring the human capital necessary to access better job opportunities. Sustained growth caused poverty and extreme poverty to fall sharply in the last ten years Over the last ten years, moderate poverty fell by more than half, and extreme poverty by more than two-thirds. As Section 1 presents, the strong and sustained economic growth of the last decade was broad-based. Households in the bottom 25 percent experienced income growth rates of about 7 percent per year between 2004 and 2014, while households in the top 25 percent saw increases of only 5 percent or less per year. The share of the population living in poverty fell from 58.7 percent in 2004 to 21.8 in 2015, a 63 percent fall in the poverty rate. Similarly, extreme poverty fell by three-fourths, reaching 4.1 percent in 2015. In absolute numbers, these changes signify that 9.3 million people exited poverty, and 3.2 million people exited extreme poverty over this ten-year period. 50 Poverty headcount (%) 2005 USD 40 30 PPP 20 10 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 $2.50 LAC $2.50 Peru $4 LAC $4 Peru Source: SEDLAC, CEDLAS, and The World Bank. Note: Income poverty is computed to be able to compare Peru’s performance with other countries in the region. 27 44. Both extreme and moderate poverty rates fell faster in Peru than the regional average. Using standardized poverty and extreme poverty lines for the LAC region, Peru started the period with poverty rates above the regional average. Peru’s rates fell to the regional average by 2009, and have since fallen even further (Figure 4). At the national level, for each percentage point increase in Peru’s GDP growth, poverty fell by 1.4 percentage points. On average, the poverty-growth elasticity was higher in Peru’s urban areas in the first half of the period (until 2010), after which rural poverty became substantially more responsive to growth. This trend is apparent especially in the Sierra and Selva regions of the country. A strong labor market generated better incomes, especially at the bottom of the income distribution 45. Labor earnings in urban and rural areas increased sharply in the last decade. Median monthly real income in urban areas increased by 45 percent between 2004 and 2015, from S/. 611 to S/.887. In rural areas, growth of the real median income was 54 percent, but median income in 2015 (S/.377) was still only 60 percent of urban median income in 2004. Hourly earnings show a similar pattern, with higher growth rates in rural areas (60 against 53 percent), reaching, in 2015, S/.2.5 per hour in rural and S/.4.8 per hour in urban areas, respectively. Nationally, median monthly wages grew by more than 50 percent in real terms during the last ten years, rising to S/.783 by 2015, and were a major driver for poverty reduction during this period. There are, however, important regional disparities. Wages in the Sierra and Selva regions are almost 40 and 30 percent lower, respectively, compared to those of Costa region. Similarly, wages in Lima are 60 percent greater than throughout the rest of Peru, and urban areas also show marked differences with rural areas. 46. Higher labor income was the primary driver of poverty reduction over the 2004 –15 period. The decomposition of changes in income into changes in employment, changes in labor and non-labor income (non-labor income split into transfers and others), and changes in the households’ dependency ratio reveals that labor income was the largest contributor to the decline in poverty, followed by changes in dependency and in non-labor income.12 In rural areas, transfers played an important role in poverty reduction, in particular during the last five-year period (Figure 5). Agriculture and services generated the largest increases in income for the poor: the contribution of agriculture was stronger in the reduction of extreme poverty, whereas labor income from services played a more significant role in moderate poverty reduction. 12 These decompositions are made using the official definition of employment used by INEI. Using the SEDLAC definition, which counts any person that receives zero income as unemployed, the employment effect is slightly higher and the labor income effect slightly lower. This is because an increase in earnings from individuals with no earnings, but classified as employed, has an employment effect under the SEDLAC definition, but not under the official definition. 28 Rural poverty reduction Urban poverty reduction 2004-2009 2004-2009 2010-2015 2010-2015 -25 -20 -15 -10 -5 0 5 -25 -20 -15 -10 -5 0 5 Labor income Participation Labor income Participation Non-labor income Transfers Non-labor income Transfers Source: Staff calculations based on ENAHO 47. As urban employment increased, rural employment fell, but the latter also diversified. Disaggregating the past decade into three sub-periods reveals distinct patterns of job creation by sector and region. Rural employment grew from 2002 to 2006, but it then stagnated and subsequently declined, driven by large contractions in agricultural employment. Nonetheless, services and construction jobs have increased in rural areas since 2010, but not enough to offset the move away from farm work. The large majority of jobs were created in urban areas, and across many sectors, especially wholesale and retail trade (24 percent of total new jobs), government services (16 percent), transportation and construction (12 percent each), and manufacturing (10 percent). Peru also made significant progress in increasing shared prosperity 48. Peru also made progress in reducing overall income inequality, albeit progress was more significant in urban areas. At the national level, the Gini coefficient fell by 5 points over the last 11 years: from 0.49 in 2004 to 0.44 in 2015 (Figure 6). Urban inequality fell by the same magnitude, dropping to 0.40 in 2014. In contrast, rural inequality—due in part to Peru’s varied geography—has proven harder to reduce, falling by only 2 points. As a result, urban inequality— which had been higher than rural inequality until 2008, fell to rural levels between 2008–10, and has been lower ever since. 49. Similarly, shared prosperity—the growth rate of real income per capita among the bottom 40 percent, increased significantly over the period. Income for the bottom 40 percent grew at a faster rate than average growth (6 percent against 4.4 percent over the period 2004–15). This pro-poor growth was remarkable within the region: it was similar to that of Brazil, and third only behind that of both Argentina and Bolivia (Figure 6). The Figure also presents the results for Peru divided in two periods. While the mean income and the income among the bottom 40 both grew at a lower rate during the most recent period (2008–13), the difference in growth rates in favor of the bottom 40 has been stronger in recent years. 29 0.53 0.51 0.49 0.47 Gini 0.45 0.43 0.41 0.39 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 National Urban Rural Source: ENAHO 5 Peru:03-08 Bolivia Brazil Colombia Peru Argentina - urban Peru:08-13 Mean Income Growth Overall 3 Chile 1 Panama Ecuador Mexico -3 -1 1 3 5 7 9 11 -1 El Salvador -3 Honduras -5 Mean Income Growth Bottom 40% Source: Authors’ elaboration with data from SEDLAC, CEDLAS and The World Bank, circa 2003 -2013 50. In line with similar patterns throughout the LAC region, the substantial fall in poverty was accompanied by a significant increase in the size of its middle class. The fall in poverty rates was accompanied by a significant increase in its middle class population, and in only a marginal increase in the share of its vulnerable population. Between 2004 and 2014, the share of its vulnerable population increased by 3.9 percentage points while the middle class grew by 18.5 percentage points (Figure 8). By 2014, the number of people in Peru’s middle class exceeded the number of its poor population. Despite this progress, the vulnerable class still accounts for the largest portion of Peru’s population. In 2014, while 40.6 percent of the Peruvian population was not considered poor, this group still faced the small but real probability of becoming poor if hit by a negative shock. Indeed, the income threshold for this group—which corresponds to US$10 per 30 capita per day—is defined so that, above it, the probability of falling into poverty is less than 10 percent (Ferreira et al. 2013). INCOME DISTRIBUTION (%) USD 2005 PPP 22.5 19.3 35 37.8 36.3 36.7 38.9 40.6 39.5 43 23.3 20.1 LAC PERU LAC PERU 2004 2014 Poor $4 Vulnerable $4-$10 Middle Class $10-$50 Source: SEDLAC, CEDLAS and The World Bank 51. At the same time, poverty is converging towards chronic poverty. Consistent with the prolonged and pro-poor growth experienced for more than a decade, upward mobility was substantial. A three-year panel analysis of households from 2007–10, shows that about 50 percent of the extreme poor and 32 percent of poor households in 2007 had exited extreme (total) poverty by 2010. Conversely, medium-term downward mobility is relatively low, and likely to be lower if considering longer periods of time. However, over the same three-year period, 67 percent of households who were the extreme poor in 2010 (and 82 percent of the moderately poor) were already poor in 2007. This suggests that there is a significant proportion of Peruvians who, despite the favorable economic conditions, were not able to leave poverty. A similar analysis over a longer period of time, applying a synthetic panel methodology, shows that nearly all of those who were poor in 2012 were also poor in 2004 (Vakis et al., 2015). There are large gaps between urban and rural poverty 52. Poverty is unevenly distributed across the country, and is concentrated in rural areas. Some Regions, such as Ica and Madre de Dios, have very low poverty rates (5 and 7 percent, respectively), while Cajamarca, Amazonas, Huancavelica and Ayacucho have poverty rates above 40 percent13. While less than a quarter of Peru’s population is rural, about half of the poor and 80 13Peru’s territory is divided into 26 administrative units, 25 Departments or “Regions” and the Callao Province. When the term “Region” is used in this report it refers to one of the country’s 25 regions, as administrative units. Regions are further di vided into provinces, which are further divided into districts. 31 percent of the extreme poor reside in rural areas. Finally, in 2015, the poverty headcount in rural areas was more than three times the urban headcount, and the extreme poverty headcount was 13 times higher (Figure 9). 83.4 48.2 45.2 41.6 14.5 13.9 5.7 1.0 MODERATE POVERTY (%) EXTREME POVERTY (%) MODERATE POVERTY (%) EXTREME POVERTY (%) URBAN RURAL 2004 2015 Source: ENAHO Source: INEI (2015) 53. Both total and extreme poverty are concentrated in relatively few districts. According to the 2012–13 district poverty map (Figure 10), almost half of the extreme poor are concentrated in approximately 8 percent of Peru’s districts, a large percentage of which are located in the Apurimac, Cajamarca, La Libertad, and Piura Regions (Genoni and Salazar, 2014). Similarly, 181 32 of the approximately 1,800 districts contain 54 percent of the total poor, and a large share of those districts are located in the Regions of Lima, Cajamarca, Piura, and La Libertad (Diaz et al., 2016). Poverty fell more in the poorest places, creating geographic convergence 54. The reduction in extreme poverty was more pronounced in Regions that had the highest poverty rates at the beginning of the period, thereby helping to diminish regional gaps. Changes in extreme poverty rates were positively correlated to the levels of extreme poverty in 2004: the higher the poverty rate in 2004, the larger the improvement by 2014 (Figure 11). The Huancavelica, Huánuco, and Puno Regions, which started the period with extreme poverty rates of over 40 percent, witnessed the best performance with reductions of over 30 percentage points. At the district level, an analysis based on estimated poverty maps between 2007 and 2012–13, suggests that progress was more likely among very poor districts in poorer Regions than among very poor districts in better-off Regions (Diaz et al., 2016). 55. There is no clear pattern of Regional convergence in the reduction of moderate poverty, but there is some convergence at the district level. There is no definitive relationship between the magnitude of moderate poverty reduction and baseline poverty rates in 2004. For example, the Piura and Cajamarca Regions started the period with similar poverty rates, of about 73-77 percent. By 2015, Piura’s poverty rate had fallen by 60 percent while Cajamarca’s progress was only about half as much (35 percent). At the district level, however, those districts with higher poverty rates at baseline experienced, in general, greater success in poverty reduction. However, many districts have not seen significant progress: 61 percent of the 1,179 districts with poverty rates above 50 percent in 2007 still have poverty rates above 50 percent six years later (Diaz et al. 2016). Madre de Dios Huancavelica Lambayeque La Libertad San Martín Moquegua Cajamarca Amazonas Ayacucho Apurimac Arequipa Huánuco Tumbes Ucayali Ancash Loreto Change in Total Poverty Rate (2015-2004) Callao Cusco Tacna Pasco Junín Puno Piura Lima Ica 0 70 Extreme Poverty Rate 2004 -10 60 -20 50 -30 40 -40 30 -50 20 -60 10 -70 0 Change in Percentage Points 95% Lower Bound 95% Upper Bound Extreme Poverty Rate 2004 Source: Authors, with data from ENAHO 33 56. Within Regions, poverty reduction is very heterogeneous. Among the Regions containing a large number of districts with poverty rates above 50 percent in 2007, some saw poverty rates in most of their districts fall below 50 percent by 2013, while others saw very few districts cross that 50 percent bar. Still others saw something in between (Figure 12). In Regions like Cajamarca and Ayacucho, more than 80 percent of districts can be classified as ‘chronically poor’: having poverty rates above 50 percent in both 2007 and 2013. At the other end of the spectrum, Cusco had more than 80 percent of districts with high poverty rates in 2007, but most of its districts had lowered their poverty rates to below 50 percent by 2013. Chronically poor Improved Worsened Below 50% always 100% 90% 80% 70% Share of Districts 60% 50% 40% 30% 20% 10% 0% An s Uc s Ap h o co Li o o sh o n o ali Tu os tí n d oq a o de a ca e H ac Ar cna a m ipa Ay rca Am iura a be M im Ic qu t Sa usc lla ua Pun ní r ta gu sc on La ore ay ca i eli im nu uc ar D Ju La equ m a Pa Ca L Ta ue ye be P jam av C az M uá ac ur L ba nc n Ca re ad H M Source: Diaz et al. 2016. 57. Mining activities at the district level are closely related to local poverty reduction, but the benefits from mining appear to be unevenly distributed across, and even within districts. Mining activity is beneficial for districts where production occurs, resulting in higher consumption per capita, and lower poverty and extreme poverty rates than in comparable non-producing districts (Loayza and Rigolini, 2016). The benefits, however, are limited to producing districts, and fall rapidly with increasing geographic distance from those districts. Mining also tends to increase inequality. For instance, the consumption Gini coefficient increases in all districts of mining provinces, and particularly in those districts where mining takes place. Moreover, inequality across districts increases as well, as mining benefits are substantially higher in producing than in non- producing districts, even those located in the same province. Furthermore, when isolating the effects of mining revenues (for example, the Canon Minero), these are insignificant at the local level, suggesting that benefits from mining are more likely to come from higher incomes in mining- related jobs rather than through improved public services and infrastructure. 34 Access to simultaneous services is still low, and water and sanitation gaps are large 58. The probability of an individual being in poverty is much higher when there is a lack of access to multiple services and assets. Asset endowments have been found to be the main factor explaining poverty reduction differences across districts, especially between rural and urban areas.14 Chronically poor districts lag behind improved and non-chronically poor districts in the number of households with simultaneous access to basic infrastructure services.15 In 2013, only 17 percent of households in chronically poor districts had simultaneous access to four basic services (water, sanitation, telephone and electricity), compared to 32 percent in improved districts and 75 percent in non-chronically poor districts. Conversely, there is evidence suggesting that households that own a combination of two or more assets perceive higher returns than the equivalent of the sum of individual returns in terms of income.16 59. Districts where access to more than one basic service increased have reduced poverty faster. Improved districts performed better than chronically poor districts in providing simultaneous access to four services between 2007 and 2013 (Diaz et al. 2016). The share of households with access to four services increased by 22 percentage points for improved districts, significantly higher than the increase of 12 percentage points among chronically poor districts. This evidence suggests that providing packages of services could trigger different services’ complementarities and increase assets’ returns. 60. One area of remarkable progress during the last decade was the increases in access to basic services for rural populations, especially electricity and telecommunications . Most Regions experienced substantial expansion in access to infrastructure services, and Regions that were below the national average experienced larger improvements. As a result, gaps in access across Peru’s diverse Regions decreased substantially between 2004 and 2015, especially in access to electricity and telephones. Still, large differences remain between urban and rural areas in the simultaneous access to basic infrastructure services. In urban areas, 81 percent of households have access to water, sanitation, electricity and telephone (86 percent if telephone is excluded). Furthermore, no urban households report a total absence of services. In rural areas, between 2004 and 2015, the share of households with no access to electricity, water or sanitation fell from 38 to 11 percent, and the share of households with access to those three services more than tripled, 14 See Escobal and Ponce (2009, 2011). However, in the short run, the authors argue that poverty may be driven by the rate of return to assets rather than access. 15 See Diaz et al. (2016). 16 See Escobal and Torero (2000). 35 increasing from 9 to 34 percent. Still, as many as 21 percent of rural households still had access to only one service—mostly electricity. 61. Nevertheless, there are still significant spatial disparities in the quality of energy services. In 2014, 2.2 million people still had no access to electricity. The distribution sector still relies on the state for service provision in Regions other than Lima, which together serve 62 percent of all users through different regional state-owned enterprises (SOEs).17 These SOEs face performance issues due to a lack of management capacity, financial restrictions for long-term investments, and inadequate cost recovery. As a result, two-thirds of the population are experiencing decreasing quality of services. A key sector challenge is thus modernizing the distribution sector to improve the quality of supply and reach the last mile in rural electrification. Moreover, a pricing policy for natural gas could promote the development of a market beyond the energy sector and deficient infrastructure constrains the hydrocarbons sector. Further, Peru’s electricity generation oversupply (with reserves margin of over 50 percent) can lead to underinvestment. While the marginal price of energy declined, the regulated tariff has not been changed, inducing unregulated consumers to switch to direct agreements with generators. This trend can force regulators to increase tariffs which, coupled with a decreasing quality of service, could raise social conflicts. The enactment of the legal framework to enable electricity exports could help solve the issue. Another important sector challenge is the lack of coordination and planning between the different entities and institutions which has resulted in the lack of a long- term energy policy definition and weak coordination between the Peru’s energy plans and its social and environmental objectives. 62. The gap in access to adequate water and sanitation is still significant. Peru is behind its structural and aspirational peers in terms of access to water and sanitation, particularly in rural areas. While 86 percent of households nationally have access to water, this figure drops to 64 percent for households in rural areas. Likewise, sanitation coverage is 77 percent nationally, but only 44 percent in rural areas, and 56 percent in the Selva region.18 Open defecation is close to 30 percent among the rural population. 63. In addition, there is still a large gap between urban, peri-urban, and rural areas in the quality of basic service provision. For example, for rural households, piped water coverage increased from 36 percent in 2004 to 65 percent in 2015, yet only 16 percent had drinking-quality water and fewer than 5 percent had water with some level of chlorine in 2015, well behind urban households (Figure 14 and Figure 15). Even though urban areas show significantly better coverage levels, in fast-growing peri-urban areas services tend to lag due to the unplanned urban expansion 17 They are under the administration of the Fondo Nacional de Financiamiento de la Actividad Empresarial del Estado, FONAFE, which is a holding company under the direct supervision of MEF. 18 ANA (2015). 36 and a lack of resources. As a result, households in these areas often pay much higher fees for access to these services. For instance, water can cost up to S/.65 per 5 cubic meters in informal settlements, compared to S/.12 for piped water to a residential area. 19 Continuity of water service is also erratic. For instance, in Trujillo, Peru’s third most important city, average service is only 17 hours per day, and in smaller cities and rural areas it can go as low as only one to two hours per day.20 Water Urban 100 80 60 40 20 Sanitation Rural 0 Water Rural Sanitation Urban Peru Structural peers Aspirational peers Source: World Development Indicators (WDI). Note: Structural Peers: Ecuador, Colombia, Mexico, Thailand, South Africa, Romania and Malaysia. Aspirational Peers: Poland, Korea Rep. and Chile. 64. Peru’s large urban-rural gaps in human development that have historically accompanied its income gaps are now somewhat smaller. For instance, the Human Development Index (HDI) shows a general increase, particularly among Peru’s poorer provinces. The average HDI grew from 0.49 in 1993 to 0.58 in 2007.21 The largest growth occurred in the southern Sierra provinces (Cusco, Puno, Ayacucho, Huancavelica and Apurimac), although areas in the northern Sierra also showed important increases. Similarly, disparities in life expectancy have also fallen significantly at the same time that the average life expectancy increased. Life expectancy in Peru went from 66 years in 1993 to 75 years in 2014. The regional gap between the lowest (Huancavelica) and the highest (Lima) was reduced three times, from 21 to 7 years. 22,23 65. Infant and maternal mortality rates have declined, but rural areas, particularly in the Sierra and Selva, have not yet achieved the national Millennium Development Goals (MDG). In 2015, national average for infant mortality fell to 15 per thousand infants born, surpassing the MDG goal of 18 (Figure 16). However, it is still much lower in urban areas (13 per 1,000 infants born) than in rural areas where it has not yet achieved the national MDG goal despite having achieved greater improvements in rural areas since 2009. Child mortality, at a national average of 19 El Comercio (2015). 20 SEDALIB S.A. (2015). 21 Remy (2015). 22 In Huancavelica life expectancy increased from 54.4 years to 71 years; in Lima it went from 76 to 77.4 years. 23 Still, there are significant life expectancy average differences between the Sierra and the Costa regions, with the Costa region having higher averages. 37 18 per thousand, has also surpassed the MDG goal (26 per thousand), but follows a similar pattern: at 28 per thousand children it is almost double in rural areas than in urban areas (15 per thousand children). The Sierra and Selva regions particularly stand out relative to the rest of Peru. Similarly, maternal mortality has decreased significantly since 1990, although in the 2004–10 period it was still considerably higher than the MDG goal of 66.3 per hundred thousand live births (Figure 17). Drinking Water Not-drinking water Safe Water Inadequate chlorine No chlorine 94 89 84 HEADCOUNT (%) HEADCOUNT (%) 38 34 29 16 11 1 5 URBAN RURAL URBAN RURAL Source: Staff calculations based on ENAHO Source: Staff calculations based on ENAHO. A substantial improvement in human development indicators has contributed to narrow the large spatial gaps 66. Stunting rates in children under five years of age have fallen considerably, but anemia rates in children remain a pressing issue. Stunting in children under 5 years of age, especially in rural areas, had been an important problem in Peru for a long time. Not until 2008 did stunting begin to decline sharply, falling to 14 percent in 2015, half of the prevalence during the decade before 2008 (Figure 18). Economic growth was an important driver in reducing stunting. In addition, over the last 15 years, the Government managed to secure an appropriate stream of resources through the Articulated Nutritional Program (Programa Articulado Nutricional, PAN— See Box 3). Resources were delivered through a results-based approach to the same districts prioritized by the conditional cash transfer (CCT) program, Juntos, providing a virtuous cycle in which supply and demand for vital health services intertwined, including a higher coverage of institutional births, the timely provision of the necessary nutrients and medicines, better performance of existing health services, improved access to water and sanitation services, among others--all policies strongly related to gains in maternal and early childhood health outcomes. This, together with increased household incomes from the economic growth allowed improvements in households’ caloric consumption and in resilience against health setbacks, in turn resulting in this remarkable reduction in stunting. However, anemia rates in children between 6 and 36 months have shown a less impressive downward trend, and have remained at around 43 percent since 2011 (Figure 19). Despite these breakthroughs, there are still marked geographical disparities in 38 outcomes of both stunting and anemia rates, which continue to be higher in rural areas, especially in the Sierra and Selva regions, and among indigenous people and the poorest population. 40 300 265 Demises per 100 thousand 250 30 200 185 20 births 150 10 93 100 66.3 0 50 National Urban Rural Lima Rest of Sierra Selva Coast 0 Infant Mortality (IM) Child Mortality (CM) 1990- 1994- 2004- Goal at IM Goal CM Goal 1996 2000 2010 2015 Source: Staff calculations based on ENDES Source: Staff calculations based on ENDES 67. Enrollment in pre-primary and secondary education has increased significantly. Net enrollment for pre-primary (3–5 years) increased from 53.5 percent in 2001 to 83.2 percent in 2015, surpassing even the LAC average in basic education coverage. This increase came with higher spending for education materials and an increase in the proportion of certified teachers, and a greater focus on formal education (that is, in an institutional setup, which represents almost 90 percent of enrollment). Primary enrollment is almost universal in Peru. In secondary education, net enrollment increased from 68.8 percent to 84.3 percent between 2001 and 2015. This increase was driven by a fall in cumulated dropout rates from 22 percent to 12 percent from 2005 to 2015 among students 13–19 years of age. Peru’s strategy to tackle malnutrition is a clear illustration of the importance of multi-pronged interventions in the area of child development. The country’s political commitment to achieve specific goals in the nutrition field led to a national strategy to reduce child malnutrition which was supported by the development of budgetary tools. In 2007, the Government adopted the Articulated Nutritional Program (PAN) to provide an umbrella for a number of multi-sectorial interventions aimed at reducing stunting. The PAN was designed to improve the efficiency of spending on nutrition and lined up national objectives, indicators and specific budgetary channels to finance key activities. The PAN setup of interventions was designed based on a UNICEF model that explains the relation between different drivers of malnutrition. In 2009, a system of incentives was put in place to resolve implementation bottlenecks and provide monetary incentives to regions and local governments to prioritize results in access to water, sanitation, health checkups, etc. Those incentives were initially piloted through the European Union’s EUROPAN and World Bank’s SWAp Results for Nutrition projects, and in 2013 they were institutionalized through the Fondo de Estimulo al Desempeño (FED). 39 Headcount (%) Headcount (%) 65 58 61 57 60 57 57 58 53 53 56 50 50 50 53 52 51 47 46 44 47 47 45 44 47 46 40 39 37 42 40 42 43 40 32 32 29 28 38 25 27 27 23 23 16 16 14 14 20 18 18 14 10 11 10 14 14 8 9 National Urban Rural National Urban Rural Source: Staff calculations based on ENDES Source: Staff calculations based on ENDES 68. Learning outcomes in primary and secondary education are low, but they are improving. According to the national test for second graders (ECE) developed by the Ministry of Education (Ministerio de Educacion, MINEDU) as part of the universal testing strategy, the proportion of students with a satisfactory level in Reading increased from a stunningly low 15.9 percent in 2007 to 49.8 percent in 2015. In Mathematics, the improvement has also been threefold, from 7.2 percent to 26.6 percent of students (Figure 20). Likewise, performance among secondary students also improved. Between 2000 and 2015, Peru’s results under the Programme for International Student Assessment (PISA) showed the largest improvement of any country. This reflects the country’s remarkable performance in increasing basic education learning outcomes in such a short period of time. Despite notable progress, there is still considerable room for improvement: still only half of the students can read at second grade level and almost three- quarters of second graders do not have a satisfactory level in Mathematics. Furthermore, in the last round of PISA in 2015, only 3.1 percent and 3.8 percent of students achieved a strong performance (level 4 or higher) in Mathematics and Reading, well behind OECD averages of 29.3 percent and 28.8 percent, respectively (Figure 21).24 These results improved in the 2015 PISA—Peru had the fourth strongest improvement in test scores. However, Peru still remains one of the worst performers among participating countries (ranked 65 out of 69 countries) and is still the worst performer among participating LAC countries other than the Dominican Republic (which participated for the first time in 2015). 69. Access to higher education has increased, but not for all. Graduation from higher education among young people (ages 22 to 24) increased from 12 to 19.8 percent between 2003 and 2013. Since 2013, however, there has been a slight downward trend (to 18.2 percent in 2015). 24 World Bank (2016b). 40 While 20.2 percent of non-poor in this age group has completed higher education, only 8.1 percent of poor youth have done so (3.3 percent among extreme poor). Furthermore, 60 percent of those graduates work in fields outside their expertise, including 54 percent of university graduates and two-thirds of technical higher education graduates. 60 Peru 6.4 47.5 49.8 50 Reading Chile 27.1 HEADCOUNT (%) Students with 40 level 1 or less 30 26.6 OECD 18.8 Students with 20 15.9 Peru 37.7 28.4 level 2 or more Mathematics 10 7.2 Chile 23.0 26.3 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 OECD 14.9 Reading Mathematics 100 0 100 % Source: MINEDU - Learning Quality Measurement Source: PISA (2016). Office. Although the labor market is a driver of poverty reduction, job quality is uneven 70. Informality is a generalized phenomenon in Peru.25 Informality rates have decreased in the past decade, but more than 70 percent of those employed were considered informal workers in 2015. The majority of new formal jobs created are fixed-term contracts. Informality is more common among young, poor, and low-educated workers. There are also regional disparities: workers located in rural areas and the Sierra and Selva regions are more likely to be part of the informal workforce. More than half of employment in almost all economic sectors is informal; the exceptions are in high skilled services, mining, and electricity and water, which combined represent a very small share of total employment (Figure 22). Moreover, informality is almost universal among self-employed workers and employees in small firms. Unlike other countries, where informal jobs constitute an entry into the labor market while workers “queue” for formal jobs, in Peru mobility from informal to formal employment is low. Education is the main driver for these transitions: workers without education are extremely likely to be stuck in informal jobs. Still, the correlations are rather low, and other demographic characteristics have even fainter associations, which points to the importance of demand side constraints. Irrespective of education, workers in certain types of jobs, economic sectors and geographic areas are more prone to remain 25For this document, we define informal employees as those not covered by social security (Seguro Social de Salud; EsSalud), and informal employers or own-account workers as those not registered with the tax authorities. All unpaid workers are considered informal, irrespective of the formal status of the enterprise that employs them. 41 in informality: farmers, the self-employed, those outside of Lima and in commerce, agriculture or, to a lesser extent, services are less likely to transit into formal jobs. Formal Informal 35 SHARE OF EMPLOYED (%) 30 25 20 33 7 15 24 8 14 10 18 12 5 8 6 10 13 5 9 5 2 4 3 5 2 3 0 2 1 1 2004 2015 2004 2015 2004 2015 2004 2015 2004 2015 2004 2015 2004 2015 2004 2015 2004 2015 Agriculture Mining Manufacturing Electricity and Construction Commerce Transports, Professional Domestic water storage and and financial Services restaurants services Source: Staff calculations based on ENAHO Indigenous, Afro-Peruvians, and women still face specific disadvantages 71. In 2010, Peru had the second largest population of indigenous people in Latin America, with its 7.6 million indigenous surpassed in absolute number only by Mexico (Figure 23).26 Indigenous people represent around one-quarter of Peru’s population, and are incredibly diverse, with the Quechua and Aymara being the largest groups. Interestingly, indigenous people are roughly equally distributed between urban and rural areas, with 53 percent living in urban and 47 percent in rural areas. Although the rural indigenous population is larger than the national average of 23 percent, in absolute numbers the indigenous are more concentrated in urban areas. However, the rural-urban split varies widely across Regions. In southern Regions like Madre de Dios, Moquegua, Arequipa, and Tacna over 70 percent of the indigenous population reside in urban areas. In contrast, in the north central Regions like Ancash, Huánuco, and Huancavelica, between 70 percent and 80 percent of the indigenous population live in rural areas. 72. Although indigenous people in Peru have historically suffered from poverty and exclusion, in recent years their incomes have improved substantially. As recently as a decade ago, over 70 percent of Peru’s indigenous people were poor or extremely poor, compared to slightly less than half of the non-indigenous population (Figure 24). The incidence of extreme poverty was almost three times higher among the indigenous than the non-indigenous, and the 26For countries without census data for 2010, the 2010 projection of the national population is used. The indigenous population was estimated using self-identification in all censuses, except for Peru, where surveys and censuses do not ask people to self- identify. Instead, the indigenous population was estimated by identifying as “indigenous” all the members of a household wher e the head of the household speaks an indigenous language. Thus, this definition of indigenous peoples includes Andean indigenous communities, traditionally self-identified and referred to as “peasant communities.” See World Bank, 2015b. 42 poverty incidence for the indigenous was 13 percent higher than for the non-indigenous. In the decade that followed, the poverty profile of the indigenous people in Peru improved dramatically. By 2015, only 31 percent of the country’s indigenous people were poor or extremely poor. However, while this improvement followed the pattern for the entire population, poverty fell more for the non-indigenous than for the indigenous (by 61 percent compared with 41 percent). As a result, the difference in poverty incidence between the two groups actually widened: in 2015, indigenous poverty was 70 percent higher than that of the non-indigenous. 18 45% 16 40% proportion of the total 14 35% 12 30% population millions 10 25% 8 20% 6 15% 4 10% 2 5% 0 0% Estimated indigenous population circa 2010 (in millions) Proportion of the total population (right axis) Source: World Bank, 2015(b). 27.0 51.5 69.0 81.7 43.1 38.1 29.5 25.3 14.9 10.5 3.4 5.7 NON-INDIGENOUS INDIGENOUS NON-INDIGENOUS INDIGENOUS 2005 2015 Extreme poor Poor Non-Poor Source: ENAHO. Note: Poverty computed with official poverty lines. 73. There is a clear link between residing in a rural area and poverty, even for the indigenous people. The higher poverty incidence among indigenous people is often driven by the fact that they live in rural areas, rather than by their ethnicity. For example, Regions like Moquegua and Ancash both have indigenous populations of about 40 percent; however, in Moquegua 70 percent live in urban areas, compared with only 30 percent in Ancash. Poverty in Moquegua is about 8 percent, while it is 24 percent in Ancash. A more rigorous regression analysis of the probability of being poor that takes into account several socio-demographic variables, including ethnicity and location, shows that the probability of being poor for an indigenous household is 43 positive and significant, all other characteristics equal. However, this probability is more than three times higher if the household is rural. In fact, conditional on being rural, the probability of an indigenous household being poor is only marginally higher, increasing by 0.016.27 A comparative analysis for other LAC countries revealed that the marginal effect of being indigenous on poverty was the lowest for Peru.28 74. Still, human development outcomes remain substantially lower among Peru’s indigenous population. A comparison between the indigenous and non-indigenous populations in 2015 shows that despite the improvements in monetary outcomes, human development still lags for indigenous people. The indigenous have significantly lower levels of education (on average, about three fewer years of education than non-indigenous, and are about 20 percentage points less likely to have a high school diploma or more) and lower access to basic services, in particular sewerage (54 percent have access compared with 73 percent among the non-indigenous). Even comparing the non-poor indigenous and non-indigenous reveals a similar pattern: the non-poor indigenous household heads have fewer years of education (6.4 years relative to 9.1), lower access to services, in particular sewerage (62 percent compared to 76 percent) and are less likely to have simultaneous access to all services.29 Among the urban indigenous, access to services is much lower than for the non-indigenous, and this difference is much more significant in Peru than in other LAC countries (Figure 25). Peru Mexico Latin America Dirt Dirt Dirt Floor Floor Floor 60% 60% 60% 40% 40% 40% No No No Slum 20% electri Slum 20% electr Slum 20% electri city icity city 0% 0% 0% No No No No No No sewer piped sewe piped sewer piped age water rage water age water Indigenous Non-indigenous Indigenous Non-indigenous Indigenous Non-indigenous Source: World Bank (2015b) 27 See Appendix B. 28 Other countries included Bolivia, Ecuador, Guatemala, and Mexico. See World Bank (2015b). 29 See Appendix B. 44 75. Children in indigenous households are also more likely to suffer from stunting, even in urban areas. Early childhood nutrition is crucial to enable the cognitive development of the child, and therefore has important implications for outcomes later in life, such as education attainment and earnings. In Peru, indigenous children are more likely to suffer from disadvantages in their cognitive development, relative to non-indigenous children, as a higher proportion of indigenous children suffer from stunting (Figure 26). In 2013, almost a third of indigenous children in urban areas suffered from stunting, close to three times more than non-indigenous children, while one in four rural indigenous children suffered from stunting. 44% 41% 39% 37% 33% 29% 29% 26% 18% 12% 10% 9% Indigena -rural No indigena - rural Indigena- urbano No indigena - urbano Source: Staff calculations based on ENDES (2013) In the beginning In Process Satisfactory 2015 54.3 27.0 18.8 TOTAL 58.4 49.7 2014 54.3 27.0 18.8 33.8 80.2 12.7 7.1 SHIPIBO 2015 29.5 2014 74.7 17.9 7.4 68.5 15.6 16.0 AWAJUN 2015 2014 79.2 13.1 7.7 INDIGENOUS INDIGENOUS NON - INDIGENOUS NON-INDIGENOUS LINGUISTI COLLAO C GROUPS 2015 54.8 22.7 22.5 OTHER 2014 56.7 26.6 16.7 QUECHUA- 2015 36.9 28.7 34.4 CUSCO 2014 40.8 34.7 24.6 26.1 28.4 45.5 AIMARA 2015 PRIMARY SECONDARY 2014 28.9 38.5 32.5 Source: Staff calculations based on MINEDU data Source: Staff calculations based on MINEDU data 76. Likewise, indigenous children have lower learning outcomes. Net school enrollment is significantly lower for indigenous students, especially at the pre-primary and secondary levels. 45 Moreover, a larger proportion of indigenous children are above-age in primary and secondary school (Figure 27). Half of indigenous students in primary and secondary school are above-age, whereas this proportion for non-indigenous children is 29.5 percent and 33.8 percent, respectively. Furthermore, although indigenous students have improved their reading comprehension both in their mother language and in Spanish, they still lag behind Spanish-speaking students (Figure 28). 77. Finally, indigenous people have worse labor market outcomes than non-indigenous. Indigenous people are less likely to be employees and more likely to be self-employed or unpaid workers. They are more likely to work in the primary sector and less likely to work in the services sector, and these patterns are similar among the non-poor indigenous. Further, even though indigenous people are more likely to participate in the labor market (about 5 percentage points, after controlling for gender, age, urban location, household size and education), their wages are about 6 percent lower than those of the non-indigenous (again after controlling for worker characteristics). For indigenous women, there is an additional 40 percent wage gap over men. 78. Recent evidence suggests that the living conditions of the Afro-Peruvian population have not improved dramatically over the last ten years. The regular household surveys in Peru are not designed to capture information regarding this minority group (which represents between 3–5 percent of the total population), and therefore there is scant quantitative evidence to track the evolution of their living conditions. However, a survey conducted in 2014 specifically targeting the Afro-Peruvian population revealed that while in some respects their living conditions improved over the last ten years, on average, their overall standard of living declined.30 Poor housing and overcrowded conditions are prevalent among the Afro-Peruvian population, relative to the rest of the Peruvian population. The share of Afro-Peruvian households with poor housing conditions increased from 5.3 percent in 2004 to 8 percent in 2014. This increase was especially large in the south Costa region (from 3.7 percent to 17.6 percent), reflecting the devastating effect of the 2007 earthquake on this population, and the slow recovery thereafter. 79. Economic and social outcomes for Afro-Peruvians still lag behind the rest of the population. For instance, labor participation has declined among Afro-Peruvians over the last ten years, and the disparity between female and male participation is larger than for other Peruvians. Job quality is generally lower for Afro-Peruvians as well: the percentage that has unskilled jobs has decreased, but it is still 10 percentage points higher than the national average. In education, enrollment, attendance, and attainment have improved but some differences remain. Finally, university and technical education attainment is expanding among the Afro-Peruvian population, but it is still lower compared to national data. In health, Afro-Peruvians show higher incidence of non-communicable diseases such as high blood pressure, cholesterol problems, diabetes, and heart 30 See Benavides et al. (2015). 46 problems. Finally, discrimination in Lima and urban areas seems to be prevalent and more likely based on racial (rather than economic or class) grounds. 80. For women, many socioeconomic human development outcomes are remarkably similar to those of men. Women head one in every 5 poor households and one in every 4 non- poor households. But overall female-headed households and male-headed households are very similar. On average, the household heads in both categories are around 50 years old (52.6 for women), they have less than complete secondary education (8.6 years for men and 7 years for women), their households have between four and five members and they are mostly urban (more so for female heads). Their housing conditions are also similar, and they have very similar labor participation rates. There are, nevertheless, some interesting differences. Female-headed households have larger shares of older adults, female heads are more likely than male heads to work as employees, and slightly less likely to be self-employed and they are more likely to work in retail and services and less likely to work in agriculture (which reflects their slightly higher urban concentration). Among the extremely poor and poor households, female-headed households have somewhat higher dependency ratios than male-headed households. 81. Basic health and education outcomes are similar among girls and boys as well. In health, indicators for girls are better than those for boys. Girls have a lower incidence of chronic malnutrition, lower incidence of anemia, and lower neonatal, infant, and child mortality. In education, enrollment levels across are extremely similar for girls and boys at all ages, and virtually universal among both sexes between the ages of 6 and 15. While overall enrollment rates are lower both at younger and older ages, there is no significant disparity between enrollment rates of girls and boys, except over the age of 18, when enrollment rates of girls exceed those of boys. 82. But there are some important gender gaps in education performance. Peru’s gender gap in education performance follows a common pattern found in almost all countries in the LAC: Boys outperform girls in Mathematics, and girls outperform boys in language. These differences are found in the results from both the national second-grade test (ECE) and international exams such as TERCE or PISA (Figure 29). 83. An important economic gender gap exists in labor market performance. About 1 in 4 women and 1 in 10 men were considered low earners in 2015, that is, workers whose monthly earnings were below the poverty line. Although these shares have fallen significantly since 2004 (when they were at 38 percent and 21 percent, respectively), the gap between the low earnings rate of women and men has widened, as men’s incomes have increased faster than women’s. Thus, women were 76 percent more likely than men to be low earners in 2004 and 140 percent more likely in 2015. This gender gap increases further when workers who need to work overtime to compensate for low earnings are counted as low earners. In other words, women are more likely to work longer hours to earn wages above the poverty line. Moreover, an analysis of differences 47 in hourly labor income that controls for individual characteristics, such as age, location, household size, education, sector of employment, and firm characteristics, also suggests that being male is associated with an additional 30 percent in hourly earnings on average. TERCE Math Test scores, boys and girls PISA Math Test scores, 15-year-old males and females Note: * denotes significant differences at 95% Note: * denotes significant differences at 95% confidence level. Mean score is 500, and standard confidence level. Mean score is 500, and standard deviation is 100. Source: GRADE (2015), using data deviation is 100. Source: GRADE (2015), using data from TERCE (UNESCO). from PISA (OECD). 84. Domestic violence is a well-known social problem in Peru that affects a large number of families, with serious individual and social consequences. Women are frequently subjected to verbal, psychological, and physical abuse, both inside and outside their households. According to the latest demographic and health survey (Encuesta Demográfica y de Salud Familiar; ENDES), which collects information on women 15–49 years old, and children between 1 and 5 years of age, domestic violence against women and girls is widespread. In 2015, 70.8 percent of women who have had a partner reported suffering some kind of violence by the husband or partner, including psychological or verbal abuse (67.4 percent), physical violence (32 percent), and sexual abuse (7.9 percent). Domestic violence is not a phenomenon restricted to low-income households: although slightly lower for the highest quintile, the share of women reporting various types of abuse is remarkably similar across income quintiles (Figure 30). Violence towards women has wider repercussions, as it affects the development of children in the household, with negative consequences on their schooling performance. Indeed, a recent study (Alcazar and Ocampo 2016) 48 shows that exposure to domestic violence towards the mother increases the probability of grade repetition for children below 11 years of age. PERCENTAGE OF WOMEN BY INCOME GROUP, 2015 70.6 70.5 69.0 67.4 66.2 59.2 36.7 35.6 32.9 32.0 30.8 22.1 9.2 9.1 8.4 7.9 7.7 4.3 PSYCHOLOGICAL/VERBAL PHYSICAL SEXUAL Lowest quintile Q2 Q3 Q4 Highest quintile Average Source: ENDES (2015). 49 Peru experienced a period of remarkable growth averaging over 5 percent annually over the last 15 years. Growth was led by factor accumulation, made possible by Peru’s rich endowment of natural resources, which attracted large (foreign) investments in mining and related sectors, despite little improvement in productivity. Peru’s low aggregate productivity stems in part from an inefficient allocation of capital and labor among firms, especially in the services sector, indicating severe constraints in product and labor markets. The large number of informal firms likely contributes to this misallocation, limiting the size and growth of more productive (informal) firms. The lack of technology adoption and integration into global value chains further limit productivity growth. The low productivity, in turn, limits firms’ export competitiveness, their ability to adopt new technologies, and the demand for well-paid jobs, likely contributing to the high labor informality. Raising aggregate productivity would have the highest growth payoff while the potential gains from increasing capital or labor are modest. Peru’s potential productivity gains are significantly larger than in other Latin American countries. The low private sector productivity is thus a key impediment to sustaining high returns to investment and creating the demand for better-paid jobs and income opportunities for the bottom 40 percent. Addressing this structural challenge is vital to continue the path of shared prosperity (see Figure 45). 85. Peru doubled its real per capita income during the 12-year period of high growth between 2002 and 2013. The country grew at an average rate of 4.4 percent per year during 1990– 2015, compared to global and LAC regional growth of around 3 percent. Growth was even faster since 2001, averaging 5.3 percent, which was one of the highest in Latin America. As a result, income per capita started to converge and Peru achieved high middle-income status in 2008. 86. This high growth marked a recovery period after the 1980s’ output collapse that was accompanied by political and social turmoil and followed by high output volatility in the 1990s. Dramatic changes in policy orientation until the year 2000 impeded economic convergence. After robust economic performance during the 1960s, averaging more than 5 percent annually, growth decelerated in the 1970s and collapsed in the 1980s―Peru’s GDP contracted by almost 15 percent between 1981 and 1990 (Figure 31, right). The economic crisis was triggered by growing domestic and external imbalances that were rooted in populist economic policies such as the nationalization of mining companies and implementation of a program of land reform. Economic mismanagement finally led to a debt crisis and hyperinflation in the late 1980s. Structural reforms implemented in the early 1990s (for example, removal of price controls, trade liberalization, and strengthening Central Bank autonomy) ensured a stable macroeconomic environment. The 50 volatility of output growth declined somewhat in the 1990s but investment uncertainty remained high as high-level corruption scandals undermined the rule of law. 87. Peru’s income per capita relative to that of the United States is still slightly below the level in the 1970s. The economic situation turned around in the 2000s, a decade characterized by macroeconomic stability, trade openness, and fewer social and economic tensions—the latter assisted by a political decentralization agenda that redistributed large parts of the mining receipts to the local governments hosting the mines. In recent years, Peru’s GDP per capita relative to that of the United States has exceeded 20 percent, still somewhat below the levels in the 1960s and 70s (Figure 31, left). While adverse external economic conditions slowed growth to about 2-3 percent in 2014 and 2015, the current growth rate is still well above the levels of many other commodity exporters in the LAC region. 50% 5.5 45% 5.3 5.6 4.8 40% 35% 3.7 3.9 30% 25% 20% 15% 10% -1.0 5% 0% 1951-1960 1961-1970 1971-1980 1981-1990 1991-2000 2001-2010 2011-2015 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Chile China Malaysia Thailand Peru Sources: World Bank Staff calculation based on Total Economy Database by The Conference Board (left); World Bank staff calculations based on data from the Central Reserve Bank of Peru (right). Note: GDP per capita based on constant 2014 US$ per capita, 1950-2015 (left); average annual GDP growth by decades (right). 88. Capital accumulation has been the main driver of growth, accounting for more than two-third of Peru’s high growth rates of about 6 percent annually since 2001. Capital accounted for 70 percent of total growth when including the contribution of investments in information technology (IT). Since 2005, capital accumulation accounted for 82 percent of aggregate growth―non-IT capital accounted for 74 percent and IT capital for 8 percent (Figure 32, right). Labor contributed about 20 percent to aggregate growth between 1995 and 2014 based on high labor force participation rates of more than 70 percent. In contrast, the contributions of 51 quality improvements in human capital (3 percent) were negligible. Similarly, increases in total factor productivity (TFP) played a relatively minor role, contributing only 11 percent to growth between 2000 and 2014. We note that we account for changes in the quality and utilization of capital and labor when computing TFP, for instance, by distinguishing between IT and other forms of firms’ physical capital. Applying this more refined measure of TFP is important in the case of Peru because it reduces the TFP contribution to growth in the past 15 years compared to previous studies not making these adjustments and allows for more adequate comparisons across countries.31 We also note that the contribution of TFP to growth should be analyzed over the longer horizon since it has been established that it is typically overestimated in (boom) periods of high commodity prices.32 The stronger contribution of TFP to growth in Peru in the 2000s (Figure 32, right) should thus be regarded with some caution, and the TFP contribution should instead be evaluated over the longer term including periods of lower prices (e.g., 2000-14). Despite the capital investment boom, the level of capital accumulation in recent years, i.e. the economy’s capital intensity, is close to the level predicted by Peru’s income per capita.33 This implies that Peru’s capital stock was significantly below its equilibrium level in the 1980s and 1990s due to the output collapse and the associated investment uncertainty. 7% 7% 5% 5% 3% 3% 1% 1% -1% -1% 1991-2000 2001-2010 2011-2015 95-99 00-04 05-09 10-14 Private Consumption Private Investment Public Spending Net Exports Labor Labor Quality Non-IT Capital GDP IT Capital TFP Growth Source: World Bank Staff calculation based on data from the Central Bank of Peru (left); World Bank Staff based on the data from The Conference Board (right). Note: Data in percentage points (pp). 31 The recent World Bank (2015a) report estimated a higher contribution of TFP to growth of about 30 percent from 2000-13. The report did, however, not account for changes in IT capital and capital utilization so that the contribution of these factors were attributed to the residual TFP measure instead of capital accumulation. Accounting for changes in the quality and utilization of capital and labor appears to be especially important in recent years―we find that TFP contributed about 30 percent to growth in the 2000s but that its contribution turned negative since 2010. This trend is consistent the findings of Céspedes and Ramírez- Rondán (2014) who also show that the TFP contribution to growth in Peru declines once one controls for the quality and utilization of capital; the authors show that it is important to account for the total hours worked which have been declining over past years. 32 The computation of TFP is based on measuring revenues (value added) which need to be adjusted for yearly price changes over time. Output prices, however, are typically only observed at the industry level and not for all individual products in an economy. Therefore, in periods of high commodity prices, the use of (to low) industry level prices leads to an overestimation of the contribution of the residual output measure (that is TFP) to output growth (see Foster et al., 2008). 33 World Bank (2015a). 52 89. Growth has been led by private investments and consumption since 2001. The share of total investment in GDP doubled from about 10 percent in the early 1990s to more than 20 percent in recent years, with private investments accounting for the bulk of capital accumulation (about 75 percent). Private investments grew by an average annual rate of about 12 percent from 2001-2015 compared to about 5 percent in the 1990s. The contribution of private investment to growth increased from 14 percent between 1991 and 2001 to 30 percent from 2001-15 (Figure 32, left). Private domestic consumption accounted for about 55 percent of total growth in both periods (Figure 32, left). The average annual growth rate in private consumption increased from 2.9 percent between 1991 and 2001 to 5 percent between 2001 and 2015. Thus, the high growth in private investment did not crowd out domestic consumption (savings) in the high-growth period since 2001. Foreign private investment in mining has been strong 90. Mining investment surged over the past 15 years, including large inflows of foreign direct investment (FDI). Peru is one the world’s largest producer of metals. It recently became the second largest producer of copper (behind Chile) and is the second largest producer of silver, third of zinc and tin, and seventh of gold. Mining accounted for about 15 percent of GDP in 2015 and mining exports totaled US$20.4 billion in 2014, over 50 percent of total exports. Mining investments accounted on average for about 20 percent of total private investments over the last 15 years and peaked at almost US$10 billion in 2013. The surge in mining investment was supported by a large inflow of FDI, which has increased four-fold in the last 15 years. Total FDI amounted to US$61 billion between 2006 and July 2016; the mining sector absorbed more than half of all FDI during this period (Figure 33). 50% 40% 30% 20% 10% 0% Peru Turkey China Brazil Mining Utilities & Other services Other manufacturing IT&Communications Chemicals Hotels&Tourism Financial services Real Estate Machinery Source: World Bank staff based on FDI markets database. Note: Share of sector FDI in total FDI into Peru between 2006 and July 2016; Peru’s total FDI in that period amounted to US$60,924 million. 53 91. Several upstream goods and service producers benefited significantly from the mining sectors’ higher domestic demand. Forward linkages from mining to other sectors are negligible in Peru―only the domestic precious and non-ferrous metals industry demands a notable fraction of the metal mining industry’s output (14 percent). Backward linkages are stronger. For instance, 68 percent of the output of the basic chemical sector is sold to mining companies (Figure 34). Several other manufacturing sectors also sell significant shares of their output to mining companies, such as manufacturers of rubber (40 percent), other metal products (33 percent), and machinery and equipment (29 percent). Several higher productivity modern service sectors also benefitted substantially from the mining boom. Mining companies accounted for 16 percent of output for electricity and gas, 12 percent for land transport, and 10 percent for financial and professional services. Financial Services Professonal services % of Output Supplied to Mining Sector Land transport Refined Oil % of total Mining Inputs Other chemicals Employment agencies Electricity and gas Iron & steel Manuf machinery & equ Other metal products Renting machinery & equ Manuf rubber products Basic chemicals 0% 10% 20% 30% 40% 50% 60% 70% Source: World Bank staff based on INEI input-output data from 2008. Note: All sectors for which the share of output supplied to mining companies exceeds 10 percent of their output (in percent of supplying sector’s output). 92. The extractive industries sector remains competitive despite lower international prices. Weak linkages with the local economy further undermine local support of large mining projects. Investment is projected to decline in 2016–17, due to the phasing out of some large mining projects and low prices of copper, gold, and zinc.34 Mining investments are projected to continue declining, reaching a new balance significantly below its peak in 2013. Still, this new, lower level of investment is expected to be above its pre-commodity boom level as a result of the sector’s high international competitiveness due to high metal grades and relatively low energy, water, and transport costs (for mining products). Mining output is projected to increase in the coming years despite lower international prices, due to the substantial increase in capacity. Further, 34One Peruvian private sector think tank projects a decline in mining investment of up to 25 percent in 2016 and another 15 percent in 2017, due to the phasing out of several major mining projects (for example, Cerro Verde). 54 there are large untapped hydrocarbon reserves with large exploration opportunities—Peru has 18 hydrocarbon basins, but 98 percent of the exploratory wells have been drilled in only four basins. Petroperú’s involvement in exploration has been small, and it has been unsuccessful in promoting foreign investment due to lack of an open and stable regulatory environment. Increases in regulatory costs and potential social tensions in the areas hosting extractive industries are considered major investments risks. 93. Low energy costs resulting from abundant domestic sources and successful sector reforms promote the competitiveness of mining and other sectors. Peru has abundant domestic sources of energy—hydroelectricity accounted for 38 percent of the installed generation capacity and natural gas for 45 percent in 2014. The Camisea gas project, for instance, has provided a reliable source of cheap and environmentally friendly energy since 2004 and has made the country an exporter of liquefied gas. Peru has a modern regulatory system for the energy sector. The activities of generation, transmission, and distribution are separated by regulation since 1992. The management and operation of the main electricity facilities in generation and transmission have been transferred to the private sector. Macroeconomic stability has supported high savings 94. Sound macroeconomic management allowed Peru to maximize the long-term benefits of the commodity super-cycle. The country benefitted from prudent fiscal and monetary policies. Throughout the period of high growth fueled by high commodity prices between 2004 and 2013, and unlike many countries, Peru saved the commodities’ windfall (Figure 35, right), leaving significant savings to provide for needed investments and strong macroeconomic buffers to face more challenging times. The countercyclical fiscal policy (Figure 35, left) contributed to public savings, which played an important role in boosting domestic savings over the last 10 years from just above 10 percent of GDP in the early 1990s to 24 percent in 2014. The Central Bank of Peru (Banco Central de Reserva del Perú; BCRP) kept an open capital account, but actively intervened in the foreign exchange market to smooth out short-term volatility and maintain a stable and appropriate real exchange rate, which has been critical to external sustainability. The modest current account imbalances were covered by FDI and portfolio investments. The BCRP has also adhered to a well-established inflation-targeting regime since 2002, which, together with the countercyclical fiscal policy, has stabilized the economy and anchored low inflation expectations. Foreign exchange reserves stood at 32 percent of GDP or 20 months of imports by the end of 2015. Public debt has remained low, around 20–25 percent of GDP in recent years and net public debt was only 3 percent of GDP in 2015, one of the lowest in the world. 55 4 6 3 Percentage points of GDP 5 1.8 Fiscal balnace (in %) 2 4 1 0.3 0 3 -1 2 3.9 3.3 -2 1 -3 0 -4 LAC5 Perú -4 -2 0 2 4 Output gap (in %) Consumo Ahorro Promedio de crecimiento anual Source: Staff calculations based on UN and BCRP data. LAC5 includes Argentina, Brazil, Chile, Colombia and Mexico. 95. The recovery of capital accumulation was supported by foreign and domestic savings. Prudent macroeconomic management and an increasingly open economy ensured access to international capital markets and attracted foreign investments. Fiscal and monetary policy generated the appropriate incentives to maintain a balance between the growth of domestic consumption and savings. In contrast to most other major commodity exporters in the LAC region, Peru’s high domestic savings helped finance the acceleration in capital accumulation, with domestic financing exceeding 80 percent of total in recent years. 96. And financial intermediation increased significantly, albeit from a low base, while the rate of dollarization declined, reducing financial sector volatility. Credit to the private sector surpassed 30 percent of GDP and the liquidity of the financial system improved. Deposits to GDP increased simultaneously from 21 percent in 2004 to 32 percent in 2013 (Figure 36, right). Even so, the volume of credit to the private sector is lower than predicted by the country’s level of development. At the same time, the rate of dollarization declined, reducing exposure to exchange rate fluctuations. The dollarization rate of net loans declined from about 50 percent in 2010 to almost 30 percent by July 2016. Deposits denominated in US$ remained at about 50 percent of total deposits for the same period. The decline in the financial sector dollarization was enabled by several measures implemented by the BCP. The recent depreciation of the domestic currency further helped to reduce the demand for US$-denominated loans. 97. As a result, the share of investment converged to pre-crisis levels. The average total investment rate dropped from 27 percent in the first five years of the 1980s to 16.7 percent in the first five years of the 1990s. The private investment share collapsed from 18 to 12.6 percent in the same period. It then started to accelerate in the early 2000s as macroeconomic stability was 56 regained and recovered to an average of 20.8 percent between 2010 and 2015, slightly above its pre-crisis levels and the high levels of the 1950s (Figure 36, left). Similarly, the total investment share has recovered to 26.2 percent of GDP in the past six years (2010–15). 25 45% Investment rate as % of GDP 20 35% 15 10 25% 5 0 15% 2004 2007 2010 2013 Deposits Peru Deposits LAC Deposits World Credits Peru Privada Publica Total Credits LAC Credits World Source: World Bank staff based on UN and BCRP data (left); World Bank staff based on the Global Financial Development database (right). But financial inclusion and structural change have been slow 98. Peru’s financial inclusion continues to be low compared to its regional peers, pointing to inefficiencies in the financial sector. Peru’s credit to GDP ratio of about 34 percent lags behind the 45 percent average for the LAC region. Only 29 percent of the population has a savings account, lower than in Bolivia (42 percent), Colombia (39 percent), Ecuador (46 percent), and Chile (63 percent).35 Only 22 percent of women report owning a deposit or transaction account in Peru--one of the largest gender gaps in the world. Peru is ranked as having one of the world's most developed microfinance sectors—aggregated assets account for 5.4 percent of the financial sector, while outstanding credit and deposits account for 10 and 8 percent, respectively. Non-bank financial institutions, especially microfinance institutions, contribute to financial inclusion, accounting for one-third of the total financial assets in 2015. Apart from limited financial literacy, the high costs of accessing financial services is the main obstacle to greater penetration. High transaction costs, the high costs associated with the opening of new branches, and limited connectivity and poor infrastructure, have all been identified as the main obstacles to broader financial inclusion.36 99. The percentage of firms with a bank loan or credit line is above the average for the LAC region but capital markets are still underdeveloped. The percentage of firms with a bank loan or credit line (60 percent) is above the average in the LAC region (45 percent) and only a 35 World Bank, Global Findex Database: http://datatopics.worldbank.org/financialinclusion/. 36 Center of Financial Inclusion (2013). 57 fraction of firms report access to finance as a major constraint. Peru also has made significant advances in the implementation of the Basel III requirements as the Superintendence of Banks, Insurance, and Pension Funds (Superintendencia de Banca, Seguros y AFP; SBS), started to implement the corresponding banking regulations. Despite the country’s continuous classification as an emerging market, however, capital markets are not liquid due to low stock trading volumes— averaging only 1.8 percent of GDP in 2014 compared to the Latin American average of 16.9 percent and the OECD average of 131 percent--and an underdeveloped bond market. The amount traded in fixed income instruments in Peru’s local market was 0.47 percent of GDP compared to 185 percent in Colombia and 101 percent in Chile. The insurance market is also underdeveloped with insurance spending amounting to only 1.7 percent of GDP in 2013 compared to 2.7 in Colombia and an OECD average of 8.4 percent. 100. Over the past 20 years, labor moved from low-productivity, rural agriculture to somewhat higher productivity services activities, often in the informal urban sector. The labor force working in agriculture declined from 32 percent in 2001 to 24 percent in 2013. At the same time, the share of labor working in retail trade, hotels and restaurants, and transport increased. This shift raised the economy’s aggregate productivity, since the declining share of labor in low- productivity subsistent agricultural activities in the Sierra and Selva was compensated by an increase in the labor share of somewhat higher productivity basic service activities, often in the urban informal economy (Figure 37, right).37 Nontradable sectors (services, utilities, and construction) grew the fastest, accounting for two-thirds of aggregate GDP growth since 1991. The growth in non-tradable sectors was spurred by a strong recovery in domestic demand. 101. But unlike other fast-growing middle-income countries, Peru’s reallocation of labor across sectors (that is structural change) has been slow overall, signaling low growth in the potentially higher productivity manufacturing and service sectors. The composition of GDP into agriculture, industry, and services has hardly changed since 1990 (Figure 37, left). All three sectors grew at a comparable pace so that their contribution to GDP was determined by their size, with services accounting for about 3 percentage points of GDP growth since 1990 (industry 1.8 and agriculture 0.3). Production structures hardly changed even within service sectors: the share of many service activities in GDP remained constant between 1994 and 2015, apart from a slight increase in retail, financial, and IT services at the expense of health, education, and other services. 37Aggregate growth of added per worker can be decomposed into labor productivity growth within sectors or structural shifts in employment from lower- to higher-productivity sectors. Measures of value added per worker and employment patterns across 15 economic sectors were used to estimate the relative contributions of these two sources. Shifts in employment from agriculture to construction and services accounted for about 1 percent of the total labor productivity growth of 4 percent between 2002 and 2012―within sector gains accounted for the rest. Following the critique of this approach, we also constructed marginal labor productivity measures by sector and conducted similar analysis (using the World Bank I2D2 database for wages). The ranking across sectors using marginal measures of labor productivity was the same. 58 1.0 Mining Utilities 6% 0.8 Log(sector prod / average prod) 0.6 Financial serv 8% 0.4 0.2 Manufacturing 35% Construction 33% 0.0 Transport 59% 59% -0.2 Gov.Serv 1990 Commerce -0.4 2015 -0.6 Agriculture -0.8 -1.0 Services Industry Agriculture -0.10 -0.05 0.00 0.05 Change in employment share (1990-2011) Source: World Bank staff based on INEI data (left); Araujo et al. (2015) based on Groningen data (right). 1.5 Annual average change in employment 1 share (percentage points) 0.5 0 -0.5 -1 -1.5 -2 TUR ZAF LKA BRB MYS CRI MNG CHN PAN GTM DOM UKR CHL NAM PAK SLV EGY GHA HND SRB JAM MUS BTN ARG BOL THA IND KAZ PER NIC PHIL UGA High-skilled occupations (intensive in non-routine cognitive and interpersonal skills) Middle-skilled occupations (intensive in routine cognitive and manual skills) Low-skilled occupations (intensive in non-routine manual skills) Source: World Bank 2016. Note: The figure displays average annual changes in employment shares (percentage points) between circa 1995 and circa 2012. Classification follows Autor (2014). High-skilled occupations include (i) legislators, senior officials and managers, (ii) professionals and (iii) technicians and associate professionals; middle- skilled occupations comprise (i) clerks, (ii) craft and related trades workers, (iii) plant and machine operators and assemblers; low-skilled occupations are (i) service and sales workers and (ii) elementary occupations. 102. The composition of employment in terms of high, medium, and low skill occupations has not changed since 1995, confirming that higher productivity activities failed to absorb more workers despite high economic growth overall. Employment shares were growing in high- skilled, high-paying occupations (managers, professionals, technicians) in most other countries over the past 20 years. This trend reflects the changes in the structural composition of employment, in particular the growth of demand for high-skill occupations that complemented new technology in other fast-growth, emerging economies that became technologically more advanced. Low- 59 skilled, low-paying occupations (elementary, service and sales workers) typically also grew while the demand for middle-skilled, middle-paying occupations (clerks, plant and machine operators) has often been squeezed since most of these occupations reflect routine work that is susceptible to automation as firms start to adopt new technologies.38 In Peru, however, the composition of employment has barely changed over the past 20 years. The absence of changes in the demand for occupations with different skill intensities highlights the lack of structural change in the economy despite substantial trade and investment liberalization. Non-mining firms trade very little, indicating low productivity growth 103. Peru has one of the world’s most liberal trade policy regimes. Trade liberalization began in 1990 when Peru cut tariffs significantly, eliminated most nontariff barriers, and opened several service sectors to foreign investment. By 2013, Peru had lowered its average tariff39 to 1.9 percent, which is one of the lowest worldwide. Moreover, only around 37 percent of imports are subject to nontariff technical barriers to trade (NTMs), including import licenses, quality inspections, and rules of origin. This figure is lower than in most comparator countries such as Mexico (43 percent), Chile (47 percent), South Africa (52 percent), Colombia (70 percent), and China (90 percent). Peru also signed free trade agreements with its main trading partners, including the United States, the European Union and China. Increased openness of the economy allowed Peru to take advantage of the rapid growth in world trade. Exports accounted for only 17 percent of GDP in 1990, immediately after the output collapse, but recovered continuously in consecutive years, reaching a peak of 33 percent in 2005 supported by favorable external conditions. The share of exports in GDP amounted to about 25 percent of GDP in 2015. 104. But non-mining firms trade very little. Peru leveraged the boom in global commodity markets for rapid growth in the 2000s but still trades much less than countries at similar income levels. Peru exports primarily raw or semi-processed goods instead of high added value products. Commodity exports accounted for 70 percent of total exports in 2014. Copper and gold—the two most important commodities in Peru—accounted for about 35 percent of total exports. The quantity and value of exports of apparel products, and of plastics and metal products—such as automotive parts—grew in recent years (albeit from a low base), partly due to better access to the U.S. market. The services trade accounts for 20 percent of world trade but only 5 percent of Peru’s exports. Travel and transport services grew the fastest and account for the bulk of services exports (see also Box 8). Overall, Peruvian firms are poorly integrated into global value chains. Peru’s low trade cannot be explained by Dutch disease effects, as the Government followed a managed 38 In developing countries, the average decline in the share of routine employment was 7.8 percentage points between 1995 and 2012. Other than technology, urbanization and trade are also likely to influence this trend. See World Bank (2016e) for details. 39 Weighted average (by trade volume) most-favored-nation tariff. 60 floating exchange rate regime, preventing a strong appreciation of the real effective exchange rate during the commodity boom years of the 2000s.40 105. The composition of exports has hardly changed in the last 50 years as Peru developed few new export successes. Commodities accounted for about 66 percent of export in 2015, the same level as in the 1970s. Five sectors (minerals, metals, vegetables, foods, and textiles and apparel) accounted for almost 90 percent of exports in 2013, virtually unchanged from the 92 percent share in 1970. In other words, the sectors that led the Peruvian exports recovery since 1990—hydrocarbons, mining, and traditional agriculture—were the same sectors that collapsed in the 1980s. Peru underwent very little structural transformation and diversification in response to its export collapse. Consistent with this trend, almost all (95 percent) of aggregate export growth from 2007–13 came from selling existing export products to existing markets―the diversification of products or markets contributed little. Peru also developed few new recent export successes despite its high growth. Peru has revealed comparative advantages (RCA) in exporting agricultural and mining-based products, including fresh or chilled vegetables, copper or zinc alloys, metallic salts, and manufacturers of asbestos. Over time, a few new RCAs emerged (colored green), primarily in garments (Figure 39). But overall Peruvian firms developed few new export successes from 2000–10. Instead, 68 percent of all exported products in which Peru has a RCA were primary and resource-based products in 2010. Peru only gained RCA in one medium-tech manufacturing sector, while it lost RCA in three medium-tech sectors from 2000-10. Source: Araujo et al. (2015). Note: The graph shows Peru’s product space in 2010. The product space is a graphical representation of the relatedness between every pair of the 775 4-digit SITC manufacturing products whereby each node represents a product and distances between two products represent the similarity between their production structures. More complex, closely-related products such as machinery and equipment, motor vehicles, and chemicals are located in the densely populated core. Peru’s main exports such as mining products, agribusiness, and textiles are in the periphery of the product space. Blue triangle: classic products that have a revealed comparative advantage 40 World Bank (2016g). 61 (RCA) in 2000 and 2010; red square: disappearing products with an RCA in 2002 but not in 2010; green diamond: emerging products with an RCA in 2010 but not 2002. Final yellow pentagon: marginal products for which Peru has not acquired an RCA (0.5