MEXICO Poverty and Equity Assessment MEXICO Poverty and Equity Assessment A reproducibility package is available for Section 1 and Annexes of this report in the Reproducible Research Repository at https://reproducibility.worldbank.org/index.php/catalog/236 Copyright © 2024, International Bank for Reconstruction and Development / World Bank 1818 H Street N.W. Washington D.C. 20433, United States of America Telephone: (202) 473-1000 Internet: www.worldbank.org In Spanish: www.bancomundial.org Email: feedback@worldbank.org Rights Reserved This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this publication. Rights and Permissions The International Bank for Reconstruction and Development/The World Bank encourages the dissemination of its work and will normally grant permission to reproduce portions of this work promptly, provided the sources are acknowledged. Attribution—Please cite the work as follows: World Bank (2024), Mexico Poverty and Equity Assessment. Cover and Interior Design Kilka Diseño Gráfico SAS Table of Contents Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Evolution of vulnerability to climate events for housing in Mexico . . . . . . . . . . . . . . . . . . . . . . . . 59 Section 1: Monetary Poverty and Inequality in Mexico . . . . . . . . . . . . . 16 A note on climate events, infrastructure and territorial context of poverty in Mexico . . . . . . . . 62 Monetary poverty . . . . . . . . . . . . . . . . . . . . . . . . 17 Some international comparisons . . . . . . . . . . . . . . 64 Chronic Poverty and Income Mobility from labor incomes . . . . . . . . . . . . . . . . . . . . . . . . 28 Section 4: What would it take to eradicate extreme Inequality of Incomes . . . . . . . . . . . . . . . . . . . . . . 30 poverty in Mexico by 2030? . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Some international comparisons . . . . . . . . . . . . . . 33 How much growth is needed? . . . . . . . . . . . . . . . . 67 Section 2: How to increase the poverty reduction impact of Social deprivations in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . 38 social public expenditures? . . . . . . . . . . . . . . . . . . 68 Advances and reversals of multidimensional How to reduce vulnerability to climate events? . . . . . 70 poverty in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . 39 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Public expenditures and its distribution explain social deprivations in Mexico . . . . . . . . . . . . . . . . . 42 Annexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Some international comparisons . . . . . . . . . . . . . . . 50 Annex 1. Comparison of monetary poverty rate estimates from CONEVAL and the World Bank . . . . . . . 77 Section 3: Vulnerability to poverty due to climate Annex 2. Description of panels using ENOE . . . . . . . . 83 events in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Annex 3. Demographic profiles of poverty and Maps of exposure and vulnerability to extreme chronic labor poverty . . . . . . . . . . . . . . . . . . . . . . 86 climate events in Mexico . . . . . . . . . . . . . . . . . . . . 54 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 List of Figures Figure 1. Poverty rate and Poverty gap under the US$ Figure 14: Minimum wages, females employment, 6.85 (2017 PPP) poverty line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 informal and agricultural employment, international Figure 2. Poverty, vulnerability and middle comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 class in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure 15: Evolution of indicators of social deprivation, Figure 3. Components of changes in poverty carencias sociales. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 rate by population groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 16: Changes in social deprivations Figure 4. Growth incidence curves by sources of by household income quintile, 2008-2014 household income . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 and 2012-2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Figure 5. Components of changes in poverty rate by Figure 17: Evolution of social expenditures within sources of income . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Mexico’s’ Federal Budget. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Figure 6. Labor market variables in recent years . . . . . . . . . . . 24 Figure 18. Average expenditure, relative and absolute incidence for Public Health and Education . . . . . . . . . . . . . . . . . 45 Figure 7. Changes in poverty rate by income sources, and by sex, informality and minimum wage . . . . . . . . . . . . . . . 25 Figure 19. Relative and absolute incidence of largest -by number of beneficiaries- cash transfer programs Figure 8: Two measures of chronic and transient labor in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Figure 20. Effectiveness and Efficiency of main cash Figure 9. Indexes of income inequality . . . . . . . . . . . . . . . . . . . . 31 transfer programs in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 10. Components of changes of income Figure 21. International comparison of public health inequality (as per Generalized Entropy index) by expenditures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 population groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Figure 22: Components of changes in housing flood Figure 11. Components of changes of income risk score in Mexico, 2010-2020, total and rural areas . . . . . 61 inequality (as per Gini coefficient) rate by sources of income . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 23. Monetary poverty rate projections through 2030 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Figure 12. Global long-term economic growth . . . . . . . . . . . . . . 34 Figure 24. Impact of closing one social deprivation Figure 13. Poverty and Inequality over time across gap on population with at least three deprivations. . . . . . . . . . 69 comparison countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 List of Tables List of Figures in Annexes Table 1: Poverty changes by growth and redistribution Annex Figure 1: Difference between CONEVAL and components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 World Bank welfare aggregates . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Table 2: World Bank’s Multidimensional Poverty Index, Annex Figure 2: Difference between CONEVAL and selected years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 World Bank welfare aggregate if using adult equivalence . . . 79 Table 3: Mexican population exposed to climate Annex Figure 3: Difference between CONEVAL hazards, c. 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 and World Bank welfare aggregate if using adult Table 4: Classification of municipalities by risk index, equivalence and excluding housing rent imputation . . . . . . . 80 2010 and 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Annex Figure 4: CONEVAL and World Bank poverty rates . . . . 81 Table 5: Risk, exposure and vulnerability to poverty Annex Figure 5: CONEVAL and World Bank poverty due to climate events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 lines in current Mexican pesos . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 List of Boxes List of Tables in Annexes Box 1: A brief history of Poverty and Equity Annex Table 6: Demographic characteristics of Assessments for Mexico by the World Bank . . . . . . . . . . . . . . . . 13 selected ENOE panel and comparison full cross section. . . . 84 Box 2: Female employment and poverty reduction . . . . . . . . 26 Annex Table 7: Poverty, Vulnerability and Middle Class Box 3: Rural poverty and territorial context . . . . . . . . . . . . . . . . 63 profiles, Mexico 2000 and 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Annex Table 8: Demographic profile of chronic poor, transitory poor and no poor, using panels from ENOE List of Maps 2013 and 2024 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Map 1: Percentage of population exposed to extreme climate hazards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Map 2: Population exposed to several hazards and poor population exposed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 List of Acronyms CONEVAL Consejo Nacional de Evaluación de la Política ISSSTE Instituto de Seguridad y Servicios Sociales de Desarrollo Social (National Council for the de los Trabajadores del Estado (Institute Evaluation of Social Development Policy ) of Security and Social Services for State ENADID Encuesta Nacional de la Dinámica Workers) Demográfica (National Survey of INEGI Instituto Nacional de Estadística y Geografía Demographic Dynamics) (National Institute of Statistics and ENASIC Encuesta Nacional para el Sistema de Geography) Cuidados MPP Multidimensional Poverty Measure ENIGH Encuesta Nacional de Ingresos y Gastos de PAM Programa de Adultos Mayores (Senior los Hogares (National Household Income and Citizens Program) Expenditure Survey) PEMEX Petróleos Mexicanos (Mexico National Oil ENOE Encuesta Nacional de Ocupación y Empleo Company) (National Survey of Occupation and PPP Purchasing Power Parity Employment) PROCAMPO Programa de Apoyos Directos al Campo IMSS Instituto Mexicano del Seguro Social (Direct Farm Support Program) (Mexican Institute of Social Security) OECD Organization for Economic Co-operation and INSABI Instituto de Salud para el Bienestar (Institute Development of Health for Wellness) MEXICO Poverty and Equity Assessment Team members and acknowledments This report was prepared by a core team led by Samuel throughout the preparation of this work; Gabriela Aguilar guided Freije Rodríguez (TTL, ECLPV) under the guidance of Carlos the dissemination and communication strategy. Rodriguez Castelan (Practice Manager, ELCPV), Rafael Muñoz The team is grateful for the dialogue, comments and sug- (Program Leader), and overall direction of Mark Thomas (Coun- gestions received, by the National Council for the Evaluation try Director for Mexico). Team members who contributed with of Social Development Policy (Consejo Nacional de Evaluación substantial inputs to the chapters include Mariel Cecilia Sir- de la Política de Desarrollo Social, CONEVAL) and by the Direc- avegna, Israel Osorio Rodarte, Gustavo Canavire and Minh C. ción General de Política y Proyectos de Productividad from the Nguyen. The team also received multiple inputs and comments Secretaría de Hacienda, during the preparation of this work. The from World Bank staff in Mexico City and in Washinton, DC: Cle- team is also grateful for conversations with various Mexican mente Avila, Dean A. Cira, Joanne Gaskell, Alejandro Medina academics from Universidad Iberoamericana, El Colegio de Giopp, Truman Packard, Marcela Silveyra, and Jeremy Veillard. Mexico and Centro de Estudios Espinosa-Yglesias consulted In addition, Gladys López-Acevedo, William Wiseman and Jean at the beginning of this process. Furthermore, two background Martin Brault, served as peer reviewers of the final version of studies from Mexican experts were specially commissioned for this report. Desiree Gonzalez, Cassandra Gonzalez, and Belen this report to the Centro de Investigación Económica y Presu- Muñoz Romero provided administrative and logistics support puestaria (CIEPP) and to SAR, Consultoría y Análisis de Riesgo. 9 Executive Summary MEXICO Poverty and Equity Assessment What would it take to eradicate poverty in Mexico in the tained reduction of some inequality indexes for the period un- short or medium term? This is a question that can be refined der study. A sustained reduction of inequality would contribute into a more specific target (for instance, cut in half or even to accelerate poverty reduction. The international experience eradicate extreme poverty rate), can also be precise about shows that inequality reduction has a powerful, though less time frame (say by 2030 or 2035), and calls for specific policy frequent, impact on poverty reduction. This report estimates actions (what it would take). This Mexico Poverty and Equity that if an annual GDP/head growth of 2 percent is paired with Assessment reviews the evidence about poverty and equity in a 1 percent annual reduction of the Gini, poverty would fall to Mexico over the last two decades, compares it to comparable 9.8 percent in 2030 (or a lower 4.3 percent with a 2 percent international experience, and identifies a set of critical areas of decline in the Gini). policy intervention to answer the opening question. The report Inclusive growth can be promoted by narrowing or closing aims at contributing to an open conversation in Mexico about three main gaps in economic participation and productivity. how to achieve this essential policy objective. First, female employment and economic opportunities in gen- This report postulates three main policy areas needed for eral is an area where policy actions are needed. This report doc- poverty eradication in Mexico: inclusive growth, efficient social uments the positive influence that female labor earnings have policy and infrastructure to confront vulnerability. The report played upon poverty reduction in the past: female labor earn- includes four sections, the first three of which collect evidence ings have represented as large a share of poverty reduction as about poverty, social deprivations, and vulnerability and how men’s. Still, female labor participation in Mexico is among the the evolution of these three correlates to patterns of economic lowest when compared to other countries. Important gaps re- growth, social protection policy and territorial development. The main in terms of participation rates, access to finance, titularity fourth section provides some quantitative benchmarks of what of assets, schooling, and time use all of which hinder economic it would take to eradicate extreme poverty in Mexico. opportunities for women. All these areas can be subject to pol- Faster economic growth is indispensable for poverty erad- icy actions (for instance, a substantive expansion of child and ication in the short to medium term in Mexico. A contrast with family care policies) that favor women’s participation/produc- comparable countries shows that Mexico has a much higher tivity and faster poverty reduction. poverty rate mostly due to a much slower GDP growth. Wider Another crucial gap in terms of inclusive growth refers to international experience shows that major reductions in pov- rural population, particularly those engaged in agricultural ac- erty are mostly explained by sustained periods of economic tivities. The report details that poverty rates are higher -almost growth. Economic growth in Mexico has been slow in the past double- for rural population and households whose head works two decades. This report estimates that a 2 percent GDP/head in agriculture. They also endure more chronic poverty, which growth between 2024 and 2030 would reduce poverty defined hints at a more entrenched problem to solve. These groups as population below the upper-middle income poverty line of represent almost 40 percent of the extreme poor, so poverty US$ 6.85. in 2017 PPP terms from 21.8 percent to 15.4 percent. eradication cannot happen without a substantive change in liv- A 3 percent GDP/head growth would push it down to 13.4 per- ing conditions of rural/agricultural households. Labor produc- cent. The US$ line 6.85 is higher than the Mexican “línea de po- tivity is a complex problem observed in many sectors of the breza exrema por ingresos” and hence considerably reducing Mexican economy, but poverty eradication requires a special poverty under the former is linked to reducing extreme poverty focus on policies to increase agricultural productivity in some defined by the latter. parts of the territory. This report focuses -instead- on social But faster economic growth is insufficient for poverty eradi- and physical infrastructure for these groups. This connects to cation, inclusive growth is also needed. By inclusive growth this the third gap towards more inclusive growth: increase formal report means growth that is faster among the poorer segments employment. of the population and closes existing gaps in economic partic- Poverty in Mexico is defined not only in monetary terms, ipation and productivity. Over the past two decades, poverty but also in a multidimensional manner that includes social reduction in Mexico has been associated with labor earnings deprivations. These are social deprivations that often define growing faster among poorer families which, together with the formal-vs-informal employment, so policy changes that close extension of some social programs has contributed to a sus- these carencias, as they are called in Mexico, would also re- 11 Executive Summary duce the informality gap. The official definition of extreme not only on guaranteeing access, but on securing completion of poverty includes those below the extreme monetary poverty the cycle and having learned the skills to finalize upper levels line and with at least three social deprivations. The report doc- of education. Investing in quality basic education, is investing uments the evolution of official measures of social deprivation in access to upper education. and underlines that access to social security and access to The third area of policy action for poverty eradication is re- health services are the more widespread social deprivations. A ducing vulnerability to poverty, particularly due to climate-re- hypothetical exercise that simulates the closing of any of these lated events. The concept of vulnerability used in this report two carencias shows that they would reduce extreme poverty, comes from other World Bank studies, as well as other interna- partly defined as having at least three carencias, in almost two tional institutions, that refer to vulnerability as lack of physical thirds. No other closing of a deprivation gap would have such a and social assets that reduce the propensity of loss or the abil- substantial impact upon extreme poverty reduction. ity to cope with losses due to a climate hazard event. It can be A more efficient social policy is the second policy area measured in terms of the ability to have access to sanitation needed for poverty eradication. As indicated in the previous and electricity which enables, for instance, clean water and paragraph, closing the social deprivations referring to social cooling appliances (so-called physical vulnerability), and to security and health are crucial for rapid extreme poverty re- have the ability to recover through education, social security, duction. The report documents that health services are very financial services or incomes, to confront the consequences of asymmetric in terms of funding and characteristics of service. a shock (so-called inability to cope). This is where investments Budgetary allocations to public health have not declined, but in infrastructure to address physical and social vulnerability they are below international standards, while private expen- come to play a role in poverty eradication. It can play a double ditures in health have increased. Thorough reforms of these impact. systems should aim to full coverage of the population, similar Investing in physical and social infrastructure has a double package of services across different groups and -very import- impact: it reduces extreme poverty and reduces climate-re- ant- adequate funding from taxation other than labor taxation lated vulnerability to poverty. The contrast of international to prevent disincentives to work in the formal sector. indexes of multidimensional poverty and of vulnerability to Mexico has multiple social programs that can be made more climate, included in this report, shows that Mexico has higher effective/efficient in poverty reduction. The report describes multidimensional poverty and climate vulnerability rates than the changes in size and budget to cash-transfer programs over comparison countries. This is so because of specific indexes the past two decades. Mexico has an outstanding experience in that show higher rates of lack of adequate sanitation systems, the design and implementation of social protection programs. access to finance, and coverage of social security. Beyond the Most of the existing programs are well targeted, but many have indexes -which are reductive by definition- the message is that a small impact on poverty because of either limited coverage investing in physical and social infrastructure is a policy action or insufficient benefits, while others could be more fiscally ef- that is needed for reducing vulnerability as well as for poverty ficient. A redesign of social programs towards increasing their eradication in Mexico. progressivity and sufficiency would increase the impact of This report shows estimates -using census data for 2020- these on poverty eradication. that 31.2 percent of the population in Mexico is exposed to at Social transfers, however, are not the only means for a more least one type of severe climate event, and an additional 7.7 efficient anti-poverty policy. The case of quality education mer- percent to at least two types of high-risk climate events. Only its special attention. The analysis of budgetary allocations to about one third of these exposed are in poverty, so there is an cash transfers to vulnerable children and youth shows that important share of the non-poor population who is exposed to these have increased rapidly in recent years, aiming to avoid climate-related high-risk events. Moreover, there are studies reductions in enrollment, particularly in secondary and tertia- that estimate that at least one third of the Mexican population ry. However, budget allocations to education services in gen- is vulnerable to fall into poverty because of climate-related eral -as a percentage of GDP- have not increased. Indicators of events. Therefore, public policy needs to take action not only quality in education have not improved either. It is important to to reduce poverty, but also to reduce vulnerability to poverty underline that access to secondary and tertiary levels depends because -if a shock strikes- it may reverse gains in poverty re- 12 MEXICO Poverty and Equity Assessment duction. This extends the policy focus towards vulnerable and quired transfer call for careful design. The programs of adaptive middle class -large groups of the population in an upper-middle social protection need to be “predictable, reliable and tailored.” income country like Mexico- who are exposed to climate-event This calls for innovative policy reform. hazards Mexico has outstanding experience in the design and imple- There are several components to policy action in building mentation of social policy. This experience can serve to accel- infrastructure to reduce vulnerability. This report is limited to erate and make more efficient efforts towards extreme poverty mention only a few. On the one hand, it is clear that investing eradication. This report documents the evolution of poverty, so- in sanitation, rural roads and access to financial services, par- cial deprivations, and vulnerability to poverty. It explains the ticularly in rural areas, is a policy action that would confront main forces that have driven this evolution, and advises that deprivations in terms of physical vulnerability. On the other many of these forces may not operate the same in the future hand, the development of adaptive social protection would as they did in the past. It provides the basis to argue that short address the need for a social security/social assistance sys- to medium term extreme poverty eradication requires newer tem that responds to growing climate hazards. Adaptive social policy actions in terms of inclusive growth, more efficient social protection concentrates on building resilience to hazards by policy, and investments in physical and social infrastructure to designing social protection mechanisms to cope before, during reduce vulnerability. The report indicates that short to medium and after these hits. This requires a different identification of term eradication to extreme poverty is a major, but within reach, beneficiaries (not only the poor, but the vulnerable), location/ development challenge for Mexico. The prize is worthy making type of shock (hazards that are growing), and size of the re- every effort to achieve the goal. Box 1: A brief history of Poverty and Equity Assessments for Mexico by the World Bank The Poverty and Equity Assessment includes a detailed analysis of poverty in the country, as well as an assessment of inequality and vulnerability. The Poverty and Equity Assessment is now one of the five core documents by the World Bank that shape dialogue with member countries. Other core documents relate to macroeconomic conditions for growth (Country Economic Memorandum), public finance (Public Finance Review), private sector development (Country Private Sector Diagnostic) and climate change challenges (Climate Change Development Report). Poverty Assessments have a long history within the World Bank and have been produced for multiple countries for several years. The earliest self-standing World Bank document studying poverty in Mexico is “Poverty Alleviation in Mexico”, produced in 1989. This report was a diagnostic of the consequences of the 1982-83 crisis and the ensuing structural adjustment program. It estimated poverty rates, using surveys for the years 1977 and 1982 with the minimum wage as the line. It described differences between urban and rural areas, and included special sections about the provision of health, education and nutrition. It mentioned policy options that would become prevalent in the future such as replacing generalized indirect food subsidies with direct cash transfers to the poor, as well as “targeting welfare programs through schools as additional incentive to send children to school” and creating “institutions for frequent monitoring of poverty with improved data collection.” It underlined goals that the country has now accomplished, such as increased coverage of primary education; but also detected persistent challenges such as the dualism in quality and coverage of health services between formal and informal workers, deficient rural infrastructure, and the need of fiscal revenues because “to increase welfare support, targeted social expenditure will increase which, in turn will require larger tax revenue to avoid fiscal deficit.” 13 Executive Summary The following study “Mexico: Poverty Reduction. The Unfinished Agenda” was produced in 1995. It acknowledged the progress of Mexico during the previous decade in terms of macroeconomic adjustment and the design of a comprehensive strategy to reduce poverty. But it also underlined the new challenges posed by the Peso crisis of December 1994. It estimated of poverty rates with surveys from INEGI for years 1984, 1989 and 1992, using separate poverty lines for urban and rural areas, and detailed poverty profiles by sex, schooling, geographic location and household composition. The report highlights the increased allocation of federal budget to social expenditures and encourages to sustain these with more attention to quality in education and health services, particularly in terms of secondary education and rural basic health care. On the other hand, it underlined the shortcomings of the strategy of community-based development programs (then called Programa Nacional de Solidaridad) and proposed a better design of temporary employment programs to cope with the Peso Crisis. The most recent in the series, “Poverty in Mexico: An Assessment of Trends, Conditions, and Government Strategy” was made public in 2004 (the other two were internal documents), and included additional studies on Urban Poverty, Rural Poverty, Social Protection Programs and Decentralization. It included poverty rate estimates and profiles for every two years between 1992 and 2002 using Mexican poverty lines defined by the Comite Técnico para Medición de la Pobreza, as well as World Bank’s dollar-a-day threshold. The report incorporated a comprehensive documentation of the variety of programs designed and implemented in the previous decade. It described the framework that Mexican authorities adopted to organize anti-poverty policy (a capabilities-approach called the CONTIGO strategy), and the many innovative programs included in it (for instance, Progresa-Oportunidades, Micro-regiones, Habitat, Seguro Popular) all with an intense result-based monitoring approach and under new legislation to advance social policy: the 2003-enacted, Ley de Desarrollo Social. This poverty assessment is the first to document the distributive incidence of different programs in social policy, underlining the progressivity of some of them (i.e, cash-transfers) and the regressivity of others (e.g., contributory pensions), as well the possibility of increased efficiency in poverty reduction through redesign or redistribution of resources across programs. The report calls for policy action in areas that have already been matter of concern: (i) quality of secondary education as a gateway to expanded tertiary education, (ii) fragmentation in coverage and quality of health services, and (iii) additional fiscal resources to sustain and expand quality and coverage of social policy. But it also underlines new challenges: the assessment warns of risks to poverty reduction because of meager economic growth, limited quality in public services and increased vulnerability to risks. Although not exactly a poverty assessment, a recent comprehensive review of poverty and inequality in Mexico is included in the “Mexico. Systematic Country Diagnostic” (SCD), published in 2018. The diagnostic includes an assessment of conditions for growth, inclusion, and sustainability in the country. The section on inclusion -using the Mexican official multidimensional measures for the first time in World Bank documents- underlines that Mexico had experienced advances in the reduction of social deprivations (access to health, social security, education, housing and food security) but little progress in monetary poverty. The diagnostic about poverty, highlights four main problems: (i) limited competition in some markets which leads to higher prices affecting real incomes of the poor, (ii) limited financial access which affects economic performance of the poor 14 MEXICO Poverty and Equity Assessment and small firms; (iii) a segmented social security system, which leads to unequal health services and disincentives to formal employment and (iv) a fiscal system with limited overall redistributive capacity. Over the years, World Bank poverty assessments have described trends of poverty and inequality, explained its root causes, promoted policy solutions, and assessed policy actions. These reports have attested the evolution of anti-poverty policy in Mexico across several crises, and over different strategies: from indirect subsidies to direct transfers, from community-based investment programs to conditional cash-transfers. They initially relied on limited data, and now use the admirable advances in data collection and poverty measurement by Mexican institutions. They started by responding to country-specific crisis situations. Now they serve as regular assessment of standards of living in Mexico, its innovations in anti-poverty policy action, and its contributions to global poverty eradication. 15 Section 1: Monetary Poverty and Inequality in Mexico Section 1: Monetary Poverty and Inequality in Mexico 16 MEXICO Poverty and Equity Assessment This Section 1 concentrates on monetary measures of pover- tween 2006 and 2008, and between 2014 and 2016.1 Therefore, ty and inequality. Section 2 will discuss non-monetary measures the analysis will be divided into there six-year periods: from 2000 of poverty. Both will make use of Encuesta Nacional de Ingresos to 2006, from 2008 to 2014 and from 2016 to 2022. Projections y Gastos de los Hogares, ENIGH, for the period 2000-2022. Over of monetary poverty for 2024 through 2030 are included in Sec- this period, technical aspects of the survey and methodological tion 4. Section 3, instead, will concentrate on vulnerability to pov- changes involve non comparability of the data series twice: be- erty due to climate hazards as of 2020 census data. Monetary poverty Over the three six-year periods under study, spanning more still below an income level to separate them from the rich (Fig- than a decade, a clear pattern can be observed in poverty rates ure 2). In this report, the vulnerability to poverty is defined as under US$ 6.85 (2017 PPP). The period 2000-2006 as well as population above the US$ 6.85 a day, but below US$ 14 a day; the period 2016-2022 experienced a rapid decline in pover- while the middle class are those above US$ 14 a day, but below ty rates, whereas the period 2008-2014 the poverty rate did US$ 81 a day.4 not undergo a statistically significant decline.2 The initial and It is important to underline that, despite these trends in final periods had similar falls in poverty rates: 10.2 and 9.6 terms of incidence of poverty, vulnerability and middle class, percentage points, respectively. The first period, experienced a the demographic profiles of these groups have not changed re- steady decline in poverty rates, the second included the 2009 markably over time. Poverty incidence has declined, between international financial crisis from which the economy took long 2000 and 2022, for every population group defined in terms of time to recuperate, and the third suffered a reversal due to the area of residence, sex, age, schooling, type of household and COVID-19 pandemic in 2020 and experienced a sudden decline employment of household head.5 At the same time, the prev- thereafter. Despite these ups and down, the macro-economy alence of vulnerability and middle class have increased for has experienced a quite stable and low 10-year average growth almost every group as well.6 Moreover, the rankings of these since the early 2000s, with very low productivity growth and no poverty, vulnerability and middle class rates within members convergence towards high-income economies.3 of each group in a given category have remained almost the If looking at the poverty gap or the poverty severity, instead, same.7 there is a slow but regular decline in all periods, although the Poverty rates are higher for rural population, children, less second period reduction is slower than the other two. This schooled people, families with more children, and households means that although the share of the population below the whose head works in agriculture.8 The same can be said about “upper-middle” income poverty line did not always decline, the the rankings in terms of incidence of vulnerability to poverty, depth of poverty and its severity did so regularly over the three and the inverse in the case of middle class status. In other periods under study (Figure 1). In other words, the distance words, the characteristics of those more likely to be poor or between the incomes of the poor and the poverty line declined vulnerable remain the same over the period. With one excep- in all three periods. tion: population aged 60 and more used to be near the top of the This decline in the poverty rate is accompanied by an in- poverty and vulnerability incidence ranking. Not anymore; and crease in the share of the population vulnerable to poverty and instead, this group is now at the top of middle class incidence the share of population that can be defined as middle class. by age group. This evolution of a declining poverty and increasing middle Changes in demographics and the evolution of these class is compatible with trends observed for the rest of the re- group-specific rates have affected the relative composition gion. Vulnerability to poverty has been defined in the academic of each socio-economic group. Demographic transition (more literature as the population with a probability to fall back into people in middle age), advances in schooling (more people poverty, within a five year period, higher than 10 percent. The with secondary and higher education) and structural transfor- middle class definition has often used different benchmarks, mation of the economy (fewer people working in agriculture) but is usually defined as those not vulnerable to poverty but have changed the composition of these socio economic groups. 17 Section 1: Monetary Poverty and Inequality in Mexico Particularly the poor. Families with children, and children them- but 40 percent in 2022. Households whose head works in ag- selves, represented around 40 percent of the poor in 2000, but riculture were nearly 40 percent of the poor in 2000, but just now they represent around 30 percent. People with less than above 30 percent in 2022. primary schooling represented 60 percent of the poor in 2000, Figure 1. Poverty rate and Poverty gap under the US$ 6.85 (2017 PPP) poverty line Poverty rate Poverty Gap and Poverty Severity Indexes 55% 0,25 50% 0,20 Percentage of total population 45% 0,15 40% 35% 0,10 30% 0,05 25% 0,00 20% 2000 2004 2020 2002 2008 2005 2006 2010 2014 2012 2018 2016 2000 2004 2002 2020 2008 2006 2022 2010 2014 2012 2018 2016 poverty gap poverty severity Source: Own calculations using ENIGH 2000 through 2022. Note: Poverty rate, poverty gap and poverty severity computed with command ipov from DASP, version 3.03 Figure 2. Poverty, vulnerability and middle class in Mexico 55% 50% 45% percenetage of total population 40% 35% 30% 25% 20% 15% 10% 5% 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 poverty rate vulnerability rate middle class Source: Own calculations using ENIGH 2000 through 2022. Note: Population proportions computed with command proportion from STATA 18 18 MEXICO Poverty and Equity Assessment Changes in poverty as explained by changes in couples) also leads to poverty reduction in the three periods population groups. (Figure 3, bottom left). Accumulation of human capital through additional years Changes in poverty rates over time can be explained by of schooling for a growing share of the population, also brings changes in the poverty rate of specific groups, or changes in positive poverty reduction effects. Population with some sec- the relative population size of each group. Intuitively, a fall in ondary education or more, went from 35.7 percent in 2000 to the total poverty rate can be due to a reduction in the poverty 57.0 percent in 2022. This involves a larger share of the popu- rate of one group, or a rise in the relative size of a group charac- lation is potentially more productive. Figure 3, top right, shows terized by a lower poverty rate, or a combination of these two that the growing share of more schooled people explains nearly forces. This decomposition illustrates how much changes in 2 percentage points of poverty reduction in all three periods poverty are due to changes in poverty incidence across groups, under consideration. or changes due to people moving from one group to another.9 In contrast, the poverty reduction effect of the distribution The report explores demographic changes (that is, changes in of employment by economic activity, fluctuates over the three the relative size of age groups), schooling changes (through the periods considered. It had a positive effect in the 2000-2006 accumulation of years of education by the population), socio- and 2016-2022 periods, but no effect in the 2008-2014 period logical changes (as the per prevalence of different household (Figure 3, bottom right). This is because the path of structur- compositions) and structural transformation (that is changes al transformation (i.e., growth of the secondary and tertiary in the proportion of employment in agriculture). sectors of employment) varied over these years. The share The aging of the Mexican population has led to an increase of people living in families whose household head works in in the share of working age adults. This implies that more peo- agriculture declined from 2000 to 2006 (from about 18.5 to ple are potentially able to find jobs, accumulate more years of about 14.5 percent), leveled off between 2008 and 2014, and schooling and work experience and families have more adults declined again between 2016 and 2022 (from about 14.5 to to sustain themselves. Figure 3, top left, shows that the change about 12.5 percent). Something similar happened to the share in the relative size of age groups has contributed to a growing of household heads who were jobless (i.e., inactive or unem- percentage of poverty reduction, in all three periods: about half ployed): it was higher in the middle period than in the other two. a percentage point in 2000-2006 and 2008-2014, and almost This underlines the importance of employment and sources of a full percentage point in 2016-2022. Additionally, a recom- family income in explaining poverty reduction. position of family structures, with a growing share of smaller families (singles, single couples and less than three children 19 Section 1: Monetary Poverty and Inequality in Mexico Figure 3. Components of changes in poverty rate by population groups By age group By schooling group 2 4 chnages in poverty rate (percentage points) chnages in poverty rate (percentage points) 2 0 0 -2 -2 -4 -4 -6 -6 -8 -8 -10 -10 -12 -12 2000-2006 2008-2014 2016-2022 2000-2006 2008-2014 2016-2022 By family type group By employment of household head group 2 2 chnages in poverty rate (percentage points) 0 chnages in poverty rate (percentage points) 0 -2 -2 -4 -4 -6 -6 -8 -8 -10 -10 -12 -12 2000-2006 2008-2014 2016-2022 2000-2006 2008-2014 2016-2022 within group between group Source: Own calculations using ENIGH 2000 through 2022 Note: Poverty change decompositions à la Ravallion and Huppi (1991) computed with command dfgtg2d from DASP, version 3.03 on STATA 18.0 . The age groups are four categories defined as ages 0-14, 15 to 29, 30 to 44, 45 to 59 and 60 or more. The schooling groups are seven categories defined in terms of no formal education, incomplete or complete primary, secondary and tertiary levels. The family type groups are seven categories defined as singles, singles with children, couples without children, couples with less or more than three children, and multi-generational. The household head categories are jobless, agriculture, industry, construction and services. Changes in poverty as explained by changes in tem), cash transfers (public through social assistance or pri- sources of family income. vate through remittances) and other sources of income (which includes housing rents and returns from financial assets). Given that monetary poverty in Mexico is measured using Growth in any income source contributes to total household household income as welfare aggregate, changes in pover- income, but growth in some components is more important for ty are necessarily associated with changes in the different some households than others (say public transfers for poorer sources of income that households receive. Four main sourc- families). es of household income can be stipulated: labor earnings (i.e., Labor incomes have had a pro-poor growth pattern in all pe- incomes associated to wage work by employed household riods. In other words, the annual growth rate of labor incomes members), contributory pensions (that is, pensions received for those living in low-income households has been higher than by members who were contributors to the social security sys- for those in higher income households (Figure 4, top left pan- 20 MEXICO Poverty and Equity Assessment el). The growth rate was faster for all groups in the first and percentile of the income distribution in the period 2000-2006. third period of analysis than in the second period. Contributory But in the period 2008-2014 all people, except those in the top pensions have an irregular pattern across deciles, but on aver- decile, experienced a decline in remittances. Finally, the period age they have grown faster in the most recent period: around 2016-2022 sees increases in private transfers for all groups 2.7 percent in 2000-2006, 2.3 percent in 2018-2014 but 4.7 and higher annual growth rates among those at the bottom of percent in 2016-2022 (Figure 4 , top right panel). the distribution (Figure 4, bottom right panel).10 This indicates Cash transfers, either public or private, have had very di- that remittances played a role in poverty reduction in the period verse growth patterns over time. Private cash transfers (mostly 2016-2022, but not so in previous periods. remittances) increased for people living between the 4 and 8th Figure 4. Growth incidence curves by sources of household income Labor Incomes Contributory Pensions 7 20 6 5 15 Annual Growth (%) Annual Growth (%) 4 3 10 2 1 5 0 -1 0 -2 -3 -5 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Income Deciles (by initial period household per capita income) Income Deciles (by initial period household per capita income) Public Transfers (social assistance) Private transfers (remittances) 60 14 12 50 10 Annual Growth (%) Annual Growth (%) 40 8 6 30 4 20 2 0 10 -2 -4 0 -6 -10 -8 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Income Deciles (by initial period household per capita income) Income Deciles (by initial period household per capita income) 2000-2006 2008-2014 2016-2022 Source: Own calculations using ENIGH 2000 through 2022 Note: Public transfers refer to cash benefits through social assistance programs, including non-contributory pensions Public transfers had a significant growth in the first period. beneficiaries expanded from covering fewer than 10,000 lo- Perhaps due to the extension of conditional cash transfer pro- calities and about 2.6 million households in 1998 to more than grams (such as Progresa-Oportunidades), whose number of 50,000 localities and about 5.2 million families by 2009. This 21 Section 1: Monetary Poverty and Inequality in Mexico high growth is observed across all deciles of the distribution in the first two periods, but a stagnation in the third because of (Figure 4, bottom left panel) because other programs benefit- the pandemic in 2020. Average labor earnings per family work- ted households in the middle and top of the distribution.11 In er (orange bars) had a positive effect on poverty reduction in the period 2008-2014 all deciles, except the top, experienced 2000-2006, and an even larger effect in 2016-2022; whereas it continued growth in public transfers, but at a slower pace than fully reversed gains from employment creation in 2008-2014. in the previous period. The period 2016-2022, however, sees This is consistent with falling real wages in the intermediate a very different growth pattern: annual growth is negative for period and growing real wages in the final period.13 These three those in the first decile, negligible in the second decile, and components, namely labor earnings as per either more employ- becomes positive in the third decile with higher growth rates ment or more wages per employed person, represent the main for those in the 7th decile than the rest of the population. This is force that explains the direction and size of poverty reduction associated with a change in social assistance programs, pivot- in Mexico for all periods. ing from means-tested conditional cash transfers in the past Non-labor incomes have a smaller, but usually positive ef- to non-conditional programs more recently (of which the ex- fect on poverty reduction. Public transfers add up to nearly one pansion of non-contributory pensions is the most important), percentage point of poverty reduction in the first two periods, as well as an increase in the budget to scholarships to upper but less so in the third period (grey bar). Remittances had a secondary students. A more extensive discussion of this pro- negative impact on poverty reduction in the second period, be- cess is included in Section 2 of the report. cause this source of income retrenched in those years (dark The previous data show sources of income growth for all green bar).14 Contributory pensions slowly increased its effect population groups, but poverty changes are explained by (light yellow bar), which is consistent with a growing percent- changes in the bottom of the distribution. Focusing only on age of above-60 population and growth in the number of bene- households at the bottom of the income distribution, changes ficiaries of the largest social security program, IMSS (more on in poverty rates can be decomposed into changes by various this in Section 2 of this report). sources of income. The calculation is done assuming that be- The period 2016-2022 is affected by the COVID-19 pandem- tween two periods of study, only one source of income chang- ic of 2020. The economic contraction produced by the pan- es and all the others remain constant. The decomposition of demic reversed the slight reduction in poverty rates observed poverty changes into income sources is produced by repeating between 2016 and 2018. It is the rebound of economic activi- this hypothetical exercise for every source of income, and av- ty in 2021 and 2022, as well as policy and budgetary actions eraging total household income across the population.12 This adopted since, that explains most of the change in poverty report explores six components of household income: adult rates for the period 2016 -2022. GDP contracted 8.7 percent in household members (a sort of inverse of the dependency 2020 -one of the largest declines ever recorded in Mexico- but rate), employment of adult members, labor incomes, contribu- then grew 5.7, 3.9 and 3.2 percent in the three following years, tory pensions, public social assistance, remittances, and other above what was initially projected.15 The rightmost bar of Figure sources of income. 5 shows that almost all the changes observed for the period The three periods under study show major differences in 2016-2022 are mirrored in the period 2020-2022. the sources of poverty reduction. The share of working age Given the predominance of labor earnings over poverty re- adults among low-income families (light blue bars in Figure 5) duction, it is worth exploring labor market trends in Mexico. Over reduced poverty in all periods. This is consistent with the pre- the period under study, there are three important trends (Figure vious message of a demographic bonus for poverty reduction. 6). First, there is a slow decline in total labor force participation The share of employed adults (dark blue bars in Figure 5) re- characterized by a rise in female participation and a decline in duced poverty in the two first periods but had a null effect in the male participation (both about five percentage points change third. This is consistent with growing waged employment rates between 2005 and 2023). Second, informal employment (de- 22 MEXICO Poverty and Equity Assessment fined as a percentage of employment without access to health between 2006 and 2016, but then rose to 63 percent in 2022. services) declined slightly since 2010, after remaining quite This is very likely the consequence of a faster rise in minimum stable for more than a decade. Although informal employment wages than in the average wages which implies an increase in is still the most prevalent, the five percentage point decline in the ratio of minimum wages to median wages of full-time work- the rate of informal employment is important because, first, ers (also known as the Kaitz index). In contrast, other aspects it is not a common trend among most upper-middle income of labor market composition such as distribution by weekly countries and, second, it suggests improving labor market hours of work, occupation, economic activity or employment conditions.16 More noticeably, third, the share of employment position have remained quite stable over time. earning up to 2 minimum wages, went from 33 to 40 percent Figure 5. Components of changes in poverty rate by sources of income 6 4 2 change in poverty rate (in percentage points) 0 -2 -4 -6 -8 -10 -12 2000-2006 2008-2014 2016-2022 2020-2022 share of adults employment share labor income contributory pensions public transfers (social assistance) private transfers (remittances) other incomes Source: Own calculations using ENIGH 2000 through 2022 Note: Poverty change decompositions à la Paes de Barros et al. (2006) computed with STATA command ADECOMP by Azevedo, João P., Minh Nguyen and Viviane Sanfelice (2012). “ADECOMP: Stata module to estimate Shapley Decomposition by Components of a Welfare Measure” Statistical Software Components S457562, Boston College Department of Economics. 23 Section 1: Monetary Poverty and Inequality in Mexico Figure 6. Labor market variables in recent years Labor Force Participation Informal Employment 85% 65% 66% 64% 80% 60% ( as percentage of working age males ) ( as percentage of working age females ) 62% Male labor force participation Female labor force participation Percentage of total employment 75% 55% 60% 58% 70% 50% 56% 65% 45% 54% 52% 60% 40% 50% 55% 35% 48% 2020 2020 2008 2009 2008 2009 2005 2006 2005 2006 2022 2022 2023 2023 2007 2007 2010 2010 2014 2014 2012 2021 2012 2021 2018 2019 2018 2019 2013 2015 2016 2013 2015 2016 2017 2017 2011 2011 Male (left axis) Female (right axis) Without access to health services Informality Minimum Wages 80% 70 70% 60 of median wage of full-time employees percentage of total employment Minmium wage as percentage 60% 50 50% 40 40% 30 30% 20 20% 10% 10 0% 0 2020 2008 2009 2005 2006 2022 2023 2007 2010 2014 2012 2021 2018 2019 2013 2015 2016 2017 2011 Proportion of employment erning up to 1 minimum wages Proportion of employment earnings between 1 and 2 minimum wages Minimum relative to median wages of full-time workers (right axis) Source: Data from INEGI, Encuesta Nacional de Ocupación y Empleo (https://www.inegi.org.mx/programas/enoe/15ymas/#tabulados on 07/15/20240). Data on ratio of minimum wage to median wage from OECD Data Explorer (https://data-explorer.oecd.org/ on 07/15/2024 ) Note: “Informality” as per INEGI definition: “Proportion of the employed population that includes the sum, without duplication, of those employed who are occupationally vulnerable due to the nature of the economic unit for which they work, with those whose labor link or dependence is not recognized by their source of work.”. Workers without access to health services as per INEGI definition: “Situation that distinguishes the employed population, depending on whether or not they have medical care in public or private institutions, derived from their main job.” Female employment, informal employment and minimum erty reduction than informal employment in 2000-2006 (2.5 wages have influenced changes in poverty rates. A decompo- and 3.3 percentage points, respectively), and in 2016-2022 sition of sources of poverty change by male and female labor (2.0 and 2.6 percentage points, respectively). For the period earnings shows that the combination of employment creation 2008-2014, formal earnings showed little change, whereas in- and higher wages by women contributes as much as men in formal earnings declined noticeably. (green and yellow bars in the period 2000-2006. In the next period, women’s earnings Figure 7, right panel). This underlines the importance of formal declined much less than men’s, and in 2016-2022 women’s employment as a more steady -less erratic - source of poverty contribution to poverty reduction is again as large as men’s (red reduction. However, informal employment is still ubiquitous versus blue bars in Figure 7, left panel). Labor earnings associ- and drives most of poverty reduction because it is the main ated with formal employment represent a smaller share of pov- source of employment among the poor. 24 MEXICO Poverty and Equity Assessment The most noticeable change in poverty reduction compo- to sustaining above the minimum wage for up to two family nents is due to the influence of minimum wages. The increase members and thus growing wages of this group -and given that of the share of workers below 2 minimum wages, or the in- there has been no employment reduction so far- helped reduce crease of their average wages had a regular influence on pov- poverty by a larger margin in the third period of analysis.17 erty reduction in the periods 2000-2006 and 2008-2014. In This growing influence upon poverty reduction by female contrast, this effect increased in size for the period 2016-2022 employment, informal employment and minimum wages may (Figure 7, bottom panel). This is due to the position of the min- also involve changes in income distribution. These are groups imum wage vis-a-vis the poverty line. The minimum wage, in that tend to be at the bottom of the income distribution so their 2017 PPP terms, was below the poverty line of US$ 6.85 per improvement may bring about poverty reduction as well as day until 2012, but it doubled since then to be around 15.9 US$ some reduction of inequality. Actually, a partition of poverty per day by 2022. When in the early 2000s more people earned changes into growth and redistribution components shows that less than two minimum wages it meant this had some contribu- the latter component has contributed to poverty reduction in all tion to poverty reduction in terms of employment and earnings. three periods, but its impact has been larger in the period 2016- By 2022, earning up to two minimum wages workers are closer 2022 than in the previous two periods (Table 1).18 Figure 7. Changes in poverty rate by income sources, and by sex, informality and minimum wage 4 20 change in poverty rate (in percentage points) change in poverty rate (in percentage points) 2 15 0 10 5 -2 0 -4 -5 -6 -10 -8 -15 -10 -20 -12 -25 2000-2006 2008-2014 2016-2022 2020-2022 2000-2006 2008-2014 2016-2022 2020-2022 men labor income men share work women labor income Above 2 MW labor income Above 2 MW share work Below 2 MW labor income women share work Everything that is not labor share of adults Below 2 MW share work Everything that is not labor share of adults 4 change in poverty rate (in percentage points) 2 0 -2 -4 -6 -8 -10 -12 2000-2006 2008-2014 2016-2022 2020-2022 formal labor income formal share work informal labor income informal share work Everything that is not labor share of adults Source: Own calculations using ENIGH 2000 through 2022 Note: Poverty change decompositions à la Paes de Barros et al. (2006) computed with command adecomp on STATA 18.0 25 Section 1: Monetary Poverty and Inequality in Mexico Table 1: Poverty changes by growth and redistribution components. 2000-2006 2008-2014 2016-2022 2020-2022 percentage percentage percentage percentage % % % % points points points points Income growth -6.26 61 2.07 2131 -5.31 55 -6.85 81 Reistribution -3.96 39 -1.97 -2031 -4.34 45 -1.65 19 Total change in poverty -10.22 100 0.10 100 -9.65 100 -8.50 100 Source: Own calculations using ENIGH 2000 and 2022. Note: Decomposition based on method by Datt and Ravallion (1992), using STATA command drdecomp Box 2: Female employment and poverty reduction Women play a very important role in defining the standards of living of families. Over the past two decades, women headship of households has experienced a sizeable increase in Mexico. As per ENIGH data, in year 2000, only 14.9 percent of the population lived in a household led by a woman but, by 2022, 29.1 percent did. Single-women headed families capture 3.5 percent of the population in 2022 (up from the 1.7 in 2000) and although a small share, this group faces severe poverty conditions. The poverty rate of this group (as per US$ 6.85 line) was 18.4 percent in 2022, whereas for single-men headed households the poverty rate was 8.7 percent. In contrast, non- single women heads have a lower poverty rate than non-single men heads: 20.6 and 22.8 percent, respectively, for the same year. This growing role of women has to do with demographic and social changes, but also with increasing labor force participation and earnings in the labor market. The female labor force participation rate has grown from 41.6 percent in 2005 to 46.5 in 2023, and the average hourly wages of women has grown as fast as for men between 2005 and 2023.19 This has translated into a key influence of women labor earnings on poverty changes. As illustrated before in Figure 7, left panel, employment and earnings from females explain as much as males in the poverty reduction experienced in periods 2000-2006 and 2016-2022 (about a fourth of total poverty rate reduction in each period accrues to women’s labor earnings and employment). In the poverty rate stagnation period 2008- 2014, changes in female labor earnings did not contribute to poverty increase whereas male labor earnings did so. Like formal employment, female earnings are a more stable source of poverty reduction. Despite these positive trends, female condition as a contributor to family earnings and poverty reduction are still limited and has potential to grow further. Female labor force participation in Mexico is the second lowest among OECD countries (only surpassed by Türkiye) and also among the lowest in Latin American countries (surpassed by Guatemala and Honduras). The average gender wage gap is also high: 16.7 percent for Mexico, whereas the average for the OECD is 11.7 (although a few countries in this group have wider gaps, like Korea and Japan). When compared to Latin 26 MEXICO Poverty and Equity Assessment American countries, Mexico has a wider gender wage gap than most countries (except Peru).20 In terms of access to education, a key productive asset to engage productively in labor markets, women have a lower enrollment rate in upper secondary and tertiary education than comparison countries. For instance, tertiary education gross enrollment rate of 48.5 percent (as of 2021) whereas the region average is 65.7 percent and the average for upper-middle income countries in the world is 68.8 percent. Net enrollment rates in upper secondary in Argentina, Brazil, Chile, Colombia and Peru are above the 90 percent, whereas in Mexico is slightly below 70 percent.21 There are no important gender gaps in basic education, which translates into generational differences in the educational gap (one of the social deprivations in the official multidimensional measure of poverty): women aged 50 and more have lower schooling attainment than men, whereas the gap is much narrower or null for younger cohorts. Women in Mexico also experience gaps with respect to other important productive assets such as access to financial services and housing property. The Encuesta Nacional de Inclusion Financiera, ENIF 2021, indicates that only 61.9 percent of women make use of at least one financial instrument, lower than 74.3 percent of men, and lower than the average for women in the region (68 percent), and women in upper-middle income countries (81.7 percent).22 Access to the titularity of the dwelling also shows a gender gap, but with a generational difference. The gender gap in titularity for people aged 15 to 44 is about 5 percentage points, but people aged 45 and more it ranges from 15 to 20 percentage points. The use of time by women, another key asset for participating in economic activities, is biased by social norms towards family care rather than to employment. For the period 2016-2022, women in poor households allocate about 27 hours per week to care other family members, whereas men only about 14 hours. This is similar to the allocation in non-poor households (27 hours against 16 hours, respectively) and to the gap seen in previous periods (same rates for both groups in 2018, and slightly lower for both sexes but with a similar gap in 2008). The difference is especially acute for women in reproductive age (14 to 44 years) and has special incidence the decision of young women about further schooling or employment. According to the Encuesta Nacional de la Dinámica Demográfica (ENADID), in 2023, 17.1 percent of the respondents (aged 15 to 34) indicated marriage or childbearing as the 4th most common reason for dropping out of school, below other reasons like lack of resources or uninterest in continuing study. Although lower than in the previous ENADID survey of 2018 (when it was 20 percent of respondents), family care responsibilities still indicate an important barrier to women getting schooling and employment opportunities.23 In terms of regulations, the most recent World Bank global report on Women Business and the Law ranks Mexico 49th out of 190 countries in terms of laws that affect women’s economic opportunities. It is surpassed by most OECD countries and a few regional countries (like Costa Rica and Peru).24 Interestingly, Mexico achieves the top score in 5 of the 8 indicators of the index, falling short in terms of the dimension of Pay (laws affecting occupational segregation and gender wage gap), Parenthood (laws affecting women’s work during and after pregnancy) and Pensions (laws affecting the size of women’s pensions). These are three areas where policy reform can close gender gaps, boost women’s employment and enhance the impact of women on reducing poverty in the country. 27 Section 1: Monetary Poverty and Inequality in Mexico What are the determinants of women’s earnings generation potential? A recent World Bank study on female labor force in Mexico lists six factors that either hinder or enhance female labor force participation: labor demand, skills, access to productive assets, childcare, legislation and social norms. Gender gaps in several of these aspects have been described in the previous paragraphs, so closing these gaps is a key policy objective to promote female employment and enhance its potential for poverty reduction. But childcare merits special elaboration. The study concentrates on childcare as “one of the main barriers, or perhaps the most important one” to promote women’s gainful participation in labor markets.25 It is the cause most often mentioned as an explanation of why not to work (above low wages, lack of transportation or family commitments), by high- and low-skill qualifications women. It cites the existence of multiple academic studies, for Mexico and for other countries, that confirm the positive impact of childcare services on female labor force participation.26 It also explains that the supply of these services in Mexico is limited (half of the municipalities do not have child care centers), fragmented (multiple types of services and limited public funding for these), and heterogenous (due to wide differences in quality of service and incomplete regulations about basic standards of service). Moreover, from the demand side, childcare services are mostly acquired by households with higher incomes and in some states (entidades federativas) more than others, which creates significant inequality in access to this service. Lack of confidence in the quality of the service, as well as the level of schooling of the mothers, are also very important explanatory factors of the demand for childcare services. This combination of demand and supply factors leaves the majority of children and mothers, particularly among vulnerable and poor population, without use of childcare services and weakens the potential for women to work and additional resources for the households. Consequently, the study recommends expanding the use of childcare centers through policy actions in the areas of regulation (e.g., defining minimum quality standards for the service, and recognize childcare as a right) and access (such as subsidies to families or communities, public provision, private provision by employer’s firms and/or extension of school time).27 Chronic Poverty and Income Mobility from labor incomes The measures of monetary poverty described so far are is entrenched, more difficult to alleviate, and thus have been static. Poverty rates refer to whether household members per poor for quite some time. Or, instead, it is a group that entered capita income is above or below a given threshold (and by how poverty recently and the declining trend in poverty is due to much, in the case of the poverty gap). These measures do not the share of people falling into poverty being smaller than the indicate whether these household members have experienced share that has been scaping poverty so that the net stock of poverty for a single or several periods. These measures do not population in poverty has declined over time. In other words, explain whether, for instance, the reduced number of poor in the question is whether the poor in a given period are mostly more recent periods are fundamentally people whose poverty chronic poor or transient poor. 28 MEXICO Poverty and Equity Assessment The measurement of chronic and transient poverty requires higher for rural population (almost twice as high than urban panel data, that is repeated surveyed observations of the same population), for women, for younger and less schooled people, households over time. Panel data are not often available, and for those living in households whose head is unemployed, or existing panel data are of variable structure in terms of time and working in the informal sector or in agriculture. The opposite is population coverage. Moreover, there are different methodolog- true for the never poor. However, the distribution of the chronic ical approaches to the measurement and analyses of chronic poor across groups depends more on their relative size, than poverty. In this assessment we make use of the rotating struc- the group-specific poverty incidence: most of the chronic poor ture of the Encuesta Nacional de Ocupación y Empleo, which live in urban areas, have less than secondary education, and interviews the same dwelling for up to 5 consecutive quarters. work in the informal sector (as is the distribution of above-15 Being a labor survey, it only includes labor conditions, and population in the panel survey). The only case that merits spe- hence we measure chronic and transient labor poverty; that is cial attention is people living in households whose head works the percentage of the population whose household labor earn- in agriculture, for whom the proportion among the chronic poor ings per capita are below the upper-middle income poverty is the highest (more about this, and the need of special atten- line of US$ 6.85 (in 2017 PPP). This report includes trends and tion to agricultural activities and rural areas, in Box 3 in section profiles of chronic labor poverty for the period 2006-2023.28 3). The proportion of people that stay in poverty every quarter Counting the number of periods a person earns less than for a full year has declined recently. If we call “chronic poor” the the poverty line is not the only way to measure chronic poverty. observations whose labor income remains below the poverty If households are able to save and smooth out consumption line for five consecutive quarters (the maximum duration of the across periods, being below the poverty line in a given period available panels), this group reached a peak of 29.3 percent by does not necessarily involve that the household is in poverty the fourth quarter of 2019, and has since declined to 17.9 per- the next: because it can use saved earnings in one period to cent in the fourth quarter of 2023 (Figure 8, left panel). On the keep consumption up in another. Chronic poverty can then be other hand, the percentage of observations that have not been defined, instead, as average income across periods compared in poverty any of the five quarters has also increased from 22.0 to the poverty line.31 Using this methodology, chronic pover- to 34.9 percent in the same period. In the middle, there is the ty is a larger component of total poverty in Mexico, but it also group of observations that have been between 1 and 4 periods shows a remarkable rapid decline in recent years (Figure 8, under the poverty line. If we call this group of observations the right panel). This decline involves that workers are less likely “transient poor”, their share has moved little over time and has to be without work in a given period and/or that losses in a given remained around 47 percent for most years of the period under period deviate less from the poverty line and can be partly com- study.29 This indicates that a still large share of the population pensated with incomes from other periods. This is consistent faces poverty at least one quarter within a given year. with the decline of unemployment and informal employment in The characteristics of the chronically poor, and the never recent years, as well as the increasing number of workers earn- poor -in labor earnings terms- have not changed over time. For ing between 1 and 2 minimum wages and this minimum wage two selected years, (panels ending in first quarter of 2014 and being closer the poverty line, described in the previous section. 2023, as examples) the profiles of chronic poverty incidence have been remarkably stable.30 Chronic labor poverty rates are 29 Section 1: Monetary Poverty and Inequality in Mexico Figure 8: Two measures of chronic and transient labor poverty 100 60% 17,9 17,8 18,2 19,6 20,1 22,6 23,3 23,4 24,4 24,8 24,1 25,0 25,7 27,3 28,4 percentage of population in panel 28,8 29,3 80 50% 40% 60 30% 40 20% 20 41,0 39,3 37,3 34,9 33,6 10% 30,5 29,0 28,2 27,5 27,2 27,3 27,1 25,8 24,4 22,6 22,5 22,0 0 0% 2020 2008 2009 2006 2022 2023 2007 2010 2020 2014 2008 2009 2006 2012 2021 2022 2018 2019 2023 2013 2015 2016 2007 2017 2010 2014 2011 2012 2021 2018 2019 2013 2015 2016 2017 2011 never poor one quarter poor two quarters poor chronic poverty transient poverty three quarters poor four quarters poor five quarters poor Source: Own calculations using quarterly ENOE from 1st quarter 2005 to 1st quarter 2023. Note: Panels constructed using rotation panels of ENOE which interviews same dwelling for up to 5 consecutive quarters. The structure of data collection was interrupted and modified during three quarters of 2020 due to the COVID-19 pandemic. For a description of the panels and comparability to full ENOE surveys see Annex 2. Description of panels using ENOE. Labor poverty refers to percentage of the population aged 15 to 64 whose labor earnings per household member are below the US$ 6.85 per day (2017 PPP). Chronic poverty in the panel to the right decompositions à la Jalan and Ravallion (1998) using STATA command dtcpov from DASP, version 3.03. Inequality of Incomes Income inequality shows a similar trend to poverty rates. and bottom left panels, show that population shifts or changes As measured by the Gini index, it declined between 2000 and in average incomes across groups represent small slivers of 2006, and declined again between 2016 and 2022. Point esti- change in inequality. In other words, most inequality change mates also show a slight decline between 2008 and 2014, but has been experienced within each of these groups. this is not statistically significant, so a stable trend seems more The evolution of schooling, however, has a larger influ- plausible for this period (Figure 9, left panel). Other measures ence upon inequality changes. On the one hand, an increase of income inequality show similar trends. For instance, the 90th of schooled people has a disequalizing effect because there to 50th percentile ratio of the income distribution (a measure is a growing share of people with higher incomes.33 On the of income differences in the upper half of the population), other hand, narrower income differences between schooling shows the same trends as the Gini. In the early 2000s, people groups have contributed to inequality reduction. For instance, at the ninetieth percentile had household incomes around 3.3 between 2016 and 2022, the ratio of average household in- times higher than people in the 50th percentile; but by the early come between people with tertiary education and people with 2020s it is about 2.7 times higher (Figure 9, right panel). Some- primary education went from 4.95 to 3.18.34 Employment of thing similar happens to the 50th to 10th percentile ratio. The the household head plays an irregular impact on inequality. It Generalized Entropy, an index of inequality especially sensitive represented a large component of the fall in income inequality to income differences at the bottom of the distribution, also has between 2000 and 2006, but it reversed this impact in 2008- a decline, level-off, decline again pattern. 2014. For the latest period, it also played an equalizing effect, Changes in income inequality can be attributed to changes but smaller than before. This is consistent with evolution of la- within and between population groups. We use a method to ac- bor markets: people working in agriculture, being out of work count how much each group has contributed to these changes, (groups with the lowest labor earnings) grew between 2008 similar to the one used in the section above.32 Demographic and 2014, contributing to an increase in income inequality, and forces, such as shifts in age composition or family type, do not declined again between 2016 and 2022.35 play a noticeable role in changing inequality. Figure 10, top 30 MEXICO Poverty and Equity Assessment Figure 9. Indexes of income inequality GINI Coefficient 0,60 4,0 1,00 3,5 0,55 3,0 0,75 Generalized entropy Inter - qunatile ranges 2,5 0,50 2,0 0,50 1,5 0,45 1,0 0,25 0,5 0,40 0,0 0,00 2000 2004 2000 2002 2020 2004 2008 2006 2002 2020 2022 2008 2006 2022 2010 2014 2010 2012 2014 2018 2016 2012 2018 2016 Generalized entropy (0) Generalized entropy (1) Inter-quantile range 90th-50th Inter-quantile range 50th-10th Source: Own calculations using ENIGH 2000 through 2022. Figure 10. Components of changes of income inequality (as per Generalized Entropy index) by population groups By age group By schooling group 1 chnages in generalized entropy (0) 4 chnages in generalized entropy (0) 0 2 -1 -2 0 -3 -2 -4 -5 -4 -6 -6 -7 -8 -8 -9 -10 2000-2006 2008-2014 2016-2022 2000-2006 2008-2014 2016-2022 By family type group By household head activity 2 2 chnages in generalized entropy (0) chnages in generalized entropy (0) 1 0 0 -1 -2 -2 -3 -4 -4 -6 -5 -6 -8 -7 -8 -10 -9 2000-2006 2008-2014 2016-2022 2000-2006 2008-2014 2016-2022 populations shifts within subgroup inequality between subgroup income change Source: Own calculations using ENIGH 2000 through 2022 Note: Change in inequality decompositions à la Mookherjee and Shorrocks (1982) with STATA command by Silva, Andrew. 2017. “MSDECO: Stata Module to Calculate the Mookherjee & Shorrocks (1982) Over-Time Inequality Decomposition by Subgroup” Statistical Software Components S458373, Boston College Department of Economics. Groups are defined as in Figure 3. 31 Section 1: Monetary Poverty and Inequality in Mexico Trends in income inequality have been particularly sensitive and third period, but increasing it in the second (dark blue bars to the performance of labor markets. Changes in inequality by in Figure 11). This impact is similar to the one observed on pov- sources of income show that employment creation and labor erty and explained by the same factors. earnings had a positive contribution to the fall in inequality be- The analysis above refers to what is called “size” distribution tween 2000 and 2006 (light blue, dark blue and orange bars in of income. That is the distribution of income between persons. Figure 11). This impact was larger in 2008-2014, and even larger Another measurement refers to factorial distribution of income; in 2016-2024.36 This is consistent with the negatively sloped that is the distribution of national income between factors of growth incidence curves shown in Figure 4, top left, which illus- production, namely capital and labor. This may be informative trated that far all periods the growth of labor earnings was faster because returns to capital are under-reported in surveys and for poorer deciles than for higher deciles of the distribution. hence size distribution from surveys mostly refers to distribu- In contrast, public transfers (mostly associated with social tion in labor incomes and transfers, but not in other sources of assistance) have a declining impact on inequality. These ex- income. The share of labor earnings within national income in plained nearly a third of the fall in Gini between 2000 and 2006, Mexico has increased in recent years. It went from 25.7 to 28.2 but had negligible impact in the two subsequent periods (grey percent between 2008 and 2023. If adding the so-called “mixed bars). This is due to the expansion of conditional, means-test- income” which is a combination of labor and capital earnings ed, cash transfer programs in the late 1990s and early 2000s. usually derived from self-employment and small firms, the More recently, growing social assistance programs are share goes from 45.2 to 51.0 percent in the same period.37 This non-conditional, non-means-tested, and universal hence have seems to indicate that workers and small-firms are garnering a a less equalizing effect. This is consistent with Figure 4, bottom larger share of total national income. However, studies that use left, which shows a positively sloped growth incidence curve tax data and other methods to investigate the incomes of the in the period 2016-2022, which contrasts with the slope of the rich underline that Mexico is still a very unequal country. It may other two periods. Private transfers (i.e. remittances) have an be that despite capital rents being a smaller share of national irregular impact on inequality, contributing to its fall in the first income, these are very unequally distributed.38 Figure 11. Components of changes of income inequality (as per Gini coefficient) rate by sources of income 2 1 change in Gini coefficinet 0 -1 -2 -3 -4 2000-2006 2008-2014 2016-2022 2020-2022 share of adults employment share labor income contributory pensions public transfers (social assistance) private transfers (remittances) other incomes Source: Own calculations using ENIGH 2000 through 2022 Note: Poverty change decompositions à la Paes de Barros et al. (2006) computed with STATA command ADECOMP by Azevedo, João P., Minh Nguyen and Viviane Sanfelice (2012). “ADECOMP: Stata module to estimate Shapley Decomposition by Components of a Welfare Measure” Statistical Software Components S457562, Boston College Department of Economics. 32 MEXICO Poverty and Equity Assessment Some international comparisons Up until 2020, global poverty reduction had completed a Mexico shows higher poverty rates than the four compari- remarkable downward trend. Global poverty rates (under the son countries for almost every year of the period under study. global poverty line) went from 30.3 percent to 8.9 percent Only recently Mexico achieved poverty rates that are similar between 1999 and 2019. If using the upper-middle income to the poverty rates these countries had when they had sim- poverty line, a higher threshold, the decline goes from 69.8 ilar GDP/head than Mexico. Indeed, when Chile, Malaysia and percent to 46.4 percent for the same period. The world expe- Türkiye had GDP/head around 20,000 US$ (2017 PPP), these rienced a rapid decline in poverty rates because, among other countries also had poverty rates of around 22 percent. All com- things, most countries experienced positive, sometimes fast, parison countries now have lower poverty rates than Mexico economic growth, Figure 12 left panel, shows the distribution of but also bigger GDP/head ( Figure 13 left panel). The four com- countries in terms of their GDP per capita in PPP terms, around parison countries now have poverty rates at the US$ 6.85 a day 1998 and 2018. Almost every country is above the diagonal, (2017 PPP) below the 10 percent mark; the four comparison which means a higher GDP per capita in 2018 than in 1998. countries have GDP/head in 2017 PPP terms at least 25 percent Mexico is among the so-called upper middle-income coun- higher than Mexico. tries. It is situated, in 2019 and in 1999, between the 40th and Moreover, if comparing measures of the evolution of income 80th percentile of global GDP/head distribution. But when com- inequality as per the Gini coefficient, most countries -except pared with many countries in the same group, it is among the Poland- have a GINI index similar to Mexico (Figure 13, right ones with the slowest growth in GDP/head (Figure 12 bottom panel). For a good part of the decade, Mexico, Malaysia, Türkiye panel). Many countries that started in a position similar or lower and Chile have had Gini indexes between 0.42 and 0.47, and than Mexico, have outperformed it. One may choose four for very little change over time (except for Mexico which has seen comparison: Chile, Malaysia, Poland and Türkiye. The last three a rapid decline of the Gini between 2006 and 2022). Poland, in share with Mexico to have a population of more than 30 million, contrast, had a lower initial Gini index (around 0.37 in the early have close trade links with large market economies, and had 2000s) and has seen a slow but persistent decline over two similar initial conditions in terms of GDP/head in 1999. Chile decades to reach about 0.29 in 2021. does not fully share some of these characteristics but seems a more suitable comparison country within the region.39 33 Section 1: Monetary Poverty and Inequality in Mexico Figure 12. Global long-term economic growth 2017 - 2019 average , GDP per capita, PPP (constant 2017 international $) 128.000 64.000 32.000 16.000 8.000 4.000 2.000 1.000 500 500 1.000 2.000 4.000 8.000 16.000 32.000 64.000 128.000 1997 - 1999 average , GDP per capita, PPP (constant 2017 international $) 40.000 2017 - 2019 average , GDP per capita, PPP (constant 2017 international $) Slovenia Lithuania 35.000 Estonia Puerto Rico Portugal Panama Poland Hungary Slovak Republic 30.000 Latvia Greece Romania Turkiye Croatia Seychelles Russian Federation Malaysia Kazakhstan Trinidad and Tobago 25.000 Chile Libya Uruguay Mauritius Argentina Bulgaria Costa Rica Montenegro Mexico 20.000 Maldives Thailand Suriname Dominican Republic Lebanon Serbia North Macedonia Barbados 15.000 Brazil Colombia Gabon Sri Lanka South Africa Peru Ukraine Indonesia Algeria Ecuador 10.000 5.000 10.000 15.000 20.000 25.000 30.000 1997 - 1999 average , GDP per capita, PPP (constant 2017 international $) Source: Own calculations using World Development Indicators (https://databank.worldbank.org/source/world-development-indicators accessed April 15th, 2024) Note: Yellow vertical and horizontal lines represent the 40th and 80th percentile of global distribution of GDP/head. 34 MEXICO Poverty and Equity Assessment Figure 13. Poverty and Inequality over time across comparison countries Evolution of Gini coefficient and GDP/head Poverty Rates (at US$ 6.85 poverty line) and GDP/head (circa 2000 through circa 2022) (circa 2000 through circa 2022) 60,0 57,0 1998 1998 Poverty headcount ratio at $6.85 a day (2017 PPP) 50,0 2000 52,0 2002 2000 2004 2005 40,0 2010 200620082014 47,0 2003 (% of population) Gini coefficient 2012 2021 2017 2016 42,0 2022 30,0 2020 2016 2003 2018 2006 2021 2004 2022 37,0 20,0 2006 2004 10,0 32,0 2021 2022 2021 2021 2021 27,0 0,0 15.000 17.000 19.000 21.000 23.000 25.000 27.000 29.000 31.000 33.000 35.000 15.000 20.000 25.000 30.000 35.000 GDP per capita, PPP (constant 2017 international $) GDP per capita, PPP (constant 2017 international $) Chile Mexico Türkiye Poland Malaysia Chile Mexico Türkiye Poland Malaysia Source: Own calculations using World Development Indicators (https://databank.worldbank.org/source/world-development-indicators accessed April 15th, 2024) Note: Segmented lines represent breaks in comparability of series because of changes in surveys. 35 Section 1: Monetary Poverty and Inequality in Mexico The lower poverty rates in comparison countries seem to 2000s and by 2019 it had surpassed Mexico by 5 percentage be mostly associated with faster economic growth. It is un- points (although Chile seems to have lost impulse because of derstood that economic growth renders faster poverty decline the pandemic). Even Türkiye, with much lower rates throughout when inequality is lower to start with or declines with econom- the period, saw an increase of this rate of nearly 12 percentage ic growth. The comparative experience of these five countries points, whereas Mexico by about 8 percentage points. These shows that most poverty reduction can be ascribed to econom- data show that the country can further increase female employ- ic growth, rather than reduced inequality; perhaps with the ex- ment which, as seen before, contributes to poverty reduction. ception of Poland where both economic growth and declining High rates of informal employment are perhaps the main inequality have led to poverty eradication.40 obstacle to inclusive growth in Mexico. Informal employment is International comparisons may also illustrate margins for associated with low productivity and lack of social protection, accelerated poverty reduction Besides faster growth. As indi- which makes workers less likely to cope with economic shocks cated in previous paragraphs Mexico has experienced rapid and then more vulnerable to poverty. Informal employment in poverty reduction in recent years partly because of growing Mexico has declined slowly over the past decade, but it is still female labor force participation, growing formal employment nearly 30 percentage points higher than in Türkiye and Chile and increased minimum wages. The country has a margin to (Figure 14, bottom left panel).42 The experience of Türkiye and accelerate further the first two trends, but perhaps less so the Chile is one of accelerated economic growth and reforms to the third. social security system. Mexico needs to reduce the excessive Minimum wages in Mexico have increased the most from a size of its informal sector.43 comparative perspective. Within members of the OECD, Mexico Another comparison of interest is the size and productivity has experienced the largest increase in a comparable measure of the agriculture sector. In previous sections of this report, the (that is, the ratio of the minimum wage to the median wage of high incidence of poverty for households whose head works full-time workers). It went from 38.1 to 68.3 percent between in agriculture has been indicated as characteristic of poverty 2005 and 2023, surpassing Korea (the second largest) which and chronic poverty. There is also a divergence of international went from 37.3 to 60.9 for the same period ( Figure 14, top experiences in this regard. Although all countries have experi- left panel). Comparison countries within this set experienced enced a decline in the share of employment working in agricul- a smaller increase (Poland went from 42 to 54 percent while ture, Mexico still has the higher share, bar Türkiye. Moreover, Chile went from 65 to 70) or even a decline (Türkiye fell from 74 average labor productivity in the primary sector is the lowest to 64). This sheer increase in minimum wages does not seem in Mexico (Figure 14, bottom right panel). More on this will be to have caused employment reduction in Mexico, instead it discussed in Box 3 in section 3 of this report. seems to have benefited poverty reduction. However, it is not These international comparisons underline that Mexico has evident what is the limit to the use of this policy tool. Further to accelerate economic growth in order to further reduce pover- increases in minimum wages could potentially erode employ- ty rates and aim towards poverty eradication in the near future. ment creation in the future if not accompanied by further eco- This growth needs to be inclusive so that inequality does not in- nomic growth (as seen in other countries).41 crease or even declines further. Faster economic growth, more Employment opportunities for women in Mexico have cer- and better jobs for women, continued policy reforms to formal- tainly increased, but there is room for further progress. In the ize labor relations, modernization of agriculture production in early 2000s, both Poland and Mexico had similar employment areas where it remains a low productivity sector and sustaining rates for women, but by 2023, Poland’s female employment the protective value of minimum wages (while avoiding these rate is 5 percentage points higher than Mexico’s (Figure 14, top to erode employability) constitute a sure blueprint towards right panel). Chile even had a lower rate than Mexico in the early poverty eradication in Mexico. 36 MEXICO Poverty and Equity Assessment Figure 14: Minimum wages, females employment, informal and agricultural employment, international comparison Minimum relative to median wage of full time workers Female Employment to population ratio (aged 15 and more) 95 55 Colombia 2023 Costa Rica 2023 Colombia 2007 85 percentage of working age women 50 75 Costa Rica 2010 Türkiye 2005 45 Chile 2023 Portugal 2023 Mexico 2023 France 2005 percentage 65 Chile 2006 Türkiye 2023 New Zealand 2023 40 France 2023 Korea 2023 United Kingdom 2023 Australia 2005 55 New Zealand 2005 Israel 2005 Poland 2023 Australia 2023 35 Portugal 2005 Spain 2023 Germany 2023 Slovak Republic 2023 Canada 2023 Israel 2023 Belgium 2023 Greece 2023 Hungary 2023 Germany 2015 Netherlands 2023 Lithuania 2023 45 Hungary 2005 Lithuania 2005 Greece 2005 Estonia 2023 Japan 2023 Belgium 2005 Netherlands 2005 30 Poland 2005 United Kingdom 2005 Latvia 2023 Czechia 2023 Slovak Republic 2005 Czechia 2005 Mexico 2005 Estonia 2005 Canada 2005 Latvia 2005 Spain 2005 35 Korea 2005 Japan 2005 25 United States 2005 25 United States 2023 20 10.000 20.000 30.000 40.000 50.000 60.000 70.000 80.000 2000 2004 2002 2020 2008 2009 2003 2005 2006 2022 2023 2007 2001 2010 2014 2012 2021 2018 2019 2013 2015 2016 2017 2011 Mexico Poland Chile Turkiye Malaysia GDP per capita, PPP (constant 2021 international US $) Employed covered in the event of work injury Employment and Labor Productivity in Agriculture GDP per worker (in thousand 2015 US$) 90 45 80 40 percentage of employemnt / percentage of total employment 70 35 60 30 50 25 20 40 15 30 10 20 5 10 0 0 Mexico Chile Poland Turkiye Malaysia Mexico Chile Poland Turkiye Malaysia 2022 2022 2022 2023 2010 2010 2021 2019 2019 2015 Chile Mexico Malaysia Poland Türkiye Employment in agriculture (% of total Agriculture, forestry, and fishing, value employment) (modeled ILO estimate) added per worker (constant 2015 US$) 1999 2022 Source: top left: OECD Data Explorer. https://data-explorer.oecd.org/ / top and bottom right: World Development Indicators, https://datatopics.worldbank.org/world-development-indicators/ / bottom left: ILO, ILOSTAT https://ilostat.ilo.org/data/# / (all retrieved August 3, 2024) 37 Section 2: Social deprivations in Mexico Section 2: Social deprivations in Mexico 38 MEXICO Poverty and Equity Assessment As indicated before, this section 2 of the report will discuss deprivations using the official measures of carencias sociales, non-monetary measures of poverty in Mexico. The country has as they are called in Mexico; followed by a discussion of public been a pioneer in devising and implementing official multidi- expenditures allocated to reduce such deprivations. Section mensional measures of poverty which include, in addition to 2 concludes -like section 1- with an international comparison monetary poverty, measures of lack of access to several basic using the World Bank own multidimensional poverty as well as services derived from a rights perspective and the Mexican Con- some other inter-country contrasts of social expenditures. stitution. The next section will describe the evolution of social Advances and reversals of multidimensional poverty in Mexico Official measures of multidimensional poverty in Mexico 2022. Similarly, deficiency in “quality and spaces of dwelling” started in year 2008. These follow a mandate of the General declined from 17.7 percent in 2008, to 12.3 in 2014 and from Law of Social Development and include, in addition to monetary 12.0 in 2016 to 9.1 in 2022 (Figure 15). These are not small poverty, measures of deprivation in the dimensions of educa- declines. They represent 6.1, 7.7 and 10.6 million people less tion, health, social security, housing quality, basic utilities and being affected by these wants, respectively, between 2008 food security. In addition to these contexts of economic and and 2022. social deprivation -but not included in the definition of poverty- The largest decline is observed in access to social securi- Mexican authorities also assess a territorial context in terms of ty. The percentage of people without access to social security social cohesion as per measures of income inequality and in went from 65.0 percent in 2008, to 58.5 percent in 2014 and terms of basic infrastructure as per measures of localities with- then from 54.1 in 2016 to 50.2 percent in 2022. This means out electricity, paved roads, trash disposal and sewerage. This 22.3 million people either got formal employment, or a family section of the report concentrates on social deprivations, while relative got formal employment, or -if being aged 65 and more- measures of monetary poverty and inequality were discussed is protected under a contributory or a non-contributory pension in the previous section and measures of territorial context will scheme. This is consistent with trends in formal employment be included in the next section. and pension beneficiaries observed in recent years and doc- The indicators to measure social deprivations changed umented elsewhere in this report. However, and despite the slightly in 2018. This was to respond to the mandate -included progress, access to social security was and still is the largest in the initial statutory rule- to review the methods if changes carencia social. It more than doubles the incidence of any other in conditions warrant so. For instance, a Constitutional change carencia, bar the case of access to health services. in 2012 extended mandatory public education from basic to In contrast, two of the carencias sociales increased in re- secondary and, consequently, the measure of educational gap cent years. The percentage of the population with an age spe- had to be adjusted.44 These changes, and modifications in the cific educational gap went from 21.9 percent in 2008 to 18.7 household survey, imply that measures between 2014 and percent in 2014, but from 18.5 in 2016 to 19.4 percent in 2022. 2016 are not strictly comparable, although the differences This is a slow moving carencia because it is defined in terms are very small.45 The text below acknowledges this break, and of what is mandatory public education given a specific age co- discusses trends from 2008 to 2014, and from 2016 to 2022; hort. Adults who did not complete mandatory public education which emulates the periods in the previous section. remain with an educational gap unless they engage in adult Four of the six carencias sociales have declined over the education. Thus, this carencia moves mainly by the educational period of study. The lack of “access to food” went from 21.9 advancement of younger cohorts. It can be partly argued that percent in 2008 to 23.4 percent in 2014, but then declined the Covid-19 pandemic affected schooling for children and from 21.9 in 2016 to 18.2 percent in 2022. Absence of “ac- those who could not engage in distance education; but the de- cess to basic services in housing” went from 22.9 percent in celeration and reversal in the trend of this indicator is observed 2008, to 21.2 in 2014 and from 19.2 in 2016 to 17.8 percent in before year 2020.46 39 Section 2: Social deprivations in Mexico The largest reversal in the downward trend of carencias is to medical services, but between 2016 and 2022 more than 30 observed in enrollment in health services. The percentage of million lost registrations. The number of unenrolled populations the population without enrollment in medical services -public in 2022 (50.4 million) is now larger than in 2008 (42.8 million). or private- fell from 38.4 percent in 2008 to 18.2 percent in This severe setback is ascribed to the yet incomplete transition 2014. It declined still to 15.6 percent in 2016 but has increased from the former Seguro Popular (a health program for those not since to 39.1 percent in 2022. This reversal means that be- enrolled in formal employment) and the new health service pro- tween 2008 and 2016, about 24.0 million people got enrolled gram called Bienestar (more on this in the next section). Figure 15: Evolution of indicators of social deprivation, carencias sociales. 70 60 percentage of total population 50 40 30 20 10 0 2008 2010 2012 2014 2016 2018 2020 2022 Access to Health Services Access to Social Security Quality and space of dwelling Access to basic services in housing Average educational gap Access to Food Source: CONEVAL https://www.coneval.org.mx/Medicion/MP/Paginas/Pobreza_2022.aspx. and https://www.coneval.org.mx/Medicion/MP/Paginas/Pobreza-2018.aspx 40 5% 0% -5% -15% -10% -35% -25% 0% -30% -20% 10% 70% 30% 50% 60% 80% 90% 20% 40% 100% Average Average educational educational gap gap Access Access to Health to Health Services Services MEXICO Access Access to Social to Social Security Security Quality and Quality and space space of dwelling of dwelling Access to Access to basic services basic services in housing in housing Access Access Poverty and Equity Assessment Social Deprivations as of 2014 to Food to Food Change in Social Deprivations between 2008 and 2014 by household income quintiles (CONEVAL welfare aggregate) by household income quintiles (CONEVAL welfare aggregate) At least on At least on deprivation deprivation National At least three At least three deprivations deprivations 1st quintile 2 0% 10% 0% 10% 30% 50% 20% 40% -10% 70% 30% 50% 60% 80% 90% 20% 40% 100% Average Average 3 educational educational gap gap 4 Access Access to Health to Health Source: Own calculations using ENIGH 2008 through 2022 Services Services Access Access 5th quintile to Social to Social Security Security Quality and Quality and space space of dwelling of dwelling Access to Access to basic services basic services Figure 16: Changes in social deprivations by household income quintile, 2008-2014 and 2012-2022 in housing in housing Access Access Social Deprivations as of 2022 to Food to Food Change in Social Deprivations between 2016 and 2022 by household income quintiles (CONEVAL welfare aggregate) by household income quintiles (CONEVAL welfare aggregate) At least on At least on deprivation deprivation At least three At least three deprivations deprivations 41 Section 2: Social deprivations in Mexico The percentage of the population with at least one carencia cline in the share of the population with at least three carencias has declined, but the percentage of the population with at least for this group than the rest (Figure 16, bottom left panel). In three has increased. The number of carencias is a key indicator contrast, the decline in access to health services, and three of multidimensional poverty in Mexico. The definition of multi- carencias or more, has increased more for the poorest families dimensional poverty or extreme poverty in Mexico depends on in the period 2016-2022 (Figure 16, bottom left panel). The whether a household suffers at least one or at least three caren- cases of “access to basic services in housing” and “quality and cias. Having one carencia went from 76.6 percent in 2008, to spaces of dwelling” do show some pro-poor focus but still, as of 72.4 percent in 2014 and from 68.5 in 2016 to 65.7 percent in 2022, still show a gap of 20 and 40 percentage points, respec- 2022. In contrast, having three carencias fell from 31.7 percent tively, across socio-economic groups.47 in 2008 to 22.1 percent in 2014, but rose from 20.0 in 2016 to It is important to underline a caveat about the social depri- 24.9 percent in 2022. These divergent trends are basically the vation indicator of access to health services. The indicator consequence of advances in the reduction of lack of access to refers to the percentage of the population that is neither en- social security and the reversal in the lack of access to health rolled to public institution of social security (like IMSS or ISSTE), services. nor in a private medical service, nor to the medical services The frequency of carencias at the national level hides major of Seguro Popular or the Instituto de Salud para el Bienestar, differences across population groups and localities. The inci- INSABI. Between 2018 and 2023, the latter were in a transition dence is always much higher for lower income households than process which, together with the COVID-19 crisis in 2020, may for higher income households. The exception is the lack of ac- have led many people to believe that they had no public health cess to health services which, in 2014, had achieved an almost coverage.48 However, this does not mean that people did not neutral incidence across all income levels (Figure 16, top left have access to health services. As will be described in the next panel); although this has been undone and now lack of access section, federal budget expenditures in health services did not to health ranges from 62 percent for the poorest households decline in these years. People who required health attention and 20.3 percent for households in the top income quintile had the right to receive it, but there is also evidence that there (Figure 16, top right panel). All other carencias show a very un- was an increase in the use of private medical services (from equal prevalence with gaps of more than 50 percentage points 28.7 to 35.3 percent of those who required medical attention, (in access to social security) down to 20 percentage points (in according to ENIGH data for 2018 and 2022).49 There is also an access to “quality and spaces of dwelling”). increase in the share of household expenditures allocated to Moreover, these gaps across groups have hardly changed health services (from 2.7 percent in 2016 rose to 4.2 in 2020 between 2014 and 2022, which implies that public policy to- and declined to 3.4 percent in 2022), and in households with wards reducing carencias have not focused enough upon the catastrophic health expenditures (more than 30 percent of dis- poorest groups of the population. Access to health services posable income) which reached 3.9 and 2.9 percent of the pop- which, during the period 2008-2014 reduced the prevalence ulation in years 2020 and 2022, respectively, still higher than of the carencia much more for families in the poorest quintiles the 2.1 percent of pre-pandemic 2018, and particularly among than for the rest of the population. This also led to a faster de- households in the lower ranks of the income distribution.50 Public expenditures and its distribution explain social deprivations in Mexico The evolution of social expenditures within the national 2009 and 2015, then declined to slightly below 10 percent un- budget partly explains the trends in social deprivations. Public til 2018. Since then, they have increased to a peak in the series expenditures in areas that are directly related to social depriva- of 12.5 percent in 2024.51 tion (that is, public health and education, pensions, social as- The component of social expenditures that has grown sig- sistance and programs associated with food security and social nificantly over this period is pensions. Pensions -contributory housing) hovered around 10 and 11 percent of GDP between and non-contributory- represented 2 percent of GDP in 2008 42 MEXICO Poverty and Equity Assessment and 5.8 percent in 2024 ( Figure 17 ). The difference in budget- crease, which coincides with nearly flat social expenditure allo- ary allocation to social expenditures is of about 2 percentage cations to education, health, social housing and food security. points of GDP (from the three-year average around 2023 and The evolution of total expenditures does not suffice to explain the three-year average around 2009). This is the net result of, trends in poverty or social deprivation because it may hide major on the one hand, allocations to contributory and non-contribu- differences in the allocation to specific social groups. It is not only tory pensions which increased by 3 percentage points of GDP total and average expenditures what matters, but focused and suf- (2 and 1 percentage points, respectively). And, on the other ficient allocations to specific groups in poverty and social depriva- hand, other expenditures experienced a relative decline (0.6, tion what would induce reductions in multidimensional poverty. 0.3 and 0.2 percentage points in the case of education, health In many cases, the social expenditures listed above allocate most and housing) while those ascribed to food security remained of its budgeted resources to the poor and vulnerable families. But constant. A loose connection between these trends and the there are cases where the bulk of the budget goes elsewhere in evolution of social deprivations suggests that the social depri- the income distribution. Yet in other cases, although the budget vation that declined the most, access to social security, is in- focusses on these most in need, it is not sufficient to move them deed connected to the social expenditure that grew the most: out of poverty or deprivation. In both cases, either because of lack contributory and non-contributory pensions. Conversely, the of focus or insufficient of funding, social expenditures may fall other social deprivations either declined or had a small in- short of reducing poverty. Figure 17: Evolution of social expenditures within Mexico’s’ Federal Budget. 14% 12% 10% Percentage of GDP 8% 6% 4% 2% 0% 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Contributory pensions Non-contributory pensions Education Health Housing and social services Food security Source: Data from Cuenta Pública 2008-2023, processed and analyzed by Macías et al. (2024). Note: Data for 2024 is from federal budget. The cases of public services in health and education show to about MX$ 6,750 in 2024 (as per budget). These health ex- a slight decline in budgetary resources and margin to improve penditures, however, are highly fragmented. As of 2022, more emphasis on the poor. Budgetary allocation to public health than half of the total budget of the health services from IMSS, declined from 3.0 percent of GDP around 2009 to 2.4 around ISSTE and PEMEX go to households in the top half of the income 2018, to increase again to 2.9 to respond to the Covid-19 pan- distribution; whereas health services funded by the Ministry of demic in 2020, where it has remained since (Figure 18, top left Health and other institutions52 go mostly to households at the panel). This has made the public expenditure allocation go from bottom of the income distribution (Figure 18, top center pan- around 6,200 MX pesos per person per year (in prices of 2022) el). In the end, households at the top of the distribution receive to around 5,650 MX pesos between 2009 and 2018, to then rise a larger share of the total health budget (13.6 percent) than 43 Section 2: Social deprivations in Mexico those at the bottom (6.3 percent). These differences in budget of a rather stable annual nominal budget growth in allocation allocations translate into differences in scope of service: IMSS, to each sub-system (except tertiary, which has decelerated), ISSSTE, and Pemex have provided coverage for 8,000 interven- together with diverging trends in enrollment for each group. tions from the ICD (International Statistical Classification of Dis- Over the long run, as part of the demographic transition, the eases and Related Health Problems), whereas Seguro Popular number of students in basic education have declined, whereas covered 1,807 interventions, and IMSS-Bienestar covered only students in upper secondary and tertiary have grown. 50. The ongoing transformation of Seguro Popular, INSABI and Expenditures in public education are pro-poor. They repre- IMSS-Bienestar aims towards a more comprehensive provision sent a larger percentage of household income for families at the of services for those not covered by IMSS and ISSTE. bottom of the distribution for all levels: basic, secondary and Despite these differences, health expenditures represent tertiary (Figure 18, bottom right panel). Although much more a larger share of the budget among the poor. Public health ex- for the first than for the other two, because poorer families penditures from FASSA/INSABI/IMSS-Oportunidades, and other have more children and primary education gross enrollment public services represent more than 20 percent of average rates are above 100 percent for the most of the period; whereas household income for households in the first decile of the in- gross enrollment rates for secondary and tertiary enrollment come distribution, but less than 5 percent for households for are also growing but lower (go from 71 to 98 percent and from 4th and higher deciles. Something similar, but less pronounced 20 to 46 percent, respectively, between 2000 and 2022).54 occurs with IMSS (Figure 18, top right panel). A larger allocation The distribution of the education budget, is also pro-poor in the of resources to programs funded by FASSA/INSABI/IMSS-Opor- case of basic education: more than half the budget goes to the tunidades and Ministry of Health, or an extension of cover- poorer half of the population. But is more neutral in the case age of program like IMSS to all the population, would reduce of upper-secondary, with 2nd to 8th deciles of the distribution the regressive distribution of the health budget, increase the receiving around 10 percent of the budget each. The budget relative progressivity of the system and repair the reversal in for tertiary education, however, is more skewed towards higher health coverage described as main trend of social deprivations income families (Figure 18, bottom center panel). above.53 Within the education expenditures there are cash transfers Trends in public education expenditures are somewhat to guarantee access to education, particularly among vulner- different. The public education expenditures as percentage able groups. These pivoted from the conditional cash-transfer of GDP have declined from 3.7 percent in 2009 to 3.1 percent program Oportunidades-Prospera (which represented around in 2023 (three year average in both cases). But average ex- 0.15 percent of GDP between 2008 and 2014) to new schol- penditures per student show different trends. Students in ba- arship programs Elsa Acuña and Benito Juárez for basic ed- sic education received from about MX$ 22,400 per student in ucation (similar share of GDP between 2020 and 2024) and 2009 to nearly MX$ 27,700 in 2024 (in MX$ of 2022); those in programs for upper secondary and tertiary, Benito Juárez, and upper secondary went from nearly MX$ 28,700 to MX$ 30,500; tertiary, Jóvenes Escribiendo Futuro (an additional 0.15 per- whereas expenditures to those in tertiary fell from MX$ 73,600 cent of GDP between 2021 and 2024). to MX$ 52,000 (Figure 18, bottom left panel). This is the result 44 MEXICO Poverty and Equity Assessment Figure 18. Average expenditure, relative and absolute incidence for Public Health and Education Public Health Expenditures Absolute Incidence of Public Health Services (2022) Relative Incidence of Public Health Services (2022) Percentage of Market Income of beneficiaries 35% 10 25% 10 8.000 8% 9 9 Percentage of total expenditures 7.000 7% 30% 8 20% 8 MX Pesos of 2022 per head 6.000 6% 25% 7 7 5.000 5% 6 15% 6 20% 4.000 4% 5 5 15% 10% 4 4 3.000 3% 10% 3 3 2.000 2% 2 5% 2 1.000 1% 5% 1 1 - 0% 0% 0 0% 0 IMSS ISSSTE PEMEX FASSA TOTAL IMSS ISSSTE PEMEX FASSA 2008 2009 2010 2012 2013 2014 2015 2016 2017 2018 2019 2020 2022 2023 2024 2011 2021 per head (left axis) percentage of GDP (right axis) Public Education Expenditures (2008-2024) Absolute Incidence of Public Education Services (2022) Relative Incidence of Public Education Services (2022) Percentage of Market Income of beneficiaries 18% 10 100% 10 78.000 6% 16% 9 90% 9 Percentage of total expenditures 65.000 5% 14% 8 80% 8 MX Pesos of 2022 per head 7 70% 7 12% 52.000 4% 6 60% 6 10% 39.000 3% 5 50% 5 8% 4 40% 4 26.000 2% 6% 3 30% 3 4% 20% 13.000 1% 2 2 2% 1 10% 1 - 0% 0% 0% 0 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Basic Secondary Tertiary Total Basic Secondary Tertiary Education Education Education Education Education Education per student in secondary education (left axis) percentage of GDP (right axis) per student in basic education (left axis) per student in tertiary education (left axis) Source: Top and bottom left, authors calculations using data from Macías et al. (2024). Center and right panels, own calculations using ENIGH 2022 data processed using MEXSIM. The horizontal axes of center and left panels refer to deciles of pre-tax and transfers household per-capita income. Note: FASSA refers to people reporting being enrolled in health services either by Seguro Popular, IMSS-Oportunidades or INSABI (see footnote 48). 45 Section 2: Social deprivations in Mexico In total, the education system public budget is still pro- ered to 65 years in 2021. By 2023, 10.2 million elderly ben- gressive in both relative and absolute terms, but ongoing de- efited, with over 95% of those over 65 receiving this pension. mographic trends -and budgetary pressures- pose challenges Some pensioners from the contributive scheme are under 60, to eradicate carencias sociales due to educational gap. Without mostly from state enterprises and the public sector, explaining an emphasis on quality, budgetary or redistributive measure there are more pensioners than population aged 65 and older. will fail to close the gap. Academic skills acquired in basic Pension payments vary considerably across systems. The education are indispensable for successful completion of up- per capita annual public pension spending has usually been per-secondary and tertiary education. A recent report indicates higher been for CFE, PEMEX, and IMSS whereas non-contributive that: “Large-scale learning assessments of students in Mexico pensions receive much lower pension.57 But the number of ben- indicate that 47 percent do not achieve the minimum profi- eficiaries in each system varies a lot. Only 52 thousand pen- ciency level at the end of primary school, proxied by data from sioners received a pension from CFE in 2022, compared to 9.8 grade 6 in 2019.” The most recent PISA for year 2022 is lower million receiving the non-contributive pension. Originally, this than previous assessments in years 2015 and 2018. Although non-contributory pension did not suffice to cover the food and among the highest in the LAC region, Mexico has not narrowed non-food poverty line; but since 2022 it has surpassed both the gap with the OECD average.55 Quality basic education is still rural and urban food poverty lines. This extension in coverage an important issue to guarantee access. Investing in quality and sufficiency of the non-contributory pension partly explains education is also investing in access to education. -together with the growing number of IMSS recipients and the Education and health used to be the largest components growth in formal employment- the decline in recent years of of social public expenditures in Mexico, but not anymore. Pen- carencia social in terms of lack of access to social security, sions, both contributory and non-contributory, now represent already reported in the previous section. the largest component of social expenditures: as large as ed- The recent growth of non-contributory pensions has trans- ucation and health together by 2024. As in the case of health, formed the context of transfers in Mexico. Contributory pen- and somewhat in education, total expenditures in pensions and sions can be interpreted as delayed incomes, not transfers, its distribution across socio-economic groups vary substantial- because these somehow accrued to beneficiaries according ly throughout components of the system. to previous contributions as workforce. Non-contributory pen- The Mexican pension system is fragmented with large, sions are cash transfers that aim to reduce poverty among medium, and small sub-systems at the federal, state, and mu- vulnerable population. As such, non-contributory pensions nicipal levels. At the federal level, the system comprises four are now the largest -both in budget and number of beneficia- coexisting pillars: pillar 0 refers to the non-contributive pen- ries- anti-poverty cash transfer program in Mexico. These are sion; pillar one is the DB (Defined Benefit) scheme; pillar two is followed by Becas Benito Juárez (a scholarship for students the DC (Defined Contribution) scheme; and pillar three allows in basic and upper secondary school, vulnerable to drop out of for voluntary savings that workers can make to their individ- school) with about 12 million beneficiaries; Producción para ual accounts of pillar two or other voluntary parallel savings el bienestar, a cash transfers to small and medium farmers, schemes.56 with about 2 million beneficiaries; and Jóvenes Construyendo Pension coverage has become practically universal. Pillars Futuro an allowance-cum-mentorship program to youth aged one and two are linked to retirees who contributed to the formal 18-29 with no school or employment activity, with about 2.9 labor sector. In Mexico about 60% of the elderly receive this kind million beneficiaries.58 of pension, funded by contributions from workers, employers, Despite its budgetary growth in recent years, the allocation and the government. In 2019, the non-contributive pension be- of transfers does not predominantly go to the poorest groups came universal, covering all persons over 68, and indigenous of the population. The notable case is the Programa de Adultos community members over 65. This age requirement was low- Mayores, whose total budget is allocated almost equally across 46 MEXICO Poverty and Equity Assessment all the population groups defined by household income decile. to receive care, but 35 percent of them do not. The lack of care Of course, given their lower income, these non-contributory pen- is higher among people with disabilities (39 percent) and the sions represent a larger share of a poorest household income ( elderly (78 percent). Women represent 3 out of 4 care givers. Figure 19, top left), and have had a sizable impact on poverty The survey shows that 68% of the economically inactive women reduction in recent years. In other words, although the program who provided care and expressed interest in working reported is very progressive in relative terms, it is neutral in absolute in- that they couldn’t due to lack of care support for children, el- cidence. This implies that, as will be explained later, the program derly, ill individuals, and persons with disabilities. There is a could have had a larger impact on poverty if its resources were predominant percentage of the population that agrees these more concentrated among the poor. are not the sole responsibilities of women and that the use of In other cases, transfers programs have little coverage these services depends crucially on quality services.60 These or benefits are insufficient to dent poverty rates. The Benito figures show an important demand for care services which pub- Juarez scholarship coverage is quite extensive, and benefits lic and private sectors could partner to provide. are progressive both in relative and absolute terms, more for There are also important public expenditures on social pro- basic education than for upper-secondary (Figure 19, top cen- grams linked to promote employment and entrepreneurship, ter). But their amount is relatively small when compared to the both for urban and rural areas. It can be said that these pro- poverty line (in 2022, 667 MX$ per month per family for ten grams concentrate on the monetary side of poverty reduction, months, but 920 MX$ in 2024). Something similar can be said rather than carencias sociales. This is the case of programs to about the program Produccion para el Bienestar (Figure 19, top create economic opportunities for economic development, right), which has progressive relative and absolute incidence, where old programs like Programa de Apoyo al Empleo (PAE) but its transfers were relatively small (4800 MX$ per year in and the Fund for Small and Medium Enterprises (Fondo PYME) 2022, but up to between MX$ 6000 and 24000 per year in have been reduced or replaced, while new programs like Jo- 2024). venes Construyendo Futuro have grown, representing about Cash transfer programs targeted to women, youth and pop- 0.7 percent of GDP both in the late 2000s and the early 2020s. ulation with disabilities merit special attention. The Madres Early programs of social protection, social assistance and ru- trabajadoras program provides a large transfer (up to MX$ ral infrastructure such as Habitat, Microregiones, Programa de 1,800 as of 2022) and its budget is well targeted to vulnerable Empleo Temporal (PET) and Sedesol-Oportunidades have been households, but its coverage is very small ( less than a quarter replaced by the new program Sembrando Vida, declining from million beneficiaries, as of 2022) leading to very small relative 0.4 percent of GDP in the late 2000s, to 0.1 percent in years incidence (Figure 19, bottom center). Something similar can 2022 through 2024.61 In terms of support to rural areas and ac- be said about the youth program Jóvenes Construyendo Futuro tivities of the primary sector, where poverty has high incidence, (Figure 19, bottom left).59 Finally, the Bienstar Discapacidad the budget allocations moved from around 0.4 percent of GDP program, which targets people with permanent disabilities, is in the late 2000s, to 0.2 percent in the late 2010s where it has progressive in relative and absolute terms, its transfers are not remained by the 2023 and 2024; with emblematic programs small, but can be expanded and better integrated into a more pivoting from PROCAMPO in the late 2000s, to Programa de comprehensive care system. (Figure 19, bottom right). Fomento a la Agricultura and Programa Integral de Desarrollo These programs can serve as a basis for a larger, better inte- Rural in the mid 2010s, towards the new programs Produccion grated and funded, family care program in the future. According para el Bienestar, Precios de Garantía, and Fomento Ganadero to the national Survey for the National Care System (ENASIC in operation since 2020. 2022), 45.2 percent of the Mexican population is susceptible 47 Section 2: Social deprivations in Mexico Figure 19. Relative and absolute incidence of largest -by number of beneficiaries- cash transfer programs in Mexico Pensión Adultos Mayores Becas Benito Juárez Producción para el Bienestar 35% 35,0% 30% 35% 3,5% Percentage of Market Income of beneficiaries Percentage of Market Income of beneficiaries Percentage of Market Income of beneficiaries Percentage of total benefits of the program Percentage of total benefits of the program Percentage of total benefits of the program 30% 30,0% 25% 30% 3,0% 25% 25,0% 25% 2,5% 20% 20% 20,0% 20% 2,0% 15% 15% 15,0% 15% 1,5% 10% 10% 10,0% 10% 1,0% 5% 5,0% 5% 5% 0,5% 0% 0,0% 0% 0% 0,0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Deciles of gross market income plus contributory pensions Deciles of gross market income plus contributory pensions Deciles of gross market income plus contributory pensions absolute (left axis) relative (right axis) relative (Educación Básica) relativa (Educación Media Superior) absolute (left axis) relative (right axis) absolute (Educación Básica) absolute (Educación Media Superior) Jóvenes Construyendo Futuro Madres Trabajadoras Bienestar Discapacidad 20% 1,2% 35% 0,3% 20% 2,0% Percentage of Market Income of beneficiaries Percentage of Market Income of beneficiaries Percentage of Market Income of beneficiaries Percentage of total benefits of the program Percentage of total benefits of the program Percentage of total benefits of the program 30% 16% 0,2% 16% 1,6% 25% 0,8% 12% 0,2% 12% 1,2% 20% 15% 8% 0,1% 8% 0,8% 0,4% 10% 4% 0,1% 4% 0,4% 5% 0% 0,0% 0% 0,0% 0% 0,0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Deciles of gross market income plus contributory pensions Deciles of gross market income plus contributory pensions Deciles of gross market income plus contributory pensions absolute (left axis) relative (right axis) absolute (left axis) relative (right axis) absolute (left axis) relative (right axis) Source: Own calculations using ENIGH 2022 data processed using MEXSIM. 48 MEXICO Poverty and Equity Assessment The differences in budget, coverage, and incidence of these pact, like Madres trabajadoras, Produccion para el Bienestar cash transfer programs lead to diverse impacts on poverty and Sembrando Vida.64 and inequality. A simulation exercise gauges what the pover- Alternatively, cash transfer programs can be compared ty rate, or the Gini coefficient, would be if removing a specific in terms of their potential efficiency. Given its current design program. Figure 20, left panel, illustrates the results of these and budget, Madres trabajadoras has a larger efficiency upon simulations, using data from ENIGH 2022, and ranks programs poverty reduction per MX$ spent than Pension Adulto Mayor according to their poverty and inequality reduction impact.62 (Figure 20, right panel). This suggests that a reallocation of Programa de Adultos Mayores, has the largest poverty re- budget from the latter to the former could increase the over- duction impact (nearly two percentage points), followed by all poverty reduction impact of the cash transfer programs. In the Benito Juarez scholarships (more than half a percentage other words, there is potential to increase the poverty reduction point, together), while the rest of the programs have a much impact of existing programs, by changing some aspects of its smaller contribution to poverty (generally below 0.2 percent- design. These and previous paragraph results refer to data for age points of poverty reduction). This is because despite its low 2022; the growth of these programs in terms of coverage and absolute incidence (by the Kakwani index in the right axis of sufficiency in 2023 and 2024 may have led to different im- the same figure)63, the Programa de Adultos Mayores, has very pact in more recent years. Newer data is needed to re-assess large coverage and budget, leading to an average benefit that these programs. Moreover, these estimates are only budgetary represents a large share of household incomes in poverty. In assessments based on administrative data. More systematic contrast, other programs with higher Kakwani index have small impact evaluation following experimental design are needed coverage and/or budget, leading to small poverty reduction im- to have a more comprehensive assessment of the true impact of these programs.65 Figure 20. Effectiveness and Efficiency of main cash transfer programs in Mexico Poverty reduction relative to budget allocation, 2022 2,5 1,25 2,0 2,0 1,00 Marginal contribution (percentage poitns) percentage points per 100 billion MX$ 1,5 Kakwani (index of progressivity) 1,5 0,75 1,0 1,0 0,50 0,5 0,5 0,25 0,0 0,00 0,0 Mayor Mayor Madres Madres Futuro Futuro Futuro Futuro Sembrando Sembrando Construyendo Construyendo Pension Adulto Pension Adulto Superior Superior Educacion Educacion Educacion Educacion Producción para Producción para discapacidad discapacidad Trabajadoras Trabajadoras Basica Escribiendo Basica Escribiendo Jovenes Jovenes Jovenes Jovenes el Bienestar el Bienestar Benito Juarez Benito Juarez Benito Juarez Benito Juarez Transferencia Transferencia Transferencia Transferencia Marginal contribution to poverty reduction Marginal contribution to inequality reduction Kakwani coe cient (right axis) Source: Own calculations using ENIGH 2022 data processed using MEXSIM. Note: The poverty reduction relative to budget (panel to the right) is computed by dividing the budget allocated to a given program as per data from Cuenta Publica for year 2022, divided by the estimate poverty rate change as per MEXSIM (panel to the left). 49 Section 2: Social deprivations in Mexico Some international comparisons Multidimensional poverty measure (MPM), an index devel- erty threshold. The multidimensional headcount ratio is the oped by the World Bank, provides a comprehensive approach primary measure of MPM.66 that aims to present a broader picture of human well-being. Two components of per-capita expenditures in social ser- MPM is a complementary effort to the Multidimensional Poverty vices are included in the World Bank’s multidimensional index: Index (Global MPI), developed by the UNDP and Oxford Poverty education and housing. No health or social security indicator is and Human Development Initiative. Both measures assess the included. The index includes measures of educational attain- crucial aspects of overall well-being, such as access to edu- ment and educational enrollment, as well as access to electrici- cation and basic infrastructure, while the MPM differentiates ty, sanitation and drinking water. The education measures refer from the MPI by incorporating monetary-based measures of to whether at least one school-age child is not enrolled in school, well-being alongside the non-monetary dimensions. and whether no adult in the household has completed primary MPM evaluates the overall well-being by measuring the education. The infrastructure indicators refer to access to a lim- deprivation in three dimensions: monetary poverty, educa- ited-standard drinking water, and sanitation, and electricity. The tion and access to basic infrastructure services. Six different monetary indicator refers to daily consumption income below indicators are aggregated into a single index, assigning equal the international poverty line of US$ 2.15 (2017 PPP). weight to three dimensions, and each indicator is also weighted Although admittedly the MPM considers very frugal thresh- equally within each dimension. A dummy variable is allocat- olds, Mexico has a higher incidence than all comparison coun- ed to each indicator where 1 implies the deprivation, and the tries in each dimension and in the multidimensional indicator. sum of the weighted deprivations is used to construct multi- It also has a faster annual decrease in the incidence in most of dimensional poverty scores. Households are categorized as indicator and the multidimensional indicator (Table 2). The larg- experiencing multidimensional poverty if they are deprived in est difference with respect to comparison countries is seen in indicators whose weights add up to 1/3 or more. Likewise, a access to sanitation, something that will be discussed further in household which is deprived in at least one dimension will be section 3 regarding the territorial aspects of poverty in Mexico. directly considered poor based on the multidimensional pov- 50 MEXICO Poverty and Equity Assessment Table 2: World Bank’s Multidimensional Poverty Index, selected years Deprivation rate (share of population) Multidimensional poverty Educational Educational headcount ratio Monetary Electricity Sanitation Drinking attainment enrollment (%) (%) (%) (%) water (%) (%) (%) 2015 0.02 0.66 0.63 0.62 13.18 1.65 0.21 2018 0.01 1.26 0.69 0.04 0.46 1.43 0.08 Malasya average annual 0.00 0.20 0.02 -0.19 -4.24 -0.07 -0.05 change 2010 0.01 0.15 3.00 0.00 4.06 0.63 0.03 2019 0.00 0.04 0.38 0.00 1.03 0.11 0.00 Poland average annual 0.00 -0.01 -0.29 0.00 -0.34 -0.06 0.00 change 2013 0.30 3.21 4.22 0.00 2.86 0.68 0.58 2021 0.44 3.29 0.00 0.00 3.02 0.31 0.47 Türkiye average annual 0.02 0.01 -0.53 0.00 0.02 -0.05 -0.01 change 2011 0.90 4.53 0.53 0.32 0.45 1.62 0.99 2022 0.40 3.21 0.38 0.29 1.17 1.11 0.46 Chile average annual -0-05 -0.12 -0.01 0.00 0.06 -0.05 -0.05 change 2010 4.70 6.68 3.57 0.78 17.40 5.45 5.73 2022 1.18 3.72 2.67 0.32 10.68 4.35 1.74 Mexico average annual -0.29 -0.25 -0.08 -0.04 -0.56 -0.09 -0.33 change Source: World Bank, Poverty and Inequality Platform, https://pip.worldbank.org/multidimensionalpoverty-measure (retrieved August 3, 2024) The share of the population covered by specific social in- This change in per capita expenditure in public services can surance, social assistance or social development programs be explained by three main forces. First, the average expendi- in general is only part of the story. The sufficiency of these ture in public social service per potential or actual beneficiary; services, when referring to cash transfers, or the quality of the second, the share of these expenditures in public services services, when referring to benefits in kind like public health within the total budget of the government; and third, the nation- and education, are of paramount importance. These elements al income per capita. In other words, inversely, how big is the of sufficiency and quality are closely related to the amount economy, how much is collected in taxes from this economy of resources allocated to each beneficiary. For example, the and how much of these tax revenues is allocated to specific per capita expenditures in public health services in the early social services.67 2000s were similar for Mexico, Chile and Türkiye but by 2021 Again, the lack of growth of the Mexican economy can be Chile and Türkiye spent more than double in health expendi- considered the main culprit of the limited advance in social tures per capita than Mexico (Figure 21, left panel). Poland, that expenditures per capita. Figure 21, right panel, shows that already started with higher health expenditures than Mexico, Mexico’s average annual health expenditures per capita (in now spends three times more. Malaysia, which spent half than US$ in 2017 PPP) grew about 2 percent between 2000 and Mexico in the early 2000s, now spends at par. 2021, whereas it grew at between 5 and 6 percent for all the 51 Section 2: Social deprivations in Mexico comparison counties. The main explanation of this difference land) than Mexico. It is economic growth what mostly explains is the growth in the size of the economy (light blue bar). All the the divergence in health expenditures per capita seen before. other countries had similar growth in tax collection and share A similar analysis would render similar results with respect to of health budget (a little higher in Malaysia, a little lower in Po- education or basic infrastructure. Figure 21. International comparison of public health expenditures Domestic General Government Health Expediture Components of average annual growth of Public Health Expenditure per capita (2000 - 2021) 1.800 7% annual US$ 2017 PPP per capita 1.600 6% 5% 1.400 4% 1.200 3% 1.000 2% 800 1% 600 0% 400 -1% 200 -2% México Chile Türkiye Malaysia Poland - annual growth of GDP/capita (2017 PPP) 2000 2004 2002 2020 2008 2009 2003 2005 2006 2022 2007 2001 2010 2014 2012 2021 2018 2019 2013 2015 2016 2017 2011 annual growth of General Government Expenditure as share of GDP annual growth of Health Expenditure as share Mexico Chile Türkiye Malaysia Poland of General Government Expenditure annual growth of Public Health Expenditure per capita (2017 PPP) Source: Authors calculations using data from World Health Organization, World Health Global Database (https://apps.who.int/nha/database/Select/Indicators/en (retrieved August 3, 2024) 52 MEXICO Poverty and Equity Assessment Section 3: Vulnerability to poverty due to climate events in Mexico 53 Section 3: Vulnerability to poverty due to climate events in Mexico The evolution of the poverty rate is the net result of some er than this threshold. In section 1 of this report, this method households exiting poverty while others suffer a setback and was used to define the size of vulnerable and middle class fall into poverty. This churning was described in section 1 of populations in Mexico. But in section 3 of this report, a special this report with panel data from ENOE indicating that poverty definition of vulnerability is under analysis. reduction in Mexico in recent years has also been accompanied Vulnerability to poverty due to climate events focuses on by a reduction in chronic poverty. But section 1 also indicates identifying the population that may fall into poverty due to that the percentage of the population experiencing poverty at income or assets losses associated with climate shocks. Vul- least one quarter within a year has remained quite stable. This nerability in general refers to any shock that may affect house- indicates that at any moment a still large share of the popula- hold living standards, but now the focus is on climate events tion is likely to fall back into poverty. only. This merits special attention because extreme weather Vulnerability to poverty can be defined as “people who are events are one type of setback that households are facing not considered poor in a specific society but face a non-neg- with increasing frequency because of climate change.69 This ligible chance of falling into poverty.” 68 The fluctuations of requires a specific method that defines the risks associated economic activity (e.g. recessions, inflation) or of household with climate events and how to identify and quantify who are life (e.g., accidents, sickness, divorce) may alter life and live- vulnerable to these risks. The next section will describe two lihoods and, for some groups in particular, these events may approaches to measure exposure and vulnerability to climate lead to poverty. Empirically, this group of people can be identi- change in Mexico and will conclude -like previous section 1 and fied by defining a probability threshold and then using survey 2- with some early international comparison of vulnerability to data to compute the income or consumption level below which climate shocks. a given household has a probability of falling into poverty high- Maps of exposure and vulnerability to extreme climate events in Mexico There are multiple aspects of vulnerability. The concept Identifying the population exposed to severe climate itself has multiple definitions and as many uses in different events relies on using maps that pinpoint if people are located disciplines. This report adopts the concept associated with defi- where these hazards are likely to occur. These maps are pro- nitions of risk as a combination of hazard, exposure and vulner- duced by identifying the percentage of the population exposed ability.70 Identifying who is at risk of climate events requires to extreme events by overlaying a map of hazards to a map of defining the hazards (i.e., the probability of extreme events that population density and other datasets. For each hazard, an ex- may cause an impact on life and livelihoods), who or what is treme event is defined in terms of intensity thresholds and re- exposed (i.e., people or assets located in hazard-prone areas turn periods. A specific intensity level (i.e., a physical measure or activities) and their vulnerability (i.e., susceptibility to harm of how severe the event is, such as wind speed for a hurricane) due to the event and lack of capacity to cope and adapt). and a return period (i.e., the likelihood that the hazard occurs Due to its geo-physical characteristics, Mexico is a country at or above this intensity level, say more than 1 percent in any with multiple climate hazards. Coasts to the Atlantic and the Pa- given year) are set to define a binary category (yes or no) of ex- cific are affected by regular hurricane seasons, its hydrological treme event areas in a gridded map of Mexico. These grids vary map has a north-south unbalance that leaves important parts between about 100 m2 and 10 Km2, depending on hazard. Maps of the territory prone to either floods or droughts. These and of population density, and census data provide population in- other events such as heat waves and landslides have become formation in each grid (about 1 Km2). Population exposed to more frequent partly due to climate change.71 These are the top each extreme event is estimated as the spatial union of hazard five most important climate extreme events because of their maps and population maps at the grid level. These numbers are regularity and size of impact upon Latin American regional pop- then aggregated at the municipality level and compared to the ulation in general, and Mexican population in particular. total population in each municipality.72 54 MEXICO Poverty and Equity Assessment The maps then report the percentage of the population in is less prevalent and more scattered in the territory, concen- each municipality exposed to one or more extreme events. trating in Tabasco, and coastal municipalities of Guerrero and This population can also be separated into poor and non-poor Michoacan (top center). Heat waves affect large shares of the hence producing estimates of exposure by poor and non-poor population in the coastal municipalities of Baja California Norte, population. When the map indicates that in a given municipality Sonora, Sinaloa and Nayarit, on the Pacific coast; Tamaulipas, population is exposed to no event this really means exposed to Veracruz and Tabasco on the Gulf of Mexico; and Quintanaa Roo no event at the intensity and likelihood of thresholds defined in the Atlantic coast (bottom left). Landslides concentrate in above. Hence the label, “below threshold” seen in the legend of the sierras of Queretaro, Puebla, Oaxaca and Michoacan in the maps and tables. Some exposure to climate event hazards exits center of the country; Chiapas in the south and parts of Jalisco almost everywhere, but these hazards are below the pre-de- and Zacatecas to the northwest (bottom center). fined thresholds of “extreme”. Almost the totality of the territory has municipalities with Almost the totality of the Mexican territory faces high per- more than 50 percent of the population exposed to extreme centages of population exposed to extreme climate events. events. Nearly all municipalities in the coasts of the Pacific, Map 1 , top left, shows that more than 50 percent of the pop- Atlantic and Gulf of Mexico have very high percentages -above ulation are exposed to high probability of severe hurricanes in 70 percent- of population exposed to extreme climate events the coastal municipalities of Yucatan and Quintanaa Roo (in the (Map 1, bottom right). However, despite the presence of mul- Atlantic coast), and Baja California Sur, Sinaloa, Nayarit, Jalisco, tiple hazards, most of the territory faces at most one high ex- Colima, Michoacan y Guerrero (in the Pacific coast). High per- posure hazard. centages of the population are exposed to extreme droughts Map 2 illustrates the localization of areas with an overlap in the central plateau of Chihuahua, Durango, Zacatecas and of high incidence of climate hazards and poverty. The legend San Luís Potosí; and plains to the east of Sierra Madre Oriental shows that 38.7 percent of the population (as of census 2020) in Nuevo León and Tamaulipas; as well as in multiple small mu- lives in municipalities with exposure to any hazard above 30 nicipalities in Puebla, Oaxaca and Chiapas (top right). Flooding percent and poverty rates above 10 percent. 55 Section 3: Vulnerability to poverty due to climate events in Mexico Map 1: Percentage of population exposed to extreme climate hazards Source: World Bank (forthcoming) “Climate hazards and poverty in Latin America” 56 MEXICO Poverty and Equity Assessment Map 2: Population exposed to several hazards and poor population exposed Source: World Bank (forthcoming). Adapting to adversity: Climate risks and Poverty in Latin America and the Caribbean. 57 Section 3: Vulnerability to poverty due to climate events in Mexico These extreme events seldom overlap. Exposure to multi- clones is higher for non-poor than poor population: 17.4 versus ple hazards is not prevalent. The percentage of the population 15.4 and 6.3 versus 5.2, respectively. exposed to only one hazard is 31.2 percent, whereas those Vulnerability to extreme climate events, interestingly, also exposed to two hazards is 6.6 percent, followed by a 1 per- affect the non poor. As indicated in previous paragraphs a cent exposed to three hazards. Population exposed to four or large share of the non-poor is also exposed to extreme climate five hazards is below the 0.5 percent mark (Table 3). If count- events; and for some hazards the non-poor are more exposed ed separately, the hazard affecting the most people is heat- than the poor. Households that are not poor are vulnerable to waves. Nearly 21 million people are exposed to heatwaves (a extreme weather events and can be made poor by these events. 16.7 percent of the population in the reference year). This is Households that are poor and exposed can be made poorer be- followed by droughts, which affect 14.7 million people (11.7 cause of a climate shock. There are two approaches to identify percent of the population), landslides with 8.8 million people and quantify vulnerability to climate shocks. One the one hand, (7 percent), floods 7.7 million (6.2 percent) and cyclones 7.5 derived from the literature on losses associated with disasters, million (6 percent). one method consists in estimating the probability and size of Population in poverty is more exposed to droughts and a loss in incomes or assets due to a shock. Those with a loss landslides. The incidence of exposure to these hazards is 13.8 larger than a given threshold are defined as vulnerable. On the and 15.4 percent of the poor population, respectively, in con- other hand, another method identifies the vulnerable as those trast to the 10.7 and 7.0 percent prevalence for non-poor pop- without access to certain assets and services which makes ulation. This is related to droughts affecting mostly rural areas them less able to cope with the impact of climate shocks on and population associated with agricultural activities which, their livelihoods. This report shows estimates of vulnerability in as described in section 1 of this report, are groups with higher Mexico using the former method in this section, and using the poverty incidence. In contrast exposure to heatwaves and cy- latter in the next section on international comparisons. Table 3: Mexican population exposed to climate hazards, c. 2020 Total pop. exposed Share of total Poor and exposed Share of Non-poor and exposed Share of (million people) population (million people) poor (million people) non-poor Below threshold 76.6 61.1 23.6 57.6 53.1 62.8 By type of hazard cyclone 7.5 6.0 2.1 5.2 5.34 6.3 drought 14.7 11.7 5.7 13.8 9.04 10.7 flood 7.8 6.2 2.6 6.3 5.20 6.2 heat 21.0 16.7 6.3 15.4 14.67 17.4 landslide 8.8 7.0 4.3 10.6 4.50 5.3 By number of hazards only 1 39.1 31.2 14.1 34.5 25.0 29.6 cyclone 3.3 2.6 1.0 2.4 2.3 2.7 drought 10.5 8.4 4.1 10.1 6.4 7.6 flood 4.3 3.4 1.4 3.3 2.9 3.5 heat 14.0 11.2 4.2 10.4 9.8 11.6 landslide 7.0 5.6 3.4 8.4 3.6 4.3 any 2 8.3 6.6 2.9 7.0 5.4 6.4 any 3 1.3 1.0 0.4 0.9 0.9 1.1 any 4 0.0 0.0 0.0 0.0 0.0 0.0 any 5 0.0 0.0 0.0 0.0 0.0 0.0 Population totals 125.4 41.0 42.4 84.4 37.2 Source: World Bank (forthcoming). Adapting to adversity: Climate risks and poverty in Latin America and the Caribbean. 58 MEXICO Poverty and Equity Assessment A recent study estimates that about 30 percent of the pop- 30.3 percent in 2022. This national average for 2022 hides a ulation is vulnerable to climate events in Mexico. This study wide gap between urban and rural areas (54.1 and 32.3 percent classifies households as vulnerable if their household income respectively), between agriculture and non-agriculture house- has a probability of falling below the poverty line of at least 50% hold head’s employment (62.0 and 26.3 percent, respective- once or twice in the next two years. This is equivalent to say- ly); and north and south states (above 60 percent in Chiapas, ing that a household is considered vulnerable if the probability Oaxaca and Guerrero and less than 15 percent in Baja California of falling into poverty in any given year is 29%.73 Using data Norte, Baja California Sur, Coahuila, Chihuahua, Nuevo Leon, from ENIGH for several years, this study estimates a decline in Sinaloa, Sonora and Tamaulipas).74 vulnerability to climate hazards from 65.8 percent in 2000 to Evolution of vulnerability to climate events for housing in Mexico Another perspective on vulnerability to climate events can 2010 and 2020, the number of municipalities classified at high focus on housing rather than household incomes. Climate risk of housing floods increased from 50 to 137. Population hazards may affect incomes but may also result in lost assets living in these 137 municipalities amount to 29.2 million peo- and investments that limit income gains in the future. Housing ple, representing 22.9 percent of total population as per 2020 is one of the most important assets that people rely upon for census. For the same period, the number of municipalities with sustaining and protecting their family, promoting human capi- low or very low risk increased from 606 and 124, to 642 and tal accumulation, and often serve as production or distribution 209 respectively ( Table 4 ). These municipalities contain 19.1 premises for family enterprises. Housing damages due to cli- million people, representing 15.0 percent of total population mate events affect family protection but also income genera- as per 2020 census. tion sources. In most severe cases, housing losses may cause The expansion of exposure and frequency of events has migration and, as a consequence, impact of climate events caused increased risks to floods in Mexico in some areas, but spills over from one region to the rest of the country or even improved housing and social conditions have reduced this other countries.75 risk in many others. The 534 municipalities that improved its Estimates of flooding risk to housing help in understanding flood risk score, did so almost entirely because of a reduction this angle of vulnerability to climate hazards. Again, based on of vulnerability: physical -an average reduction of 0.244 out the concept of risk as a combination of hazard, exposure, and of 0.492- and social -0.239 points out of 0.492 (Figure 22, top vulnerability, an index of housing risk related to floods and hur- left). In contrast, the 440 municipalities that worsened its flood ricanes can be computed for all municipalities, using housing risk score, did so because of increased exposure -0.095 out construction materials and location from Mexican censuses of 0.474- and a raise in hazard frequency -0.322 out of 0.474 of 2010 and 2020. This index ranges from 0 to 3 and is the (Figure 22, top right). These results are eloquent. Social devel- addition of three scores (each ranging from 0 to 1) measuring opment in terms of poverty reduction and improvements in hazard as likelihood of the event (as per combination of data on housing infrastructure (both of which reduce vulnerability) are susceptibility and frequency of floods), exposure (measured the main factors behind improvement in this flood risk score. as per density of housing per municipality) and vulnerability Contrariwise, increases in exposure or frequency in events, (physical, as per building materials of the dwelling, and social without reductions of vulnerability (either because poverty as per multidimensional poverty index). The three components prevalence has not changed or housing construction material are equally weighted and changes in the risk score over time have not improved) explain increased flood risk score. can be attributed to changes in each of its components.76 Rural areas have a similar pattern. Actually, most of the There is some churning with some municipalities going up municipalities that experienced a reduction in the flood risk and others going down, but the general trend looks positive: score are rural (362 of 534) but less than half of those who 534 municipalities moved down in the score (meaning a re- see increased risk score are rural (185 of 440).77 The causes of duction of flood risk to housing), while 440 moved up. Between changes in the risk score are similar to the rest of the country ( 59 Section 3: Vulnerability to poverty due to climate events in Mexico Figure 22, bottom panels). Those with a reduced risk score did events frequency, with little or no changes in vulnerability. so because of a reduction in poverty and increases in quality of There are other aspects of living conditions in rural areas, how- housing building materials. Those that see an increased flood ever, that do not sow these signs of progress (see more in next risk score did so because of increased exposure and hazard section on territorial context and Box 3 on rural poverty) Table 4: Classification of municipalities by risk index, 2010 and 2020 Municipios totales 2020 Muy Bajo Bajo Medio Alto Muy Alto Very low 101 19 4 0 0 Low 95 374 124 12 1 2010 Medium 13 237 661 177 30 High 0 12 160 326 73 Very high 0 0 2 15 33 Total Municipalities 2469 Municipalities whose score moved down (lower risk) 534 Municipalities whose score leveled off 1495 Municipalities whose score moved up (higher risk) 440 Poblacion (en millones) 2020 Muy Bajo Bajo Medio Alto Muy Alto Very low 2.5 0.7 0.2 0.0 0.0 Low 1.2 8.7 5.9 1.3 0.0 2010 Medium 0.2 5.4 29.0 18.2 3.1 High 0.0 0.3 5.2 18.4 13.2 Very high 0.0 0.0 0.1 1.0 12.8 Total population 127.4 Population whose score moved down (lower risk) 13.4 Municipalities whose score leveled off 71.4 Population whose score moved up (higher risk) 42.6 Source: Authors calculations using data from (Alonso Beltrán, et al. 2024). 60 MEXICO Poverty and Equity Assessment Figure 22: Components of changes in housing flood risk score in Mexico, 2010-2020, total and rural areas Areas with declining risk (534 municipalities) Areas with increasing risk (440 municipalities) 0.0 0.5 -0.010 0.001 0.058 0,474 0.5 -0.001 0.095 -0.1 0.4 0.4 Change in flood risk score Change in flood risk score -0.2 0.322 0.3 -0.239 0.3 -0.3 0.2 -0.4 0.2 0.1 -0.5 -0.244 -0.492 0.1 0.0 -0.6 hazard exposure social physical TOTAL hazard exposure social physical TOTAL vulnerability vulnerability vulnerability vulnerability Rural areas with decreasing risk score (362 municipalities) Rural areas with increasing risk score (185 municipalities) 0.0 0.5 -0.003 -0.006 0.5 0.019 0,431 0.092 -0.1 -0.003 0.4 0.4 Change in flood risk score Change in flood risk score -0.2 0.323 -0.185 0.3 -0.3 0.3 0.2 -0.4 0.2 0.1 -0.5 -0.297 -0.491 0.1 -0.6 0.0 hazard exposure social physical TOTAL hazard exposure social physical TOTAL vulnerability vulnerability vulnerability vulnerability Source: Authors calculations using data from (Alonso Beltrán, et al. 2024) A policy focus on reducing vulnerability will render lower in other forms of public infrastructure such as rural roads and risks in areas that face high hazard exposure. The previous ev- sanitation, that help coping with climate hazards will render a idence from changes in flood risk score underlines the impor- double impact on poverty reduction (through cutback of caren- tance of this policy. Particularly in areas where, as explained cias sociales associated with quality housing) and poverty vul- in the previous section, there is higher likelihood of climate nerability due to climate hazards. related hazards. Investments in improved housing, but also 61 Section 3: Vulnerability to poverty due to climate events in Mexico A note on climate events, infrastructure and territorial context of poverty in Mexico Climate hazards have a specific geographic localization. The percent of localities have low or very low access to paved roads previous section has used this local specificity in combination and frequent transport, representing about 5 percent of the with socio-economic information to identify population vulner- population. Again, these concentrate in municipalities of Sierra able to climate hazards and measures of flood risk on housing. Madre Occidental, and sierras of Guerrero, Oaxaca and Chiapas. Similarly, Mexican authorities have underlined the importance These are located in isolated or desertic terrain which make it of a territorial context in poverty measurement. As indicated at difficult the extension of this infrastructure. On the other hand, the beginning of section 2, in addition to monetary poverty and nearly 90 percent of the population live in localities with very social deprivations, multidimensional poverty measurement in high access to paved roads.81 Mexico also includes a territorial context. These measures help Another crucial infrastructure to protect living conditions in understand the geographic dispersion of monetary poverty or the face of climate events are sanitation and electricity. These social deprivation, and may guide public policy in locating spe- should guarantee access to clean water, air conditioning or food cific policy programming and design.78 preservation in the case of climate hazards. These are two key Some measures of the territorial context of poverty in Mex- components of the carencia social defined as lack of access to ico indicate vulnerability to climate hazards only in a few spe- public utilities. In section 2 of this report it was described that cific areas of the country. Vulnerability to climate hazards is this is one of the carencias affecting a sizeable percentage of associated with lack of physical and social assets that reduce the population, particularly among households of low income the propensity of loss or the ability cope with losses. Measures (42 percent for those in the first income quintile, as of 2022 of vulnerability include indicators of access to sanitation or Figure 16 top right) and particularly among rural households transportation infrastructure as well as health and financial ser- (see Box 3). But these national averages hide wide differenc- vices.79 These are of special importance because they indicate es across the territory. According to Longitudinal Social Gap the capability of people affected by a climate shock to at least Index, 19.4 percent of dwellings in Oaxaca lacked sewerage, partly recuperate their life and livelihood conditions. According followed by Guerrero and Chiapas with 11.8 and 8.5 percent, to official sources, there are 602 municipalities (out of 2469 respectively, whereas less than 1 percent do in Aguascalien- total) that do not have access to financial services (i.e., puntos tes, Colima, Mexico City, Jalisco or Nuevo Leon.82 If focusing on de acceso, corresponsales or sucursales, in Spanish).80 These municipalities the dispersion is even starker: 10 percent of the concentrate in the states of Oaxaca, Chiapas, Puebla and Yu- municipalities (most of them in Oaxaca, but many in Guerrero, catan which, as shown before, are also areas of high exposure Veracruz, Chiapas and Chihuahua), where 2.5 million people to drought and landslide hazards. In terms of health services, live, have more than 40 percent of dwellings without sewerage. about 10 percent of localities, with a population of half a million It should be acknowledged that there’s been a lot of prog- people, would take 30 minutes or more by car to reach a first ress about infrastructure in Mexico over the past two decades. level health care center. This concentrate near coastal areas The municipal average percentage of houses without sewerage of Michoacan and Guerrero, as well as municipalities in Sierra fell from 49.5 in 2000 to 13.8 in 2020. The percentage of hous- Madre Occidental. If the distance is to a 2nd level health care es without piped water declined from 27.3 to 6.0 percent; those center, then the numbers are 95.2 thousand localities and 68.9 without sanitation from 24 to 4 percent. Those without electric- million people (about 50 percent of total in each case). Most of ity from 13.8 to 1.8 percent, making it practically of universal the Mexican territory outside medium and large urban centers access. This has led to most of the municipalities improving would fall in this category. their Longitudinal Social Gap Index. In 2000 around 20 million A key element to have access to health, financial or oth- people lived in localities with medium, high or very high social er services for socio-economic development is the quality of gap, but by 2020 the number declined to about 2 million. The transportation infrastructure. According to CONEVAL about 40 analysis of these data indicates that all the localities with a high 62 MEXICO Poverty and Equity Assessment or very high social gap index in 2020, are rural. And about two for these areas which concentrate in the states of Chihuahua, thirds of them have been catalogued as high or very high social Chiapas, Guerrero, Durango and Oaxaca. gap since 2000, indicating a serious chronic poverty condition Box 3: Rural poverty and territorial context The most important territorial aspect of poverty in Mexico has to do with rural poverty. Under the World Bank’s upper-middle income country poverty line of US$ 6.8583, the rural poverty rate in Mexico declined from 81.4 percent in 2000 to 61.1 percent in 2006, from 68.6 to 62.2 percent between 2008 and 2014, and from 53.4 to 35.8 percent in 2008 and 2022 (the three comparison periods postulated in section 1 of this report). It is certainly a rapid decline for all three periods, but rural poverty has kept an incidence of about double the poverty rate in urban areas for all these years (35.8 and 17.3 percent in 2022). Given its stable share of the total Mexican population, rural poverty has also kept a relative size of about 40 percent of total poverty for the period of analysis. The rural/urban divide partly explains the persistent territorial gap in poverty rates across states (entidades federativas). Southern states like Chiapas, Guerrero and Oaxaca had poverty rates above 65 percent in early 2000s and now range between 40 and 50 percent in 2022. On the other hand, northern states, like Baja California Norte y Sur, Nuevo Leon and Ciudad de Mexico went from poverty rates around 20 percent back then, to rates below 10 percent in 2022. Over the years, however, there has been convergence and most states with high poverty rates have experienced faster poverty reduction (absolute and relative). Labor income is the most important source of income in terms of poverty reduction in rural areas. In this sense, it is not very different from the rest of the country: both areas have growth in employment and average wages explaining half or more of the change in poverty reduction. But the difference seems to be in terms of stability of labor incomes. Using rotating panels from the ENOE, rural workers are more likely than urban workers to experience labor income poverty for more than one quarter in a year. Chronic labor poverty defined as being poor for five consecutive quarters affects 22 percent of rural workers, but 12 percent of urban workers as of 4th quarter of 2023. Transitory poverty defined as being labor poor at least one of five consecutive quarters affects 49 percent of rural workers, and 44 percent of urban workers. This difference may be the results of a concentration of agricultural employment -which is highly seasonal- in rural areas: about 50 percent of employment in rural areas is dedicated to agriculture. But the difference is not only due to agricultural activity. The median labor income of a worker in agriculture in Chiapas, Guerrero, Puebla or Oaxaca is between 40 and 50 percent of the median income of an agriculture worker in Nuevo Leon, Coahuila or Chihuahua.84 Northern states are more characterized by capital-intensive, export-led agriculture, whereas the south by small- area, subsistence agriculture. There is a vast difference in terrain, location to markets, production systems and investment.85 But there are also important differences in social infrastructure. Rural areas suffer much deeper carencias sociales. For the period 2016-2022, the rural/urban divide in terms of education gap, access to social security, quality of housing and food security, range between 15 and 30 percentage points and have remained unchanged. The two other carencias merit 63 Section 3: Vulnerability to poverty due to climate events in Mexico special attention. Social deprivation due to lack of basic utilities (i.e., sanitation and electricity) has a 40 percentage points divide between urban and rural areas. Only 50 percent of rural population has access to these utilities. Access to health services, had almost no gap between urban and rural areas (it was about 15 percent of population on both areas), but the phasing out of Seguro Popular has led to an increase in this deprivation, much faster in rural than urban areas: about 55 and 35 percent, respectively, in 2022. Consequently, poverty in terms if carencias sociales is much higher in rural areas. As of 2022, nearly 90 percent of the rural population has at least one social deprivation, while almost 50 percent have at least three carencias. The deterioration of access to health services has led to an increase in rural population with at least three deprivations (from 41.6 to 49.1) between 2016 and 2022. This contrasts with the period 2008-2014 when the rural population with three carencias declined (from 62.4 to 46.0 percent). Another social aspect of importance in rural areas, is the presence of indigenous communities. About 20 percent of the rural population is indigenous (defined as people who live in a household where the household head, her/his spouse and or any ancestor, father, mother, grandfather or grandmother declares to speak an indigenous language). And nearly half of the indigenous population live in rural areas. About 40 percent of workers of indigenous identification work in agriculture. Many indigenous groups locate in areas that have been identified as high-risk for climate hazards and or a limited territorial development. Moreover, indigenous population have endured social exclusion that has limited their participation in social and economic development.86 All these factors are associated with indigenous peoples having double the poverty rates than non-indigenous peoples (32.1 versus 17.1 percent in 2022, under the US$ 6.85 poverty line). Nearly 90 percent have at least one carencia social, and 60 percent have at least three (2022). All these factors underline the importance of a territorial perspective in anti-poverty policy, with a special focus on rural areas. Population in these areas face higher poverty rates, often chronically, with severe gaps in social and economic infrastructure, and with an important representation of indigenous population who have traditionally faced social exclusion hindering their economic development. In addition, rural population rely more on agricultural activities and are located in areas of growing climate hazard, which makes them more vulnerable, in general, and to the potential impact of climate change. Poverty eradication in Mexico obligates a major focus on rural areas. Some international comparisons Country by country comparisons of exposure and vulner- nancial assets, access to electricity, etc.; and geographical data ability to climate change are not produced regularly yet. Most on population density and likelihood of climate events at spe- studies have a regional and global scope. However, recent cific geo-referenced grids of global territory. This method allows research, allows for a comparable count of people exposed to count people exposed and vulnerable, based on proxies for and vulnerable to risks associated to climate shocks across high vulnerability associated with extreme weather events.87 countries. The method consists of combining country-specific The international comparison here relies on counting the household survey data with information about indicators asso- number of people at high risk from climate disasters. It uses the ciated with vulnerability such as education, social security, fi- IPCC definition of risks as the combination of hazard (i.e., the 64 MEXICO Poverty and Equity Assessment likelihood of an extreme event), exposure (i.e., people in harms comparison countries, a higher percentage of those exposed way because of the hazard) and vulnerability (i.e., the propen- in Mexico are also in poverty: 14.6 million in Mexico, versus be- sity of these people to be adversely affected). The measure tween a quarter million and three million in comparison coun- includes four hazards (hurricanes, heatwaves, droughts and tries. Namely, nearly one in three exposed to climate disasters floods), which are defined as high-risk under specific intensity are also in poverty in Mexico ; whereas less than 10 (or even and likelihood thresholds. Vulnerability is measured in terms 5) percent in comparison countries. of the ability to have access to clean water and cooling appli- In terms of vulnerability, it also ranks lower than compar- ances (so-called physical vulnerability), and to have ability to ison countries. A little more than 30 percent of Mexicans are recover through education, social security, financial services exposed and vulnerable to at least one type of climate disaster; or incomes, to confront the consequence of a shock (so-called whereas comparison countries have rates ranging between 4.5 inability to cope). Measures consist in counting number of peo- and 17.7 percent. This higher rate of vulnerability, seems to be ple in a given territory, facing at least one hazard, and failing associated with higher rates of inability to cope, rather than to to have access to at least two of the vulnerability indicators.88 physical vulnerability. Although Mexico shows higher numbers In comparative terms, Mexico appears to have higher expo- and percentages of population vulnerable under each dimen- sure to climate-related disaster events. This higher risk is due sion, the differences seem much higher in Social Protection and to several factors. Mexico, due to its geographic characteristics Access to Finance. About 24 percent of the population is ex- has a higher percentage of the population exposed to climate posed to disasters and lack social security in Mexico, whereas events: 38.4 percent, versus exposure rates between 18.9 and only 1.4 percent are in Chile. Nearly 20 percent of Mexicans are 35.1 percent of the population for the comparison countries exposed and lack access to finance, whereas only 1.5 percent (Table 5). Moreover, because of its higher poverty rate than are in Poland.89 Table 5: Risk, exposure and vulnerability to poverty due to climate events Population exposed and Share of Poor people exposed Population Population exposed and vulnerable on the following dimensions (million) vulnerable on any one population (million) exposed dimension (million) exposed to (million any shock people) Social (%) US$ 2.1 US$ 6.85 Income Education Finance Water Electricity (million) (% of total) Protection Chile 4 18.93 0.02 0.25 0 0 0 0 0 0 1 5.1% Poland 13.48 35.13 0.05 0.35 0 0 5 1 0 - 6 15.8% Malaysia 7.62 23.45 0.00 0.28 0 0 2 1 0 0 3 9.0% Türkiye 23.14 28.27 0.08 2.92 0 1 12 6 0 - 15 17.6% México 48.18 38.36 1.35 14.64 3 2 30 24 2 0 41 31.6% Source: Doan et al. (2023), table A3 This comparative evidence of higher risk to climate disas- prevent poor and non-poor to fall into poverty due to disasters, ters in Mexico reiterates the importance of policies enhancing whose growing occurrence due to climate change pose one social protection and inclusive growth, but also a territorial per- of the newest challenges to poverty reduction in the country. spective for policy action. This was mentioned in sections 2 and This section 3 adds a territorial perspective and underlines the 1 before, but now, not only because of the influence these re- benefits of investment in infrastructure to cope with climate forms can have upon poverty reduction, but also on forestalling hazards with a double impact in terms of reduction of poverty vulnerability to poverty due to climate disasters. These policies and vulnerability. 65 Section 4: What would it take to eradicate extreme poverty in Mexico by 2030? Section 4: What would it take to eradicate extreme poverty in Mexico by 2030? 66 MEXICO Poverty and Equity Assessment Mexico should be able to eradicate poverty in the short ularly. Although this target is not within reach at global level, term. This is an objective that can be defined with a precise it is certainly a valid target for specific countries like Mexico.90 time frame, specific targets and concrete policy actions. The In this report, we have used the US$ 6.85 (2017 PPP) poverty international experience shown in previous sections indicates line, which is a little higher than the extreme monetary poverty that Mexico is underperforming in terms of reducing poverty, line in Mexico. We have also used the social deprivations and vulnerability or coverage-plus-quality of social services when reported CONEVAL’s estimates of population with at least one contrasted to selected countries. Even if compared more wide- and at least three carencias sociales. This report postulates a ly (for instance with all upper-middle income countries in the target of extreme poverty eradication as less than 3 percent world), Mexico has a higher than expected poverty rate. Never- of the population with household income below the US$ 6.85 theless, Mexico is a country with abundant insight and experi- (2017 PPP) poverty line and three or more carencias sociales.91 ence in terms of macroeconomic management, social policy The following subsections will include an examination of design and disaster risk prevention. Mexico can put these tal- what it would take. First, a set of projections of the needed ents and resources to eradicate poverty in the short term. changes in economic growth and income inequality that would What would it take to eradicate extreme poverty in Mexico lead to substantive reduction of monetary poverty rates. Sec- by 2030? This sentence includes a precise time frame (year ond, aspects of social policy design that would make existing 2030) a specific target (eradicate extreme poverty) and calls social programs more effective and efficient in terms of mon- for concrete policy actions (what it would take). The 2030 time etary poverty reduction, as well as the areas of concentration frame is one that has been often used by the World Bank. It in social policy that would lead to no family living with three or is the target of Sustainable Development Goal 1 (eradicate more carencias. Third, a brief review of existing programs that global poverty by 2030) which the World Bank monitors reg- can reduce vulnerability to climate hazards. How much growth is needed? The evidence shown in section 1 underlines that further and then forecasts average household incomes using the GDP/ poverty reduction in Mexico requires faster economic growth. capita growth in year 2023 (2.5 percent), the most recent pro- Reducing inequality has certainly played a role, in Mexico and jection for 2024 (1.6 percent), and two hypotheses about its in other countries, although its effect has not been predom- growth between years 2025 and 2030. One exercise adopts a inant. Growth seems to be the main driver of poverty reduc- 2 percent annual growth of GDP/head and another at 3 percent tion. But a favorable combination of faster growth and lower annual growth. In addition to these growth assumptions, alter- inequality, often dubbed inclusive growth, accelerates pov- native scenarios with annual changes in the Gini coefficient of erty reduction. Hence, the following projections will consider 1 and 2 percent, are also included.92 changes both in growth of the average incomes and changes Growth alone will bring poverty reduction, but not enough in its distribution. to eradicate extreme poverty by 2030. A growth rate of 2 per- A method to produce these projections has been devised cent per year, for all income groups the same, would lead to by World Bank staff, and is used for the World Bank regular a poverty rate of 15.4 percent by 2030. A faster growth of 3 monitoring of worldwide poverty. In this case, the average per percent per year, would lead to a 13.4 percent poverty rate (Fig- capita income is assumed to grow as fast as GDP/head. A “neu- ure 23).93 Both will be notable achievements, because they will tral” distribution, assumes that every individual in the sample bring similar poverty reductions than the growth component of grows at the same rate: the rate of growth of GDP/head. This poverty declines in the past (4.5 and 6.5 percentage points re- implies no change in inequality over time. An inequality-sen- spectively, versus 6.3 and 5.3 percentage points in the periods sitive distribution implements some assumptions about how 2000-2006 and 2016-2022, as reported in Table 1). inequality changes affect the income growth of each person in Inclusive growth is needed to eradicate poverty by 2030. the sample. The projections in this section start with the pover- Projections that combine average growth with reductions in ty rate in 2022 (21.8 percent under the US$ 6.85 poverty line), the Gini coefficient, lead to poverty rates closer to the poverty 67 Section 4: What would it take to eradicate extreme poverty in Mexico by 2030? eradication goal. A 2 percent growth with a 1 percent annual fall growth of 3 percent with 1 percent fall in Gini ends with an 8.1 in Gini render a poverty rate of 9.8 in 2030, but if Gini declined percent poverty rate, but with a faster 2 percent decline in Gini by 2 percent the poverty rate would reach 4.3 percent. Faster would lead to a poverty rate of 3.2 percent in 2030. Figure 23. Monetary poverty rate projections through 2030 30 30 Poverty rate under US$ 6.85 Poverty rate under US$ 6.85 25 25 20 20 (2017 PPP) line (2017 PPP) line 15 15 10 10 5 5 0 0 2022 2024 2026 2028 2030 2022 2024 2026 2028 2030 Neutral distribution 2% growth in Gini 1% growth in Gini Neutral distribution 2% growth in Gini 1% growth in Gini 1% decline Gini 2% decline Gini 1% decline Gini 2% decline Gini Source: Own calculations using ENIGH 2022 and code from Lakner, et al. ( 2022). Note: The left (right) panel assumes an annual GDP/head growth of 2 (3) percent. What is the order of magnitude of these projections? A 2 How plausible are these projections? The median GDP/head percent in economic growth would lead to a Mexican GDP/head growth for the period 2002-2022 among upper middle income of US$ 23,732 (2017 PPP, similar to Chile’s in 2014) and a 3 countries has been 2.2 percent; and the median decline in the percent growth would end in a GDP/head of US$ 27,720 (sim- Gini coefficient for this group of countries has been 0.9 percent. ilar to Chile’s in 2022 and Malaysia’s in 2021). A one percent Hence a 3 percent GDP/head growth and a 2 percent Gini decline decline in Gini, starting from 0.435 as per ENIGH 2022, would for Mexico -which would push monetary poverty to 3 percent by lead to a Gini of 0.401 in 2030 (similar to the Gini of Malaysia 2030- would be above the medians for this group of countries. A in 2021) and an annual decline of 2 percent would end in a 2 percent GDP/head growth and a 1 percent Gini coefficient de- Gini of 0.370 (similar to the Gini of Poland in 2003). A 1 per- cline are pretty close to the group’s medians, and would lead to cent decline in Gini has been the average experience of Mexico more than halving the current poverty rate -that is from 21.8 in since 2000, and there are six-year periods with 1.3 percent per 2022 to 9.8 percent in 2030. Namely, a performance similar to year Gini declines (the reference periods of 2000-2006 and past experience of upper middle-income countries would halve 2016-2022). But, as indicated in sections 1 and 2, economic the poverty rate; but an accelerated pattern (still within the range growth has been meager during the past two decades in Mexi- of observed marks of upper-middle income countries)94 would co. Achieving substantially higher growth rates, and sustaining lead to a reduction of the extreme poverty rate to close 3 percent. reductions in inequality as growth accelerates, is the main chal- These are the bounds for significant poverty reduction in Mexico. lenge to significant poverty reduction in Mexico. The goal requires important economic and social policy efforts. How to increase the poverty reduction impact of social public expenditures? The evidence described in section 2 has shown that public locations to cash transfers (e.g. scholarships, subsidies) and expenditures allocated to social policy have increased in recent in-kind transfers (e.g., public services in education and health) years. This increase is particularly steep in funding allocated have been relatively stable. There have been important chang- to pensions (both contributive and non-contributive), while al- es in the design of programs through the introduction of new 68 MEXICO Poverty and Equity Assessment ones (like Becas Benito Juárez, Jóvenes Construyendo Futuro, ple could be the Programa de Adultos Mayores which, as shown Sembrando Vida, Produccion para el Bienestar) and phasing in section 2, it has an almost neutral absolute incidence; name- out others (like Oportunidades, and Seguro Popular). ly, its budget is allocated almost equally across deciles of the Mexico has extensive experience in the design and imple- income distribution. This means that there are beneficiaries of mentation of social programs. As shown in section 2, most of the program in middle and upper class groups. In these groups them are well targeted and benefit the poorer groups of the there are people who receive the Programa de Adultos May- population. However, in some cases, the coverage of certain ores and some other pension benefit as well. Reallocating part vulnerable groups is limited (e.g., working mothers), in others of the budget from this program to other programs, or creating the budget and sufficiency of the program is limited (like in the different size of benefits according to different size of need of case of Becas Benito Juárez), and in others the program is not the beneficiary to reach the poverty line, could lead to more focused on vulnerable population hence eroding its budgetary efficient poverty reduction. efficiency, although still be very effective in poverty reduction What about in-kind transfers in terms of public services? (e.g. Programa de Adultos Mayores). The extensive experience The previous paragraphs mostly refer to cash transfers and of the country in devising, implementing and renovating social subsidies. These certainly affect monetary poverty, and may programs opens the opportunity of adjusting existing programs also influence social deprivations like educational gap or ac- in order to increase its effectiveness and efficiency towards cess to social security. But the largest components of social poverty eradication. policy are education and health. Where should social policy Changes in the incidence and sufficiency of some programs reform concentrate to eradicate extreme poverty? may free resources to increase the impact of others. An exam- Figure 24. Impact of closing one social deprivation gap on population with at least three deprivations. 0,0% -2,0% -4,0% -6,0% -8,0% -10,0% -12,0% -14,0% Average Access to Access to Quality and Access to Access educational Health Social space of basic services to Food gap Services Security dwelling in housing Source: Own calculations using ENIGH 2022 and MEXSIM Note: The figure illustrates the change in percentage of the population with at least three carencias, given the full closing of one the social deprivations gaps. If one concentrates in extreme poverty in terms of caren- cess to one of the six dimensions), what would be the impact cias, then social policy reform should focus on two of them: upon the number of people living with at least three carencias access to social security and access to health services. As ex- ? Figure 24 shows that closing the access to social security plained in section 2, these two carencias are the most prevalent gap would reduce the population with three or more carencias and affect around 50 and 40 percent of the population, respec- by nearly two thirds (i.e., 13.5 out 24.9 percentage points). tively, as of 2022. It is clear that extension of coverage in these Closing the gap in access to health services would reduce two is of importance to eradicate extreme poverty in terms of the share of population with three or more carencias by half carencias. A simple simulation exercise would illustrate this. (12 out of 24.9 percentage points). The other four carencias Assuming, hypothetically, that by certain policy action one of would have a smaller impact: about 25 percent in the case of the social deprivations is fully closed (i.e., nobody lacks ac- educational gap, and less than 15 percent the other three. This 69 Section 4: What would it take to eradicate extreme poverty in Mexico by 2030? doesn’t mean there should be no efforts in closing the other of their parameters to secure their long term financial sustain- four carencias, but it means that closing the gaps in access to ability and also partly serve as an instrument of redistribution health and social security play an indispensable role in eradi- to avoid poverty among the elderly. cating extreme poverty in the short term. In the case of education, it is also important to underline Thorough reforms of these systems should aim to full cov- that guaranteeing access through scholarship and grants would erage of the population and similar package of services across not suffice. The various programs that distribute scholarships different groups. The pension and health systems are frag- certainly aim to enable school enrollment at various education mented in terms of quantity and quality of service. Section 2 levels but if the quality of education does not increase, mere described the disparities in incidence of public expenditures access would not do. If students in primary do not get quality across the different health systems (IMSS, ISSTE and INSABI) education, they will not be able to finish secondary. If quality in which renders very unequal services across population groups. secondary is deficient, even with additional scholarships and Something similar occurs in terms of the pensions system. In grants, students may fail to finish tertiary education because the case of health, a universal health coverage program funded insufficient fundamental study skills or academic knowledge. with proceeds from taxation other than labor taxation to prevent This is of key importance because the demographic trends in disincentives to work in the formal sector is one avenue for re- Mexico are increasing the demand for secondary and tertiary form. In the case of pensions, the existing four pillars (non-con- education. Investing in quality basic and secondary education tributory, benefit-defined contributory, contribution-defined is also an investment in access to higher levels of education. contributory and voluntary savings) may be adjusted in some How to reduce vulnerability to climate events? Section 3 has described that there is a large share of pop- are areas of the territory that still lack sufficient access to fi- ulation exposed to climate-related hazards and vulnerable nancial services, or transportation services to primary health to poverty because of these hazards. It is estimated -using centers. Rural areas have a wide gap in access to basic utilities: census data for 2020- that 31.2 percent of the population in nearly 50 percent of the rural population has a social depriva- Mexico is exposed to at least one type of severe climate event, tion in access to basic services for housing. and an additional 7.7 percent to at least two types of high in- It is then clear that investments in infrastructure in areas of tensity climate events. Only about one third of these exposed the territory characterized by high exposure to climate events are in poverty, so there is an important share of the non-poor and social deprivations related to basic services, would have population who is exposed to climate-related high-risk events. a double beneficial impact. On the one hand, it would reduce There are studies that estimate that at least one third of the vulnerability to climate events by providing the physical infra- Mexican population is vulnerable to fall into poverty because structure to cope with the consequences of these hazards.96 of climate-related events.95 Therefore, public policy needs to On the other, it would reduce extreme poverty as defined by take action not only to reduce poverty, but also to reduce vul- social deprivations, particularly in rural areas where these are nerability to poverty. so prevalent. In recent years, public expenditures in rural roads Vulnerability refers not only to potential losses of income declined from 15 billion pesos in 2014 to 8 billion in 2023. Pub- and consumption, but also of productive assets such as hous- lic expenditures in sanitation, water and sewerage went from 8 ing, which means that these losses may last and affect stan- billion pesos in the late 2000s, to less than two billion pesos in dards of living in the medium- and long-term. Section 3 also 2019, although it has seen a rapid increase since to reach 21 described that, despite progress in some areas, there are other billion pesos in 2023.97 areas that have seen an increase in the risk associated with Expanding coverage of social protection is also needed, housing losses due to floods and hurricanes. These are areas but a new version of it: adaptive social protection. As already characterized by growing occurrence of this type of events and mentioned in the previous sub-section, expanding access to limited investment in housing materials improvement. There social security and health services is the most effective way 70 MEXICO Poverty and Equity Assessment to reduce the number of households with at least three caren- the possibility of improved targeting of the use of these funds cias sociales. International comparisons also indicate that this through direct payouts to the poor and vulnerable, which could is one of the areas where Mexico falls short when compared expedite the distribution of resources and improve the effec- to other countries. But it is important to understand that so- tiveness of relief and adaptation after a shock. This requires, cial protection programs tend to be reactive, rather than pro- however, clear identification of beneficiaries, location/type of active. This is particularly important to address vulnerability shock, and size of the required transfers, as well as intense to climate-related events. On the one hand, social assistance inter-institutional coordination, for careful design. In this regard, programs may not be appropriate because they target the poor, the programs of adaptive social protection need to be “predict- not the vulnerable to poverty due to a climate shock. On the able, reliable and tailored”.100 Building a social registry to iden- other hand, social insurance programs might not include cli- tify those vulnerable to climate change and to facilitate quicker mate risks as triggers to specific benefits or may be insufficient responses to shocks is a first step in this direction. to cover growing risks due to more severe climate events or Mexico has experience in several instruments of disaster due to the changing conditions that mitigation and adaptation risk financing from risk retention to risk transfer. In terms of risk policies -in Mexico or in the rest of the World- may bring upon transfer instruments, (e.g., regular insurance, parametric insur- workers and their livelihoods. There are some experiences of ance and catastrophe bonds) Mexico also has experience with these programs reacting and responding to shocks, but these several catastrophe bond such as the MultiCat Mex 2012-2015 experiences have varying degrees of success and underline bond was partially activated in 2015 due to the damage caused the importance of thinking and implementing of a more sys- by Hurricane Patricia and more recently the 2020 parametric tematic design of social protection systems to confront vulner- catastrophe bond worth $485 million that provides earthquake ability to climate events.98 and hurricane disaster protection. Risk retention instruments Adaptive social protection concentrates on building resil- assume certain levels of risk to be assumed with government ience to hazards by designing mechanisms to cope before, resources, through regular budget allocations, intermittent during and after these hit. Adaptive social protection “helps to post-shock budget reallocations, and disaster funds. Mexico’s build resilience of poor and vulnerable households by investing General Civil Protection Law and the until-2021 operating FON- in their capacity to prepare for, cope with, and adapt to shocks DEN101, are examples of regular budget allocations to these this protecting their wellbeing and ensuring that they not fall into type of instruments. These funds have seen irregular allocation poverty as a result of the impacts”. This requires identifying the in the federal budget, and an important decline in the last three existing or new programs to be redesigned to confront climate years.102 Mexico has experience in the various instruments of risks, collect the necessary data about peoples, localities and disaster risk finance and innovative social protection program. their relation to these events, select or create the institutional It should put this valuable knowledge into the design of an arrangement to implement the new or reformed programs, and adaptive social protection system to cope with the growing secure sufficient and sustainable financing.99 Mexico has expe- occurrence of climate-related hazards and protect its sizeable rience and capacity to advance these policy actions. population vulnerable to these risks. A key element of adaptive social protection is to link it to existing or new sources of disaster risk finance. Some suggest 71 References References Abel, M., E. Carranza, K. Geronimo, y M.E. Ortega. 2022. «Can VAT Changes on PricesL Evidence from Mexico.» Labour Temporary Wage Incentives Increase Formal Employment? Economics 80. Experimental Evidence from Mexico.» IZA Discussion Pa- Campos Vasquez, R., y N. Lustig. 2019. «Labour income in- per Series 15740. equality in Mexico: Puzzles solved and unsolved.» Journal Alonso Beltrán, S., D. Cruz Rubí, A. Montes León, y J. Soriano of Economic and Social Measurement 44: 203-219. Escobedo. 2024. Housing Vulnerability and Flood Hazard. Campos-Vasquez, R., y G. Esquivel. 2023. «The effect of the Min- background paper to regional study Climate and Poverty imum Wage on Poverty: Evidence from a Quasi-Experiment in Latin America, SARC, Consultoría en Análisis de Riesgo, in Mexico.» The Journal of Development Studies 59 (3). Mexico City: Mimeograph. Campos-Vázquez , R., E. Chávez, y G. Esquivel. 2018. «Estimat- Alvarado Pérez, R., P. P. Orraca Romano, y F. Cabrera-Hernández. ing Top Income Shares Without Tax Return Data: MExico 2023. «El efecto de duplicar el salario mínimo en la brecha Since the 1990.» Latin American Policy 9 (1): 139-163. de género en empleo y salarios en México.» El Trimestre Canavire-Bacarreza, G., y L. Gayoso de Ervin. 2024. «House- Económico XC (4): 961-999. hold’s vulnerability to Poverty and Natural Hazards in the Azevedo, J.P., V. Sanfelice, y M.C. Nguyen. 2012. «Shapley face of Climate Change in Mexico.» Mimeograph. Decomposition by Components of a Welfare Aggregate.» Canavire-Bacarreza, G.J.; Conconi, A.; Olivieri, S.D. ; Serio, M. MPRA Paper. https://mpra.ub.uni-muenchen.de/85584/. (2024) Understanding Vulnerability to Poverty and Nat- Banco Mundial. 2020. La participación laboral de la mujer en ural Disasters in Latin America and the Caribbean. Policy México. Washington, DC: Banco Mundial. Research working paper ; no. WPS 10825; Washington, D.C. Bergstrom, Katy. 2022. «The Role of Income Inequality for Pov- Carter, M., y M. Martínez Sugastti. 2024. «Shock-responsive so- erty Reduction.» The World Bank Economic Review 36 (3): cial protection in Latin America and the Caribbean: What we 583-604. know and what we need to learn.» background paper to Bhalotra, S., y M. Fernández. 2024. «The Rise in Women’s La- Regional Report on Climate and Poverty in Latin America bor-Force Participation in Mexico. Supply vs. Demand Fac- mimeograph. tors.» The World Bank Economic Review 38 (2). Chaudhuri, S. 2003. Assessing vulnerability to poverty: con- Birdsall, N., F. Ferreira, L.F. López-Calva, y J. Rigolini. 2011. The cepts, empirical methods, and illustrative examples. Middle Class in Developing COuntries: concept, measure- Mimeo, New York, New York: Department of Economics, ment and recent trends. Mimeo, Washington DC: World Bank. Columbia University. Bowen, T., C. del Ninno, C. Andrews, S. Coll-Black, U. Gentilini, K. Chort, I., y M. de la Rupelle. 2022. «Managing the impact of Johnson, Y. Kawasoe, A. Kryeziu, B. Maher, y A. Williams. climate on migration: evidence from México.» Journal of 2020. Adaptive Social Protection: Building Resilience to Population Economics 35: 1777-1819. Shocks. International Development in Focus, Washington CONEVAL, Consejo Nacional de Evaluación de la Política de DC: World Bank. Desarrollo Social. 2023. «Documento de Análisis de la Po- Brunckhurst, B., R. Hill, G. Mansuri, T. Nguyen, y M. Doan. 2023. breza, 2022.» https://www.coneval.org.mx/Medicion/MP/ Climate and Equity: A framework to guide policy action. Pov- Paginas/Notas_pobreza_2022.aspx. erty and Equity Global Practice, Washington DC: World Bank. CONEVAL, Consejo Nacional de Evaluación de la Política de Calderón, G. 2014. «The effects of Child Care Provision in Mexi- Desarrollo Social. 2024. Sistema de Indicadores sobre Po- co.» Banco de Mexico Working Papers 2014-07. breza y Género en Máxico 2016-2022. Ciudad de México: Calderón, M., J. Cortés, J. Pérez Pérez, y A. Salcedo. 2023. CONEVAL. «Disentangling the Effects of Large Minimum Wage and 72 MEXICO Poverty and Equity Assessment CONEVAL, Consejo Nacional de Evaluación de la Política Social. De la Fuente, A., y M. Serio. 2024. «Vulnerability to poverty cli- 2021. Informa sobre Pobreza y Género 2008-2018. Una mate, risk induced by hazards and climate change in Latin década de medición multidimensional de la pobreza en America and the Caribbean.» Mimeograph. México. Ciudad de México: CONEVAL. De Lima, G.N., R.A. Abeldaño Zuñiga, y D.O. Jerez-Ramírez. 2023. CONEVAL. Consejo Nacional de Evaluacion de la Política de De- «Climate Change and the Increase in Disasters Related to sarrollo Social. 2023. «El CONEVAL presenta información Hydrometeorological and Climatic Events in Mexico: Main sobre el rezago social longitudinal del 2000 al 2020.» Co- Impacts on the Population.» En Climate Change and Health municado No. 3, 28 de marzo. Hazards, de W, Leal Filho, D.G. Vidal y M.A.P. Dinis. Springer. CONEVAL. Consejo Nacional de Evaluacion de la Política de De- Del Valle, A. 2024. «Saving lives with indexed disaster funds: sarrollo Social. 2019. Metodología para la medición mul- evidence from Mexico.» American Economic Journal: Eco- tidimensional de la pobreza en México. (tercera edición), nomic Policy 16 (2). Ciudad de México: CONEVAL. Del Valle, A., A. De Janvry, y E. Sadoulet. 2020. «Rules for Re- CONEVAL. Consejo Nacional de Evaluación de la Política Social. covery: Impact of indexed disaster funds on shock coping 2023. Mapas Temáticos. Plataforma para el Análisis Ter- in Mexico.» American Econommic Journal: Applied Eco- ritorial de la Pobreza. Mexico City: CONEVAL. https://www. nomics 12 (4): 164-95. coneval.org.mx/Medicion/MP/Documents/PATP/2023/ Doan, M.K., R. Hill, S. Hallegatte, P. Corral, B. Brunckhorst, M. Mapas_PATP_abr_2023.pdf. Nguyen, S. Freije-Rodriguez, y E. Naikal. 2023. «Counting —. 2021. «CONEVAL presenta el grado de accesibilidad a car- People Exposed to, Vulnerable to, or at High Risk From Cli- retera pavimentada 2020.» Nota Informativa, 15 de junio. mate Shocks.» Policy Research Working Paper 10619. Costella, C., M. van Aalst, Y. Georgiadou, R. Slater, R. Reilly, A. Mc- Fernandez, Jaime, Sergio Olivieri, y Diana Sanchez. 2023. «A Cord, R. Holmes, J. Ammoun, y V. Barca. 2023. «Can social Methodology for Updating International Middle-class Lines protection tackle emerging risks from climate changem, for the Latin American and Caribbean Region.» World Bank and how? A framework and a critical review.» Cliamte Risk Policy Research Working Paper (10447). Management 40. Ferreira, F., J. Messina, J. Rigolini, L.F. López-Calva, M. Lugo, y R. Covarrubias, A. 2023. «Gender and multidimensional poverty Vakis. 2013. Economic Mobility and the Rise of Latin Amer- at the individual level in Mexico.» Developmenet Studies ican MIddle CLass. Washington DC: World Bank. Research 10 (1). Fuchs, A., y H. Wolff. 2016. «Drought and Retribution. Evidence Croke, K., M.B. Mohd Yussof, Z. Abdullah, A.N. Mohd Hanafiah, K. from a Large-Scale Rainfall Indexed Insurance Program in Mokhtaruddin, E.S. Ramli, N.F. Borhan, Y. Almodovar-Diaz, R. Mexico.» World Bank Policy Research Working Paper 7565. Atun, y A.K. Virk. 2019. «The political economy of health Fuchs, S., y T. Thaler. 2018. Vulnerability and Resilience to Nat- financing reform in Malaysia.» Health Policy and Planning ural Hazards. Cambridge University Press. 732-739. Hallegatte, S., A. Vogt-Schilb, M. Bangalore, y J. Rozemberg. Cuellar, C. Y. and J. O. Moreno (2022), Employment, wages, and 2017. Unbreakable. Building Resilience of the Poor in the the gender gap in Mexico: Evidence of three decades of Face of Natural Disasters. Climate Change and Develop- the urban labor market, Latin American Journal of Central ment, Washington, DC: The World Bank. Banking, Volume 3, Issue 2. Hallegatte, S., M. Fay, y E. Barbier. 2018. «Poverty and Climate Datt, G., y M. Ravallion. 1992. «Growth and redistribution com- Change: Introduction.» Environmental and Development ponents of changes in poverty measures: A decomposition Economics 23: 217-233. with applications to Brazil and India in the 1908s.» Journal Iacovone, L., Rafael Muñoz Moreno, E. Olaberria, y M. de la Paz Perei- of Development Economics 38 (2): 275-295. ra López. 2021. Productivity Growth in Mexico: Understanding De La Fuente, A., E. Ortiz-Juárez, y C. Rodríguez-Castelán. 2018. Main Dynamics and Key Drivers. Mexico: World Bank. «Living on the edge: vulnerability to poverty and public Ibarrarán, M.E., J. Chavarría, y C. Zúñiga. 2020. «Extreme Events transfers in Mexico.» Oxford Development Studies 46 (1): in Mexico: Impacts on Public Health and Development.» En 10-27. Extreme Weather Events and Human Health, de R. Akhtar. Springer. 73 References Jalan, J., y M. Ravallion. 1998. «Transient Poverty in Postreform Mookherjee, D., y A. Shorrocks. 1982. «A Decomposition Anal- Rural China.» Journal of Comparative Economics 26 (2): ysis of the Trend in UK Income Inequality.» The Economic 338-357. Journal 886-902. Jantti, M., y S. Jenkins. 2015. «Income Moblity.» En Handbook Munguía Corella, L.F., y Marco Antonio Gómez Lovera. 2023. of Income Distribution, 807-935. Amsterdam: Elsevier B.V. The Minimum Wage Impact on Poverty in Mexico. mimeo- Jollife, Dean, Daniel Gerszon Mahler, Christoph Lakner, Aziz At- graph, Available at SSRN. doi:http://dx.doi.org/10.2139/ amanov, y Samuel Kofi Tetteh_Baah. 2022. «Assessing the ssrn.4626139. Impact of the 2017 PPPs on the International Poverty Line Murray-Tortarolo, G.N., y M. Martínez Salgado. 2021. «Drought as and Global Poverty.» World Bank Policy Research Working a driver of Mexico-US migration.» Climatic Change 164 (48). Paper (9941). Neumark, D., y L.F. Munguía Corella. 2021. «Do minimum wag- Kakwani, N. 1977. «Measurement of Tax Progressivity: An inter- es reduce employment in developing countries? A survey national comparison.» The Economic Journal. and exploration of conflicting evidence.» World Develop- Kharas, H, y H Kholi. 2011. «What is the MIddle Income Trap, ment 137. Why Do Countries Fall into It and How Can it be Avoided?» Padilla Romo, M., y F. Cabrera Hernández. 2018. «Easing the Global Journey of Emerging Market Economies 3 (3): constraints of motherhood: the effect of all-day schools 281-289. on mother’s labor supply.» Economic Inquiry. Lakner, C., D.G. Mahler, M. Negre, y E.B. Prydz. 2022. «How much Paes de Barros, R., M. Carvalho, S. de Franco, y R. Mendoca. does reducing inequality matter for global poverty?» The 2006. «Uma Análise das Principais Causas da Qeeda Re- Journal of Economic Inequality 20: 559-585. cente na Desigualdade de Renda Brasileira.» Revista Eco- Levy, S. 2010. Good Intentions, bad outcomes: Social policy, nomica 8 (1): 117-147. informality and economic growth in Mexico. Brookings Paul, S. 2004. «Income Sources Effects on Inequality.» Journal Institution Press. of Development Economics 73 (1): 435-451. Levy, S., y Luis Felipe López-Calva. 2023. «¿Qué falló? ¿Qué Puggioni, D., M. Calderón, A. Cebreros Zurita, L. Fernández Bujan- sigue? Mexico 1990-2023. Version extendida.» Nexos, 1 da, J.A. Inguanzo González, y D. Jaume. 2022. «Inequality, de Agosto. income dynamics and worker transitions: The case of Mex- Leyva, G., y C. Urrutia. 2023. «Informal labor markets in times of ico.» Quantitative Economics 13. pandemic.» Review of Economic Dynamics 47: 158-185. Ravallion , Martin. 2016. The Economics of Poverty. Oxford Uni- López Acevedo, G., S. Freije Rodríguez, M. Vergara Bahena, y D. versity Press. Cardozo Madeiros. 2021. «Changes in female employment Ravallion, M., y M. Huppi. 1991. «Measuring Changes in Pover- in Mexico: demographics, markets and policies.» Estudios ty: A Methodological Case Study of Indonesia during and Económicos 36 (1). Adjustment Period.» The World Bank Economic Review 5 López Calva, L.F., y E. Ortiz Juárez. 2014. «A vulnerability ap- (1): 57-82. proach to the definition of the middle class.» Journal of Reyes, M., G. Teruel, y M. López. 2017. «Measuring True In- Economic Inequality 23-47. come INequality in Mexico.» Latin American Policy 8 (1): Macías, A., R. Cantú Calderón, J. S. Méndez Méndez, y E. Sán- 127-148. chez Salazar. 2024. Social Policy Public Financing and Samaniego de la Parra, B., y L. Fernández Bujanda. 2024. «In- Non-Monetary Poverty in Mexico. mimeograph, Mexico creasing the Cost of Informal Employment: Evidence from City: Background Paper to Mexico Poverty Assessment. Mexico.» American Economic Journal: Applied Economics Monroy-Gómez-Franco, L., L.F. Vélez-Grajales, y L.F. López-Cal- 16 (1). va. 2022. «The potential effects of the COVID-19 pandemic Sánchez, L., D. Echevarría, P. Wooders, K. Kuehne, T. Lean, C. on learnings.» International Journal of Educational Devel- Beaton, S. Shama, y Y. Oharenko. 2018. Mejorando y refo- opment 91. calizando los subsidios a la electricidad. Convergencia de la Política Energética y de Cambio Climático (CONECC) y la Alianza Energética entre México y Alemania. 74 MEXICO Poverty and Equity Assessment Scott, J. 2007. «Agricultural Policy and Rural Poverty in Mexi- WMO. 2023. State of CLimate in Latin America and the Caribbe- co.» Documentos de Trabajo del CIDE 395. an 2022. Geneva: World Metereological Organization. Serván-Mori, E., O. Gómez-Dantés, D. Contreras, L. Flamand, D. World Bank. forthcoming. Climate hazards and poverty in Latin Cerecero-García, H. Arreola-Ornelas, y F.M. Knaul. 2023. America. Washington DC: World Bank. «Increase of catastrophic and impoversihing health World Bank. 2023. Health Financial Sustainability and Resil- expenditures in Mexico associated to policy changes ience Assessment Report. World Bank. and the COVID-19 pandemic.» Journal of Global Health. World Bank. 2015. Indigenous Latin America in the Twenty First doi:10.7189/jogh.13.06044. Century. . Washington DC: World Bank. Skoufias, E., K. Vinha, y B.M. Beyene. 2023. «Quantifying Vul- World Bank. 2018. Mexico. Systematic Country Diagnostic. nerability to poverty in the drought-prone lowlands of Ethi- Washington, DC: World Bank Group. opia.» Journal of African Economies. World Bank. 1996. Mexico: Poverty Reduction. The Unfinished Solórzano, A., y M. Goode. 2023. Linking disaster risk financing Agenda. Report No. 15692 ME, Washington , DC: The World with social protection: An overview of concepts and con- Bank. siderations. Rome: World Food Programme. World Bank. 1989. Poverty Alleviation in Mexico. Report No. Stampini, M., P. Ibarrarán, C. Rivas, y M. Robles. 2021. «Adap- 7785-ME, Washington DC: The World Bank. tive, but not by design: Cash Transfers in Latin America and World Bank. 2018. Poverty and Shared Prosperity 2018: Piecing the Caribbean before, during and afte the COVID-19 pan- Together the Poverty Puzzle. Washington, DC: World Bank. demic.» IDB Technical NOte 2346. World Bank. 2009. «Poverty and Social Impact Analysis of Teruel, G. 2022. Dinámicas de la pobreza en México. Mexico groundwater Over-exploitation in Mexico.» mimeograph. City: Centro de Estudios Espinosa Yglesias. World Bank. 2004. Poverty in Mexico: An Assessment of Trends, Vaalavuo, M. 2015. «Poverty Dynamics in Europe: From What to Conditions, and Government Strategy. Mexico City: The Why.» EG-EMPL Working Paper (Publications Office of the World Bank. European Union ). World Bank. 2023. Women, Business and the Law. Washington, Villareal, H., y A. Macías. 2020. «El sistema de pensiones en Méx- DC: World Bank. ico. Institucionalidad, gasto público y sostenibilidad finan- ciera.» CEPAL, Macroeconomía del Desarrollo (CEPAL) 210. 75 Annexes Annexes 76 MEXICO Poverty and Equity Assessment Annex 1. Comparison of monetary poverty rate estimates from CONEVAL and the World Bank The World Bank regularly produces global estimates of pov- Mexico has its own official measures of monetary and erty headcounts. This requires the estimation of poverty rates multidimensional poverty. Actually, Mexico has been one of for all countries in the World in a manner that is consistent over the most prominent countries in the world producing regular time, and comparable across countries. This also allows as- measures of monetary poverty and one of the first to innovate sessing a country relative performance vis-à-vis the World, its and adopt an official multidimensional measure of poverty. The region or other countries, and drawing lessons from compara- official Mexican measures are different to the World Bank mea- tive development experiences. Full comparability is not always sures, but very similar at the same time. They follow the same possible and, sometimes, the measurement method adopted principles and theoretical background, but use different quanti- by the Bank differs from the method adopted by the country. tative approaches. This report contrasts both sets of measures Here we describe the differences between World Bank’s esti- to give a full assessment of the trends and patterns over time mates of poverty for Mexico and the country own official pover- and across countries. ty estimates, produced by the Consejo Nacional de Evaluación The estimation of a poverty measure requires the definition de la Política de Desarrollo, (CONEVAL). Both entities produce of three main components: a welfare aggregate, a poverty line, multidimensional measures of poverty, but here we will only and a poverty measure or index. CONEVAL and World Bank mea- compare monetary measures of poverty. 103 sures differ in the two first components, but both use the pov- As an upper middle income country, the World Bank mea- erty headcount, poveryt gap and poverty severity as measures sures poverty in Mexico using a poverty line of US$ 6.85 per of monetary poverty. Regarding the welfare aggregate, both in- day, in 2017 purchasing power parity terms. This is one of the stitutions use the same welfare aggregate (that is, household different poverty lines that the World Bank has been using to total monthly income as captured by Encuesta Nacional de In- assess progress in poverty eradication with worldwide, regional gresos y Gastos de los Hogares, ENIGH) but with one exception: and country-specific focus. The “international poverty line”, for- the World Bank includes housing rents (observed and imputed) merly called the “dollar-a-day” line, has been regularly updated in the welfare aggregate, whereas CONEVAL does not. The ob- and serves as the first sustainable development goal for glob- jective is to measure the value of occupying a dwelling for a al poverty eradication.104 But as countries grow and develop, month, and avoid two households with the same consumption what is socially defined as poverty changes, and consequently patterns have different welfare aggregate just because one higher poverty lines are used for measuring monetary poverty pays rent and the other does not. This difference would make in countries with higher standards of living. Given its current poverty estimates from the World Bank lower than CONEVAL’s. living standards, Mexico is an “upper-middle” income country Moreover, the World Bank welfare aggregate is defined in and hence its current challenge is to eradicate poverty under household member per capita terms, whereas CONEVAL uses the US$ 6.85 poverty line. an adult equivalence. An adult equivalence defines the income Moreover, poverty is a condition that not only refers to lack need of each household member, often defined in terms of age of monetary resources, but also to insufficient functioning and and household size, relative to a benchmark. CONEVAL uses the capabilities which hinder the full development and freedom of OECD-modified equivalence scale, which assigns a weight of people. This has led to the formulation of multidimensional pov- 1.0 to the first adult in the household, 0.5 to each additional erty measures -both in academia and official national statis- adult, and 0.3 to each child. For instance, a household with two tics- to better assess progress in poverty reduction. The World adults and two children would have a welfare aggregate divided Bank has also adopted a multidimensional poverty measure by four in the World Bank measure, but divided by 2.1, in CONE- since 2018. It includes measures of access to health, educa- VAL’s measure. This difference would make poverty estimates tion and housing, as well as monetary poverty.105 from the World Bank higher than CONEVAL’s. 77 Annexes Finally, CONEVAL uses two monetary poverty lines: the “línea would have a difference of about 2 percentage points. The dif- de pobreza” and the “línea de pobreza extrema”. These are up- ference would rise to a maximum of nearly 3 percentage points dated every month using a specific consumption basket for ur- for a poverty rate close to US$ 15 day (the official “linea de po- ban and rural areas which leads to four different poverty lines. breza por ingresos” was around MX$ 4100 per month for urban The World Bank uses several poverty lines that are deemed ap- areas in 2022, about US$ 13 in 2021 PPP terms). Annex Figure propriate for different countries depending on their income level. 2, left panel, compares CONEVAL’s aggregate to a version of the For the case of Mexico, this is the “upper-middle income” coun- World Bank’s aggregate that uses adult equivalence rather than try poverty line of US$ 6.85 per day in 2017 purchasing power per capita adjustment. In this case, the difference (shown in the parity terms.106 The World Bank does not use different lines for right panel) would reach up to seven percentage points around urban and rural areas, or any other region-specific line. Instead, the upper-middle income poverty line. This underlines the fact two adjustments are made to the measure of well-being. First, that using adult equivalence increases the welfare aggregate time adjustment is necessary because not all households and and reduces the poverty rate. Finally, Annex Figure 3, left panel, individuals are interviewed in the same month of the year. This shows a comparison between the CONEVAL’s aggregate to a ver- involves converting the nominal monetary and nonmonetary in- sion of the World Bank’s aggregate that uses adult equivalence come of individuals interviewed in different months to August’s rather than per capita adjustment and does not include the im- values, using the same detailed CPI as CONEVAL for their welfare putation for housing rents. This would make both aggregates as aggregate, which is obtained from INEGI. Second, the spatial ad- close as possible in terms of definition. Indeed, the right panel justment is applied after the measure of wellbeing is calculated shows that the difference reaches at most a one percentage and it is used to account for the differences in the cost of living point around the upper-middle income poverty line and less between urban and rural areas. This involves increasing the in- than half a percentage point elsewhere.109 come of rural households and individuals by 15 percent, as the In summary, the poverty rates computed by CONEVAL and cost of living in rural areas is significantly lower than in urban the World Bank are bound to be different because of distinct areas. To accomplish this, a new variable is created, with a val- definitions of the welfare aggregate, the poverty line and the ue of 1 for urban areas and 0.8695 for rural areas. The welfare definition of family size. Once these discrepancies are taken aggregate is deflated by this factor and then compared to the into account, the poverty rates are very similar. The time trends upper-middle income poverty line.107 of poverty rates from CONEVAL and from the World Bank for the Given all these differences it is natural to ask if the pover- period 2000-2022 are similar too (Annex Figure 4). Under all ty rates from these methods are fundamentally different. A the lines, there is a clear decline for the period 2000-2006, a method to do this comparison is called stochastic dominance. stable trend between 2008 and 2014, and another decline be- It consists in comparing the poverty incidence curve of the two tween 2016 and 2022. measures. This curve plots the poverty rate in the vertical axis The World Bank poverty rates were closer to CONEVAL’s and the poverty line on the horizontal axis (which is the same poverty rates in the early 2000s whereas they are closer to as the cumulative distribution function of the welfare aggre- CONEVAL’s extreme poverty rates in the early 2020s. This is gate). It illustrates whether poverty rate is higher in one aggre- due to price adjustments of the extreme poverty line that are gate or the other for every possible poverty line.108 faster than the average consumer price index in Mexico (CPI), Annex Figure 1, left panel, shows these incidence curves which serves as deflator of the US$ 6.85 upper-middle income curve for the welfare aggregate of CONEVAL (in black) and the poverty line for the country. Consequently, the World Bank’s welfare aggregate of the World Bank (in blue). The difference upper-middle income poverty line was closer to CONEVAL’s rural between these two is seen in the panel to the right and it shows extreme poverty line in the early 2000s, but is now closer to that around the World Bank’s upper-middle income poverty line the CONEVAL’s urban extreme poverty line ( Annex Figure 5 ). of US$ 6.85/day the poverty rate between these two aggregates 78 MEXICO Poverty and Equity Assessment Annex Figure 1: Difference between CONEVAL and World Bank welfare aggregates FGT Curves Difference Between FGT Curves Household income per capita SEDLAC vs. Household income per capita SEDLAC vs. Household income per adult equivalent scale CONEVAL Household income per adult equivalent scale CONEVAL 1 .03 .8 .02 Poverty rate .6 Difference .4 .01 .2 0 0 0 20 40 60 80 100 0 20 40 60 80 100 poverty line (2017 USD PPP) poverty line (2017 USD PPP) per capita SEDLAC per adult equivalent CONEVAL confidence interval (%95) per capita - per adult equivalent Source: World Bank staff calculations using ENIGH 2022 Note: FGT curves and differences using command cfgt from DASP 3.03 in STATA, version 18.0. Annex Figure 2: Difference between CONEVAL and World Bank welfare aggregate if using adult equivalence FGT Curves Difference Between FGT Curves Household income adult equivalent SEDLAC vs. Household income per capita SEDLAC vs. Household income adult equivalent scale CONEVAL Household income per adult equivalent scale CONEVAL .1 1 .08 .8 FGT Curves (a=0) Difference .06 .6 .4 .04 .2 .02 0 0 0 20 40 60 80 100 0 20 40 60 80 100 poverty line (2017 USD PPP) poverty line (2017 USD PPP) adult equivalent SEDLAC adult equivalent CONEVAL confidence interval (%95) per adult equivalent- per adult equivalent Source: World Bank staff calculations using ENIGH 2022 Note: FGT curves and differences using command cfgt from DASP 3.03 in STATA, version 18.0. 79 Annexes Annex Figure 3: Difference between CONEVAL and World Bank welfare aggregate if using adult equivalence and excluding housing rent imputation FGT Curves Difference Between FGT Curves Household income per adult equivalent SEDLAC vs. Household income per adult equivalent SEDLAC vs. Household income per adult equivalent scale CONEVAL Household income per adult equivalent scale CONEVAL 1 .01 .8 .008 FGT Curves (a=0) Difference .6 .006 .4 .004 .2 .002 0 0 0 20 40 60 80 100 0 20 40 60 80 100 poverty line (2017 USD PPP) poverty line (2017 USD PPP) per adult equivalent SEDLAC per adult equivalent CONEVAL confidence interval (%95) per adult equivalent - per adult equivalent Source: World Bank staff calculations using ENIGH 2022 Note: FGT curves and differences using command cfgt from DASP 3.03 in STATA, version 18.0. 80 MEXICO Poverty and Equity Assessment Annex Figure 4: CONEVAL and World Bank poverty rates Poverty rates in Mexico under different methodologies 60 50 percentage of total population 40 30 20 10 0 2000 2002 2004 2005 2006 2008 2010 2012 2014 2016 2018 2020 2022 under "linea de pobreza extrema de ingresos" (new series) under "linea de pobreza de ingresos" (new series) under "linea de pobreza extrema de ingresos" (old series) under "linea de pobreza de ingresos" (old series) under "linea de pobreza alimentaria" under "linea de pobreza patrimonio" under US$ 6.85 (2017 PPP) Source: World Bank staff calculations using ENIGH 2002-2022 and CONEVAL (several publications) Note: All poverty rates produced by CONEVAL, except under the US$ 6.85 (2017) produced by World Bank. 81 Annexes Annex Figure 5: CONEVAL and World Bank poverty lines in current Mexican pesos Official CONEVAL Poverty Lines and World Bank Poverty Line for upper-middle income countries in monthly current Mexican Pesos 4.500 4.000 3.500 monthly MX$ per head 3.000 2.500 2.000 1.500 1.000 500 - 2000 2002 2004 2005 2006 2008 2010 2012 2014 2016 2017 2018 2020 2022 rural extreme poverty line urban extreme poverty line rural poverty line urban poverty line US$ 6.85 (2017 PPP) Source: World Bank staff calculations and CONEVAL (several publications) Note: poverty line at US$ 6.85 in current Mexican pesos computed using Purchasing Power Parities for GDP (base 2017). 82 MEXICO Poverty and Equity Assessment Annex 2. Description of panels using ENOE This annex provides an overview of the methodological Construction of the Panel of Workers changes introduced in the National Survey of Occupation and Employment (ENOE) household surveys in Mexico and their im- The construction of a panel of workers involves selecting pact on constructing indicators, assembling a panel of workers, individuals who appear in consecutive survey waves, allow- and understanding the effects of labor outcomes and poverty ing for tracking labor market transitions at the individual level. by income. The differences in data collection methods may Quarterly data was obtained from INEGI110 from the first quarter result in variations in the survey representativeness, particu- of 2005 to the first quarter of 2023. To keep consistency, we larly for workers in the informal sector or those with irregular keep individuals between 15 and 64 years old and exclude employment patterns. telephonic surveys. Hence, we combine face-to-face interviews in the ENOE and ENOE-N surveys. We follow the methodology provided by the National Council for the Evaluation of Social Evolution of ENOE Methodology: ENOE, ETOE, Development Policy (CONEVAL) to merge survey modules by ENOE-N constructing unique household and individual identifiers that link individuals across survey waves111. The CONEVAL method The ENOE household survey in Mexico has undergone sig- allow us to easily distinguish from face-to-face and telephonic nificant methodological changes in recent years. This includes interviews. This process also includes excluding some cases the transition from the original ENOE to the Telephone Survey of with incomplete or unreliable data. Lastly, the construction of Occupation and Employment (ETOE) during the second quarter demographic and labor variables is based on the World Bank of 2020, the introduction of ENOE New Edition (ENOE-N) to ad- Global Labor Database harmonization template112. dress emerging data collection challenges and the eventual re- turn to the original ENOE framework in the first quarter of 2023. The shift to ETOE was necessary due to the COVID-19 pan- Implications for Research on Income Inequality demic-related constraints, allowing data collection to continue only via telephone surveys. However, this change introduced The methodological changes in ENOE and the construction potential biases, particularly in representing informal sector of the worker panel have significant implications for research workers and those with limited access to telecommunications. on income inequality in Mexico. The constructed panel is valu- The ENOE-N aimed to mitigate these biases by using a mixed- able for analyzing labor market dynamics; however, it remains mode approach, combining telephone and face-to-face inter- critically to assess the potential biases and limitations inherent views. Despite these efforts, the comparability of data across in the data. The potential discontinuities in the data series, cou- these phases remains a concern, as changes in methodology pled with the challenges in maintaining a consistent panel of may affect the measurement of labor market outcomes, partic- workers, underscore the need for a cautious approach to inter- ularly for vulnerable groups. preting trends in inequality. Attrition, non-response, and poten- The return to the original ENOE framework restored some tial biases introduced by the survey’s rotating panel design can consistency in data collection. However, caution is needed affect the robustness of the panel estimates. Lastly, the shift when interpreting trends, especially when analyzing periods in survey methodology introduces challenges in maintaining a spanning different survey methodologies or when a large event consistency over time. disrupts data collection in specific regions, such as the 2023 Hurricane Otis in the state of Guerrero. These potential for dis- Comparison of panels to cross-sectional data continuities in the data series requires careful consideration of the implications for longitudinal analysis, especially in studies A comparison of these panels and the cross sectional data of poverty and income inequality. shows that they are similar. The distribution of population ac- cording to area of residence, sex, sector of employment and 83 Annexes economic activity have differences with respect to the refer- ary level) and bigger households (those with more than four ence cross section below the 1 percentage point. There are members). The latter is due to a smaller proportion of singles some differences as well. The panels have a smaller share of and single-headed households in the panels (perhaps, family young population (aged 15-44) and a higher percentage of arrangements more likely to move from one dwelling to another more senior people (aged 45 and more). Panels also have a over time). Annex Table 6 shows the comparison for two panel larger share of people with more schooling (at least second- data and a comparison cross section. Annex Table 6: Demographic characteristics of selected ENOE panel and comparison full cross section. Distribution of sample in cross panel cross panel error standard percentage of sample (full section Q1’2013/ difference section Q1’2022/ diferencia standard error survey or panel) Q1’2014 Q1’2014 Q1’2023 Q1’2023 area rural 14.46 15.91 -1.45 0.20 14.99 15.02 -0.03 0.21 urban 85.54 84.09 1.45 0.20 85.01 84.98 0.03 0.21 sex (of household head) women 24.41 24.86 -0.45 0.40 31.27 31.29 -0.02 0.45 men 75.59 75.14 0.45 0.40 68.73 68.71 0.02 0.45 age (of household head) 15-29 12.65 9.41 3.24 0.28 10.19 7.84 2.35 0.26 30-44 41.31 41.43 -0.13 0.46 35.38 33.82 1.56 0.46 45-59 37.46 39.96 -2.49 0.45 42.54 45.33 -2.78 0.48 60 and more 8.58 9.21 -0.62 0.27 11.89 13.01 -1.12 0.33 schooling (of household head) a lo mas primaria 31.87 34.31 -2.45 0.45 22.50 24.67 -2.17 0.42 completa secundaria completa o 28.85 29.40 -0.55 0.43 30.49 31.04 -0.55 0.46 incompleta preparatoria completa o 17.04 16.71 0.34 0.35 21.50 21.19 0.31 0.40 incompleta profesional completa o 19.89 17.35 2.54 0.36 22.56 20.52 2.03 0.40 incompleta postgrado completo o 2.35 2.23 0.12 0.14 2.96 2.58 0.38 0.16 incompleto economic sector (of household head) inactive 13.46 14.27 -0.81 0.32 15.44 16.20 -0.76 0.36 unemployed 3.00 2.63 0.37 0.15 1.49 1.27 0.21 0.11 informal 39.82 40.55 -0.73 0.45 38.26 39.17 -0.91 0.47 formal 43.72 42.55 1.17 0.46 44.81 43.35 1.46 0.48 84 MEXICO Poverty and Equity Assessment Distribution of sample in cross panel cross panel error standard percentage of sample (full section Q1’2013/ difference section Q1’2022/ diferencia standard error survey or panel) Q1’2014 Q1’2014 Q1’2023 Q1’2023 economic activity (of household head) inactive 13.46 14.27 -0.81 0.32 15.44 16.20 -0.76 0.36 unemployed 3.00 2.63 0.37 0.15 1.49 1.27 0.21 0.11 agricultura (11) 8.80 9.93 -1.13 0.28 7.75 8.40 -0.65 0.27 mineria, energia y 1.26 1.17 0.09 0.10 1.13 1.13 0.00 0.10 agua (21 + 22)i construccion (23) 8.67 8.47 0.19 0.26 8.38 8.60 -0.22 0.27 manufactura 13.36 13.52 -0.16 0.32 13.97 13.98 -0.01 0.34 (31-33) comercio (43 13.80 13.49 0.31 0.32 13.94 14.04 -0.10 0.34 + 46) transporte (48-49) 5.10 5.20 -0.10 0.21 5.25 5.65 -0.39 0.22 comunicaciones 0.67 0.58 0.09 0.07 0.64 0.59 0.04 0.07 (51) servicios financieros e 1.39 1.19 0.20 0.10 1.43 1.30 0.13 0.11 inmobiliarios (52-53) profesionales, cientificos y 2.06 1.92 0.14 0.13 2.23 1.82 0.41 0.13 corporativos (54-55) educacion (61) 4.38 4.25 0.13 0.19 4.20 4.09 0.11 0.19 salud y asistencia 2.36 2.18 0.18 0.14 2.45 2.45 0.00 0.15 social (62) recreacion, cultura y deportes 0.77 0.73 0.04 0.08 0.68 0.72 -0.03 0.08 (71) restaurantes y 5.15 4.94 0.21 0.20 5.73 5.00 0.73 0.21 hoteles (72) gobierno (93) 5.84 5.73 0.11 0.22 4.50 4.63 -0.13 0.20 otros ( 99, 81, 56) 9.93 9.80 0.12 0.28 10.80 10.14 0.66 0.29 Household characteristics 1 Household MONTHLY per- capita income 2.565.36 2.405.38 159.97 27.99 4.818.95 4.491.71 327.23 69.37 (Mexican pesos) - CONEVAL Family size (shares) Four or less 68.34 65.54 2.80 0.44 75.39 73.30 2.10 0.43 members More than four 31.66 34.46 -2.80 0.44 24.61 26.70 -2.10 0.43 members 85 Annexes Annex 3. Demographic profiles of poverty and chronic labor poverty Annex Table 7: Poverty, Vulnerability and Middle Class profiles, Mexico 2000 and 2022 Distribution Distribution of Distribution of Distribution Vulnerability Middle class of population Poverty Rate population in middle class of total Rate rate vulnerable to poverty population population poverty 2000 2022 2000 2022 2000 2022 2000 2022 2000 2022 2000 2022 2000 2022 area of residence: rural 81.4% 35.8% 13.5% 40.6% 4.9% 23.2% 40.8% 39.8% 11.8% 25.5% 6.0% 14.7% 25.0% 24.3% urban 39.4% 17.3% 33.7% 38.0% 25.5% 43.3% 59.2% 60.2% 88.2% 74.5% 94.0% 85.3% 75.0% 75.7% sex women 49.8% 22.2% 28.9% 38.9% 20.3% 37.7% 51.9% 53.2% 52.4% 52.5% 51.7% 51.1% 52.0% 52.0% men 49.9% 21.3% 28.4% 38.2% 20.5% 39.1% 48.1% 46.8% 47.6% 47.5% 48.3% 48.9% 48.0% 48.0% age 0-14 61.6% 34.2% 24.8% 42.0% 13.4% 23.5% 41.6% 37.0% 29.2% 25.7% 22.3% 14.4% 33.7% 23.6% 15-29 44.2% 20.0% 32.7% 40.4% 22.0% 38.8% 23.7% 22.1% 30.5% 25.2% 28.9% 24.3% 26.7% 24.1% 30-44 46.3% 21.7% 30.0% 38.0% 22.8% 38.9% 18.1% 20.4% 20.4% 20.2% 21.8% 20.8% 19.5% 20.5% 45-59 38.1% 14.5% 28.4% 34.9% 29.9% 48.6% 9.2% 11.9% 12.0% 16.1% 17.7% 22.5% 12.1% 17.8% 60 and more 46.1% 13.3% 28.3% 35.2% 23.6% 49.2% 7.4% 8.6% 7.9% 12.8% 9.3% 18.0% 8.0% 14.1% schooling aged less than 61.6% 34.2% 24.8% 42.0% 13.4% 21.7% 41.6% 37.0% 29.2% 25.7% 22.3% 14.4% 33.7% 23.6% 15 no formal 73.8% 36.0% 19.6% 43.1% 6.3% 22.3% 9.6% 5.6% 4.5% 3.8% 2.0% 1.8% 6.5% 3.4% incompleted 64.6% 28.0% 26.1% 43.3% 9.1% 25.0% 15.7% 8.7% 11.0% 7.6% 5.4% 5.0% 12.1% 6.8% primary completed 52.3% 24.3% 31.3% 42.4% 16.1% 32.4% 12.6% 10.3% 13.1% 10.2% 9.5% 8.0% 12.0% 9.3% primary incomplete 41.2% 21.4% 37.0% 42.9% 21.2% 33.9% 16.1% 26.5% 25.2% 30.0% 20.3% 24.8% 19.5% 27.0% secondary completed 21.5% 13.5% 38.1% 38.0% 38.6% 47.5% 3.0% 7.6% 9.3% 12.0% 13.2% 15.1% 7.0% 12.2% secondary incomplete 12.3% 8.0% 30.9% 33.4% 53.1% 56.9% 1.1% 2.5% 4.7% 5.8% 11.3% 10.0% 4.3% 6.7% tertiary completed 3.1% 3.5% 18.1% 17.2% 67.1% 72.8% 0.3% 1.8% 3.1% 4.9% 16.0% 20.9% 4.9% 11.0% tertiary 86 MEXICO Poverty and Equity Assessment Distribution Distribution of Distribution of Distribution Vulnerability Middle class of population Poverty Rate population in middle class of total Rate rate vulnerable to poverty population population poverty 2000 2022 2000 2022 2000 2022 2000 2022 2000 2022 2000 2022 2000 2022 by type of household single person 19.6% 5.4% 25.5% 18.7% 47.3% 68.7% 0.7% 0.9% 1.6% 1.8% 4.1% 6.8% 1.8% 3.8% couple without 24.9% 7.7% 31.1% 27.9% 36.4% 59.5% 1.9% 2.5% 4.1% 5.1% 6.7% 10.9% 3.8% 7.0% children couple with up to three 39.4% 22.0% 36.1% 41.7% 23.9% 35.5% 15.1% 16.4% 24.0% 17.5% 22.4% 14.9% 19.1% 16.2% children couple with three children 71.5% 50.7% 16.6% 35.4% 11.8% 13.6% 27.7% 19.5% 11.2% 7.7% 11.2% 3.0% 19.3% 8.4% or more single parent 51.9% 31.7% 23.4% 41.0% 24.8% 26.8% 2.9% 5.3% 2.2% 3.8% 3.3% 2.5% 2.7% 3.6% with children multi- generational 50.1% 20.4% 30.1% 41.6% 18.5% 37.2% 37.6% 36.2% 39.3% 41.7% 34.1% 37.5% 37.4% 38.7% family other 44.5% 18.8% 31.6% 38.6% 23.2% 42.0% 14.2% 19.2% 17.6% 22.3% 18.2% 24.4% 15.9% 22.3% by employment sector of household head agriculture 84.6% 45.1% 10.6% 35.7% 4.7% 18.8% 38.2% 32.7% 8.7% 15.2% 5.4% 8.6% 18.6% 12.7% industry 39.5% 17.0% 39.2% 41.4% 19.5% 40.3% 14.1% 11.8% 25.5% 16.9% 17.9% 17.7% 14.7% 12.1% construction 62.4% 26.9% 27.8% 45.3% 9.2% 27.4% 14.5% 13.4% 11.8% 13.3% 5.5% 8.6% 9.6% 8.7% services 35.1% 16.6% 32.7% 38.4% 30.7% 43.3% 31.2% 37.9% 52.9% 51.5% 70.3% 62.2% 36.6% 39.9% unemployed / out of labor 68.7% 32.0% 20.0% 39.6% 10.0% 27.5% 2.0% 4.0% 1.1% 2.9% 0.8% 2.2% 1.2% 2.2% force Source: Own calculations using ENIGH 2000 and 2022. 87 Annexes Annex Table 8: Demographic profile of chronic poor, transitory poor and no poor, using panels from ENOE 2013 and 2024 Incidence Distribution of total Distribution of population in No Poverty Transient Poverty Chronic Poverty No Poverty Transient Poverty Chronic Poverty panels 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 Total 27,23 34,86 47,8 47,2 24,97 17,90 area rural 11,7 19,8 44,9 49,5 43,4 30,6 8,6 11,8 18,8 21,8 34,7 35,6 15,9 15,0 urban 31,1 38,8 48,5 46,6 20,4 14,6 91,4 88,2 81,2 78,2 65,3 64,4 84,1 85,0 sex (of household head) women 23,6 30,3 47,4 46,8 29,1 22,9 17,8 24,2 20,3 27,7 23,9 35,7 24,9 31,3 men 28,2 36,6 47,9 47,4 23,9 16,0 82,2 75,8 79,7 72,3 76,1 64,3 75,1 68,7 age (of household head) 15-29 25,5 33,9 44,8 45,7 29,8 20,4 10,5 9,3 10,5 9,3 13,3 10,9 9,4 7,8 30-44 25,7 30,3 46,4 49,0 27,9 20,7 44,3 35,7 45,6 42,5 52,4 47,4 41,4 33,8 45-59 31,3 40,7 49,8 45,9 18,9 13,3 41,5 48,4 37,7 40,3 27,4 30,9 40,0 45,3 60 and more 17,9 28,8 52,2 46,8 29,9 24,4 3,8 6,6 6,3 7,9 6,9 10,9 9,2 13,0 schooling (of household head) no formal - - - - - - - - - - elementary, complete or 16,3 24,9 49,3 48,4 34,4 26,6 22,6 18,8 39,1 26,9 52,2 39,1 34,3 24,7 incomplete secondary, complete or 23,2 28,9 50,0 52,3 26,7 18,8 25,8 28,5 31,6 38,0 32,3 36,1 29,4 31,0 incomplete high-school, complete or 33,2 37,6 52,3 47,7 14,5 14,8 17,9 21,2 16,0 19,8 8,5 16,2 16,7 21,2 incomplete university, complete or 56,2 55,7 36,3 37,0 7,5 7,3 25,6 24,9 9,4 12,2 3,7 6,4 17,3 20,5 incomplete post-grade, compete or 77,7 71,8 18,0 23,8 4,3 4,4 3,6 3,4 0,5 0,8 0,2 0,4 2,2 2,6 incomplete N/A 34,2 45,1 45,0 41,9 20,8 13,0 4,6 3,3 3,4 2,3 3,0 1,8 88 MEXICO Poverty and Equity Assessment Incidence Distribution of total Distribution of population in No Poverty Transient Poverty Chronic Poverty No Poverty Transient Poverty Chronic Poverty panels 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 2013-2014 2022-2023 economic sector (of household head) inactive 10,2 14,7 46,6 47,9 43,3 37,4 4,1 5,5 10,7 13,3 19,0 27,3 14,3 16,2 unemployed 11,6 12,0 61,0 61,0 27,3 27,0 1,2 0,6 3,5 2,3 3,0 2,7 2,6 1,3 informal 16,9 24,5 50,6 52,7 32,5 22,8 29,8 30,2 50,7 48,0 62,2 54,8 40,6 39,2 formal 46,0 52,7 43,7 40,8 10,3 6,5 65,0 63,7 35,2 36,4 15,8 15,2 42,6 43,4 economic activity (of household head) inactive 10,2 14,7 46,6 47,9 43,3 37,4 4,1 5,5 10,7 13,3 19,0 27,3 14,3 16,2 unemployed 11,6 12,0 61,0 61,0 27,3 27,0 1,2 0,6 3,5 2,3 3,0 2,7 2,6 1,3 agricultura 8,6 15,6 37,8 41,2 53,6 43,2 4,4 5,2 11,2 10,1 30,2 27,9 9,9 8,4 (11) all other economic 34,0 42,0 49,4 47,7 16,5 10,2 90,3 88,7 74,7 74,4 47,8 42,1 73,2 74,1 activities Household characteristics 1 Family size (shares) Four or less 34,4 40,9 47,2 45,1 18,4 14,1 65,4 72,1 51,1 58,6 38,1 48,3 65,5 73,3 members More than 19,5 25,3 48,4 50,7 32,1 24,0 34,6 27,9 48,9 41,4 61,9 51,7 34,5 26,7 four members Source: Own calculations using quarterly ENOE from 1st quarter 2005 to 1st quarter 2023. Note: Panels are constructed taking advantage of the rotation scheme of ENOE which interviews same dwelling for up to 5 consecutive quarters. For a description of the panels and comparability to full ENOE surveys see Annex 2. Description of panels using ENOE. Labor poverty refers to percentage of the population whose labor earnings per household member are below the US$ 6.85 per day (2017 PPP). 89 Notes Notes 1 The household survey used for poverty measurement in Mexico, 6 The exception is people living in single person households or couple ENIGH, had a major change in 2016 aimed to better capture household without children households whose vulnerability rate also declined. income, making it representative at more disaggregated regions 7 Full demographic profile of the poor, the vulnerable and the middle class and capturing aggregate household income closer to national in Annex 3. Demographic profiles of poverty and chronic labor poverty. accounts figures, all of which made the survey non-comparable All these profiles using a poverty line of US$ 6.85 per day at 2017 PPP between 2014 and 2016. For a technical description of such change terms. see https://en.www.inegi.org.mx/programas/enigh/nc/2016/. The 8 Poverty, vulnerability and middle class profiles show no statistically measures of multidimensional poverty started in 2008 and also had signifficant difference between men and women. This is due to the fact a methodological adjustment in 2018 which made the data in 2014 that this processing of survey data does not separate household income not-strictly comparable to the numbers in 2016. See Diario Oficial de la per capita by gender or other household member characteristics. Federación, 30 de Octubre 2018: https://www.dof.gob.mx/nota_detalle. Given that there is, on average, the same number of members of php?codigo=5542421&fecha=30/10/2018#gsc.tab=0 ). each sex, these poverty profiles do not differ.For a special treatment These created two separate six-year periods starting in 2008 and of multidimensional poverty indexes to gauge gender differences in in 2016. To maintain comparability and symmetry, we considered a Mexico see (Covarrubias 2023). For a discussion of women and poverty third six-year period starting in 2000 for the study of the evolution of in Mexico in this report see Box 2 below. monetary poverty. For measures of non-monetary multidimensional 9 The formulation of this decomposition is by (Ravallion and Huppi 1991). poverty the report concentrates on the two six-year periods for which 10 Personal remittances went from 1.0 to 2.6 percent of GDP between 2000 national data are available. and 2006, but declined from 2.3 to 1.9 percent between 2008 and 2014 2 For a discussion of the World Bank upper-middle income poverty line (with a decline in absolute number of total remittances between those two and its comparison to Mexico’s official poverty line see see Annex 1. years). They increased again from 3.6 to 4.2 percent of GDP between 2016 Comparison of monetary poverty rate estimates from CONEVAL and the and 2022. Data from World Development Indicators (visit https://datatopics. World Bank. worldbank.org/world-development-indicators/ on July 15, 2024). 3 The three six-year periods have very low rates of real GDP growth (1.8, 11 Our processing of microdata indicates a particular growth of transfers 1.5 and 0.8 percent respectively), and even slower growth rates of associated with the PROCAMPO program to households at the upper average productivity in terms of GDP/head (0.3, 0.2 and 0, percent half of the income distribution. For study on the high concentration of respectively). The three periods suffered from a recession (a mild one PROCAMPO subsidies among the top of the distribution see: (Scott 2007) in 2001 due to the deceleration of the USA economy, and two severe 12 The formulation of this decomposition is by (Paes de Barros, et al. 2006) ones in 2009 because of the international financial crisis and in 2020 and (Azevedo, Sanfelice and Nguyen 2012). because of COVID-19). But, beyond the recessions, the slow growth 13 The median hourly wage for employed workers rose 2.1 percent per of the Mexican economy has more profound causes. For an extensive year between 2008 and 2014, whereas the CPI grew at 4 percent discussion see World Bank study on the sources of economic growth in that period. Between 2016 and 2022, the rates were 6.3 and 5.3, and productivity in Mexico over the last three decades, see (Iacovone, respectively. Data on wage employment rates and hourly wages from et al. 2021). Another assessment about policies to accelerate growth National Survey of Occupation and Employment (ENOE), population through competition, innovation and inclusion is included in (World Bank aged 15 years and over Interactive query of strategic indicators 2018). Poverty reduction despite slow GDP growth, as will be explained (InfoLaboral): (https://www.inegi.org.mx/sistemas/Infoenoe/ later, is due to progressive rates of growth in labor earnings and social Default_15mas_en.aspx ). protection transfers, which favored poorer households. 14 This is consistent with the evolution of macro data about remittances. 4 The technical definition of vulnerability comes from (Ferreira, et al. See footnote 9. 2013) and (López Calva and Ortiz Juárez 2014). For an approach to the 15 World Bank Macro Poverty Outlook Fall 2022. definition of the middle class see (Birdsall, et al. 2011). Extension and 16 The large size of informal employment in Mexico has been described as one reelaboration of these methods by (Fernandez, Olivieri and Sanchez of the main causes of persistently low economic productivity and deep- 2023) render the thresholds used in this report. These authors define rooted social deprivations. A complex set of labor regulations, payroll taxes the vulnerability line as the median per capita household income over and social security programs have been usually blamed for the high levels households who were not poor in the initial year and became poor in the of informality in the country. See (Levy 2010) and (Levy and López-Calva final year, i.e., a household i in j = 2 moves into poverty if yi2 < z and yi1 2023). It is not yet determined, however, what can explain the small but > z, where z is the international poverty line of 6.85 a day in purchasing persistent decline observed in informality rates in recent years. Recent power parities (PPP) 2017. The middle class upper bound is defined with studies find that more regular inspections may increase the probability of the same method but using the 99th percentile of household incomes being formal (see (Samaniego de la Parra and Fernández Bujanda 2024)), that slipped into poverty from one period to the next. and that higher entry-level wages may also induce more workers to become 5 The only exception is the poverty rate of people with university degrees formal (see (Abel, et al. 2022)). There is also evidence of informality rates which was 3.1 percent in 2000 and 3.5 percent in 2022. The confidence declining partly due to the effect of the COVID-19 pandemic (see (Leyva intervals of these numebrs, however, overlap. Moreover, the comparison and Urrutia 2023). Some elements of these three factors, as well as some should be made with caution because the two datasets are not fully recent reforms on regulations about unions and about SMEs, may have comparable. played a role but further research is needed to identify the causes, and potential continuation, of declining informality rates. 90 MEXICO Poverty and Equity Assessment 17 There is a counter effect in the share of workers earning above two of up to four years to identify economic and demographic drivers of minimum wages. This is a group whose size has declined in relative poverty. She finds that poverty duration varies even for countries with terms (they represented more than 60 percent of employment in 2014, the same static poverty rate, hence the value of studying poverty but a little more than 30 percent in 2023) and hence, their relative dynamics. It is senior adults (aged 65 and more) who are more likely contribution to poverty reduction has diminshed. to be chronically poor in most countries (except in Romania, Spain, 18 The formulation of this decomposition is by (Datt and Ravallion 1992). Netherlands, Belgium and Portugal, where it is children who are more 19 Average hourly wage for employed women and men rose 4.7 and 4.5 per exposed to chronic poverty). This study also finds evidence of poverty year, respectively, between the 4th quarter of 2005 and the 4th quarter of persistence, that is most of the poor have been so for several periods 2023. Data from ENOE, Consulta interactiva de indicadores estratégicos: and their probability to exit poverty declines with poverty duration. The https://www.inegi.org.mx/sistemas/Infoenoe/Default_15mas.aspx ( rotating panels from ENOE used in this report extend for no more than a accessed August 15th, 2024). year (five consecutive quarters), and hence measure very short term 20 Data about gender wage gap from OECD (https://data-explorer.oecd. poverty dynamics, mostly associated with seasonal drivers of changes org/) and from CEPALSTAT (https://statistics.cepal.org/portal/cepalstat/ in household income. For a description of the construction of these index.html?lang=en ), refers to the difference between the median panels, see Annex 3. Description of panels using ENOE. wage of men and women as a percentage of the median wage of men, 29 A rotating panel is not necessarily representative of the whole for employees working full time (in the case of OECD) and all employed population. Annex 3 includes an argument in favor of these panels being in the case of ECLAC. Recent academic study by Cuéllar and Moreno not statistically different from the full cross sectional data in terms of the (2022) shows a decline in the gender wage gap in urban employment in distribution of key population groups. However, it is not argued that the Mexico. panels represent the whole population, and hence results are described 21 Data on tertiary enrollment rate from Mexico Gender Landscape brief, in terms of shares of the panel survey, rather than shares of the total downloaded from World Bank Gender Data Portal: https://genderdata. population. Moreover, the panel refers to the population aged 15 to 64, worldbank.org/en/economies/mexico . Data on upper-secondary net so it concentrates on the labor poverty status of those potentially active enrollment rate from CEPALSTAT (https://statistics.cepal.org/portal/ in the labor market. cepalstat/index.html?lang=en ), accessed (August 15th, 2024). 30 Demographic profiles of the chronic and transient labor poor in Annex 3. 22 Data from ENIF at https://www.inegi.org.mx/programas/enif/2021/ and Demographic profiles of poverty and chronic labor poverty. from World Bank Gender Data Portal: https://genderdata.worldbank.org/ 31 For a discussion of this definition of chronic poverty see (M. Ravallion en/economies/mexico . 2016), pages 231-233. The formulation of a decomposition of poverty 23 Data on dwelling titularity and time use from two CONEVAL studies into chronic and transient components as per this definition is by (Jalan Sistema de Indicadores sobre Pobreza y Género en Máxico 2016-2022 and Ravallion 1998). and Informe sobre Pobreza y Género 2008-2018. Una década de 32 The Gini coefficient is not decomposable by population groups whereas medición multidimensional de la pobreza en México. And from INEGI the Generalized Entropy is. (see M. Ravallion (2016), page 209). The ENADID, 2023 https://www.inegi.org.mx/programas/enadid/2023/ method used here for decomposing changes of inequality over time is 24 (World Bank 2023) by (Mookherjee and Shorrocks 1982). 25 (Banco Mundial 2020) can be found at: https://www.bancomundial.org/ 33 This is temporary as the share of more educated people becomes more es/region/lac/publication/la-participacion-de-la-mujer-en-el-mercado- prevalent, the effect would diminish and even reverse. laboral-en-mexico 34 Data from INEGI, in prices of 2022, Comunicado de Prensa Núm. 420/23, 26 The academic study cited is by (Calderón 2014). A more recent study of July 26th 2023, Table 9. on childcare services and female employment in Mexico is by (López 35 Unemployment rates averaged 4.8 between 2008 and 2014, and Acevedo, et al. 2021). A long term study on the forces driving changes then declined to 3.0 percent in 2022. Jobs in agriculture remained at in female labor force participation in Mexico by (Bhalotra and Fernández around 13.8 percent of total employment between 2008 and 2014, and 2024) underlines the importance of supply factors such as higher levels declined to 10.8 percent in 2022. Data for 4th quarter 2005-2023 from of schooling and lower fertility rates, but also indicate that other labor Interactive query of strategic indicators (InfoLaboral): (https://www. market factors such as wages, work benefits and changing social norms inegi.org.mx/sistemas/Infoenoe/Default_15mas_en.aspx ). may also play an important role, particularly in more recent years. 36 For this latter period, however, the impact is solely due to labor incomes. 27 A study on the impact of extended school time on female employment Changes in employment rates (green bar) would have increased is by (Padilla Romo and Cabrera Hernández 2018). inequality, though. In fact, employment rates declined from 75.2 to 28 The most comprehensive study on poverty dynamics in Mexico is 74.2 percent between 2016 and 2022. There is some evidence of by (Teruel 2022) who uses a purpose-defined panel over ten years, wage compression, as the proportion of workers earnings up to two including not only measures of monetary poverty but also of non- minimum wages increased from 40.2 percent to 64.6 for the period. monetary poverty defined as “carencias sociales”. It includes interviews A decomposition of changes in inequality à la Paes-De-Barros, shows in 2002, 2006 and 2012 and identifies factors that explain transitions that the share and average income of population earning leass than into and out of poverty in the medium term. One of the main findings two minimum wages also represent a growing component of changes of this study is that between 23.9 and 48.3 percent of the population in income inequality (figure not shown in this report, but available (depending on the method and definition) live in chronic poverty, and upon request from the authors). Data from INEGI, Encuesta Nacional 25 percent in transient poverty. The chronic poor have a demographic de Ocupacion y Empleo, (https://www.inegi.org.mx/programas/ profile where children, indigenous population, and population with enoe/15ymas/#tabulados on 07/15/24). low schooling are prevalent. Exiting poverty is mostly associated with 37 INEGI, Quarterly Gross Domestic Product, Income Method. Base good jobs (stable and well paid), whereas agricultural employment is year 2018, (https://www.inegi.org.mx/temas/pibti/#tabulados on closely associated with chronic and transient poverty. Studies based on 07/15/2024) rotating panels tend to concentrate on more short-term economic and 38 Recent studies including top incomes in Mexico (Campos-Vázquez , demographic factors. For instance, for a group of European countries, Chávez and Esquivel 2018) and (Reyes, Teruel and López 2017) find (Vaalavuo 2015) uses the SILC panel structure of repeated interviews higher and increasing levels of inequality when adjusting for under- 91 Notes captured incomes at the top of the distribution. Likewise, (Campos access to government financed-and-provided health care and allowed Vasquez and Lustig 2019) and (Puggioni, et al. 2022) report the for private sector provision for wealthier consumers. In this sense, a problems of measuring inequality in Mexico if using surveys with a measure of informality based on access to health insurance would entail large share of non-response data (these studies refer, instead, to the that Malaysia has a lower informality rate than Mexico. labor survey ENOE which is not intended to capture capital incomes, 43 See footnote 16. but suffers a similar problem of misreporting and non-response among 44 For an extensive discussion of the official method for measuring top-earners). multidimensional poverty in Mexico see (CONEVAL. Consejo Nacional 39 Brazil is closer to Mexico in terms of population size, but not in terms de Evaluacion de la Política de Desarrollo Social 2019). For a desciption of initial conditions. Argentina, had similar initial conditions and of the changes to the methods between 2010 and 2018 see Diario population size similar to the other comparison countries, but a very Oficial de la Federación, 30 de Octubre 2018: (https://www.dof.gob.mx/ unstable macroeconomic performance that hinders comparability. Chile, nota_detalle.php?codigo=5542421&fecha=30/10/2018#gsc.tab=0 ). instead, has been usually perceived as one of the most stable and best 45 For instance, educational gap affected 17.4 percent of the population in performing countries in the region. 2016 under the previous methodology, but 18.5 percent under the new. 40 There is a well-known formal mathematical connection between the Measure of lack of housing quality remained at 12.0 percent under both evolution of poverty rates and changes in average income and in methdos. In contrast, access to social security fell from 55.8 to 54.1. income inequality. In a recent paper, using global data for three decades, Historical series from CONEVAL at https://www.coneval.org.mx/Medicion/ (Bergstrom 2022) finds that “…the inequality elasticity of poverty MP/Paginas/Pobreza-2018.aspx (visited on July 1st, 2024). reduction is larger, on average, compared to the growth elasticity of 46 (Monroy-Gómez-Franco, Vélez-Grajales and López-Calva 2022) report poverty reduction and that the growth elasticity declines steeply that due to COVID-19 “the last quarter of the 2019-2020 cycle and the with a country’s initial level of inequality. However, we find that prior totality of the 2020-2021 educational cycle were conducted remotely”. changes in poverty were, in large part, explained by changes in mean The Mexican government put in place distance learning initiatives, but incomes.” The results from comparison countries above conforms it is estimated that important declines in the hours of study occurred, with the predominance of the growth effect upon poverty reduction as particularly among low income families. The authors estimate losses that documented by Bergstrom’s study. range from “a learning loss equivalent to the learning during a third of a 41 In general, minimum wages can potentially have a negative impact upon school year in the short run translates into a learning loss equivalent to employment, particularly in competitive labor markets, but not so if an entire school year further up the educational career of students” up monopsony conditions or other labor market imperfections characterize to “an entire school year and becomes a loss of two years of learning in the labor market, which may be the case particularly in developing the long run.” A World Bank note on the results of the PISA text for 2022 countries. See (Neumark and Munguía Corella 2021). For the case of indicates: “The pandemic exacerbated existing challenges of ensuring Mexico, (Campos-Vasquez and Esquivel 2023) and (Munguía Corella and equitable access to and completion of compulsory education. School Gómez Lovera 2023) find a positive effect of the increase of minimum dropout, already acute at the secondary level, was worsened, especially wages upon poverty reduction. Another recent study, by (Alvarado among vulnerable youth whose educational trajectories are most at risk.” Pérez, Orraca Romano and Cabrera-Hernández 2023) finds no significant Cited from Mexico PISA 2022 Brief, May 2024. impact of recent minimum wages increases on labor force participation 47 There are important differences in social deprivation, not only across or employment rates, but a positive impact on real wages, particularly groups defined by household incomes, but also across groups defined large for workers -both formals and informal- in the first quartile of the by sex, area of location and ethnicity. Data from CONEVAL for year wage distribution, which would be consistent with an important poverty 2022 shows that women have slightly lower rates of social deprivation reduction effect of minimum wages among low-earnings workers. than men: 24.0 (65.3) percent of women and 25.8 (68.6) percent Minimum wages can also affect real wages through price increases of of men have at least three (one) deprivations. But higher monetary products with a large share of minimum wage workers. The minimum poverty rates for women make that 36.9 percent of women be in wage policy was introduced in combination with a reduction of VAT in multidimensional poverty, whereas the rate is 35.6 percent for men. some parts of the country. This combination can help compensate the The rural-urban gap is wider: 50.6 (88.2) percent of population in rural price increase impact that minimum wages could produce. (Calderón, areas has at least 3 (1) deprivations, but 16.9 (58.4) percent do in et al. 2023) find that, for workers in the Free Zone area of the Northern urban areas, a difference of at least 30 percentage points. In the case of Border, real wages increased for all type of workers (low and high wage, ethnic groups, the difference is the starkest. Extreme monetary poverty formal and informal) confirming the compensatory effect of the VAT rates for indigenous and non-indigenous groups are 68.0 and 40.8 reduction in this case. percent, respectively, while the percentages of population with three 42 Informal employment may have different definitions in different deprivations are 60 and 21 percent respectively. Consequently, the countries. A measure produced by ILO to monitor the Sustainable extreme poverty rate among indigenous populations is 26.3 percent, Development Goal (SDG) 8.3.1 shows that Mexico has an almost double whereas for non-indigenous is 5.0 percent. See https://www.coneval.org. informality rate than the other countries (see SDG indicator 8.3.1, mx/Medicion/MP/Paginas/Pobreza_2022.aspx proportion of informal employment in total employment by sex and 48 For a brief description of the transition between Seguro Popular and sector at https://rshiny.ilo.org/dataexplorer21/?lang=en&id=SDG_0831_ INSABI, see footnote 48 below. SEX_ECO_RT_A ). ILO, however, does not produce this indicator for 49 There is academic evidence of an increase in catastrophic health Malaysia or Türkiye for comparable periods. A proxy of formal expenditures during the initial stages of the COVID-19 pandemic and the employment (percentage of workers covered by insurance against institutional transformation of Seguro Popular into INSABI. See (Serván- work injuries) is included in Figure 14, bottom right panel. This indicator Mori, et al. 2023). seems to show that Malaysia has as high an informality rate as Mexico. 50 Data from (CONEVAL, Consejo Nacional de Evaluación de la Política de However, in the case of Malaysia, (Croke, et al. 2019) argue that – in Desarrollo Social 2023), pages 43 to 53; INEGI. Encuesta Nacional de contrast to Latin American countries where a fragmented system Ingresos y Gastos de los Hogares. ENIGH 2016, 2018, 2020 and 2022. provides unequal coverage to different social groups – Malaysia started Tabulados básicos (https://www.inegi.org.mx/programas/enigh/nc/2022/ from a progressive system, a system that achieved early universal accessed August 15, 2024) and (World Bank 2023) pages 24-27. 92 MEXICO Poverty and Equity Assessment 51 Data for 2024 corresponds to the federal budget. Most of the analysis in 62 These simulation exercises are done with MEXSIM microsimulation this section draws from background paper specially commissioned for model. The microsimulation consists in comparing poverty/inequality this Poverty and Equity Assessment by (Macías, et al. 2024). indexes of the population with and without the specific cash transfer. 52 There are several institutions linked to the funding and provision of These are accounting exercises and assume no behavioral reaction, on health services to the uninsured. Institute of Health for Wellbeing (in the part of beneficiaries, to changes in the transfers received. Hence Spanish, Instituto de Salud para el Bienestar – INSABI) was created these simulations should be interpreted as upper-bound limits of the to replace the National Commission on Social Protection in Health (in potential poverty/inequality effect of these programs. Spanish, Comisión Nacional de Protección Social en Salud – CNPSS), 63 The Kakwani index is a widely used summary measure of progressivity commonly known as Seguro Popular. In principle, INSABI is responsible in the academic literature. It is calculated by subtracting the for channeling federal resources and coordinating the care to be concentration coefficient of the policy intervention from the pre- provided by the state health services for the population without access intervention Gini of the welfare aggregate in the case of a transfer, and to social security. As of April 29th, 2023, an initiative to disappear INSABI the inverse in the case of a tax. Progressive (regressive) interventions and integrate its functions and resources into IMSS-Bienestar and the have a positive (negative). See (Kakwani 1977). Ministry of Health, had been approved by both the Mexican Chamber 64 The average transfer for a population in a given group “q” , which is of Deputies and the Senate. IMSS-Bienestar was originally created in connected to the concept of sufficiency because the higher this average 1979 and is a federally funded program providing health services to benefit the closer it may be to the poverty line, can be formulated as: deprived populations (and thus without access to social security), both benefitsq benefitsq populationq budget in rural and urban settings, and run by IMSS. Provision of the service is = x x beneficiariesq budget beneficiariesq populationq also provided by the Ministry of Health, through the National Institutes of Health and other Federal and High-Specialty Regional Hospitals, as where the first term to the right represents absolute incidence (the well as by State Health Services (in Spanish, Secretarías de Salud de share of the budget allocated to group “q”), the second term is the los Estados – SESA) which are 32 local Ministries of Health, who provide inverse of the coverage of the program of group “q” (the proportion care through their own facilities for the population without access of population of group “q” who are beneficiaries of the program) and to social security. Cited from World Bank (2023), “Health Financial the third is a measure of available resources. The average transfer to Sustainability and Resilience Assessment Report” beneficiaries in group “q” can be made higher (and hence increase 53 Relative progressivity refers to higher (lower) transfers/subsidies as the sufficiency of the transfer to reduce poverty) the higher any of a percentage of household income/consumption for those with lower these three components. Given the population in group “q” is fixed, and (higher) household income/consumption. Absolute progressivity full coverage of target group “q” is intended (otherwise some of the refers to a higher percentage of the total budget in transfers/subsidies beneficiaries do not get benefits and would not reduce or exit poverty) to households in lower (higher) quantiles of the household income/ the policy levers depend on either increasing absolute incidence, or the consumption distribution. budget of the program, or both. 54 Gross enrollment rates by education level from World Development 65 These are estimates of budgetary efficiency of a program. These are Indicators (https://databank.worldbank.org/source/world-development- derived from an accounting exercise of dividing budgetary allocation indicators accessed August 15th, 2024) as per administrative data by estimates of poverty rate change with 55 Data from “Mexico Learning Poverty Brief” (World Bank, June 2022) and or without the program as per survey data. These are different from “Mexico, PISA 2022 Brief” (World Bank April 2024). impact evaluation of the programs which involve experimental design 56 For a detailed analysis of the Mexican pension system, see (Villareal and to identify a statistically significant difference in outcomes of interest Macías 2020) (e.g., poverty rate, education level, employment) between comparable 57 This is partly due to different funding and distribution designs across beneficiaries and non-beneficiaries of a program. Given its recent pension systems. Under DB, a worker can achieve up to a 100 implementation, most of these programs do not yet have a scientific percent replacement rate, sometimes more (like for Pemex and CFE). impact evaluation. This should be an important research program in Conversely, most workers under DC typically achieve a 40 percent coming years. replacement rate, with voluntary contributions potentially increasing 66 A full description of the MPM is in (World Bank 2018), also at: https:// this rate. After the 2021 pension reform, the replacement rate for DC www.worldbank.org/en/topic/poverty/brief/multidimensional-poverty- increased for those who began contributing after 1997. See (Macías, et measure al. 2024) and (Villareal and Macías 2020). 67 Formally, these three components are derived from the following 58 Number of beneficiaries as of 2023, as per official report “Informe de identity: Avances” at: https://programasparaelbienestar.gob.mx/wp-content/ uploads/2024/01/25012024-Avances-Programas-Para-El-Bienestar.pdf SX SX BD GDP = x x (accessed on October 1, 2024). POP BD GDP POP 59 In this case however, the program coverage is more sizeable (about 1.5 where SX stands for social expenditures, BD for fiscal budget, GDP for million youth); but ENIGH 2022 fails to capture most of them, and these gross domestic product (or some other measure of aggregate national estimates can be biased downwards. income) and POP stands for population. Taking logarithms and a first 60 Data from Presentation of Results documentation at https://en.www. derivative renders the following approximation: inegi.org.mx/programas/enasic/2022/#documentacion (accessed on October 1, 2024). SX SX BD GDP ∆% = ∆% x ∆% x ∆% 61 These two groups of programs correspond to expenditures under POP BD GDP POP chapter 4000 of ramos 20, 10 and 14, modalidad S. See (Macías, et al. namely, the percentage change of the term to the left equals the sum of 2024). the percentage changes of the three components to the right. 68 M. Ravallion (2016), page 270. 69 Hallegatte, Fay and Barbier (2018). 93 Notes 70 For a review of different definitions in different disciplines, see Fuchs 82 Data mentioned refers to year 2020. The methodology and data for the and Thaler (2018). The World Bank approach to the relation between Longitudinal Social Gap Index (Rezago Social Longitudinal) is at https:// climate hazards and equity is from (Brunckhurst, et al. 2023) www.coneval.org.mx/Medicion/Paginas/IRS_longitudinal.aspx 71 For recent documentation of increased climate disaster events in Mexico 83 In its regular estimates for Latin American countries, the World Bank and its impact on the population see (Ibarrarán, Chavarría and Zúñiga uses a single poverty line for urban and rural areas but adjusts the rural 2020) and (De Lima, Abeldaño Zuñiga and Jerez-Ramírez 2023). See welfare aggregate by a factor of 15% to compensate for price differences also about the warmest conditions in Mexico vis-avis other countries in across regions. See Annex 1. Comparison of monetary povery rate the region at (WMO 2023). estimates from CONEVAL and the World Bank 72 A detailed explanation of the method and of thresholds defintion of 84 Data processed using ENIGH 2022.. extreme events in (Doan, et al. 2023). In this report we use maps 85 An important element related to agricultural productivity in Mexico for Mexico included in the forthcoming World Bank regional study is that public investments in irrigation systems have declined over (Climate hazards and poverty in Latin America) which follows this the course of the last 10 years and there are also unsustainable same methodology but with two differences: the maps are produced subsidies (electricity for pumping, for instance) that have a regressive at municipalities rather than provinces, and an additional hazard is distributional impact. Large commercial farmers with high technology included (i.e., landslides). and irrigation systems benefit, while family farmers receive a smaller 73 The probability of a household’s income to be higher than the poverty share because they don’t have access to or cannot afford water pumps. line in any given year is P = Prob (ln (income) i > ln (poverty line)). Then, In addition, in the face of climate change which puts additional pressure the probability of becoming poor in the next two years is represented by on water resources, poor farmers are likely to be most affected by vij,t+2 = 1 - P 2 ≥ 0.5, for a 0.5 threshold. Solving for P then P = √0.5 = 0.71, aquifer overexploitation in the future as they lack the capital to adapt which denotes that P = 0.71. This further implies that the probability of to falling water tables. Rethinking these subsidies and investment a household’s welfare falling below the poverty line in any given year is programs to cope with climate vulnerability could further help with 29% or 0.29, as 1 - P = 0.29. poverty reduction among rural population and agricultural workers. See 74 See Canavire-Bacarreza and Gayoso de Ervin (2024). This study uses (World Bank 2009) and (Sánchez, et al. 2018). a long list of climate-related events: floods, droughts, fires, electrical 86 A World Bank study on social inclusion of indigenous peoples in storms, landslides, etc. Another study, De la Fuente and Serio (2024), Mexico in Latin America is “Indigenous Latin America in the Twenty First using a similar methodology but only four hazards (floods, droughts, Century”. hurricanes and heatwaves) finds 27 percent of the population are 87 The method for this counting is by (Doan, et al. 2023). Other methods vulnerable to poverty which is close to the estimates by Canavire and estimate vulnerability as probabilities of loss of assests or loss of Gayoso. A similar study including Mexico and other Latin American incomes associated with specific events, climate or not climate related. countries is by Canavire et al. (2024). For climate vulnerability see (Hallegatte, Vogt-Schilb, et al. 2017) or 75 A recent study documenting the impact of climate events, migration and (Skoufias, Vinha and Beyene 2023), for vulnerability to poverty in public programs in Mexico is by (Chort and de la Rupelle 2022). Also see general (Chaudhuri 2003) . (Murray-Tortarolo and Martínez Salgado 2021). These studies find that 88 For full details of the measurement of hazards and vulnerabilities, as extreme events (particularly droughts) are associated with migration well as data sources, see (Doan, et al. 2023). patterns from Mexico to the US, and also suggest that policies oriented 89 It could be argued that the international comparison in this case is not to reduce the negative impacts of these events could be effective in completely valid because Mexico’s geography determines its exposure. reducing migratory peaks. However, Poland has a population exposure rate similar to Mexico’s 76 For details on the method, see Alonso Beltrán, et al. (2024). This is a and has a much smaller vulnerability rate. More similar, however, is the background paper for the regional study World Bank (forthcoming) case of Romania with a 51 percentage of the population exposed, and on climate hazards and poverty in Latin America. The study offers a still a lower 29.4 percent vulnerability rate, due to much lower rates of comparison between Mexico, Colombia and Chile. Its simplicity allows population without social protection or access to finance. Egypt, with for comparison across countries with different abundance and quality an exposure rate of 44 percent and a vulnerabiltiy rate of 36.7 percent of data, but at the same time gives a granular and intertemporal is more similar to Mexico, not because of similar exposure rates but assessment of changes in climate risks. because of similar lack of access to finance and higher poverty rates 77 A municipality is defined as rural if more than half its population is rural, than Mexico.Data from (Doan, et al. 2023). as per 2020 census data. 90 Global poverty eradication referred to an extreme poverty rate of less 78 For the definition of measures of territorial context see Mexico than 3 percent. The World Bank Poverty and Shared Prosperity reports see (CONEVAL. Consejo Nacional de Evaluacion de la Política de for years 2018 and 2020 warned that the global goal was slipping Desarrollo Social 2019) and also Diario Oficial de la Federación, away. The World Bank Poverty and Shared Prosperity reports for 2022 30 de Octubre 2018: https://www.dof.gob.mx/nota_detalle. and 2024 estimate that the COVID-19 pandemic has made the goal php?codigo=5542421&fecha=30/10/2018#gsc.tab=0 ). This practically unattainable. subsection makes use of data on territorial context by CONEVAL from: 91 Although not fully comparable to a target of eradication as per official https://www.coneval.org.mx/Medicion/Paginas/Plataforma-Analisis- poverty measures, this exercise aims to illustrate the bounds of Territorial-de-la-Pobreza.aspx and also from (CONEVAL. Consejo Nacional necessary policy action to achieve extreme poverty eradication. de Evaluación de la Política Social 2023). Given that the US$ 6.85 (2017 PPP) is higher than the official extreme 79 See method on counting the vulnerable to climate hazards by (Doan, et monetary poverty line, this target would be very similar to a target of al. 2023). extreme poverty eradication as per CONEVAL’s definition of extreme 80 (CONEVAL. Consejo Nacional de Evaluación de la Política Social 2023), poverty. For a comparison of World Bank and Mexico’s official monetary page 7. extreme poverty line see Annex 1. Comparison of monetary poverty rate 81 The methodology and data for the index of access is at https://www. estimates from CONEVAL and the World Bank. coneval.org.mx/Medicion/Paginas/Grado_accesibilidad_carretera.aspx . 92 The projections use methods and code developed by (Lakner, et al. 2022). The authors propose a method that implies a fixed tax rate to 94 MEXICO Poverty and Equity Assessment every individual, whose total proceeds are then equally divided among 102 See (Macías, et al. 2024). all the population. It can be proved that, under certain conditions, 103 Methods adopted by CONEVAL are at: https://www.coneval.org.mx/ this tax rate is equal to the percentage change in the Gini coefficient. Medicion/MP/Paginas/Metodologia.aspx. For methods adopted by the Hence a decline of 1 percent in the Gini is implemented in the data as World Bank vist: https://datanalytics.worldbank.org/PIP-Methodology/ a 1 percent tax upon income for all people whose proceeds are equally 104 The “upper-middle income” poverty line amounts to US$ 6.85 per day, distributed across the population. in terms of 2017 purchasing power parity rates and is derived from 93 These projections do not include changes in demographic structure. the median values of national poverty lines of upper-middle income GDP growth and nowcasts for years 2023 and 2024 comes from Macro countries. The global poverty line is defined as US$ 2.15 per day, in and Poverty Outlook from Spring 2024 (https://www.worldbank.org/en/ terms of 2017 purchasing power parity rates. It is an update of the publication/macro-poverty-outlook accessed May 15, 2024). famous “dollar a day” benchmark introduced in the World Development 94 A 3 percent GDP/head growth is the 68th percentile of the distribution Report of 1980 on Poverty, as a combination of poverty lines of the among upper middle income countries for the 2002-2022 period. poorest countries in the world. These lines have been computed and A 2 percent annual decline in the Gini is the 95th percentile of the updated using a consistent methodology, see (Jolliffe, et al. 2022).. distribution for upper-middle income countries. Author’s calculations 105 For a description of the monetary and multidimensional measures using data from World Development Indicators ( https://databank. poverty adopted by the World Bank, and its global, regional and country worldbank.org/source/world-development-indicators accessed August estimates, see (World Bank 2018). For more recent estimates visit World 1, 2024) Bank Poverty and Inequality Platform https://pip.worldbank.org/home. 95 (Canavire-Bacarreza and Gayoso de Ervin 2024) and (De la Fuente and 106 For a description of the official poverty lines in Mexico see: https://www. Serio 2024) using econometric methods, and (Doan, et al. 2023) using a coneval.org.mx/Medicion/MP/Paginas/Lineas-de-Pobreza-por-Ingresos. counting of indicators method. aspx. For a descripton of the differnet poverty lines by the World Bank 96 As explained in (Doan, et al. 2023), reducing vulnerability involves and their valuation under the 2017 PPPs see (Jollife, et al. 2022). having access to physical infrastructure, as well as social protection, 107 This procedure is adopted to data for all Latin American countries, to cope with the consequences of a climate shock. These are the following the protocols of SEDLAC, Socio-Economic Database for Latin components that provide resilience to those vulnerable to these events America and The Caribbean, a joint project by The World Bank and Centro so that they can recuperate their standards of living as soon as possible de Estudios Distributivos, Laborales y Sociales, CEDLAS, Universidad after the shock, preventing them from falling into poverty. Nacional de La Plata, Argentina. As indicated in the methodological 97 Data from Cuenta Pública 2008-2023, processed and analyzed by guide: “In this database all rural incomes are increased by a factor of 15% (Macías, et al. 2024). to capture differences in rural-urban prices. That value is an average of 98 A critical review of the potential of social protection systems to address some available detailed studies of regional prices in the region. Although the risks from climate change is by (Costella, et al. 2023). A review of certainly arbitrary, we believe this alternative is better than (i) ignoring how social protection responded to the COVID-19 shock is by (Stampini, the problem of regional prices altogether, or (ii) using for each country et al. 2021). A description of the vulnerable population and their the available price information, despite the enormous differences in insufficient coverage by existing social programs in Mexico is by (De La methodology, scope, and results“ (see https://www.cedlas.econo.unlp. Fuente, Ortiz-Juárez and Rodríguez-Castelán 2018) edu.ar/wp/en/estadisticas/sedlac/metodologia-sedlac/ ) 99 A global perspective on the four building blocks of an Adaptive Social 108 For a technical explanation see (M. Ravallion 2016), section 5.5 “The Protection System is in (Bowen, et al. 2020). Definition from Op. Cit. page robustness of Poverty Comparisons”. 3. 109 A similar exercise was done for the ENIGH 2014 and the results were 100 (Solórzano and Goode 2023) and (Carter and Martínez Sugastti 2024) qualitatively similar. In addition, a stochastic dominance of Lorenz 101 Two important experiences of climate disaster mechanisms in Mexico curves was done for years 2014 and 2022. In both years, the Lorenz are FONDEN and CADENA. FONDEN (Fondo para Desastres Naturales) curves cross and hence there is no stochastic dominance of one method was created in 1996 and phased out in 2021. It was established with on the other and no systemaitc message about Lorenz-consistent the primary purpose of providing financial resources for the rapid inequality indicators can be said about the two methods. The differences rehabilitation and reconstruction of federal and state infrastructure, low- were minimal: between 0.4 and -0.2 percentage points of cumulative income housing, and certain components of the natural environment welfare. Results are available from the authors upon request. affected by adverse natural events. It has often been analyzed in 110 Data was downloaded on February 15, 2024. the academic press as an example of ex-ante disaster risk finance 111 The Stata code is available here: https://www.coneval.org.mx/Medicion/ instrument effective in reducing vulnerability to climate shocks. See Paginas/ITLP-IS_pobreza_laboral.aspx (Del Valle, De Janvry and Sadoulet 2020) and (Del Valle 2024). CADENA 112 The World Bank staff harmonized the first quarter of each year, from (Componente de Atención a Desastres Naturales), launched in 2003 and 2005 to 2020. We incorporated harmonization code for the second, terminated in 2021, offered relief in the event of natural disasters, such third, and fourth quarter of each year. Additionally, we added all surveys as droughts, floods, and hurricanes for small-scale, vulnerable farmers after 2020. The harmonization code is in the process to make it publicly who lacked access to credit or commercial insurance. At its peak in 2018 available. it provided index insurance protection to about 2.5 million small crop and livestock producers across 31 states. Its impacts are documented by (Fuchs and Wolff 2016). 95 Notes 96