IMPROVING HOUSING RESILIENCE SYNTHESIS REPORT September 2019 This Synthesis document was produced by Luis Triveno, Gabriel Arrisueno and Sarah Antos based on (i) the technical work produced by Build Change, Miyamoto International, SARC, Development Seed, Analisis Geograficos SAC, Alejandro Cantera, Carlos Ariel Cortes, Elena Tames and Salvador Herrera; and (ii) the data collection effort that the team conducted in the Istmo de Tehuantepec Region in Oaxaca. Ming Zhang (Practice Manager, LAC Urban and DRM) and Horacio Terraza (Lead Urban Specialist) provided guidance. Carina Lakovits, Phoebe Spencer, Miguel Rodriguez and Jonathan Hasoloan, provided support during the data collection and/or the elaboration of this report. The drone imagery was produced by Adam Benjamin. Xavier Conesa was responsible for the design and most photographs. Claudia Pacheco and Diana Jimenez provided logistic support. The data presented in this document has been obtained from official sources that we consider reliable. Estimates have been produced based on conservative assumptions. The drone and street-view imagery, training data and algorithms are all open-source and available upon request. 2 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | TABLE OF CONTENTS Executive Summary5 The Challenge11 Affordable Housing Program (Programa de Vivienda Social)16 The National Reconstruction Program 16 The Urban Upgrading Program 18 The Opportunities 23 Gap identification 23 —Institutional gap 23 — Policy gap 26 — Resource gap 30 Methodology development and case study 31 Conclusions41 Annex 1: Imagery Collection and Analysis 43 Annex 2: Main actors in the Mexico housing sector 45 Annex 3: A note on terms and definitions 47 Annex 4: Glossary of abbreviations 49 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 3 FOR THOSE WITH THE LOWEST INCOME LEVELS, THE QUALITY OF THE HOUSING SOLUTIONS OFFERED IN THE MARKET IS TOO LOW AND THE UNITS ARE TOO SMALL AND DISTANT. 4 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | EXECUTIVE SUMMARY This synthesis report summarizes the main findings of the MX Im- HOUSING POLICY IN NUMBERS 1973-2018 proving Housing Resilience ASA (P169278). The ASA was prepared • 45 years. to provide technical guidance to the National Housing Commission • 122 entities. (CONAVI) and the Ministry of Agrarian, Land and Urban Develop- • 30.4 million of housing actions. ment (SEDATU) in their efforts to increase resilience by (a) improving • New housing: 52% of the housing actions, 88% of the targeting of housing subsidies; (b) strengthening municipal ur- the budget. • Home improvement: 40% of the housing actions, ban management and (c) improving the design and implementation 4% of the budget. of CONAVI’s existing home improvement subsidy scheme. • Other: 8% of the actions, 8% of the budget. SOURCE: CONAVI Mexico has developed a solid legal and institutional housing framework over the past decades and made significant progress in terms of expanding access to formal and affordable housing. Mexico’s housing policy has been successful in increasing access to housing for low and middle-income families, through the provision of targeted technical assistance and increased access to housing loans or microcredits. The creation of the National Housing Com- mission (CONAVI), the adoption of a new housing law in 2006 and the introduction of a demand subsidy program targeted at low-in- come families by CONAVI all contributed to expanding access to There are 5 million abandoned homes in affordable housing in Mexico. Mexico. 80% abandoned due to bad location and lack of access to jobs, However, important challenges in terms of meeting existing education and health services. housing needs in Mexico remain. In 2016, according to the SNIIV Mexico had a housing deficit of 9.2 million households, which rep- resented 28% of the total number of households estimated for that year (32.8 million). The housing deficit among indigenous commu- nities is severe, reaching 79.1% according to CONEVAL. Households headed by women, the elderly or disabled are particularly likely to live in inadequate homes (CONEVAL, 2018). In terms of targeting, most housing subsidies were provided to formal workers affiliated to the social security system. In 2016, 68 million Mexicans (53% of the population) were not affiliated to the social security system and therefore ineligible to receive loans from entities such as INFON- AVIT or FOVISSSTE. In addition, the share of housing finance tar- AN EXAMPLE OF MEXICO’S PIONEERING geted at self-construction —the main way the poor access housing— HOUSING SOLUTIONS Piso Firme Program offered households with dirt floors remains very small. In the period 2007-2019, only 14% of CONAVI’s up to 50 square meters (538 square feet) of concrete subsidies went towards self-construction (CONAVI, 2019). cement flooring. IMPACT: • Reduction in diarrhea (49%), anemia (81%) and Exposure to natural and climatic disaster risk exacerbates the parasitic Infestations (78%) In children. precariousness of people’s housing. Around 41 percent of Mexi- • Reduction in depression rates (52%), increase their co’s territory and 31 percent of its population are exposed to hur- satisfaction with their homes (59%) and their lives (69%) in adults. ricanes, storms, floods, earthquakes, and volcanic eruptions. In economic terms, this translates into 30 percent of GDP considered | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 5 HOUSING ACTIONS to be at risk from three or more hazards and 71 percent at risk from Subsidies and Loans two hazards or more. Today, 7 out of 10 Mexican live in areas of YEAR TOTAL ACTIONS (MILLION) TOTAL (USD BILLION) high natural and climatic disaster risk. The location of housing in 2007 1.27 13.2 high risk areas, and the precarious structure of dwellings, expos- 2008 2.01 14.2 es especially vulnerable groups and low-income communities. As 2009 1.66 12.3 more and more people move to the cities, the percentage of people 2010 1.74 12.9 2011 1.59 12.9 living in areas exposed to floods, landslides, tsunamis and other 2012 1.78 13.2 natural phenomena continues to increase. 2013 1.42 13.9 2014 1.38 16.1 For low-income Mexican families already with limited access to 2015 1.37 17.0 mortgage markets, self-construction and incremental housing 2016 1.26 17.4 have become the preferred solutions to their housing needs. 2017 1.17 18.0 However, the quality of this type of housing is often low, with units 2018 1.22 18.5 small and distant. Furthermore, the neighborhoods lack basic ser- 2019* 0.47 8.65 vices, green spaces and access to jobs, education and health ser- TOTAL 18.35 188.3 vices. Mexico’s high number of abandoned homes suggest that that * Until June 30th, 2019 SOURCE: SNIIV these homes were not built where they were needed: according to INEGI, there are 5 million abandoned homes in Mexico. For low-in- come Mexican families already with limited access to mortgage markets, self-construction and incremental housing have become the preferred solutions to their housing needs. The Mexican Government that took office in January 2019 rede- fined the focus of Mexico’s housing policies and prioritized re- construction, resilience and social inclusion as its most import- ant objectives. This renewed focus will enable to the government to reach the poorest and most vulnerable populations that so far Mexico: 64.2% of housing units were self have largely been excluded from official housing assistance pro- or incrementally built with or without grams. The aim is to address existing geographical imbalances in the provision of housing assistance; improve the institutional sup- technical assistance. port for home improvement and housing retrofit; and strengthen household-led self-construction processes, amongst others. The 6 States that explain 48.2% of the housing deficit (2016), concentrated In order to assess the vulnerabilities of the housing stock and 24.4% of the housing actions between identify opportunities for simple, affordable and effective hous- 2007 and 2018. ing retrofit, the World Bank conducted a pilot intervention in se- lected municipalities of the Istmo de Tehuantepec region of Oax- aca. The Bank team customized the methodology developed by the Global Program for Resilient Housing to capture detailed informa- tion about the housing stock and identify opportunities for strength- ening existing structures and other improvements. The aim was to test a methodology that could be scaled up and replicated in other parts of the country, particularly regions exposed to high natural disaster and climate risks. This pilot intervention was complement- ed by an extensive review of the Mexican experience of providing housing assistance to identify what works and what doesn’t work in public housing assistance for the poor and most vulnerable. The piloted approach combines the application of geospatial technology to characterize the housing stock and assess its lev- els of vulnerability, analysis of census data and hazard risk maps, 6 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | OPPORTUNITIES TO CLOSE THE INSTITUTIONAL, POLICY AND RESOURCE GAPS INSTITUTIONAL GAP POLICY GAP RESOURCE GAP Strengthen interinstitutional Calibrate the portfolio of housing subsidies. Establish a comprehensive diagnostic of the housing coordination. sector in the country. Strengthen hazard and planning Adjust housing assistance programs to effectively Adopt a robust and simple methodology to assess information and integrate housing address housing needs of the most vulnerable the vulnerability of the housing stock. and urban policies. segments of the population. Build the capacity at the federal, Strengthen housing resilience by integrating Design technically and economically viable self- state and municipal level in the retrofit measures into home improvement construction and home improvement instruments. housing and planning sector. activities. Prioritize those geographical areas with the largest housing deficits. Strengthen the ecosystem for the Harmonize and coordinate urban and housing Promote financial inclusion by designing savings, provision of specialized technical policies, as well as housing retrofit interventions credit and insurance instruments that suit the needs assistance and tailored housing with urban upgrading and other programs. of the different population segments. assistance. and selective qualitative assessments to provide a picture of the THE COMPONENTS OF THE METHODOLOGY FOR HOUSING DATA reconstruction and retrofit needs in the area with an unprece- CAPTURING dented level of detail. All existing hazard maps were reviewed and integrated in a single geospatial database. In addition, drone imagery captured roughly 30 sq km of land, and street view cameras cap- tured 580 km of road. To identify potential alternatives for structural strengthening, more than 100 general and 40 detailed engineering assessments were performed. Probabilistic risk models were used to evaluate the efficacy of allocating more public funds in an area of interest in Oaxaca (26 districts in Salina Cruz) to making homes safer. The piloted approach demonstrated that simple retrofit inter- vention targeting the most vulnerable housing structures could Drone imagery. reduce estimated potential losses in case of earthquakes by half. For example, it was found that targeted retrofit interventions fo- cusing on the 18% of unreinforced or semi-reinforced housing units and completing their confinement reduces the estimated potential losses in case of an earthquake by half. Based on these characteri- zations, the team proved that simple and small investments can, at the same time, improve the quality of life of families and significant- ly reduce the expected losses in the event of a disaster. Based on the review of Mexico’s experience of providing housing assistance and the findings generated by the successful pilot inter- Street view imagery. vention in the Istmo de Tehuantepec region, the Team arrived at the following conclusions: Logo 90% Building: confined masonry 93% Window 72% Buildings 85% Merchandise 77% Commerical sign 85% • The Mexican government has correctly identified the need to Window 75% Street sign 69% Commerical sign 45% focus on sustaining reconstruction efforts and promoting resil- Door 96% Person 72% Car 69% Truck 92% ience and social inclusion. Car 98% Person 81% Person 78% Person 96% Merchandise 83% Merchandise 83% Merchandise 71% • However, the Government needs to address three major gaps in the country’s housing assistance framework for its new strate- gy to succeed; these include: (i) an institutional gap, which hin- ders inter-agency coordination, effective policy implementation Machine learning. and private sector participation; (ii) a policy gap, characterized by uncoordinated housing and urban policies, shortcomings in the design of housing subsidies and assistance programs, and a lack of policy provisions for residential retrofitting; and (iii) a | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 7 THE POSITIVE IMPACT OF RETROFITTING resource gap, referring to a lack of available data and analysis With retrofitting (confined masonry) Without retrofitting about the housing stock, as well as need to expand the toolbox USD MILLION of the Government to tailor housing assistance mechanisms to 1,000 the needs of the most vulnerable population and expand access to flexible financial services to those most in need. 800 600 • The Government can close those gaps by leveraging the expe- rience and reach of Mexico’s institutional setting, the lessons 400 learned through decades of implementation of home improve- ment programs, international efforts at house retrofitting, along 200 with the existing hazard information and the application of new 0 5 10 20 50 70 100 200 500 700 900 1000 technologies and geospatial data collection methods. The inter- PERIOD OF RETURN vention in the Istmo de Tehuantepec piloted an innovative meth- odology that could help CONAVI and SEDATU reach the most vulnerable segments of the population with a strong value prop- osition, namely one that combines legal and physical security, quality of life and social welfare. It was found that through sim- ple yet targeted home improvement and retrofit interventions, household’s exposure to risk could be reduced by half. At the same time, the data and information generated by this method- ology about the living conditions of the beneficiaries can inform government programs in other sectors. Confined masonry • Budget constraints on government spending, including in the housing sector, highlight the need for an active participation of the private sector (households and companies). This is consis- tent with international best practices, which show that structur- al reinforcement programs can incentive private participation if properly designed and communicated. • The Bank’s methodology used in this ASA confirmed that achiev- ing these goals is affordable, reliable, and scalable. The neces- Unconfined masonry sary fieldwork, which we conducted in Oaxaca in just weeks, could be replicated in the States that have been designated as HOUSING UNITS AFFECTED BY THE 2017 EARTHQUAKES priorities in the National Reconstruction Program (Puebla, Guer- FEDERAL STATE HOUSING UNITS rero, Morelos, Chiapas, Mexico, Oaxaca); as well as the States in PARTIAL DAMAGE TOTAL DAMAGE the South which concentrate the country’s poorest populations Estado de México 6,059 2,468 (Oaxaca, Chiapas, Guerrero, Veracruz, among others). Morelos 5,765 2,273 Puebla 27,812 3,214 • Experiences from countries with extensive data on housing risk Oaxaca 63,336 21,823 management, such as the United States and Japan, indicate that Chiapas 59,397 18,058 it is key to use cutting-edge technology complemented with other Tlaxcala 34 0 sources of information, to identify risks and report an efficient al- Guerrero 2,976 1,451 location of resources for prevention. Acceptance of stakeholders SENATE FROM MEXICO, 2017. is essential; the design and execution of effective communication AN EXAMPLE OF A STATE IN THE SOUTH and coordination strategies must be a priority; as well as carrying • Housing deficit in Chiapas: 78.2% of the total. out strengthening activities to improve regulatory capacity; offer • In rural housing: 96.5%. a flexible menu of financial assistance schemes; maintain a long- • In indigenous communities: 79% term horizon for housing programs; develop and update regulato- ry instruments such as codes and guidelines; and pilot innovative schemes that allow gradual adjustments to existing programs. 8 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | THERE IS A CLEAR OPPORTUNITY TO DESIGN INNOVATIVE HOUSING SOLUTIONS CAPABLE OF REACHING THE BOTTOM-40 SEGMENT OF THE POPULATION. For future Bank engagements in Mexico, there is a clear opportuni- ty to work together with authorities (i) to preserve and leverage the solid institutional framework in the housing sector that Mexico has developed in recent decades; (ii) to design innovative housing solu- tions capable of reaching the bottom-40 segment of the population; and to (iii) help the country to expand its infrastructure —instead of continually having to rebuild it after disasters. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 9 FOR THE MOST VULNERABLE, THE “MEXICAN MET. HOUSING MIRACLE” IS STILL A PROMISE UN­ 10 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 1. The Challenge Mexico is well-recognized in Latin America for its success in ex- POPULATION GROWTH (%) panding the supply of formal and affordable housing in the face of 1950= rapid urbanization: between 1950 and 1970, Mexico’s urban popu- 1970: > lation more than doubled. Over the past 45 years, due to the action 2010: of multiple public and private institutions, more than 30 million 2030: > homes in Mexico have been built or improved through different fi- nancing mechanisms. In the last 12 years alone, a total of USD 182 URBAN GROWTH: 2010-2030 billion was allocated to finance the construction or improvement of CITY HIERARCHY 2010 2030 more than 18 million houses. 80% of financing for home purchas- Mega City 10 million 1 1 es and improvements came from public entities, including public Major Cities 1 to a 5 million 10 17 housing funds, and only 18% from private sector finance1. Intermediate Cities 500k to 1 million 22 18 These remarkable results, nonetheless, have been insufficient to Medium Cities 100k to 500k 62 76 address the current housing deficit, while accommodating the for- Small Cities 50k to 100k 40 102 mation of new households —particularly from the lower income seg- Urban Centers 15k to 50k 249 747 ments of the population or the most vulnerable— in a country where TOTAL 384 961 78% of the people live in cities. And there are reasons for this: in the last 12 years, while 51% of households in the country were classified THE HOUSING DEFICIT IN MEXICO as poor, only 27% of all housing actions were directed towards them. In 2000, the housing deficit was estimated to be 44.9% of the housing stock. As a result, for the most vulnerable population, such as indige- % OF TOTAL MILLION OF nous people, the “Mexican housing miracle” is still a promise un- UNITS met. For example, while 45% of all Mexicans experience a housing SNIIV, based on census data 2015*, 28.1% 9,2 own methodology deficit, it’s 79% among the indigenous. And while only 15% of Mex- Federal Mortgage Society, based on 28,1% 9,2 ican households lack access to basic services in their homes, 56% census data 2015, own methodology of indigenous households experience this problem. A similar gap National Housing Program 2014-2018, 49% 15,3 based on census data, 2012 exists between formal workers and unaffiliated ones. While only Coneval, based on census data, 2015, 44.7% 14 25% of households where the breadwinners have formal jobs face own methodology a housing deficit, almost 75% of the households where unaffiliated UN Habitat 2018, based on their own 38% 12,6 estimations workers live face this problem. * https://www.gob.mx/cms/uploads/attachment/file/444719/ For those Mexicans with access to mortgages, the greater avail- Rezago_habitacional_2016.pdf ability of housing finance significantly improved their quality of life. In some cases, though, not so much. The construction boom driven by private housing developers led to significant urban sprawl in the outskirts of Mexican cities where land prices were low and econo- mies of scale achievable. The reason is relatively simple: it is esti- mated that land in Mexico’s major cities represents between 6.8% and 18.2% of the total housing project cost (SHF and CIDOC, 2015); however, in consolidated areas of the largest cities, land could rep- resent up to 50% of the total housing project cost (SEDATU, SEMAR- NAT and GIZ, 2017). As a result, many housing developments that emerged during this period were so distant from urban amenities and other public services, including transport, water and sanitation, that often fami- 1 CONAVI, SNIIV, 2019.      | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 11 HOUSING DEFICIT: 2016 lies simply abandoned them. In a survey of its borrowers conducted Percentages of the total for each category. by INFONAVIT in 2015, for example, almost half expressed dissat- Urban 38.5% isfaction with the quality of life in the new housing developments2. In transition 18.0% According to a 2017 study by CONEVAL based on the interpre- Rural 43.5% tation of census data, five million houses, or 14% of the housing Affiliated 25.5% stock in Mexico, were unoccupied. More than 90% of these houses Non affiliated 74.5% were located in urban areas. According to INFONAVIT, 80% of its From 0 to 2 MW 41.5% non-performing loans are linked to housing developments in the From 2 to 4 MW 32.8% outskirts of urban areas that are not only overcrowded but also dis- From 4 to 6 MW 13.9% connected from employment opportunities3. From 6 to 8 MW 5.7% The introduction of the city perimeters program (Perímetros From 8 to 10 MW 2.6% de Contención Urbana, PCU) in 2014 —aimed at directing govern- 10 MW or more 3.5% ment-backed mortgages to city centers— was an attempt to ad- Active 92.0% dress this imbalance. But it was only partially effective because Inactive 8.0% SOURCE: SHF BASED ON MCS 2016 INEGI. the value of the subsidies could not compensate for the increase in land values in the urban core. Of the 2.2 million housing units financed by the public housing funds between 2014 and 2017, only ALLOCATION OF PUBLIC FUNDS TO HOUSING BY LOCATION: 2014-2017 8% were located in consolidated areas of the city (PCU1). Only 3% Of 2.2 million new homes constructed: of the housing units intended for low-income residents (vivienda Outside of the urban perimeter zones 21.% económica for those earning below 2.6 UMAs) were located in the Peripheral Zones (PCU 3) 43.6% PCU1 (ONU-Habitat, 2018). In contrast, the homes for those bet- Consolidates Zones (PCU2) 26.9% ter-off (vivienda media and vivienda residencial) were predominant- Urban Core (PCU1) 8% ly located in well-consolidated areas (52%), with only a fraction of SOURCE: UN HABITAT BASED ON CONAVI, 2018. them (12%) built outside the PCU. Faced with a housing supply that was insufficient, unaffordable or unsuited to their needs, city newcomers continued building their own dwellings, without government assistance. They relied on available materials and workers, building their homes incre- mentally, as financial resources became available. According to the National Household Survey, 68% of houses in Mexico are built through incremental self-construction (ENIGH, 2010). As land be- comes scarce and more expensive, mirroring what has happened in the formal sectors, families start settling in distant areas discon- nected from jobs and services, thus exacerbating social and spatial 41% of Mexico’s territory and 31% of its segregation. Those at the bottom of the income distribution are population are exposed to hurricanes, often left with no option but to occupy areas highly vulnerable to storms, floods, earthquakes, and hazards such as floods, tsunamis, and earthquakes. 41% of Mexi- volcanic eruptions. co’s territory and 31% of its population are exposed to hurricanes, storms, floods, earthquakes, and volcanic eruptions4. While those lucky enough to find a relatively safe spot often end up building unsafe homes, due to budget restrictions and lack of access to tech- nical expertise to make their homes resilient. While different sources disagree on the exact size of the housing defi- cit, three things are crystal clear about Mexico’s housing challenges: 2 Survey of Satisfaction of the Accredited, cited in What works and what does not in housing? 3 INFONAVIT, 2014, cited in UN Habitat, 2018. 4 “Strengthening DRM in Mexico” (P146241).      12 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | First, the reduction in the qualitative deficit deserves more at- POOR QUALITY HOUSING tention. According to the SHF, out of a housing deficit of 9.2 million, Inadequate housing is defined by a lack of (1 or more of) these criteria: 19.3% is due to by overcrowding; 78.5.% is due to low quality of ma- -- Access to water. terials; and 2.2% is due to the lack of a toilet inside the house. The -- Access to sanitation. -- Sufficient living area. most important improvements required are related to walls, roofs -- Structural quality. and floors that need to be reinforced or replaced, the lack of bath- -- Durability and location. -- Security of tenure. rooms or sanitary installations, and insufficient living space. Most -- Affordability. of these shortcomings could be addressed through simple, afford- -- Accessibility and cultural adequacy. able and quick fixes, home expansions or self-construction; only a SOURCE: UN-HABITAT, 2018. minority of homes would need to be replaced by new units (3.4%). HOUSING DEFICIT WITH TYPE OF MAIN Thus, the reasonable way to proceed would be to allocate funds to PROBLEM: 2016 help these families improve their existing homes. Between 2008 COMPONENTS 2016 and 2016, however, 48.5% of the number of subsidies and loans and Housing units 32,873,588 89.6% of the resources, were directed to the acquisition of a new With low-quality and 9,224,147 overcrowding problem housing unit. The characteristics of the products and incentives in the market show a concentration in the allocation of resources in % of total housing units 28.1% favor of home purchases: on average, the resources available from % Overcrowding 19.29% subsidies and loans to buy a new home were USD 20,900; and to (1,779,034) build your own home USD 7,000. To improve a housing unit, resi- % Low-quality materials 78.49% dents received average subsidies of only USD 1,270. (7,240,422) % Toilettes 2.22% Second, a focus on the more vulnerable would require new (204,691) geographic priorities and improved operational rules. Even if SOURCE: WORLD BANK ESTIMATES. housing programs have reached and improved the lives of millions, they have also bypassed the most vulnerable segments of the pop- HOUSING DEFICIT BY TYPE OF SOLUTION REQUIRED ulation. In addition to the disadvantageous situation of the indige- 2016 (% of housing units). nous population, households headed by women, the elderly or dis- Only home improvement 60.2% abled are particularly likely to live in inadequate homes (CONEVAL, Self-construction 18.9% 2018). In the meantime, most housing subsidies were provided to Expansion and improvement 9.1% Only expansión 8.5% formal workers affiliated to the social security system. However, in New 3.4% 2016, 68 million Mexicans, (53% of the population) were not affiliat- SOURCE: CONAVI. ed to the social security system, according to CONEVAL; and were therefore ineligible to receive loans from entities such as INFON- AVIT or FOVISSTE. In addition, the share of housing finance target- ed at self-construction —the main way the poor access housing— re- mains very small. In the period 2007-2019, for example, only 14% of CONAVI’s subsidies went towards self-construction (CONAVI, 2019). The Mexican Government could increase its reach to the most vul- nerable by: (i) reviewing the size of subsidies allocated for home improvement; (ii) focusing on geographic areas with higher concen- tration of target groups; (iii) adapting the operational rules to the characteristics of those target groups; and (iv) reviewing its existing tools to incentivize social housing provision in urban centers. Regarding home improvement subsidies: International best practices indicate that lower property values and incomes would justify higher allocations both in absolute and relative terms. This is particularly important because affordability remains a significant challenge in Mexico and would require targeted adjustments of the subsidy amounts available to low-income households. According to CONEVAL, more than half the households in the country cannot buy a home in the formal market because their income is less than | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 13 HOUSING NEEDS OF THE INDIGENOUS the government requirement of 5 times the minimum wage. Ac- POPULATION cording to CONEVAL estimates, most Mexicans earning below the VARIABLE % % INDIGENOUS NATIONAL POPULATION equivalent of five minimum salaries in Mexico —73 million people Housing deficit 44.7 79.1 (57% of the population) would need some sort of public subsidy to Needs a new house 11.6 15.2 acquire a new home in the formal housing market. Families earn- Needs a better house 33.1 63.9 ing below four or five times the average minimum salary in Mexico Low quality materials and lack of 10.2 30.2 space would not be able to buy the most popular home option offered Stoves with Wood or Coal 14.8 58.8* by INFONAVIT in recent years (the Vivienda popular) —priced at With property titles 83.7 76.4 USD 34,000. Nor can they afford home improvement: the average With a woman registered as owner 40.8 30.9 support package for home improvement offered by institutions like Lack of access to basic services 15.5 56.3 CONAVI is USD 1,577) —insufficient considering that the average cost inside the house Water connections inside the house 94.5 87.2* of improving a home in Mexico is USD 2,700. Access to internet 33.0 11.3 Regarding the geographical focus, currently 8 Mexican States Access to any type of phone 84.8 58.2 that concentrate 50% of all the homes in deficit have received *CDI: 2017 SOURCE: EIC 2015 AND MEC FROM MCS-ENIGH 2016 only 26% of all housing actions (subsidies and credits). A stronger emphasis on underserved States would be required both through MAXIMUM AFFORDABILITY housing subsidies and urban infrastructure investments. PER FAMILY INCOME Regarding the operational rules, they would need to be adapted MINIMUM RANGE OF MINIMUM MAX AFFORDABILITY SALARY SALARY IN USD (HOUSE VALUE = 5X to (i) incentivize private sector participation in geographical areas MINIMUM SALARY) 0-1 From 0 to 135,2 USD 8,112.00 and with products that are currently not the most preferred; (ii) en- 1-2 From 135,2 to 270,4 USD 12,168.00 sure the technical standards of the interventions; and (iii) avoid ex- 2-3 From 270,4 to 405,6 USD 20,280.00 cluding large segments of the population that have a high need for 3-4 From 405,6 to 540,8 USD 28,392.00 housing interventions (eg. the non-affiliated, the elderly, monopa- 4-6 From 540,8 to 811,2 USD 40,560.00 rental households, etc). 6-8 From 811,2 to 1.081,6 USD 56,784.00 Regarding the incentives to increase the supply of social housing in central areas: There is ample space to introduce changes that could make the financial numbers work for construction compa- nies and families. A World Bank study conducted in Merida con- cluded that mixed housing projects including affordable housing in the city center are possible if simple changes in zoning and other regulations were considered, such as the increase of permitted den- sities, and the reduction of required parking spaces for housing and commercial buildings. Third, resilience needs to become a central part of the Govern- ment’s housing policies and programs to ensure sustainability, particularly given how exposed Mexico is to different hazards and climate-related events: One out of three Mexicans are vulnerable hazards and climate related events. Worse still, there events are becoming more frequent, damag- ing, and deadly —with the housing sector suffering a significant part of the damages caused by these adverse events. When a catastroph- ic event occurs, the poor tend to be disproportionately affected; and, since they have fewer assets, they are more likely to lose every- thing. The poor’s loss of wealth is two to three times greater than that of the non-poor, which puts their livelihoods and their ability to escape poverty at risk (Hallegatte et al 2017). Since the homes of the poor in Mexican cities tend to be more precarious or located in areas more vulnerable to disasters, poor residents are also more likely to lose their lives in a major disaster. Furthermore, the lack of effective preventive measures, regulatory frameworks, and strong 14 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | THE POTENTIAL TO GENERATE SOCIAL HOUSING IN CENTRAL AREAS MÉRIDA, MÉXICO. BUSINESS AS USUAL SCENARIO 1 SCENARIO 2 (WITHOUT INCENTIVES) (WITH INCENTIVES) (WITH INCENTIVES + FOVIM) Benefits None It allows for the construction of It allows for a higher proportion low-income housing (Vivienda of low-income housing (Vivienda economica) in the city center economica) in the city center It uses an existing fiduciary fund from the municipality Density 77 units per hectare 153 units per hectare 153 units per hectare Housing Units 208 420 420 Total area 17,250 square meters 26,850 square meters 26,475 square meters % of social housing 33% 36% 42% (USD 15.7K–USD 18.3K) Uses Residential Residential and commercial Residential and commercial Parking requirements 1 per housing unit and 1 per 0.5 per housing unit and 0.5 per housing unit and every 30 square meter in com- 1 per every 50 square meter 1 per every 50 square meter mercial units in commercial units in commercial units Parking spaces needed 515 267 267 Leveraged returns 2% annual 18% annual 18% annual Unfeasible Feasible Feasible safety enforcement makes people more vulnerable to catastrophes. North America, Central America and How the Mexican government allocates housing subsidies can thus the Caribbean accounted for 83% of either increase or decrease the risk of the housing stock. total losses and 93% of insured losses It is useful to understand how exposed Mexico is to hazards: Al- caused by catastrophic events most two-thirds of the country’s surface is exposed to significant worldwide in 2017. seismic risk (Programa Nacional de Protección Civil 2014-2018). Ac- cording to CENAPRED, 35.5% of the population is exposed to flood risk. In States like Tabasco (100%), Colima (92.1%) and Campeche CAPACITY FOR RISK MITIGATION IS LIMITED (85.9%) virtually all families are exposed to flood risk. Other States 2017: 370/2,459 municipalities (frequently outdated) like Aguascalientes and Baja California Sur are exposed to different Risk Atlas. events including droughts and desertification. Housing actions need 2017: 329/2,459 have updated construction norms (CENAPRED, 2016). to ensure that dwellings are resilient enough to withstand the force Current definition of the housing qualitative deficit of floods and earthquakes that seem to be increasing in both fre- does not include structural assessments. quency and severity. For example, the earthquakes that took place on 17 and 19 of September 2017 left more than 180,000 housing Between 1999 and 2017 disasters units partially or completely damaged; more than 60,000 of these caused damages and losses to some damaged houses were located in the state of Oaxaca. The creation 1.7 million housing units for of the National Reconstruction Program and the emphasis that the USD 8.7 billion. Government has placed on this process is consistent with the fact that reconstruction costs were estimated to amount to more than USD 2.5 billion, with an estimated USD 520 million needed for the reconstruction of private homes (Senado de la República, 2018). Nonetheless, investing in mitigation —through strengthening and improving existing homes— could pay-off significantly for the Government, while helping the housing authorities achieve their objectives. The reason is simple: mitigation will ensure that di- sasters do not end up wiping out Government efforts to close the | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 15 housing gap: disasters registered between 1999 and 2017 caused an estimated USD 8.7 billion in damages and losses to 1.7 million hous- ing units (CENAPRED, 2019). In Oaxaca alone, it is estimated that small investments in strengthening homes could cut the poten- tial losses by half. HOUSING IS A KEY PRIORITY TO ACHIEVE THE OBJECTIVES OF THE NATIONAL DEVELOPMENT PLAN: The Mexican Government that took office in January 2019-2024: “TO ACHIEVE SUSTAINABLE DEVELOPMENT 2019 redefined the focus of Mexico’s housing policies. AND TO IMPROVE THE LIVING CONDITIONS FOR ALL”. To respond to the challenge of delivering social inclusion, resilience, and affordable housing, the Government’s new priorities are: • A focus on in-situ new construction and home improvement; • A focus on the non-affiliated sector and other segments in re- gions that were previously not reached by the housing programs; • A focus on housing resilience. The objectives of the Government’s “Social Policies” as defined in the National Development Plan 2019-2024 are to achieve sustain- able development and to improve the living conditions for all. In this context, housing is defined as a strategic priority for addressing many other social programs. The Plan defines the Affordable Hous- ing Program, the National Reconstruction Program and the Urban Upgrading Program as the major vehicles through which to reach vulnerable and poor families and those affected by recent disasters, with focus on the territorial, neighborhood and house scales. Affordable Housing Program (Programa de Vivienda Social) The Right to Adequate Housing as defined by the United Nations Committee on Economic, Social and Cultural Rights underpins the guidelines and operating rules of the Government’s Affordable Housing Program. The Government set out the following strategies: −− To facilitate the access of the most vulnerable groups to ade- quate housing, especially those that have traditionally been excluded from the formal housing market; −− To help meet urban an poverty challenges through the inte- gration of housing policy with other federal programs and projects, with a focus on regions with the highest poverty, groups with the highest vulnerabilities and special attention to population exposed to and affected by hazards and climate related events; −− To strengthen self-construction processes (Producción Social de Vivienda Asistida) recognized as the most common prac- tice amongst poor families and communities to incrementally build, improve and expand their dwellings. The National Reconstruction Program The earthquakes of September 2017 and February 2018 caused damage and losses to more than 180,000 private homes. Since taking office in December 2018, the Government prioritized the assistance to people whose houses were partially or completely 16 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 17 OBJECTIVES OF CONAVI’S WORK damaged, particularly in the states of Oaxaca, Chiapas, Puebla and PROGRAM (CONAVI, Programa de Labores, 2019) Morelos. The program also provides for the reconstruction of edu- cation and health infrastructure, as well as any structures related to CONAVI’s 2019 work plan reflects the priorities to pro- mote resilient and adequate housing for the poor and to “cultural heritage” that were damaged. redirect efforts towards home improvement and housing reconstruction: 1. Address the housing deficit in particular as it relates The Urban Upgrading Program to insufficient living space and inadequate housing The Urban Upgrading Program (Programa de Mejoramiento Urbano) amongst low-income families focuses on cities with high concentrations of poverty and housing 2. Provide home improvement assistance to families that cannot afford new housing deficits. The Program seeks to attend neglected regions through a 3. Strengthen the sustainability and resilience of peo- combination of policies for upgrading urban housing and housing ple’s living environments assistance policies. These include cities with a large indigenous 4. Promote coordination and co-responsibility amongst federal and state entities and the private sector to community and cities along the Northern border with high levels address the housing deficit of violence and social marginalization. Though many of these cities 5. Provide reconstruction assistance and relocation to families affected by disasters have significant housing deficits, they have received few housing assistance subsidies in the past. 2019 BUDGET ALLOCATION FOR HOUSING PROGRAMS The budgets allocated to these programs reflect the Government’s AVERAGE objective moving towards more targeted support to the most vul- SUBSIDY PER PROGRAM BENEFICIARY HOUSEHOLDS AMOUNT IN USD HOUSEHOLD IN USD nerable groups and their culturally rooted home building practic- Urban upgrading 20,081 115,9 million 5,764 es. More than half of the 2019 budget was allocated to the National program Reconstruction Program and to addressing the damages caused by National 52,311 267,5 million 5,114 reconstruction landslides and flooding in Nayarit in 2018; 24% of the budget was program allocated to the Urban Upgrading Program and the remaining funds Reconstruction 1,310 5,6 million 4,288 were allocated to the Affordable Housing Program. —Nayarit However, for these programs to effectively target their beneficia- Affordable 34,413 77.6 million 2,255 housing program ries and develop adequate and swift responses to Mexico’s housing TOTAL 108,115 466,6 MILLION 4,316 challenge, three gaps must first be closed: (i) an institutional gap, SOURCE: CONAVI characterized by lack of inter-agency coordination, which hinders effective policy implementation; (ii) a policy gap, epitomized by uncoordinated housing and urban policies, shortcomings in the design of housing subsidies and assistance programs, and lack of support for residential retrofitting; and (iii) a resource gap, refer- ring to the lack of insufficient and outdated information about the housing stock, as well as a lack of standardized instruments for data collection, home improvements, and the provision of flexible and adequate financial services. 18 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | THE GOOD, THE BAD AND THE UGLY OF INCREMENTAL HOUSING Incremental housing is a step-by-step, staged develop- of financing allocated to this building modality remained ment starting from an initial small starter core, and grow- small. CONAVI subsidies allocated to self-construction ing as needs mandate and resources allow. In the early never exceeded 10% of the overall program in any given 1960’s, John Turner first coined the term, when he ob- year (FIVASE, 2012). Starting in 2012, what was originally served the self-made reconstruction of housing after the conceived as a program promoting self-directed home Arequipa earthquake in Peru. He recognized that people production with the family at the center of the construc- produced the most efficient housing solutions when left to tion process increasingly also devolved into the provi- their own devices. Since the 1960s, the process of incre- sion of finished housing units for low-income families. In mental housing has been growing exponentially in cities 2017, 74% of CONAVI subsidies allocated to self-construc- with different cultures and socioeconomic backgrounds. tion were again administered by construction firms that In the 1970s, incremental housing projects —then called delivered finished houses to beneficiaries (SHF, 2017). “sites and services” or “serviced land” projects— were im- plemented by governments who recognized that many While incremental housing and self-construction suit the citizens migrating from rural areas could not access for- needs and financial possibilities of the most vulnerable mal markets of housing. However, these projects came to population, without appropriate technical assistance, it a halt around the 1980s (Turner, 1972). can increase the structural vulnerabilities of their dwell- ings. Most houses in Mexico are built without the exper- The National Housing Law enacted in 2006 for the first tise or technical assistance to ensure the resilience of the time recognized self-construction as a predominant mode housing structure at all stages of the construction. Hous- of housing construction in Mexico, particularly amongst ing units that are incomplete or not well-built are more low-income communities. When the subsidy program ad- susceptible to partial or complete collapse during seis- ministered by CONAVI was launched in 2007, it allowed mic events or other events. For example, in Oaxaca and consumers to select amongst different housing solutions Chiapas, after the M 8.2 earthquake in 2017, 72% of the such as acquisition of new or existing units, improvement housing units that were built with unreinforced masonry or expansion of an existing unit, purchase of a serviced experienced partial or total collapse, a construction type lot, or self- construction rather than only new home pur- that is not accepted by the existing building code in Mex- chase. Although housing finance for both self-construc- ico. In contrast, only 20% of homes made with confined tion and home improvement steadily increased, the share masonry experienced partial or total collapse. THE INCREMENTAL PROCESS OF HOUSING CONSTRUCTION Income Bedrooms Family Change of job 1980 1990 2000 2010 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 19 WHAT ABOUT POORER FAMILIES THAT CANNOT 20 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | AFFORD NEW HOMES AND MORTGAGES? | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 21 HOME IMPROVEMENT INTERVENTIONS CAN BE DESIGNED TO STRENGTHEN STRUCTURES AND IMPROVE QUALITY OF LIFE. 22 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 2. The Opportunities The Mexican Government is already equipped with robust institutions capable of providing financial and technical as- sistance to low-income families to make their homes more resilient. In fact, its years of experience expanding access to formal and affordable housing at a large scale makes Mexico an excellent model for the rest of the region. There is a wealth of lessons and best practices to build upon. But before the Government can proceed to adjust housing policy to serve the country’s neediest population in the most effective and evidence-based way, some critical gaps must be filled in institutions, policies and resources. GAP IDENTIFICATION 1. Institutional gap Mexico boasts several public and private organizations already involved in the design, financing and development of social housing. In the last 45 years, more than 120 entities have had a stake in Mexico’s housing policy. The federal gov- ernment is responsible for formulating and conducting the national policy of human settlements, territorial planning and urban development5 in coordination with federal entities and municipalities, with the participation of the social and private sectors. States and municipalities must apply and adjust their planning processes to the national strategy and administer territorial planning and urban development programs, as well as monitor and evaluate their compliance with the participation of dependencies of the Federal Public Administration and society, and establish comprehensive risk management strategies, promoting and executing actions to prevent and mitigate the risk of human settlements and increase their resilience to natural and anthropogenic events. To achieve its objectives, the new administration must leverage this experience and expertise. Mexico has been successful at producing standardized, finished housing units at a grand scale. While there is also plenty of experience in the country on home improvement interventions at a smaller scale, these entities are less familiar tailoring their services to individual home- owners, along with the right kind of technical assistance. To close this institutional gap, we recommend that the Government: • Strengthen coordination between institutions. A wide range of entities are involved in the design and implemen- tation of housing assistance programs in Mexico. But there is too much duplication of efforts and too many bottle- necks. There is an opportunity to revise the responsibilities of these entities to exploit synergies in the selection of target populations and in the effective design of home improvement and retrofit interventions, subsidies and credit schemes. While National Housing Organizations (ONAVIS) such as INFONAVIT, FONHAPO, FOVISSTE and CONAVI, which operate throughout the country are the main source of housing subsidies, assistance programs can be diver- sified and expanded not only through increases in the provision of subsidies at the federal level, but also through a greater participation of the State Housing Organizations (OREVIS), the financial and microfinance sector and the resources of the population itself 6. • The new National Housing Program currently being prepared by SEDATU could play a key role in addressing these issues. Effective housing policies that prioritize social inclusion and housing resilience build on adequate and updated territorial risk and planning information. To date, housing policy instruments include construction norms, but maps documenting hazards and risks are either incomplete, inexistent or not enforced in most Mexican states and municipalities. CENAPRED must work with State and local governments to update state and municipal risk maps and incorporate them into the National Risk Atlas. The use of this information for investment planning purposes could ensure that the government budget is used to decrease risk and not increase it. 5 According to the General Law of Human Settlements, Territorial Planning and Urban Development of 2016.      6 For more information on the actors of the housing sector in Mexico, see Annex 2.      | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 23 THE INSTITUTIONAL SETTING FOR THE HOUSING SECTOR IN MEXICO 1960 1970 1980 1990 2000 2010 2020 1963 1976 1982 1992 2001 2006 2019 Fondo de Secretaría de Secretaría de Secretaría de Comisión Nacional Comisión Nacional Secretaría Operación y Asentamientos Desarrollo Urbano Desarrollo Social de Fomento de Vivienda Desarrollo Financiamiento Humanos y Obra y Ecología (SEDESOL) a la Vivienda (CONAVI) Agrario, Territorial Bancario a la Pública (SAHOP) (SEDUE) (CONAFOVI) y Urbano Vivienda (FOVI) (SEDATU) NATIONAL HOUSING PROGRAM 1979 1980 1990 2000 2010 2019 Programa Nacional Programa Nacional Programa Nacional Programa Sectorial Programa Nacional *Programa Nacional de Vivienda de Desarrollo Urbano de Vivienda 1990- de Vivienda 2001- de Vivienda 2014- de Vivienda 2020- y Vivienda 1984- 1994 and Programa 2006 and Programa 2018 2024 1988 Sectorial de Vivienda Nacional de Vivienda 1995-2000 2007-2012 THE ROLE OF THE STATE IN THE HOUSING SECTOR 1972-1992 1985-? 1993-2000 2006-NOW 2001-2012 2013-2018 2019 2018-2019 Market Maker: Social housing INFONAVIT’s Promotor of A force to Social Changer. Elimination of The Citizen at INFONAVIT provision by social mortgage. better located reduce the INFONAVIT public trusts the center of all starts both FONHAPO. INFONAVIT housing. housing deficit. as a financial FONHAPO. housing policies building and BANOBRAS in abandons the INFONAVIT as a institution with a and programs. financing charge of the construction banker. social purpose. housing. construction. business. 24 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | THE POLICY INSTRUMENTS FOR THE HOUSING SECTOR IN MEXICO 1981-2006 2001-2006 2006 2008 2013 2016 2018-NOW Implementation Implementation SEDESOL’s Program for the Creation of SEDATU merge Program of FONAHPO’s of the National subsidies started Support and CONAVI’s the POTER and for Shared credit program. Housing Program to be used for Regularization RENARET. S237 and creates Responsibility (PNV). new housing. of Informal the “Program for continues. 1989-1994 Settlements 2014 Risk Prevention”. Home 2001 2007-2008 (PASPRAH). Publication of the 2018-NOW Improvement INFONAVIT’s CONAVI starts National Urban 2016-2018 Implementation Actions by Support Program 4 programs 2009 Development Implementation of the Program the Solidarity that introduces (DUIs, DUIs, FOVISSSTE Program 2014- of the Program for the Program from mortgage building code, starts issuing 2018 that creates for Access to develution of SEDEU-SEDESOL. guarantees to NAMAS, NAMAS, certificates that the “urban Financing for resources from loans generated sustainability). can be traded perimeters”. Housing. INFONAVIT’s 1992 by commercial in the Stock Savings Fund Creation of the banks. 2007 Exchange. 2014 2016 72-92 to retirees Special Program Creation of Implementation from the IMSS. 2002 2009 SEDATU for the Promotion INFONAVIT’s implements the of the Program and Deregulation Opportunities Green Mortgage Cretion of the for Housing 2018-NOW Program for the of the Housing Program from Program, Vida Program EcoCasa Resettlement Support. Program for Sector (PEFDV). SEDESOL, Integral and from the SHF. of People living Indigenous focused on home Qualitative in hazard areas 2016 Infrastructure 1992 improvement Evaluation of 2009-2010 (REPZOR). CONAVI creates that provides 100 Cities and acquisition Housing (ECUVE). SHF expands the National housing to Program from for vulnerable its scope and Registry for indigenous SEDESOL that groups. 2007-2015 includes home 2014-2018 Territorial households. included housing Implementation improvement and Creation of the Reserves subsidies. 2003 of the Federal self-construction (RENARET). 2019 Program for Implementation Subsidy and with technical a Sustainable MEJORAVIT 1995 of INFONAVIT’s Financing Schem assistance in its Housing 2017 continues. Total Solution credit program for housing. portfolio. Development. Creation of Program from for low-income the Programs 2019 FOVISSTE, housing. 2007-2012 2009-NOW for Credit and ARRENDAVIT 2014 created after the Reengineering of SHF responds to Annual Financing ends. 2003 Beginning of the 1994 economic the Programs for mortgage crisis, approved by Program for the crisis. Implementation Rural Housing and buys assets from the Executive 2019 Consolutation of Program Savings, Subsidy SOFOLES and Commission from Housing of Urban Land 1995-2000 “Savings, Subsidy and Credit for SOFOMES. FOVISSSTE, the Insurance Reserves by Implementation and Loan” for the Incremental Total Solution Program by SHF SEDATU. of the National Incremental Housing —Tu 2012-2017 Program for to continues. Housing Program Housing (VIVAH). Casa. MEJORAVIT from expedite the 2014 payment of debts (PNV). INFONAVIT starts 2019 2003 2008 operating. The Opportunities and transfer of properties and Green Mortgage 2000 Military Housing INFONAVIT- Program is the Program Program by SHF Fund (FOVIMI) FOVISSSTE start 2013 renamed FONHAPO stops for Housing continues. recognized by giving loans to and becomes giving loans. Creation of the Reconstruction ISSFAM. married couples. PROSPERA. Program for to respond 2019 It includes 2000 Decent Housing. to the 2017 Start of the 2003 2008-2012 incentives and Start of the Piso housing financing earthquakes. Program for the Firme Program Creation of Implementation 2013 Regularization support for from SEDESOL. Habitat —SEDATU of the Program Creation of the 2016-2018 of Informal families living in Program. for Federal Program for Settlements poverty. SEDATU improves 2001 Subsidies by Rural Housing (PRAH) by INSUS. 2005 CONAVI. the Program for Housing Sectoral and Program for 2015 Risk Prevention. Program 2001- FOVISSSTE starts the Promotion 2019 2008-2012 Pilot Program 2006 CONAFOVI- cofinancing of Rural RENARET ARRENDAVIT CONAVI. operations with Implementation of Consolidation. continues from INFONAVIT banks. the Program for through the to recover the 2001 the Modernization housing units Program for 2006 of Property Social Housing Implementation that were Implementation Registry and its occupied before 2019. of the Tu Casa linkage with the Program from of the Rural 2015. Housing Program ­ SHF. Cadaster — CONAFOVI. —FONHAPO. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 25 • Build the capacity of state entities at the federal, state and municipal level in the housing and planning sector. There is an opportunity to strengthen coordination between the federal and municipal levels so that municipalities have the capacity to implement housing and planning programs as effectively as possible. This will also empower municipalities in several ways: to make more efficient use of intraurban land; promote land value capture and initiate new housing developments with private sector partners; promote normative changes as required; increase their tax base; and promote more sustainable urban development patterns. • Strengthen the ecosystem for providing specialized technical assistance and tailored housing assistance. One place to start is by improving the mechanisms by which public and private organizations can coordinate their efforts to provide specialized technical assistance to guide and accompany home improvement and self-construction ef- forts. OEOs (Works Executing Entity) and AOs (Operation Administrators) could be used more frequently as certified intermediaries to train and develop intermediary advisors, institutionalize and professionalize technical assistance schemes through home improvement and retrofit guidelines and standardized plans, while developing a long-term presence in the community that will help generate confidence among homeowners. 2. Policy gap The Government is now formulating the details of new housing programs that will guide its efforts in the coming years. Here are some recommendations for measures that we believe will help realize the new administration’s vision: • Improve the use of granular data of the housing sector. Sound diagnostics must be institutionalized as the basis for designing and implementing all aspects of the housing programs, including: −− For the Affordable Housing Program, it would be critical to identify the characteristics of the housing stock to inform decisions about the allocation of resources through different instruments and target audiences, with a focus on the most vulnerable. A sound diagnostic could also help the country move from discussing the size of the housing gap (while some estimate it to be 9 million units, other equally credible sources estimate it to be 14 million) to evaluating different options to close it. −− For the National Reconstruction Program, a focus not only on damaged properties but also undamaged houses that present building deficiencies —with the objective in mind “to build better before,” would help significantly reduce the future estimated losses the country faces in case of disasters. −− For the Urban Upgrading Program, more granular data could be of great benefit achieve the Government’s ob- jective of helping the poorest and most vulnerable population who are not eligible for support under the other programs and who need not only access to better homes but also integration to society through the access to basic services, jobs, education and health services and other amenities. • Increase housing resilience by integrating structural strengthening as part of home improvement investments. The Government, for example, could use the Affordable Housing Program to pair structural retrofits with home improvement measures as a critical step to achieving a resilient housing assistance framework in disaster-prone ar- eas. To effectively increase housing resilience, the Government needs to complement home improvement measures with targeted actions to reduce housing vulnerability. This includes institutionalizing vulnerability assessments and identifying a set of retrofitting standard options that could complement existing home improvement programs (see Annex 1), as well as streamlining administrative procedures to obtain building permits for retrofit interventions and/ or to regularize properties. • Support self-construction. It is well-known that, most Mexican families opt for building their homes themselves, without technical assistance, rather than buying a new or existing unit. These self-built houses make up a significant share of the qualitative deficit of Mexico’s housing stock. The Affordable Housing Program could benefit families by providing housing assistance and subsidies specifically designed to support home improvement and do-it-yourself construction. 26 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | FOR EVERY DOLLAR INVESTED ON INFRASTRUCTURE STRENGTHENING BEFORE, IT’S BEEN PROVEN THAT AT LEAST 4 DOLLARS ARE SAVED IN RECONSTRUCTION EFFORTS LATER. • Prioritize geographical areas with the largest housing deficits. Today, the Mexican states with the highest hous- ing deficit are Chiapas, Oaxaca, Guerrero, Veracruz and Campeche (CONAVI 2015). Over the last 12 years, however, these same states did not receive a comparable share of overall housing finance. In fact, states with low housing deficits received higher amounts of housing finance (e.g. Nuevo León). A larger budget allocation to states with higher housing gaps could also help develop the capacity of the providers to operate in the targeted market segments. However, it is advisable to use a gradual and programmatic approach that allows these changes to be properly planned, so that they do not suspend or stop serving needy populations regardless of the size of the deficit in the state in which they live. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 27 Satellite (Digital Globe - Worldview 3) 3-band True Color, Spatial Resolution: 50 cm 28 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | Drone (Ebee —3D S.O.D.A) 3-Band, True Color, Spatial Resolution: 3.65 cm | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 29 • Strategically position subsidies and incentives to promote energy efficiency and renewable energy. Between 2016-2018, CONAVI administered a program to subsidize investments that increase energy efficiency or renewable energy as components of the Affordable Housing Program. Based on this experience and on comprehensive tech- nical and economic analysis of the sector, the Government could design new instruments that can be scaled up to contribute to more efficient, affordable and sustainable energy consumption patterns of target households. The experience of other entities like INFONAVIT and SHF on this topic could inform CONAVI’s effort. • Coordinate urban and housing policies with urban upgrading and other programs. There are synergies in infor- mation collection, communication, civil society participation, etc. that can be leveraged to enhance the benefits of resilience, home improvement and other programs to improve urban housing, such as upgrading neighborhoods, land titling, resettlement, and infrastructure for basic services. Coordination in the planning and implementation stages of such activities is key to achieve better results and avoid overlaps and inefficiencies. For example, an effective coordination of the Affordable Housing and Urban Improvement Programs can enhance the goals of both programs. 3. Resource gap When designing a housing program to meet the needs of the poorest segments of the population, the first requirement is a comprehensive diagnosis of the state of the housing sector, which involves: (i) a consistent methodology to measure the housing deficit; (ii) a transparent and accessible tracking and monitoring system of all housing interventions; and (iii) an updated assessment of Mexico’s housing stock that includes a housing vulnerability assessment. To ensure the needs of the poorest are being targeted, the following activities should be performed: • Review Mexico’s operational definitions for the housing sector. This includes revising the official methodology, definition and measurement of the housing deficit, while updating and adding to databases such as the Single Hous- ing Registry (RUV). • Develop a robust and simple methodology to characterize and assess the vulnerability of the housing stock, based on a consensus among national experts. Rapid and effective housing assessments consist of three steps: −− Identify houses where risk cannot be reduced, thus ruling out retrofitting. The minority of homes located in ar- eas where risks cannot be mitigated, including those in flood plains or landslide prone slopes, must be declared uninhabitable and the residents resettled. −− Identify houses with structural risks that can be repaired. Housing engineers can visually inspect rows of substan- dard housing units and rapidly identify the common structural inadequacies. −− Identify —with geospatial tools— patterns of structural vulnerability that are common and can be addressed with standardized and scalable retrofitting interventions designed by engineers. The large majority of housing units in Mexico have been built with safe materials and common construction methods. Addressing any remaining struc- tural vulnerabilities/deficits can be done at relatively low-cost, transforming a deadly structure into one strong enough to resist future stress. Both the Urban Improvement Program and the Reconstruction Program can bene- fit from a geospatial information system that informs decisions about home and urban improvements, as well as about reconstruction, resettlement and formalization. • Design technically and economically viable home improvement instruments. Bringing existing houses up to a decent standard of safety is not only a more efficient way to deploy limited government funds but an effective strate- gy for saving lives. The Government program can simplify and improve the process for designing and implementing home improvement subsidies by: −− Developing standardized home improvement options, where possible. −− Incentivizing the development of private sector solutions. −− Leveraging private resources, such as household investments and home improvement microfinance markets. 30 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | • Promote financial inclusion by designing financial instruments for AREA OF INTEREST savings, credit and insurance that suit the needs of Mexico’s differ- ent population segments. Access to quality financial services, such savings accounts, credit schemes and home insurance, can soften the VERACRUZ impact of catastrophic events —even prevent people from falling into poverty or becoming poorer. In Mexico, according to the Global Findex (2018), among the poorest 40%, only 5.2% borrowed from a financial in- Area of interest stitution or used a credit card in 2017; more than 75% of them reported that they couldn’t come up with funds to deal with an unforeseen emer- CHIAPAS gency7. Mexico’s housing programs should consider the use of multiple OAXACA financial instruments to provide flexibility, such as: long-term low-in- terest-rate loans, municipal or national grants, compensation through Juchitan property taxes, more favorable financing terms for safer buildings, some pass-through costs, and reduced property insurance. Salina Cruz Pacific Ocean METHODOLOGY DEVELOPMENT AND CASE STUDY Mexico has started to develop a policy framework that aims at redirecting its housing policies to include the poor and most vulnerable. An important, cost-effective element of this change consists in rebalancing its programs not only to produce new housing but also to increase the resilience of existing housing. Therein lies another challenge: To improve existing housing belonging to the poor will require Government to operate in areas that are not only more exposed to different hazards but also for which data is scarce, outdated or non-existent. To explore ways of adding value to the authorities’ efforts, CONAVI and the Bank Team agreed on an area of interest and applied a methodology developed by the Global Program for Resilient Housing to capture detailed information about the housing stock and to identify opportunities for strengthening existing structures and other improvements. This methodology was already successfully applied in a range of other countries, including: −− In Colombia, assistance is being provided to the government for the design and implementation of its recently launched housing improvement subsidy “Casa Digna, Vida Digna” (Dignified House, Dignified Life). −− In Guatemala, buildings at risk of collapse in case of an earthquake were mapped and identified. −− In Indonesia, the government is working on including resilience as a central element of its housing improvement program, one of the largest in the world. −− In Peru, technical assistance was provided for the realization of automated real estate valuations, the improvement of local tax collection and the identification of structural vulnerabilities. −− In St. Lucia, roof damage caused to homes by Category 5 hurricanes was assessed. −− In Sint Maarten, the government is working on estimating the costs of scaling up its resilient housing program. The selected area was Oaxaca’s Istmo de Tehuantepec Region because of its long history of seismic activity. In 2017, an 8.2 M earthquake struck southern Mexico. Damages were concentrated in the states of Chiapas and Oaxaca, with the epicenter in Pijijapan, Chiapas. Only twelve days later, another M 7.0 earthquake struck Puebla and Morelos, shaking the center of the country. A total of 5,735 seismic events were recorded in Mexico for the month of September, damaging a total of 186,526 houses. The hardest hit state was Oaxaca —with damage to 80,310 private houses8. Then came its neigh- bors, Chiapas with 46,546 damaged houses and Puebla with 28,343. In the aftermath of those earthquakes, structural engineers sent to the field found that most structures built after 1985 performed well. Why? After the massive M 8.0 earthquake in Mexico City in 1985, authorities began to pay more attention to structural vulnerabilities and pushed for adopting advanced construction techniques that would keep buildings from crumbling during earthquakes. In the state of Oaxaca, for example, and more specifically in the Istmo de Tehuantepec Region, the housing stock consists of predominantly one —and two— story masonry buildings that fare relatively well 7 Defined as 1/20 of the GNI per capita in local currency.      8 Includes damages of the earthquake in February 2018.      | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 31 EARTHQUAKE HAZARD FLOOD HAZARD Seismic Hazard 86 87-121 122-155 Flood 156-225 Susceptibility Urban land High Soil Acceletarion Very High (in Gals) Urban land FLOOD VULNERABILITY FLOOD VULNERABILITY Flood Susceptibility Flood Susceptibility High High Very High Very High Salina Cruz Salina Cruz Rooftops Rooftops 1 story No data 2 stories Mixed 3 stories Non residential 4+ stories Residential during earthquakes —if they are well-built and located on safe terrain. Such structures are therefore good candidates for low-cost and life-saving retrofitting, along with investments that improve the quality of life of their inhabitants. The team’s first conclusion was that for SEDATU and CONAVI to succeed with the type of housing policies they are pro- moting, they needed data on the housing stock at the right scale. Their policies and goals of serving the poorest segments of the population are aimed at families who live in homes for which good data is not usually available. The recommended approach for better identifying these housing structures, therefore, must begin from high above and then zoom down to a more detailed picture of the housing and its terrain. Hazard maps are useful tools for the identification of flood, seismic, tsunami, and landslide events across a large area. Leveraging satellite imagery, topographic information and past events, they reveal where the hazard exposure is highest. By overlaying these hazard layers with population information, the Government can get a general idea of the intensity and likelihood of hazards affecting cities. For example, it is possible to compare the distribution of the urban population in Istmo de Tehuantepec Region with the intensity of earthquakes and flooding hazards. But hazard maps alone are not enough to help the Government prioritize the allocation of home improvement sub- sidies that will increase housing resilience. More granular data is needed. And we can now get that, more quickly and affordably. Using the advanced technology of drones, 360-degree street cameras, and machine learning algorithms you can now screen houses —remotely— for vulnerabilities generated due to the use of substandard materials or inadequate construction techniques. SEDATU and CONAVI need tools that will identify housing vulnerabilities at every level —national, neighborhood and individual housing unit. During this ASA, the World Bank Team gathered data in the Istmo de Tehuantepec Region and applied a 3-step method- ology to inform programs aimed at improving living conditions while increasing housing resilience. 32 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | ADVANTAGES USING DRONES IN NEIGHBORHOOD LEVEL ANALYSIS Irregular layout Absence of sidewalk Paved roads Existence of water tanks Low-density housing Unpaved roads 1. A geospatial housing needs assessment tool that incorporates disaster risk information to better prioritize and target subsidies; 2. A multi-stage and flexible subsidy design that is conducive to incremental and resilient construction and enables families access multiple subsidies over time. These subsidies need to include a condition and/or incentive to com- plete the necessary structural reinforcements before any other improvement and a strong linkage with microloans to allow households to pool resources to combine home improvements with structural reinforcements in a single intervention. 3. Specialized technical assistance capable of delivering standardized and affordable technical solutions to lower the cost of engaging with citizens, while encouraging private sector investments in construction, credit and insurance markets, including the identification of standardized solutions, tools for calculating works’ budgets and identification of reliable suppliers. A geospatial housing needs assessment tool CONAVI and SEDATU need to adopt a methodology that combines well-designed policy schemes with disruptive technol- ogy for targeting and prioritizing areas that should receive home improvement subsidies and/or neighborhood upgrad- ing. Operating at three levels, the approach will seek to save lives, protect assets and shield economies. 1. National level: The Government of Mexico can use this methodology to identify municipalities with a high concen- tration of low-income urban households that could greatly benefit from risk reduction investments. Using national scale geospatial hazard layers, combined with population density and socio-economic information would make it easier to | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 33 AREAS OF CHANGE/TRANSITION prioritize municipalities with a high percentage of their population living in Salina Cruz hazard prone land. For example, if the urban center of a municipality exists mainly in a flood plain or has a long history of seismic activity, that munic- ipality should be a candidate for substantial funding. In Oaxaca’s Istmo de Tehuantepec Region, the team began by compiling an inventory of national level spatial data layers: landcover, census data, hazard risk (generated by Mexico’s Comprehensive Risk Management Strategy (EGIR) in 2014), and infrastructure locations. With a complete package of layers, data dictionar- ies and metadata in hand, it is possible to determine the scale and distri- bution of the multiple risks a region faces. While this database provides a strong baseline for assessing housing needs at the regional level, it lacks the detailed spatial resolution to inform decision makers about which homes Buildings under should receive a subsidy first, or which are good candidates for a specific construction 0-1 intervention. 2-10 11-20 21-57 2. Neighborhood level: Geospatial data from drones and satellite can be analyzed to ensure interventions are reaching clusters of vulnerable homes. For this level of analysis, drones or high-resolution satellite imagery is need- ed. Satellite imagery allows for the detection of construction in risky land, by revealing building location (count and perimeter of rooftop), roads (length and width) and topography (slope of the terrain). Overlaying this informa- tion with hazard layers reveals neighborhoods that are prone to flooding (lo- cated in low lying land near rivers), or areas that are too steep for building. In addition, information about the neighborhood’s general connectivity and building footprint density can be calculated. However, for more in-depth analysis, such as the true size of the homes (calculations including height) or the quality of public infrastructure —drone images are required. Drone Buildings imagery could provide policymakers with the type of detailed information abandoned 0-1 they need to improve the effectiveness of their policies and programs. 2-10 11-20 Such data can also reveal neighborhood growth dynamics by showing areas 21-57 with high concentration of homes that are under construction or abandoned. 3. Housing unit level. Leveraging drone and street view imagery provides information about not only the location, dimension, quality, and material of each building but also information on the front of the building (use, wall material, features). All the attributes that are extracted when machine learning is applied to street view and drone images can be presented in a Housing Passport. With the rich household level database derived from street view and drone imagery, home improvement programs can better target certain types of at-risk homes. For example, if CONAVI were interested in allocating housing subsidies, such imagery would help them quickly rule out homes built along earthquake faults or on the edge of cliffs as too risky; but not homes in areas with medium or high seismic risk that have been built with good quality materials and reason- able construction techniques. In Salina Cruz, 2% (2,211) of the buildings (98% residential) were found to have been built on ground with a slope greater than 15%. With this data, homes that have been constructed on dangerous slopes on the hillsides near the city center can be located and quantified. The structural integrity of such homes should be a priority in any retrofitting program. Homes constructed on solid ground but adjacent to a road that frequently floods due to poor drainage should be prioritized for mitigation investments. It is in these areas that public investments in home improve- ment subsidies can be combined with disaster risk management best practices to deliver the maximum impact. In addition to improving disaster risk manage practices, a detailed database of the housing stock will help Mexico prioritize scarce engineering personnel by optimizing routes and making the best use of their time by sending them to assess those homes that need more detailed assessments with data and images about the structures in hand. For exam- ple, in Salina Cruz, when the database is filtered to show buildings that are potentially soft story (at least two stories high and with large first floor openings), out of an estimate of the 15,261 structures that were captured on street view, 904 of 34 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | ROOF QUALITY ASSESMENT IDENTIFICATION OF SUITABLE LAND Roof quality Abandoned Damaged Fair Slope of terrain Good 0.00°-5.00° Poor 5.01°-15.00° Under 15.01°-25.00° construction 25.01°-30.00° IDENTIFICATION OF POTENTIAL SOFT-STORIES THE POSITIVE IMPACT OF RETROFITTING With retrofitting (confined masonry) Without retrofitting USD MILLION 1,000 800 600 400 Soft-story ground floor 200 Possibly 1 story 1+ story 0 5 10 20 50 70 100 200 500 700 900 1000 PERIOD OF RETURN LESSONS FROM MEXICO’S HOME IMPROVEMENT PROGRAMS Mexico’s most popular home improvement programs fo- often times it resulted in further overcrowding, as entire cused on targeted interventions to reinforce floors, walls families occupied the newly won space. However, this and ceilings, install bathrooms or add additional living type of intervention is still very much needed; its opera- space. One of the most successful home improvement tional delivery model should be improved. programs was Piso Firme, administered by SEDESOL to re- place mud floors. Between 2006 and 2013, SEDESOL was FONHAPO’s experience with Cuarto Adicional also, has il- able to improve 3.1 million homes in this way, thus reduc- lustrated that the verification of the progress, completion, ing Mexico’s qualitative deficit significantly. Although this maintenance and use of the investments is critical to the intervention was low-cost and simple (it could be com- program’s success. To that end, FONHAPO, SEDESOL and pleted by the homeowners), it had a huge impact in terms CONAVI developed strong monitoring and information sys- of improving health outcomes in children and adults. tems for registering and tracking home improvement mea- sures, which, in turn, facilitated the verification of each In contrast, FONHAPO’s Cuarto Adicional, which sought to intervention. The Housing Registry (RUV) administered by address problems of overcrowding by adding additional CONAVI tracks progress of construction alongside other living space was not as successful. Although the program relevant data on the national housing supply, location and reached scale and upgraded almost half a million homes in housing quality and has become an important source of 5 years, independent evaluations show that in most cases, information to improve decision making of both public and the additional room was not always used as intended; private actors involved in housing provision. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 35 BUILDING PASSPORT LOCATION ATTRIBUTES ROOF MATERIAL Concrete-plastered ROOF CONDITION Good AREA 650.11 m2 HEIGHT 11.20 m VOLUME 1,601 m3 TERRAIN SLOPE 2,5° USE Commercial WINDOWS 5 DOORS 2 GARAGE 1 WALL MATERIAL Concrete WALL CONDITION Good SEGURITY No COMPLETE Yes RISK VALUE ANALYSIS them had more than 1 garage and were at least two stories high. This could indicate the presence of soft-story. Such a map would guide engineers to visit these homes first, or, depending on resources, require a field visit only if the ‘potential’ soft story building is requesting a home improvement subsidy. Upon arrival, engineers would have pictures and basic statistics about the home and its neighborhood. Such a system would link all information to a precise location and allow for the tracking of all beneficiaries through a spatial referenced household database. Authorities could later return to the homes to ensure after the works have been completed and ensure that technical standards have been met. A multi-stage and flexible subsidy design The main federal entities to finance home improvement in Mexico over the past decades have achieved a good balance in terms of coverage. Mexico’s Social Development Ministry (SEDESOL) and the Affordable Housing Fund (FONHAPO) had a clear mandate to reach the poorest segments and those that are excluded from the formal housing sector9. The home improvement subsidies and credit from INFONAVIT and CONAVI targeted low-income but affiliated workers with access to mortgage loans. Most programs were focused on improving finishes, fixing roofs and adding habitable space. However, in terms of the quality of the interventions and their resilience, the results of these programs have been mixed. To calibrate the costs and potential savings of interventions aimed at reinforcing the housing stock against different catastrophic events, it is necessary to conduct an assessment of potential losses. For this task, the geographical analysis of hazard and exposure is useful but limited since it only presents information on exposed assets but not about the range of possible losses or the probability of their occurrence. For the Istmo de Tehuantepec, catastrophe models were devel- oped to better understand the probability of observing large losses. This analysis consisted of performing three major processes: a. Threat modeling carried out through simulation of individual scenarios, assigning each one a frequency of occur- rence; b. Vulnerability modeling in terms of average percentage of damage depending on the intensity of the threat and; c. Risk estimation in terms of expected losses through the construction of a probabilistic model. To calculate the average and extreme losses, a simulation process of each of the scenarios considered was carried out to probabilistically integrate a curve with the results obtained using the frequency of occurrence in each scenario as weight factors. Through this analysis it was found that seismic risk was the most relevant in the Istmo de Tehuantepec Region. 9 SEDESOL’s housing-related programs were transferred to SEDATU in 2013 and FONHAPO’s programs will be implemented by CONAVI from 2020.      36 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | Different simulations were performed with the main construction variables (number of stories, wall materials, type of roofs, etc.), resulting in no major changes in the composition of the portfolio versus total risk contribution, which suggests that construction technique instead of building materials presents the greatest contribution to vulnerability. One of the most important findings was the behavior of unconfined masonry: while these structures represented 18% of probable losses, with an earthquake, they increased by almost 3 times to 56% of the total portfolio risk. To verify this finding, a simulation was performed assuming that there was a massive retrofit for 18% of homes in the study area, which revealed that estimated losses would decrease by half, saving almost 1,400 million pesos —only due to the conversion of unconfined masonry homes to confined masonry ones. Thus, the value proposition of home improvement programs is two-fold: for housing authorities, they deliver better quality of life for the most vulnerable segments of the population; for economic and finance authorities, they can reduce significantly the contingent liability that is concentrated in the housing sector in the event of a disaster. Specialized technical assistance While incremental housing and self-directed construction respond to the needs and financial means of the most vul- nerable population, unless accompanied by technical assistance to ensure the resilience of the intervention, they could lead to an allocation of housing subsidies that increases the risk of the housing stock instead of decreasing it. Unfinished housing units collapse during seismic events or other disasters. This highlights the need to integrate structural retrofit in standard home improvement processes that focus on improving the family’s living conditions. Tailored home improvement measures and structural modifications need to be accompanied by specialized technical assistance, while putting families at the center of the decision-making and execution processes. CONAVI was particularly successful in developing the program Esta es tu casa combining specialized technical assistance with self-directed con- struction processes. The strength of this approach was that it placed the family at the center of the process as each home improvement measure was designed and implemented by the beneficiary. This was in stark contrast to the traditional housing assistance models in Mexico where executing entities and private developers would deliver the finished housing unit to the beneficiary. CONAVI has collaborated with a range of entities, including non-profit organizations, microfinance institutions and social cooperatives, to generate a niche market for the provision of technical assistance to beneficiary families. These organiza- tions needed to be accredited to provide these services. The program administered by CONAVI was very flexible at the ini- tial stages and a substantial dynamism and diversity developed in the sector in terms of who was able to provide technical assistance and under what terms, including non-profit organizations dedicated to the provision of technical assistance, and cooperatives and microfinance institutions that hired architects and engineers to provide technical assistance in-house. In order to design tailored technical assistance schemes, it is necessary to know the different retrofit needs of the hous- ing stock. In the Istmo de Tehuantepec, engineers visited the study area to evaluate in-situ the conditions of the housing stock and assess the feasibility of structural strengthening. In two weeks, a team of three engineers performed both rapid assessments in more than 100 homes and in-depth assessments of some 40 homes. Many of the homes visited had been affected by the September 2017 seismic events, and therefore provided valuable insights into local housing vulnerabilities. Following the field visits the team developed an inventory of the housing stock that included the following informa- tion: (i) location criteria; (ii) construction characteristics; (iii) estimated needs for reconstruction and retrofit. The predominant housing typology in the region is one and two stories masonry construction. This masonry is mostly composed of confined, or partially confined, solid concrete block walls. Confined hollow concrete block, and unrein- forced fired clay brick walls are also present. Other housing typologies exist as well, in much lower numbers, such as adobe and other earth construction like bahareque, and temporary shelters built out of lightweight improvised materi- als. The analysis identified the adobe and lightweight material houses as the ones that suffered a higher level of damage in the earthquake. However, several confined masonry homes presented structural and seismic deficiencies, such as: −− Incomplete confinement connections, mostly missing vertical reinforcement of window and door openings. −− Poorly framed timber structures in roofs, with loose heavy clay tiles. −− Construction quality issues, such as concrete cover of reinforcement in slabs, promoting rebar corrosion, and low masonry unit strength. −− Configuration issues. As open fronts, lack of cross walls in interior spaces, missing door and window concrete lintels, or sills. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 37 A full report of these vulnerabilities and solutions was also completed. The team used the menu of retrofitting solutions provided by the engineers and match them with the household level typology derived from the drones and street view images. For example, all homes built with concrete blocks and a maximum of two stories that present building defects can be considered as candidates for reinforcement consisting of simple measures, such as the installation of missing confinements in the wall openings, including windowsills, jambs and lintels, new walls and fillings, strengthening of selected walls with plaster, and installation of missing confinement columns. The initial estimate of the costs of a typical modification of an undamaged masonry house is USD 30 per m2, a price tag significantly lower than the reconstruction one (USD 300 per m2). This number, when combined with the rooftop footprint derived from the drone, provides the government with an essential ingredient in the design of a housing resilients program. AVERAGE COSTS FOR RETROFIT* BY STRUCTURAL SEISMIC DEFICIENCY TYPICAL CONFINED MASONRY HOUSE EXAMPLE 1 TYPE OF INTERVENTION AVERAGE COST AVERAGE FOR 70 M2 COST USD/M2 New house built with subsidy presenting deficient connections HOUSE USD between the roof and walls and poor construction detailing of reinforced concrete elements (insufficient cover and excessive Type 1: New Build (excludes demolition, plot, 21,000 300 distance between confinement bars). and temporary housing) Type 2: Basic Retrofit (excludes repair of EQ 2,100 30 damages and roof replacement) -Installing missing confinement at wall 10 openings, including sill, jambs and lintels -New walls and infills 3 -Strengthening select walls with mesh plaster 7 -Installing missing confining columns 9 (corners/free ends/bracing long panels) Type 3: Basic RF + EQ Damage Repair 2,660 38 -Installing missing confinement at wall 10 openings, including sill, jambs and lintels -Demolition and construction of new walls to 12 repair EQ-damaged ones; opening infills Photos: MEX-BC_e4600491, home of Marina Jacinto Suárez, Tehuantepec 2019. -Strengthening select walls with mesh plaster 7 -Installing missing confining columns 9 EXAMPLE 2: (corners/free ends/bracing long panels) Type 4: Basic RF + Slab Roof Replacement** 6,370 91 New house built with subsidy. -All interventions as listed for the Basic 30 Observed deficiencies: Retrofit (Case 2) −−Lack of confinement around wall openings that might generate wall failure in case of an earthquake. -Replace roof with reinforced concrete slab 61 −−Thin locally procured block, with low compressive strength. roof −−Lack of shear wall area. Type 5: Basic RF + Light Roof Replacement** 6,580 94 -All interventions as listed for the Basic 30 Retrofit (Case 2) -Replace roof with lightweight roof system 64 TYPICAL UNREINFORCED BRICK MASONRY HOUSE Basic Retrofit (excludes repair of EQ damages) 5,530 79 -New roof system*** 49 -Strengthening of corners 19 -Fixing missing lintels and columns at free 11 ends * Values listed include costs for materials and labors but do not include costs related to trans- port of material to the construction site or technical assistance. ** Roof replacement is necessary in many cases where the existing slab is deteriorated due to weathering and poor construction quality, and/or damaged by the earthquake, or where the existing lightweight roof did not comply with the required quality. Homeowners indicated a preference for slab roofs due to interior temperature and waterproofing concerns. *** The majority of these buildings have an existing tile roof, supported on a timber frame that is generally in poor condition, not sufficiently connected to the ring beam. For this reason, the retrofit intervention proposed generally includes replacing the roof with a structurally compliant roof. Photos: MEX-BC_8c4bd59b, home of Alfredo Rasgado Ordaz, Ixtaltepec 2019. SOURCE: BUILD CHANGE. 38 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | LEARNING FROM SUCCESSFUL INTERNATIONAL HOUSING RETROFITTING EXPERIENCES Mexico’s home improvement initiatives can also learn This menu could include long-term low-interest-rate from programs implemented in other countries. A review loans, local or national government grants, compensa- of retrofit programs implemented in the United States tion through property taxes, more favorable financing and Japan commissioned by the World Bank (Miyamoto terms for safer buildings, some pass-through costs, and 2019), offers several valuable lessons and recommenda- reduced property insurance. tions: • Long term programs can help ensure a higher rate of • Digital imaging and machine learning can be used to compliance. In some urban centers with massive num- identify the different existing building types. Other bers of vulnerable houses, a retrofit program could take sources of information such as census data, city devel- years or even decades to be completed. Setting a long- opment patterns, and site surveys can supplement this term schedule can also ease the securing of funds by data and verify its accuracy. Once the building types house occupants, thus promoting compliance. A phased are known, a prioritization of those that cause larger approach can be used; for example, allow some time for losses of lives should be carried out. screening, followed by development and submittal of retrofit plans, and then a number of years to complete • Buy-in from the various relevant stakeholders is criti- the actual work. cal in the project development phase, during which in- puts from developers, government officials, engineers, • It is important to develop regulatory instruments, such potential financers, and the public can be obtained to as codes for new construction, retrofit guides for multi- shape the program. The lack of understanding of disas- -unit buildings, and handbooks for single-unit building ter risk and knowledge on safe construction is a pri- and enforce their compliance. For smaller units, resi- mary reason for the lack of action on mitigation. In this dents can do some of the basic retrofits themselves at phase, communication and information strategies are a low cost if they have some basic training or booklets key to create awareness about the importance of ensur- to reference and if they can borrow tools from the city. ing that buildings are structurally safe. Local forums for the exchange of ideas among the various stakeholders • Develop the methodology for creating a pilot project in as well as between different cities can be organized. a vulnerable city. This task involves project develop- Web-based information portals can also help educate ment, including all key components such as regulations, the public and disseminate information about the proj- financing, engineering, and outreach. References listed ect and serve as repository for relevant documents. in this document can provide more detail on the vari- ous stages of the program development. • Regulatory capacity and compliance of housing retrofit programs can be enhanced by capacity building activi- • Implement a small-scale pilot project for a district in a ties targeted at government staff and stakeholders, in- city that is based on the developed methodology. Use cluding engineers, architects, masons and contractors, in it as a test case to refine the program for more general both the formal and informal sectors in the target area. application. The success of such a program will help to obtain buy-in from various stakeholders, to ensure that • A flexible menu of financial assistance schemes can help the program is transparent, and to encourage partici- convince the population to participate in the program. pation by incorporating suggestions for improvements. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 39 HOW THE MEXICAN GOVERNMENT ALLOCATES SUBSIDIES CAN EITHER INCREASE OR DECREASE THE RISK OF THE HOUSING STOCK. 40 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 4. Conclusions After completing the research on Mexico’s housing policies and institutional framework, as well as the pilot intervention in the Istmo de Tehuantepec, the team has arrived at the following conclusions: • The Mexican government has correctly identified the need to focus on sustaining reconstruction efforts and promot- ing resilience and social inclusion. Although it is necessary to continue working in the production of new housing, Mexico’s qualitative housing deficit is significant. In a country where the majority of the population inhabit areas vulnerable to catastrophic events, resilience should be an important component of every home improvement inter- vention. Governments can either increase or decrease risk with the allocation of housing subsidies. • However, several policy aspects require further development before the government’s vision can be realized. It cur- rently does not possess the necessary data to inform policy decisions on where and which houses to intervene, the retrofit interventions required for each type of vulnerability, and the financial products that meet the needs of the different population segments. To further develop its policy approach, the government needs to collect this informa- tion more efficiently. • The experience and reach of Mexico’s institutional set up, lessons learned through decades of implementation of housing policies and programs, complementary hazard information and recent technological advances can help close the existing institutional, policy and information gaps. But there is also a need to act at the housing unit level, for which detailed and complete information is required. Traditionally, obtaining this information has entailed long, labor intensive and costly processes. Today, however, the methodology proposed by the bank can save money and time and increase precision. • The Bank’s methodology for the assessment of the housing stock proved to be affordable, reliable, and scalable. The fieldwork, which was conducted in two weeks, can be replicated in other areas of the country. The cities of Salina Cruz and Juchitán de Zaragoza present several similarities but also differences on size, terrain, and prevalence of dif- ferent construction materials. These differences provided an opportunity to test the bank’s methodology on different environments. • The pilot intervention in the Istmo de Tehuantepec showed that the application of this methodology could help CONAVI reach the most vulnerable segments of the population with a strong value proposition that combines physi- cal security, quality of life and respect for the use of local materials and construction practices. NEXT STEPS The next step is the proof of concept of the feasibility of applying the vulnerability assessment methodology at scale, extending the scope to a greater area. Based on the results of this pilot, a functional model can be adjusted for applica- tion in other municipalities in the state of Oaxaca or in other regions that are exposed to different hazards, with distinct geographical, sociocultural and economic conditions, so that the model can be used for multiple hazards and locations. As a result of the proof of concept, valuable input can be provided to the government of Mexico for improving the tar- geting and delivery efficiency of housing programs, including resilience at their core. The approach to “build better before” provides a great opportunity to address housing vulnerabilities, improve coordination between public and private organi- zations and provide specialized technical assistance to guide and accompany home improvement and self-construction pro- cesses. Additionally, the methodology an technological tools applied can can be used for other objectives such as property valuation and tax collection, as is being done in Peru and Colombia, thus addressing the challenge of better linking housing, urban governance, and disaster risk management policies. Most importantly, a renewed focus on housing resilience allows to put the most vulnerable and largely underserved population at the core of Mexico’s housing assistance model. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 41 HAZARDS POSE THEIR GREATEST THREATS TO SUBSTANDARD HOUSES, AND THE FAMILIES LIVING INSIDE. 42 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | ANNEX 1 Imagery Collection and Analysis To inform the allocation of home improvement subsidies, spatially explicit and comprehensive data is needed. In the cities of Salina Cruz and Juchitán the team created a rich housing stock database to demonstrate the technical feasibility and resulting benefits of such data. Images of the urban housing stock were taken from the sky (with drones) and from the ground (with a street view camera). Later machine learning algorithms were applied to extract pertinent information about each building. This Annex provides technical information on this exercise. Drones Fixed-wing eBee X drones were flown over the city of Salina Cruz and DRONE COVERAGE Juchitán —producing 3.6 cm resolution natural color images and a 7.1 cm Salina Cruz Juchitan digital elevation model (surface and terrain). This imagery was then mosa- iced and ortho-mosaiced to achieve a positional accuracy of ~50 cm with the use of Real-time Kinematic (RTK) positioning. More specifically, these EBee X endurance activated drones were carrying SODA 3D sensors so that build- ing and terrain height could be modeled. The pilots faced several challenges —strong wind (up to 12 mph daily), steep terrain, and minimal in the field ­ access to take-off and landing sites. Despite these challenges, within 6 days two pilots collected 31.3 km2 of land. The drone imagery extracted covered 20 km2 in Salina Cruz and 11.3 km2 Juchitán). Street view For the street view images, a Trimble MX7 camera was mounted to an SUV STREET VIEW COVERAGE resulting in 360-degree image every 2 meters. This camera has 6 lenses and Salina Cruz Juchitan can enhance the precision of location data derived from satellite-based posi- tioning systems (global navigation satellite systems, GNSS) and Real-time Ki- nematic (RTK). In the city of Salina Cruz every navigable road was driven re- sulting in a total of 579 km, and in Juchitán, roughly 136 km were captured. The street view team also faced challenges —such as a tropical depression (heavy rain), steep terrain, unpaved roads and dead-end streets. Protests and deteriorating security conditions resulted in less than antici- pated street view data collected in Juchitán. Despite these challenges —337,263 panoramic images, 2,023,578 cube images and more than 5.63 TB of data were collected. Analysis Once drone and street-view images were collected, machine learning algorithms were trained to extract certain charac- teristics of each house. To analyze the street view, the team stood up an open source annotator (CVAT) and ran the model on using Tensor- WORKFLOW FOR STREET VIEW MACHINE LEARNING Street level images Computer Vision CVAT XML Training data Annotation Tool (CVAT) | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 43 flow’s Object Detection API. The model was built using over 50,000 labels and predicted 9 different characteristics, resulting in a total of 140,625 predictions in Salina Cruz and 65,754 predictions in Juchitán. From the drone images, rooftops were delineated, then 6 characteristics were generated: height of roof, area of roof, volume of building, terrain slope, roof material and condition. Finally, all street view characteristics were attributed to the build footprint layer using the cardinal direction and trajectory information from the street view image. The result of all this imagery collection and machine learning is that for each building with have up to 15 attributes. Results This exercise demonstrated that within a short amount of time, it is possible to generates a detailed geospatial database that provides decision makers with valuable information regarding the housing stock that they are responsible for. In Salina Cruz and Juchitán, where seismic activity frequently damages homes, understanding the quality and materials being using in the construction is vital. Furthermore, this approach allows for the detection of buildings that are likely to collapse during an earthquake —i.e tall buildings with large weak walls (large windows and garages). Such structures can now be flagged and examined further by engineers in the field. Lastly, when integrated with field work, this database allows for the creation of a building typology. For example, during this exercise experts in the field determined that res- idential buildings that are 100 square meters large, constructed of brick/plaster, will cost USD 50 a meter for retrofitting. This database can be queried to locate and provide a cost estimate for retrofitting all such buildings in the City of Salina Cruz. The ability to extract such knowledge as the potential to inform Mexico’s urban management policies in the field of disaster risk management (DRM), infrastructure planning, and climate change mitigation. 44 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | ANNEX 2 Main actors in the Mexico housing sector Mexico’s housing sector is made up of multiple public and private actors that are involved in the design, financing and development of social housing. In the last 45 years, more than 120 entities have had a stake in Mexico’s housing policy. Today, the main institution responsible for policy design in the housing sector is the Ministry of Agrarian, Territorial and Urban Development (SEDATU). The main executing body of SEDATU in the housing sector is the National Housing Commission (CONAVI), alongside other public and private entity including the National Housing Funds, state housing agencies at state level and other entity with co-responsibility in the housing sector. NATIONAL HOUSING AGENCIES (ONAVIS) • SEDATU (Ministry of Agrarian, Territorial and Urban Development). It is the agency of the Federal Public Administra- tion that has as its main purpose is to plan, coordinate, manage, generate and execute public land planning policies, guarantee decent housing to all, and manage urban and rural development. • CONAVI (National Housing Commission). Created in 2006, it has been an important operator of federal housing sub- sidies. Previously it acted as coordinator of housing policy but those responsibilities have been transfered to SEDATU and CONAVI is currently in charge of operating the federal subsidy programs. • INFONAVIT (Institute of the National Housing Fund for Workers). Created in 1972 as a housing fund for salaried workers in the private sector. Its resources come from the mandatory savings of employees and companies that are discounted via payroll. INFONAVIT is the largest mortgage lender in Mexico; it mostly focuses on financing new housing developments. • SHF (Federal Mortgage Company). It is the second most important lender by volume of credits. In 2011, it became a second-tier banking entity. • FOVISSTE (Housing Fund of the Institute for Social Security and Services of State Workers). It is the equivalent of INFONAVIT for public sector employees and is approximately one tenth in size, according to the number of affiliates. • FONHAPO (National Fund for Popular Rooms). It was created in 1982 with the objective of serving the poorest seg- ment of the population through targeted housing interventions. FONHAPO was closed in 2019 and its functions were transferred to CONAVI. STATE HOUSING ORGANIZATIONS (OREVIS) AND MUNICIPALITIES Each of the 32 States of Mexico has a State Housing Agency financed with state revenues, as well as several municipal institutes. The volume of operation of these entities represent only a fraction of the subsidies provided by the federal government. OTHER ENTITIES WITH HOUSING ASSISTANCE SCHEMES A number of specialized housing funds operate alongside INFONAVIT and FOVISSTE, such as ISSFAM (Institute of Social Security for the Mexican Armed Forces). Other institutions also operate dedicated housing assistance programs such as CFE (Federal Electricity Company), PEMEX (Petróleos Mexicanos) and some non-profit organizations such as Habitat for Humanity Mexico. In addition, the Ministry of Social Development (SEDESOL) has operated significant volumes of housing improvement actions focused on the poorest population segments. PRIVATE SECTOR Many private entities such as micro-finance institutions, cooperatives and popular savings banks provide access to fi- nance for home improvement. Some private entities also offer technical assistance services. The private sector also plays a role in the provision of construction materials. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 45 46 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | ANNEX 3 A note on terms and definitions The purpose of this note to guide the reader about the terms and definitions used in this report. Regarding the housing deficit, official sources point to different estimations of Mexico’s qualitative and quantitative housing deficit. The Govern- ment refers to both the deficit identified in the National Information System and Housing Indicators (SNIIV) (9.2 million households, representing 28% of the total number of households based on 2015 census data) and the deficit calculated by CONEVAL (2018), which estimates the national housing deficit to amount to 14 million, or 45% of the overall housing stock. In the report, a table is included in page 11, which contains the housing deficit estimates used by different entities. For this report, the quantitative housing deficit is understood as the gap existing between the demand for housing and the existing housing stock. Furthermore, the qualitative housing deficit refers to the housing units that are inadequate as defined by UN Habitat, namely as lacking one or more of the following elements: access to water, access to sanitation, sufficient living area, structural quality, durability and location, security of tenure, affordability, and accessibility and cultural adequacy.1 The report also refers to home improvement interventions. They can be defined as those interventions that seek to ex- pand housing, make repairs, provide basic services or any other improvement in the quality of housing. It includes ret- rofit interventions, which are understood as those consisting on repairs, reinforcements or additions of housing features to reduce the risk of collapse. Finally, this report revolves around resilient housing in the context of an increasingly urban Mexico. A World Bank report (2014) defined resilience in an urban context as the “ability of a system, entity, community, or person to adapt to a variety of changing conditions and to withstand shocks while still maintaining its essential functions”.2 In relation to housing, re- silience can be understood as the capacity of housing units to protect lives and/or assets when unfavorable events occur. 1 UN-Habitat (2018). SDG Indicator 11.1.1 Training Module: Adequate Housing and Slum Upgrading. United Nations Human Settlement Pro- gramme (UN-Habitat), Nairobi. Available at: https://unhabitat.org/wp-content/uploads/2019/04/Indicator-11.1.1-Training-Module.pdf 2 World Bank. 2014. “An expanded approach to Urban Resilience: Making Cities Stronger.” Washington DC, in GFDRR (2015) “Investing in Urban Resilience”. Available at: https://www.gfdrr.org/sites/default/files/publication/Investing%20in%20Urban%20Resilience%20Final.pdf | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 47 48 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | ANNEX 4 Glossary of abbreviations AO Administrador de Operaciones —Operations Administrator. ASA Advisory Services and Analytics. CENAPRED Centro Nacional de Prevención de Desastres —National Disaster Prevention Center. CIDOC Centro de Investigación y Documentación de la Casa —Housing Research and Documentation Center. CONAVI Comisión Nacional de Vivienda —National Housing Commission. CONEVAL Consejo Nacional de Evaluación de la Política de Desarrollo Social —Nacional Council for Evaluation of the Social Development Policy. EGIR Estrategia de Gestión de Riesgo Integral de México —Integral Risk Management Strategy of Mexico. ENIGH Encuesta Nacional de Ingresos y Gastos de los Hogares —National Survey of Household Income and Expenses. FONDEN Fondo de Desastres Naturales —Natural Disasters Fund. FONHAPO Fondo Nacional de Habitaciones Populares —National Fund for Popular Housing. FOVISSSTE Fondo de Vivienda del Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado —Housing Fund of the Institute of Social Security and Services of State Workers. GIZ German Corporation for International Cooperation. INEGI Instituto Nacional de Estadística y Geografía —National Institute of Statistic and Geography. INFONAVIT Instituto del Fondo Nacional de la Vivienda para los Trabajadores —Institute of the National Housing Fund for Workers. OEO Entidad Ejecutora de Obras —Works Executing Entity. ONAVIS Organizaciones Nacionales de Vivienda —National Housing Organizations. UN Habitat United Nations Human Settlements Program. OREVIS Organizaciones Estatales de Vivienda —State Housing Organizations. PCU Perímetros de Contención Urbana —Urban Containment Perimeters. RUV Registro Único de Vivienda —Single Housing Registry. SEDATU Secretaría de Desarrollo Agrario, Territorial y Urbano —Ministry of Agrarian, Territorial and Urban Development. SEDESOL Secretaría de Desarrollo Social —Ministry of Social Development. SEMARNAT Secretaría de Medio Ambiente y Recursos Naturales —Ministry of Environment and Natural Resources. SHF Sociedad Hipotecaria Federal —Federal Mortgage Society. SNIIV Sistema Nacional de Información e Indicadores de Vivienda —National Information System and Housing Indicators. UMA Unidad de Medida y Actualización —Unit of Measurement and Update. USD US Dollar. | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 49 50 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) | 51 52 | MEXICO | IMPROVING HOUSING RESILIENCE (P169278) |