AGO – 2024 State of the Art of Social Registries in Latin America and the Caribbean State of the Art of Social Registries in Latin America and the Caribbean Acknowledgments This technical note was prepared by Luis Eduardo Castro (Consultant) and Luz Stella Rodríguez-Novoa (Senior Social Protection and Jobs Specialist) from the World Bank. The note benefited from recommendations from William Wiseman (Practice Manager for Social Protection and Jobs in Latin America and the Caribbean), Nelson Gutierrez (Senior Social Protection Specialist) and suggestions from Phillipe Leite and Luis Iñaki Alberro (Senior Social Protection Specialists). The team appreciates the lo- gistical support provided by Mathew David Romero and Paula Flores. The team is also grateful for contributions from Carlos Encalada Nicolalde (General Technical Coordina- tor, Social Registry Unit, Ecuador), Germán David Briceño Villalobos (Deputy Director of Poverty and Targeting, Social Registry of Households, Colombia), Irene Hernández Carazo (Psychologist), and Natalia Rojas (IT Professional, IT Coordinator) from SINIRU- BE, Costa Rica, Tereza Cotta (Project Manager, Cadastro Único, Brazil), Alexis García (Coordinador Padrón de Beneficiarios, Panama), Francisca Moreno Montealegre (Head of the Social Policy Division, Undersecretariat of Social Evaluation, Ministry of Social Development and Family, Chile), Viviana Cruzado de la Vega (Director, Design and Methodology Directorate, General Directorate of Targeting and Social Information, Peru), Jefrey Lizardo (Director SIUBEN, Dominican Republic) and our World Bank col- leagues: Gastón Blanco, Santiago de la Cadena, Pascal Jaupart, Ailo-Klara Manigat, Úr- sula Martínez, Julieta Trías, Marcela Salvador, Marla Spivack, Briana Wilson, Clemente Ávila, Nancy Rocío Banegas, Marina Petrovic, Miriam Montenegro, Raquel Tsukada, and Ignacio Apella for their valuable contributions. Catalina Murcia carried out the layout of this technical note. The note was originally written in Spanish and translated into English using the WBG-Translate 2.0 tool. Readability was improved using mAI WBG GPT. -3- Acronyms CCSS Costa Rican Social Security Fund CSE Socioeconomic Classification of Households (Chile) DNP National Planning Department (Colombia) ECLAC Economic Commission for Latin America FIBE Basic emergency file (Chile) GAD Departmental Autonomous Governments (Ecuador) IFI International Financial Institutions IMAS Joint Institute of Social Assistance (Costa Rica) IRS-2018 Social Registry Index (Ecuador) LAC Latin America and the Caribbean RHS Social Registry of Households (Chile) System for the Identification of Beneficiaries of Social Programs SISBEN (Colombia) SIUBEN Single System of Beneficiaries (Dominican Republic) SIFO Targeting System (Peru) URS Social Registry Unit (Ecuador) -4- Content Pag. Executive Summary 6 I. Introduction 8 II. The role of Social Registries in social policy 9 III. Evolution of Social Registries in LAC 14 IV. The role of Social Registries during the COVID-19 pandemic 18 V. Social Registries in LAC today 20 1. Institutional arrangements 22 2. Collection and updating of household information 24 3. Generation of socioeconomic indices of households and their characterization to determine eligibility for programs 33 4. Information systems and interoperability 38 5. Performance measures 41 VI. Challenges and recommendations for LAC Social Registries 45 References 50 -5- Executive Summary Social registries are fundamental tools for social protection. These regis- tries make it possible to identify people in situations of poverty and vulnerability, fa- cilitating in an efficient way the targeting of resources, the widespread use of infor- mation for public policy decision-making, and promoting the population’s access to relevant services and benefits. Social registries are the gateway for potential beneficiaries to access certain social pro- tection programs. They are inclusion sys- tems, since they allow the registration of citizens and residents of a country who need support for enrollment in the social programs that are available, are used to assess the socioeconomic profiles (iden- tifying the population in poverty and vul- nerability), and the needs of the house- hold, so that they can be considered as potential beneficiaries of certain public policies. They are also information sys- tems, as they consolidate and store all the data and relevant information of these individuals and households. In this context, the social registries serve not only the country’s social institutions, but also serve with their information the dif- ferent programs and interventions at the national and subnational levels. Over the past two decades, social reg- gaps in care, (iii) identify potential sets istries in Latin America and the Carib- and sequences of interventions and con- bean (LAC) have experienced significant ditions for exiting social programs, and growth in terms of coverage, interoper- (iv) quickly identify potential beneficia- ability, and usage. Except for countries ries and mobilize responses to shocks such as Argentina and Uruguay, which and emergencies. have very robust social security informa- tion systems and have developed other Despite their accelerated development types of information systems linked to and the innovations introduced in re- their social security systems, most coun- cent years, LAC’s social registries face tries in the region have begun to develop significant challenges in expanding their social registries that consolidate informa- use and consolidating themselves as tion on the poorest and most vulnerable instruments of social policy and man- populations and have gradually expanded agement. In particular, it is necessary to: their coverage to other income groups. (i) strengthen the legal frameworks and Thus, the coverage of social registries has the institutions that support them, (ii) in increased significantly, reaching more many cases, identify the optimal coverage than 80% in countries such as Chile, Co- of the registries and improve the quali- lombia and Costa Rica. ty of the information, (iii) optimize the mechanisms for collecting and updating Social registries play a decisive role household information, which includes along the delivery chain of social pro- improving interoperability and its role in tection programs, especially in the pro- keeping household information current cesses of dissemination, admission and and updated, (iv) create a new relation- registration of households in the stage ship and fluid communication with citi- of evaluation of the socio-economic con- zens and other users, including national ditions of individuals and households. and subnational public institutions that Therefore, they are more frequently used use registries to allocate benefits. This beyond targeting of individual policies or technical note seeks to contribute to the cash transfers. Social registries fulfill dif- knowledge of the social registries of the ferent functions that allow them to: (i) LAC region, to understand their evolution identify eligible populations for different and the legal, institutional, technological types of benefits and services according and operational challenges they face, and to program eligibility criteria, (ii) nomi- to contribute to the knowledge agenda of nally track social investment in individu- delivery systems at the global level. als and households and identify potential -7- I. Introduction The World Bank has been working for several decades to strengthen social registries in LAC and globally. This technical note aims to contribute to reducing the knowledge gap on the state of the art and challenges of social registries in LAC, identify where they are in relation to various analysis criteria, and generate recommendations on strategies to strengthen and consolidate their role in the design, implementation, and evaluation of social policies in the region. This note draws on the evidence and knowledge generated by other previous World Bank publications, particularly the recent work of Leite et al., 2017, Lindert et al., 2021 and Grosh, Leite et al., 20221 as well as the work of Berner, H. et al., 2020 of ECLAC, which served to inspire the conceptual framework and provided existing pri- or evidence. Based on the proposals for analysis in these documents, a framework for information collection and analysis is proposed. Similarly, the note benefited from the development of a survey completed by officials working in the various entities in charge of social registries and by World Bank social protection specialists in LAC. This survey has been complemented with semi-structured interviews with regional experts with relevant experience. The technical note collects information from 17 countries2. This technical note is organized into five sections, including this introduction. The second section presents the role of social registries in social policy. Section three pres- ents the evolution of social registries in LAC and how they adapted to the challenges that arose to provide immediate responses during the COVID-19 pandemic. The fourth part generally describes the state of the art of social registries in the region, analyzed from five pillars: (i) institutional arrangements, (ii) structural aspects3, (iii) implemen- tation chain, (iv) information and interoperability systems, and (v) performance mea- sures. The fifth section presents the main challenges facing the registries and the rec- ommendations of the study. 1 World Bank, Social Registries for Social Assistance and Beyond: A Guidance Note and Assessment Tool (Leite et al., 2017), Fundamentals of Social Protection Implementation Systems: A Reference Book (Lindert et al., 2021), and Revisiting Targeting in Social Assistance: A New Look at Old Dilemmas (Grosh, Leite et al., 2022). 2 Argentina, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Haiti, Jamaica, Mexico, Panama, Paraguay, Peru, and Uruguay. 3 It refers to the relationship of the Social Registry with other institutions and with the citizen. -8- II. The role of Social Registries in social policy II. The role of Social Registries in social policy Social protection and employment systems are essential for accumulation and protection of human capital throughout the life cycle. There are different so- cial protection interventions that enable people to maintain and increase their accu- mulation of human capital, which in turn translates into improvements in individual and household living conditions. In this context, social registries are essential tools for social protection and economic development. They support the identification of indi- viduals experiencing poverty and social vulnerability, enabling their participation in so- cial protection programs. This facilitates efficient resource targeting and promotes the widespread use of information for public policy decision-making, all within a life-cycle approach. Figure 1. Life Cycle and the evolution of human capital Source: World Bank (2020). -9- State of the Art of Social Registries in LAC Social registries are fundamental tools for the implementation of social protection systems. They are inclusion systems, as they allow support the first three steps of the delivery chain to disseminate information, register citizens who need support for their enrollment in the social programs that are available, and assess the socioeco- nomic profile needs of the household, so that they can be considered as potential beneficiaries of certain public policies. In some countries, social registries also support other parts of the program delivery chain, including case management. They are also information systems, since they store all the data and relevant information from these households, in order to assess their condition and allow interventions to be managed more efficiently (Leite et al.,2017). In this context, the social registries in each country serve with their information to the different social protection programs and interven- tions and public policies at the national and subnational level. Figure 2. The delivery chain and social registries (a) Delivery chain of social protection programs (b) Role of the Social Registry in the implementation system Source: World Bank (2020), Sourcebook Foundations of Social Protection Delivery Systems. - 10 - II. The role of Social Registries in social policy Social registries play an important role throughout the delivery chain of social pro- tection programs, especially (but not exclusively) in the processes of dissemination, admission and registration of the households, in the stage of evaluation of socio-eco- nomic conditions of individuals and households. In this stage, the goal is to identify the eligible population to target interventions most effectively. With adequate mecha- nisms for interoperability, the information in social registries also enables the nominal and individual monitoring of beneficiary households for each intervention. Within the delivery chain, social registries seek to be more inclusive and respond more rapidly, especially at times of emergencies. A social registry aims to reduce operational complexity of having different databases for different programs for the same type of beneficiaries. This reduction in complexity leads to cost and time savings associated with data collection, cleaning, and transformation, thereby achieving econ- omies of scale. All these efficiency gains aim to improve the citizen experience by hav- ing a single institution collect and manage information. This institution then informs other government agencies for the purposes of analysis, application, and selection of various social protection programs and public policies. For example, in the event of catastrophic events, the availability of an integrated and up-to-date registry helps reducing the need to raise parallel registries to respond to the emergency, or even allows preventive interventions to be anticipated if the data of households have been cross-referenced with climate or seismic vulnerability data. Information is the primary input and output of social registries. Inputs include data provided by households, such as identification documents, geospatial location, home characteristics, demographic details, employment information, and specifics about special conditions. This may include information on disabilities, sanitary conditions, and household health. This information may be collected in the field on mass surveys or census sweeps, in citizen service desks often operated by national authorities, or by a combination of the two methods, according to the characteristics of the geographi- cal area (for example: in areas of greater vulnerability it is advisable to carry out census sweeps and areas of greater institutional presence and higher development can prior- itize over the counter data collection)4. Computer exchanges (or interoperability) with existing administrative registries databases (e.g., with identification databases, social security, credit bureaus, among others) also serve to update the information in the registries or to validate information previously collected. Therefore, the information becomes the output of social registries that allows the socioeconomic classification of individuals and households. Likewise, sets of indicators associated with risk conditions in the face of possible impacts of natural disasters are included in the social registries. 4 Self-reported information can also be collected via digital interfaces such as apps or websites. - 11 - State of the Art of Social Registries in LAC All this information is standardized and shared with the different social protection pro- grams. Each of the programs has the responsibility to define – independently – the conditions of eligibility of their different interventions (according to the objectives, coverages, and availability of resources). Finally, this information can also be used to perform analysis for both monitoring and public policy decision-making. In this way, the information cycle is completed. Based on these analyses, it is possible to identi- fy new information needs, make improvements, reduce redundancies, and optimize data. This allows for the redesign of inputs and the resumption of the information production chain from the social registries. Figure 3. Information and its main role in social registries Source: Authors. - 12 - II. The role of Social Registries in social policy Social registries are key elements of the digital public infrastructure5. By interoper- ating with identification information, social registries help promoting or strengthening foundational identification systems through the creation of “identity mappers” that allow the authentication of the identity of users and provide them with access to dif- ferent products (i.e. financial products). On the other hand, by interoperating in coor- dination with payment platforms, social registries help to reducing time and costs both for the government and for the customer; and to guarantee deposits of social benefits in bank accounts or financial products chosen by the beneficiary6. Social registries enable bidirectional data flow and generate benefits through econ- omies of scale by reducing operational complexity. Departments and agencies at the local or subnational level can access a universal database and use it for functions related to their objectives. This allows them to design programs that reduce inclusion and exclusion errors by identifying households, their socioeconomic characteristics, and potential needs. 5 Defined as the digital network that enables countries to safely and efficiently offer economic opportunities and social services to all residents. 6 Payment platforms generally interoperate with the programs’ own systems since these are those who select the final beneficiaries based on the information in the social registries. Consult the “Playbook on Digital Social Protection Delivery Systems Towards Dynamic Inclusion and Interoperability” for greater reference on the interaction between the different digital social protection systems. - 13 - State of the Art of Social Registries in LAC III. Evolution of Social Registries in LAC Social registries in LAC have been conceptually discussed for more than four decades. In 1976, emerges in Chile the idea of generating a system to targeting social benefits. Similarly, in the mid-1990s, Colombia introduced the System for the Iden- tification of Beneficiaries of Social Programs (SISBEN). This system was designed to collect household information by filling out forms that enumerate the needs of each household. At the end of this same decade, Argentina launched the National Tax and Social Identification System (SINTyS) with a novel concept for its time: the use of ad- ministrative records for accessing and targeting benefits. However, most of the social registries in LAC have emerged in the last two decades, always linked to institutions with responsibility in the design and implementation of social policy. The emergence of registries stems from the need for information to sup- port the processes of determining eligibility for social programs in a consistent manner and to reduce costs for both the government and citizens. A good part of the registries initially emerged from registries of beneficiaries of cash transfer programs, such as the Social Registry of Ecuador and the Household Targeting System (SISFOH) of Peru, and then moved towards broader and more comprehensive models that were used to target a larger universe of programs. However, it is important recognize that the experience of all countries has been different and that several countries have had intermittent registries, which have been triggered in response to certain government programs and then have been discontinued or even restarted (i.e. Honduras, Guate- mala, and Mexico). Since their creation, social registries have aimed to generate information sets that promote the inclusion of the population in social programs. As a result, its coverage has increased over time, although it has experienced important variations in some - 14 - III. Evolution of Social Registries in LAC countries, such as Ecuador, Peru, and Colombia. These countries, that during the 2000s, had coverage rates that ranged between 30% and 60% of the population, ex- perienced declines due to the difficulties in keeping the information up to date during the 2010s. However, with the specialization of functions and advancements in data- base exchange and interoperability, multiple registries have managed to maintain or improve their coverage. In this context, Colombia has successfully included 100% of its population in the social registry. This achievement is based on innovations in in- teroperability introduced during the COVID-19 pandemic, the creation of the Master Base, and subsequently, the Social Registry of Households. These developments have become the operational framework for SISBEN and other available registries (Table 1). Figure 4. Creation of LAC social registries Source: Authors. - 15 - State of the Art of Social Registries in LAC BOX 1. Institutional evolution to the creation of the Social Household Registry in Colombia Colombia’s social registry has significantly evolved and vulnerable groups, it has developed into an over the past three decades. Since the creation of integrated system covering 100% of the country’s the System for the Identification of Beneficiaries population. This system estimates household in- of Social Programs (SISBEN) in 1995, which aimed come based on relevant variables. to target health and education spending on poor Figure B.1. Colombia: Social Registry’s Evolution 1995-2023 In its first three iterations (1995, 2005, 2011), SIS- The Social Households Registry (RSH) currently BEN manually collected household information contains more than 55 million records and in- through surveys or sweeps focused on the poor- teroperates with 43 databases and over 1,800 est areas, establishing a socioeconomic classifica- subnational government databases. This exten- tion of households to target social benefits. In its sive repository of information not only helps iden- fourth iteration (2021), SISBEN IV, managed by the tify potential beneficiaries but also characterizes National Planning Department (DNP), initiated a them, identifies supply gaps, and verifies features process of digital data capture for household in- or compliance with conditionalities. formation. For the first time, it also georeferenced household locations, a transformative element for The architecture of the Social Households Regis- managing public policies, and established a virtual try (RSH) is supported by three major databases window for online procedures. that enable the identification of the population: (i) identification of Colombian citizens (National The COVID-19 pandemic generated unprecedent- Registry), (ii) people affiliated to the Health Sys- ed demand for information and, for the first time, tem (Unique Database of Affiliates of the Minis- facilitated the exchange and interoperability of try of Health), and (iii) identification of regular data. In response, the Base Maestra was created migrants (Migración Colombia). Similarly, the RSH in 2020 as a large repository of administrative and interoperates with administrative records from household information. This process led to the various sectors, including health, education, labor, official launch of the Social Households Registry social assistance, and census lists, among others. in 2022, established as an integrated registry that This is achieved through different technological combines population information (demand) with exchange mechanisms (VPN, Batch, Web Ser- subsidies and other benefits (supply) (Decree 890 vice) and by establishing clear rules for data hi- of 2022). erarchies, business rules, and feedback from the sources. - 16 - III. Evolution of Social Registries in LAC Starting in 2023, as part of the National Devel- clude identifying more reliable data sources for opment Plan, the government of Colombia has income information and designing mechanisms to embarked on the project of implementing the encourage citizens to report their income to the Universal Income Registry (RUI). This initiative will government. This effort will require a delicate and enable Colombia to transition from a revenue pre- systematic coordination process with subnational sumption model to a hybrid model based on ex- governments, as well as a communication strategy isting administrative records and self-declaration, to dispel myths surrounding the targeting instru- particularly for the informal and lower-income ment. population. The challenges in this new stage in- Source: Briceño, German (DNP), 2024. Improvements in coverage of social registries have resulted in an increase in the number of programs who use these registries to target their profits. Initially, the use of social registries was primarily focused on cash transfer programs. However, in most countries, this has evolved to include targeting in areas such as education, health, com- bating chronic malnutrition, economic inclusion, and employability programs, among others. Additionally, social registries have extended their application beyond social pro- tection to intersectoral programs, including energy subsidies, housing, public services, and other subnational government programs (Figure 5). Figure 5. Evolution of the coverage and use of social registries Source: Authors. - 17 - State of the Art of Social Registries in LAC IV. The role of Social Registries during the COVID-19 pandemic The COVID-19 pandemic put a heavy strain on the social protection of LAC and confirmed the usefulness of social registries. The countries with robust social registries were able to respond more quickly and effectively to the needs of its citizens. Table 1. Innovations in social assistance and social registries due to the COVID-19 pandemic Type of social Innovations made protection measures to social registries (number) (number of measures) Country Innovation in New action or Changes in the Program/ Incorporation the sources of service in existing selection of users new measure of technology information program of social benefits being used Argentina  3 6 1 1 2 Belize 1 1 0 0 0 Bolivia 0 6 0 1 0 Brazil  2 3 0 1 3 Chile  1 8 2 1 1 Colombia  1 17 0 2 2 Costa Rica  3 12 3 1 2 Dominican Rep.  1 5 0 1 1 Ecuador  2 3 0 0 1 El Salvador  1 3 1 0 1 Guatemala 1 1 1 1 1 Haiti 1 3 0 1 2 Mexico  1 0 0 0 0 Panama 0 7 0 3 0 Paraguay 2 5 0 3 3 Peru 1 12 1 4 6 Uruguay  2 4 0 1 4 Total 21 94 8 20 28 Source: Berner, H., & Van Hemelryck, T. (2020), Palacios, R. (2020), 18 - U., Almenfi, M. B. A., López, A. V., & Zafar, U. (2022) -Gentilini, IV. The role of Social Registries during the COVID-19 pandemic In LAC, 64% of the population was covered by some type of protection measure during the pandemic (Gentilini et al, 2022). To reach these measures, countries implemented two major types of innovations in their social registries: (i) incorporation and cross-ref- erencing with new sources of information, such as bank data, cellular and mobile tele- phony, and satellite mapping for more detailed information on the situation of house- holds; and (ii) use of cutting-edge technologies such as machine learning to analyze the information and classify households more precisely. The countries of the region chose to expand their social registries through the pro- motion of interoperability with administrative data. They can be identified three main ways of expanding: (i) Combining interoperability with new applications of citi- zens to the registry (as in the case of Brazil and Chile), (ii) exclusively interoperability with administrative data (such as Colombia, Ecuador, Argentina, Peru7 and Uruguay), and (iii) expansions based solely on new applications of citizens to specific programs (such as El Salvador, Haiti, Mexico, and Paraguay) (Berner et al., 2020). Table 2. Pop-up expansions in social registries during the COVID-19 pandemic Expansion based on Expansion based on Country administrative data new applications Argentina  • Bolivia • Brazil  • • Chile  • • Colombia  • Costa Rica  • Ecuador  • El Salvador  • Haiti • Mexico • Paraguay  • Peru • Uruguay  • Source: Authors’ compilation. 7 In the case of Peru, the expansion was achieved because the Census could be used. - 19 - State of the Art of Social Registries in LAC V. Social Registries in LAC today Leite et al. (2017) propose a detailed and structured tool to assess social regis- tries based on five dimensions (see Figure 6): • Characterization of the current institutional status analyzes the laws and the framework legal in which it works. • Structural aspects: refers to its relationship with other institutions and with the citizen. • Business process aspects: related to process of data collection, verification, and socioeconomic classification. • Aspects of information systems: related to the type of information system, its design, the supporting architecture, the security framework of the current in- formation and interoperability protocols for both consumer and consumer and for sending information. • Performance: develops an analysis of coverage, equity and access, quality of the data and use of the Social Registry. Based on this methodological framework and considering the information restrictions, this note proposes the simplification of the five dimensions of Leite et al. (2017) to present an overview of the state of the art of LAC’s social registries8 around five areas: 1. Institutional arrangements: aspects related to the institutional framework, role and governance of social registries. 2. Mechanisms for collecting and updating household information: The pro- cesses of collecting household information and its updating over time, differ- 8 This simplification implies that elements such as the governance of information systems, the mechanisms of the specific characteristics of the databases, the management of the information system and measures such as data quality, system integrity, and performance measurements have not been considered in this technical note. This note seeks to provide general elements to compare the different LAC countries and surely it will require further in-depth analysis. - 20 - V. Social Registries in LAC today entiating between census sweeps and those that have open over the counter windows for the citizens. 3. Generation of socioeconomic indices of households (Targeting): assesses the methodology that each registry uses to classify and identify the socio-economic profile of households. 4. Information systems and interoperability: the existence of interoperability protocols is evaluated for both consumption and shipping of data from social programs and administrative records. 5. Performance measures: the analysis focuses on the coverage and usage of in- formation of social registries9. Figure 6. Dimensions of the analysis of social registries (Leite et al, 2019) Source: Leite et al. 2017. 9 The information for the analysis of the five components described above was collected using two Instruments: (i) A survey sent to officials of the social registries of the region, which was complemented by with semi-structured interviews; and (ii) review of secondary sources that have documented cases. - 21 - State of the Art of Social Registries in LAC 1. Institutional arrangements In terms of institutional arrangements, although each country operates under a dif- ferent framework, in general, there is a close link between social registries and the ministries of social policy or planning. While most country social Figure 7. Institutional models registries in the region, from of social registries in LAC which information for this note was obtained, are managed by the Ministry of Social Devel- opment, only two countries have their registries under the Ministry of Planning. Notably, Ecuador has an independent institution responsible for its social registry (Figure 7). Regarding the legal frame- work, in Argentina, Brazil, Chile, Colombia, El Salvador and Peru, the institutionality of its registries originates from related laws always within the scope of social protection. This means that its institutions are rooted in the core of social policies. On the other hand, in coun- Source: Authors’ compilation based on own tries such as Ecuador, Bolivia questionnaire and references. and Uruguay, the registries have been established through presidential or executive decrees, which gives them similar functionalities than their peers in the first group, but this institutional design makes them more vulnerable as they are susceptible to being unilaterally eliminated, without the need for discussions with the legislative branch. In all cases, social regis- tries maintain close interaction with all institutions in the field of protection in order to coordinate their actions and fulfill their functions effectively. - 22 - V. Social Registries in LAC today Reflections 1. Challenges of the institutionality of registries: Need to understand the role of human talent and the budget as the key resources that allow the operation of social registries There is a growing need to understand the ment (DNP), employs 55 people at the national institutional framework of social registries by level. Information collection is the responsibility assessing their human talent and operation- of subnational governments, which conduct sur- al budget. However, there is not homogeneous veys or sweeps and operate citizen service win- information that allows us to make an adequate dows for updates or new survey requests. The comparison since in many cases the officials re- core staff of the RSH focuses on information sys- ported that the information was not available or tems, as much of the household data is collected that was not public. through interoperability with various information sources. The central operation of the RSH is fully This table summarizes key elements extracted funded by the national budget, estimated at USD from the information available at the time of pre- 1 million per year. Subnational governments co-fi- paring this technical note for some countries. nance the information collection and citizen en- gagement activities. Comparative analysis in three countries: In Costa Rica, SINIRUBE, created in 2013, oper- ates with 14 staff and 3 contractors for IT support. In Ecuador, the URS, created in 2018, has 85 staff It does not have a defined organizational struc- members distributed across various areas: data ture. Costa Rica has placed significant emphasis collection, information monitoring, information on promoting interoperability with administra- analysis, information systems, and administrative tive registries. Despite having only three people and legal support. The URS supervises data col- from information systems, SINIRUBE is support- lection operations conducted by operators and ed by the hiring of more than 10 programmers certified universities, selected based on cost and and a project leader, which allows it to address quality. Currently, the government of Ecuador is the needs that arise for system improvement and transitioning to a model that will involve depart- data collection. Households’ data collection is mental autonomous governments (GADs) in the carried out by the Joint Institute of Social Assis- collection of household and individual informa- tance (IMAS) and by the Costa Rican Social Secu- tion. The operation of the URS and data collection rity Fund (CCSS). There are ongoing efforts to ex- efforts are financed by the government, including pand the number and type of institutions that can a loan from the World Bank. The estimated aver- collect households’ information, including edu- age annual operating budget is 1.8 million USD, cation institutions and subnational governments. excluding census sweeps. However, SINIRUBE does not have citizen service windows. Costa Rica does not have an estimate of In Colombia, the RSH, which originated in 1995 the cost of the household information collection and is managed by the National Planning Depart- that is conducted externally. - 23 - State of the Art of Social Registries in LAC 2. Collection and updating of households information In most countries in the region, the inclusion of households in the social registry is still static, that is, it only occurs at the time of the sweeps and does not respond to dynamic criteria of permanent updating. Census sweeps and surveys continue to be the predominant collection mechanism with large periods between each sweep, and therefore may have limitations on the use of the register for shock responses. Despite this, a good part of the countries of the region is moving towards hybrid models, which combine the self-declaration of households in physical or virtual windows and the col- lection of information via administrative data. In Ecuador, the URS is launch- Figure 8. Households ing a pilot program with mu- data collection models nicipalities called the “Modelo de Actualización Permanen- te.” This program aims to open service windows for citizens to update their information and collect data on new house- holds. Currently in its initial phase, the model seeks to en- sure that information collec- tion processes adhere to the highest quality standards. If home visits are required, they will be conducted by operators certified by the Social Registry, including municipal staff. In Peru, municipalities collect household information with- in their jurisdictions through their Local Census Units (ULE). They employ two strategies for data collection: (i) upon request by the household, Source: Authors’ compilation based on own questionnaire and references. - 24 - V. Social Registries in LAC today and (ii) mass registration in Figure 9. Models to update targeted areas. The model is households information being improved to allow the use of administrative data for real-time validations and en- hancements of coverage. Key considerations for maintaining high standards and ensuring constant updates include the possibility of setting goals for municipalities within the Mu- nicipal Incentives Plan and establishing more regular mechanisms for transferring resources from the central government. Regarding the mechanisms for updating information, home visits remain the pri- mary method, but there is a shift towards models that respond to continuous de- mand. This shift includes the opening of citizen service win- Source: Authors’ compilation based on own questionnaire and references. dows, both in-person and vir- tual, where individuals can apply for benefits, consult their information, and, in some cases, update their records. This approach is seen in countries with more advanced operational registries, such as Chile and Brazil, where user-centric models enable citi- zens to play a significant role in continuously updating their information. The quality of the information in the social registries is vital to their operation. The social registries in countries such as Chile and Brazil have internal and external verifi- cation processes to assess the consistency and quality of their data. In the same line, the registries of Ecuador and the Dominican Republic are also submitted to external audits, for example, carried out by academic centers or international organizations. In the case of Colombia, there are protocols to assess the quality of the records collect- ed in the field and verification of analysis of consistency in compiled administrative records. - 25 - State of the Art of Social Registries in LAC Despite these advances and the commitment to the quality of the records, achieving accuracy in determining household composition is a challenge of great proportions. The countries they face many difficulties to accurately and automatically map the in- dividuals that belong to a family nucleus or a household, depending on each country’s definition. Therefore, the authorities of the countries have had to resort to interview processes and validation of information, regardless of the initial capture or update model. - 26 - V. Social Registries in LAC today BOX 2. Chile: The role of information in the RSH and the actors involved in their local management In Chile, the RSH manages to integrate all as- activities at the public level, or exploitation and pects of the information provided by citizens at analysis of information by part of universities, re- the municipal level. The information is verified, search centers or civil society organizations (Fig- validated and managed to inform policy decisions ure B.2). access to benefits, and for aspects of monitoring Figure B.2. Data and information management The RSH management model is decentralized and each municipality, offers the same trainings, an IT requires the participation of a variety of actors platform, technical support and resources (based who fulfill complementary functions for provi- on contributing performance indicators to the fi- sioning, updating, and maintenance of the data. nancing of the management of the RSH to local On the one hand, the municipalities, through of level). The commitment of the municipalities is to their social management areas are responsible have a municipal manager of the RSH, enumera- for receiving, processing (input, updating, provi- tors or social workers, and staff in the social areas sioning, of supplementary information) and man- to process applications. age the requests from citizens (monitoring, com- munication, resolution of applications). For these On the other hand, other government agencies effects of the Ministry of Social Development and and entities that do not necessarily deliver ben- Family (MDSF), which signs an agreement with efits or services but have relevant information - 27 - State of the Art of Social Registries in LAC about the households (i.e. taxes, social security, Chile has leveraged technological advances to severance pay) operate as providers of adminis- gradually replace self-reported data with ad- trative data and sign exchange agreements with ministrative data for measuring the standard of the MDSF information that regulate the data to living of individuals and households. By the end be shared, the periodicity of the delivery, and of 2023, less than 10% of RSH households were data protection and exchange safeguarding stan- self-reporting their income. This shift has allowed dards. eligibility decisions for potential beneficiaries to be based not solely on income but on a compre- Social welfare institutions and providers of social hensive analysis of household socioeconomic vul- benefits share information about the design of nerability. their programs (e.g., eligibility criteria for their respective benefits) with the MDSF. In turn, the The changes introduced since 2016 have been MDSF provides them with technical assistance facilitated by the widespread access to the inter- and information available on the RSH to: (i) Veri- net and mobile devices, which has enhanced dig- fy compliance with the agreed eligibility criteria, ital communication with households. As a result, including Socioeconomic Qualification, (ii) Gen- more than 70% of RSH applications in 2023 were erate a list of potential users based on defined submitted digitally. prioritization criteria, (iii) Use the socioeconomic qualification assigned to a household to target a Another factor that has improved data interoper- benefit, sometimes in addition to other targeting ability between institutions is the existence of the criteria. Single Identification Registry (RUN), which serves as a unique identifier extending even to the mi- Additionally, providers of social benefits, wheth- grant population, and the unique key system er or not they use RSH information, must report (digital identification system) that standardizes their payrolls or lists of beneficiaries to the MDSF the relationship between citizens and the state. according to agreed terms and periodicity, based Updates to information, whether generated by on the characteristics of the service. citizens or through interoperability pathways, are not implemented instantaneously. Instead, they Finally, citizens play a key role in the operation of are verified using data hierarchy protocols. the RSH. As primary users, registered individuals and households are responsible for keeping their Despite these significant advances, challenges self-reported information up to date to ensure for Chile’s RSH persist: (i) determining household that beneficiary targeting decisions are based on composition, which relies on self-reported data reliable and valid information. In Chile, citizens and contends with the incentive for households have access to information that the RSH has over to report smaller or single-person households; (ii) them. This web platform allows the owner of the Improving interoperability processes and proto- personal data and adult member of a household cols; (iii) reaching remote and difficult-to-access access, after authenticating of their identity, to a areas; (iv) effectively communicating with citi- set of basic information1. zens, who often perceive that the registries limit their ability to access social benefits. 1 This information includes the following information: (i) identification of the owner, (ii) household composition, without personal information of the members, (iii) information that affects their Qualification Socioeconomic (income, number of older adults, number of dependent people or with some degree of disability, among others), (iv) Benefits, programs and/or social benefits to which persons who have previously accessed make up the household, (v) benefits, programs and/or benefits social Source: World Bank (2018), Chile’s networks that people could eventually access that make Households Social Registry and up the home. Poblete, Paula, Chile’s MDSF (2024). - 28 - V. Social Registries in LAC today BOX 3. Dominican Republic: The process of households’ data collection in SIUBEN and its quality control process The Dominican Republic has consistently aimed 2. Quality control in processes to ensure that its processes for collecting infor- of information gathering: mation from households meet the highest quality standards. To achieve this, three types of controls SIUBEN monitors compliance with all standards have been established. established in its manuals and protocols for the following processes: 1. Quality certification to the • Training of Interviewers Registry Business Processes • Fieldwork: This includes the strategy for data SIUBEN has 3 international certifications. collection, distribution, workloads, online • General Management: The central data col- fieldwork monitoring, and random re-inter- lection and usage processes are ISO 9001 views in the field. certified. This certification helps the institu- • Review of Data Captured Online: This involves tion monitor and manage continuous quality controlling the daily flow of data, monitoring across all its processes, identifying areas for data deviations and field productivity, remote improvement. monitoring, and managing alerts within secu- • Information Management: The institution rity frameworks. holds an ISO 27001 certification, which es- tablishes best practices for implementing an 3. Data certification: information security management system. As shown in Figure B.3, SIUBEN has an internal This ensures that citizens’ data is protected, audit process that validates information, partic- managed as SIUBEN’s greatest asset, main- ularly identity, against administrative records. tains privacy, and generates trust with both The consistency of updated data is also verified collaborating institutions and the citizens who with surveys and national household survey pat- provide the data. terns. Additionally, an external audit is conducted • Business Process Continuity: The institution is to validate the database, the Quality-of-Life Index, also ISO 22301 certified, which aims to guar- and the integrity of the information. This external antee business process continuity through a verification process is intended to be carried out continuous management system. This certifi- by universities, international financial institutions cation builds security and resilience under the (IFIs), United Nations organizations, or other in- current operating scenario and provides tools ternational technical cooperation agencies. to respond to potential shocks, ensuring the continuity of processes, particularly the col- lection and generation of household informa- tion. - 29 - State of the Art of Social Registries in LAC Figure B.3. SIUBEN Data and Information Management in the Dominican Republic Source: SIUBEN (2023), Presentation “Social Information System: Guarantee for the Implementation of Adaptive Social Protection Policies for the Most Vulnerable Households”. - 30 - V. Social Registries in LAC today Reflections 2. Need to understand the validity (duration) of the household information and the cost associated with data collection The dynamics of information collection for the associated collection costs to ensure that the social registries are evolving, which generates data reflects the current characteristics of house- controversies. Institutions responsible for social holds and individuals, especially during responses registries are analyzing the validity of the infor- to shocks and emergencies. mation and implementing practices to ensure their databases are dynamically updated. The de- 2. Use of administrative data bates focus on two main elements: On the other hand, the use of administrative data 1. Permanent update or allows for significant cost reduction and stream- periodic census sweeps? lines the information collection process. It also enhances the accuracy and reliability of data, en- Since data is the greatest asset of social regis- abling more complex analyses. However, much of tries, countries are making efforts and seeking this interoperability is limited to individuals in the strategies to ensure that the information remains formal labor market, excluding a significant seg- current and reliable. Constant updating, whether ment of the population in the region, namely the through cross-referencing with administrative re- informal sector. Thus, using administrative data cords or incentivizing users to update their infor- presents challenges in both obtaining and harmo- mation regularly (virtually or in person), enhances nizing the data, as well as ensuring and protecting data accuracy and availability. This leads to better the privacy and security of the information. Addi- targeting of social programs and a greater capaci- tionally, there are risks of algorithmic discrimina- ty to respond to socioeconomic changes. Howev- tion associated with this strategy. er, constant updating presents several challenges, including increased technical and administrative Several reflections emerge from this analysis: complexity, the need for robust quality control mechanisms, and the risk of bias in the informa- Duration or validity of information: Having a tion, among other issues. household registry interconnected with data and a service window for the public (virtual or Census sweeps or periodic surveys are costly, physical) open every day of the year represents logistically complex, and sometimes invasive for significant challenges in terms of information citizens. Establishing virtual or physical service management. It is necessary to dimension the ex- points at the local level requires the participation pectation of updating the information by house- of various actors, demands technological, physi- holds and ensure that it remains valid. cal, and financial resources, and necessitates the development of specific technical and administra- • Chile and Brazil are pioneers in enabling up- tive capacities. Additionally, it requires rigorous grades based on changes in household com- coordination between national and subnational position and location. The key to this model government levels. is that households have a significant incentive to update their data (e.g., birth of a new child, Despite these challenges, keeping the informa- death of a member, loss of employment). This tion current is critical to ensure the accurate allo- update is a requirement to apply for national cation of benefits. It is essential to understand the and subnational social protection and employ- validity (duration) of household information and ment programs, either as new users or for the - 31 - State of the Art of Social Registries in LAC renewal of benefits. In Brazil, if households do form, the national or regional scale of data not update their information, the maximum collection, geographical conditions and logis- term for updating household files is two years. tics of the process in the field, among others. A factor that supports continuous or dynamic updating is the existence of an extensive net- • In countries such as Brazil, Chile, Colombia, work of service windows (the CRAS in Brazil), and Peru, where municipalities actively par- which is essential for keeping self-reported ticipate in data collection, calculating the unit data up to date. cost becomes even more complex because it often involves resources from subnational • In contrast, countries like Ecuador and Cos- governments. In Costa Rica, where informa- ta Rica do not necessarily provide a clear in- tion is collected by government institutions centive for households to update their data. (IMAS and CCSS), the cost of collection is dif- Therefore, their social registries have opted to ficult to calculate as it is integrated into other define a term or period of validity for the in- tasks carried out by the institutions in the ex- formation in the household registry, typically ercise of their competencies. between 3 to 5 years. However, this approach limits the ability to reflect dynamic changes • The cost of maintaining administrative records in household composition, such as transitions is associated with the development of infor- into or out of informality and unemployment. mation and interoperability protocols, as well as the constant maintenance of technological • Chile, Brazil, and Colombia are making faster tools and information storage platforms. progress in using records or administrative data. However, in these countries, there is an In this context, it is evident that countries need to ongoing discussion about whether decisions calculate or approximate their costs of data col- should be based solely on the information the lection and the resources available for permanent government already has about citizens or if updating. Being transparent about these costs additional household consent should be re- will allow for a clear understanding of the glob- quired (through a physical or virtual window). al cost of social registries, helping to identify the Additionally, there is debate on how quickly population and secure the necessary resources to changes in a person’s or household’s classifi- ensure their sustainability. cation should be reflected—whether imme- diately with better data or with a margin to avoid abrupt loss of benefits. Cost associated with collecting households’ data: • One of the frequent consultations when ex- plore or evaluate processes of sweeping infor- mation is the estimate of the unit cost (e.g. household survey) on which there are large information gaps. For 2023, Ecuador and Co- lombia reported for example, a unit cost per household of between USD 11 and USD 12, respectively for the survey via sweeps. How- ever, the cost of the survey is determined by multiple variables including extension of the - 32 - V. Social Registries in LAC today 3. Generation of socioeconomic indices of households and their characterization to determine eligibility for programs Information collected from individuals and households, whether it is gathered for the first time or through continuous updates, is essential for profiling the living con- ditions and well-being of the household. Social programs that use the social registry may apply different criteria to identify potential beneficiaries. For programs that need to identify the monetarily poor, models that measure household livelihoods using proxy variables (PMT) are predominant in the region. Some countries, such as Brazil, use income statements directly and administrative records for information verification, while Chile uses administrative records to conduct direct evaluations of livelihoods (Table 2). Figure 10. Models to assess socio- Therefore, it is necessary economic status of households for registries to reach agree- ments with social programs to ensure they use the same income estimation method. However, it is important to remember that the social reg- istries support not only pro- grams serving the poor but must also ensure the quality of information for all variables beyond income. In the specific case of defin- ing the level of household well-being in the region using a monetary value, technical innovations have been intro- duced in the estimation of the PMT. These innovations have significantly improved household classification mod- els based on machine learning techniques. Colombia, Costa Source: Authors’ compilation based on own Rica, and Ecuador are clear questionnaire and references. - 33 - State of the Art of Social Registries in LAC examples of countries using these techniques to model household income or con- sumption information. The methodological approach aims to minimize errors in both household inclusion and exclusion by training algorithms to more accurately predict household classification. Despite these methodological advances for the estimation of the PMT model, coun- tries are moving towards hybrid models that combine data from Income from ad- ministrative sources, such as social security or registries tax and pension taxes, with self-declared information in household files. Although countries are looking to rely much less on self-reported data, high rates of labor informality in the region prevent reliable information of household income, which is why there is still a need to collect data directly of families through cards or forms, since there is a segment of the pop- ulation for which there are no administrative data that allow effective identification in the registries. - 34 - V. Social Registries in LAC today BOX 4. Best Practices for Income Estimation to Support Eligibility Determination for Social Programs in LAC: Ecuador and the Use of Machine Learning for the Socioeconomic Classification Algorithm of Households As of 2019, Ecuador updated the methodology mographic characteristics, location, housing, and for its PMT by adopting statistical learning mod- household goods. Using machine learning, the els. The URS, with support from the World Bank, survey data were divided into training and testing conducted a methodological exercise to calculate samples. Statistical techniques were applied with per capita consumption and income. This was the goal of minimizing inclusion and exclusion er- based on surveys of employment and household rors during training. living conditions to generate a new Social Registry Index (IRS-2018). Based on the models obtained, predictions of per capita consumption or income were made and The development of the model involved exploring transformed into percentiles, resulting in indices five modeling techniques for consumption and ranging from 0 to 1. Figure B.4.1 summarizes the income prediction, using variables based on de- flow of the methodology. Figure B.4.1. Summary methodological flow of the exploration to obtain the IRS 2018 Table B.4.1. Summary of results IRS 2013 IRS 2018 Nonlinear Principal Component Analysis Extreme gradient model with robust Methodology (NPCA) specification using computational learning Survey for model Survey on Household’s Socioeconomic Survey on Living Conditions (ECV), 2013- estimation Situation (ESSHO, 2012) 2014) Poverty line 34.67905 25.83095 Extreme poverty line 24.08766 5.74848 Rating error: 12.98% Improvement Classification error: 35% Inclusion error: 7.24% Exclusion error: 5.74% Source: URS (2018), “Methodology for the Construction of the Social Registry Index 2018”, Official Presentation for Discussion with National Entities of the Central Government. - 35 - State of the Art of Social Registries in LAC Among the models explored, the gradient boost- nomic Classification (CSE). The total monthly ing technique proved to produce the least error income values reported by households are com- in terms of inclusion and exclusion, with a general pared against the values obtained and computed classification error of 12.98% for per capita con- by the Social Household Registry (RSH) through sumption. This technique is based on a series of administrative income records. These records in- regression trees, where predictions are made us- clude tax data, mutual fund information, private ing household and member variables, along with pension payments, and other sources. custom weight parameters for household profiles. It is noteworthy that the classification error in the For the calculation of per capita income, the to- Ecuadorian case is the lowest recorded compared tal household income is adjusted by considering to other countries in the region for which infor- the household size. This value is then weighted mation is available: Colombia (30%), Costa Rica by an adjustment factor based on the age of its (46%), and Panama (54.2%). members. The objective is to reflect the higher resource demands that occur in households with In Chile, the means test for socioeconomic clas- children and/or elderly adults. The CSE places sification of households is based on income. The each household in a certain bracket of income country uses a means test that relies on self-de- and vulnerability (Table B.4.2). clared household income to build the Socioeco- Table B.4.2. CSE brackets and description Bracket Percentile Description of socioeconomic classification Households that belong to the 40% of households with the lowest income or the Bracket of 40 0% – 40% greatest socioeconomic vulnerability. Households between the 41% and 50% of households with the lowest income or Bracket of 50 41% – 50% greater socioeconomic vulnerability. Households between the 51% and 60% of households with the lowest income or Bracket of 60 51% – 60% greater socioeconomic vulnerability. Households between the 61% and 70% of households with the lowest income or Bracket of 70 61% – 70% greater socioeconomic vulnerability. Households between the 71% and 80% of households with the highest income or Bracket of 80 71% – 80% less socioeconomic vulnerability. Households between the 81% and 90% of households with the highest income or Bracket of 90 81% – 90% less socioeconomic vulnerability. Households between the 91% and 100% of households with the highest income Bracket of 100 91% – 100% or less socioeconomic vulnerability. Source: World Bank (2018). Chile’s Households Social Registry. - 36 - V. Social Registries in LAC today At least 110 services directly or indirectly use When reviewing the information by CSE section, information from the Social Household Registry the distribution of benefits by CSE bracket is as (RSH). Of these, 91 use the Socioeconomic Classi- follows: fication (CSE) for the allocation of corresponding benefits and services. • 40% bracket (highest vulnerability): More than 95% of the resources allocated corre- • Transfers (monetary and in-kind) and sub- spond to transfers and subsidies. sidies: Represent 74.7% of the resources as- signed. • Upper brackets of lower vulnerability: The proportion of transfers and subsidies decreas- • Services: Include training, labor intermedia- es. tion, technical assistance for economic activ- ities, social services, among others, and rep- • Exception – 90% bracket: The proportion of resent the remaining 25.3% of the resources transfers and subsidies is close to 100% due assigned. to the Housing Subsidy for middle-income sectors, which is the program with the largest budget for this vulnerability bracket. Figure B.4.2. Distribution of allocated resources according to the type of benefit per CSE bracket (year 2017) Source: World Bank (2018). Chile’s Households Social Registry. - 37 - State of the Art of Social Registries in LAC 4. Information systems and interoperability As discussed throughout this technical note, registries benefit from promoting interop- erability. As for the progress on this front in the countries of the region, three groups can currently be observed: • Greater progress: The region’s registries show steady growth in interoperability. Chile, Brazil – along with Argentina and Uruguay that have registries generated from their tax ans social security information systems – lead this aspect, not only in terms of programs, but also in the collection of information on income, assets and pensions, with a periodicity that varies from monthly to semi-annual. • Medium progress: Colom- Figure 11. Interoperability models bia, Costa Rica, Ecuador and Peru. In these coun- tries, interoperability is moving forward and seeks to focus on the correct use of the administrative records associated with identification, household income, assets (land, vehi- cles, etc.). In addition, they use additional information from records of beneficia- ry registries of social pro- grams. • Incipient interoperability processes: For the rest of the countries, the mini- mum objective is to estab- lish automated communi- cation with the institutions responsible for carrying out the identification pro- cesses. Source: Authors’ compilation based on own questionnaire and references. - 38 - V. Social Registries in LAC today BOX 5. Brazil: Interoperability and usage of administrative data Brazil’s Cadastro Único operates Figure B.5. The architecture of Cadastro Único based on citizen declarations, which to respond to the increased coverage demands are verified through interoperability resulting from the COVID-19 shock with various administrative records. (a) Internal level This system was put to the test during the COVID-19 pandemic, where not only interoperability processes but also those of application, data ex- traction, analysis, and beneficiary se- lection were challenged to provide a quick response. In Brazil’s case, this was possible due to the robust archi- tecture of its information system. Figure B.5 illustrates how internal data management is divided into database servers that operate via application servers and connect through an Ap- plication Programming Interface (API) with external services to consume administrative log data. This internal operation is complemented by the ex- ternal operation with citizens, where benefit requests are compared with monthly salary payment bases, Tax (b) External level Administration Data, and social pro- grams beneficiary registries. The expansion of interoperability al- lowed the use of 27 data sets, mainly on income, taxation, micro-enterpris- es, unemployment, and work insur- ance, in collaboration with the Single Registry of Beneficiaries of social pro- grams. As a result, it was possible to identify most of Brazil’s population, totaling 34.7 billion pieces of data. The availability of this information en- abled a response to 151.1 million re- quests for emergency benefits. These requests were collated with the results, allowing for an expansion beyond the common framework of Cadastro Único. The results of this interoperability and information analysis made it possible to identify a set of 68 million beneficiaries. Source: Ministry of Social and Family Development (2023), Presentation “Talking Interoperability Focus Brazil”. - 39 - State of the Art of Social Registries in LAC Reflections 3. Alternatives of social registries for the development of their information systems and data storage Information Systems proprietary • Security: Proprietary solutions offer improved versus non-proprietary usability and flexibility in uploading and trans- mitting encrypted information (physical or There is an ongoing discussion about whether cloud services). Non-proprietary solutions the tools used for social registries’ information require more development programming to systems should be proprietary or non-propri- ensure security. etary. Proprietary Systems are based on licensed software and equipment owned by a copyright • Maintenance: Proprietary systems have an- holder. Non-Proprietary Systems have an open ar- nual license costs or other arrangements to chitecture platform, allowing the use of software maintain the operability of database engines and hardware from different providers, offering and user interfaces, in addition to human re- more flexibility. sources costs. Non-proprietary systems have personnel costs, which can be more intensive. The decision to develop either type of system is usually based on each country’s policies regard- Choices and Challenges: Countries face complex ing software and information technologies. decisions in this area to adapt to national policies and budgetary and operational constraints. The In the consultations carried out for this tech- challenge for the future is to maintain operation- nical note, information from ten countries was al information systems and databases with high identified: 3 use systems non-proprietary (Ecua- availability and sustainable costs. dor, Guatemala and Haiti) and 7 use proprietary systems (Argentina, Brazil, Colombia, Costa Rica, Types of information storage Chile, Panama and Peru). and capacity to processing The choice between proprietary and non-propri- The storage of information is a significant chal- etary systems depends on several factors: lenge for all social registries, especially because they handle large-scale volumes of information. • Budget: Non-proprietary tools generally re- In the region, much of the consumption data, quire more development and a more robust, such as administrative records, is stored in the trained technical team, but they do not re- information systems of the entities that generate quire license payments. them. These can be physical storage, cloud stor- • Design: Non-proprietary tools offer great- age, or hybrid storage. The initial strategy of many er flexibility but are often more expensive registries, which aimed to become large reposito- in terms of man-hours spent on system pro- ries of information, is being rethought. Instead of gramming. maintaining a single, large repository of informa- tion, the focus is shifting towards consuming only • Scalability and Modularity: There is debate the data necessary for their work. over which option allows for better scalability and modularity of data and tools. Non-propri- Colombia consumes selected information, spe- etary systems offer greater design flexibility, cifically administrative data from relevant house- while proprietary systems provide better pro- holds. This approach involves consulting house- grammability and integration with other visu- hold files without duplicating information storage. alization tools, business intelligence, and data On the other hand, Brazil and Chile use a storage processing. model where they query data online and, at the - 40 - V. Social Registries in LAC today end of the day, generate and store only the per lect data associated with the business flow of reg- capita income of the household. These countries istries, (ii) information of identification: This type have separate systems for beneficiary registries, of information is expected to increase in volume, storing information relevant to programs, bene- enhancing security for individual data access but fits, and services. Costa Rica, Ecuador and Peru also increasing responsibilities for managing sen- are currently analyzing their data growth and con- sitive information, (iii) ideal server configuration ceptual models for administrative records, seek- with processing capacity: This will need to in- ing an appropriate storage model. It is important crease not only for reading, writing, and trans- to note that Brazil includes in social registration nationality of the data but also for real-time ana- only for poor and vulnerable people and not to lytics. This means having increased RAM capacity the entire population. for management and using business intelligence tools and online data processing. In this way, the challenges for storage of the data must take into account certain key elements: (i) Selection of variables and information: Ingest se- 5. Performance measures The first measure of the performance of a social registry is its coverage, especially of the poor and vulnerable population. The coverage of the social registries of LAC exceeds that of almost all other regions globally (only at the same level of MENA, but with 2.5 times more countries in LAC). Importantly, the coverage varies with countries with large coverage: Colombia, Costa Rica, Argentina, Uruguay and Chile with coverage of more than 80%. On the other hand, there is a group of lagging countries, mainly in Central America and Bolivia, which they have very low coverage, less than 10%10. How- ever, it is important to clarify that the depth of coverage within poor and vulnerable groups, who will benefit most from the social programs, is perhaps more crucial than the overall coverage itself. 10 In Bolivia, several categorical social assistance programs maintain their own separate registries and are not integrated with the Platform for the Integrated Registration of Social Programs (PREGIPS), which was established in 2011. - 41 - State of the Art of Social Registries in LAC Figure 12. Coverage of social registries (a) LAC countries with information Sources: Colombia: Authors’ survey; Costa Rica: Authors’ survey; Argentina: Berner, H., & Van Hemelryck, T. (2020); Uruguay: Berner, H., & Van Hemelryck, T. (2020); Chile: WB Delivery Systems GSG Data Base; Peru: Authors’ survey; Dominican Republic: WB Delivery Systems GSG Data Base; Ecuador: Authors’ survey; Brazil: Authors’ survey; Panama: Authors’ survey; Belize: WB Delivery Systems GSG Data Base; Honduras: WB Delivery Systems GSG Data Base; Haiti: Authors’ survey; El Salvador: Berner, H., & Van Hemelryck, T. (2020); Mexico: WB Delivery Systems GSG Data Base; Jamaica: WB Delivery Systems GSG Data Base; Bolivia: WB Delivery Systems GSG Data Base; Granada: WB Delivery Systems GSG Data Base; Guatemala: Authors’ survey. (b) Comparison of LAC with other regions Source: Authors’ compilation based on their own questionnaire and references. - 42 - V. Social Registries in LAC today Another key performance indicator is the use of information generated by social reg- istries. In general, in the region there is evidence of a widespread use of information to target cash transfer programs. When analyzing the data from the region, we can ob- serve a first group of leading countries that have expanded the use of social registries information for national public policies. Brazil and Chile have integrated the use of data not only for central government programs, but also for subnational governments, being leaders with at least 33 to 25 national programs that take advantage of the information from their social registries. A second group of countries that have moved beyond cash transfers and have expanded the use of information to areas such as health, education, and subnational governments’ programs. This group includes Colombia, Costa Rica, Ecuador and Peru, which are in a process of transitioning towards greater and more effective use of information. Finally, there is a lagging group that includes most of the countries of Central America and Mexico, which is still mainly focused on cash transfer programs. Table 3. Number of Programs Using Social Registry Information by Country and Sector in 2023 Source: Authors’ compilation based on own questionnaire and references. - 43 - State of the Art of Social Registries in LAC BOX 6. Applications of Social Registry Information for Adaptive Social Protection (ASP) Information collected by social registries en- national level to the individual one. The IVACC ables them to play a crucial role in activating pub- scores to all households in the SIUBEN database lic policy actions during shocks and emergencies. with values from 0 to 1. Variables that are con- sidered include the characteristics of the housing Chile and the Basic Fact and equipment, household income, and location Sheet of Emergencies (FIBE) in risk areas. As for its use, the IVACC is Used by various institutions, including: The FIBE is a survey collected by the RSH that al- lows for the gathering of information from house- • The National Emergency Commission (CNE):- holds affected by emergencies such as earth- For disaster planning and response. quakes, tsunamis, floods, forest fires, among • The National Integrated Information System others, that have suffered damage to their be- (SINI): For the identification and monitoring of longings and homes. The FIBE collects informa- threats, vulnerabilities and risks to phenome- tion declared by a member of the family group na Natural. over 18 years old, including variables such as: (i) Household location, (ii) Composition of the fami- • The Emergency Operations Center (COE): For ly group and characteristics of its members, (iii) storm preparedness and hurricanes. Special needs, (iv) Support networks and partici- pation, and (v) Damage to the home. • The Pacific Disaster Center (University of Ha- waii): For the National Reference Assessment Each survey is valid for a single emergency. If a for Disaster Preparedness. household experiences subsequent emergencies, new surveys will be conducted. Additionally, the • The United Nations System: For the formu- allocation of benefits depends on the type of lation of projects that seek to strengthen di- emergency and its impact on the families. saster preparedness in local governments and communities. Dominican Republic and its Index of • The National Housing Institute (INVI): For the Vulnerability to Climate Shocks (IVACC) identification of beneficiaries of their housing The IVACC is an index that calculates the prob- programs. ability that a household is vulnerable to the oc- • The Nature Conservancy: For Identification of currence of a climatic phenomenon such as hur- environmentally vulnerable communities. ricanes, storms and floods, depending on their socioeconomic and geographic characteristics. • The SUPERATE Program: The main program of cash transfers in the country, to estimate the The unit of analysis of the IVACC is the house- potential number of households to be affect- hold. This allows the geographical disaggregation ed by climatic shocks. of information at any cartographic scale, from the - 44 - VI. Challenges and recommendations for LAC Social Registries VI. Challenges and recommendations for LAC Social Registries In light of the reviewed evidence, the progress that social registries have made in the region is remarkable, not only in terms of coverage and use but also in innova- tions for the collection, storage, and quality assurance over time. If social registries operate in coordination with robust identification systems (for both nationals and mi- grants), with transparent and inclusive benefit payment platforms and systems, and with user-centered social case management mechanisms, they have the potential to become a fundamental pillar of the digital public infrastructure of the countries. De- spite the advances, there is still a long way to go, not only in development but also in the consolidation and sustainability of the registries at the national and regional levels. It is imperative to strengthen the legal frameworks that regulate social registries and strengthen the institutional framework that supports them, including the pro- tection of personal data. This will enable the promotion of coordination mechanisms at different levels of government and public policy intervention. Social registries are an integral part of the digital public infrastructure of governments, which means they require a solid framework to ensure their long-term stability and sustainability. There- fore, it is important to invest in the development of human resources with skills in handling and analyzing large data sets, promoting interoperability schemes, and en- gaging and interacting with citizens. Robust teams with a combination of technological expertise and an understanding of social policies are necessary. Similarly, it is import- ant to foster a culture of personal data protection among citizens and institutions that collect, share, or process such data. This strengthening is not limited to the national level; it is essential to expand capa- bilities at the subnational level, involving key actors in social policy and public man- agement. Additionally, adequate human resources, technology, and training (not only - 45 - State of the Art of Social Registries in LAC in the use of forms and data collection but also in the use of data for public policy decision-making) must be provided so that the registries can effectively carry out their functions. These functions include promoting mechanisms that facilitate interoperabil- ity with different programs and areas, as well as interaction with citizens. It is very important to establish cost-efficient and sustainable mechanisms to keep the information in the registries updated and to fulfill the ultimate goal of ensuring greater access to services and benefits for the poorest and most vulnerable popu- lations. As discussed, surveys or sweeps are generally complex and costly processes, and household information becomes outdated relatively quickly. In this regard, it is vital to identify alternative strategies for collecting household information through cit- izen service points (both in-person and virtual). While promoting the exchange and interoperability of information with other sources of administrative data is crucial, it is clear that the levels of informality and the nascent development of administrative records in many countries in the region pose a barrier to accessing updated and qual- ity information. Therefore, dual developments (improvement of registries and citizen service points) should be promoted. One immediate task is for countries to work on understanding or calculating the integrated cost of operating social registries: having a closer and more comparable measure to determine how much needs to be invested upfront in the correct identification and targeting of households is vital for achieving their sustainability. The costs should include the operating expenses of public admin- istration to implement and operate the registry (costs of communicating with citizens, data collection and updating costs, costs for IT systems, national and subnational infra- structure, among others) and the out-of-pocket costs for citizens to participate (travel costs, access to copies of identity documents, opportunity costs for foregoing other activities to complete this process, among others). It is necessary to periodically review and update the metrics of social registries and ensure the quality of the data to accurately determine the needs of applicants (potential beneficiaries), as well as to determine the potential benefit package and ser- vice strategy for individuals and households. In most social registries, this assessment is usually automated through algorithms or metrics that establish scores or ranges. Similarly, it is critical for households to frequently update their information so that the metrics reflect any changes. It is important to note that the determination of eligibility by the Social Registry is dif- ferent from enrollment in social programs. The determination of eligibility involves the assessment of needs and conditions with respect to basic eligibility criteria for social - 46 - VI. Challenges and recommendations for LAC Social Registries programs, while enrollment decisions involve the formal inclusion of eligible individ- uals or families in a specific program. Eligible enrollees (applicants) do not always or automatically enroll in a particular program due to: (i) budget constraints that can limit space and coverage, so that enrollment it is determined on a first-come, first-served basis with waiting lists; or (ii) additional criteria that guide enrollment decisions, be- yond socioeconomic needs and conditions, such as geographic prioritization, other categorical risk factors, etc. In addition, the institutional jurisdiction of these phases may differ: the decisions of registration are the competence and legal responsibility of the user programs, not of the Social Registry. This is particularly important if the Social Registry is administered by an institution other than the one responsible for imple- menting the programs (Leite et al, 2017). On the other hand, it is vital that the entities in charge of social registries have stan- dards, protocols, and validation frameworks to ensure that the data is complete for all registered individuals, that the information does not present inconsistencies or dupli- cations, and that mechanisms are in place to mitigate fraud. It is necessary to improve and enhance the interoperability of information, which not only contributes to improving spending efficiency but also facilitates a rapid re- sponse in times of emergency and shocks. There is no single model for managing da- tabases and interoperability. Information can be fragmented across multiple systems and teams and respond to various IT architectures. Similarly, it is important to under- stand which exchanges have the greatest potential not only in terms of coverage but also in terms of information quality and storage efficiency, and to define the protocols and depth of data exchange. For example, interoperability with risk and disaster data can facilitate collaboration between social protection agencies, natural disaster risk management agencies, and non-governmental organizations in these shock response strategies. It is worth building on the lessons learned during the COVID-19 pandemic. This has taught us the need to establish clear digital public infrastructure platforms, where social registries can easily identify a phone number, a bank account, and a national ID per household. This will not only improve communication with users but also facilitate citizens’ access to information about programs and potential benefits, especially in emergency situations. Similarly, interoperability with financial information or access to mobile phone services can help map households and provide them with access to the financial services they need. - 47 - State of the Art of Social Registries in LAC There is a need to further expand the coverage of registries, especially in areas with high rates of poverty and risk and ensuring that at least the population with the highest rates of poverty and vulnerability is covered. This involves the implementa- tion of active search strategies and the promotion of mechanisms for more frequent updates. However, it is important to note that this adaptation must be viewed through the unique lens of each country, as it will require a specific assessment to determine the ideal size of the registries social needs and particularities of programs and policies public of each country. Finally, social registries have the mission of creating a new relationship with house- holds fostering a culture that reinforces the importance of information. In this sense, it is essential to re-establish communication and interaction mechanisms with citizens, establishing clear rules on communication, duration of the benefits offered by social protection or other programs, the shared responsibilities by citizens and the fundamental role they play social registries in the chain of provision of social services. It is crucial that these aspects are understandable to citizens and that they clearly associate and see how social registries contribute to improving their quality of life. To achieve this, it is necessary to strengthen capacities at both local and national levels. An example of this approach can be found in the case of Chile, where very clear in- formational guides are provided to citizens, addressing topics such as the duration of benefits and the overall role of the social registry in the country. - 48 - State of the Art of Social Registries in Latin America and the Caribbean References (2007). González, M. S. SISFAM, a register of social vulnerability. From political ends to scientific ends. In IX Jornadas Argentinas de Estudios de Población. Association of Population Studies of Argentina. (2014). Aristy-Escuder, J. Vulnerability index to climate shocks: Dominican Republic. Santo Do- mingo: United Nations Development Programme (UNDP). (2017). Leite, P., George, T., Sun, C., Jones, T., & Lindert, K. Social registries for social assistance and beyond: a guidance note and assessment tool. World Bank. (2017). Marini, A., Gallagher, P., & Rokx, C. Standing tall: Peru’s success in overcoming its stunt- ing crisis. World Bank. (2018). National Council for the Coordination of Social Policies - Argentina. SINTyS 20 years. (2018). World Bank. Social Registry of Households in Chile. (2020). Berner, H., & Van Hemelryck, T. Social Information Systems and Registries of Recipients of Non-Contributory Social Protection in Latin America: Progress and Challenges in the Face of COVID‐19. (2020). Palacios, R. Scaling up social assistance payments as part of the Covid-19 pan- demic response. World Bank. URL: https://documents1.worldbank.org/curated/ en/099830009302217091/pdf/P1731660f8c52f062092ac00d53c648bac7.pdf (2022). Gentilini, U., Almenfi, M. B. A., López, A. V., & Zafar, U. Social Protection and Jobs Re- sponses to COVID-19: A Real-Time Review of Country Measures (Living paper version 16). World Bank. (2022). Grosh, M., Leite, P., Poi, M., & Tesliuc, M. Revisiting Targeting in Social Assistance: A New look at Old Dilemmas. World Bank. (2022). Leite, P., George. The importance of implementation mechanisms in social protection programs. World Bank. (2022). Poblete, M., Martínez, N. Social registries in social protection systems. Library of the National Congress of Chile. (2022). Social Protection and Employment in LAC 2022. (2022). World Bank. The Role of Digital in the COVID-19 Social Assistance Response. World Bank. (2023). Interoperability in Action: Social Protection Information System Interacting with Social Registry. World Bank presentation at Digital Convergence USP 2030. (2023). Social Information Registry of Chile. Ministry of Social Development and Family presen- tation and brief at Digital Convergence USP 2030. - 50 - (2023). Social Registry of Households. National Planning Department of Colombia presentation at Digital Convergence USP 2030. (2023). Talking Interoperability Focus in Brazil. DATAPREV presentation and brief at Digital Con- vergence USP 2030. (2023). Vedavalli, P., Kwatra, N., Srinivasan, S., & Sinha, V. Leveraging Digital Public Infrastruc- ture for building inclusive social protection systems. (2024). World Bank. Playbook on Digital Social Protection Delivery Systems: Towards Dynamic Inclusion and Interoperability. https://www.argentina.gob.ar/politicassociales/siempro/sisfam https://www.gub.uy/ministerio-desarrollo-social/siias https://www.mds.gov.py/index.php/noticias/mds-habilita-plataforma-para-acceder-constan- cia-del-rsh-en-el-marco-del-arancel-cero - 51 -