Jamele Rigolini • Sarah Coll-Black • Rafael de Hoyos • Ha Thi Hong Nguyen PROTECTING HUMAN CAPITAL THROUGH SHOCKS AND CRISES Jamele Rigolini • Sarah Coll-Black • Rafael de Hoyos • Ha Thi Hong Nguyen PROTECTING HUMAN CAPITAL THROUGH SHOCKS AND CRISES How lessons learned from the COVID-19 response across Eastern Europe and the South Caucasus can be used to build better and more resilient human development systems © 2023 International Bank for This work is a product of the staff of the World Bank with external contri- Reconstruction and Development / butions. The findings, interpretations, and conclusions expressed in this The World Bank work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments that they represent. 1818 H Street NW, Washington DC 20433 This publication was produced with the financial support of the European Telephone: 202-473-1000 Union. Its contents are the sole responsibility of World Bank staff and do Internet: www.worldbank.org not necessarily reflect the views of the European Union. The World Bank does not guarantee the accuracy of the data included in Rights and Permissions this work. The boundaries, colors, denominations, and other information The material in this work is subject presented or shown on any map in this work do not imply any judgment to copyright. Because The World on the part of The World Bank concerning the legal status of any territory Bank encourages dissemination of or the endorsement or acceptance of such boundaries. its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. iv PROTECTING HUMAN CAPITAL through Shocks and Crises Table of Contents Acknowledgments vii Executive Summary ix 1. Introduction 1 2. A Brief Primer on Resilience 3 What Do Resilient Systems Look Like? 6 3. Human Capital in the Time of COVID-19 9 4. Learning from COVID-19: Delving into Countries’ Pandemic Responses 13 Health 13 Social Protection 16 Education 19 5. Going Forward: Building Better and More Resilient Human Development Systems 23 References 27 v List of Figures: Figure 1: Factors Influencing People’s Resilience 3 Figure 2: Resilient and Non-resilient Human Development Systems 4 Figure 3: The Four Pillars of An Effective Shock and Crisis Response 6 Figure 4: GDP growth by quarter, 2019-2021 9 Figure 5: Growth in Remittances by quarter, 2019-2021 9 Figure 6: Employment Rates in Selected Countries 10 Figure 7: Poverty and Inequality Rates in Selected Countries 10 Figure 8: Government Expenditure on Education, 2000-2019 11 Figure 9: Government Expenditure on Health, 2000-2019 11 Figure 10: Harmonized Learning Outcomes in Selected Countries 11 Figure 11: Non-communicable Disease Burden in Selected Countries 11 Figure 12: Estimated Learning Losses in Different Countries 12 Figure 13: Out-of-pocket Health Expenditures 14 Figure 14: Approximate Allocations of Emergency Resources 17 Figure 15: Coverage and Spending of Poverty-targeted Programs 18 Figure 16: School Closures and School Learning Outcomes 20 List of Tables: Table 1: Foregone Care in Moldova 15 Table 2: Timeline of School Closure 19 vi PROTECTING HUMAN CAPITAL through Shocks and Crises Acknowledgments This work was led by a team consisting of Jamele Rigolini, Sarah Coll- Black, Rafael De Hoyos Navarro, and Ha Thi Hong Nguyen under the guidance of Fadia Saadah (Regional Director, Human Development, Europe and Central Asia Region), Arup Banerji (Country Director, Eastern Europe), Sebastian Molineus (Country Director, South Caucasus), Rita Almeida (Practice Manager, Education, Europe and Central Asia Region), Tania Dmytraczenko (Practice Manager, Health Nutrition and Popula- tion, Europe and Central Asia Region), and Cem Mete (Practice Manager, Social Protection and Jobs Global Practice, Europe and Central Asia Region). The report benefitted from the work of an extended team who co-au- thored the companion notes for each sector, comprising Stefanie Brodmann, Marjorie Chinen, Alejandro Gonzalez-Aquines, Anna Koziel, and Cornelius von Lenthe. The team would also like to thank Elvira Anadolu, Olena Doroshenko, Renata Freitas Lemos, James Gresham, Maddalena Honorati, Christine Lao Pena, Shiro Nakata, Mirey Ovadiya, Efsan Ozen, Lucian Pop, Naomi Rupasinghe, Sirma Seker, Tigran Shmis, Lucia Solbes Castro, Roberto Sormani, Adanna Deborah Ugochi Chukwuma, and Roman Zhukovskyi for great feedback and contribu- tions. Other World Bank staff and external experts also contributed to the preparation of the companion notes, where they are acknowledged. This report also benefitted from valuable feedback from Caryn Breden- kamp, Carolin Geginat, Mattia Makovec, Sarah Michael, Harry Patrinos, Mersedeh Tariverdi, William Wiseman, and Levent Yener. The team would also like to thank representatives of the European Commission for useful feedback and inputs, in particular Hoa Binh Adjemian, Thibault Charlet, and Corinne Deleu. The note was made possible by generous funding from DG NEAR’s Europe 2020 Trust Fund. vii viii PROTECTING HUMAN CAPITAL through Shocks and Crises Executive Summary Risk and uncertainty are on the rise, and coun- of shocks and crises means that it is necessary not tries across Europe and Central Asia (ECA) are not only to invest in human capital but also to protect immune from it. The region is being hit by crises, human capital from the effects of shocks and crises. conflicts, and continued uncertainty that are neg- Failing to protect human capital gradually erodes atively affecting people’s livelihoods in the short people’s health and skills and the ability of firms to term and prosperity in the long term. The 2008 climb the productivity ladder and, ultimately, limits financial crisis was followed by instability and a long-term prosperity. What is equally worrisome is series of regional conflicts that significantly reduced that shocks such as the COVID-19 pandemic are growth and prosperity. Then COVID-19 hit, inflicting profoundly regressive and thus exacerbate preex- massive damage on people’s wellbeing, livelihoods, isting inequalities in an already polarized world. and human capital. Lockdowns prevented people from working, school closures caused students to This report explores how to increase the resilience fall behind in learning, and overwhelmed hospitals of health, education, and social protection systems had to postpone important medical treatment. By to protect people’s human capital from the long- January 2023, the COVID-19 pandemic had caused, term impact of recurrent shocks and crises. Resil- directly or indirectly, more than 3.5 million excess ience generally relates to the capacity of people deaths across the region,1 and profoundly affected or systems to face and recover from difficulties. people’s livelihoods and human capital across the Given our focus on human capital, we define it as life cycle, often leaving long-term damage. the ability to protect human capital and the liveli- hoods of poor, vulnerable, and middle-class house- Unfortunately, COVID-19 will not be the last pan- holds during shocks and crises. People’s resilience demic nor the last crisis or shock affecting human to shocks and crises is influenced by three major capital and long-term wellbeing. The pandemic was factors: (i) their personal characteristics; (ii) the envi- not even over when Russia launched the full-scale ronment in which they live; and (iii) the services that invasion on Ukraine. The impacts of the war are they receive. The characteristics and endowments devastating for Ukraine, but the war is also having of individuals and households are major factors a considerable negative effect on the region and affecting people’s ability to prepare for, adapt to, the entire developing world through, for example, and cope with shocks and crises. The environment higher food and energy prices and lost remittances. within which people live – that is, the strength of New crises will also keep emerging – economic, the private sector, civil society, and communi- political, climate-related, or another pandemic – ties – also affect their ability to cope with shocks. that will keep negatively affecting people’s well- Finally, the quality and resilience of the services being and livelihoods. If no action is taken, they will that people receive also substantially affect their cause long-term damage across generations, fore- ability to cope with shocks and crises, especially the stalling the long-term prosperity of individuals and poor and vulnerable who may not be able to rely as countries. much as others on their own coping strategies or on the communities in which they live. This report One of the casualties of crises (especially if they focuses on this last, third pillar (the delivery systems are frequent) is human capital, especially among to effectively respond to crises) given its importance the poor. Human capital can be defined as a pop- in supporting the most vulnerable. ulation’s health, skills, knowledge, experience, and habits and is essential for increasing a country’s pro- The report draws from a series of companion notes2 ductivity and for contributing to long-term growth. that review the impacts and responses of COVID-19 Building citizens’ human capital is one of the best in the health, social protection, and education sectors investments that countries can make. Unfortu- of the countries of Eastern Europe and the South nately, investing in human capital produces rewards Caucasus (Armenia, Azerbaijan, Georgia, Moldova, only in the long term and tends to be substan- and Ukraine). It draws lessons from these analyses tially weakened by crises. The frequent recurrence on how to protect people’s human capital and min- The Economist COVID-19 tracker, lower bound estimate (as of February 2023). 1 ix See Koziel et al (2023) for health, Coll-Black et al (2023) for social protection, and Chinen and De Hoyos (2023) for education. 2 imize the long-term effects of recurrent crises. The example, where cross-sectoral steering committees countries in Eastern Europe and the South Cauca- existed, they did not always include representatives sus are facing similar economic trends as the ECA of key sectors to effectively coordinate the lock- region as a whole, and the COVID-19 pandemic had down measures with economic and income support an equally massive impact on the human capital of measures or they did not have enough power to their populations. By January 2023, the pandemic counter-balance political pressures. Some coun- had caused, directly or indirectly, between 300,000 tries rarely carry out routine simulation exercises and 400,000 excess deaths in these five countries.3 to prepare for emergencies, either at the ministry/ Many elements of the response were also similar to government level or the hospital/community level. those adopted across the region. Therefore, the Communication failures and mistrust in government lessons learned from this review can provide valu- have also affected the extent to which the popula- able insights into the best ways to strengthen resil- tion was willing to abide by travel bans and get vac- ience across the region and beyond. cinated. Limited digitalization of health information systems in some countries prevented the effective monitoring of the pandemic. And the pre-existing Delving into COVID-19’s impact and structure and capacity of the healthcare system also countries’ pandemic responses had a negative effect on its ability to respond effec- tively to the pandemic. On the financial side, while Countries in Eastern Europe and the South Cauca- the costs of immediate COVID-19 treatments were sus enacted swift measures throughout the human covered in most countries, high co-payments and development sectors to contain the spread of the costly or poor access to medicines for outpatient pandemic, while at the same time helping house- treatments made it difficult for the public to seek holds to protect their human capital and livelihoods. and receive adequate care, particularly given that How these measures were designed and imple- so much hospital care for non-COVID conditions mented and what we can learn from them to better were being deferred. Despite countries’ efforts to prepare for future crises is the main focus of the hire more healthcare staff, staff shortages remained report. widespread, especially in rural areas, which already had a deficit of healthcare professionals, and poli- Health. The countries of the region all adopted cymakers may have paid too little attention to the similar approaches to limiting the spread and the likely long-term impacts of forgone care. impacts of COVID-19. The measures that they adopted included: lockdowns and bans on travel, Social protection. All countries in the region passed in particular internationally; the establishment of emergency spending measures between March and intersectoral steering groups to facilitate taking the May 2020 and initiated emergency programs to right actions at the right time; the optimization of reach those who had lost income during the pan- hospital care (for example, by deferring some treat- demic. Depending on the country, these included ments) and facilities (for example, by expanding the increasing the size of social assistance benefits, supply where needed); the acquisition and local easing eligibility and/or documentation require- production of medical equipment and pharma- ments to access benefits, providing assistance to ceuticals, including vaccines; financial measures to new categories of beneficiaries, channeling funds alleviate the cost of COVID-related care; the expan- to the unemployed, subsidizing wages to retain sion of the healthcare workforce by training people, jobs, and increasing the benefits provided by hiring students and retirees, and facilitating the various social insurance programs. All countries in immigration of healthcare personnel; the provision the region expanded last-resort income support of financial compensation for healthcare facilities programs by increasing the level of benefits, allow- and personnel; the expansion of testing and contact ing automatic extensions for existing participants, tracing capacity; the use of digital resources and or widening eligibility criteria and accelerating the health information systems to monitor the spread enrollment process for new beneficiaries. New pro- of the pandemic; and the use of communication grams were created to provide one-time payments campaigns to effectively inform the population of to low-income individuals, disability beneficiaries, or the importance of abiding by public health meas- pregnant women and students. Other ways in which ures and of being vaccinated. countries tried to support incomes included efforts to control the prices of key foods and subsidies to While on the surface countries’ responses appear cover households’ utility bills, food purchases, and to be similar, their effectiveness depended very tuition. Most countries also expanded support to much on pre-existing conditions and implementa- unemployed workers, and some did the same for tion challenges that varied between countries. For pensioners. To retain formal sector jobs, some coun- x The Economist COVID-19 tracker (lower bound estimate, as of February 2023). 3 PROTECTING HUMAN CAPITAL through Shocks and Crises tries made direct payments to firms, gave them tax between students and teachers while lessons and exemptions, or covered the payment of social secu- instructional content are being shared) and asyn- rity premiums for their workers. Some countries also chronous instruction (students learning at their own attempted to create jobs, for example, by expand- pace with or without interacting with the teacher). ing public works programs. To assist poor and vulnerable students with online learning, the countries also offered subsidized or As with the health sector, the breadth and prompt- free internet access, provided students with text- ness of countries’ response mask implementation books in electronic format free of charge, or made and design challenges that may have hampered its access to distance learning platforms available effectiveness. Some countries had to established through mobile phones. A few countries also sub- new, ad hoc institutional structures to manage sidized or provided devices for free, while others their overall COVID-19 response, including social introduced online support measures such as flexi- protection measures. Most of the social protec- ble and self-paced platforms (asynchronous learn- tion responses only covered the initial months of ing) and psychosocial and mental health support to the pandemic and provided no long-term protec- learners (such as online counselling). tion. Because of a lack of planning, the allocation of spending among programs did not necessar- While distance learning provided students with ily reflect countries’ priorities or economic struc- essential learning opportunities during school clo- tures. For example, many countries channeled sures, it has been a poor substitute for face-to-face most emergency funding to formal sector firms teaching, especially for poor and vulnerable children or workers, leaving them with limited resources to who may not have had the necessary supportive protect the poor and vulnerable throughout the learning environment at home. Implementation pandemic. Similarly, a lack of preparedness and lim- challenges may also have hampered further its itations within existing delivery systems meant that effectiveness. Common challenges faced by teach- the emergency social assistance response in many ers and principals included poor connectivity and/ countries focused on providing more aid to existing or a lack of equipment, difficulties in adapting the beneficiaries rather than on reaching those made curriculum to remote education, and an increased poor by the pandemic. Administrative constraints, workload with little guidance from education min- overburdened social workers, and a lack of strong istries. On the receiving end, parents and students outreach also contributed to the limited expansion faced similar challenges, including poor or no con- of social assistance programs – even in cases where nectivity at home, limited student interest and/or the program’s design was modified with the aim fatigue with online learning, and deficient or incon- of expanding coverage. Finally, the fact that some sistent guidance about remote education models important aspects of service delivery in some coun- from education ministries. tries – such as targeting, profiling and enrollment – were not digitized also hampered the effectiveness However, the most important failures that led to and promptness of the response. massive learning losses were not necessarily in the design and implementation of distance learning Education. The pandemic’s main impact in the modalities but in the excessive time during which education sector was the closure of schools and the schools were closed in many countries and the need to adopt distance learning.4 Between March absence of strong post-pandemic measures needed 1, 2020, and October 31, 2021, the number of days to close learning gaps. While school closures, at the when schools were fully closed in the region (affect- onset of the pandemic, may have been an effec- ing at least 80 percent of the student population) tive and inevitable public health measure, schools ranged from 64 to 205.5 These school closures have at times remained closed despite restrictions being led to substantial learning losses, which, for some lifted earlier for main economic activities (such as students, are equivalent to a full year of schooling office-based work). Countries also failed to experi- or even more. To provide some continuity to the ment with potentially promising approaches such as learning process while schools were closed, gov- introducing differentiated return-to-school policies ernments adopted remote learning policies. These by academic level or allowing schools to open earlier were based on a combination of online platforms in rural areas where the student density in schools is and media broadcasts in the form of television edu- significantly lower than in urban areas. Furthermore, cation programs. The online platforms included a since schools reopened, countries have not invested variety of educational resources for teachers and enough in helping students catch up on what they students (such as learning materials and homework missed, leaving many poor and vulnerable children resources). The countries of the region also used a facing lifetime damage to their human capital and combination of synchronous (real-time interactions labor market opportunities. 4 The analysis in this report focuses mostly on school closures in basic education, missing higher education and also early childhood development xi (ECD). The pandemic and related school closures clearly affected all education levels; nevertheless, the lack of systematic learning assessments and, for ECD, the decentralized nature of the systems make the analysis much more challenging, especially on a regional basis. Many of the lessons learned apply however across education levels. 5 UNESCO Institute for Statistics based on UNESCO map on school closures. https://en.unesco.org/covid19/educationresponse Learning from COVID-19: building in need of support (for example, social registries or better and more resilient human effective on-demand programs) are able to iden- tify and enroll new beneficiaries more rapidly when development systems crises hit. And education systems that accompanied The response to the COVID-19 pandemic in Eastern their decentralization and school autonomy reforms Europe and the South Caucasus has yielded some with both financial and capacity-building sup- useful lessons for increasing the resilience of human port measures are better equipped to tailor crisis development systems and for protecting human responses to local needs. Resilience and long-term capital from ongoing and future crises. development to a large extent go hand in hand; it is not necessary to sacrifice the long-term process of Prepare, prepare, prepare investing in human capital development in order to build more resilient systems. As the response framework above shows, human development systems should not only be strong but …and continue responding after the crisis is over should also prepare for crises. This simple mantra to close long-term, crisis-induced human capital seems obvious, but many countries in the region gaps. By many measures the worst of the pan- do not follow it. Most of them have disaster risk demic is over, and new crises are already hitting the management policies in place, but these often are region. Therefore, it would be tempting to move on not tailored to economic crises and tend to neglect to addressing other emerging priorities. However, the human development sectors. However, good that would be a mistake. Crises and shocks leave preparation is essential for an effective response to profound scars on people’s human capital. Effec- crises. Countries should have institutions to coor- tive crisis responses must support people not only dinate emergency responses already in place. The in the immediate aftermath of a crisis but also in the public health sector should have effective monitor- medium to long term to help them to make up for ing systems and legislation/financing that can be crisis-induced human capital losses. quickly triggered in the event of a pandemic, and hospitals should conduct regular emergency drills. Explore the potential of digitalization Social protection programs should have up-to- while building back better date and accurate social registries and monitoring and evaluation (M&E) systems that can facilitate The crisis has created opportunities to “build back the identification of affected populations during better” and both to improve the quality of service crises. Legislation should also be passed to enable delivery in normal times and increase its resilience to the immediate scaling up of programs in response future crises. An emerging, cross-sectoral theme is to crises, with potential changes to both eligibility the potential of digital technologies for improving and benefit rules set out in advance. As for educa- delivery and monitoring. Health information systems, tion, distance and other learning modalities should social registries, integrated monitoring and evalu- be piloted in normal times, ensuring that teachers, ation systems, and payment and learning/training principals, students, and parents all receive proper platforms can all significantly improve the quality of training to increase the effectiveness of the new service delivery – and at the same time increase the modalities and minimize learning losses during crises. crisis-responsiveness of delivery systems. Respond with an eye to the future...… Protect financing When a series of crises hits countries, they may find This is not the moment to reduce human develop- it easiest to focus their attention on short-term ment investments. A few exceptions notwithstand- needs. Yet it is important not to lose sight of the ing, spending on the human development sectors medium and long-term need to strengthen human by countries in Eastern Europe and the South Cau- development systems. Over and over, evidence casus was low to begin with, and the COVID-19 pan- has shown that systems that can deliver quality demic has put further strain on public finances and services in normal times tend to respond better to fiscal space. However, when crises are expected to crises. Effective public health systems are better recur, it is vital to protect as well as increase fragile equipped to monitor the spread of a pandemic and human capital endowments to support greater pro- adopt swift precautionary measures, and a health ductivity, growth, and prosperity. Ideally, countries workforce that is properly trained, managed, and should invest more in human capital given the need remunerated needs less additional support during to both strengthen systems and close crisis-gener- stressful times. Social protection systems that have ated gaps. At the bare minimum, they should ensure invested in administrative tools to identify people that spending does not fall further. xii PROTECTING HUMAN CAPITAL through Shocks and Crises Improve governance arrangements Boost data, information, and and the design of responses evidence-based policymaking to support effective responses The literature on disaster risk management puts considerable emphasis on implementation issues. In the immediate aftermath of a crisis, there is little However, poorly designed responses – even if deliv- time for thinking and evaluating. However, moni- ered effectively – can substantially hamper their toring systems should be already in place to gather impact at the ground level. In social protection, the the information needed by policymakers to make amount that counties invested in specific programs evidence-based decisions about ways to respond. and the duration of those investments did not Moreover, many crises, especially man-made ones, always reflect the vulnerability of different popula- can last for a long time (COVID-19 itself affected the tion groups or the amount and length of support social sectors and economic activity for two years), that they needed. Design flaws also reduced the meaning that there is time for countries to improve effectiveness of specific programs. Some countries, their crisis responses based on emerging evidence. for instance, chose to increase their safety net sup- Evidence-based policymaking should be the norm port to existing beneficiaries but did not expand it in dealing with crises. However, despite the length to many new beneficiaries. The education sectors in of the COVID-19 pandemic, countries’ responses the region quickly adopted distance learning as a did not change much over time. In the education necessary first response; over time, but many failed sector, little effort was made to assess the extent to experiment with different approaches that could of learning losses, for example, by administering have reduced learning gaps, such as varying school standardized tests and learning assessments. In opening times for different grades or spacing stu- the social protection sector, after initial tweaks to dents more widely apart in classrooms. In the health respond to the emergency, the design of response sector, countries implemented massive vaccination programs changed little over time. And while the campaigns, but vaccine hesitancy among the pop- health sector has been more responsive to emerg- ulation hindered their success. ing international evidence and has adapted treat- ments accordingly, the duration and rules of public Develop cross-sectoral, household- health measures such as social distancing were set centered governance approaches as much in response to political considerations as to the available evidence. Ultimately, the goal of disaster responses is to help households to cope with crises and shocks. Put- ting the household at the center requires a solid intersectoral governance structure from the outset because decisions in one sector affect how house- holds react in other sectors. For instance, lockdowns required not only social protection measures to help households to cope with lost incomes but also measures to support parents, children, and teachers with distance learning and remote working. Taking an effective household-centered approach requires coordination across a wide range of ministries, a good cross-sectoral referral or case management system, regular monitoring of implementation and outcomes, and, given the fiscal and capacity con- straints faced by most countries in the region, the careful consideration of spending and implementa- tion tradeoffs across sectors and programs to max- imize the positive impact of responses. xiii xiv PROTECTING HUMAN CAPITAL through Shocks and Crises 1. Introduction Risk and uncertainty are on the rise, and coun- 18,000 civilian casualties,8 while close to 8 million tries across Europe and Central Asia (ECA) are not Ukrainian refugees were recorded across Europe.9 immune from it. The region is being hit by crises, The impacts of the war also reverberated across conflicts, and continued uncertainty that are nega- the region through, for example, higher food and tively affecting people’s livelihoods in the short term energy prices and lost remittances. Moreover, the and prosperity in the long term. The 2008 finan- poor and vulnerable are bearing most of the suf- cial crisis was followed by instability and a series of fering because, in proportional terms, they spend regional conflicts that significantly reduced growth a much higher share of their income on food and and prosperity. While real per capita GDP in the energy than the rest of the population.10 New crises region grew by 2.7 percent per annuum between will also keep emerging – economic, political, or cli- 1995 and 2008 (and by 4.6 percent in low- and mate-related or, yet again, a new pandemic11 – that , middle-income countries), growth reduced sub- if no action is taken, will do further damage to peo- stantially after 2008 to average a mere 0.5 percent ple’s wellbeing and livelihoods across generations, between 2008 and 2020 (or 1.6 percent in low- and hindering the long-term prosperity of individuals middle-income countries).6 and countries. Then COVID-19 hit, inflicting massive harm on One of the casualties of crises (especially if they are people’s wellbeing, livelihoods, and human capi- frequent) is human capital – the health, skills, knowl- tal. Lockdowns prevented people from working, edge, experience, and habits of the population.12 school closures prevented students from learning, Undeniably, people’s education, health, resilience, and overwhelmed hospitals had to defer impor- and access to opportunities have intrinsic value and tant treatments. By January 2023, the COVID-19 can be considered to be human rights, but, from pandemic had caused more than 3.5 million excess an economic perspective, they are equally essen- deaths across the region, either directly or indi- tial for increasing productivity and contributing to rectly.7 The pandemic has also inflicted long-term long-term growth.13 In consequence, building citi- damage on people’s livelihoods and human capital zens’ human capital is one of the best investments over their lifecycles. While the worst may be over, that countries can make. Unfortunately, investing millions of people are still coping with the long- in human capital produces rewards only in the long term impacts of employment loss. Students, espe- term and tends to be substantially weakened by cially those from poor and vulnerable backgrounds, crises.14 The COVID-19 pandemic, for instance, led may never be able to fully recover from the learning to massive learning losses, and while some losses losses caused by school closures. The dramatic rise may have been inevitable, too little is now being in mental health issues among young people may done to help students to catch up. have adversely affected their school-to-work tran- sition and professional careers. Also, the deferral of The frequent recurrence of shocks and crises means medical treatments may have long-term negative that it is necessary not only to invest in human capital health effects, especially in countries with a high but also to protect human capital from the immedi- prevalence of non-communicable diseases (NCDs). ate and long-term effects of these shocks and crises. Failing to protect human capital gradually erodes Unfortunately, COVID-19 will not be the last pan- people’s health and skills and the ability of firms to demic nor the last shock affecting human capital climb the productivity ladder and, ultimately, limits and long-term wellbeing. The pandemic was not long-term prosperity. What is equally worrisome is even over when Russia launched the full-scale inva- that shocks such as the COVID-19 pandemic are sion of Ukraine. The impacts of the war are devas- profoundly regressive and thus exacerbate pre- tating for Ukraine, but the war is also substantially existing inequalities in an already polarized world. affecting the whole region and the entire develop- ing world. As of January 2023, in addition to mas- This report explores how to strengthen the resil- sive destruction across the territory, the war caused ience of health, education, and social protection 6 World Development Indicators, per capita GDP in 2017 PPP. 10 Alberini and Umapathi (2021). 1 7 The Economist COVID-19 tracker, lower bound estimate (February 11 MSC Special Edition (2021). 2023). 12 Becker (1964). 8 OHCHR (2023). 13 Flabbi and Gatti (2018). 9 UNHCR, Operational data portal: Ukraine Refugee Situation, 24 January 14 See, for instance, Schady et al (2023). 2023 update. systems to better protect people’s human capital sons learned from this review can provide valuable from the long-term effects of recurrent shocks and insights into the best ways to strengthen resilience crises. The report draws from a series of companion across the region and beyond. notes15 that review the impacts of and responses to COVID-19 in the health, social protection, and edu- The report is structured as follows. Section 2 dis- cation sectors of countries from Eastern Europe and cusses the concept of resilience for the purpose the South Caucasus (for the purposes of this report, of protecting human capital. Section 3 reviews the this comprises Armenia, Azerbaijan, Georgia, Mol- impact that the COVID-19 pandemic had on the dova and Ukraine). It draws lessons from this anal- economic, social, and human capital in the countries ysis on how to protect people’s human capital and of Eastern Europe and the South Caucasus. Section minimize the long-term effects of recurrent crises. 4 discusses the ways in which countries responded The countries in Eastern Europe and the South Cau- to these challenges and Section 5 draws lessons casus are facing similar economic trends as the ECA from the analysis on how countries can best pre- region as a whole, and the COVID-19 pandemic had pare for and respond to future crises. Because most an equally massive impact on the human capital of of the information used in the report is drawn from their populations. By January 2023, the pandemic the companion notes, which themselves draw from had caused, directly or indirectly, between 300,000 focus group interviews, background work by local and 400,000 excess deaths in these five countries.16 experts, and local reports, the reader should assume Many elements of the response were also similar to that, unless specified otherwise, the information those adopted across the region. Therefore, the les- presented is taken from the notes. 2 15 See Koziel et al (2023) for health, Coll-Black et al (2023) for social protection, and Chinen and De Hoyos (2023) for education. 16 The Economist COVID-19 tracker, lower bound estimate (February 2023). PROTECTING HUMAN CAPITAL through Shocks and Crises 2. A Brief Primer on Resilience Resilience is defined as the capacity of people or The environment in which people live – that is, the systems to face and recover from difficulties. For strength of the private sector, civil society, and com- the purposes of this report, we will define it as the munities – also substantially affects people’s ability ability to protect human capital and the livelihoods to cope with shocks.19 If the private sector is strong, of poor, vulnerable, and middle-class households it may shed fewer jobs during a crisis or may recover during shocks and crises. The concept of resilience more quickly than if it were weaker, and civil soci- has gained traction in the policy world in recent ety organizations with a solid local presence help years, and a growing body of literature focuses addressing the immediate needs of the population on resilience in human development.17 While many in the aftermath of a shock. When communities frameworks have been developed, those covering are strong, they may also be more able to organ- human development tend to share a few common ize a response than weaker ones and to ensure that features. Most acknowledge, for instance, that peo- their voices are heard by those leading the wider ple’s resilience to shocks and crises is influenced by response. For example, a society that values schools three major factors: (i) their personal characteristics; and education would be likely to prioritize opening (ii) the environment in which they live; and (iii) the schools sooner than others. services that they receive (Figure 1). Finally, the quality and resilience of the services that The characteristics and endowments of individuals people receive (that is, the quality and resilience and households are major factors that affect their of delivery systems) are crucial factors in helping ability to prepare for, adapt to, and cope with shocks households to cope with shocks and crises, espe- and crises. Living in poverty and having limited sav- cially for the poor and vulnerable who may not be ings, for instance, substantially hinders households’ able to rely on their own coping strategies or on ability to manage shocks effectively by significantly the communities in which they live. Effective social reducing their incomes in the short term and pre- protection systems, for instance, can provide much- venting them from building human capital, which, needed income support – if they are able to respond in turn, reduces their long-term wellbeing.18 Other and expand fast enough. Health systems should factors that affect their ability to protect their liveli- be able to respond to health emergencies, while at hoods and human capital when shocks hit are where the same time minimizing the deferral of necessary they live, who they can rely on, and the type of jobs medical care that is not related to the crisis. Also, that they have. the more that education systems can continue to Figure 1: Factors Influencing People’s Resilience Individual & RESILIENCE OF PEOPLE households characteristics Resilience of Savings services and The ability to protect human capital and the livelihoods of Education systems poor, vulnerable and middle-class households during crises Network Type of Job Type of dwelling Preparedness Civil society, private sector & … communities’ characteristics NGOs, Social capital, Wealth, Ability to organize ... 17 See, among others, National Academy of Sciences (2012), World Bank (2013), World Bank (2016), Kruk et al (2017); Linkov et al (2019), Bowen et al 3 (2020), and Thomas et al (2020). 18 Dercon (2002) and Vakis et al (2016). 19 World Bank (2013). provide effective teaching during crises, the lower The next stage, when a shock or crisis hits, is the will be the learning losses and long-term scars that response phase. While the quality of the response crises may leave upon children. depends substantially on the level of preparedness, no country is ever fully prepared for a shock, particu- Notwithstanding the fundamental importance of larly those that are unexpected such as the COVID- all three of these pillars in promoting resilience, this 19 pandemic. Shocks and crises always bring unex- note will focus on the third – the resilience of deliv- pected issues that will have to be dealt on the spot, ery systems – for two important reasons: (i) it is dis- and the response will strain financial and human proportionately important for supporting the poor resources of any government and will require strong and vulnerable during crises; and (ii) it is under the coordination between sectors and effective leader- direct control of governments and policymakers. ship to set priorities. It is also important to acknowl- edge that the response phase, especially in human Resilient delivery systems respond effectively to development, may continue well beyond the end of shocks and crises at all stages of the response a crisis, given the imperative to close human capital chain: preparing, responding, and adapting and gaps that may have been caused by the crisis – an building back better (see Figure 2). A lot of the work endeavor that can take several years. should be done in advance of any shock. Policymak- ers can prepare systems to face shocks by drawing Finally, in order to be resilient, a system also has from international best practices, lessons learned to have the capacity to both adapt and build back from the impacts of and responses to past crises, better. Because each crisis yields valuable lessons and forward-looking analyses. For instance, social on how to respond better to future crises, it is protection policymakers should design and imple- important to analyze and learn lessons from each ment methods to rapidly identify and financially response and reform delivery systems accordingly. support victims of disasters and crises and also At the same time, analysts should update existing enact the legislation needed to ensure that pro- forward-looking analysis to be used in contingency grams can be rapidly expanded and pre-position planning or to forecast future needs. The recovery budgets to quickly fund such a response.20 Public and reconstruction phase also creates an opportu- health policymakers should ensure that systems are nity to improve the quality of delivery systems and able to closely monitor the emergence and evolu- strengthen long-term human capital outcomes by tion of pandemics and that healthcare systems can incorporating lessons learned during the crisis. For respond to temporary surges in demand. Education example, some social protection innovations devel- policymakers should prepare the sector for the need oped during the crisis may also increase the effec- to teach and learn under various crisis scenarios and tiveness of support during ordinary times. Modal- – equally importantly – should design programs ities developed to continue teaching and learning that will ensure that no students drop out because during crises (such as blended learning models) can of a crisis and that in the aftermath of the crisis chil- also be adapted and integrated to improve teach- dren can recover their learning losses. ing practices and learning outcomes at other times. Figure 2: Resilient and Non-resilient Human Development Systems Disruption Disruption RESILIENT SYSTEMS Human Capital Adapt & Minimize Build Back Better NON RESILIENT SYSTEMS Prepare Respond Adapt & Build Back Better Source: Adapted from National Academy of Sciences (2012), Linkov et al (2019), and Bowen et al. (2020). 4 Bowen et al (2020). 20 PROTECTING HUMAN CAPITAL through Shocks and Crises Ultimately, as seen in Figure 2, building resilient gency procedures need to be in place that involve delivery systems by preparing, responding, adapting pulling staff from other areas, making back-up staff- and building back better contributes positively to ing arrangements, or calling on a network of well- long-term human capital formation and prosperity. trained volunteers. However, if these additional staff Human capital formation (and, in fact, development) are poorly trained, they may not be able to make has never been a linear process: crises and shocks – effective use of response technologies and may lack whether natural or man-made – will always affect the training or the confidence required to impro- us, and if crises are poorly managed, human capital vise in the face of the countless unexpected events may even deteriorate either throughout the popula- that always arise during emergencies. Therefore, it is tion or among particularly vulnerable groups. There- important to prepare staff for crises by conducting fore, it is important to consider resilience not as an regular drills and contingency planning exercises to add-on feature of delivery systems but as an essen- help them to become familiar with emergency pro- tial element for promoting long-term prosperity. cedures. The “Prepare, Respond, and Build Back Better” A third pillar that has recently been emphasized in framework in Figure 2 is useful for understanding the literature is the importance of data and infor- the evolution of effective and less-effective shock mation. Data and information help boost the effec- responses, but it does not address how to achieve tiveness of service delivery in normal times, but effective responses. The literature typically identifies regular monitoring and evaluation of implementa- three fundamental pillars of any effective response,21 tion and outcomes during crises can also continu- to which we feel the need to add a fourth (Figure ously increase the effectiveness of the preparation, 3). The first pillar is, of course, financing. The effec- response, and recovery phases. In education, having tiveness of any country’s crisis responses hinges on a standardized assessment system is an important its ability to finance the rapid implementation or way to identify students who are lagging behind expansion of emergency and response programs. their peers and to assess the nature and size of the Various financial tools can be used for emergency learning gap. In social protection, the existence of financing, including reserve funds, the insurance social registries, which contain relevant information of public assets, catastrophe (cat) bonds, or con- about households, can help to speed up enrollment tingent credit lines.22 Ensuring financing for crisis in social assistance programs and verification of responses also requires the enactment of legislation eligibility during a crisis – provided that these data that allows for the extraordinary financing of emer- are kept up to date. Similarly, effective public health gency interventions. Mexico’s FONDEN is an exam- information systems can track the spread of a pan- ple of a contingency fund that has been established demic to inform the response. During the delivery to respond to regular natural disasters.23 of the response, effective monitoring can be used to rapidly identify any bottlenecks (such as lagging A second fundamental pillar that is at the center of or poorly targeted enrollments or shortages of staff most resilience frameworks is delivery. Adequate or materials) that need to be addressed. In a more software, equipment, infrastructure, and human general sense, reliable information is also essential resources must be in place to enable programs and for communicating with affected populations about sectors to respond promptly and effectively. For the response. For example, vaccination campaigns example, in the social protection sector, online regis- across the region have been hampered by a lack of tration and digital payment systems facilitate faster trust (among many other factors), which highlights and more effective responses by making it possi- the importance of effective – and transparent – ble to quickly identify who has been affected and communication. to compensate them. In the education sector, reli- able internet access and a protocol for conducting A fourth important but often neglected pillar is the remote education are the backbone of any effective governance and design of the institutions respon- learning program while schools are closed. In the sible for coordinating shock responses as well as of health sector, contingency planning for ensuring the response programs themselves. Over and over, adequate equipment and pharmaceuticals is a key assessments of shock responses have mentioned pillar of effective and prompt responses. The human the need to coordinate cross-sectoral responses, dimension is as important – if not more important – but not all countries have established coordina- for guaranteeing effective responses. When staffing tion committees, and lines of accountability and is inadequate, personnel rapidly become overbur- decision-making remain unclear, which reduces the dened, which is a problem that not even sophisti- speed and effectiveness of the response. Moreover, cated technologies can solve. Also, surge capacity institutions are only one element of proper design. is often needed in crises, which means that emer- Effective communication strategies and protocols 21 Bowen et al (2020). 5 22 World Bank and ADB (2017). 23 World Bank (2012). Figure 3: The Four Pillars of An Effective Shock and Crisis Response Financing Ability to finance rapid expansion of emergency and response programs Governance and Design PREPARE Clear Governance structure in place within and across sectors to respond to crises Data and Information RESPOND Program and response designs adequately cover all Ability to identify rapidly magnitude and impacts of ADAPT population groups shocks and crises and people a ected & BUILD BACK Balance of programs across territory and population Ability to monitor and evaluate quality of response BETTER groups Presence of feedback loops from monitoring & evaluation of responses to design and implementation changes Delivery Adequate tools (software, equipment and infrastructure) to respond to shocks and crises Adequate sta ng with trained personnel Performing of regular drills and contingency planning exercises Source: adapted from Bowen et al (2020). that distribute responsibilities and legal attributes some dimensions but still face challenges in others. to different decision-making levels (such as national Therefore, we highlight selected examples of resil- versus subnational governments) are also essen- ient responses as opposed to whole systems. tial. Also, different population groups and parts of the country have different needs, and thus, to be Response elements can be developed ahead of effective, interventions should be designed to take crises, but sometimes they are also developed as into account idiosyncratic vulnerabilities. For exam- the crisis evolves and learning occurs. Italy and the ple, in social protection, not enough emphasis is put UK were two countries that were most affected by on how to balance support between, say, formal the first wave of the COVID-19 pandemic in West- and informal workers or on the generosity of bene- ern Europe. However, by using evidence generated fits versus the duration of support or the extent to from their initial responses and from the interna- which, within a given budget, more support should tional experience of dealing with COVID-19, both be provided to existing beneficiaries of social assis- countries took important steps that helped to tance (i.e. a vertical expansion) or to people who fell control the spread of the virus during the subse- into poverty as a consequence of the crisis (i.e. a quent waves. Response elements included ensur- horizontal expansion). Similarly, not enough thought ing effective political leadership and cross-party has been given to how to gradually reopen schools consensus: Decisions such as those imposing lock- based on growing evidence. downs or mobility restrictions were usually criticized or opposed by different political parties. In the UK, To illustrate this four-pillared framework (see Figure the creation of the All-Party Group on Coronavirus 3), we now discuss examples of effective responses facilitated agreement on the pandemic’s response; to the COVID-19 pandemic. We then explore how Strengthening monitoring systems: Both countries the pandemic affected human capital in the region have developed robust monitoring systems that and how financing, delivery, data, and governance produce weekly reports on the number of cases and design all substantially influenced the response. and deaths due to COVID-19; Establishing scientific We conclude by discussing how, based on this anal- advisory groups to favor evidence-based decisions: ysis, the ability of countries in the region to respond The UK used its previously existing Scientific Advi- to future crises could be strengthened. sory Group for Emergencies (SAGE) to evaluate the emerging evidence and disseminate its findings to What Do Resilient Systems Look Like? policymakers and the public, while Italy established a new expert advisory group to do the same; Ensur- The most successful systems have financing, deliv- ing sufficient and stable funds: The mobilization of ery, data and information, and governance and funding from both unearmarked and earmarked design elements that facilitate rapid and effective resources helped the UK government to ensure responses to shocks and crises. However, it can adequate funding for the COVID-19 response; be difficult to identify whole systems that can be Implementing innovative approaches to health showcased as best practices. As the pandemic has workforce management: Innovations in skill-mixing shown, any crisis involves a substantial amount of enabled the UK and Italy to increase their capacity uncertainty, and systems need to learn and adapt. to conduct public health interventions. For instance, Moreover, some systems can respond effectively in dentists with sedation skills supported the National 6 PROTECTING HUMAN CAPITAL through Shocks and Crises Health Service (NHS) during COVID-19 surges in to help with processing claims. Despite the huge the UK, and members of the public were trained to increase in claims during the early months of the administer vaccines or otherwise assist in admin- COVID-19 crisis, claims were processed more quickly istering vaccination programs. In Italy, volunteers than before, with 93 to 95 percent of claimants were recruited to support the delivery of medication receiving their first payments on time as opposed to and food to vulnerable groups and those who were the pre-pandemic figure of 85 percent.27 Australia’s self-isolating; and Developing efficient vaccination existing governance arrangements also helped with programs: By mid-2021, the European Union (EU) its pandemic response. It already had a single deliv- had already secured 4.65 billion doses to cover an ery structure, Service Australia, in place to enroll EU population of 446 million people. Although the beneficiaries and deliver benefits for its social insur- program's roll-out was slow, by the end of July 2021, ance and social assistance interventions, includ- over 70 percent of adults had already been vacci- ing for those schemes that were handled by other nated against COVID-19. In the UK, the vaccina- departments, which strengthened the coordinated tion roll-out started on December 8, 2020, and by delivery of benefits.28 August 2021, over 75 percent of adults had already received two COVID-19 vaccine doses.24 Finally, effective responses in the education sector have demonstrated the importance of collecting Analyses of social protection responses have also data to facilitate evidence-based policymaking. In shown how strengthening delivery systems by Denmark, schools were closed for more than half a introducing crisis-sensitive elements has helped school year, but students did not experience a sub- to increase the effectiveness of these responses. stantial loss of learning. The reason why Denmark Global experience has shown that procedures for was better prepared than other countries to handle expanding access to crisis-affected populations the educational disruption was because it had relia- need to be set out in advance and to be supported ble digital infrastructure and broadband connectiv- by the required implementation capacity, infor- ity, high levels of digital skills among teachers and mation systems, governance arrangements, and students, and online learning support platforms pre-positioned financing as discussed above. Here, that were available to schools. Denmark also moni- we focus on the process by which new people are tors the acquisition of foundational skills by holding identified or apply for benefits, which is crucial for regular standardized tests that can identify any stu- expanding social protection support in a crisis. For dents who are lagging behind their peers. To close instance, Ukraine’s pandemic response was facili- the pandemic learning gap, Denmark financed the tated by the use of the digital Diya platform, which provision of compensatory interventions such as facilitated the use of online applications and case additional teaching and remediation for lagging management. By the end of 2021, 12 million indi- students, which improved students’ well-being and viduals, or around a quarter of Ukraine’s population, reduced inequalities among children from different had used Diya.25 In Chile, the government was able socioeconomic backgrounds.29 to harness its pre-existing Household Social Regis- try that connects several administrative databases In France, keeping schools open has been a policy to rapidly identify, target, enroll, and support 14 mil- priority. French schools were closed for only 10 lion Chileans (73 percent of the population) in only a weeks, one of the shortest periods in Europe. France few months. The first payment of Emergency Family also had a strong national assessment system that Income (IFE) to 1.3 million households was made on measured student learning during the pandemic May 23, 2020, just two weeks after the first manda- across multiple grades and cohorts. This gener- tory lockdown in the capital city, Santiago.26 In the ated robust data on students’ academic progress United Kingdom, the Universal Credit is targeted to before and during the pandemic. The country also low-income households or people out of work and collected demographic data (on, for example, stu- not in receipt of a pension. From its introduction, it dents’ socioeconomic background, ethnicity, and has been managed and accessed almost entirely immigrant status) that made it possible to conduct digitally, with claimants applying online. This was performance gap analyses that also used informa- an advantage during the pandemic as applications tion from France’s extensive censuses or surveys on could proceed with no need for face-to-face con- topics ranging from mental health, attitudes, and tact. Given the increase in claims, all staff at local home and life situations during lockdown. It then job centers were repurposed as claims processing provided support for lagging students in aspects teams to support the processing of new claims, of learning that were not able to be taught during while more than 8,500 staff from the Department lockdown. France also cut class sizes in half for stu- of Work and Pensions (DWP) and 1,000 staff from dents in priority schools, which helped bring learn- other government departments were redeployed ing back up to pre-pandemic levels.30 24 Sagan et al (2021) and UK Government press release (August 2021). 7 25 World Bank (2022a). 26 Gonzalez (2022). 27 Mackley (2021) and Ross and Clarke (2021). 28 Sandford (2021). 29 OECD (2020a), Reimer et al (2021), Chinen and de Hoyos (2023). 30 OECD (2020b), Thorn and Vincent-Lancrin (2021), and Chinen and de Hoyos (2023). 8 PROTECTING HUMAN CAPITAL through Shocks and Crises 3. Human Capital in the Time of COVID-19 Loss of human life is a tragedy but is definitely not The COVID-19 crisis has profoundly affected eco- the only scar that the pandemic has left on people. nomic activity, labor markets, and human capital It has also had a profoundly negative impact on live- across the Eastern Europe and South Caucasus lihoods and on the accumulation of human capital region, with average GDP dropping by more than 5 in both the short and long term. While hopefully the percentage points in 2020 (Figure 4). Remittances, worst is over, millions of people are still coping with which are a substantial source of income for many the long-term impacts of employment loss, while families (in 2021 they amounted to 8.4 percent of students, especially from poor and vulnerable back- GDP in the region),32 also took a serious hit, drop- grounds, may never be able to fully recover from ping by up to 40 percent in selected South Cau- the learning losses caused by school closures. The casus countries33 (Figure 5). According to balance of dramatic rise in mental health issues among young payments data, remittances dropped by an average people31 may have adversely affected their school- of 18 percent and 16 percent relative to the 2019 to-work transitions and professional careers. Also, average in the first two quarters of 2020 respec- the deferral of medical treatment may have long- tively. However, once lockdown measures were term negative health effects, especially in countries relaxed, growth and also remittances rebounded – with a high prevalence of non-communicable dis- although this was faster in some countries (Georgia eases (NCDs). Women – who were often required and Moldova) than in others (Azerbaijan). Prior to to care for children and the elderly as schools and the war in Ukraine, experts had estimated that the other services closed – and the poor and vulnerable region’s GDP would have returned to its pre-crisis were disproportionately affected by the pandemic level by 2022, with growth of 3.6 percent and 2.9 as they had fewer coping options. percent in 2021 and 2022 respectively.34 Figure 4: GDP growth by quarter, 2019-2021 Figure 5: Growth in Remittances by quarter, 2019- 2021 140 200 180 130 (Indexed avg. 2019 = 100) (% of previous quarter) 160 Evolution of GDP 120 140 Remittances 110 120 100 100 80 90 60 80 40 19 Q1 19 Q2 19 Q3 2 Q4 20 Q1 20 Q2 20 Q3 2 Q4 21 Q1 21 2 3 19 Q1 19 Q2 19 Q3 2 Q4 20 - 1 2 Q2 20 Q3 2 Q4 21 Q1 21 Q2 3 20 Q 20 - Q Q Q 20 - 20 0 - 20 1 - 20 - 20 0 - 20 1 - 20 - 20 - 20 - 20 - - 20 - 20 - 20 - 20 0 - - 20 - 20 - 20 - 20 - 19 19 20 20 Armenia Azerbaijan Moldova Armenia Azerbaijan Moldova Georgia Ukraine Georgia Ukraine Sources: Armenia Statistics Office, State Statistical Committee Sources: Central Bank of Armenia, Central Bank of the of the Republic of Azerbaijan, Georgia Statistical Office, Republic of Azerbaijan, National Bank of Georgia, National Moldova Statistical Office, State Statistics Service of Ukraine. Bank of Moldova, National Bank of Ukraine. Note: Numbers are indexed, with 100 = 2019 average remittances. 31 CDC (2022). 9 32 Simple country average. 33 World Bank (2021). KNOMAD Remittances Data available at: https://www.knomad.org/data/remittances 34 EBRD (2021). The crisis also had a substantial negative effect on While growth, labor markets and – in part thanks labor markets and incomes, with a corresponding to the deployment of many social protection pro- rise in poverty. Even before the pandemic, labor grams – even poverty may recover relatively quickly, markets in most countries in the region were frag- the effects of the pandemic on people’s human ile (Figure 6). The employment rate for all countries capital are likely to be felt for much longer. Progress except Azerbaijan remained below the EU average. in human capital development was already slow There were also substantial gender differences in in the region even before the pandemic hit. Apart labor markets, with gaps in labor force participa- from the high (but declining) spending levels on tion ranging from 6 to 22 percent across the region. health in Moldova and Ukraine, public spending on Pre-COVID-19, poverty rates differed substantially both health and education remained well below the between the countries of the region. As measured European Union average through 2019. Moreover, by the $5.50 a day poverty line, poverty in Azerbai- spending trends are flat, or even declining (Figure jan and Ukraine was quite close to the average of 8 and Figure 9). Although human capital outcomes new EU member states, but in other countries such started from a relatively high level, they have been as Georgia, it accounted for more than 40 percent stagnating or even declining. In all countries in the of the population (Figure 7). Most countries also region, the rate of NCDs is higher than the average struggled with low youth employment rates and rate in countries at similar levels of development. high proportions of youths who were not in employ- In education, learning outcomes in the region are ment, education, or training (NEET). better than those of their development peers, but they have been stagnating and, for the most part, While social protection measures helped to contain are well below the EU average. some of the poverty and labor market impacts of the crisis, unemployment rates still increased by an The pandemic added therefore pressure on sys- average of 1 percent between 2019 and 2020, and tems that were already struggling to promote more labor force participation decreased in all countries human capital investments. While all countries in in the region. Poverty reduction also took a serious the region implemented prompt health, social pro- hit. In Georgia, for instance, the poverty headcount tection, and education measures to counter the was expected to decrease in 2020, while in fact it impacts of the pandemic (these will be the focus of climbed.35 Similarly, poverty in 2020 increased by 2.4 the next section), their effectiveness varied consid- percentage points in Armenia and 1.6 percentage erably among countries and socioeconomic groups, points in Moldova.36 In Ukraine, in contrast, poverty and evidence is emerging that suggests that the was estimated to have maintained its downwards pandemic left profound scars on people’s human trend – although this trend has since been reversed capital, especially among the poor and vulnerable. by the current conflict. Figure 6: Employment Rates in Selected Countries Figure 7: Poverty and Inequality Rates in Selected Countries 70 Headcount ratio at $5.50 a day (2011 PPP) (LH) 60 Gini index (RH) 50 50 38 40 40 34 30 30 30 20 20 26 10 10 0 0 22 19 7 9 9* 8 9 EU ia ia e a n 7 EU EE 1 1 1 EE ov in ja en g 20 20 20 1 20 20 ai a or ST d ST rm kr rb ol Ge - - - - U - ne M ze ia ia va A n en g A ja do ai or ai kr m ol Ge rb U Ar M e Total 15+ Male 15+ Female 15+ Az Source: Eurostat (employment), World Development Indicators and World Bank PovCalNet (poverty). Notes: Figures are for 2019. The methodology to estimate employment differs in Ukraine and Georgia (see Eurostat). Poverty in Azerbaijan is a World Bank estimate as the latest comparable data is from 2005. STEE7 includes Bulgaria, Croatia, Estonia, Latvia, Lithuania, the Slovak Republic, and Slovenia. 10 World Bank (2021). 35 World Bank (2022c). 36 PROTECTING HUMAN CAPITAL through Shocks and Crises Figure 8: Government Expenditure on Education, Figure 9: Government Expenditure on Health, 2000-2019 2000-2019 10% 8% 8% 6% Percentage of GDP Percentage of GDP 6% 4% 4% 2% 2% 0% 0% 2000-2004 2005-2009 2010-2014 2015-2019 2000-2004 2005-2009 2010-2014 2015-2019 ARM AZE MDA ARM AZE MDA GEO UKR EU GEO UKR EU Figure 10: Harmonized Learning Outcomes in Figure 11: Non-communicable Disease Burden in Selected Countries Selected Countries 600 45 Disability-adjusted life-years (DALY) 40 550 EU average 35 (NCD DALY per 100,000) EU average 500 30 25 450 20 400 15 10 350 5 300 0 TUR ALB AZE KAZ SVN ARM CZE POL HRV MDA LTU SRB ROU GEO HUN UKR BGR KOS TJK GEO MKD AZE KAZ MNE MDA ROU ARM SRB TUR UKR HRV LVA EST Harmonized test scores Average of development peers NCD DALYs Average of Development Peers Source: World Bank (2023).World Development Indicators and OECD (2016 and 2019). Note: The average of development peers was obtained by using a kernel regression. In the health sector, COVID-19 had both a direct pre-scheduled health service appointments post- and an indirect impact on human capital. A study poned, and 5.2 percent had been denied healthcare covering 81 countries around the world found that services after the outbreak.38 Similarly, a 2022 study the initial waves of COVID-19 were responsible for of 39 low- and middle-income countries found the loss of 20.5 million years of life (as of January that almost one-fifth of the population reported 2021), which is two to nine times higher than the not being able to access healthcare when needed. average for the seasonal flu.37 The indirect impact Interestingly, the underlying reasons were not only was mainly on forgone medical care. While it is too COVID-19 but also the financial constraints that less early to assess the impacts with any precision, the affluent households were facing.39 sheer magnitude of forgone care suggests that it will also have long-term effects, especially among Unfortunately, there is not much information on people with fragile health and those who suffer forgone healthcare in Eastern Europe and the South from non-communicable diseases. In a survey of Caucasus, but focus group discussions with doctors people aged 50 and over in 25 European countries and nurses have suggested a similar picture. Doc- and Israel, 12.4 percent of respondents reported tors and nurses in Georgia, for instance, experienced having forgone medical care, 27 percent had had a massive increase in their workload, leading to 37 Pifarré i Arolas et al (2021). 11 38 Smolić et al (2021). 39 Kakietek et al (2022). crowded hospitals and staff burnout. Doctors men- Figure 12: Estimated Learning Losses in Different tioned that forgone care was a problem, especially Countries at the beginning of the pandemic, and that this 3.5 increased the likelihood of patients’ suffering from long-term complications. Some patients appear to NPL 3 have attempted to treat their own medical condi- tions at home for fear of catching COVID-19 in clin- ics and hospitals, resulting in complications. 2.5 Learning losses (years) In the education sector, the main impacts have 2 IND MWI resulted from school closures, which had a signifi- KHM ARG cant negative effect on student learning outcomes, 1.5 TURMEX KEN especially those from poor and vulnerable back- BRA BWA POL grounds in homes with limited access to online 1 CHL GHACOL BGD NOR BEL learning and parents who did not have the time FIN CHNGRC USA ZAF CHE ITA DEU GBR NLD RUS CAN or skills to support them in their learning. In addi- 0.5 HUN UGA tion, many countries with high shares of students ETH IDN CZE with poor learning outcomes prior to the pandemic ESP KOR DNK FRA JPN 0 SWEAUS closed their schools for longer than other countries. 0 20 40 60 80 100 As a result, the COVID-19 pandemic exacerbated learning inequalities both within and between School closure (weeks) countries.40 Source: Authors’ calculations based on Patrinos et al (2022). The length of school closures varied considerably Note: We conservatively considered that one year of formal schooling is approximately 0.25 standard deviations in terms across the region, from 63 to 205 school days (more of learning outcomes (Evans and Yuan, 2019). than a full school year). Taking partial closures into account, schools across the region remained closed It will take a decade or more to start understanding for between 87 and 342 school days.41 Unfortu- the impacts of all the lost learning. However, given nately, as yet there are no precise estimates of learn- the limited opportunities that many students have ing losses for the region, but the available interna- had to catch up and given that each year of addi- tional evidence points towards massive losses. In tional schooling is estimated to increase average Mexico, for instance, where schools were closed for returns by 9 percent,45 the income losses due to the 53 weeks, studies have estimated an increase in lost learning are likely to be in double digits for many learning poverty (the inability of a 10-year-old child students, especially those from poor households. to read and understand a text adequately) of 25 For example, a four-month school closure has been percentage points among students from disadvan- estimated conservatively to result in aggregated taged socioeconomic backgrounds and 15 percent- lifetime losses ranging from 43 percent of GDP in age points among students from more advantaged low-income countries, 15 percent in middle-income backgrounds.42 In Brazil, where schools were closed countries, and 6 percent in high-income countries.46 for 38 weeks, secondary dropout rates increased more than threefold, and students in remote learn- While COVID-19 was an especially large shock, it ing appear to have learned only a quarter of what highlights how countries need to prepare for and they would have learned face-to-face.43 Countries be able to respond to shocks to protect and build in Eastern Europe and the South Caucasus were human capital. The COVID-19 crisis also created a facing several challenges in implementing distance unique opportunity to implement long-awaited learning, including a lack of equipment and infra- reforms capable of overcoming structural con- structure (especially in remote areas and among the straints to human capital accumulation. Most poor), teachers who had not been trained to deliver human capital challenges already existed before the the curriculum through remote education, and the pandemic and were simply exacerbated by COVID- absence of adequate learning environments at 19. For instance, even before the pandemic, around home for many children. It is therefore very likely half of 15-year-old students in the region knew how that the magnitude of the negative effects of clos- to read but could not comprehend simple texts; or ing schools is similar to those that have been esti- knew how to solve basic math problems but could mated for other countries and regions (see Figure not use them in everyday life. Hopefully the pan- 12).44 demic created the conditions needed to design and implement the interventions required to address these challenges. 12 40 Azevedo et al (2022). affected all education levels; nevertheless, the lack of systematic learn- 41 UNESCO Institute for Statistics based on UNESCO map on school ing assessments and, for ECD, the decentralized nature of the system closures. https://en.unesco.org/covid19/educationresponse make the analysis much more challenging, especially on a regional basis. 42 Hevia et al (2022). Many of the lessons learned apply however across education levels. 43 Lichand et al (2021). 45 Psacharopoulos and Patrinos (2018). 44 The analysis in this report focuses mostly on school closures in basic 46 Psacharopoulos et al (2020). education, missing higher education and also early childhood devel- opment (ECD)The pandemic and related school closures clearly PROTECTING HUMAN CAPITAL through Shocks and Crises 4. Learning from COVID-19: Delving into Countries’ Pandemic Responses Countries in Eastern Europe and the South Cau- grated approach the to the crisis response. In Geor- casus enacted swift measures in all of the human gia, coordination seemed to have been particularly development sectors to contain the spread of the good. As mentioned by a senior health policymaker pandemic, while at the same time helping house- during focus group interviews with social protection holds to protect their human capital and livelihoods. program recipients and social workers, “coordina- How these measures were designed and imple- tion during the COVID-19 response was unprece- mented and how to draw on lessons learned to pre- dentedly good. This is an outstanding lesson to see pare for future crises are the focus of this section, how effective the government can be because of which draws on the framework set out in Figure 3. the joint and synchronous action and what tangible results the government could achieve.”47 Health Another aspect related to governance is com- munication. Effective communication is essential Countries took a similar approach to limiting the for the success of public health measures such as spread and impacts of COVID-19. Measures taken mobility restrictions and vaccination, in particular included lockdowns and bans on travel, in particular in countries where mistrust may be already high. internationally; the establishment of intersectoral Failure to effectively communicate can undermine steering groups to facilitate taking the right actions public health campaigns, whereas a well-designed at the right time; the optimization of hospital care communication campaign can have a considerable (for example, by deferring some treatments) and impact. In Armenia, a poll conducted in March 2021 facilities (for example, by expanding the supply revealed that only one-third of respondents were where needed); the acquisition and local production willing to get vaccinated. Consequently, the gov- of medical equipment and pharmaceuticals, includ- ernment developed a communication strategy that ing vaccines; financial measures to alleviate the cost involved training media representatives to present of COVID-related care; the expansion of the health- information on COVID-19 to the public, and medi- care workforce by training people, hiring students, cal staff were also trained to provide general infor- and facilitating the immigration of healthcare per- mation about COVID-19 and protocol compliance. sonnel; the provision of financial compensation to These interventions contributed to a 10 percent healthcare facilities and personnel; the expansion increase in the proportion of Armenians willing to of testing and contact tracing capacity; the use of get a COVID-19 vaccine.48 digital resources and health information systems to monitor the spread of the pandemic; and the use of On the design side, the level of advanced planning communication campaigns to effectively communi- and preparedness to address health emergencies cate to the population the importance of abiding by also affected the quality of response. Even the best public health measures and of being vaccinated. health systems need to plan and prepare to address crises and health emergencies. Institutions such as Governance and Design. From a governance per- departments of public health, hospitals, and emer- spective, effective intersectoral coordination has gency medical systems need to routinely practice proven to be essential to coordinating the public their responses to emergencies under the super- health response. In Georgia and Armenia, for vision of well-defined coordination bodies. For instance, the creation of the Interagency Coordina- instance, Taiwan’s experience with the 2003 SARS- tion Council and of the Commandant Office facil- COV outbreak was followed by the passing of the itated coordination among sectors in response to Infectious Disease Control Act 2007 that provided the pandemic. In Moldova, the government intro- a legal framework to respond to future outbreaks duced the National Extraordinary Commission for of disease, including access to people’s travel his- Public Health that contained representatives from tory and National Health Insurance cards to rapidly all ministries and departments to ensure an inte- identify and isolate individuals who are likely to have Notably, many countries bypassed the disaster risk management committees and structures establishing dedicated coordinating mechanism. See 13 47 the companion social protection note: Coll-Black et al (2023). RFE/RL (2021). 48 been exposed to the virus. Similarly key to Taiwan’s jan, mandatory health insurance was expanded from preparedness for COVID-18 was the establishment covering only three pilot areas representing over 65 of a National Health Command Center in 2004 as percent of the population in March 2020 to cover the agency with the power to coordinate inter-sec- the whole population by April 2021. This mandatory toral work and acquire additional personnel in the health insurance now covers a comprehensive list of case of emergencies.49 In Eastern Europe and the 2,550 medical services that include emergency and South Caucasus, the lack of planning and drills in tertiary care. Similarly, although Georgia put restric- normal times may have hindered the COVID-19 tions on the number of publicly funded proce- response. Countries’ experiences in combating this dures, it expanded the benefits package of primary pandemic should serve as an essential foundation healthcare services, strengthened the capacities for developing long-term and institutionalized crisis of primary health care (PHC) teams, and improved management mechanisms. the coordination and increased the use of diagnos- tic and specialized services. These changes to PHC Financing. The quality and strength of health services will increase the access of the remote and financing and the organization of health care ser- rural population through the use of digital services, vices also strongly influenced the response. It is which has the potential to achieve universal access difficult to achieve universal access to health care to healthcare in the medium to long term. when households have to make high out-of-pocket (OOP) payments to access preventive and curative Delivery. Although various countries had undergone care, which was the case in all Eastern Partnership reforms to optimize their service delivery networks, countries before the pandemic. OOP payments their health systems before COVID-19 remained as a share of total health expenditure were two to skewed toward hospital services. Having a large four times higher than the European Union average number of hospitals allowed for a swift response in (Figure 13), which limits the ability of these countries terms of capacity to treat COVID-19 patients, but to respond to shocks.50 The need to extend health- hospitals did not always have the essential equip- care coverage was fully recognized in Azerbaijan, ment needed to deliver care. In Ukraine, for exam- where localized pilots for mandatory health insur- ple, some of the designated facilities for COVID-19 ance were scaled up nationwide in the midst of the patients lacked therapeutic oxygen, and in Mol- pandemic, which helped to protect the most vul- dova, the high concentration of hospitals in Chis- nerable population from financial hardship. inau made it difficult to deliver care throughout the whole country.51 Concentrating the delivery of The crisis also created an opportunity to push care in secondary and tertiary facilities also left pri- through the challenging long-term reform meas- mary care networks with scarce human and finan- ures that were essential for delivering an effective cial resources to respond to the pandemic, despite response. Even some more immediate actions also being responsible for contact tracing and remote had long-term positive effects. For instance, because care, two of the most important and cost-effective responding to the pandemic required a coordinated tasks for controlling the spread of the virus. system with wide healthcare coverage, in Azerbai- Figure 13: Out-of-pocket Health Expenditures 100% Share of current health expenditure 80% 60% 40% 20% EU average 0% HRV SVN CZE TUR SVK ROU POL EST BLR HUN BIH LTU KAZ MLT SRB RUS LVA MNE MDA BGR MKD ALB CYP GEO UKR KGZ UZB TJK AZE TKM ARM Out of pocket health expenditures Average of Development Peers Source: World Development Indicators. Averages are population-weighted. LIC & MIC do not include Russia. Note: Data are for 2018. The average of development peers is obtained by using a kernel regression. 14 49 Summers et al (2020). 50 El-Khatib et al (2020). 51 Koziel et al (2023). PROTECTING HUMAN CAPITAL through Shocks and Crises Workforce management has been another major tial care has happened in virtually every country challenge during the pandemic. Most countries during the pandemic, there has been wide variation were facing staff shortages even before the pan- in the extent to which it has happened, suggest- demic hit, as their health workforce was aging and ing that it might be partially avoidable if appropri- was concentrated mainly in cities. However, the ate measures are put in place. A survey of people pandemic exacerbated the situation. In our focus aged 50 and over in Europe and beyond indicated group discussions with health care workers, nurses that the unweighted proportions of unmet health- from Armenia stressed that there were periods care needs ranged from 4.2 percent in Spain to during the pandemic when only two nurses had to 22.9 percent in Israel (for patients forgoing medical care for 42 patients. Doctors also highlighted staff treatment), from 1.5 percent in Bulgaria to 50.4 per- shortages as one of the main problems in dealing cent in Luxemburg (for patients postponing sched- with COVID-19: “We have only one doctor left in uled medical appointments) and from 0.7 percent addition to the Head Physician. We have a staff of in Bulgaria to 11.1 percent in Lithuania (for facilities two to cover a vast territory, and it is very difficult to cancelling medical appointments).54 During the work.” In Georgia, nurses mentioned that each nurse COVID-19 pandemic, while health systems across was doing the work of three, leading them to leave Eastern Europe and the South Caucasus provided the health sector, thus exacerbating the shortage and expanded healthcare to COVID-19 patients, of health workers. In Moldova, a doctor mentioned some non-COVID patients were discharged from that “At my age, I had some thoughts of retiring hospital or their elective care was postponed. For at first [when the pandemic started], but I under- example, in Georgia, the government defined three stood no one would be left to work in my place.”52 To stages of standby readiness depending on the address these staff shortages and the low morale number of COVID cases in the country at any given of the remaining workforce, countries took drastic time. When a threshold of cases was reached, hos- actions. Azerbaijan raised the salaries of healthcare pitals were notified and required to discharge non- workers, and recruited retired personnel or hired COVID patients and empty some beds within 48 to senior medical students to cope with the increased 96 hours to be able to accept COVID-19 patients. demand caused by the pandemic.53 Although staff Also, some publicly funded elective procedures were shortages have been a prevailing problem in most postponed between November 2020 and March health systems, the advent of COVID-19 gave them 2021. The general public’s fear of getting infected an urgent wake-up call to increase the quantity of also meant that they postponed appointments to healthcare professionals and improve the condi- receive essential healthcare, which also contrib- tions in which they work. Primary healthcare clinics uted to the reduction in the delivery of care.55 In and hospitals had insufficient personal protective Moldova, data from the National Health Insurance equipment to protect their staff and even lacked Company (NHIC) showed a decrease in outpatient essential drugs such as therapeutic oxygen. Moreo- visits of almost 60 percent among certain groups ver, the negative impact of COVID-19 on the mental of patients (Table 1). The extent of forgone care was health of healthcare workers was unprecedented, also evident in the case of antenatal visits in Geor- leading to high rates of burnout. In all countries, too gia, where the number of women completing eight little attention may also have been paid to the long- antenatal visits declined in 2020 compared to the term impacts of forgone care. While forgone essen- previous year.56 Table 1: Foregone Care in Moldova 2019 to 2020 Condition 2019 2020 difference (%) Cancer 22,119 18,586 -16.0% Diabetes 9,062 4,995 -44.9% Hepatitis chronic 725 420 -42.1% Hepatitis viral chronic 3,142 1,269 -59.6% Cataracts 4,367 3,294 -24.6% Sprain/strains 237 173 -27.0% Infarct myocardia (chronic or over four weeks) 410 224 -45.4% Source: Koziel et al (2023). Note: Number or patients with chronic diseases who accessed care services, first 9 months of the year. 52 Koziel et al (2023). 15 53 Koziel et al (2023). 54 OECD (2021). 55 Koziel et al (2023). 56 See Koziel et al (2023). Staff calculations based on administrative data. Some actions taken by governments, while enacted This is an illustration of how information systems to support the immediate needs arising from the with up-to-date data are crucial for the close mon- pandemic, will also support the long-term devel- itoring of health services and for setting up innova- opment and resilience of their health systems. tive delivery models of care. However, more could These actions have included safeguarding long- be done to reduce the backlog of care as has been term oxygen supply by importing power generators demonstrated in some interesting examples around that produce oxygen as a by-product (Georgia) and the world. In Austria and Ireland, waiting lists have increasing local production of medical equipment to been updated by creating pre-triage clinics to find reduce dependence on the external production of alternative forms of care and support where appro- essential equipment (Azerbaijan and Georgia). priate for people who have been waiting a long time for needed care. In the Netherlands, a centralized Data and Information. The extent to which digi- data hub containing real-time data has been intro- tal technologies were used to deliver services also duced to evaluate demand for inpatient care and to affected the speed and quality of the response. As redistribute capacity accordingly.58 the pandemic developed, countries with efficient monitoring systems were able to make data-driven decisions. In Azerbaijan, for instance, collabora- Social Protection tion between different sectors made it possible to implement a “one-stop” digital platform that pro- All countries quickly passed emergency spending vided access to essential health services during the measures between March and May 2020 and initi- pandemic as well as information about e-services ated emergency social protection programs to sup- from other sectors such as commerce, education, port those who had lost income or were particularly food, and entertainment. Georgia made use of an badly affected by the pandemic. Countries used electronic information system that it had begun the full range of their social protection systems, developing as early as 2015 that integrates human from social assistance (including social services) to and animal health surveillance. In Armenia, addi- social insurance and labor programs. All countries tional analytic functions were added to the pre-ex- expanded their last-resort income support pro- isting e-health system (ArMed) to facilitate pan- grams in various ways, by increasing the level of ben- demic management and monitor the vaccination efits, by allowing automatic extensions for existing campaign.57 Although electronic information sys- participants, or, to a lesser degree, by changing the tems were also available in Moldova and Ukraine, eligibility criteria to expand coverage to new ben- these systems were only partially digitized, which eficiaries. Some countries also launched new social prevented them from being used effectively and assistance programs to provide timebound pay- which meant that pandemic managers had to rely ments to support children who lived in poor house- on multiple different sources of information or tra- holds or whose parents had lost their jobs, disability ditional paper-based data systems. In general, elec- beneficiaries, pregnant women, or students. Other tronic data systems were found to be crucial for approaches to supporting incomes included con- agile decision-making during the pandemic, which trolling the prices of key foods and subsidizing util- has emphasized the need for all countries to move ity bills, food purchases, or tuition. Most countries from paper-based to digital data systems to make also expanded their support for unemployed work- information rapidly available, not only for public ers and some to pensioners through their social health purposes and hospital care but also for the insurance schemes.59 Wage subsidies were used to monitoring of the supply of medical equipment, encourage employers to retain their formal sector human resources for health, and forgone care. The worker, which took the form of direct payments to COVID-19 pandemic has made GovTech even more firms, tax exemptions, and the financing of social urgent, accelerating the impetus to promote more security payments. Some countries also attempted effective, efficient, transparent, and accountable to create jobs, for example, by expanding public public services. works programs. Finally, many countries modified their social services to protect existing beneficiaries Countries have also used data and digital technol- from the pandemic, with some countries introduc- ogies to reduce forgone care and to resume pre- ing new programs to serve particularly vulnerable COVID levels of service delivery. For instance, during groups, such as bringing food to elderly people who the pandemic, Georgia used its Perinatal Registry, were isolating at home. developed in 2015, to reach out to all pregnant women (if they were detected in a timely manner) Financing. How emergency spending was allocated to provide them with educational as well as individ- among the different pillars of the social protec- ual and group consultative services through Zoom. tion system varied significantly between countries. 16 57 Koziel et al (2023). 58 Koziel et al (2023). 59 While these additional payments were delivered through social insurance schemes, they were government funded. PROTECTING HUMAN CAPITAL through Shocks and Crises Wage subsidies featured prominently in most of tiveness of the response in many countries. Some the countries, accounting for almost 60 percent countries had to establish new, ad hoc institu- of emergency spending in 2020 in Armenia, to 17 tional structures to manage their overall COVID-19 percent in Ukraine (Figure 14). This is not perhaps response, including social protection. Few, if any surprising given the widespread concern about job countries, had established “crisis windows” within losses in the early days and weeks of the pandemic. their existing social protection programs or the Armenia, Azerbaijan, and Ukraine allocated a rela- systems and procedures to expand access. Also, tively small proportion of their emergency spending because of this lack of planning, the allocation of to social assistance, while Georgia devoted the bulk spending between social protection programs did of its emergency spending to this purpose. Pay- not necessarily reflect countries’ priorities or eco- ments through the pension systems were an impor- nomic structures. For instance, many countries tant part of the emergency packages in Azerbaijan channeled most of their emergency funding to and Ukraine. Unemployment benefits varied from formal sector firms or workers, thus depleting the 14 percent to 26 percent of emergency spending in amounts available to protect the poor and vulnera- Azerbaijan, Georgia, and Ukraine. ble throughout the pandemic. This reflects the fact that emergency programs were not simply based Figure 14: Approximate Allocations of Emergency on reaching those most in need or even those most Resources affected by the pandemic but were also constrained 100% by the opportunities and limitations of existing pro- Allocation of emergency resources 90% grams and their supporting administrative struc- 80% tures. Similarly, the lack of preparedness of and 70% limitations within existing delivery systems meant 60% that the emergency social assistance response in many countries focused on providing more aid to 50% existing beneficiaries or those for whom data was 40% readily available (for example, in tax records), which 30% probably led to many of those made poor(er) by the 20% pandemic being overlooked. 10% 0% The breadth and promptness of a country’s response Armenia Azerbaijan Georgia Ukraine was also enabled or hindered by the coverage, ben- efit generosity, and, ultimately, flexibility of its exist- Social Assistance (other than Utility) Pensions ing social protection system. When COVID-19 hit, Unemployment Benefits Wage Subsidies (WS) countries in the region had well-established social ALMPs (other than WS) + Tax Rebates Utility Subsidies protection systems, which received, in some cases, significant levels of funding. However, this spend- Sources: Gentilini et al (2020), Guliyev (2021), IMF Policy ing was mostly focused on social insurance, espe- Responses to COVID-19 (2022), Background Notes (Koziel et al cially pensions. The coverage of social assistance (2023) for health, Coll-Black et al (2023) for social protection, and Chinen and De Hoyos (2023) for education), and Annual programs varied greatly between countries, reflect- Report of the Ministry of Finance for Azerbaijan.60 ing their differing mixes of programs and their size, adequacy, and objectives. In many instances, these However, despite the breadth of the response, the programs were categorically targeted and reached amount allocated to emergency social protection relatively few poor households (Figure 15). Low cov- was modest compared to routine social protection erage rates among the poor – coupled with contin- spending. The additional amounts of emergency ued high rates of informality – limited the ability of financing allocated to social protection in 2020 systems to provide immediate protection to those ranged from around 0.4 percent of GDP in Armenia poor and vulnerable households negatively affected to 1.8 percent in Georgia. In Georgia, however, the by the COVID-19 pandemic who were not already pandemic coincided with the roll-out of a reform of covered by the system. its social protection schemes, as a result of which, spending on social protection almost doubled to 9 Delivery. In social assistance, most pandemic sup- percent of GDP in 2020, although expenditures on port was channeled through existing programs. the emergency schemes only accounted for 1.8 per- Much of the increase in spending consisted of cent. increased benefits for existing beneficiaries (“ver- tical expansions”) or the allocation of benefits of a Governance and Design. Limited planning and new program to beneficiaries enrolled in an exist- preparation ahead of the crisis affected the effec- ing program (“piggybacking”). For example, in Mol- 60 Notes: Social assistance includes any direct payment which is not based on the level of previous social security contributions. These are approxi- 17 mations based on the description given in the emergency packages and are broad categorizations; these figures cover emergency packages from 2020 only. In Armenia, figures are based on actual spending in 2020 using administrative data. Armenia did provide some assistance to unem- ployed workers, but these are grouped with social assistance because they were not based on contributions. Moldova is excluded due to lack of disaggregated or detailed data. Figure 15: Coverage and Spending of Poverty- port programs was probably due to several factors, targeted Programs including explicit policy decisions to support exist- 45% 0.60% ing beneficiaries, rigidities within existing programs that prevented them from expanding in response 40% to increased demand, a lack of shock-responsive 0.50% 35% elements within the design of the programs, and delivery challenges. For instance, few efforts were 30% 0.40% made to inform potential beneficiaries of changes % of GDP Coverage 25% to program eligibility, which reinforced a general 0.30% neglect of outreach and undermining the effective- 20% ness of on-demand systems. These limited efforts 15% 0.20% were surprising given the stronger outreach efforts 10% that governments made for labor and employment 0.10% programs. In Ukraine, for instance, unemployment 5% and sickness benefits were widely promoted and 0% 0.00% made easily accessible through the Diya platform. AZE - UKR - MDV - ARM - GEO - In Georgia, staff of the State Employment Sup- 2015 2018 2018 2019 2018 port Agency felt that citizens were reasonably well Coverage total population (LH) Expenditure (RH) informed about labor programs as a result of exten- Coverage Q1 (LH) sive advertising and dissemination of information in Source: SPEED (Social Protection Expenditure and Evaluation various media. Furthermore, even when countries Database, 2022).61 took steps to simplify the recertification process dova and Ukraine, additional payments were made for benefits or, more importantly, change eligibility to pension beneficiaries, while Armenia provided an criteria, they put less effort into helping new ben- energy subsidy as a “top up” to existing beneficiar- eficiaries to apply for support. While Ukraine took ies of the poverty-targeted Family Living Stand- steps to expand the GMI to poor people affected ards Enhancement Benefit (FLSEB), and Georgia by the pandemic, its application rules were not provided an educational grant to students in poor eased, and still required in-person visits for first- households under Targeted Social Assistance. In time applicants to register. In Moldova, the govern- contrast, adding new beneficiaries to existing pro- ment raised the income threshold for eligibility, and grams (“horizontal expansions”) or introducing new the program was automatically extended to those programs for uncovered groups happened less households already in the management information often. One notable exception was in Ukraine, where system that were found to be eligible. However, not the eligibility criteria for the country’s Guaranteed all households who fell into poverty were already Minimum Income (GMI) program were modified. in the database, and some were unable to apply Ukraine also simplified the administrative require- because welfare offices were closed and procedures ments for enrolling in the Housing and Utilities for applying by phone were inconsistent. Subsidy (HUS) program, the country’s largest social assistance program, which had reached close to Overall, the level of support provided to households half of the country’s households in 2017 (6.5 million). through both regular and emergency social protec- Similarly, in Georgia the share of households in the tion programs tended to be low, thereby undermin- bottom quintile receiving Targeted Social Assis- ing the effectiveness of the response. Even when tance (TSA) or emergency-related transfers is esti- the monetary value of the benefits were increased, mated to have increased from 39.5 percent in 2018 focus group interviews with social protection pro- to 62.6 percent in the last quarter of 2020.62 Other gram recipients and social workers showed that horizontal expansions were more limited in scope.63 participants appeared to have mixed views on their true value. Armenian participants viewed the ben- The emphasis on serving existing beneficiaries is efits as being “quite irrelevant” to recipients’ needs, likely to have reduced the ability of the social protec- and Georgian participants differed among them- tion system to help those people who were driven selves, but the Moldovans asserted that, although into poverty by the crisis. Vulnerable informal sector the social aid was not enough to meet all basic workers may have been particularly badly affected, needs, it was nevertheless very helpful. The dura- since, with a few exceptions, few countries used pov- tion of the additional support was generally limited erty-targeted social assistance programs to reach to the length of the country’s state of emergency. poor and vulnerable workers who had lost their In Georgia, for example, support was provided for jobs. The lack (with some exceptions) of substantial six months. In Moldova, the increase in the income horizontal expansions of last-resort income sup- threshold used to determine eligibility and the 18 61 Note: Coverage is derived from each country’s Household Budget Survey (HBS). Expenditures are derived from administrative data and may be more up to date than the indicated survey year. The poverty-targeted programs in the shown countries are: AZE - TSA; UKR - GMI; MDV - Ajutor Social; GEO - TSA; ARM - FLSEB. 62 World Bank staff calculations. 63 Coll-Black et al (2023). PROTECTING HUMAN CAPITAL through Shocks and Crises higher benefit level for Ajutor Social, the main tar- payments accelerated the distribution of payments. geted social assistance program, was only effec- However, access to commercial banks differs from tive in April and May of 2020. Similar trends were country to country in in the region, which limited also seen in other benefits. For example, in Geor- the extent to which social benefits could be paid gia the temporary unemployment allowance to out electronically. In Ukraine, for example, 65 per- the self-employed instituted during the crisis was cent of payments to 20 million beneficiaries were a flat GEL 200 (US$64) a month, or 16.8 percent made through bank accounts. In Moldova, the exist- of the average wage in 2020, equivalent to about ence of a government service for electronic pay- 64 percent of the subsistence minimum for a typi- ments (MPay) made it possible to make payments cal household. In Ukraine, the temporary increase in in a variety of different ways, including credit cards, unemployment benefit was limited to the period of payment terminals, e-banking, and cash payments. the lockdown.64 In Armenia, most social assistance benefits and pensions are paid in cash, and beneficiaries who Data and Information. The degree to which impor- received bank transfers mentioned the need for tant aspects of service delivery in some countries them to travel to the bank to check whether or not such as targeting, profiling and enrollment were their transfers had arrived because of a lack of SMS digitized affected the effectiveness and prompt- notifications. ness of the response. Overall, the integration of databases substantially increased the ability of countries to assess existing and new beneficiaries Education effectively. The horizontal expansions of the last-re- sort income support programs in Georgia and Mol- The pandemic’s main impact on the education dova were enabled by the social registry (in Georgia) sector was the closure of schools and the need to and integrated beneficiary registry (in Moldova). adopt distance learning. Between March 1, 2020 In Ukraine, the verification of application informa- and October 31, 2021, the number of days when tion was automated, which meant that it could be schools were fully closed (affecting at least 80 per- done in minutes. Armenia and Georgia used the cent of the student population) ranged from 64 to data held by their tax authorities to provide benefits 205 (Table 2). When the number of days of “aca- to people who had lost employment. Similarly, the demic breaks” and “extended academic breaks” are use of banking systems to make social protection also included, South Caucasus countries are among Table 2: Timeline of School Closures 2020 2021 Country Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan Feb Mar. Apr. May June July Aug. Sept. Oct. Armenia Azerbaijan Georgia Moldova Ukraine Belgium Norway Denmark COVID-19 guidance for safe in-person learning released AAP ECDC Fully closed Government-mandated closures of educational institutions a ecting most or all (or at least 80%) of the due to COVID-19: student population enrolled from pre-primary through to upper secondary levels. Most schools across the country are on scheduled academic breaks for periods of at least one week. All study Academic Break: during this period is suspended. Schools are (a) open/closed in certain areas only, and/or (b) open/closed for some grade levels/age groups Partially open: only; and/or (c) open but with reduced in-person class time, combine with distance learning (hybrid approach). For the majority of schools (at least 80%), classes are being held exclusively in person, noting that measures Fully open: to ensure safety and hygiene in schools vary considerably from context to context and/or by level of education. Source: Chinen and de Hoyos (2023) based on UNESCO’s Tracker of school closures caused by COVID-19.65 Coll-Black et al (2023). 64 19 Note: The American Academy of Pediatrics (AAP) released, on June 24, 2020, the first school guidance for safe in-person learning (https://www. 65 aap.org/en/pages/2019-novel-coronavirus-covid-19-infections/clinical-guidance/covid-19-planning-considerations-return-to-in-person-edu- cation-in-schools/). The European Centre for Disease Prevention and Control released its recommendation on August 6, 2020 (https://www.ecdc. europa.eu/sites/default/files/documents/COVID-19-schools-transmission-August%202020.pdf). Figure 16: School Closures and School Learning Outcomes Student Learning SCHOOL INPUTS (cumulative process) FAMILY INPUTS Ine ciency e ect due to Inequity e ect due to more misallocation. relevance of family inputs in the production of learning. Well-o households have enabling conditions to Vulnerable households do not have the minimum mitigate (or partly), the e ects of school closures: conditions to learn: • Enough and appropriate devices • No or inadequate devices • Good connectivity • No or poor connectivity • Suitable space for children to study • No suitable space for children to study • Learning materials / books • No or minimal learning materials / books • Educated parents, with flexible schedules to support • No educated parents, with fixed schedules to support their children their children Source: Chinen and de Hoyos (2023). the 10 countries in Eastern Europe and Central Asia combination of online platforms (which included a that closed schools for the longest periods during variety of educational resources for teachers and the pandemic. students, such as learning materials and homework resources) and media broadcasts, especially televi- The impact of school closures on learning outcomes sion education programs. The countries also used a can be characterized by the effects that they had combination of synchronous (real-time interactions on learning inputs (Figure 16). When schools closed, between students and teachers while lessons and learning inputs had to be used to deliver remote instructional content are being shared) and asyn- learning activities, which significantly reduced chronous instruction (students learning at their own their ability to enable students to learn. For exam- pace with or without interacting with the teacher). ple, within just a few days, teachers had to deliver classes through online platforms with no previ- However, the most important failures that led to ous experience. With schools closed, family inputs massive learning losses were not necessarily in the became much more important in the production design and implementation of distance learning of learning, putting marginalized students at a dis- modalities but in the excessive time during which advantage compared to their more affluent peers. schools were closed and the overall absence of Students in well-off households were more likely to strong post-pandemic measures needed to close have the enabling conditions to mitigate, at least learning gaps. While school closures, at the onset of partly, the effects of school closures, including elec- the pandemic, may have been an inevitable public tronic devices available at home with access to the health measure, if the inefficiencies and inequities internet, parents with sufficient schooling and time that characterized remote education in the first to assist them during their learning process, and few months after the pandemic began would have quiet spaces to study, concentrate, and learn. Most been recognized, the reopening of schools would disadvantaged students did not have even these have been prioritized. Countries failed to adjust minimum conditions to facilitate learning. The shift their policies in response to the emerging evidence to distance learning, therefore, resulted in a general and, thus, minimize learning disruptions. As early as decline in learning outcomes (even more so in places June 2020 in the United States and August 2020 where distance learning was poorly delivered), with in Europe, evidence had emerged suggesting that a strongly regressive aspect because learning losses many girls and boys under the age of 12 who con- were much larger for children from poor and vulner- tracted COVID-19 did not develop symptoms and able backgrounds.66 that only a small proportion had severe symptoms. Soon afterwards, evidence also emerged showing Governance and Design. To provide some conti- that schools were not the virus hotspots that had nuity to the learning process while schools were initially been thought.67 Findings from Germany,68 closed, the governments of the region all adopted Spain,69 and the United States showed that the remote learning policies. These were based on a rate at which children transmitted the virus among 20 66 Agostinelli et al (2020). 67 Lewis et al (2021). 68 Isphording et al (2020). 69 Catala et al (2020). PROTECTING HUMAN CAPITAL through Shocks and Crises themselves and adults was significantly lower than students attending public schools. Nevertheless, transmission rates among adults. Additionally, in Armenia received donations of laptops that were early 2021, evidence showed that primary school distributed among students, and in Moldova, the teachers did not have higher infection and mortal- Ministry of Education distributed 10,000 laptops to ity rates than the general population, suggesting teachers and students.72 that teachers were not necessarily at greater risk.70 Around the same time – between the start of the Delivery. A significant challenge faced by education pandemic and late 2020 – both theoretical and systems while schools were closed was to ensure empirical studies concluded that school closures that teachers had the necessary equipment, skills, came at a considerable cost in terms of learning loss, and training to deliver online teaching. Countries particularly for disadvantaged students. It became adopted a range of measures such as offering free also clear that children were not only losing learn- connectivity and some forms of professional devel- ing in core areas such as numeracy and literacy but opment training to help teachers to transition into were also suffering from depression, anxiety, and remote learning. In Armenia, for instance, the dis- other mental health issues.71 Despite the evidence tance education platform provides teachers and available by early 2021, school closures remained an school management with links to educational plat- essential part of the strategy to contain the spread forms and instructions for using ICT tools. Teachers of the virus in many countries, including those in were also offered special mobile and fixed internet Eastern Europe and the South Caucasus. tariff plans and free access to the main applica- tions and platforms used for distance learning and Moreover, despite the urgency to address post- to accelerated courses on e-learning platforms. COVID learning losses, countries in Eastern Europe Georgia offered webinars to provide teachers with and the South Caucasus made only minor adjust- some pedagogical and methodological recommen- ments to school calendars to make up for the school dations. In Azerbaijan, the Ministry of Education days that had been lost and did not report having provided 14,000 educators with online training any plans to revise the duration of instruction to improve their ICT skills, while 40,000 teachers time or the content of curriculum. Some measures were given mobile internet access free of charge. In were taken to identify students’ learning gaps and Moldova, the Ministry of Education provided over address their cognitive and socioemotional wellbe- 12,500 laptops to both teachers and students from ing, but the efforts and investments did not match disadvantaged groups. Similarly, in 2021, the Minis- the magnitude of the challenge. try of Digital Transformation in Ukraine announced the Laptop for Every Teacher project to equip sec- Financing. While little information is available on ondary teachers with computers. Nonetheless, how countries’ core education budgets may have implementation challenges and the limited cover- changed because of school closures, most coun- age of some of these initiatives may have hampered tries designated spending aimed at facilitating the their effectiveness. In Ukraine, for instance, no lap- move towards distance learning. To assist poor and tops had been yet distributed by September 2021, vulnerable students with online learning, countries while in Armenia, training had been provided to only offered various forms of support such as subsi- 50 to 75 percent of the teachers by the beginning dized or free internet access and the provision of of 2021.73 textbooks in electronic format free of charge. For example, Armenia offered special mobile and fixed Some countries also introduced measures to sup- internet tariff plans to schoolchildren and teach- port populations at risk of being excluded. These ers and free access to the main distance learning included flexible and self-paced learning platforms applications and platforms. All 181 general second- (asynchronous learning) and psychosocial and ary and high school textbooks were also posted in mental health support for learners (for example, electronic format free of charge. Georgia introduced online counseling) to minimize the impact of school subsidized internet access to children in January closures on the well-being of students. Many coun- 2021, although this was introduced relatively late in tries also offered psychosocial and mental health the pandemic and was not widely publicized. Mol- support to learners by phone or online. For exam- dova offered 80 gigabytes of free internet to each ple, Azerbaijan created a hotline for students and individual in need, and in Ukraine, national mobile teachers through which school psychologists pro- operators provided free access to the education vided psychological counseling on various issues platform’s website and its mobile application. Fewer such as aggressive behavior among young students countries subsidized or provided devices for free, and fear of admission exams. Also, Moldova intro- with the exception of Georgia where, since 2011, duced additional in-kind support for lower-income laptops have been distributed for free to Grade 1 households. 70 Fenton et al (2021). 21 71 Lewis et al (2021). 72 Chinen and De Hoyos (2023). 73 Chinen and De Hoyos (2023). While distance learning has provided an essen- It is not too late to strengthen catch-up invest- tial service during school closures, implementation ments. The COVID-19 crisis has created not only the challenges affected its effectiveness, which, even urgency to implement learning recovery strategies under an ideal scenario, is less than that of face-to- but also the opportunity to institute forward-look- face learning. Common challenges faced by teach- ing policies to address a learning crisis that was ers and principals included limited digital skills, poor already emerging before the pandemic. Countries connectivity and/or the lack of equipment, diffi- could still reverse the learning losses caused by the culties in adapting the curriculum to remote edu- pandemic by: (i) prioritizing foundational skills in the cation, and an increased workload with little or no curriculum; (ii) using standardized tests to identify guidance from education ministries. On the receiv- students whose performance is below the minimum ing end, parents and students faced similar chal- learning standard; and (iii) implementing compen- lenges, including poor or no connectivity at all at satory or remediation policies.74 Compensatory or home, limited student interest and/or fatigue with remediation programs – such as tutoring, grouping, online learning, and deficient or inconsistent guid- or computer-assisted learning (CAL) – have been ance from education ministries about remote edu- proven effective when the instruction is adjusted cation. Poor connectivity and limited access to dig- to students’ learning level (to teach at the right ital devices were particularly challenging in remote level) rather than their age or grade. For instance, areas and among low-income families. In Azerbai- in India, when volunteers provided instruction at jan, for instance, middle- and low-income families the level of the student for 40 days, this resulted struggled to make full and effective use of online in significant learning gains.75 Also, in Italy, a pro- classes due to the shortage of digital devices at gram that provided free individual tutoring online to home, the lack of uninterrupted high-speed inter- disadvantaged middle school students during the net, and the overlapping class schedules of their COVID-19 lockdown increased students’ academic children. In Georgia, internet connection was a chal- performance by an equivalent of more than a full lenge, especially beyond Tbilisi. In rural Georgia, year of formal schooling as well as enhancing their children mainly accessed distance learning through socio-emotional skills, aspirations, and psychologi- mobile phones. The relatively small scale of some cal well-being.76 remote education programs was also an issue. Data and Information. The lockdowns associated with the COVID-19 pandemic led to massive exper- iments and investments to find digital technologies to substitute for face-to-face learning (initially) and later to complement it and to develop effective training systems. Notwithstanding these massive investments, many countries – well beyond Eastern Europe and the South Caucasus – failed to use the full potential of data and information to improve their policies. As mentioned above, despite the rig- orous evidence available by early 2021 about virus transmission in schools, many countries, including those in Eastern Europe and the South Caucasus, continued to use school closures as an essential part of their strategies to contain the spread of the virus. Moreover, many countries failed to use the full potential of core MIS systems to track the perfor- mance of both schools and students or to admin- ister learning assessment either during or after the pandemic to identify vulnerable schools and stu- dents who were lagging behind their peers. Also, countries were lax about integrating education and social protection support to the most vulnerable households, which would have helped to improve learning environments at home for students most in need. 22 74 The strategies discussed in this section are aligned with the World Bank’s framework for learning recovery and acceleration. See World Bank (2022b). 75 Banerjee et al (2016). 76 Carlana and La Ferrara (2021). PROTECTING HUMAN CAPITAL through Shocks and Crises 5. Going Forward: Building Better and More Resilient Human Development Systems The COVID-19 pandemic was a crisis on an unprec- Some responses that were implemented during the edented scale, and it is important to recognize the pandemic are also worth mainstreaming in the pro- rapid response and immense resources that were cess of building back better. This is particularly the mobilized. Countries across the globe and in the case for digital technologies, the use of which has Eastern Europe and the South Caucasus region expanded dramatically during the pandemic and implemented swift and drastic measures to limit the which, if properly mainstreamed, have the potential spread and effects of the pandemic, which included to significantly increase both the effectiveness and emergency healthcare measures, the enactment of efficiency of service delivery. bans on travel, and school closures. To help people to cope with these drastic measures, most countries In addition to the sector-specific recommendations also invested massively in social protection programs that are discussed in the companion notes, we here and distance learning modalities. These measures recommend some reforms that have the potential were implemented relatively quickly and, at least to increase the resilience of human development in the initial phase, the only available option was to systems in all sectors. improvise and rely on intuition as it was extremely challenging to base decisions on solid evidence. Prepare, prepare, prepare Yet, more crises will come – in fact the invasion The importance of preparing for future crises cannot of Ukraine is already worsening the human capi- be stressed enough. This simple mantra seems obvi- tal impacts of the pandemic. To safeguard human ous, but many countries do not follow it. Thorough capital gains, it is important to learn from the preparation is beneficial across all pillars in Figure response to the COVID-19 pandemic to strengthen 3. Most countries have disaster risk management the resilience of human development systems to policies in place, but they are often outdated, not future crises and to ensure that recurrent crises tailored to economic crises, or neglect the human will not affect human capital accumulation, par- development sectors. The public health sector ticularly among the poor and disadvantaged. The should have effective monitoring systems and leg- effectiveness of the response varied substantially islation/financing that can be quickly triggered in among countries, and, as time passed, many coun- the event of a pandemic, and hospitals should con- tries failed to learn from past mistakes to improve duct regular emergency drills. Institutions to coor- their responses. For instance, schools were often dinate emergency responses should also already not re-opened despite the overwhelming evidence be in place. For instance, in Moldova, the Commis- showing that it was relatively safe to do so, and sion for Exceptional Situations and the Extraordi- social protection programs – especially those aimed nary National Commission for Public Health have at the poor and vulnerable – reduced their support both played an essential role in coordinating and after the initial months of the pandemic despite evi- planning intersectoral measures to respond to the dence of the protracted nature of the crisis. pandemic at the national level. In social protection, legislation should be passed to enable programs to Overall, there is potential for improving responses be rapidly scaled up in response to crises, with rel- to future crises by improving governance and the evant changes to both eligibility and benefit rules designs of interventions, adapting elements of ser- set out in advance. Operational procedures should vice delivery to emergency situations, and making also be established to ensure business continuity, better use of data and information to boost evi- for example, by determining how additional staff dence-based policymaking in both normal times will be allocated quickly to help to identify and enroll and crisis situations. There is also a need to find new beneficiaries and to ensure that payments can ways to mitigate the long-lasting impacts of crises be made even in the face of a disaster. Sources of (including the COVID-19 pandemic) on human capi- financing should be identified in advance and, if tal and to develop policies to close the human cap- needed, pre-positioned through the adoption of ital gaps that emerged. a disaster risk financing strategy. As for educa- 23 tion, distance and other learning modalities should shows that resilience and long-term development be piloted in normal times, investments should be to a large extent go hand in hand. It is not neces- made in proper enabling environments for the new sary to sacrifice the long-term process of human modalities, and teachers, principals, students, and capital investment to build more resilient systems. parents should all receive proper training to increase This highlights the critical importance of strength- the effectiveness of the new modalities and mini- ening systems and reversing the downward trend in mize learning losses during crises. human development spending in the region. Respond with an eye to the future… …and continue responding after the crisis is over to close long-term, crisis-induced human capital When crises hit countries, they may find it easiest to gaps. By many measures, the worst of the pan- focus their attention on short-term needs. Yet it is demic is over, and new crises are already hitting the important not to lose sight of the medium and long- region. Therefore, it would be tempting to move term need to strengthen human development sys- on to new priorities. However, forgetting about tems by protecting, financing, and improving deliv- the past would be a mistake. The pandemic left ery. Taking a longer-term outlook supports not only profound scars on people’s human capital; learn- long-term development and human capital accu- ing losses among children and youths are massive, mulation, but also increases a country’s resilience deferred medical treatment is having a profound to crises. Over and over, evidence has shown that negative effect on the health of many people, and systems that can deliver quality services in normal precarious employment, particularly among young times tend to respond better to crises. Effective workers or those who are about to enter the labor public health systems are better equipped to mon- market, can also leave long-term scars. Proper crisis itor the spread of a pandemic and adopt swift pre- responses must support people in the medium to cautionary measures, and a health workforce that is long term, not only in the immediate aftermath of a properly trained, managed, and remunerated needs crisis. There is an imperative need to address human less additional support during stressful times. Data- capital deficits by investing in remedial education driven health policies also enable faster and more programs and catching up on the provision of med- effective responses. Social protection systems that ical care. If nothing is done to heal the scars, there provide adequate protection to poor and vulnera- is a strong risk that human capital in the region will ble households during normal times not only help deteriorate further over time (the red line in Figure to ensure that those households are able to better 2), affecting countries’ long-term prosperity pros- cope but also provide a stronger foundation for a pects. To be sure, countries are developing recovery response to shocks. Also, social protection systems measures, but often they are not adequate given that have invested in administrative tools to iden- the magnitude of the challenge. For example, in tify people in need of support (for example, social Georgia, assessments were used to diagnose stu- registries or effective on-demand programs) are dents’ learning outcomes and socio-emotional able to identify and enroll new beneficiaries more wellbeing once schools reopened. However, survey rapidly. For instance, during the pandemic, because results suggest that these diagnostic tests were not the verification of application information in Ukraine implemented everywhere and that not all schools was automated, it could be conducted in minutes. may have remedial activities for students who had Armenia used the data held by its tax authority to fallen behind. pay benefits to formal workers who had lost their jobs, and in Georgia, the existence of a social reg- Explore the potential of digitalization istry enabled the country to rapidly expand tar- while building back better geted social assistance to additional households. Countries with unemployment insurance schemes The crisis has identified opportunities to “build back also had a foundation – and a governance system – better” both by improving the quality of service upon which to expand their support to formal work- delivery in normal times and by increasing its resil- ers and firms. Similarly, education systems that had ience to future crises. The companion notes provide managed to accompany their decentralization and concrete examples for each sector, but an emerg- school autonomy reforms with both financial and ing, cross-sectoral theme is the potential of digital capacity-building support were better able to help technologies for improving delivery and monitor- schools to tailor their responses to local needs. In ing. Health information systems, social registries, contrast, those education systems in which a large integrated monitoring and evaluation systems, and proportion of students had poor learning outcomes payment and learning/training platforms can all sig- faced a higher risk of students falling further behind nificantly affect the quality of service delivery – and during school closures. The pandemic response at the same time improve the shock-responsiveness 24 PROTECTING HUMAN CAPITAL through Shocks and Crises of delivery systems. Ukraine’s widespread commu- for instance, increased benefits for existing social nication and digital applications for unemployment assistance beneficiaries but were less successful and sickness benefits offer insights into the kind of in expanding coverage. This left many vulnerable innovations that could be used to strengthen the households affected by the crisis with inadequate on-demand nature of social assistance programs support, which was particularly problematic given in the region. Armenia has several examples of how that many of the people who fell into poverty as a digital information collected by government agen- result of the pandemic had different characteris- cies can be used to inform the pragmatic and rapid tics than those who were already in poverty. In the selection process of social assistance beneficiaries, education sector, countries quickly moved to adopt including the use of tax data to identify workers, the remote learning, a necessary first response. How- data on pregnant women collected by the Ministry ever, over time, many failed to experiment with dif- of Health, and the student enrollment records held ferent approaches that could have reduced learning by the Ministry of Education. While digital invest- gaps, such as varying school opening times for dif- ments have great potential, it is important to adopt ferent grades or putting more space between stu- a gradual approach and ensure that any infrastruc- dents in classrooms. In the health sector, countries ture and software investments are sustainable and implemented massive vaccination campaigns, but are accompanied by skills and training and govern- vaccine hesitancy among the population hindered ance and legislative reforms. Digitalization can have their success. The use of behavioral approaches a long-term impact only if systems are properly as were used in Armenia could have improved the maintained and updated and if people know how to design of information campaigns and increased use digital technologies effectively. vaccination rates. Protect financing Develop cross-sectoral, household- centered governance approaches This is not the moment to reduce human develop- ment investments. A few exceptions notwithstand- Ultimately, the goal of disaster responses is to help ing, spending on the human development sectors households cope with crises and shocks. This, again, by countries in Eastern Europe and the South Cau- seems to be an obvious statement, but the pan- casus was low to begin with, and the COVID-19 demic responses often failed to adopt a house- pandemic has put further strain on public finances hold-centered approach. Putting the household at and fiscal space. When crises are expected to recur, the center requires countries to have a solid inter- it is vital to protect as well as increase fragile human sectoral governance structure from the outset capital endowments to support greater productiv- because decisions taken in one sector affect how ity, growth, and prosperity. Ideally, countries should households will react in other sectors. For instance, invest more in human capital given the need to both lockdowns require not only social protection meas- strengthen systems and close crisis-generated ures to help households to cope with lost incomes gaps. At the bare minimum, they should ensure that but also measures to support parents, children, and spending does not fall any further. teachers with distance learning and remote work- ing. In Moldova, the Ministry of Education provided Improve governance arrangements over 12,500 laptops to teachers and students from and the design of responses disadvantaged groups and offered teachers 80 gigabytes of free internet data. In parallel, the gov- The literature on disaster risk management and on ernment modified the poverty-targeted social pro- adaptive social protection heavily emphasizes deliv- tection scheme to provide additional benefits to ery issues. However, poorly designed responses – households with children. Some countries also put even if delivered effectively – can substantially limit in place solid cross-sectoral institutions to coordi- their impact on the ground. To take social protec- nate all aspects of the responses, but others mostly tion as an example, the countries of the region did focused on health measures, leaving the other min- invest in a variety of programs during the pandemic istries to design responses on their own. Decisions – from social assistance to pensions to unemploy- about what support to provide were also influenced ment and wage subsidies. However, the amounts by the voices of specific population groups, who invested, their allocation among programs and the were not necessarily the groups with the greatest duration of the investments did not always reflect needs and vulnerabilities. Taking an effective house- the severity of the vulnerability of different popu- hold-centered approach requires coordination lation groups or the amount and length of support among a wide range of ministries, a good cross-sec- that they needed. Design flaws also reduced the toral referral or case management system, regular effectiveness of specific programs. Many countries, monitoring of implementation and outcomes, and, 25 given the fiscal and capacity constraints faced by and complex education, health, and social protec- these countries, the careful consideration of spend- tion systems consisting of many actors with inter- ing and implementation tradeoffs across sectors ests that are not always aligned, implementation and programs to maximize the positive impact of can be as challenging as – if not more challenging responses. than – design and budget allocation. Politicians and policymakers have an enormous responsibility Boost data, information, and to create coalitions with relevant actors and search evidence-based policymaking to for consensus among stakeholders about how best support effective responses to improve human capital outcomes among disad- vantaged groups and construct more prosperous, In the immediate aftermath of a crisis, there is little equitable, and resilient societies. time for thinking and evaluating. However, moni- toring systems should be already in place to gather the information needed by policymakers to make evidence-based decisions about ways to respond. Moreover, many crises, especially man-made ones, can last for a long time (COVID-19 itself affected the social sectors and economic activity for two years), meaning that there is time for countries to improve their crisis responses based on emerging evidence. However, despite the length of the COVID-19 crisis, countries’ responses did not change much over time. In the education sector, little effort was made to assess the extent of learning losses. For exam- ple, administering large-scale student assessments would have enabled the design of more effective responses by assessing the extent of the learning damage caused by the crisis and identifying the children who were most badly affected children. In the social protection sector, after initial tweaks to respond to the emergency, the design of response programs changed little over time, and few, if any, longer-term reforms were adopted in response to shortcomings in program design or implementa- tion. More could have been done to identify gaps in coverage and needs as well as inclusion and exclu- sion errors and to modify programs accordingly. In the health sector, the picture was more nuanced. Health sectors in the region were more responsive than other sectors during the course of the pan- demic and adapted their crisis interventions in reac- tion to emerging international evidence in addition to implementing massive vaccination campaigns. However, the duration and rules of public health measures such as social distancing were set as much in response to political considerations as to the available evidence. Exploiting the opportunities created by COVID-19 to build more resilient human development sys- tems will require evidence-based policies, finan- cial resources, and political compromise. More resources should also be allocated to where they are most needed – serving disadvantaged house- holds and individuals. Unfortunately, well-designed policies with sufficient budgets are not a guarantee of success: implementation also matters. In large 26 PROTECTING HUMAN CAPITAL through Shocks and Crises References Agostinelli, F., M. Doepke, G. Sorrenti, and F. Catala, M., C. Prats, S. Alonso, E. Alvarez, M. Zilibotti (2020). “When the Great Equalizer Marchena, D. Conesa, and D. Lopez (2020). Shuts Down: Schools, Peers, and Parents in Analysis and prediction of COVID-19 for Pandemic Times.” National Bureau of Economic EU-EFTA-UK and other countries. Situation Research. No. 28264. Report. No. 136. https://biocomsc.upc.edu/en/ shared/20201002_report_136.pdf Alberini, Anna and Nithin Umapathi (2021). “What Are the Benefits of Government Assistance Center for Disease Control (CDC) (2022). Press with Household Energy Bills? Evidence from Release. New CDC data illuminate youth Ukraine.” Policy Research Working Paper mental health threats during the COVID-19 Series. No. 9669. The World Bank. pandemic. 31 March 2022. https://www.cdc. gov/media/releases/2022/p0331-youth- Azevedo, J.P., M. Gutierrez, R. de Hoyos, and J. mental-health-covid-19.html Saavedra (2022). “The Unequal Impacts of COVID-19 on Student Learning.” In: Reimers, Chinen, Marjorie and Rafael de Hoyos (2023). F.M. (eds), Primary and Secondary Education “COVID-19 as an Opportunity to Build During Covid-19. Springer, Cham. https://doi. Resilient Education Systems: Drawing Lessons org/10.1007/978-3-030-81500-4_16. from 5 Countries in Eastern Europe and South Caucasus. ” World Bank. Banerjee, A., R. Banerji, J. Berry, E. Duflo, H. Kannan, S. Mukherji, M. Shotland, and M. Walton (2016). Coll-Black, Sarah, Cornelius von Lenthe, Stefanie “Mainstreaming an Effective Intervention: Brodmann, William Shaw, Judith Sandford, Evidence from Randomized Evaluations of Alejandro Gonzalez, and Jamele Rigolini (2023). Teaching at the Right Level in India.” NBER Social protection in a world of crisis: Learning Working Paper. No. 22746, National Bureau of from the response to the Covid-19 pandemic Economic Research, New York. in the Eastern Partnership Countries. World Bank. Becker, Gary S. (1964). Human Capital: A Theoretical and Empirical Analysis, with Special Reference Dercon, Stefan (2002). “Income Risk, Coping Strat- to Education. National Bureau of Economic egies and Safety Nets.” World Bank Research Research, New York. Observer 17 (2). https://doi.org/10.1093/ wbro/17.2.141. Bowen, Thomas, Carlo del Ninno, Colin Andrews, Sarah Coll-Black, Ugo Gentilini, Kelly Johnson, EBRD (2021). “Regional Economic Prospects.” Yasuhiro Kawasoe, Adea Kryeziu, Barry Maher, European Bank for Reconstruction and Devel- and Asha Williams (2020). Adaptive Social opment, Washington, D.C., April 2021. Protection: Building Resilience to Shocks. Adaptive Social Protection: Building Resilience The Economist COVID-19 tracker, lower bound to Shocks. https://doi.org/10.1596/978-1- estimate. https://www.economist.com/graph- 4648-1575-1. ic-detail/coronavirus-excess-deaths-tracker. Accessed in February 2023. Carlana, M. and E. La Ferrara (2021). Apart but Connected: Online Tutoring and Student El-Khatib, Z., A. Out, U. Neogi, and S. Yaya (2020). Outcomes during the COVID-19 Pandemic. “The Association between Out-of-Pocket HKS Faculty Research Working Paper Series. Expenditure and COVID-19 Mortality Globally.” No. RWP21-001. Harvard University. J. Epidemiol Glob Health. 10(3): 192-193. 27 Evans, David K. and Fei Yuan (2019). “Equivalent Kakietek, Jakub Jan, Julia Dayton Eberwein, Years of Schooling : A Metric to Communicate Nicholas Stacey, David Newhouse, and Learning Gains in Concrete Terms.” World Bank Nobuo Yoshida (2022). “Forgone Healthcare Policy Research Working Paper. No. 8752. During the COVID-19 Pandemic: Early Survey World Bank, Washington, DC. Estimates from 39 Low- and Middle-Income Countries.” Health Policy and Planning. https:// Fenton, Lynda, Ciara Gribbon, David, Caldwell, doi.org/10.1093/heapol/czac024. Sam Colville, Jen Bishop, Martin Reid, Jane White, Marion Campbell, Sharon Hutchinson, Koziel, Anna, Ha Thi Hong Nguyen, and Alejandro Chris Robertson, Helen M. Colhoun, Rachael Gonzalez-Aquines (2023). A way forward for Wood, Paul M. McKeigue, and David, A. building resilient health systems. Lessons McAllister (2021). “Risk of Hospitalization learned from Eastern Europe and the South with COVID-19 among Teachers Compared Caucasus. World Bank. to Healthcare Workers and Other Work- ing-Age Adults. A Nationwide Case-Control Kruk, Margaret E., Emilia J. Ling, Asaf Bitton, Study.” Preprint. Epidemiology. https://doi. Melani Cammett, Karen Cavanaugh, Mickey org/10.1101/2021.02.05.21251189 Chopra, Fadi El-Jardali, et al (2017). “Building Resilient Health Systems: A Proposal for a Flabbi, Luca, and Roberta Gatti (2018). “A Primer Resilience Index.” BMJ (Online) 357. https://doi. on Human Capital.” 8309. World Bank Policy org/10.1136/bmj.j2323. Research Working Papers. Lewis, S.J., A.P.S. Munro, G.D. Smith, and A.M. Gentilini, Ugo, Mohamed Almenfi, Ian Orton, and Pollock (2021). “Closing schools is not evidence Pamela Dale (2020). “Social Protection and based and harms children.” BMJ, n521. https:// Jobs Responses to COVID-19 : A Real-Time doi.org/10.1136/bmj.n521 Review of Country Measures.” World Bank, Washington, DC. Lichand, Guilherme, Carlos Alberto Dória, Onicio Leal Neto, and João Cossi. (2021). “The Impacts Guliyev, Farid (2021). “Azerbaijan's Social Policy of Remote Learning in Secondary Education: Response to Covid-19: Swift Response, Evidence from Brazil during the Pandemic.” Modest Benefits.” Covid-19 Social Policy IDB-TN-02214. IADB Technical Notes. Response Series. No. 8. Universität Bremen. Linkov, Igor, Benjamin D. Trump, and William Hynes Gonzalez, Alejandro (2022). “Social Protection (2019). “Resilience-Based Strategies and Response to the COVID-19 Pandemic in Chile.” Policies to Address Systemic Risks.” Organisa- Unpublished case study. tion for Economic Co-Operation and Develop- ment, no. September. Hevia, Felipe J., Samana Vergara-Lope, Anabel Velásquez-Durán, and David Calderón (2022). Mackley, Andrew (2021). “Coronavirus: Universal “Estimation of the Fundamental Learning Loss Credit during the Crisis.” House of Commons and Learning Poverty Related to COVID-19 Library Brief Paper. No 8999. Pandemic in Mexico.” International Journal of Educational Development 88. https://doi. MSC Special Edition (2021). Beyond Westlessness: org/10.1016/j.ijedudev.2021.102515. Renewing Transatlantic Cooperation. 57th Munich Security Conference Meeting Global International Monetary Fund (2022). Policy Challenges. https://securityconference.org/en/ Responses to COVID-19. https://www.imf.org/ msc-special-edition-2021/ en/Topics/imf-and-covid19/Policy-Respons- es-to-COVID-19 National Academy of Sciences (2012). Disaster Resilience: A National Imperative. https://doi. Isphording, I.E., M. Lipfert, and N. Pestel (2020). org/10.17226/13457. “School Re-Openings after Summer Breaks in Germany Did Not Increase SARS-CoV-2 Cases.” OECD (2016). PISA 2015 Results (Volume I): Excel- IZA, 13790. https://docs.iza.org/dp13790 lence and Equity in Education. PISA. https:// doi.org/10.1787/9789264266490-en. 28 PROTECTING HUMAN CAPITAL through Shocks and Crises ——— (2019). “PISA 2018 Results (Volume I): What Reimer, David, Emil Smith, Ida Gran Andersen, and Students Know and Can Do.” Paris. https:// Sortkær Bent (2021). “What happens when www.oecd.org/pisa/publications/pisa-2018- schools shut down? Investigating inequality in results-volume-i-5f07c754-en.htm. students’ reading behavior during Covid-19 in Denmark. ” Research in Social Stratification and ——— (2020a). “School Education During COVID-19: Mobility. Vol 71. Were Teachers and Students ready? .” Denmark Country Note. https://www.oecd. RFE/RL (2021). Health Minister Sees Falling Vaccine org/education/Denmark-coronavirus-educa- Hesitancy In Armenia. https://www.azatutyun. tion-country-note.pdf am/a/31440654.html ——— (2020b). “School Education During COVID-19: Ross, Antrim and Anna Clarke (2021). “Credit Where Were Teachers and Students ready? .” It’s Due? Claiming Universal Credit during the France Country Note. https://www.oecd. Covid-19 Pandemic.” Peabody. org/education/France-coronavirus-educa- tion-country-note.pdf Sagan, Anna, Erin Webb, Natasha Azzopardi-Muscat, Isabel de la Mata, et al (‎2021)‎. Health systems ——— (2021). Strengthening the frontline: resilience during COVID-19: lessons for building How primary healthcare helps health back better. European Observatory on Health systems adapt during the COVID-19 Systems and Policies, World Health Organiza- pandemic. https://read.oecd-ilibrary.org/ tion. Regional Office for Europe. https://apps. view/?ref=1060_1060243-snyxeld1ii&ti- who.int/iris/handle/10665/348493 tle=Strengthening-the-frontline-How-pri- mary-healthcare-helps-health-sys- Sandford, Judith (2021). “Social Protection Response tems-adapt-during-the-COVID-19-pandemic to the COVID-19 Pandemic in Australia.” Unpublished case study. Office of the United Nations High Commissioner on Human Rights (OHCHR) (2023). Ukraine: civilian Schady, Norbert, Alaka Holla, Shwetlena Sabarwal, casualty update 3 January 2023. https://www. Joana Silva, and Andres Yi Chang (2023). How ohchr.org/en/news/2023/01/ukraine-civilian- the COVID-19 Pandemic Eroded Human Capital casualty-update-3-january-2023 and What to Do about It. World Bank. Patrinos, H.A., E. Vegas, and R. Carter-Rau (2022). An Smolić, Šime, Ivan Čipin, and Petra Međimurec Analysis of COVID-19 Student Learning Loss. (2021). “Access to Healthcare for People Aged World Bank Policy research Working Paper. No. 50+ in Europe during the COVID-19 Outbreak.” 10033. The World Bank. Washington, DC. European Journal of Ageing. https://doi. org/10.1007/s10433-021-00631-9. Pifarré i Arolas, Héctor, Enrique Acosta, Guillem López-Casasnovas, Adeline Lo, Catia Nicodemo, SPEED (Social Protection Expenditure and Evalua- Tim Riffe, and Mikko Myrskylä (2021). “Years of tion Database). 2022. Database, Washington, Life Lost to COVID-19 in 81 Countries.” Scien- DC: World Bank tific Reports 11 (1). https://doi.org/10.1038/ s41598-021-83040-3. Summers, Jennifer, Hao-Yuan Cheng, Hsien-Ho Lin, Lucy Telfar Barnard, Amanda Kvalsvig, Nick Psacharopoulos, George, and Harry Anthony Wilson, and Michael G Baker (2020). “Potential Patrinos (2018). “Returns to Investment in lessons from the Taiwan and New Zealand Education: A Decennial Review of the Global health responses to the COVID-19 pandemic.” Literature.” Education Economics 26 (5): The Lancet Regional Health - Western Pacific. 445–58. https://doi.org/10.1080/09645292.2 No. 4. 018.1484426. Thomas, Steve, Anna Sagan, James Larkin, Jonathan Psacharopoulos, George, Victoria Collis, Harry Cylus, Josep Figueras, and Marina Karanikolos Anthony Patrinos, and Emiliana Vegas (2020). (2020). “Strengthening Health Systems Resil- “Lost Wages: The COVID-19 Cost of School ience: Key Concepts and Strategies.” European Closures.” World Bank Policy Research Working Observatory on Health Systems and Policies. Paper. No. 9246.TheWorld Bank.Washington, DC. 29 Thorn, William and Stéphan Vincent-Lancrin (2021). ——— (2021). “Poverty and Welfare Impacts of Schooling During a Pandemic: The Experience COVID-19 and Mitigation Policies in Georgia.” and Outcomes of Schoolchildren During the Washington, D.C. First Round of COVID-19 Lockdowns. OECD. https://www.oecd-ilibrary.org/education/ ——— (2022a). “Ukraine Crisis Social Protection schooling-during-a-pandemic_1c78681e-en Briefing Note.” Internal document, Washington, D.C. UNHCR (2023). Operational data portal: Ukraine Refugee Situation, 24 January 2023 update. ——— (2022b). “The RAPID Framework and a Guide https://data.unhcr.org/en/situations/ukraine for Learning Recovery and Acceleration.” https:// thedocs.worldbank.org/en/doc/e52f553225 UK Government press release (2021). 3 in 4 UK adults 28903b27f1b7e61238e416-0200022022/ receive both doses of a COVID-19 vaccine. 10 r e l a t e d / G u i d e - f o r- L e a r n i n g - R e c o v - August 2021. https://www.gov.uk/government/ ery-and-Acceleration-06-23.pdf news/3-in-4-uk-adults-receive-both-doses- of-a-covid-19-vaccine. ——— (2022c). Poverty and Inequality Platform. https://pip.worldbank.org/home Vakis, Renos, Jamele Rigolini, and Leonardo Lucchetti (2016). Left Behind: Chronic Poverty World Bank and Asian Development Bank (2017). in Latin America and the Caribbean. https://doi. “Assessing Financial Protection against org/10.1596/978-1-4648-0660-5. Disasters: A Guidance Note on Conducting a Disaster Risk Finance Diagnostic.” Washington, World Bank (2012). FONDEN : Mexico's Natural D.C. Disaster Fund - A Review. Washington, D.C. ——— (2013). World Development Report 2014 : Risk and Opportunity — Managing Risk for Develop- ment. Washington, D.C. ——— (2016). “Education Resilience Approaches.” http://wbgfiles.worldbank.org/documents/ hdn/ed/saber/supporting_doc/brief/SABER_ EDR_Brief.pdf. ——— (2020a). “Poverty and Welfare Impacts of COVID-19 and Mitigation Policies in Armenia.” Washington, D.C. ——— (2020b). “Poverty Impacts of COVID-19 and Mitigation Policies in Azerbaijan.” Washington, D.C. ——— (2020c). “The Human Capital Index 2020 Update: Human Capital in The Time of COVID- 19.” International Bank for Reconstruction and Development, The World Bank, Washington, DC. ——— (2020d). “Welfare Impacts of the COVID-19 Pandemic in Moldova: Main Channels and Impact of Mitigation Measures.” http:// d o c u m e n t s .wo r l d b a n k .o rg /c u r a t e d / en/909531621926588990/Welfare-Impacts- of-the-COVID-19-Pandemic-in-Moldova- Main-Channels-and-Impact-of-Mitigation- Measures-Summary-Note 30 PROTECTING HUMAN CAPITAL through Shocks and Crises 31 © 2023 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org