CASH TRANSFERS IN PANDEMIC TIMES Evidence, Practices, and Implications from the Largest Scale Up in History Ugo Gentilini ACKNOWLEDGEMENTS The paper is the result of two years of work on Covid-19 responses at analytical and operational level as part of the World Bank’s Social Safety Nets Global Solutions Group. The report benefited enormously from conversations with policymakers, practitioners, and researchers emerging over the course of numerous meetings and events held in 2020-2022. The author is grateful to Michal Rutkowski for his strategic guidance; to the paper’s peer reviewers for their advice, including Iffath Sharif, Thomas Walker, Fernanda Brollo, and Ralf Radermacher; and to the organizers and participants of the 6th Zurich Conference on Public Finance in Developing Countries (December 2021), where the basic tenets of this report were presented. Mohamed Almenfi, Hrishikesh TMM Iyengar, Usama Zafar, Emilio Raul Urteaga, Giorgia Valleriani, Sheraz Aziz, and Jimmy Vulembera Muhindo provided excellent research assistance. The paper only reflects the author’s personal views. ABOUT THE AUTHOR Ugo Gentilini serves as Global Lead for Social Assistance with the Social Protection and Jobs Global Practice at the World Bank. Email: ugentilini@worldbank.org Published in July 2022 EXECUTIVE 2 SUMMARY CASH TRANSFERS IN PANDEMIC TIMES I INTRODUCTION 6 Evidence, Practices, and Implications from the Largest Scale Up in History A CRITICAL II 8 JUNCTURE? ANATOMY OF RESPONSES III 12 AND EMERGING PERFORMANCE THEMATIC ISSUES IV AND COUNTRY 46 PRACTICES LOOKING AHEAD: V EMERGING INSIGHTS 65 AND LESSONS VI REFERENCES 72 VII ANNEX 86 VIII ENDNOTES 94 EXECUTIVE SUMMARY Is Covid-19 a “game changer” for cash transfers? This tantalizing question has animated a large body of recent literature and over 60 virtual panels. This paper offers some clues to address the question by bringing together data, evaluations and practical experiences generated over the course of the pandemic. In particular, the paper flashes out differences between Covid-19 and other crises; it lays out an anatomy of global responses and offers novel data analysis around stylized international trends; synthesizes fresh empirical evidence on response effectiveness based on over 40 evaluations; discusses country- level operational practices as emerging from an array of high and lower-income contexts; and distills key insights with possible future implications. Ten facts emerge from global trends in cash transfer responses: 1. Cash transfers reached unprecedented levels of coverage, but unevenly so across countries and largely concentrated in the early phase of the pandemic. About 66% of those measures were introduced in the first semester of 2020. Cash transfers reached 1.36 billion individuals, that is, one out of six people in the world received at least one cash transfer payment. About half of the population of East Asia and North America were covered, while about one-tenth of Africa’s was so. 2. Cash transfers were of short duration and uncertain in extension. The average duration of cash transfers was only 4.5 months. Where these were extended, or in 16% of cases with available data, they were so for an additional 6.3 months on average. Program extensions were highly uncertain in their outcome. Duration choices may be in no small measure due to the uncertainty surrounding the direction of the crisis and its waves. Yet there seems to be no correlation between the duration of severe containment policies and the duration of cash transfers. 3. Transfers were generous, but likely not enough to counter forgone labor incomes. Covid-19 cash transfers were 70% higher than in the pre-2020 period. The amount of cash received by beneficiaries represented an average of 46% of median income. Yet evidence shows that such generosity may have not been sufficient to offset the temporary or permanent loss of livelihoods incurred over the pandemic, especially among vulnerable but not the poorest people. This is corroborated by empirical studies showing that cash transfers provided significant support in mitigating the effects of the pandemic, but without countering its full impact. 2 EXECUTIVE SUMMARY 4. Governments simplified the design of cash transfer programs in several ways. In-kind transfers were in some cases converted into cash; transfers were made easier to access, were mostly unconditional, and eligibility was simplified. In specific cases, transfers were adjusted in subtle ways to ensure relatively fast and localized spending, but without turning them into quasi-cash (or voucher) transfers. A range of practical innovations introduced during Covid-19 were absorbed into routine delivery systems. 5. Most responses relied on introducing new programs through various identification and registration mechanisms. Most cash transfers were introduced afresh. The identification of beneficiaries mostly relied on open registration, social registries, and tax databases. Among registration mechanisms, online platforms were the most widely utilized. 6. Most cash transfers were paid digitally. A total of 763 million people were reached by digital cash transfers. Digital payments yielded a number of benefits, including making people more likely to receive transfers on a timely basis. However, access to digital payments can be hindered at several levels and can amplify exclusion. It is key to ensure the alignment of payment models with contextual user capabilities and payment ecosystems. 7. Preexisting social protection delivery systems helped responses significantly, but they were not always a necessary nor sufficient condition for scale-up. The social protection responses to Covid-19 underscored the fundamental importance of delivery systems. However, these were often unable to reach people at the middle or so of the income distribution. This speaks to the salience of recognizing the importance of all delivery building blocks, and not just some (e.g., payments or registries), the relevance of interoperable government databases across sectors, as well as considering both public and private sector infrastructure in delivery capacity assessments. 8. The timeliness of transfers varied substantially. On average, coverage scale-up (horizontal extension) took 26 days, with timeliness ranging from 2 to 119 days. The increase in benefits (vertical expansion) was faster, taking an average of 18 days (varying between 0 and 67 days). Programs that were preexisting and using social registries were faster. The delivery of payments in high income countries shed light on a range of challenges also in those settings. 9. Governments moved centerstage, but civil society actors played an important role. A qualitative reading of the overall Covid experience indicates that governments embraced the crisis response in ways that are more visible and deliberate than in other circumstances. This is likely because of the role states played in managing the crisis. The pandemic has accelerated linkages between international humanitarian assistance and domestic social protection. National and local civil society actors were also called for widespread engagement in direct provision of assistance. 10. Record-level global spending was highly uneven across countries, with financing based on a range of sources. Average spending on social protection amounted to an average of 2% of GDP, three-quarters of which was devoted to social assistance. With a global average of $313/capita, high-income countries display a level of injection about 93 times higher than low-income countries. Domestic financing of responses occurred via transfers from the existing budgets, budgetary reallocation, contingency CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 3 funds, and in some cases taxation. These were complemented by extrabudgetary mechanisms with various degrees of alignment with national budgets. Private donations were also an important part of the financing mosaic. Global and country-level responses point to ten lessons for the future: 1. Whether Covid-19’s response is a “historical game changer” depends on a range of factors. These include where the bar for change is set – that is, whether it’s the scale of response; the capacity of the response to withstand the crisis; the influence it exerts on existing, routine delivery systems and practices; the effects on fomenting or amplifying public debates; and the ability to alter the structural configuration of systems. Framed in those terms, the pandemic has clearly been a game changer in some aspects, but not in others. 2. The pandemic has illuminated major blind spots in the labor market. There is a need for better protecting unemployed and precarious workers across a range of contexts, including strengthening, establishing, or rethinking unemployment insurance, social assistance, and their interactions for workers in non-standard employment, the self-employed and informal sector workers. 3. Future crises scale up may be enhanced by a wider use of social protection automatic stabilizers. Social protection responses relied heavily on discretionary choices – “who to cover, how much to provide, for how long” – and to a limited extent on automatic stabilizers, especially in low- and middle- income countries. Automatic stabilizers are nested within an “anticipatory” framework featuring early warning systems, scale-up triggers, protocols for program expansion, and pre-positioned financing. As such, they are anchored on transparent, predictable criteria, while also leaving room for the policy space and discretionary spending (to manage basis risk, etc.). These “thermostatic” measures could be further explored moving forward. To identify gaps in response readiness, social protection systems should be routinely stress-tested. 4. The pandemic displayed creative and diverse financing modalities, with trade-offs and fiscal headwinds looming on the horizon. A range of financing innovations were recorded, such as contingency financing and new funds being established in lower-income contexts. Private support also played a significant role in financing responses, such as contributions from the Zakat system. Yet many of those financing practices also present drawbacks. For instance, responses were in some contexts financed by redirecting resources from other social sectors. High debt levels is being recorded across country income groups. This indicates that sharp trade-offs might loom on the horizon. The crisis might present an opportunity to reenergize the debate on social protection financing across a portfolio of options, including for example the relationship between subsidies and transfers, the role of indirect taxes, and innovative earmarked taxation. 5. The expected multipliers of cash transfers helped fiscal and unconventional monetary policy converge. Given the macro contexts, central banks favored universal or quasi-universal cash transfers over other options as an effective route to reignite consumer demand and economic activity. Emerging 4 EXECUTIVE SUMMARY evidence from higher-income contexts shows that transfers may have generated economic multipliers via augmented consumption. Such evidence may also affect debates on the benefits of cash transfers in non-crisis settings. It is likely that future recessions entailing strong reductions in consumer demand, temporary large-scale cash transfers may become a default option in emergency response. 6. Capitalize on the new generation of urban social assistance to enhance the salience of social protection in cities, large informal settlements and throughout the urban-rural continuum. This would imply adapting conceptual, design and implementation practices, as well as devising fresh partnerships and novel framings. Operating in urban environments opens the door for accelerating the discussion on mobility, portability of benefits, and individualized assistance. 7. Covid-19 points to the need for universal delivery systems that could potentially reach the entire country population. A key role will be played by large-scale, dynamic, and interoperable information systems. This also entails a judicious digital approach to delivery functions like identification, enrolment and payments based on the surrounding ecosystem and user preferences. At what pace countries would travel toward utilizing universal systems in routine programs – i.e., ensuring social protection coverage, adequacy, and comprehensiveness for everyone in society – would depend on a variety of evolving political, economic, operational, and societal factors. 8. Simple cash transfers as possible benchmark. Given the simplicity of Covid-19 cash transfers benefit structure, there is renewed interest in benchmarking programs against basic unconditional design. This entails contrasting benefits, costs, and implementation requirements of different design variants pursuing similar goals for comparable populations. These involve a whole spectrum of options ranging from “plain vanilla” cash transfers to more comprehensive, integrated interventions. Other user-friendly delivery choices tested in the pandemic (e.g., remote applications) could also be benchmarked against standard practices. 9. Enhance information on global social protection responses. The experience from extensive Covid-19 social protection tracking suggests the need for more proactive, comprehensive, coherent, and coordinated reporting systems. Perhaps the development of global standards for social protection response reporting might enhance future tracking endeavors. Initiatives that may help institutionalize the collection, analysis, and dissemination of global data in crisis times might also be better explored. 10. Historical antecedents and current experience suggest four scenarios or pathways that countries may take post-Covid. Pending on if and how they act on pandemic lessons, country scenarios may include “the ostrich” (ignoring lessons), “the refresher” (marginal improvements), “the constrained reformer” (relatively significant enhancements), and “Beveridge redux” (comprehensive reform). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 5 I. INTRODUCTION It is hard to overstate the current vibrance of social protection debates. If Covid-19 shed light on gaps in social protection systems, the unprecedented scale of country responses reenergized the conversation on the future of those systems. As Covid-19 programs were tapering out in early 2022, shockwaves from the Ukraine conflict and global inflation reembarked governments on a new round of crisis interventions (Gentilini et al 2022a,b). Clear lessons from recent pandemic responses are needed to inform both short- term action and longer-run trajectories. To be sure, there has been no paucity of efforts in capturing emerging learning. Insightful perspectives were offered by country leaders involved in Covid responses firsthand (Behsudi 2022; Nishtar 2021). Various contributions synthesized global experiences (Bastagli and Lowe 2021; Gentilini 2021; Hammad et al 2021; Abdoul-Azize and El Gamil 2020). Suites of assessments were regionally geared (Blofield et al 2021a; Devereux 2021; ILO 2021a; Barrientos 2020) or covered country-specific responses (Fiala et al 2021; Arruda et al 2021; Beazley et al 2021). A branch of literature examined responses to particular profiles of beneficiaries (Alfers et al 2022, 2020; Barca et al 2021; Holmes 2021; Mukherjee 2020), and specific operational and delivery issues were investigated (Gelb et al 2022a; Ohlenburg 2022; GIZ et al 2021). Debates in dedicated commissions and conferences flourished (e.g., UN 2022; Peres et al 2021), while working groups – such as in the EU – were formed to “… build on the lessons from the COVID-19 pandemic and consider the interaction between social protection systems and other welfare policies”.1 Academic journals released special issues (Beland et al 2021; Cook and Ulriksen 2021), collections of working paper series were launched, 2 and institutions have been articulating their visions, roles, and experiences (Rutkowski 2020; UNICEF 2020; WFP 2020a; UNHCR 2020). Bibliographies and reference guides helped navigating the vast production of Covid-related social protection literature (Barca 2021). Drawing from my keynote address at the 6th Zurich Conference on Public Finance in Developing Countries, 3 this paper brings together trends, evidence and practices generated over the course of the pandemic. The focus revolves primarily around social assistance in general and cash transfers in particular, with both anchored to the broader remit of social protection. The objective is to give a compact, evidence- based, and coherent overview of issues and experiences unfolding in different contexts, and present a balanced discussion of recent performance and possible future directions. Why a particular premium on, but not an exclusive attention to, cash transfers? Reasons are threefold. First, cash transfers have been a core protagonist of the global Covid-19 response (see annex 1 for program- level descriptive statistics). As we will discuss in section 3, those instruments have been the single most-widely used intervention in governments’ social protection toolbox, accounting for about a quarter 6 I. INTRODUCTION total global responses across social assistance, insurance, and active labor market programs. Second, the responses itself presents an historical record: cash transfers coverage reached an unprecedented level of 1.36 billion people or 17% of the world’s population. Finally, research on cash transfers increased enormously in recent years, a fact that might help explain the massive surge in practice. A Google Scholar search conducted for this paper4 yielded nearly 1.3 million papers produced over the past three decades, a level that increased 26-fold since 1980 and reached 95,000 titles in 2021 (figure 1). Figure 1. Number of cash transfers-related publications, 1980-2021 100,000 95,175 90,000 80,000 70,000 60,000 50,000 40,000 30 ,000 20,000 10,000 3,644 0 2000 2008 2006 2004 2009 2005 2003 2002 2020 1980 1988 2007 1986 1990 2001 2010 1984 1989 1998 2018 1996 2016 1994 2014 1985 1999 2019 1983 1982 1995 2015 1993 2013 1992 2012 2021 1987 1981 1997 2017 1991 2011 Source: entries based on Google Scholar. Overall, the paper finds that social protection responses mitigated the effects of the pandemic significantly. Countries diversified their responses, and social assistance played a key role in reaching people and areas uncovered by social protection programs. The crisis displayed the value of pre-Covid investments in social protection capabilities, and in a number of cases innovations spurred by the pandemic were absorbed into routine, permanent delivery systems. However, responses were also uneven in provision across time and space, and didn’t often offset the full impact of the crisis. Assistance has been generally short-term and uncertain in extension over time, in part because of the unpredictable nature of the crisis itself. Enmeshed technical and political economy considerations might also have contributed to the volatility of responses. Covid-19 exposed structural gaps in serving particular population profiles and locations, as well as in connecting social protection with other sectoral policies. The reminder of the paper is structured as follows. The next section briefly contextualizes Covid-19 within the broader discussion on crises and their role in sparking change. Section 3 offers ten stylized facts emerging from global responses, including novel analysis chiefly drawing from a large public dataset of nearly 4,000 social protection programs with about a hundred indicators and 223 economies. The section then offers a review of design and delivery practices, compounded by synthesis of the effectiveness of social assistance measures as emerging from over 40 evaluations on welfare and economic effects. Section 4 turns into practical country-level experiences across an array of themes. These include how cash transfers were linked to labor markets, spatial policies, and human capital sectors (especially health and education). Finally, section 5 distills ten key insights and lessons for the future. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 7 II. A CRITICAL JUNCTURE? Over the course of 2020-2021, social protection was rapidly elevated as a fundamental component of countries’ response framework. 5 This has energized an intense debate on the future of social protection, including whether a radical disruption like the pandemic could alter the long-term, structural equilibrium of social protection coverage, adequacy, and comprehensiveness. For example, The Economist (2021) mused about “how to make a social safety net for the post-Covid world”, a question that animated at least 62 virtual panels in 2020-2022.6 Such vibrant conversation is motivated by multiple factors. For once, Covid might have expanded the notion of real and perceived vulnerability. The pandemic itself worsened wellbeing in relation to specific dimensions of wellbeing, such as poverty, human capital, mental health, and social cohesion7 (Kuhfeld et al 2022; Lampraki et al 2022; Abbott 2021; Sumner et al 2020). At the same time, while quantitative data is limited it might be plausible that a wide range of populations might perceive being increasingly exposed to social, health and economic risks (Verhaeghe 2020). For example, it has been argued that in Bangladesh there is a more acute awareness around the income vulnerability of informal sector workers, despite their average earnings being significantly above the poverty line (Johnson et al 2022). Covid-19 has injected a sense of urgency in “casting the net wide” for tackling longstanding deficiencies in current social protection systems as well as restoring trust in social contracts (Barca et al 2021; Perry 2021). Bierbaum et al (2021) argued that pandemic measures should be “… understood and utilized as building blocks for establishing rights-based national social protection floors, rather than remaining mere stopgap responses” (p.1). Following Razavi (2021), “… [w]e can turn the COVID-19 crisis into an opportunity to build robust, comprehensive and universal social protection [or] we can stumble zombie-like through this crisis and leave ourselves exposed to and unprepared for future shocks” (p.15). Perceptions of vulnerability, however, may vary over time. In Germany, a study by Ebbinghaus et al (2022) examined public attitudes towards social protection during the first year and half of the pandemic. It found short-term boosts in support of policies like unemployment insurance and family-related measures; but within a year, public opinions returned to prepandemic levels: “… the pandemic has induced a “veil of ignorance” and led to the widespread perception that everyone might be at risk of losing a job or requiring health care. (…) With decreasing uncertainty and increasing risk inequality along already existing social divisions, the cost argument – social insurance becomes more expensive for a majority – weights in, and social solidarity with the vulnerable is receding to normal levels” (ibid, p.10). 8 II. A CRITICAL JUNCTURE? Global social protection responses themselves might have fueled further hope for comprehensive enhancements. If the scale of cash transfers were unprecedented, the profile of a large share of beneficiaries is also historically noteworthy: in fact, 203.7 million workers received unconditional cash transfers, including 184 million in the informal sector and 19.7 million in formal jobs (Gentilini et al 2022a). Such approach may have broken a “last taboo” in providing cash transfers beyond traditional “deserving” populations (box 1). This leads to a more fundamental point: in order to properly interpret government pandemic responses, it is key to recognize that the nature of the Covid-19 pandemic and rationale for government responses are fundamentally different from previous shocks. Box 1. Breaking the last taboo? The history of social protection has been characterized by policymakers’ constant efforts of balancing the act of supporting people in need while managing unintended behavioral effects (Cooper and Szreter 2021; Lindert 2021). In particular, fears of inducing labor supply disincentives have dominated and profoundly influenced approaches towards cash transfers for centuries. For hundreds of years, governments have tried to (and still are) navigating the tension between supporting people while not make them depended on such support: the level of coverage, the size of benefits, and ancillary design measures (e.g., work requirement) are part of efforts to reconciling those somewhat competing objectives. As a result, social assistance has often been provided to “deserving” populations (e.g., children, seniors, disabled people, etc.) and seldom to working-age “able bodied” people – unless on a small scale or in the form of public works, themselves also geared to reduce interference with labor markets (via below-market wages, etc.). By explicitly supporting over 200 million workers, countries have broken a longstanding taboo in supporting “able bodied workers” with unconditional cash transfers. If before Covid-19, policymakers were concerned of labor supply disincentives, with the pandemic response those disincentives were induced intentionally. In most cases, cash transfers were designed precisely to keep people out of the labor market. (For example, Levy and Filho (2022) show that in Brazil, Auxilio Emergencial had a relatively small, statistically significant effect on women labor supply (by 3.4 percentage points.) Of course, such remarkable development has to be put in perspective. If reaching hundreds of million workers with unconditional cash might be a departure from longstanding practices, governments had basically no other choice. Yet, the fact that workers received cash no- strings-attached withholds a historical and symbolic value that may potentially help alter persistent negative rhetoric on cash assistance in the long run. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 9 Sao Paulo, SP, Brazil - May 12, 2021: People wait in big line to receive food donations for lunch in a downtown street during a severe economic crisis due to Coronavirus, COVID-19 pandemic. By Nelson Antoine. Stressing the different nature of the Covid-19 may appear obvious, but it’s an aspect that is sometimes overlooked in policy debates. Because of the unique way in which Covid affects human decisions and interactions, containment policies to curb the spread of the virus were adopted8 (Romer 2021; Goolsbee and Syverson 2020). As such, programs like state-provided cash transfers were a response to state-induced needs. Such implicit social contract establishes a direct relationship between the role of governments in reducing economic activity and compensating for it. This association is not present in other types of crises, and it motivates the basic rationale for Covid social protection globally. As put by Barrientos (2021), “… social assistance transfers have come to the fore in the pandemic in support of government- imposed restrictions on social and economic interaction”. This combination of historical scale occurring within a highly peculiar crisis made the discussion on whether Covid represents a critical juncture quite contested and controversial. When discussing future directions, most contributions fall along a spectrum. This includes scenarios where Covid has possibly been a “game changer” for social protection (e.g., Anderson et al 2022); views arguing that “it’s too early to say” whether such change might happen (e.g., Devereux 2021); and more skeptical stances about future prospects for change (e.g., Barrientos 2021). There is truth in all those perspectives: Covid-19 shed light on structural gaps (Aaron 2020; ILO 2021b); it suggests a need for change (Cichon 2021; Oxfam 2020); it has opened the space for bold ideas, even if on a temporary basis (Gray Molina and Ortiz-Juarez 2020); and it provides a window of opportunity for public debates on future directions (Gronbach et al 2022). The question is one of degree across those dimensions. This applies for big-picture issues for specific themes, such as leveraging social protection responses for improving children and youth’s mental health (Bauer et al 2021). However, forward-looking discussions may have sometimes unfolded along preexisting policy and advocacy lines, and perhaps without being grounded on a fuller theory of change (Leisering 2021). 10 II. A CRITICAL JUNCTURE? In principle, the thesis that Covid-19 could act as a “positive disruptor” is not without conceptual and historical grounding: an important branch of scholarship has identified the occurrence of covariate crises as enablers of change in social protection systems (Obinger et al 2018; Rixen and Viola 2015; Farnsworth and Irving 2011; Kaasch et al 2010; Pierson 2004). Gupta and Tovar Jalles (2022) look at the question through a financing lens: based on data for 45 low- and middle-income countries over 2000-2015, they examine whether recent pandemics or epidemics (e.g., SARS, H1N1, Ebola, and Zika) have led to tax reforms, like changes in tax rates, tax base, and exemptions. Their analysis reveals that “… crisis situations present an opportunity for a country to implement tax reforms”9 (p.2297). Another line of inquiry, however, suggests an opposite outcome: under uncertain circumstances, policymakers may tend to resort to existing, path-dependent approaches as default options (Starke et al 2013; Chung and Thewissen 2011). This may not necessarily be undesirable: building on existing systems could be ideal if preexisting performance is strong. For example, a review of case studies from East Asia argued that while “… policy responses seem to be a déjà-vu (…), past crises have nurtured ingredients for the ‘immune system’ of the East Asian welfare state that tends to reinforce its institutional resilience and to strengthen the efficacy of social policy responses to the pandemic crisis” (Soon et al 2021, p.384). A further consideration relates to whether interventions were intended to foster change in the first place. According to Cook and Ulriksen (2021), government social protection responses “… do not in themselves change the basic characteristics of the existing welfare system” (p.389). Similarly, Seemann et al (2021) noted that the Covid response was not intended to bring about systemic change to overall redistribution regimes. In fact, country measures “… have a relatively clear anchoring in the welfare state architecture (…). [Responses] have not modified the existing configuration of state responsibility towards citizens, and that behind the temporary crisis measures, the pre-crisis mechanisms of the welfare states have overwhelmingly remained intact” (ibid, p.551 and 562). Initial conditions matter: examining the extension of social protection to urban areas in Africa, Gronbach and Seekings (2021) argued that the pandemic “… strengthened pro-reform coalitions only when those coalitions already existed. Where pro-reform coalitions were weak, Covid-19 was generally not a ‘game-changer’” (p.450). When pondering whether the Covid experience has been (or still is) a critical juncture, it is key to ask the right questions. Ultimately, whether Covid is a catalyst for change would depend on where the bar is set; what is meant by “change” across different dimensions; and whether expectations to meet those expectations were realistic. Change, for instance, could be interpreted in terms of the scale of response, its adequacy, the influence on existing systems and practices, fomenting or amplifying public debates, and altering the structural configuration of systems. Covid-19 has been a game changer for some of these issues, but not for others. The paper will revert to such discussion in section 5 after reviewing key facts and experiences from pandemic responses. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 11 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE This section presents ten facts on global cash transfer responses, it offers a rapid tour of select practical and delivery matters, and reviews emerging evidence evaluating the effectiveness of cash transfer programs on welfare and the economy. 3.1 Unpacking cash transfer responses: ten stylized facts Fact 1. Cash transfers reached unprecedented levels of coverage, but unevenly so across countries and largely concentrated in the early phase of the pandemic Over the two-year period spanning from January 2020 to December 2021, a total of 3,856 social protection and active labor market measures were planned or implemented by 223 economies (Gentilini et al 2022a). About 66% of those measures were introduced in the first semester of 202010 (figure 2). From a peak of 876 measures announced in March 2020, the launch of programs steeply declined thereafter. Coverage was sustained by a growing, although limited number of extensions: this included 449 programs (or 16% of programs with available information), with the maximum number of extensions (44) reached in June 2021. By February 2022, only about one-fifth of measures with data available was still active. 12 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Figure 2. Monthly social protection measures announced and extended, 2020-2021 Programs announced Programs extended 1000 50 900 45 800 40 700 35 600 30 500 25 400 20 30 0 15 200 10 100 5 0 Oct 0 Oct Jan Jan Jul Jul Mar Apr May Mar Apr Nov May Nov Dec Dec Feb Sep Feb Sep Jun Jun Aug Aug 2020 2021 Source: data from Gentilini et al (2022a) Nearly 61% of social protection measures took the form of non-contributory social assistance, with the rest being equally split among active labor market programs and social insurance. At program level, “cash was king”: with 1,023 measures in 203 countries, cash transfers represented 26% of social protection and 43% of social assistance measures, respectively. Their coverage has been nothing short of extraordinary: over 2020-2021, cash transfers reached 1.36 billion individuals, that is, one out of six people in the world received at least one cash transfer payment.11 About half of the population of East Asia and North America were covered, while about one-tenth of Africa’s was so (figure 3). Only two low-income countries – Gambia and Sudan – reported cash transfers coverage rates of over 20% of the population. Figure 3. Coverage of cash transfers by country income group and region, % of population 51% 49% 50% 50% 44% 40% 40% 33% 30% 30% 25% 24% 23% 19% 20% 20% 15% 10% 50% 8% 50% 0% 0% HIC UMIC LMIC LIC EAP NAC LAC MNA ECA SAR AFR (n=22) (n=40) (n=31) (n=15) (n=20) (n=1) (n=22) (n=11) (n=21) (n=6) (n=27) Source: data from Gentilini et al (2022a) CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 13 Fact 2. Cash transfers were of short duration and uncertain in extension The average duration of cash transfers was only 4.5 months. Where these were extended, or in 16% of cases with available data, they were so for an additional 6.3 months on average. This means that out of the period for which response tracking is available, or roughly 24 months, transfers supported people for nearly 13 months at best. For at least 10 months, people were left uncovered. Program extensions were highly uncertain in their outcome. In fact, extensions were surrounded by a degree of unpredictability on whether and when they would be enacted, with decisions often taken “last minute”. For example, Indonesia’s unconditional cash transfer program (BLT Dana Desa) was implemented from April to June 2020. Its extension (until September 2020) was reported a month in advance of expiration, or June 3. But other cases gave limited advance notice: Jersey’s Covid cash transfer (CRESS) was executed over the same period, i.e., April-June 2020. But it was only on its last day, June 30, that the program was eventually extended until August 31. In the Maldives, the Income Support Allowance – which also ran over the same April-June period – was reintroduced in October after it had lapsed. Similar uncertainty was observed for other social protection instruments: in the United States, the Pandemic Unemployment Assistance program was set to expire on December 31, 2020, and its continuation was confirmed on December 27. The program was extended again on September 6, 2021, a few days before the program was discontinued (Romer and Romer 2021). Those considerations need to be put in perspective: for once, the pandemic itself followed an uncertain trajectory. Determining the right duration of programs was no easy task as governments tried to balance health and economic needs in the wake of sudden new variants and surges in contagion. Also, cash transfers were the premier response modality, but other instruments were present. Especially in the upper bound of the country income spectrum, programs other than cash transfers (including labor market policies, insurance, as well as non-cash social assistance) have cushioned the effects of the pandemic (see section 3.3 for an assessment of response effectiveness). While it could be argued that even in high income settings social insurance reached structural limits in scale-up (see section 4), the reach of components other than social assistance declines as country income dwindles, hence making social assistance the cornerstone of response. In fact, in low income countries social assistance measures represented 78% of the Covid-19 relief – a level 28 percentage points higher than its share in high-income country responses. Were cash transfers aligned with the severity of the pandemic? Figure 4 plots the average duration of the Stringency Index (number of months that the index is above a score of 75) and the average duration of cash transfers (number of months) for 80 low- and middle-income countries over 2020-2021. The figure shows no correlation between the duration of severe containment policies and the duration of cash transfers.12 In other words, when interpreted over the full course of the pandemic as it unfolded in 2020 and 2021, cash transfers duration and adequacy may, in hindsight, have been insufficient to counter the effects of the crisis. Clearly, the severity of the pandemic varied over the biennium, and duration choices may be in no small measure due to the uncertainty surrounding the direction of the crisis and its waves. Under such conditions, short-term choices were made, which may have allowed governments 14 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Figure 4. Country Stringency Index and cash transfers duration, average 2020-2021 (n=80) 14 Duration of stringency above 75 (in months) 12 10 8 6 4 2 0 0 2 4 6 8 10 12 14 16 18 20 COVID cash duration in months Source: data from Gentilini et al (2022a) and Hale et al (2021) to continuously reassess the situation as it evolved. Yet the limited number of extensions and the volatility underpinning them may not have been optimal for impacted beneficiaries. Fact 3. Transfers were generous, but likely not enough to counter forgone labor incomes By pre-Covid standards, cash transfers were generous in size: on average, transfers were 70% higher than in the pre-2020 period.13 The amount of cash received by beneficiaries represents an average of 46% of median income,14 ranging from nearly 80% in low-income countries to 35% in upper middle- income settings (a level that was below 30% for East Asia and Latin America) (figure 5). Figure 5. Cash transfers as share of median income by country income group and regions (n=135) Adequacy (median income) 90% 78% 80% 67% 70% 58% 60% 49% 46% 50% 46% 40% 37% 37% 40% 35% 28% 25% 30 % 20% 10% 0% AFR SAR ECA NAC MNA EAP LAC LIC HIC LMIC UMIC Grand (n=38) (n=6) (n=42) (n=2) (n=11) (n=15) (n=21) (n=22) (n=34) (n=33) (n=46) Total (n=135) Source: data from Gentilini et al (2022a) and World Bank PovCal. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 15 Yet such generosity may have not been sufficient to offset the temporary or permanent loss of livelihoods incurred over the pandemic. An analysis from ten Latin American countries may illustrate this point. Busso et al (2021) show that with the exception of Brazil and Bolivia, social assistance replaced only a small share of potential earning losses of households. Average replacement rates – or total transfers that households receive as share of their regular labor income – were nearly 80% among the poorest beneficiaries. Yet such rates were only about 30% and 20% among households in the second and third quintile, respectively (figure 6). In the words of the authors, social assistance “… could not replace the potentially foregone incomes of a large fraction of families in the informal workforce that were forced to shelter in place and could not work. This insufficient compensation has surely limited the ability of governments to sustain extended lockdown periods, and it may limit the ability to enforce another wave of lockdowns if there were a resurgence of contagion” (ibid, p.12). Figure 6. Coverage and replacement rates for social assistance programs in LAC Figure 6. Coverage and replacement rates for social assistance programs in LAC 80 Coverage incluing all programs 60 Coverage incluing only Percentage expansion of existing safety nets 40 Replacement Rate 20 0 1 2 3 4 5 Labor Income Quintiles Source: adapted from Busso et al (2021); data refers to regional unweighted regional averages. Fact 4. Governments simplified the design of cash transfer programs in several ways Country adaptations included extending coverage by either relaxing eligibility criteria of existing programs to enroll new beneficiaries, or introducing ad hoc new interventions (horizontal expansion of 727 programs across 192 countries); transferring more assistance to existing beneficiaries or anticipating payments due at later stage (vertical expansion of 146 programs in 90 countries); making programs simpler to access, e.g., digital or remote applications and benefit management (50 programs in 43 countries); and revising design to make interventions more pandemic-sensitive (e.g., largely by making transfers unconditional).15 As mentioned, cash transfers were clearly the dominant form of social assistance transfers relative to in-kind measures (figure 7). Part of such dominance resulted from the temporary conversion of in-kind transfers into cash. A case in point are school feeding programs which switched to cash-based transfers in 25 countries, e.g., Brazil, Egypt, Malawi, Niger, Spain, and UK.16 16 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Figure 7. Share of cash and in-kind transfer programs out of total social assistance (n=1,583) HIC (n= 456) 71.4 28.6 UMIC (n= 523) 72.8 27.2 LMIC (n= 400) 66.0 34.0 LIC (n= 195) 60.0 40.0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Cash Inkind Source: data from Gentilini et al (2022a) Source: data from Gentilini et al (2022a) Some hybrid modalities also emerged. While nominally “cash”, transfers were not always usable in any location or at any point in time. In specific cases, transfers were adjusted in subtle ways to ensure Some hybrid relatively modalities fast also and localized emerged. spending, While nominally but without “cash”, turning them transfers (or into quasi-cash were not always voucher) usable transfers. In in a way, such forms of “restricted cash” represent a novel transfer modality at the intersection of cash and any or at any locationtransfers quasi-cash 2). in time. In specific cases, transfers were adjusted in subtle ways to ensure point (Box relatively fast and localized spending, but without turning them into quasi-cash (or voucher) transfers. Box 2. Hybrid modalities at the intersection of cash and quasi-cash transfers In a way, such forms of “restricted cash” represent a novel transfer modality at the intersection of cash Several practices emerged in terms of restricting the use of cash transfers (Soon et al 2021; Shi and Soon 2020). and e-moneytransfers quasi-cash Korea’s payments, (Box which2). provided about $900 for families of four or more members, had an expiration date (August 31, 2020). Also, the transfers could only be used at small-medium enterprises within the residential areas of beneficiaries. Similarly, Taiwan’s Triple Stimulus Voucher reached 96% of the population. Under the scheme, people could “purchase” $100 worth of transfers by paying one third of that level, or about $3317: the entitlement was expendable nationwide, but also in this case, transfers were timebound, i.e., they had to spent by December Box 2. Hybrid modalities at the intersection of cash and quasi-cash transfers 31, 2020. More traditional vouchers, however, have been leveraged to play an important role in recovery plans. For example, Several Australia’s Newpractices emerged South Wales in terms government of restricting provides the at least 6 types ofuse of cash voucher transfers programs 18 , such as the et (Soon al 2021; “Before and after school care voucher scheme” (a one-off voucher of $500 per child aged 4 to 13 years) – an approach that also Shi and Malaysia Soon – adopted 19 2020). and theKorea’s “First Lap learn to payments, e-money swim vouchers” which provided (a $100 voucher about towards $900 for families the cost of swimmingof lessons for children aged 3 to 6 years). Some of these schemes are recent, although evaluations exist for the four or programs’ more members, precursors, such as thehad NSW anActive expiration date (August Kids voucher 31, 2020). the program subsidizing Also, the cost transfers of children's could only registration in structured physical activities (Reece et al 2021). be used at small-medium enterprises within the residential areas of beneficiaries. Similarly, Taiwan’s Another Triple way in whichStimulus Voucher design was reached eased involved 96% of making the population. transfers Under theAmong more unconditional. scheme, people cash-based programs, unconditional cash transfers represent 89% of cases (the rest being conditional cash transfers public“purchase” andcould works). This$100 worth of constitutes an transfers by paying one 11%, or 9 percentage third points, of that increase level, or relative to about $33:17 the the approximate entitlement pre-Covid share was nationwide, expendable cash. of unconditional If we but in this also the enlarge case, sample transfers to were include all timebound, Covid-related i.e., social 20 assistance transfers21, the share of Covid unconditional transfers (whether provided in cash or in-kind) was they spent had to86%, on average by December ranging from about 2020. 31,78% in LICs to 89% in HICs (figure 8). Such average share is about an 8 percentage points increase relative to pre-Covid levels. However, this doesn’t mean that prior skepticism toward unconditionality entirely dissipated. For example, Aidukaite et al (2021) argue that in More traditional vouchers, however, have been leveraged to play an important role in recovery Wales government provides at least 6 types of plans. For example, Australia’s New South 17 voucher programs,18 such as the “Before and after school care voucher scheme” (a one-off voucher of $500 per child aged 4 to 13 years) – an approach that also Malaysia adopted19 – and the “First Lap learn to swim vouchers” (a $100 voucher towards the cost of swimming lessons for children aged 3 to 6 years). Some of these schemes are recent, although evaluations exist for the programs’ precursors, such as the NSW Active Kids voucher program subsidizing the cost of children’s registration in structured physical activities (Reece et al 2021). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 17 Another way in which design was eased involved making transfers more unconditional. Among cash- based programs, unconditional cash transfers represent 89% of cases (the rest being conditional cash transfers and public works). This constitutes an 11%, or 9 percentage points, increase relative to the approximate pre-Covid share of unconditional cash.20 If we enlarge the sample to include all Covid-related hoice for response options other than unconditional transfers was rooted in beliefs “… social ased society’ that claims assistance that transfers, no handouts 21 the share should be provided of Covid to the unconditional workless (…). In line transfers (whether provided in cash or in- e options other than unconditional transfers policy paradigm, the public works programme was was rooted in beliefs extended “… simple physical providing ft the claims unemployed” kind) was that no handouts should (p.363). onbeaverage provided86%, to theranging worklessfrom (…). Inabout line 78% in LICs to 89% in HICs (figure 8). Such average share the public works programme was extended providing simple physical d” (p.363). is about an 8 percentage points increase relative to pre-Covid levels. However, this doesn’t mean that e 8. Share of unconditional and conditional transfers out of social assistance (n=1583) prior skepticism toward unconditionality entirely dissipated. For example, Aidukaite et al (2021) argue onditional and conditional transfers out of social assistance (n=1583) that in Hungary, the choice 10.8 other than unconditional transfers was rooted in beliefs for response options 11.1 21.5 18.8 18.8 ‘work-based society’ “… about a 11.1 10.8that claims that no handouts should be provided to the workless (…). In line with this social policy paradigm, the public works programme was extended providing simple some of physical work toHungary, the the unemployed” choice (p.363). for response options other than unconditional transfers was rooted in beliefs “… about a ‘work-based society’ that claims that no handouts should be provided to the workless (…). In line 88.9 policy paradigm, 89.2 with this social the public works programme was extended providing simple physical 78.5 81.3 88.9 the unemployed” (p.363). work to some of 89.2 81.3 Figure 8. Share of unconditional and conditional transfers out of social assistance (n=1583) Figure 8. Share of unconditional and conditional transfers out of social assistance (n=1583) 100% 11.1 10.8 LIC (n= 195) LMIC (n= 400) UMIC (n= 523) 21.5 HIC (n= 456) 18.8 LMIC (n= 400) UMIC (n= 523) Conditional Unconditional 80% HIC (n= 456) m Gentilini et al (2022a) Unconditional Conditional 60% 22a) nt of simplification is evident in the targeting of transfers. Out of 374 programs with mation, methods based on occupations were used in the majority of cases (25.9%). These 88.9 89.2 on is evident s transfers toin the targeting workers based of on transfers. type of40% Out of 374 programs employment, activities,with 78.5 and sector 81.3 without based on s-based occupations eligibility were criteria used (e.g., in the majority frontline workers, of cases sector (25.9%). specific These at informal and targeting workers,based workers etc.). on Such type of employment, methods are followed activities, and sector based without by transfers on income (22.7%) and criteria (e.g., frontline workers, sector 20% targeting at informal and specific eria (21.6%), like families with children, age, and disability. Methods based on nd mixed methods represent 14.1% of the cases. Other methods areand uch methods are followed by transfers based on income (22.7%) single digits22 (table ke families with children, age, and disability. Methods based on 0% ds represent 14.1% of the cases. Other methods are single digits (table 22 LIC (n= 195) LMIC (n= 400) UMIC (n= 523) HIC (n= 456) Table 1. Targeting methods of cash transfer programs (n=374) Unconditional Conditional Means Proxy Food geting methods of Source: cash transfer Community- programs (n=374) Self- upation- Universal Geographical data from Gentilini et Source: al (2022a) data test, from means Gentilini et al (2022a) security, Combos Means based Proxy income targeting based Community- Self- testFood nutrition sal Geographical test, means security, Combos 25.9% 4.28% based 1.87% Another 1.6% test element targeting 22.73% of simplification 2.94% is evident 4.28% in the 14.17% 0.53% targeting of transfers. Out of 374 programs with income nutrition % 97 1.87% 1.6% element Another 22.73%of simplification is evident in 1. Targeting Tablewere available information, methods based on occupations methods used in the of majority of cash cases transfer (25.9%). These 16 7 were 2.94% defined as 6 4.28% transfers0.53% 85 14.17% to workers based on type of employment, activities, and sector without 11 16 2 53 the targeting of transfers. particular Out16of 374 needs-based programs eligibility programs criteria (e.g., frontline (n=374) workers, sector specific targeting at informal and m Gentilini et7al (2022a) 6 85 11 2 53 self-employed workers, etc.). Such methods are followed by transfers based on income (22.7%) and 22a) with available information, methods categorical criteria based (21.6%), like on ponses relied on introducing new programs through various identification and registration % out families with children, age, and disability. of MethodsNumber based on combinations and mixed methods represent 14.1% of the cases. Other methods are single digits22 (table occupations ntroducing new programs through were1). used various in the majority identification of cases and registration total transfers (25.9%). These were defined asTable to Categorical 21.60% 1. Targeting methods of cash transfer programs (n=374) 81 18 workers based on type of employment, activities, Means Occupation-based Community- Proxy Food 25.90% Self- 97 security, 18 Categorical Occupation- Universal Geographical test, means Combos based targeting and sector without particular based needs-based Universal income test 4.28% nutrition 16 % out 21.6% 25.9% 4.28% 1.87% 1.6% 22.73% 2.94% 4.28% 0.53% 14.17% eligibility criteria (e.g., frontline workers, sector of total Geographical 1.87% 7 specificNumber 81 targeting at 97 informal and 16 self-employed 7 6 85 11 16 2 53 Source: data from Gentilini et al (2022a) Community- based 1.60% 6 workers, etc.). Such methods are followed by Fact 5. Most responses relied on introducing new Means test, programs income through 22.73% and85 various identification registration transfers based on income (22.7%) and categorical mechanisms Proxy means test 2.94% 11 criteria (21.6%), like families with children, age, Self-targeting 4.28% 16 and disability. Methods based on combinations 18 and mixed methods represent 14.1% of the cases. Food security, nutrition 0.53% 2 Other methods are single digits22 (table 1). Combos 14.17% 53 18 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE e cash transfers response hinges on launching new programs. On average, 92% of programs and 48% of them were one-off transfers. New programs include three-quarters of cash transfer Fact 5. Most ponse hinges on launching new programs. relied responses on introducing On average, new programs through various 92% of programs identification and ns in LICs, and almost all those in HICs (figure 9). Such large scale of novel schemes speaks to re one-off transfers. New programs include three-quarters of cash transfer registration e limitations of preexisting mechanisms programs in meeting the scale of the crisis. st all those in HICs (figure 9). Such large scale of novel schemes speaks to xisting programs in meeting the scale of the crisis. ure 9. Horizontal expansion of cash transfers via new and existing programs (n=725) Most of the pansion of cash100 transfers via new transfers cashand response existing hinges programs on launching new programs. On average, 92% of programs (n=725) were new, and 48% of them were one-off transfers. New programs include three-quarters of cash transfer interventions 80 in LICs, and almost all those in HICs (figure 9). Such large scale of novel schemes speaks % out of total programs to the possible limitations of preexisting programs in meeting the scale of the crisis. Most 74.7 of the cash transfers response hinges on launching new programs. On average, 92% of programs 60 and 48% of them were one-off transfers. New programs include three-quarters of cash transfer were new, 90.6 74.7 94.1 97.9 interventions in LICs, and almost all those in HICs (figure 9). Such large scale of novel schemes speaks to 90.6 the possible limitations 94.1 of preexisting programs in meeting the scale of the crisis. 40 97.9 Figure 9. Horizontal expansion of cash transfers via new and existing programs (n=725) Figure 9. Horizontal expansion of cash transfers via new and existing programs (n=725) 20 100 25.3 9.4 2.1 25.3 5.9 0 80 LIC (=75) 9.4 2.1 UMIC (=236) HIC (=233) LMIC (=181) % out of total programs 5.9 LIC (=75) Existing LMIC (=181) UMIC (=236) programs New HIC (=233) programs 60 74.7 Existing programs New programs from Gentilini et al (2022a) 90.6 94.1 97.9 022a) cation of beneficiaries for old and new programs relied on a variety of methods, often involving 40 ns of mechanisms across 543 social protection responses in low- and middle-income countries es for old and new programs relied on a variety of methods, often involving et 2021). Before Covid-19, social registries reached about 21% of the population in low- and ross 543 social protection responses in low- and middle-income countries ome countries (Grosh et al 2022). As shown in figure 20 10, such databases and other program- id-19, social registries reached about 21% of the population 25.3 in low- and of beneficiaries were used for 117 programs during the pandemic, constituting 22% of h et al 2022). As shown in figure 10, such databases and other program- 2.1 on mechanisms (e.g., Brazil, Ecuador, Mauritania, Peru, and Pakistan). Social 23 9.4 security and 5.9 tax re used for 117 programs during the pandemic, constituting 0 22% of accounted for 27% of identification methods. The remaining programs used more half ofLMIC g., Brazil, Ecuador, Mauritania, Peru, and Pakistan). Social security and tax (=181) UMIC (=236) HIC (=233) LIC (=75) onal mechanisms, such as open registration and new enrolment campaigns. The latter of identification methods. The remaining half of programs used more Existing programs New programs the single largest share of identification methods (35%), hence signaling the need that such as open registration and new enrolment campaigns. The latter aced in expanding the identification Source: dataSource: frontier data from Gentilini from et of existing Gentilini al (2022a) options. et al (2022a) Civil registries were used in 7% share of identification methods (35%), hence signaling the need that s (e.g., Namibia and Togo). Registries of informal and self-employed registries were tapped in e identification frontier of existing options. Civil registries were used in 7% ions (6%), such as in urban TheSierra Leone and identification Sri of Lanka. beneficiaries for old and new Togo). Registries of informal and self-employed registries were tapped in programs relied on a variety of methods, often involving The urban Sierra Leone and combinations Sri Lanka. of beneficiaries for identification of mechanisms across old 543 new social and protection programs responses relied on a in low- and variety of middle-income methods, often countries Figure 10. Identification mechanisms (Hammad of social et 2021). Beforeprotection Covid-19,measures (n=543) reached about 21% of the population in low- and social registries ication mechanismsinvolving protectionof combinations middle-income of social mechanisms countries measures (Grosh (n=543)across 543 et al social 2022). As protection responses shown in figure in low- 10, such and middle-income databases and other program- level lists of countries (Hammad et 2021). Before beneficiaries were used Covid-19, social registries reached about 21% of the population in 22% of for 117 programs during the pandemic, constituting identification mechanisms23 (e.g., Brazil, Ecuador, Mauritania, Peru, and Pakistan). Social security and tax low- and middle-income databases accounted countries (Grosh for 27% of et shown in figure al 2022). As methods. identification 10, such databases The remaining and other half of programs used more unconventional mechanisms, such as open registration and new enrolment campaigns. The latter program-level lists of beneficiaries were used for 117 programs during the pandemic, constituting 22% constituted the single largest share of identification methods (35%), hence signaling the need that of identification countries mechanisms 23 faced in expanding (e.g., the Brazil, Ecuador, Mauritania, identification Peru, and frontier of existing Pakistan). options. Social security Civil registries were used in 7% of the cases (e.g., Namibia and Togo). Registries of informal and self-employed registries were tapped in and tax databases accounted for 27% of identification methods. some situations (6%), such as in urban Sierra Leone and Sri Lanka. 19 Figure 10. Identification mechanisms of social protection measures (n=543) remaining half of programs used more unconventional mechanisms, such as open registration and The19 new enrolment campaigns. The latter constituted the single largest share of identification methods (35%), hence signaling the need that countries faced in expanding the identification frontier of existing options. Civil registries were used in 7% of the cases (e.g., Namibia and Togo). Registries of informal and self-employed registries were tapped in some situations (6%), such as in urban Sierra Leone and Sri Lanka. 19 CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 19 Figure 10. Identification mechanisms of social protection measures (n=543) Social registry or Others (19) existing program- 3% level beneficiary databases (117) 22% Open registration/new enrolment campaign (191) 35% Social security or tax Informal database (148) workers/self- 27% employed registry (31) 6% Civil registry (ID, voting ID, birth certificate) (37) 7% Source: Hammad et al (2021) Source: AmongHammad et al (2021) registration mechanisms, online platforms were the most widely utilized. According to Hammad et al (2021), out of a sample of 516 social protection measures an average of 38% utilized online Among registration mechanisms, online platforms were the most widely utilized. According to Hammad registration (e.g., Brazil, Iran, Morocco, Kuwait, and Jordan). Such method was followed by automatic et al (2021), out of a sample of 516 social protection measures an average of 38% utilized online enrollment registration (24%) (e.g., andIran, Brazil, manual self-registration Morocco, Kuwait, and (18%). Some Jordan). countries Such method wastook also stepsby followed for making online automatic enrollment (24%) registration and manual more self-registration inclusive: for instance, (18%). Some Pakistan’s countries Ehsaas also took Emergency steps Cash for making Transfer onlinemade program registration more inclusive: for instance, Pakistan’s Ehsaas Emergency Cash Transfer program made the application messaging service free; Namibia allowed up to ten individuals to submit an application the application messaging service free; Namibia allowed up to ten individuals to submit an application using using a single SIM card; and hotlines and mixed strategies complemented online registration in Bhutan, a single SIM card; and hotlines and mixed strategies complemented online registration in Bhutan, Dominica Dominica andand Guatemala Guatemala (Hammad (Hammad et al et al 2021). 2021). FactFact 6. Most 6. Most cashcash transfers were paid transfers weredigitally paid digitally The crisis has witnessed a large share of transfers being paid digitally, i.e., via fully functional accounts, limited purpose accounts, or electronic non-account-based solutions. A total of 107 programs (96 of which The crisis has witnessed a large share of transfers being paid digitally, i.e., via fully functional accounts, in low- and middle-income countries) present data on both beneficiaries and payment modalities. In total, 763 millionpurpose limited accounts, people were or electronic reached non-account-based by digital cash transfers, or 62%solutions. A transfers of all cash total of 107 programs available(96 of with information (figure which in low- and11). of those digital Only 5 millioncountries) middle-income payments present data onoccurred in low income both beneficiaries and countries, while payment modalities. high income countries adopted some combination of both manual and digital payments (reaching In total, 763 million people were reached by digital cash transfers, or 62% of all cash transfers 162 with million people). available information (figure 11). Only 5 million of those digital payments occurred in low income countries, while high income Figure adopted countries by 11. Beneficiaries some payment combination modalities of both (n=107 manual and digital payments programs) (reaching 162 million people). A range of innovative practices have emerged. In Brazil, 14 million new accounts were opened via a 20 smartphone app by the public bank Caixa Economica; in Bangladesh, employment records were used to 20 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Figure 11. Beneficiaries by payment modalities (n=107 programs) 800 763 600 Million people 400 285 200 175 0 Digital Manual Combination Source: Gentilini et al (2022a) Source: Gentilini et al (2022a) open accounts of garment workers; in the Philippines, ID requirements were waived for those lacking range Athose andpractices of innovative documents; have emerged. Ecuador permitted In Brazil, 14 the involvement of million new accounts non-financial were actors such as opened via a grocery shops smartphone app by the public bank Caixa Economica ; in Bangladesh, employment records and pharmacies to act as ‘cash-in-cash-out’ agents (Hammad et al 2021). Changes in regulations were were used to open accounts of garment workers; in the Philippines, ID requirements were waived for those lacking also observed: the Central Bank of West African States allowed for onboarding via USSD text or phone those documents; and Ecuador permitted the involvement of non-financial actors such as grocery shops (such and as in the case pharmacies to actof asthe Democratic Republic ‘cash-in-cash-out’ agentsof Congo’s Solidarité (Hammad par et al 2021). Transferts Changes Economiques in regulations werecontre also la Pauvreté observed: theàCentral Kinshas STEP-KIN), a, or of Bank while West African Jordan’s States Central allowed for Bank relaxed onboarding KYC via requirements: USSD this text or phone help (such as in the boost case the of the share Democratic of NAF Republic cash transfer of Congo’s beneficiaries Solidarité paid par Transferts via mobile Economiques wallets from contre la 5% pre-pandemic to Pauvreté à Kinshasa , or STEP-KIN), while Jordan’s Central 68.5% in 2021 (Gentilini et al 2022a; Hammad et al 2021). Bank relaxed KYC requirements: this help boost the share of NAF cash transfer beneficiaries paid via mobile wallets from 5% pre-pandemic to 68.5% in 2021 (Gentilini et al 2022a; Hammad et al 2021). Digital payments yielded a number of benefits. For example, a survey of old age and widow allowances Digital payments of social pensions yielded a number of in Bangladesh benefits. found thatFor example, those a survey with access toof old age digital and payments cash widow allowances were moreof social pensions in Bangladesh found that those with access to digital cash payments were more likely to likely to receive transfers on a timely basis and had lower consumption and income risks (Shonchoy et receive transfers on a timely basis and had lower consumption and income risks (Shonchoy et al 2021). al 2021). However, access to digital payments can be hindered at several levels. For example, Gelb et al However, access to digital payments can be hindered at several levels. For example, Gelb et al (2022a) a survey a (2022a) examine examine survey of India’sof India’s program PMGKY PMGKY program transferring transferring benefits benefits to the accounts to the accounts of over 200of over 200 million beneficiaries. They found million beneficiaries. that among They found respondents that among withwith respondents specific characteristics specific (like characteristics being (like female, being female, illiterate, and illiterate, andliving livingininaahousehold household with nosmartphone), withno smartphone),thetheprobability probabilityof reportingthe ofreporting theuse useof ofdigital digital payments is only between 1 and payments is only between 1 and 3.9%. 3.9%. Therefore, it would be important to ensure flexibility in payment systems, including aligning payment models withit Therefore, contextual would be user capabilities important and payment to ensure ecosystems. flexibility in payment In this regard, systems, a typology including proposed aligning by payment Gelb et al (2022b) envisages four models: (i) direct manual payment of cash; (ii) an assisted models with contextual user capabilities and payment ecosystems. In this regard, a typology proposed bymodel, where multiple providers and instruments with assistance at the point of service may help reduce exclusion; (iii) Gelb et al (2022b) envisages four models: (i) direct manual payment of cash; (ii) an assisted model, where an assigned model, whereby selected providers with limited instruments to provide payments; and (iv) a multiple providers and instruments with assistance at the point of service may help reduce exclusion; choice model, where multiple providers and instruments offer beneficiaries choice over payment receipt (iii) an and useassigned model, (figure 12). whereby While selected in principle the providers choice model limited withoffers instruments the to provide widest utility, payments; countries and may follow different model, towards trajectories (iv) a choice such model where multiple basedand providers on prevailing instrumentsconditions. offer beneficiaries choice over payment receipt and use (figure 12). While in principle the choice model offers the widest utility, countries may Figure 12. Typology of payment models follow different trajectories towards such model based on prevailing conditions. 21 CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 21 Area by Digital and High Financial Coverage Assisted Choice Model Model Figure 12. Typology of payment models Cash/hybrid Assigned Model Model Low Low High Population with Digital Capacity Source: Gelb et al (2022b) Fact 7. Preexisting social protection delivery systems helped responses significantly, but they were not always a necessary nor sufficient condition for scale-up There is clear evidence that social protection delivery systems play a pivotal role in both crisis and “normal” times: program quality and performance heavily hinge on effective delivery building blocks (Grosh et al 2022); and their pre-existing presence is part of the larger quest for adaptive, “anticipatory” social assistance (Pople et al 2021; Bowen et al 2020). In general, the social protection responses to Covid-19 underscored “… the fundamental importance of information systems, which enabled identification of beneficiaries (e.g., social registries), and digital payments, that allowed the effective delivery of the transfers” (Stampini et al 2021, p.19). However, the pandemic presented the unique challenge of not only reaching people at the lower end of the distribution, but also those in the middle and beyond. This meant that, with social registries reaching about one-fifth of the population on average, existing delivery infrastructure may have helped to reach part of the population, but not the large swaths not enrolled in neither social assistance nor insurance. For example, while India’s initial response has been found to be “broad-based” in reaching households (Bhattacharya and Sinha Roy 2021), others like Bhan et al (2020) pointed out that in the earliest stages, “… [i]t was soon evident that reaching those already within the system was not enough” 24 (p.8). More generally, the degree to which pre-existing delivery systems were key for scale up depends how those systems are defined. A key difference, for instance, is whether such definition is strictly related to social protection, or if encompassing national capabilities more widely, including sectors other than 22 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE social protection as well as private sector capabilities. A range of countries were able to scale up – or even introduce – cash transfers without a preestablished social protection delivery building blocks in place (Kidd and Sibun 2020). In fact, figure 13 shows only a moderately significant correlation between the coverage of cash transfers and that of social registries among 48 countries. 25 Figure 13. Coverage of cash transfers and social registries (% population) Source: data from Gentilini et (2022a) Several countries that had relatively limited registry coverage achieved sizable coverage levels – e.g., notably Morocco and El Salvador, but also Bolivia and Mauritania. A review of responses in Latin America found that only 4 out of 15 countries has registries with high coverage and high interoperability (Berner and van Hemelryck 2020). As we will discuss in section 4, in the case of Togo the large scale up of cash transfers tapped the voter registry and an ad-hoc payment platform created during the pandemic (Aiken et al 2022; Lawson et al forthcoming). Similarly, in Kinshasa a registry was built real-time based on anonymized data from private mobile phone companies (Bance et al 2021). Others, like Pakistan and the Philippines, had large-scale registries and achieved high levels of cash coverage, although data wasn’t necessarily up-to-date and needed supplementation (Beazley et al 2021). In Peru, households already listed in the social registry began to receive their Bono Familiar Universal payments in May 2020; however, those who needed to be registered received the first payment only in August 2020 (Bastagli and Lowe 2021). Following Lowe et al (2021), registries “… provided a viable starting point for a fast first wave of payments to a sub-set of the affected population. However, (…) even a large-scale social registry cannot be relied upon as the sole basis for targeting crisis support, unless it is both universal and up to date, among other characteristics” (p.19). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 23 These considerations may speak to the salience of interpreting scale up within the broader delivery capabilities at country level. This would include recognizing the importance of all delivery building blocks, and not just some (e.g., payments or registries), the relevance of interoperable government databases across sectors, as well as considering both public and private sector infrastructure in delivery capacity assessments. Fact 8. The timeliness of transfers varied substantially How much time did it take to programs to reach beneficiaries? Gentilini et al (2022a) present data on program timeliness, that is, the number of days between announcement of the program and first payment. On average, data from 98 programs shows that coverage scale-up (horizontal extension) took 26 days, with timeliness ranging from 2 to 119 days. 26 The increase in benefits (vertical expansion) was Fact 8. The timeliness of transfers varied substantially faster, taking an average of 18 How days much time did it take between (varying to programs0to and 67 reach days). An important beneficiaries? Gentilini et al distinction pertains (2022a) present data on program timeliness, that is, the number of days between announcement of the program and first to whether expanded programs were preexisting or new: the extension of pre-existing programs was payment. On average, data from 98 programs shows that coverage scale-up (horizontal extension) took on average 42% faster than the 26 days, introduction with from programs of new timeliness ranging (18 2 to 119 days. 26 versus The 31in increase days) (figure benefits 14, (vertical left graph). expansion) was faster, taking an average of 18 days (varying between 0 and 67 days). An important distinction pertains to The larger role played by existing whether systems expanded programsin were lower-income preexisting or context reflects new: the extension of the attention pre-existing programsplaced was onon average maintaining and sustaining 42% delivery their faster than the introduction under of new programs challenging (18 versus conditions 31 days) et (Beazley (figure al. 14, left graph). 2021). As put The by larger role played by existing systems in lower-income context reflects the attention placed on Slater (2022) in the context of East maintaining Africa, and thetheir sustaining initial phases delivery underof the pandemic challenging focused conditions “… (Beazley et on al. adaptations 2021). As put by Slater (2022) in the context of East Africa, the initial phases of the pandemic focused “… on adaptations required to existing operations to ensure all those currently receiving support could continue to do so, required to existing operations to ensure all those currently receiving support could continue to do so, despite newand despite new access, monitoring funding access, constraints” monitoring (p.11). Relatedly, and funding constraints” thethe (p.11). Relatedly, use useof social of social registries registries was more than three times faster than identification mechanisms based on other databases, new enrollment, was more than three times faster than and combinations identification thereof mechanisms based on other databases, new (figure 14, right graph). enrollment, and combinations thereof (figure 14, right graph). Figure 14. Timeliness of cash transfers by new/existing programs (n=73) and by identification mechanism (n=98) Figure 14. 120 120 Timeliness of cash 119 100 102 102 transfers by new/ 100 existing programs 87 80 80 (n=73) and by Speed (days) identification 60 60 mechanism (n=98) 40 40 25 35 36 31 25 26 20 20 15 19 18 11 16 2 3 3 3 0 0 2 2 Existing New programs ) 4) 4) ) 7) 14 =5 programs (n=4) (n=69) =4 =2 n= n= (n (n (n t( t( ry en e s en st se pl llm gi am llm ba Re ro ta ls ro al en da ta en ci To w er So w ne th Ne O & es as ab at rd he Ot Source: data from Gentilini et al (2022). Note: squares indicate average values. 24 24 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE There is a range of reasons explaining the timeliness of new programs, some of which will be tackled in section 4 in the context of reaching informal sector workers in low- and middle-income countries. But bottlenecks were not unique to those contexts. The delivery of payments in high income countries shed light on a range of challenges also in those settings (Spasova et al 2022; Romer 2021; Moffitt and Ziliak 2020). For example, Japan was one of the 13 cases implementing a universal cash transfer scheme for nearly 125 million people (Ando et al 2020). Evidence shows that payment performance varied significantly due to within-country diversity in administrative capacity of local governments. While some payments started in May 2020, the majority of deposits occurred between late June and early July of that year – with some residual payment stretching until August (figure 15). According to a survey of 43 large municipalities, a key bottleneck was the amount of time needed handle numerous inquiries and check mailed envelopes (Kaneda et al 2021). The same study also laid out a counter- intuitive finding, namely that “… postal applications were significantly faster than online applications due to insufficient preparation” (ibid, p.414). Figure 15. Weekly distribution of universal cash transfer payments in Japan, May-August 2020 Figure 15. Weekly distribution of universal cash transfer payments in Japan, May-August 2020 0.12 0.10 Relative Frequency 0.08 0.06 0.04 0.02 0.00 4/May 1/Jun 29/Jun 27/Jul 24/Aug Week of Transfer Source: Kaneda et al (2021) The fact that delivery hurdles were underpinning countries across the income spectrum may, to some extent, be akin to Rodrik’s (2022) argument that “… developed countries’ problems increasingly resemble the problems found in poor countries”, and “… models and frameworks used to study developing economies are increasingly relevant to the problems confronting rich countries”. Covid-19 may have illuminated not only on clear differences between countries, but also similarities that countries across the income spectrum share on select social protection issues (e.g., supporting the “missing middle”, as discussed in section 4). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 25 Fact 9. Governments moved centerstage, but civil society actors played an important role Social assistance responses to crises in lower-income countries is typically bolstered by domestic and international assistance. A qualitative reading of the overall Covid experience indicates that governments embraced the crisis response in ways that are more visible and deliberate than in other circumstances. This is likely because of the role states played in managing the crisis (see section 2). For instance, there has been significant allocation of domestic financing for cash transfers in contexts where those programs are traditionally heavily donor-supported, such as Mauritania and Kenya (Zeufack et al 2022; Almenfi et al 2020). The uprise in state engagement is also visible in complex situations: Blattman et al (2020) documented that in Colombia’s Medellin city, community leaders reported greater welfare services rendered by state authorities (including police and municipal officials) than criminal gangs (which may hold a strong neighborhood-level presence as service providers). There is a vibrant literature documenting the ways in which international humanitarian assistance and domestic social protection could better converge, especially around cash transfers (Jodar et al 2020). The linkages between those agendas have received heighted attention during Covid-19 (Lawson-McDowall et al 2021). Such attention occurs against a backdrop of structural challenges (Levine et al 2022). For example, Development Initiatives (2021) estimates that only a fraction of global humanitarian assistance is channeled through national structures – i.e., $756 million or 3.1% of total assistance in 2020. 27 Such level is about 8 times lower than the 2016 World Humanitarian Summit target of 25% (figure 16). Recent guidance notes have further emphasized humanitarian linkages and coordination with national social protection (ECHO 2022; Grand Bargain 2022). Figure 16. Direct funding of local and national actors, % of total humanitarian assistance WHS 2016 TARGET (=25%) 25 20 15 10 5 3 3.6 3.1 2.8 2.1 0 2016 2017 2018 2019 2020 Source: data from Development Initiatives (2021) 26 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE The pandemic has accelerated further linkages in two directions (Hagen-Zanker and Both 2021; Smith 2021; Albaddawi et al 2020; Longhurst and Smith 2020). On one hand, governments have been extending social protection to typical “humanitarian beneficiaries”. This includes making refugees eligible for flagship pandemic responses in Argentina, Brazil, Cameroon (including public works), Colombia (displaced Venezuelans), Chile, Congo, Djibouti (vouchers), Panama, South Africa, and Trinidad and Tobago. 28 Furthermore, there are instances where state authorities utilized “humanitarian systems”: for example, Jordan tapped UNICEF’s RapidPro tool developed for the organization’s Hajati cash transfer for contacting, validating identity and confirming e-wallets presence for daily workers supported under the NAF social protection pandemic responses (figure 17). Two weeks after commencing the process, about 188,000 out of 200,000 beneficiaries had been identified and received their payments (Hammad et al 2021). Figure 17. To further confirm your identity RapidPro NEW MESSAGE! please reply with [NAME]’s National ID Number. system for RIGHT PERSON refugees RESPONSE leveraged in [NAME] This is a message from the Jordanian governement R Jordan’s NAF regarding requesting support E WRONG PERSON MATCH NO MATCH OTHER cash transfer for affected day labourers. S P program We need you to confirm your O identity. Are you [NAME] or live N OTHER You are confirmed to be eligible with them? S for emergency government E support. The payment will be Respond with 1 if yes or 2 if no. done through e-wallets. Please answer all questions to ensure you receive the assistance. • SMS alternative phone number • Phone call to troubleshoot Source: Albaddawi et al (2020) On the other hand, humanitarian assistance has been connecting with government social protection platforms deliberately, e.g., in terms of accessing government database (e.g., WFP in Cambodia and Colombia); aligning with national program parameters like transfer size, duration and payment (e.g., UNHCR in Peru, Turkey and Morocco). Cases were humanitarian agencies directly channeled cash as top- up to existing national schemes was observed in Armenia and Ethiopia, while participation in working group on cash and social protection emerged in Iraq and Yemen. The cases of Dominican Republic, Tajikistan and the Philippines offer further examples of joint assessments and options analysis. The pandemic also illuminated other aspects of the typical humanitarian process, including for example the wider empowerment accorded to local and domestic responders as a result of movement restrictions of international staff (Brubaker et al 2021). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 27 National and local civil society actors were also called for widespread engagement in direct provision of assistance (Gerard et al 2020). This occurred also in contexts of high social protection coverage: for example, during the early stages of the pandemic part of the South Africa’s response relied on civil society organizations for food and in-kind assistance (Gronbach et al 2022). In the Philippines, two rounds of high frequency monitoring surveys show that as the crisis unfolded, public assistance became less progressive (Cho and Johnson 2022). Data revealed that between March and December 2020, government support for the poorest quintile diminished by almost 7 percentage points, while help from informal networks of family and friends diminished relatively little. Remarkably, assistance from nongovernmental and faith-based organizations increased over time to seemingly fill gaps in public provisions29 (figure 18). Similarly, in the United States the use of food pantries (food banks or soup kitchens) operated by faith- based and other non-governmental organizations soared by 2.3 percentage points in 2020. The increase reached about 7 percentage points among very food insecure households, with nearly half (45.5%) accessing food pantries the first year of the pandemic (Coleman-Jensen and Rabbitt 2021). In France, food banks increased distributions by 25% in 2020, with an estimated 8 million people requiring such assistance that year compared to 5 million in 2019 (Legros 2021). Figure 18. Sources of assistance in the Philippines, % of poorest quintile Figure 18. Sources of assistance in the Philippines, % of poorest quintile 90 80 70 60 50 40 30 20 10 0 Government Friends and Family NGO/Church Wave 1 (since March) Wave 2 (since August) Source: Cho and Johnson (2022) Fact 10. Record-level global spending was highly uneven across countries, with financing based on a range of sources Data on social protection and labor program spending is available for 23% of programs. Over 2020-2021, countries invested a total of over $3 trillion in those interventions. Such level of expenditures represents around 18% of overall stimulus packages, and it is about 4.5 times higher than the estimated level of social protection spending during the Great Recession of 2008-09. 30 The United States accounts for 64% of global social protection spending (Gentilini et al 2022a). 28 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Average spending on social protection amounted to an average of 2% of GDP, three-quarters of which was devoted to social assistance (figure 19, panel a). Ravallion (2020) considered such level of social protection spending a minimum basis for response: “… [a]s a rough rule of thumb (…) a near-term fiscal injection of transfers less than 2% of GDP should be judged as inadequate”. Spending varied between 1.1% of GDP in MENA to 7.4% of GDP in North America. Interestingly, East Asia and Pacific present a level of investment that is over twice that of ECA. Similarly, there are stark differences when considering spending in per-capita terms: with a global average of $313/capita, high-income countries display a level of injection about 93 times higher than low-income countries – i.e., about $715 and $7.7 per capita, respectively; regionally, per capita spending ranges from about $27.5/capita in Sub-Saharan Africa to 4,130/capita in North America (figure 19, panel b). Figure 19. Social protection spending by region (average % of GDP and $/capita) 8% 8% 7% 7% 1.7% 1.7% 6% 6% PANEL A 5% 5% 1.4% 1.4% % of GDP 0.4% % of GDP 0.4% 4% 4% 0.6% 0.6% 3% 3% 0.3% 0.3% 4.3% 0.1% 2% 0.5% 0.3% 4.3% 0.1% 2% 0.1% 3.7% 3.7% 0.5% 0.3% 0.0% 0.2% 0.1% 0.3% 0.0% 0.2% 1% 0.1% 0.2% 0.3% 1% 0.0% 0.1% 0.2% 1.2% 0.1% 1.6% 1.6% 1.5% 1.0% 0.0% 1.0% 0.8% 1.2% 0.1% 0.7% 1.5% 0% 0.8% 0.7% 0% AFR (n=43) EAP (n=32) ECA (n=49) LAC (n=37) MENA (n=16) N. America SAR (n=8) Global AFR (n=43) EAP (n=32) ECA (n=49) LAC (n=37) MENA (n=16) N. America SAR (n=8) average (n=2) Global (n=2) average (n=187) (n=187) Social assistance Social insurance Active labor markets Social assistance Social insurance Active labor markets Panel b Panel b 6000 6000 5000 5000 4129.90 Average $/capita Average $/capita 4129.90 4000 Spedning ($/capita) 4000 ($/capita) 3000 PANEL B 3000 2000 Spedning 2000 687.04 1000 687.04 286.61 269.38 313.63 1000 172.89 45.74 313.63 286.61 269.38 27.53 0 172.89 45.74 EAP ECA LAC MNA North SAR 27.53 AFR Global 0 america America EAP ECA LAC MNA North SAR AFR Global Source: Gentilini et al (2022a) america Source: Gentilini et al (2022a) How are responses financed? Analysis by the IMF (2022) shows that between 2019 and 2021, the world’s Source: Gentilini et al (2022a) gross debt as percentage of GDP rose by about 13 percentage points. And an early contribution examining How are responses financed? Analysis by the IMF (2022) shows that between 2019 and 2021, the world’s sources of domestic financing for social protection responses found that debt and deficit spending gross debt as percentage of GDP rose by about 13 percentage points. And an early contribution examining constituted 38% of sampled cases (Almenfi et al 2020). Pre-pandemic analysis by Romer and Romer (2017) sources of domestic financing for social protection responses found that debt and deficit spending predicted that a country’s prior debt-to-GDP ratio had a large contractionary effect on the fiscal response constituted 38% of sampled cases (Almenfi et al 2020). Pre-pandemic analysis by Romer and Romer (2017) to a crisis, and vice-versa.30 This seems not have been the case during Covid-19, when countries debt- predicted that a country’s prior debt-to-GDP ratio had a large contractionary effect on the fiscal response fiscal response relationship was severed. Yet, in some cases political considerations may have curbed the to CASH TRANSFERS crisis, aIN and PANDEMIC appetite for vice-versa. TIMES: debt increases 30 This EVIDENCE, and notAND seemsimplications. PRACTICES, related have been the IMPLICATIONS In the case case ofduring FROM Covid-19, THE LARGEST Southern when SCALE European countries UP IN HISTORY debt- governments, for 29 fiscal response relationship was severed. Yet, in some cases political considerations may have curbed the example, Moreira et al (2021) argued that “… [o]ne possible interpretation of the initial lower immediate How are responses financed? Analysis by the IMF (2022) shows that between 2019 and 2021, the world’s gross debt as percentage of GDP rose by about 13 percentage points. And an early contribution examining sources of domestic financing for social protection responses found that debt and deficit spending constituted 38% of sampled cases (Almenfi et al 2020). Pre-pandemic analysis by Romer and Romer (2017) predicted that a country’s prior debt-to-GDP ratio had a large contractionary effect on the fiscal response to a crisis, and vice-versa. 31 This seems not have been the case during Covid-19, when countries debt- fiscal response relationship was severed. Yet, in some cases political considerations may have curbed the appetite for debt increases and related implications. In the case of Southern European governments, for example, Moreira et al (2021) argued that “… [o]ne possible interpretation of the initial lower immediate fiscal stimulus payments is that national governments treated any additional loan [from the European Central Bank] with caution because the stigma of fiscal surveillance remains politically toxic” (p.354). At more granular level, a comprehensive analysis by Hammad et al (2021) shows that domestic public fiscal was financing most prevalent the payments stimulus source is that national of funding, governments including treated 74% of responses any additional loan [from the (with the reminder European Central equally splitBank] with caution between because external the stigma assistance of fiscal (13%) and surveillance remains Domestic other sources). politically toxic” (p.354). financing occurred via transfers from the existing budgets (79.5% of the 524 recorded measures), budgetary reallocation (5.7%), At more granular level, a comprehensive analysis by Hammad et al (2021) shows that domestic public financingfunds contingency (3.4%), was the most and in some prevalent cases source taxation. of funding, These including were 74% of complemented by reminder responses (with the extrabudgetary equally split between external assistance mechanisms (6.8% of cases) (figure 20). (13%) and other sources). Domestic financing occurred via transfers from the existing budgets (79.5% of the 524 recorded measures), budgetary reallocation (5.7%), contingency funds (3.4%), and in some cases taxation. These were complemented by extrabudgetary mechanisms (6.8% of cases) (figure 20). Illustrative Figure 20.Figure Covid-19 20. Illustrative budgetary Covid-19 and extrabudgetary and budgetary financing extrabudgetary financing forprotection social protection for social BUDGETARY FUNDS (select practices) EXTRA-BUDGETARY FUNDS Budget reallocation Off-budget, using Off-budget (e.g., Albania) On-budget regular PFM channels - Public (e.g., Sierra - Public (e.g., Zambia) - Public (e.g., Colombia) Leone) Contingency funds - Mix (e.g., Mauritania) - Private (e.g., Ghana) - Private (e.g., Jordan) (e.g., Philippines) - Mix (e.g., Togo) - Mix (e.g., Kenya) Taxation (e.g., Uruguay) Source: adapted from Hammad et al (2021) and Rahim et al (2020) Source: adapted from Hammad et al (2021) and Rahim et al (2020) In Timor-Leste, social protection was in part funded by reallocating resources from the national Petroleum Fund; Bolivia redirected financing for fuel subsidies toward pandemic responses; in Ukraine, the repurposing of funds for social protection affected the planned census, among others; Kenya’s initial In Timor-Leste, social protection was in part funded by reallocating resources from the national Petroleum rollout of Kazi Mtaani – a public works program for youth in informal settlements – was financed by Bolivia redirected Fund; redirecting financing infrastructure sector for subsidies fuel for funds toward pandemic urban development; responses; and Albania in Ukraine, reallocated the repurposing $18 million of of fundsdefense spending for social for, interaffected alia, unemployment protection the planned benefits census,and social assistance among (Duran-Valverde others; Kenya’s et al 2020). initial rollout of Kazi Mtaani Liberia and Uruguay used earmarked taxes on public sector wages as a source for their Covid-19 funds – a public works (Hammad 2021). for youth in informal settlements – was financed by redirecting infrastructure program et al sector funds for urban development; and Albania reallocated $18 million of defense spending for, inter Contingency financing alia, unemployment was another benefits financing and social modality. At assistance the core of this agenda (Duran-Valverde is how et al to strike 2020). a balance Liberia and Uruguay of ex ante and post-shock risk financing instruments for a given country risk profile (figure 21). Such earmarked used approach hastaxes on public advanced in areas sector at highwages as climatic a source risk for their (particularly Covid-19 droughts funds (Hammad and typhons), et al 2021). and recently receive further impetus under the Covid crisis (Calcutt et al 2021; Longhurst et al 2021; World Bank 2021d; Hill et al 2020). The basic tenet of the agenda is that risks with a certain probability and occurrence may be tackled more effectively and efficiently with anticipatory action (Pople et al 2021). This includes a 30 firmer grounding of scale up decisions on early warning systems. III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Figure 21. Risk financing instruments Contingency financing was another financing modality. At the core of this agenda is how to strike a balance of ex ante and post-shock risk financing instruments for a given country risk profile (figure 21). Such approach has advanced in areas at high climatic risk (particularly droughts and typhons), and recently receive further impetus under the Covid crisis (Calcutt et al 2021; Longhurst et al 2021; World Bank 2021d; Hill et al 2020). The basic tenet of the agenda is that risks with a certain probability and occurrence may be tackled more effectively and efficiently with anticipatory action (Pople et al 2021). This includes a firmer grounding of scale up decisions on early warning systems. Figure 21. Risk financing instruments Occurence of Event e.g., Danish Red Cross (10 Volcanoes eruption risk) Ex-ante instrument Ex-post instrument (arranged before disaster) (arranged after disaster) e.g., Me(crop and livestock), Africa ARC Capital markets e.g. catastrophe bonds Ois.cretionary post·disaster reliefe.g. Humanitarian aid, Parametric insurance Crisis response grants/loans e.g., Humanitarian CERF Risk Pools from development banks Traditional re/insurance e.g., IMF Rapid Credit Facility and Rapid e.g., WBG IDA CAT-DOS Financing Instrument; Contingent credit Tax increases WBGCRW e.g., Philippines Reserve fund Post-disaster credit (NDRRMF); Peru (4 funds) Budget reallocation e.g., Jordan, Ukraine Budget allocation Time Source: adapted from Longhurst et al (2021) Those principles have bene put in practice in the Covid response, albeit at a small scale. In the Philippines, lockdowns on the Luzon island in the Philippines triggered the use of nearly $26 million from its Quick Response Fund for inkind provisions. In Sierra Leone, the Ep Fet Po (EFP) cash transfer program was introduced in 2014. The program supported 12,000 households with quarterly payments of $45 via mobile money combined with distribution points based on QR codes. Based on the Ebola experience as well as flooding of Freetown in 2017, the EFP was reformed in 2019 for pre-arranging external financing for emergency response. This involved a risk layered approach and a contingency budget of $4 million that was activated in 2020 enabling faster disbursement (Sandford et al 2020). Extrabudgetary funds were used to finance 36 programs in 18 countries. These present different levels of alignment with budgetary outlays. At one extreme are fully off-budget mechanisms running in parallel to national budgets. While separate from budgets, those models tend to come with strong oversight, transparency, and accountability measures: for example, Sierra Leone’s off-budget fund was established in an Emergency Operations Center under the President. Jordan’s Himmat Watan Fund was financed by public actors (e.g., Bank of Jordan), private sector and individual donations. The fund financed the national emergency cash transfer for daily wage workers, including covering the first payment in full and about 88% of the cost toward the end of the program (12% being funded by the Treasury). In Kenya, a weekly CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 31 cash transfer was implemented for 6 months with funds from the COVID-19 Emergency Response Fund. In some countries with significant external assistance, like Uganda, pandemic funds were supported by donors to ringfence contributions for Covid-19 relief and lower fiduciary risks. An off-budget practice that used routine public financial management channels is Colombia’s Fund for the Mitigation of Emergencies financed via public debt bonds for $3.1 billion. In Togo, a dedicated portal tracked payments made under the Tovissi cash transfer program on a daily basis. Lastly, Mauritania’s National Solidarity Fund offers an example of an on-budget funding fully integrated into country’s public financial management structures (Almenfi et al 2020). In mid-2020, about three-quarters of the fund’s $125 million were devoted to subsidies and utility fee waivers, as well as partly supporting the Taazour cash transfer scheme. Private donations were also an important part of the financing mosaic. For example, Zakat-related contributions accounted for at least 11 measures. Hammad et al (2021) report that the Zakat fund in Jordan funded 13.6% of its in-kind transfers; the equivalent fund in Egypt financed monthly cash transfers of $110 for 30,000 informal workers; the Kuwait zakat fund contributed in assisting over 312,000 stateless individuals and migrants with cash and in-kind transfers; and in Indonesia, “… the central Zakat Agency (BAZNAS) accepted zakat payments through QR codes, e-wallets and bitcoin (…) and created ZakatHub, a common crowdfunding platform to raise zakat to be used by the agency’s regional committees (…). Regional zakat collection and distribution agencies set up a COVID-19 crisis centre to synergise the response and implement a number of social assistance programmes across all regions. One of them was a cash for work programme in which the agency contracted daily wage workers to disinfect public spaces” (ibid, p.60). 3.2 Cash transfers in practice32 One tool, many applications Cross-country practices reveal the multiple objectives pursued by cash assistance, and the diverse population profiles served during the crisis. For example, one area where governments used cash transfers was to offset the competing demands of work and childcare. Eighty-two countries, such as Argentina and Armenia, have provided childcare benefits to parents with children. Myanmar and others 16 countries have implemented a one-off targeted cash transfers to pregnant women or with children below 3 years of age. Malta introduced childcare benefits for private sector employees who took unpaid leave for childcare. Similarly, Poland provided additional 14 days of childcare allowance to parents with children below 8 years of age. Other 12 governments provided universal childcare allowances. For example, Iceland announced a one-off childcare payment to all families with children below 18: parents with an average monthly income below ISK 927,000 in 2019 received ISK 40,000 (about $300) per child, while those with higher income received half of such amount. 32 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Cash assistance was also provided in connection to stay-at-home orders. Government provided cash assistance to people living in lockdown-imposed areas or regions, e.g., in Azerbaijan, Fiji and other 6 countries. In Malaysia, for example, the Federal Territories Islamic Religious Council provided advance payments and a one-off top-up to its monthly beneficiaries in response to lockdown measures. Frontline workers received special cash transfer payments in 36 countries (either as one-off or monthly top-ups). For example, in April 2020 Brunei announced a monthly special allowance of B$400 (nearly $290) for doctors, nurses, paramedics, ambulance drivers, and hospital cleaners and security guards. The Seychelles provided additional transfers to their immigration, customs, and related workers. Similarly, countries like Peru, Timor-Leste, and Tuvalu provided payments to their military, police, and civil defence forces. Other economies provided bonuses to their civil servants (e.g., Cayman Islands, France, Malaysia, and Zimbabwe). Cash transfers were also provided on a more traditional needs-based basis. Fifty-eight countries delivered cash assistance to poor and vulnerable people. Belize, for example, implemented a temporary cash transfer to low-income families who were not covered by any other social assistance program. Similarly, countries made payments to other vulnerable groups such as people with disabilities (55 countries), single parents (13), and the elderly (31). These also include 12 cases where transfers made to foster families or custodians that took care of disabled people, orphan children and seniors. Governments also provided additional assistance to people with a higher risk of Covid-19 due to comorbidities. This also involved governments offering home delivery of cash to such populations (e.g., Grenada). Other special transfers were made to people affected by HIV or tuberculosis (e.g., Myanmar, St Lucia, and Tajikistan), chronic kidney (e.g., Sri Lanka), and various other diseases (e.g., Cuba). A total of 11 countries (e.g., Argentina, Eswatini, France, Gambia, Iran, Madagascar, Uganda) have implemented some form of Covid-related conditional cash transfers based on absence of domestic violence during quarantine (e.g., Colombia); transfers connected to human capital development (e.g., Congo); or linked to training programs (e.g., Benin, Myanmar). Conditionalities were sometimes tied to benefit temporal extensions: for example, the Cook Island’s Emergency Hardship Fund had extended the eligibility period for support by 3 months, with an option for further extension by 3 months conditioned on recipients participating in volunteer work or community service for at least 5 hours/week. A few countries, such as Kiribati, Mongolia and 7 others, helped stranded citizens living abroad. Assistance was also provided domestically: China, India and 6 others had taken active steps to disburse cash to affected mobile populations. For instance, India’s Bihar state transferred money to bank accounts of workers stranded in other states due to lockdowns (Mukherjee 2020). Finally, governments at federal and local levels incentivized the uptake of vaccinations through a range of creative programs33 (box 3). Also, countries provided financial support to people who were infected by Covid-19: the Philippines and other 9 countries compensated people who were tested positive for Covid-19 while on duty. India’s Delhi government provided Rs5,000 (about $65) per person (and maximum of Rs10,000 CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 33 per family) to Covid-positive construction workers. Similarly, Australia and 7 other countries assisted those who were asked to self-isolate. For example, UK’s Test and Trace Support Payment Scheme provided a cash assistance of £500 (nearly $613) to low-income, self-isolating individuals. Finally, 13 governments provided cash transfers to support families with the expenses related to funerals and last rites (e.g., Bangladesh). Box 3. “Cash for jabs” Governments have provided cash assistance in a variety of forms to incentivize people to take-up vaccination. A total of 55 cash-for-jab programs were recorded in 6 economies, namely Cambodia, Japan, Russia, Serbia, United States, and Virgin Islands. Among the schemes, organizing lotteries and cash prizes were the most widely used mechanism, including 26 programs in 3 countries. For example, New York State announced in May 2021 a new vaccination program that provided free lottery scratch-off tickets to individuals aged 18 and above with a grand prize of $5 million. While 18 US states had some form of lottery, others like Arkansas and New Mexico offered $100 incentive programs. The town of Miyashiro, in Japan’s Saitama prefecture, passed a supplementary budget that included a plan to distribute JY 1,000 coupons (about $7.5) to all residents eligible for vaccination. Serbia offered about $30 to all those who got their Covid-19 vaccine by May 31, 2021. Other methods to offer financial nudges include providing university scholarships (e.g., New York, Ohio, West Virginia); gift cards (e.g., California, Maine, North Carolina, Michigan); $100 saving bonds (e.g., West Virginia); and discount coupons (e.g., Japan). Source: Gentilini et al (2022a) Implementation across delivery building blocks Programs were implemented in a variety of ways during the crisis. This section briefly reviews those arrangements following a typical delivery systems framework (Lindert et al 2020). In terms of outreach, programs were publicly announced through media and community to promote the awareness among potential beneficiaries (e.g., Pakistan and Bhutan). For example, Bhutan’s Druk Gyalpo’s Relief Kidu cash program was provided widespread publicity though television, radio, press and social media. Similarly, Colombia had YouTube videos explaining the Ingreso Solidario program. 34 For intake and registration, various modalities were pursued, often as complementary strategies within the same country. Those options included registration via digital and web portals (e.g., Bhutan, Cape Verde, Czech Republic, Guam, Honduras); via mobile platform (e.g., SMS in Brazil, Pakistan, India, Togo); manual registration at public local offices (e.g., Jordan, South Korea, Ukraine); manual registration via outreach/enrolment provided in mobile vans or similar initiatives (e.g., Algeria, Côte d’Ivoire, Kenya, Sudan). 34 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Some countries have retained the need for verification (in the form of home visit or inviting the beneficiary to the government office) for first-time applicants. This included Cook Islands, Kosovo, and San Marino. Others, instead, simplified need assessments for first-time applicants. In case of Jordan (Takaful-3), for instance, verification of new applicants by a social worker was conducted virtually. For existing beneficiaries, Kosovo and other 12 countries simplified the process by suspending all the re-application and verification requirements – such as to visit government office, fill out the form, etc – thereby allowing to automatically extend cash for existing beneficiaries whose benefits expired. Kosovo, for example, had suspended the verification process for social assistance (and pensions) benefits of re-applicants. However, new applicants were required to complete full application process. Ukraine automated the verification of information that the applicant indicated in the application: the time taken to complete the registration process through a software was less than 5 minutes. Countries enriched their social registries with information collected during the pandemic. For example, in addition to census data and local government data, Colombia’s Bogotá Solidaria en Casa used data from cell phone operators to identify its potential beneficiaries. Ghana, Indonesia, Mali, Mauritania, and Nigeria integrated information collected during Covid -19 into their social registry. In Indonesia, for example, beneficiary data collected from Bantuan Sosial Tunai (an emergency unconditional cash transfer program) has been integrated into the social registry. Brazil created the ExtraCad platform combining different information sources and databases. Some other governments (such as El Salvador and Guatemala) used other proxies of welfare – e.g., electricity consumption – to identify beneficiaries. For example, Guatemala introduced the emergency cash transfer Bono Familia to reach 2.7 million households (80% of population) for a duration of 3 months: in such case, 90% of beneficiaries were targeted based on electricity consumption less than 200whz for areas with proper electricity access; the remaining 10% of beneficiaries were selected based on socio- economic characteristics. Uzbekistan increased the age limit of the child grant from 14 to 18 years old. Similarly, 7 countries relaxed the eligibility threshold (by disregarding certain incomes and categories, or increasing the threshold) for registrants to receive allowances. For example, Belize’ BOOST cash transfer program relaxed the eligibility threshold to include vulnerable households above the poverty line. And North Macedonia relaxed the eligibility criteria of guaranteed minimum income scheme for materially deprived households, allowing the program to expand horizontally to 15,500 new households from the informal economy and 30,000 existing beneficiaries. Also, governments have suspended or reduced the eligibility proof requirements (such as reporting of income, wealth, etc.) for their cash assistance. For example, in Germany parents who lost income due to the pandemic could receive easier access to child grant (Kinderzuschlag) with simplified eligibility requirement of one-month income proof instead of 6 months. Similarly, the Kyrgyz Republic’s suspended the requirement to report income for low-income families and extended the standard one-year enrolment term for the poverty-oriented cash transfer. France waived the quarterly declaration of resources for its guaranteed minimum income. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 35 Some countries also enrolled people who were previously on waiting list of social assistance program. For example, West Bank and Gaza enrolled 10,000 new families who were previously on the waiting list of the Cash Transfer Programme. Some other countries have targeted and enrolled previously unsuccessful applicants. Albania, for example, provided cash assistance of about $140 per unsuccessful applicants of the Ndihma Ekonomike program since July 2019 (and who did not receive any assistance until April 2020). In other cases, e.g., Colombia’s Familias en Acción, programs have re-enrolled registrants who previously failed to comply with conditionalities. Such approach in Colombia has, according to Alvarez-Iglesias et al (2021), presented an opportunity to “… understand how to reduce barriers experienced by vulnerable families and young people when attempting to access support from conditional cash transfers”. Belize and 6 others have used SMS to notify regarding the result of their application and to inform about the onboarding process. The Kuwaiti Zakat House/Fund sent text messages to all registered families to update their private data through an attached link. Similarly, Egypt notified accepted beneficiaries via a SMS regarding the location and time to collect their first payment and ATM card. This was done to avoid overcrowding, and compounds other Covid protocols established at payment locations. Other Covid- sensitive practices are laid out in box 4. Cartagena, Bolívar/Colombia; 06/06/2020: Retired men are queuing up and keeping a distance to enter a local city bank in times of pandemic. by Luz Zuluaga Photography 36 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Box 4. Social distancing protocols in cash transfers In the case of Japan’s special cash transfer, beneficiaries were encouraged to fill application online and in person applications were only restricted to only special case in a attempt to avoid the spread of the virus. Also, in Philippines, “Assistance to Individuals in Crisis Situation (AICS)” program processed only limited application every day in order to adherence with the social distancing measures. Egypt sent SMS regarding payment collection details such as location and time of visit to avoid overcrowding and ensure the safety of beneficiaries. WFP, in Egypt, increased the number of distribution stations to avoid overcrowding. Inua Jamii beneficiaries, in Kenya, were required to maintain 1.5 meters of social distance, while standing in queue at the payment centers. Yemen’s Emergency Cash Transfer beneficiaries were requested to cover their nose and mouth, and maintain social distancing while queuing inside the payment sites. They were also requested to use the provided hand sanitizer before entering and after collecting their cash. Likewise, all project staff were using protective gear (such as masks and gloves). Finally, strong scheduling and crowd management mechanisms were put in place to prevent crowding at the payment sites. Montenegro’s Material security program processed the payments through post office and home deliveries, to enforce social distancing. Similarly, Grenada’s SEED program provided cash assistance delivery at home for elderly while rest of them had to collect it from allotted payment stations. Some countries such as India and Latvia used a unique way of dispersing money to its beneficiaries. For example, to maintain social distancing, India’s Pradhan Mantri Jan Dhan Yojana (PMJDY) program used last digit of the account numbers to provide the delivery of the cash transfer. For account numbers with last digit as 0 and 1 will get the money in their account on 3 April 2020, 2 and 3 will get on 4 April 2020 and so on. Similarly, Latvia’s childcare transfer dates for each child varied as the pay-out took place gradually according to the child’s year of birth, starting with the oldest children. Source: Gentilini et al (2022a) Program design is largely “back to normal” Basic design of Covid cash transfers has been “lean and simple”, including weaving conditionalities and work requirements. As of late 2021, a number of countries have reinstated pre-Covid design features. This includes the reactivation of conditionalities, like in Indonesia (Family Development Sessions for the PKH program), Dominican Republic (soft conditions for food component of Superate program), Brazil, Chile, Hong Kong, Norway, Panama, and Tanzania. Work requirements have also resumed in public works schemes in Rwanda, Ethiopia and, again, Tanzania. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 37 Such “back to normal” is also visible in core trends in coverage and adequacy. Brazil is a case in point: the national Auxilio Emergencial (AE) program gradually reduced provisions in three phases (figure 22). Out of the initial AE-1 scale up reaching 68.3 million individuals (10.5 million via Cadastro Unico, 19.5 million as Bolsa beneficiaries, and 38.2 million via mobile apps), gradually more stringent criteria were established. To enable cross checks, the number of databases tapped increased from 17 in AE-1 to 42 in AE-3. Overall, this led to reduction by 42% in coverage and 58% in benefit levels in the 18 months spanning from April 2020 to October 2021. Figure 22. Evolution of Auxilio Emergencial coverage and adequacy in Brazil, 2020-2021 Figure 22. Evolution of Auxilio Emergencial coverage and adequacy in Brazil, 2020-2021 R$ 600.00 (=$115) 70.00 R$ 600.00 60.00 R$ 500.00 50.00 R$ 400.00 40.00 R$ 300.00 68.3 R$ 250.00 R$ 300.00 30.00 56.8 (=$47) R$ 200.00 20.00 39.4 10.00 R$ 100.00 0.0 R$ 0.00 AE 1 (April – Aug 2020) AE 2 (Sept – Dec 2020) AE 3 (Apr– Oct 2021) Coverage (M) Coverage (M) Source: data from Arruda et al (2021) Similarly, Australia’s Corona Virus Supplement gradually reduced the benefit amount. In March 2020, Australia announced a six-month Coronavirus Supplement to be paid at a rate of AUD550 (about $380) on a fortnightly basis. The transfer was directed to recipients of the jobseekers payment, parenting payments, youth allowances and other payment types. The cost was AUD 14.1 billion ($10 billion). In July 2020, the program was extended until December 2020, with payments reduced to AUD 250 ($173). On December 10, 2020, the government further extended the duration of the payment until March 2021, but reduced the payment to AUD 150 ($104) (Gentilini et al 2022). Early signs of a Covid operational legacy As the wave of country Covid-19 responses recede, an operational “Covid legacy” emerge. For instance, in Brazil the crisis sealed the permanent migration from debit cards to digital bank accounts for recipients of AE as well as increasing cash benefit size (relative to pre-2020) on a more permanent basis. Others, like Iceland, New Zealand, and Moldova have either made such increase permanent or are planning to do so. Similarly, there is a growing cadre of programs that leveraged the pandemic to enhance routine social protection delivery systems (figure 23). 35 38 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Figure 23. Covid-induced innovations in delivery systems, select countries ASSESS ENROLL PROVIDE MANAGE Eligibility Determination Provision of Beneficiaries Exit Decisions, Assessment and of Benefits Notification Benefits Compliance, Notifications Intake & of Needs and Enrollment and Service and and/or Updating and and Case Outreach Registration Conditions Decisions Package Onboarding Services Grievances Outcomes RECURRING 1 2 3 4 5 6 7 CYCLE 8 9 Covid- Online Digital Enriched Interoperability of Higher Institutionalizing Built-in applications registration social registry registry with other benefit digital payments contingency induced (Kosovo) (Honduras) data with databases size in fiull or in part transfers in case permanent information (ongoing in DR) (Brazil) (Ukraine, Togo, of crises changes collected in Rwanda, Egypt, (DR; UCT window the pandemic Change in legislation (North Macedonia Brazil, Honduras) in Ethiopia UPSNP) (Indonesia, changing/relaxing eligibility criteria Mali, Nigeria) when emergencies are declared) Source: author based on internal survey of WBG task team leaders in late 2021; framework is from Lindert et al (2020) Upgrades include, among others, online applications in Kosovo, digital registration in Honduras and digital payments in Rwanda and Egypt; newly collected data in Covid times enriching pre-existing registries, such as in Nigeria; enhanced interoperability of social registries, e.g., Dominican Republic; legislative provisions activating the relaxation of eligibility criteria in crisis, like in North Macedonia; and advancements in using built-in contingency transfers in emergencies, e.g., Ethiopia. In Lithuania, responses featured higher amounts of monthly social assistance benefits and easing of eligibility criteria: as reported by Aidukaite et al (2021), such measures “… will remain in force after the COVID-19 crisis” (p.365). And in New Zealand, if an employee must self-isolate because of Covid-19 and can’t work from home during that period, the leave support scheme provides the person with cash transfers for $359-600/week. 36 Glen Innes, Auckland / New Zealand - March 24 2020: People queue outside discount supermarket Pak’nSave two days before the coronavirus covid-19 self isolation level four shut down of New Zealand. By Glen Innes. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 39 Cape Town, South Africa - August 2020: Covid Relief Funds in South Africa. long lines for people waiting for money to feed themselves. Poverty in Africa. People waiting hours for 20 Dollars. By Chadolfski 3.3 Effects of social assistance responses Welfare effects A growing volume of assessments is examining the extent to which social protection has been effective in mitigating the pandemic’s detrimental welfare consequences. 37 Table x presents findings from a compilation of over 30 cross-country and country-specific studies. 38 Impacts are reported for health as well as other welfare dimensions, particularly food security. 39 An indicative typology of performance is offered along three basic categories: social assistance responses that mitigated most or even the entire negative impact of the pandemic (green cells); responses that halted the negative effects of the pandemic significantly, but not sufficiently to offset most of the damage (in blue); and social assistance that resulted in modest or limited mitigation effects (grey cases). Most studies found that responses helped significantly, but were enough to offset the pandemic’s effects. This includes 11 out of 22 blue cells of table X.40 Table 2. Summary effects of social assistance responses on welfare WELFARE DIMENSION: HEALTH Countries Indicative Summary effects Source effects of social assistance 43 countries A standard deviation difference in poverty leads to +25% Aminjonov et al 2021 Covid cases. With cash transfers, however, the rate is only 11% increase 178 countries -21.8–47.7 pp Covid growth rate; -17.1–29.7 pp Covid mortality Asfaw 2021 growth rate; -3.69M Covid cases; 166,690 lives saved (-4.4–8.29 pp in workplace mobility; -6.6–11.6 pp in job search effort levels) Brazil Auxilio Emergencial (AE) ($155 x 5) reduced the chance of de Leon et al 2021 reporting Covid by 15% (by reducing work between 2.7 to 3.6 hours/week) Togo Improved access to health care services for the most Tossou 2021 vulnerable households 40 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE WELFARE DIMENSION: HEALTH: CONSUMPTION, FOOD SECURITY, POVERTY, INEQUALITY Countries Indicative Summary effects Source effects of social assistance South Africa Poverty reduced between 56% (food poverty) and 23% (upper- Barnes et al 2021; bound poverty line); performance improved over time Gronbach et al 2022 Brazil Offset and reversed poverty (it would have increased by 5.6M Lustig et al 2020; Blofield in the absence of cash; response of 3.3% of GDP); benefits with et al 2021a,b,2020; IMF relatively high breadth and sufficiency. 2022; Neri 2021 AE reduced poverty from 23 million people in 2019 to 9.8 million in September 2020. Chile Benefits with relatively high breadth and sufficiency Blofield et al 2021a,b Ethiopia PSNP largely mitigated food insecurity (from + 11.7pp to 2.4pp; size Abay et al 2020; food gap from 0.47 months to 0.13 months); response generally Duchoslav and Hirvonen progressive in rural areas 2021 Argentina With a response of 1.48% GDP, poverty increased by 0.6M people, Lustig et al 2020; Blofield but it would have been 1.6M without cash et al 2021a,b,2020 United States Pandemic EBT reduced the share of families in SNAP households Bauer et al 2021 where children experienced very low food security by 17%; reduced food insufficiency among SNAP households by 28% India In India’s Bihar state, households that received cash transfers had Makkar et al 2022 a lower proportion of food insecurity (43.6%) compared with those who did not (49.7%) Armenia, Cambodia, Cash transfers more effective than in-kind transfers; but overall, Dasgupta and Robinson Chad, Djibouti, analysis of High Frequency Phone Surveys found “… food insecurity 2022 Ethiopia, Kenya, that government safety nets have not fully mitigated” Malawi, Mali, Nigeria, South Africa, Uganda Colombia Ingreso Solidario had limited overall effects on food consumption; Vera-Cossio et al 2020 but among households that experienced large labor market shocks, it mitigated more than 50% of drops in food consumption South Africa Special Covid grant reduced poverty by 2pp (till 5.3%) and Kohler and Bhorat (2021) inequality by 1.3-6.3% Costa Rica, Sufficiency of benefits varies from medium to low sufficiency, often Blofield et al 2021a,b Uruguay, Peru, with differences between programs Bolivia Kenya Beneficiaries were 4.9-10.8pp less likely to report experiencing Banerjee et al 2020 hunger during the last 30 days Uruguay Positive effects, but limited in size Brum and De Rosa 2021 Colombia Response of 0.87% of GDP, poverty increased by 2.3M with cash Lustig et al 2020; Blofield (instead of 2.5M without it et al 2021a,b, 2020 Bolivia Enhanced probability that households had a week's worth of food Bottan et al 2020 stocked by 25%; reduced probability of going hungry by 40% India Cash transfers decreased moderate food insecurity by 2.4% and Kumar et al 2022; severe food insecurity by about 0.92%; NREGA program only Narayanan et al 2022 provided 13.5 days per rural household Ghana, Limited mitigation effects (simulations based on tax-benefit Adu-Ababio et al 2021; Mozambique, models) Cardoso et al 2021; Jara Ecuador, Zambia, et al 2021; Kalikeka et al Tanzania, Uganda 2021; Kisanga et al 2021; Lastunen 2021 Malawi, Liberia Modest effects on consumption and dietary diversity Aggarwall et al 2020 Mexico No significant social assistance response (9M additional people Lustig et al 2020; Blofield fell into poverty) et al 2021a,b, 2020 Source: compiled by author CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 41 Importantly, as benefits were short term and mostly phased out in 2021, welfare effects may follow a U-shaped trajectory. In the case of Brazil’s example highlighted in the earlier section, Auxilio Emergencial played a critical role in reducing poverty from 23 million people in 2019 to 9.8 million in September 2020. But as support receded, Figure reemerged poverty of 18. Sources in early assistance 2021 in the affecting nearly Philippines, % of 27.7 millionquintile people (figure 24) poorest (IMF 2020; Neri 2021). 35 without cash transfers Figure 24. Poverty in Brazil with and 30 25 Poverty headcount ratio (%) Based on labor income only (quarterly) 20 Based on labor income only (monthly) 15 Based on labor income with EA 10 Based on labor income with EA and other transfers 5 0 2019: 19: 19: 19: May June July Aug. Sep. Oct. Nov. 21: Q1 Q2 Q3 Q4 2020 20 20 20 20 20 20 Q3 Source: IMF (2022); AE = Auxilio Emergencial Analysis of cost effectiveness of interventions fell outside the remit of most evaluations. This stems from the particular challenges that cost identification and attribution may pose in emergency contexts, as well as the priority accorded to speed as opposed to efficiency that may characterize early phases of response. Future work may shed light on options for optimizing cost-effectiveness. This may involve the discussion on anticipatory transfers and automatic stabilizers, which the paper will examine in section 5. Economic effects The economic multipliers of cash transfers have received renewed attention in the literature. Before the pandemic hit, emerging evidence shows that the local economic effects of cash transfers range between $0.8 and $2.6 per dollar injected.41 Such increase in interest has grown in tandem with the adoption of unconventional monetary policies in times of economic downturns. From this perspective cash transfers have been part of macro debates whereby governments, Central Banks, and other macro authorities identify channels to increase liquidity in the economy and stimulate demand. In other words, during Covid-19 debates on cash transfers were, particularly in high-income countries, often located at the intersection of fiscal and monetary policy. 42 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE Conventional monetary policy instruments generally include open market operations, adjustments of interest rates, and reserve requirements. But when interest rates are stuck at the lower bound, authorities may lead to consider more radical, unconventional options (Furman 2020; Christiano et al 2011). Such unconventional policies include indirect and direct measures: the former involves “quantitative easing” (QE), under which central banks buy government bonds or other financial assets; the latter concerns the direct provision of cash not to intermediary financial institutions, but to people or consumers. This is also known as “helicopter money” (HM), which is the large-scale provision of cash transfers (see box 5). During 2020-2021, a total 13 economies provided universal transfers to whole population (or for adult members only and variants thereof) with the premier goal of boosting consumer spending.42 Box 5. Quantitative easing, helicopter money and design choices In addition of providing indirect and direct ways of injecting liquidity into the economy, there are a number of differences between quantitative easing (QE) and helicopter money (HM). For instance, there are diverse implications for central bank balance sheets: under QE, central banks create reserves by purchasing bonds or other financial assets; while under HM, central banks distribute the money created without increasing assets on balance sheets. The extent to which QE and HM are consumed depends on many factors, including for example the profile of recipients, supply-side responses and, intriguingly, whether money is perceived as free or coming at a cost. On the later point, HM provides newly issued, ‘windfall’ cash; instead, QE would need to be repaid (e.g., via future taxes). There is also debate on whether HM should occur via (more) transfers or (less) taxes: for example, the American Recovery and Reinvestment Act of 2009 included tax rebates. While a matter of intense empirical debate, there is suggestive evidence that the provision of relief via taxes may have not been fully “realized” by consumers. For example, Sahm et al (2012) found that he reduction in withholding in 2009 boosted spending at roughly half the rate (13%) as the one-time payments (25%) in 2008. One reason for the difference was “visibility”: “… a change in withholding is simply an adjustment by one’s employer of a recurring, periodic flow. It might not be noticed, especially for individuals whose paycheck routinely fluctuates for other reasons” (ibid, p.4). Furman (2020) argued that in 2009, the option of withholding less taxes was selected because of concerns regarding the risks and administrative feasibility of swift, nation-wide distribution of “debit cards” for relief transfers. Differently from 2009, the choice for tangible, visible “cheques” to nearly all the US population in 2020 might have enshrined an element of public visibility. Those cheques were not helicopter money per se, but they illustrate the point of tax versus benefits: as put by Matthews (2021), “… the flat cash checks to almost every American were by far the most visible part of the country’s policy response to the Covid-19 pandemic”. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 43 Were large scale injections of cash transfers effective in bolstering consumer spending? How money is allocated by recipients would hinge on a host of factors, including the profile of beneficiaries (e.g., their share of budget devoted to food expenditures, etc.); the design of programs (e.g., size, frequency, and duration of payments); and contextual factors, like inflation. For example, Roll et al (2022) show how cash transfers usage in the United States changes based on alternative design parameters like frequency of payments and their amount. Clear public communication can influence spending patterns: using high-frequency panel survey of Euro area consumers’ expectations and behaviors, Georgarakos Kenny (2021) found that that “… [b]y communicating clearly and factually about the magnitude, nature and ultimate goals of fiscal interventions, it is possible to improve the public’s perceptions about the adequacy of such policies. This, in turn, can help stimulate consumer spending and enhance the benefits of fiscal policy in terms of stabilising the economy”. In fact, one of the findings of the study is that expenditures on non-durables (e.g., clothing) increase when consumers perceive government support to be more adequate. Recent studies also suggest that the source of transfers matters: for example, an experiment by Crossley et al (2022) shows that in the UK, cash transfers provided from the government are more likely to be spent than from other sources. Evidence on marginal propensity to consume (MPC) out of large-scale Covid emergency cash transfers is available for a sample of (at least) three high income countries: estimates show that MPC in Germany, Japan, and Unites States range between 10-40%, with an average of 25.1% -- that is, a quarter of additional benefits received as cash transfers were consumed or spent on durable and/or non-durable goods (figure 25). The rest of transfers was used for others non-consumption purposes, such as savings and debt repayment. Figure 25. Use of cash transfers in select countries, marginal propensity to consume (MPC) Figure 25. Use of cash transfers in select countries, marginal propensity to consume (MPC) US (Mishra et al) 40 US (Coibion et al) 40 Japan (Kubota et al) 31 US (Baker et al) 29 Japan (Kaneda et al) 16 Germany (Goldfayn et al) 10.3 Japan (Hattori et al) 10 0 5 10 15 20 25 30 35 40 MPC Source: compiled by author. Note: when estimates include a range, the average is here reported. Source: compiled by author. Note: when estimates include a range, the average is here reported. In Japan, the effects of the one-off, universal infusion of $100B ($950/capita) were studied by three papers: Kubota et al (2021) tracked 2.8 million personal accounts at Mizuho Bank and used cash withdrawal as In Japan, the effects of the one-off, universal infusion of $100B ($950/capita) were studied by three papers: consumption proxy. Their estimated MPC was 0.31. Another study by Kaneda et al (2021) estimated an MPC of (0.06 et Kubota 0.16 foral (2021) lower and 0.27 for2.8 tracked million upper bound personal accounts estimates) within 6 at Mizuho weeks Bank and of receipt. used Finally, cash Hattori etwithdrawal al (2021) as calculated an MPC consumption of 10%. proxy. The Their latter level estimated is broadly MPC in line was 0.31. with findings Another from study by Germany: Kaneda et alGoldfayn et al (2022) (2021) estimated an MPC showed that the first among the three payments (for a total of E450/child) boosted spending by 9.5-11.1%, with effects higher for low-income families. Cash transfers have been subject to extensive empirical scrutiny in the United States. Coibion et al (2020) studied 44 III. ANATOMY OF RESPONSES AND EMERGING PERFORMANCE choices of cash allocation among 12,000 CARES Act beneficiaries. They found that recipients spent or planned to spend about 40% of transfers on average (mostly on non-durables); 30% of cash was directed for debt of 0.16 (0.06 for lower and 0.27 for upper bound estimates) within 6 weeks of receipt. Finally, Hattori et al (2021) calculated an MPC of 10%. The latter level is broadly in line with findings from Germany: Goldfayn et al (2022) showed that the first among the three payments (for a total of E450/child) boosted spending by 9.5-11.1%, with effects higher for low-income families. Cash transfers have been subject to extensive empirical scrutiny in the United States. Coibion et al (2020) studied choices of cash allocation among 12,000 CARES Act beneficiaries. They found that recipients spent or planned to spend about 40% of transfers on average (mostly on non-durables); 30% of cash was directed for debt repayment.43 Additional econometric work by Baker et al (2020) estimated a spending increase by 29% using the FinTech app (mostly spending on non-durables, driven by liquidity constrained households). Also, Misra et al (2020) examined debit card transaction data and documented an MPC of 29-51%. Furthermore, Karger and Rajan (2020) had a study following for two weeks recipients of the $1,200 payments. They found that $546 of the transfer was spent, which translated into an MPC 46%.44 Other studies don’t necessarily estimate MPC, but still provide insights in spending patterns. Again, in the United States, a report by the Federal Reserve examined spending of cash transfers under the Child Tax Credit Program (US Federal Reserve Board 2022). Findings confirm that most cash transfers were saved (43% of recipients), followed by expenses for children (40%), food (31%), rent, mortgage, or utilities (29%) and debt repayment (21%).45 Interestingly, wealthier recipients were three times more likely to save than the poorest (earning less than $25,000/year) (table 3). Table 3. Use of cash transfers by US Child Tax Credit beneficiaries by income Use of cash < $25,000 $25,000- $50,000- >$100,000 $49,999 $99,999 Saved it 18 15 33 54 Paid off debt 14 9 9 10 Spent on child 22 22 22 16 Spent on rent, mortgage, or utilities 29 33 18 6 Spent on food 13 19 15 7 Other expenditures 5 2 5 7 Source: US Federal Reserve Board (2022) In Thailand and Vietnam, Bui et al (2022) used a survey of about a thousand respondents in each country to estimate the effect of transfers. These amount to up to $240/month in Vietnam and $412/month in Thailand. Their findings show that the likelihood of buying durable goods was 22 percentage points (pp) higher among beneficiaries in Vietnam and 13 percentage points higher among those in Thailand. Finally, in Israel Feldman and Heffetz (2020) showed that at the time of their study (mid-2020), 25-45% of beneficiaries of a universal transfer ($220/adult and $150/child) had already mostly spent or were planning to spend money by year’s end. About 36-52% of beneficiaries paid down debts, while 10-18% of them had rerouted transfers to others voluntarily. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 45 Spent on rent, 29 33 18 6 mortgage, or utilities IV. THEMATIC ISSUES AND Spent on food 13 19 15 7 Other expenditures 5 2 5 7 COUNTRY PRACTICES Source: US Federal Reserve Board (2022) In Thailand and Vietnam, Bui et al (2022) used a survey of about a thousand respondents in each country to estimate the effect of transfers. These amount to up to $240/month in Vietnam and $412/month in Thailand. Their findings show that the likelihood of buying durable goods was 22 percentage points (pp) higher among beneficiaries in Vietnam and 13 percentage points higher among those in Thailand. Finally, in Israel Feldman and Heffetz (2020) showed that at the time of their study (mid-2020), 25-45% of beneficiaries of a universal transfer While devised withing a social protection framework, cash transfers are de facto cross-sectoral. Examined ($220/adult and $150/child) had already mostly spent or were planning to spend money by year’s end. About through 36-52% of practical lens a beneficiaries and paid country down experiences, debts, while 10-18%this section of them explorestransfers had rerouted the deliberate to otherslinkages between voluntarily. cash transfers and measures in other sectors and themes. A guiding compass for the section is laid out 4. Thematic issues and country practices in figure 26. While devised withing a social protection framework, cash transfers are de facto cross-sectoral. Examined through a practical lens and country experiences, this section explores the deliberate linkages between cash transfers and measures in other sectors and themes. A guiding compass for the section is laid out in figure 26. Figure 26. Thematic content of section 4 Figure 26. Thematic content of section 4 Government-provided cash transfers linkages with… Labor markets Place-based, spatial policies Human capital sectors 4.1 Non-standard and self-employed workers (multiple cases) 4.3 Urban areas and 4.4 Health and informal settlements education (Morocco, US) 4.2 Informal workers (DRC, multiple cases) (Colombia, South Africa, Philippines, Pakistan, Togo, Peru) 4.1 Non-standard and self-employed workers Social protection systems in advanced economies tend to display a variety of forms and provisions varying by worker contracts. Within such diversity, some common themes emerge. In the context of formal sector wage 4.1 Non-standard employment, in large number workers and self-employed countries invested of jobs retention programs (Carranza et al 2020). Those schemes are part of the family of wage-oriented, employment-related interventions that also include measures like public works and earned income tax credits. As such, they fall under supply-side, labor market activation measures 42 . Over the Social protection period 2020-2022, systems in advancedwage subsidies involved economies tend to233 interventions display in 117 a variety countries of forms and and were provisions particularly prevalent in more formalized economies43 (Gentilini et al 2022a). Importantly, those programs were varying by worker contracts. Within such diversity, some common themes emerge. In the context of “… adapted to the needs of the COVID-19 crisis, with the schemes extended to include larger segments of the formal economy sector wage as well employment, as benefit countries levels for workers invested in large number and compensation ofemployers rates for jobs retention programs raised” (Seemann(Carranza et al 2021, p.563). et al 2020). Those schemes are part of the family of wage-oriented, employment-related interventions that also include measures like public works and earned income tax credits. As such, they fall under supply-side, labor market activation measures.46 Over the period 2020-2022, wage subsidies involved 233 43 interventions in 117 countries and were particularly prevalent in more formalized economies47 (Gentilini 46 IV. THEMATIC ISSUES AND COUNTRY PRACTICES et al 2022a). Importantly, those programs were “… adapted to the needs of the COVID-19 crisis, with the schemes extended to include larger segments of the economy as well as benefit levels for workers and compensation rates for employers raised” (Seemann et al 2021, p.563). More to the point, a core common challenge across social protection regimes in higher-income settings included reaching and adequately supporting categories of workers that don’t conform to the “norm” of full time, regular, open-ended employment with a single employer. Those profiles include the self- employed and workers in non-standard employment (NSE) such as part-time, temporary, and zero-hour contracts. A rich literature documents how those broad categories of workers tend to face constrained access to contributory measures, like unemployment insurance, while not qualifying for social assistance schemes such as guaranteed minimum income programs (OECD 2020a,b). Also, particular implementation modalities can stifle the uptake of benefits even if granted in principle, such as limited awareness of benefits, cumbersome application procedures, stigma, and various other barriers. For example, the United Nations Human Rights Council (2022) reports that in the United Kingdom, “… an estimated 280,000 to 390,000 individuals wrongly thought they were ineligible for Universal Credit between June and August 2020 (…). [Furthermore] 41.6% of individuals who thought they were eligible [didn’t claim the benefit] because of stigma or perceptions of not deserving it” (p.12-13). There are differences between countries and within NSE categories. On one hand, recent evidence presented by Immervoll et al (2022) shows that contrary to conventional wisdom, for half of the sample of 16 OECD countries (i.e., Australia, Austria, Belgium, Germany, France, Hungary, Spain, and UK) there is only a limited gap between coverage provisions accorded to standard versus non-standard workers. Conversely, differences were more pronounced in other countries, including non-standard workers being 50-60% less likely to be covered during periods of unemployment than standard workers (e.g., Korea, Poland, Portugal, Spain, Italy, Latvia, Lithuania, and Estonia). On the other hand, there are qualitative differences between atypical workers. Most NSEs tend to have access to unemployment benefits, but face high risk of exclusion. For example, in the European Union Matsaganis et al (2016) estimated that such risk is only 0.1% for standard workers, while it soars to 31.9% and 38.7% for temporary full-time and temporary part-time workers, respectively. Self-employed populations, instead, tend to be excluded more often, but with some caveats: for example, exclusion from unemployment insurance is single digit in countries that mandate such coverage (e.g., 7.9% in Poland); but it skyrockets where unemployment insurance is voluntary (e.g., 93.7% in Romania). An EU-wide review by Spasova et al (2021) shows that 13 countries offer the selfemployed no formal access to insurance, 6 states do so on voluntary schemes, and only 7 nations mandate it. Against this background, part of the responses in the European context involved the adaptations of wage subsidies to meet the needs of vulnerable wage employed workers. For example, France extended wage subsidies to categories of workers like artists and freelance journalists, while in the UK the definition of ‘employee’ was made broader (Seemann et al 2021). There are some examples of inclusion of NSE workers in the routine system, e.g., Sweden lowered the time to access unemployment benefits (50% over the past CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 47 6 months) so to allow a larger number of part-time workers to become eligible. Also, Belgium leveraged its relatively generous unemployment system, including by halting the reduction in benefits over time (Cantillon et al 2021). Otherwise, a large number of measures for those workers involved cash transfers for, among others, seasonal occupations (e.g., Greece provided €534 for June-September 2020), part-time workers (e.g., Spain for temporary workers with contracts ending after lockdown), and domestic workers (e.g., 1.3 million returning workers in Romania). A key role was played by social assistance: while a range of measures for the unemployed were called ‘unemployment benefits’, ‘bridging rights’ and ‘worker allowances’, in practice they were ad hoc non- (figure 27 panel a). contributory Instead, cash unemployment transfers. insurance In other words, was more temporary widely social increased, assistance including filled with a range the structural of new gaps in Covid-specific cash allowances (panel b). (figure 27 panel a). Instead, unemployment insurance was more widely increased, including with a range of new Figure 27. Recipients of guaranteed Covid-specific (panel b). income (panel a) and unemployment benefits (panel b) prior minimum cash allowances Figure 27-A. Recipients and during the of guaranteed pandemic (% of minimum income working-age prior and population, during select the OECD pandemic countries) Figure 27. Recipients of guaranteed minimum income (panel a) and unemployment benefits (panel b) prior (% of working-age population, select and during the OECD pandemic countries) (% of working-age population, select OECD countries) Panel a Panel a Panel b 45 Panel b Source: OECD Social Benefit Recipients High Frequency Database (SOCR-HF), accessed June 2022. Note: the selected nine countries are those with graphs available for both indicators. 48 IV. THEMATIC ISSUES AND COUNTRY PRACTICES social insurance. The need for such ad-hoc measures is also evident in the share of new cash transfer programs out of total expansions: as discussed, these are almost 98% in HICs compared to 75% in LICs. In fact, according to the OECD Social Benefits Recipients High-Frequency (SOCR-HF) database48 with a few exceptions (e.g., Italy) existing minimum income guarantee programs were only marginally scaled up in 2020 and 2021 compared to 2019 (figure 27-A). Instead, unemployment insurance was more widely increased, including with a range of new Covid-specific cash allowances (figure 27-B). (figure 27 panel a). Instead, unemployment insurance was more widely increased, including with a range of new Covid-specific cash allowances (panel b). Figure 27-B. Recipients of unemployment benefits prior and during the pandemic Figure 27. Recipients of guaranteed minimum income (panel a) and unemployment benefits (panel b) prior (% of working-age population, select and during the OECD pandemic countries) (% of working-age population, select OECD countries) Panel a Panel b 45 Source: OECD Social Benefit Recipients High Frequency Database (SOCR-HF), accessed June 2022. Note: the selected nine countries are Source: those OECD with Social graphs Benefit available Recipients for both High Frequency indicators. Database (SOCR-HF), accessed June 2022. Note: the selected nine countries are those with graphs available for both indicators. One reason for the introduction of tailored new measures is the perceived speed and administrative simplicity of launching One reason programs ex novo. Forof for the introduction tailored the example, new TOZO measuresscheme the Netherlands in perceived is the replaced speed and preexisting administrative measures since the latter “… were seen as too complex, the administration as demanding and therefore slow simplicity of launching programs ex novo. For example, the TOZO scheme in the Netherlands replaced and the coverage as too limited” (Cantillon et al 2021, p.333). And in Ireland, Hick and Murphy (2021) noted that governmentmeasures the preexisting “… chose not the since latter on to reply “… were seen existing complex, the as toopayments welfare but administration as demanding to create a specific Pandemic and therefore Unemployment slow and Payment coverage thebased (…) on as too limited” a one-page (Cantillonform application et al 2021, p.333).rapid to enable And in Ireland, Hick and administration, with Murphy to compliance be monitored (2021) noted that ex-post. (…) [PUP] “… the government created chose a not clearer discontinuity to reply with on existing the existing welfare socialbut security payments to system, both contributory and means-tested [and] reflected the primacy placed on responsiveness” (p.315 and 318). There are various examplesIN CASH TRANSFERS of cash transfer PANDEMIC programs TIMES: EVIDENCE, for protecting PRACTICES, the self-employed AND IMPLICATIONS . For example, Germany 45 SCALE UP IN HISTORY FROM THE LARGEST 49 provided support for 6-9 months without a means test, with retroactive review undertaken under particular Europe, Italy , Milan December 2020, poor Italian and foreign people with masks lined up at Pane Quotidiano for free food during Covid-19 Coronavirus, increase in unemployed people and economic crisis. By DELBO ANDREA create a specific Pandemic Unemployment Payment (…) based on a one-page application form to enable rapid administration, with compliance to be monitored ex-post. (…) [PUP] created a clearer discontinuity with the existing social security system, both contributory and means-tested [and] reflected the primacy placed on responsiveness” (p.315 and 318). There are various examples of cash transfer programs for protecting the self-employed.49 For example, Germany provided support for 6-9 months without a means test, with retroactive review undertaken under particular circumstances. The UK’s Universal Credit program suspended the ‘minimum income floor’.50 Some countries reached specific categories, e.g., Italy targeted craftsmen and traders, and Cyprus excluded the least affected (e.g., doctors and grocers). Other countries set ceilings on the number of employees, e.g., in Luxemburg assistance was provided to self-employed with less than ten employees. Austria focused on family-run firm, while Finland reached the solo self-employed. Furthermore, targeting was in some cases dynamic, such as the three TOZO packages in Netherlands. 51 These initially provided broad-based support and encouraged uptake “only if really needed”, and then gradually assisted the hardest hit and solo self-employed exclusively. Italy’s Emergency Income program was designed to “fill the gaps” – i.e., it was intended for vulnerable people not participating in other social protection programs and provided between €400-840/month over a two-month period (Eurofound 2021). The Irish PUP program offers some interesting insights into linkages between social assistance and active labor market interventions. A week after PUP’s launch in mid-March 2020, a wage subsidy scheme (TWSS) was also introduced for wage employment. A provision in the PUP allowed for payments via employers which, because of PUP and TWSS being mutually exclusive, “… created a momentum towards the out- of-work payment, PuP, especially given its ease of administration administrative simplicity” (Hick and Murphy 2021, p.318). To correct such outflow, the government took four steps: (i) about a month later, it raised TWSS from to 70% to 85% of net wages; (ii) in June 2020, a two-tier PUP was created, with those 50 IV. THEMATIC ISSUES AND COUNTRY PRACTICES earning less than €199/week receiving benefits aligned with the Jobseeker Allowance of €203/week; (iii) in the same month, activation and case management measures were reestablished; and (iv) in July, PUP’s benefits were further reduced. As a result, PUP claimants dropped from 600,000 in May 2020 to 310,000 in June 2021. In October, however, a new Covid wave, a PUP top-up was included bringing it back to its March level (€350/week). But also wages subsidies were now set for an equal (maximum) amount of support, maintaining parity between schemes. 52 Over time, research underscored the challenge of aligning Covid and pre-Covid assistance. Cantillon et al (2021) noted that Belgium’s approach may have generated inequality in treatment among the unemployed, with those losing jobs due to Covid receiving higher benefits than pre-pandemic claimants. In Ireland, PUP’s initial benefits were not only much higher than the Jobseeker Allowance, but also equivalent to 100% of average pay in low-income sectors (Hick and Murphy 2021). Yet as just discussed, additional lockdowns “… thwarted the preferred policy direction of equalising pandemic and pre-pandemic supports” (ibid 2021, p.320). Yet there are some signs of convergence in better connecting contributory (insurance) and non-contributory arms of social protection systems. The example of Ireland itself shows that as party of the gradual phase out, remaining PUP beneficiaries were to be transferred to the routine system (a social insurance-based jobseeker’s benefit and a means-tested jobseeker’s allowance). In other cases, linkages were established more explicitly: Portugal’s cash transfer of €438 for informal workers established a requirement to join and remain in the social protection system for 30 months after the end of the support. Similar arrangements were adopted in Timor-Leste. Furthermore, Lithuania had a hybrid model whereby new, ad-hoc “job seekers” benefits (€257/month) to unemployed people excluded from insurance (which were about 60% of all unemployed) were combined with job search requirements. Also, eligibility was conditional upon payment of social contributions other than for unemployment (in fact, self-employed normally pay social contributions to some schemes, e.g., health insurance, sickness benefits). Schemes that were introduced as stop gap measure for informal workers can also open up new space for reimaging unemployment protection. For example, before the pandemic Georgia had limited unemployment insurance in place. However, in June 2020 the government announced that those in the formal sector who had lost their jobs could receive up to six monthly payments of $65. This was then extended in January 2021. By September 2020, more than 161,000 people had benefitted from the scheme. Meanwhile, the government gave a one-off payment of $97 to informal or self-employed workers who could show that they had lost work because of the crisis. This might have generated an opportunity to capitalize on the experience to develop a more permanent unemployment scheme (Anderson et al 2022). Such model also emerged in Argentina, which provided a one-off cash transfer for low-income, self-employed domestic workers (Bista et al 2021). This might also apply for programs like wage subsidies which, like in the case of Thailand and Vietnam, have acted as de-facto unemployment insurance (ILO 2020a,b). Similarly, in the Maldives an Income Support Allowance of $320/month was provided to 23,000 people (or about 5% of the population). The transfer was aligned with the Old Age Basic Pension Scheme, and it was close to the minimum wage level. Interestingly, such experience is generating momentum for extending contributory unemployment insurance (World Bank 2021a). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 51 4.2 Reaching informal sector workers Colombia During the Covid-19 crisis, the Colombia faced the challenge of filling structural coverage gaps of a social protection system with limited support to adult workers. To address such gap, the government introduced the temporary Ingreso Solidario program (Escobar Correa et al 2021). Reaching an additional 3 million households, Ingreso helped extend coverage in two ways: first, it supported populations that wouldn’t be included in permanent schemes due to budgetary restrictions (i.e., Familias en Acción, Jóvenes en Acción, and Colombia Mayor programs). Secondly, it covered profiles of beneficiaries not eligible in those routine social protection interventions, such as low-income households without children or elderly members (figure 28). Figure 28. Mapping of Colombia’s main social assistance programs across the lifecycle, 2021 Socioeconomic Early Childhood Childhood Adulthood Youth (14-18) Youth (19-28) Elderly (>=60) status/life cycle (0-5) (6-14) (29-59) Extremely poor and poor Vulnerable Non-vulnerable Familias en Acción (FA) Coverage gap of current programs: Population who would get out of social VAT Compensation assistance if Solidarity Income is only temporary. Columbia Mayor (CM) Population who might need on-demand access to social assistance in case of Jóvenes en Acción (JA) both covariate and idiosincratic (especially those closest to the poverty line). Ingreso Solidario (IS – gap filler, temporary) Source: Escobar Correa et al (2021) About 964,000 beneficiaries were brought into the banking system via payments (Prieto 2022). Such financial inclusion is confirmed by Vera-Cossio et al (2020), whose evaluation found that Ingreso Solidario increased the chances of opening of bank accounts by 14 percentage points and the likelihood of using digital tools for financial transactions by 7.7 percentage points. 53 South Africa Approximately 18 million people, or nearly one-third of the population, receive a monthly social assistance payment. These chiefly include old age social pensions, disability grants, and child support grants (CSG) at a cost of 3.4% of GDP. Yet, these programs don’t explicitly support the working age population. For example, the largest scale intervention, the CSG, only covers 44% of informal workers by virtue of reaching caregivers that happen to be in such occupation. For instance, Jain et al (2020) found that in April 2020 about 37% of those who were put on “paid leave” or laid off temporarily laid-off” were not covered by no social protection program. 54 52 IV. THEMATIC ISSUES AND COUNTRY PRACTICES In order to reach such category of workers deliberately, the new ‘‘Special COVID-19 grant” (or Co-G, or R350 program) was launched for one year (with extension of 8 months) targeting adults not formally employed nor being covered by existing social grants (Gronbach et al 2022). The benefit was about 40% of monthly median income. This was akin to a scheme introduced in Namibia. 55 The novelty of the Co-G should not be underplayed: as noted by Kohler and Bhorat (2021), “… it is the first in the post-Apartheid era to target unemployed adults – a group who had previously been largely unreached by the system” (p.7). Applications could be made by email, mobile phones (through USSD codes) and WhatsApp (following an earlier unsuccessful test when the app crashed after 91,000 applications). 56 Applicants had to simply record their name, ID number and address, and confirm that they have no other income. No supporting documents were required (Seekings 2020b). While final coverage was significant, the rollout encountered several bottlenecks. Devereux (2021) reported that when announced in April 2020, an estimated 12-15 million people were eligible: one month later, 5.1 million people had applied, 2.6 million applications were processed, and only 11,000 – or 0.2% of applicants – were “about to be paid”. Delivery improved over time: three months after announcement 3 million people had been paid, a level that rose to 5.2-6 million individuals after 6 months (Baskaran et al 2020). The grant, alongside pre-existing grants, may have covered 36 million individuals, or 63% of the population (Bhorat et al 2021). Once the process improved, the performance of the grant was considerable. Kohler and Bhorat (2021) found that in October 2020, 58% of recipients were unemployed, and that about 70% of those employed were informal sector workers. As shown in figure 29, the grant reached all deciles: among applicants, 23% of individuals in the poorest 10% of households were successful, while only 0.71% in the richest decile. 57 Such finding is broadly consistent with findings by Bassier et al (2021), who noted that “… when combined, the CSG and the Co-G offer effectively universal coverage of informal-worker households” (ibid, p.7). Figure 29. Coverage of special Covid grant in South Africa by decile, November/December 2020 Figure 29. Coverage of special Covid grant in South Africa by decile, November/December 2020 100 80 39.28 52.49 61.30 61.96 65.22 % of Individuals 71.94 73.54 60 78.89 83.00 20.21 97.09 40 16.29 17.76 10.92 14.63 7.51 13.25 10.79 13.15 20 9.93 13.28 9.24 11.61 4.23 22.75 24.71 7.71 4.22 0.59 4.63 14.14 16.33 3.44 0.71 12.29 7.21 8.96 6.07 8.14 1.61 0 Poorest 10% 2 3 4 5 6 7 8 9 Richest 10% Per capita household income decile, June 2020 Source: Kohler and Bhorat (2021) Applied + successful Applied + unsuccessful Applied + pending Never applied CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 53 The Philippines The Philippines illustrates a valuable case study on the opportunities and challenges of a massive scale up of social assistance. The social protection response (SAP) involved a phased approach at a cost of $4B and 1.1% of GDP, with a rich set of lessons The documented Philippines in the literature illustrates a valuable case study (Cho and on the Johnson 2022; opportunities andCho et al of a massive sca challenges social assistance. The social protection response (SAP) involved a 2021; Gudmalin et al 2021; Ramos 2021). The first tranche (SAP-1) targeted about 18 million households, phased approach at a cost of $4B and GDP, with a rich set of lessons documented in the literature (Cho and Johnson 2022; Cho et al 2021; G or 75% of the population. Within such caseload, about 4.3 million were reached via the existing flagship et al 2021; Ramos 2021). The first tranche (SAP-1) targeted about 18 million households, or 75% 4Ps programs (via social registry based on population. data). 2015 such Within Such payment caseload, about 4.3 top-up million was completed were in the reached via 13 days. By flagship 4Ps progr existing social registry based on 2015 data). the end of June 2020, the cash benefits had reached 98% of the target beneficiaries of the first tranche, By the end of Ju Such payment top-up was completed in 13 days. the cash benefits had reached 98% of the target beneficiaries of the first tranche, or 18 million househ or 18 million households; for the second tranche, by August 2020, 2 months after the start of distribution the second tranche, by August 2020, 2 months after the start of distribution of the second tranche, the of the second tranche, the program had coveredhad million 13.3 covered 13.3 million families families or 94% of targetor 94% of target beneficiaries beneficiaries (Gudmalin et al 2022). (Gudmalin et al 2022). Given the absence of registry data for the reminder of beneficiaries, paper applications (vias Amelioration Card, figure 30) and manual registration were undertaken. Local government units p Given the absence of registry data for the payments reminder while collecting SACs, with audits conducted at later stage. Various errors in registration Figure 30. Sample Social considerable of beneficiaries, paper applications (vias in backlog a payment, where the delivery of assistance for non-4Ps beneficiaries took over 3 Social Amelioration Card, figure 30) and manual FigureAmelioration Card in the Philippines 30. Sample Social Amelioration Card in the Philippines registration were undertaken. Local government units provided payments while collecting SACs, with audits conducted at later stage. Various errors in registration led to a considerable backlog in payment, where the delivery of assistance for non- 4Ps beneficiaries took over 3 months. The second phase of social protection response, or SAP-2, scaled up assistance further. This involved 23 million households, reaching nearly an additional 5.3 million “waitlisted” families, mostly in urban areas, that couldn’t receive SAP-1. There were efforts to enhance registration: for example, an app for digital registration (ReliefAgad) was developed by USAID-facilitated volunteers. This was used by both local government units to digitize SACs (replacing Excel/CSV) and beneficiaries to register via mobile phones, tables, etc. The innovation, however, didn’t address the core business process problem, as card with beneficiaries needed the SACSource: barcodes Gudmalin et al (2021) to start the process (such car/barcode could be The second phase of social protection response, or SAP-2, scaled up assistance further. This involved 2 damaged, non-readable, etc.). Also, the platform Source: Gudmalin et al (2021) households, reaching nearly an additional 5.3 million “waitlisted” families, mostly in urban areas, that could not be connected to functional receive SAP-1. (e.g., were efforts to enhance registration: for example, an app for digital reg IDsThere (ReliefAgad) was PhilHealth card) and other programs. As a result,developed by USAID-facilitated volunteers. This was used by both local government digitize SACs (replacing Excel/CSV) and beneficiaries to register via mobile phones, tables, etc. The inn only a portion of ReliefAgad data could however, be used. didn’t address the core business process problem, as beneficiaries needed the SAC card with b 50 54 IV. THEMATIC ISSUES AND COUNTRY PRACTICES Political pressure to avoid SAP-1 delays and move to digital payments was mounting. This led to a combination of financial service providers including three banks and three e-money issuers (i.e., PayMaya, Gcash, and StarPay). Each involved possible particular challenges at each step of the process (figure 31). To enhance speed and avoid different experiences in the same municipality, the Central Bank assigned financial service providers to areas, i.e., people could not choose FSPs. Whether in the form of e-money or basic bank deposits, the decision was to have “restricted” transaction accounts. These can be opened without IDs; but have a credit ceiling ($2000/year), they cannot be used for remittances, and can be closed if dormant, e.g., no movements in 12 months. Some provisions to “regularize” accounts through KYC is ongoing. Figure 31. Mapping beneficiaries’ journey via banks and e-money issuers (EMIs) Source: Cho et al (2021) Source: Cho et al (2021) Pakistan CASH The Ehsaas TRANSFERS strategy, IN PANDEMIC which TIMES: provided theEVIDENCE, backbone PRACTICES, AND IMPLICATIONS of the social FROM THE LARGEST protection response, SCALE built and UP IN HISTORY complemented 55 the NSER social registry (with information on 85% the population, and update since 2010/11) and the regular Pakistan The Ehsaas strategy, which provided the backbone of the social protection response, built and complemented the NSER social registry (with information on 85% the population, and update since 2010/11) and the regular BISP/Ehsaas Kafaalat program58 directly tied to it (reaching 7 million households). The main device for scale up was the Ehasaas Emergency Cash (EEC). Overall, the EEC reached up to 16.9 million families, or about 43% the population (Beazley et al 2021; Hassan 2021; Lone et al 2021; Nishtar 2020). The benefit included a one-off transfer ($71) equivalent to Figure 32. Performance of EEC in about 16.5% of expenditures/consumption of the poorest 40% of the population. While one off, the response would cover Pakistan based on district-level needs for about 4 months. The response cost $1.3 billion, with data (2020) assistance from IFIs for about $225 million for cash transfer benefit top up. The EEC included beneficiaries from Kafaalat, an additional 4 million drawn from the NSER (higher eligibility cut-off), and about 3.5 million non in NSER and identified by provincial/district administrators (Khan and Jamy 2020). In the latter case, administrators would use a poverty scorecard if PANEL A available, or average mobile/landline monthly bills less than $0.6. Provincial and district quotas were introduced, resulting in an estimated 3.7 million potential affected workers and beneficiaries being excluded from potential assistance. The Punjab province also included about 200,000 beneficiaries via local zakat, while various humanitarian programs were also active, e.g., 21,000 beneficiaries reached by WFP in FATA, UNHCR covering 75,000 Afghans, and IRC utilizing NSER for 1250 households. PANEL B The EEC launched 10 days after country’s lockdown. BISP payments use biometric verification system via two commercial banks, combining a Limited Mandate Account (LMA) with savings. Within 1 week, 46 million requests were processed, and 7.3 million people enrolled (Lone et al 2021). EEC used the same platform, but without savings (LMA only allows cash withdrawal, not other services). Analysis of district-level data by Markhof (2020) offers a glimpse at the EEC performance in the first 5 months of the pandemic: it found that, consistently with population size, most beneficiaries were located in the provinces of Punjab and Sindh (panel a of figure 32); that an average of 86% of identified beneficiaries PANEL C have received their payments (panel b); and that 66% of a district’s vulnerable population were covered (panel c). Source: Markhof (2020) 56 IV. THEMATIC ISSUES AND COUNTRY PRACTICES Pakistani people wears face mask to protect themselves Corona virus ( COVID-19 ) before entered in the historical walled city of Lahore, Pakistan during Government imposed Lock down on June,06,2020. By A M Syed Interestingly, data analytics was used to exclude 0.8 million “wealthy” households. This involved considering criteria like international travel; expensive ID/passport NADRA processing; car ownership (and land in the case of Punjab); government employment; declared income at tax agency higher than $285; monthly phone bills over $6 from telecom authorities. Such “targeting from the top” was also used in other settings (e.g., Bolivia), although at times through taxes such as in Germany. 59 Togo The experience of Togo embodies the positive performance of a trailblazer. At the outset of the pandemic, Togo’s social protection systems were limited. This involved a database with only about 90,000 households in rural areas (or 5% of the population). Because of the country’s substantial scale up was achieved through innovative analytics and evidence-based leadership, Togo has attracted widespread interest in academic and policy circles (Aiken et al 2022; Chowdhury et al 2022; Lawson et al 2022; Debenedetti 2021). Togo’s response can be divided in two main phases. The first stage of the Novissi program (April- September 2020) provided cash transfers to about 570,00 informal sector workers in the greater Lomé area, the Tchaoudjo prefecture, and the Soudou canton. A new digital platform was created in 10 days, establishing an end-to-end system for enrollment and payments via an 8-step process mobile phones. The approach tapped the national voter registration database (including data for over 90% of the 3.8 million adults and inclusive of information on location and occupation). One week after the program was launched, 450,000 beneficiaries had already received their first payment. After a year, Novissi created around 170,280 new mobile money accounts. There were implementation variants. For example, as lockdowns widened, the program added other workers, including from the public sector like community schoolteachers and public transport operators. Furthermore, in some cantons (Soudou), the program went beyond targeting worker categories and reached 85% of the population (except for civil servants). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 57 The second phase of Novissi shifted to a more spatially tailored, needs-based approach targeting 57,000 beneficiaries in the poorest 100 cantons (World Bank 2021b). This involved various steps. In terms of selecting locations, deep-learning algorithms were applied to high-resolution satellite imagery to produce wealth estimates of 2.4km grid cell (left image in figure 33); these estimates were then combined with information on population density of each grid cell (middle image); finally, such information was used to identify the 100 poorest cantons (right image). Figure 33. Identifying poorest cantons in Togo Source: World Bank (2021b) as provided by J. Blumenstock Source: World Bank (2021b) as provided by J. Blumenstock In terms of identifying people, phone surveys provided “ground truths” of living conditions of 10,000 people. Based In terms on of such data, machine-learning identifying people, phone algorithms surveys estimated provided wealth of truths” the“ground each mobile subscriber of living (left image conditions ofin10,000 figure 34). In the 100 poorest cantons (red area in right image), those estimated to consume less than $1.25/day people. Source: Based are prioritizedon World Bank such for data, (2021b) machine-learning as provided the Novissi program by(dotted algorithms J. Blumenstock vertical estimated line). Such the beneficiaries wealth are of each significantly mobile poorer thansubscriber the average person in Togo (blue area). (left image in figure 34). In the 100 poorest cantons (red area in right image), those estimated to consume In terms of identifying people, phone surveys provided “ground truths” of living conditions of 10,000 people. less than $1.25/day are prioritized for the Novissi program (dotted vertical line). Such beneficiaries are Figure 34. Identifying Based on such data, machine-learning poorest algorithms people in the poorest estimated wealth ofcantons in Togo each mobile subscriber (left image in significantly figure 34). In poorer the 100than the poorest average cantons (red area inin person Togo right (blue image), area). those estimated to consume less than $1.25/day are prioritized for the Novissi program (dotted vertical line). Such beneficiaries are significantly poorer than the average person in Togo (blue area). Figure 34. Identifying poorest people in the poorest cantons in Togo Figure 34. Identifying poorest people in the poorest cantons in Togo Source: World Bank (2021b) as provided by J. Blumenstock Further research compared different targeting methods, including machine learning and mobile phone data, geographic, occupation-based mythologies (for informal workers and poorest jobs), and a proxy means test (Aiken et al 2022). It found that machine-learning reduces errors of exclusion by 4-21%. However, when relative to scenarios with a hypothetical social registry machine-learning approach would increases exclusion errors by Source: 9-35%. World Bank (2021b) as provided by J. Blumenstock Source: World Bank (2021b) as provided by J. Blumenstock Peru research compared different targeting methods, including machine learning and mobile phone data, Further geographic, occupation-based mythologies (for informal workers and poorest jobs), and a proxy means test 58 IV. THEMATIC ISSUES AND COUNTRY PRACTICES 55 errors of exclusion by 4-21%. However, when relative (Aiken et al 2022). It found that machine-learning reduces to scenarios with a hypothetical social registry machine-learning approach would increases exclusion errors by Further research compared different targeting methods, including machine learning and mobile phone data, geographic, occupation-based mythologies (for informal workers and poorest jobs), and a proxy means test (Aiken et al 2022). It found that machine-learning reduces errors of exclusion by 4-21%. However, when relative to scenarios with a hypothetical social registry machine-learning approach would increases exclusion errors by 9-35%. Peru The experience of Peru reveals the gradual expansion of coverage and the construction of a framework incorporating three sequenced programs (Bastagli and Lowe 2021; Lowe et al 2021; Olivera 2021). The first is Bono Yo Me Quedo en Casa (Stay at Home Grant) funded through the Reserve Fund, IFI support, and treasury bonds. Payments were made within a week to 2.7 million beneficiaries (31% of the population) with two payments of $104 based on the existing social registry (itself including information on 75% of the population) (poor in urban areas). Subsequently, the Bono Independiente program targeted 773,000 informal sector workers, or 9% of population, with a single payment of $209. These included people considered “non-poor” in the social registry, but with income less than $330/month, living in “high health vulnerability areas” and not participating in other programs. Furthermore, the Bono Rural (again a single $209 payment) focused on rural areas and reached 980,000 beneficiaries considered “poor” in the registry, but also not receiving other support. Finally, the Bono Familiar Universal integrated programs into a single framework, including the three Bonos schemes plus those not enrolled in any. This brought the total coverage to 8.6 million households, or 68% of the population. Beneficiaries received two payments of $209, compounded by a mechanism of “targeting from the top”, i.e., all were considered eligible except those in formal employment, excluding interns, in public office, and with income over $824/month. Lima, Peru - April 4 2020: Women lining in a queue practicing social distancing amid coronavirus outbreak in South America. Only woman day, gender-based quarantine, women lined up to buy groceries. By Myriam B CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 59 4.3 Spatial extensions to urban areas While the effects of the pandemic had severe consequences for livelihoods in both rural and urban areas (Sitko et al 2022), urban settings had been identified as powerful vectors for spreading the virus: for example, Brown et al (2020) estimated that 96% of housing settlements in Africa weren’t compliant with WHO standards. Most countries didn’t have an established social protection operational footprint for responding to heightened urban vulnerability. Those structures had in many cases to be introduced ex-novo, especially ban and rural coverage of cash and in-kind transfers in India in large informal settlements.60 In fact, pre-pandemic data shows that the coverage of safety nets is generally higher in rural than in urban areas: for the poorest quintile, such difference amounts to about 7 percentage points on average (Gentilini et al 2021b). Many factors may contribute to such spatial difference,61 and the Covid-19 crisis may have mirrored those uneven spatial provisions. For instance, Choudhuri et al (2022) estimated that in India’s Delhi National Capital Region the likelihood of urban residents being covered by cash transfers was 8 percentage points lower than among rural dwellers (figure 35). Figure 35. Urban and rural coverage of cash and in-kind transfers in India Figure 35. Urban and rural coverage of cash and in-kind transfers in India 80% 70% 60% Percentage of Households Ration Cash Either Both 50% Rural Urban 40% 30% 20% 10% 0% Ration Cash Either Both Rural Urban Source: Choudhuri et al (2022) The pandemic crisis prompted the emergence of a new generation of urban cash transfer programs (Gentilini et al 2021b; Roelen et al 2021). Based on 2021 data from 36 programs, urban social assistance planned to cover about 66.5 million beneficiaries globally, with 18.6 million been actually reached (Gentilini et al 2021a). A range of lessons are emerging from those experiences. At design level, satellite images at granular street level are becoming more routinely used and integrated in program design (e.g., see Kinshasa example in this section). Simplified questionnaires for beneficiary intake have been produced (e.g., Liberia comprised 10 questions, Mauritania’s “PMT plus” included 15 questions). Also, some delivery functions have been undertaken simultaneously to minimize time between beneficiary registration and payment (e.g., Sierra Leone, Nigeria). Ongoing cross-country studies will provide further insights on the use of big data in urban areas, including comparative effectiveness, efficiency, speed.62 Interestingly, 60 IV. THEMATIC ISSUES AND COUNTRY PRACTICES programs have established partnerships with formal and semi-formal local actors to help overcome information constraints on urban informal sector workers, e.g., local city councils, labor and tourism ministries, and MFI and trade associations in Sierra Leone. Yet challenges remain. These include the difficulty in defining a ‘household’, a ‘community’ and even a ‘neighborhood’ in fluid, informal urban settlements (e.g., Liberia). This posed some hurdles in the applicability of community-based targeting (e.g., Mauritania). But countries shows that urban communities can also be mobilized, such as via Ethiopia’s neighborhood committees. In Madagascar, the Loharano Committee (inclusive of community leaders and chaired by the head of the Fokontany) played a major role in program pre-registration. In Cote d’Ivoire, community leaders or neighborhood chiefs collected information on people’s IDs, vulnerability information, phone number, etc. Political pressure can be exerted and offer both an advantage in rallying efforts as well as present challenges in aligning communication (e.g., Madagascar, Sierra Leone). How to make programs attractive for youth is often a key priority: Kenya’s Covid public works were deliberately designed for that purpose, while Ethiopia’s USPNP scheme found it hard to enroll youth at pre-Covid stage. Furthermore, Covid responses underscored the importance to locate social protection within the framework of city administrators, e.g., how to frame urban public works as part of cities’ sustainable waste management systems. Turning to specific country cases, Kinshasa’s Step-KIN program was an innovative intervention in a context with no previous exposure to cash transfers. Operating in informal slums, the selection of locations was conducted based on the most flood-prone areas as identified by satellite images (figure 36). Partnerships with mobile phone providers allowed to receive lists of subscribers living in targeted areas. Figure 36. Stylized visualization of selecting Kinshasa’s informal settlements Source: Bance et al (2021) Source: Bance et al (2021) Because of social distancing protocols, another interesting development during the pandemic has been the acceleration in interest for service-oriented, digital public works (Bance and Gentilini 2020; Weber 2020). In 2020, seven digital public works were piloted under a coherent framework in four African countries, i.e., Mali, Tanzania, Kenya, Sierra Leone CASH TRANSFERS (Deparday IN PANDEMIC et al 2022; TIMES: EVIDENCE, World PRACTICES, Bank AND 2021c). FROM IMPLICATIONS The initiative engaged THE LARGEST SCALE UPlow-income IN HISTORY and 61 vulnerable populations possessing meeting minimum digital skills, including about 1,300 workers completing Non-disclosure agreements with telecom operators were signed to receive those anonymized lists. This generated a “quasi-registry” of potential beneficiaries. Filters were added to eligibility rules (those with smartphones, purchased data plans, or spent more than $5 on voice/data/SMS combined were excluded). Subsequently, SMS, audio IVR messages were sent to anonymous phone holders to participate and self- register in the program. Once enrollment was completed, mobile money accounts were opened for 100,000 individuals (with simplified KYC). Transfers included $25/month for a semester. Recent expansion includes, in partnership with Afrimoney, the provision of cash transfers for an additional 150,000 beneficiaries between Nov 2021 to April 2022. Also, a pilot of Mobile Aid model with Give Directly was launched in November 2021 reaching 17,000 beneficiaries (Bance et al 2021) Because Source: Bance of social distancing protocols, another interesting development during the pandemic has been et al (2021) the acceleration in interest for service-oriented, digital public works (Bance and Gentilini 2020; Weber Source: Bance et al (2021) Because of2020).socialIn distancing digital public 2020, seven protocols, works were another piloted under interesting a coherent development framework during in four African the pandemic has been the acceleration countries, i.e., in interest Mali, Tanzania, Kenya, for service-oriented, Sierra Leone digital (Deparday public works et al 2022; World (Bance and Bank 2021c). Gentilini The initiative 2020; Weber 2020). In Because of social distancing protocols, another interesting development during the pandemic has been the 2020, seven in acceleration digital engaged public forworks low-income interest and were piloteddigital vulnerable service-oriented, a coherent underpublic populations possessing works framework meeting minimum (Bance and four African in Gentilini digital skills, 2020;countries, including Weber 2020).i.e., Mali, In Tanzania, 2020, seven Kenya, about Sierra 1,300 digital workers public Leone works (Deparday completing 2022; World et al under over were piloted 2.8 million digitalBankframework a coherenttasks. 2021c). The Conducted in ininitiative partnershipsengaged four African with low-income private countries, and i.e., Mali, vulnerable Tanzania, populations sector Kenya, (e.g., Google, Sierra possessing Leone MapSwipe, meeting (Deparday etc.), alminimum et activities digital 2022; involved World skills, geolocating Bank including 2021c). slum The about services 1,300 (Nairobi), initiative engaged workers streetview completing low-income and over 2.8 million vulnerable digital possessing prediction populations tasks. Conducted verification (Zanzibar) in partnerships classifying meeting minimumaerial with imagery digital to private identify skills, sector solid including waste (e.g.,1,300 about Google, (Bamako), and MapSwipe, building workers completing etc.), activities over involved height 2.8 million geolocating validation digital (Dar slum services Es Salaam) tasks. Conducted (figure (Nairobi), streetview 37). in partnerships with private prediction sector verification (e.g., Google, (Zanzibar) MapSwipe, classifying etc.), aerial imagery activities involvedto identify geolocating waste solid slum (Bamako), services and building (Nairobi), height streetview validationverification prediction (Dar Es Salaam) (figure (Zanzibar) 37). classifying aerial imagery to identify solid waste (Bamako), and building height validation (Dar Es Salaam) (figure 37). Figure 37. Digital public works in Africa Figure 37. Digital public works in Africa Figure 37. Digital public works in Africa Worker first selected a task, then found the relevant infrastructure to geolocate and characterize them. Source: World Bank Source: (2021c) World Bank (2021c) Source: World Bank (2021c) 58 62 IV. THEMATIC ISSUES AND COUNTRY PRACTICES 58 4.4 Linkages with health and education With transfers being made largely unconditionally (see section 2), some of the more direct connections between social assistance, health and education may have been reduced in programmatic terms. Yet those linkages were still present in a number of ways. For example, Morocco’s cash transfer program reached about 5.4 million households, or 65% of the population (Kessaba and Halmi 2021; Mansour 2021). As part of such caseload, the program served 2.8 million RAMED beneficiaries, or those enrolled in the national non-contributory health insurance program, itself reaching 34% population. Differently from the other 2.6 million households, RAMED beneficiaries were already enrolled in database. A new digital platform was created for enrollment. Launched in April 2020, it involved payments via mobile phones. Beneficiaries on the RAMED database received payments within 3 days from inception. RAMED beneficiaries were able to enroll by sending a message to a new SMS platform (1212), while non-RAMED households had to submit a simple form through a new digital platform. The phone-based payment system with 2 options. Beneficiaries received the code by SMS and could withdraw benefits at ATMs. This option was available for beneficiaries who provided a phone number in their name and having ATMs in their locality. Other beneficiaries received the code by SMS and were invited to withdraw the benefit at postal and bank counters. A call center provided support, and by end of May, potential beneficiaries previously rejected could reapply (about 13% of the population. By July, 91% of the announced coverage target was met. Financing included a mix of traditional and non- traditional sources: a Special Fund was set up by the King, including $1 billion from the treasury and $2.3 billion from regional authorities and private donors (UN 2021). Similar arrangements were present in countries like Mauritania and Uzbekistan.63 In the United States, school closures meant a substantial scale up in food assistance programs. A record spending was recorded on food assistance across 15 programs, including a total of $122.1 billion in January 2021. In the short term, participants of SNAP, the flagship e-voucher programs, increased by 5 million (from 36.9 million in February 2020 to 42.5 million for the rest of FY20. To facilitate scale up, states allowed a number of waivers, most of which expired in Dec 2021. These allowed for shifts from household eligibility home visits to interviews by phone; no assessment and recertification requirements (with reporting of identity and income only); and adapted telephonic signature procedures. Those streamlined procedures don’t seem to have altered performance negatively, and even increased uptake (CBPP 2021). In parallel, another mainstay of then US safety nets, school meals, were interrupting provisions to nearly 30 million children (Marcus and Yewell 2022). The program was temporarily replaced by the P-EBT scheme: this involved delivering food voucher transfers in lieu of school feeding for a value of at least $5.7/day. The technology deployed was compatible with SNAP’s standard EBT transfers, hence allowing for the upload of benefits on the same EBT card (Waxman et al 2021a). CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 63 P-EBT underscores the importance of programs that provide automatic eligibility to each other. For example, Neuberger et al (2021) noted that children in families receiving SNAP qualify to school feeding, hence facilitating automating P-EBT enrollment. Also, 30 states used data on children supported by other schemes that also confer eligibility for free school meals, hence again allowing the issuance of P-EBT benefits directly and without requiring any step taken by perspective beneficiaries. In those circumstances – whether via other programs or because information was provided by schools – P-EBT benefits were issued “directly”; otherwise, they issued new cards through an application process, or some combination of both (figure 38). Figure 38. Methods for issuance of P-EBT benefits Source: Neuberger et al (2021) The arrangement involved collaboration between education administrators and state SNAP agencies that had not previously existed (Kone cons. 2020). As noted by Waxman et al (2021b), “… P-EBT requires a level of coordination with SNAP that is unprecedented” (p.33). Over an initial year (FY 19/20), main challenges included changing addresses during the pandemic, the logistics of issuing new EBT cards in a short time frame, and handling benefit payments. In the subsequent year, benefits had to be tailored to the wide variety of student learning models, including a shifting mix of fully virtual, hybrid, and in- person formats (Toossi 2021). 64 IV. THEMATIC ISSUES AND COUNTRY PRACTICES V. LOOKING AHEAD: EMERGING INSIGHTS AND LESSONS Whether Covid-19 is a “historical game changer” depends on where the bar is set. The discussion on whether the pandemic is leading to permanent changes in social protection hinges on the definition of “game changer”, and whether expectations are realistic. For example, part of the debate centered on the scale of assistance. If that’s the standard for “making history”, the coverage of assistance is no doubt unparalleled. With 1.3 billion people covered with cash transfers, support has been of unparalleled, historical proportions. Notwithstanding the special nature of the Covid-19 crisis, the fact that over 200 workers were supported with unconditional transfers is also remarkable. If the bar is set on whether assistance was adequate and “fit for purpose”, it is reasonable to be more skeptical: data shows that responses helped, but were in many cases not in sync with the depth and duration of the crisis. What if the bar was set on “improving” existing systems? Covid has led to some permanent improvements in delivery systems, although the majority of business processes seem to have been “back to normal”. This may appear as a missed opportunity to challenge some assumptions around the need for “control” through particularly demanding design and delivery procedures. What if the bar was about fomenting new or amplifying preexisting public debates on the key role of social protection? Out of the plethora of experiences, some have had profound longer-term effects within countries. For instance, Togo is building a social protection system that was largely absent before 2020; Central Banks in a dozen economies have displayed a readiness to use cash transfers as unconventional monetary policy; and in South Africa, the R350 grant – which is still active at the time of writing this paper – has provided new fuel to pre-existing structural debates on weaving a universal portfolio of social assistance programs (GoSA 2022). In some contexts, those debates on the outlived the programs that sparked them, such as the child tax credit in the US. Finally, if the bar was represented by altering the structural configuration of systems, that was a tall order to meet. Social protection is a means for an end, and the question of fundamental changes in social protection reach and financing would largely depend on changes in power balance in societies. Then the question is not whether Covid led to social protection changes, but whether Covid has altered a power equilibrium – and how that translates into change in social protection. Those power and political dynamics should be fully considered when interpreting future directions for social protection. V. LOOKING AHEAD: EMERGING INSIGHTS AND LESSONS 65 Assisting unemployed and precarious workers. During Covid-19, social assistance acted as de facto insurance against severe welfare losses. To cover for gaps in the social protection systems, countries extended social assistance to large swaths of the population. While formal sector workers benefited from wage subsidies or various forms of unemployment insurance, workers in non-standard employment contracts, self-employed and informal sector workers received cash transfers. This was not without hurdles. Enduring policy and operational difficulties in scaling up Covid responses were rooted, on one hand, in the limits of social insurance for workers in employment arrangements other than stable, formal employment. In fact, insurance mechanisms have traditionally faced a number of limitations as automatic stabilizers in crises. This includes implementation lags and eligibility criteria, e.g., for individuals reaching long-term unemployment (Romer and Romer 2021; Chodorow-Reich and Coglianese 2019). But barriers also stemmed from a prevailing architecture of social assistance geared for particular populations, which may not cover most workers. This largely includes countries with programs directing resources near-exclusively to the bottom of the distribution as well as encompassing contexts with generous categorical programs for children or seniors. In both cases, social assistance had to be “invented” for informal sector workers that were largely missed by current configurations. Such experience calls for reenergizing of efforts to adapt an extend unemployment (and underemployment) insurance for workers that are currently unreached by those measures. And in doing so, it opens up the possibility of novel approaches and fresh perspectives on how to reimagine the boundaries, interface and connections between social assistance and insurance. Towards automatic stabilizers. Decisions to scale up cash transfers have been largely based on policy preferences, available information, and previous experiences. In a way, choices over “who to cover, how much to provide, for how long” largely relied on informed, discretionary choices. They were much less based on automatic stabilizers, especially in low- and middle-income countries. Those stabilizers are “thermostatic” measures that scale up (and down) based on prearranged, insurance-like triggers.64 An automatic scale up of social assistance presents three key steps: (i) the selection of risks against which automatic transfers would be triggered; (ii) the operationalization of an architecture of triggers and scale up protocols across institutions; and (iii) the calibration of design choices, particularly on coverage, size, duration of benefits and pre-positioned financing. To inform these steps, more deliberate learning from unemployment insurance would be desirable. Because of possible limitations of insurance measures in crisis situations, an automatic scale up of transfers would build upon insurance principles, but adapting insurance design (e.g., need for a certain number of years of employment experience, duration of unemployment spells, etc.). There is nascent experience in similar anticipatory approaches in ex ante responses to droughts and other climatic hazards (Calcutt et al 2021; Pople et al 2021; Bowen et al 2020). Perhaps there is scope to connect, extend and foment “lateral learning” between such climatic experiences and scale up in the face of different events. One such contingency might be, for example, when unemployment hits a certain level, or if particular decisions are taken such as a declaration of health emergency (Furman 2020; Moffitt 66 V. LOOKING AHEAD: EMERGING INSIGHTS AND LESSONS and Ziliak 2020; Boushey and Shambaugh 2019). According to Furman and Summers (2020, p.15), “… [t]his automatic recession insurance could include (…) increased unemployment insurance benefits, or other transfers like nutritional assistance or even across the-board cash transfers”. Others like Hoynes and Schanzenbach (2019) discussed establishing national triggers not to scale up a program per se, but to automatically simplify design (e.g., waiving conditionalities during downturns). Just like financial institutions are regularly stress tested to ascertain their ability to withstand large shocks, countries should stress test their social protection systems. Despite overwhelming evidence on the benefits of crisis preparedness, few tools were available to assess the level of readiness and potential of extending social protection in the short term. Some recent analytical work offers such tool for cash transfers, which also offer the potential for better aligning social protection and humanitarian assistance65 (World Bank 2021d). Creative and diverse financing responses, with trade-offs and fiscal headwinds looming on the horizon. With a global investment of at least $3 billion and average response of 2% of GDP, the fiscal pressure exerted by the pandemic was significant. The Covid-19 response displayed a key role played by domestic funding including in low- and middle-income countries. A range of innovations were recorded, such as contingency financing and new funds being established in lower-income contexts (e.g., Mauritania). Private support also played a significant role in financing responses, such as contributions from the Zakat system. Yet many of those financing practices also present drawbacks. For instance, responses were in various contexts financed by redirecting resources from other sectors. Some of such repurposing might be desirable, like fuel subsidies, while in other circumstances it might be more problematic (e.g., reallocations from other social services and infrastructure). New funds and financing models operating in parallel to government systems might enhance short-term speed and transparency, but may also undermine national structures in the longer run. If debt wasn’t a constraint in crisis response, it may be in the post-crisis stage. This indicates that sharp trade-offs might be on the horizon. Romer (2021) articulates this point clearly for high income settings: “… [as policymakers] seek to tackle issues such as climate change, crumbling infrastructure, and persistent poverty, they may find increased opposition to further spending. Thus, one potential legacy of the extraordinary fiscal actions to fight the pandemic may be that the country fails to deal with other pressing needs” (p.107). Those pressing needs are rising as the ongoing Ukraine war-induced crisis amplified a wave of global inflation, with a palpable fatigue in social protection responses (Gentilini et al 2022b). Gupta and Tovar-Jalles (2022) call for bold reforms in low- and middle- income countries: with rising debt and widening of budget deficits that may persist over time, policymakers are urged to “… reconsider their revenue-raising strategy in favour of an approach that embraces a comprehensive reform package, including policies that have encountered political opposition in the past” (ibid, p.2299). The crisis might present an opportunity to reenergize financing of social protection across a portfolio of options, including for example the relationship between subsidies and transfers, the role of indirect taxes, and innovative earmarked taxation. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 67 The expected multipliers of cash transfers helped fiscal and unconventional monetary policy converge. A dozen economies, most of which in East Asia, provided cash payments to all adult individuals or even the entire population. Given the macro contexts, cash transfers were deemed the surest route to reignite consumer demand and economic activity. Emerging evidence shows that up to about 40% of transfers were consumed, possibly generating economic multipliers. It is likely that future recessions entailing strong reductions in consumer demand, temporary large-scale (including universal or quasi-universal) cash transfers would likely become a default option in emergency response – whether through fiscal or monetary policy. Furthermore, the Covid-19 experience may further increase interest in documenting economic multipliers of cash transfers in non-crisis contexts. This, in turn, may help forge a broader definition of fiscal sustainability in ways that account for both economic costs and benefits over time (Aizer et al 2022). Capitalizing on a new generation of urban social assistance programs. Over the past couple of decades, social protection had largely held a rural footprint. The pandemic is providing new impetus in adapting and extending social assistance to populations in peri urban and urban areas. This does not only imply modifications in design and implementation, but also in partnerships and framing. In fact, the urban agenda is inherently multisectoral and involving connections with sectors with which social protection has played a limited role, like urban development, housing, and transport. Emerging experiences from a dozen countries in Africa – e.g., DRC, Liberia, Nigeria, Madagascar, and Togo – suggest that new partnerships with private sector delivery, management of big data, and local actors (local governments, local trader associations, and civil society) may help make assistance more dynamic and apt for urban settings. The jury is still out on the full performance of such novel experiences, but this is an agenda – the extension of social protection in cities – of enormous untapped potential. Operating in urban environments opens the door for accelerating the discussion on mobility, portability of benefits, and individualized assistance as opposed to more traditional, static, and household-based provisions. Building universal delivery systems. Countries entered the pandemic with a range of unsolved questions on social protection. For example, what is an optimal level of social assistance coverage? What would be an ideal composition of social assistance and insurance? More to the point, if social assistance is skewed towards the poorest and the expansion of social insurance is halted by informality, precarious labor contracts, and other barriers, how should universal social protection be attained? The pandemic doesn’t solve these puzzles, but it may present some clues to start answering them. To begin with, it might be reasonable to expect that a mere return to pre-pandemic social protection provisions may not address the problem. Policy ideas often rest on worldviews and priorities, and an altered landscape should not be interpreted through the 68 V. LOOKING AHEAD: EMERGING INSIGHTS AND LESSONS lens of preexisting narratives. While future possibilities need to be fully mapped out, recent experiences point to practices that may hinder the path to universality. These may include, for example, approaches premised on the traditional nuclear family; programs informed by collection of data occurring with large time lags; benefits that are not spatially portable; bypassing provisions to large informal urban settlements; scale up of assistance based on discretionary policy decisions; and fostering an overall climate of suspicion and distrust towards low-income populations. Covid has enlarged the share of populations exposed to risk, potentially encompassing the entire country population. This means that while countries build their trajectory toward universality, a precondition for such journey would be the availability of universal delivery systems. This would include the ability to outreach, identify, enroll, pay, and have interoperable databases that would operate on a universal basis. The experiences reviewed in this paper show that delivering cash transfers on a universal or quasi-universal basis is no easy task even in high income countries. The extent to which those functions should be digitized depends on both the contextual digital ecosystem and user perspectives and preferences. At what pace countries would travel toward utilizing universal systems in routine programs – i.e., ensuring social protection coverage, adequacy, and comprehensiveness for everyone in society – would depend on a variety of evolving political, economic, operational, and societal factors. Benchmarking simplicity. One aspect that characterized Covid responses is the simplification of design. Would such simplicity compromise program performance? While the documentation of those practices has been extensive, relatively little is known on whether leaner design may have affected performance compared to pre- pandemic features. The preliminary evidence reviewed in this paper points to the significant contribution of transfers to welfare; but as countries restore traditional design, the idea that programs can be made simpler, leaner, and customer-friendly without necessarily jeopardizing effects is a hallmark of the pandemic response. While comparative evidence on the performance of specific program “with and without” Covid simplification is limited, the pandemic might further foment the benchmarking of programs against the plain-vanilla form of unconditional cash transfers (Kondylis and Loeser 2021). It might be interesting to extend this line of inquiry beyond overall design choices (conditionality and integrated approaches) and benchmark particular delivery aspects as well – for example, whether enrolment could be kept remote instead of in-person; or whether checks on declaration of income could be pursued ex-post on a sample basis instead of ex-ante for everyone. Information on global social protection responses. Extensive global response monitoring and tracking initiatives were put in place. These entailed a significant level of effort in identifying, codifying, cross-checking, and explaining social protection data. This includes data on basic performance metrics like coverage, adequacy, duration, extensions, and spending. Information was not always published, searchable and accessible. When made available, CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 69 countries often adopted different definitions and reporting procedures. The experience from extensive Covid-19 social protection tracking suggests the need for more proactive, comprehensive, coherent, and coordinated reporting systems. Perhaps the development of global standards for social protection response reporting might enhance future tracking endeavors. Initiatives that may help institutionalize the collection, analysis, and dissemination of global data in crisis times might also be better explored. Similarly, there appears to be growing demand for knowledge sharing across the country income spectrum: as governments of Pakistan, Nigeria and others have argued,66 the pandemic has signaled the need for a permanent virtual room or “knowledge lighthouse”. This might include a regular, virtual government- to-government community of practices offering real-time exchange of strategic and operational issues on social protection. If the present can be interpreted through a global lens, the future can be best understood at country level. The response is global in the sense that a large constellation of different lessons and experiences becomes available on the global stage almost real-time. Such knowledge base can be tapped, studied, and modelled to shape present decisions in different contexts. Global knowledge is in a fluid, constantly evolving state that can, if examined systematically, help take the pulse of the current state of social protection. But it is at country level that the fusion of specific political, social, institutional, and fiscal factors occurs – and these are the forces molding the prospects for social protection. Global developments and actors can, to varying degree, influence the formation of future country trajectories; but ultimately, it’s the set of country factors that steer the direction. That’s where we should look for clues on the future. Historical antecedents and current experience suggest four scenarios on if and how countries will build on the Covid response experience:67 • “The ostrich”. Under this scenario, governments bury their heads in the sand and no change happens. The implication would be to sleepwalk into the next crisis. Policymakers may not grasp the depth and structural nature of problem; and a priori adversarial, ideologically driven stances against social protection harden. Policymakers may be skeptical or dismissive of evidence, and may not be pressed enough for change. This scenario further cements a logic of distrust toward social protection and the need for more social protection in the future. • “The refresher”. This scenario involves governments that do take action, but in a reactive mode. They make some improvements at the margin, but largely keep structures intact. No major system-wide enhancement is made, with the extension of or some increase in coverage of particular programs being the main post-Covid changes. • “The constrained reformer”. This and the former scenario are those that most closely resemble current models. Governments are more proactive than in the refresher scenario, including seriously re-examining programs and initiatives. There is a wider appetite for substantive public discussion and consultation. Yet the constrained reformer “doesn’t pull the trigger” for more thorough and 70 V. LOOKING AHEAD: EMERGING INSIGHTS AND LESSONS deliberate innovation. Improvements are less system-wide, and more driven by pragmatic results. The agenda is mostly championed by “social” ministries in a state of constant justification to more conservative and influential actors shaping budgetary decisions. • “Beveridge redux”. Here governments embark on a comprehensive agenda of reform informed by a thorough assessment and rethinking of options. This involves a Beveridge-like, system-wide vision with the ambition to seize Covid as a true opportunity for change. Countries going this path would become trailblazers on particular innovations, and may likely bring to scale ideas that were hitherto only considered as aspirational and conducted on a pilot basis. Under this scenario, countries would combine technical reforms with changes in mindsets. The state would embrace change, take more responsibility, and accept more accountability. It successfully makes the case that social protection encourages common interests, and frames it as an integral part of wealth generation instead of an antagonizing force to economic upward mobility. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 71 VI. REFERENCES Aaron, H. (2020) “The Social Safety Net: The Gaps that COVID-19 Spotlights”. In Bethel, T. (ed) The Future of Social In- surance: Insights from the Pandemic – Reflections from 15 Years of Ball Award Recipients. National Academy of Social Insurance. Washington DC. Abay, K., Berhane, G., Hoddinott, J. and Tafere, K. (2020) “COVID-19 and Food Security in Ethiopia: Do Social Protection Programs Protect?” IFPRI, Discussion Paper 1972. Washington DC. Abbott, A. (2021) “COVID’s Mental-Health Toll: How Scientists Are Tracking a Surge in Depression”. Nature 590: 194-195. Abdoul-Azize, H. and El Gamil, R. (2020) “Social Protection as Key Tool in Crisis Management: Learnt Lessons from the COVID-19 Pandemic”. Global Social Welfare 8: 107-116. Abdulai, A., and Ibrahim, M. (2021) “Ghana’s Social Policy Response to Covid-19: Leaving the Poorest Behind?” CRC 1342, Covid-19 Social Policy Response Series 31/2021. Bremen. Acemoglu, D. (2020) “The Post-COVID State”. Project Syndicate (June 5). Adu-Ababio, K., Osei, R., Pirttilä, J. and Rattenhuber, P. (2021) “Distributional Effects of the COVID-19 pandemic in Gha- na”. UNU-WIDER, Policy Brief 4/2021. Helsinki. Aggarwal, S., Jeong, D., Kumar, N., Park, D., Robinson, J. and Spearot, A. (2020) “Did COVID-19 Market Disruptions Dis- rupt Food Security? Evidence from Households in Rural Liberia and Malawi”. NBER, Working Paper 27932. Cambridge. Aidukaite, J., Saxonberg, S., Szelewa, D., and Szikra, D. (2021) “Social Policy in the Face of a Global Pandemic: Policy Responses to the COVID‐19 Crisis in Central and Eastern Europe”. Social Policy and Administration 55(2): 358-373. Aiken, E., Bellue, S., Karlan, D., Udry, C. and Blumenstock, J. (2022) “Machine Learning and Phone Data Can Improve Targeting of Humanitarian Aid”. Nature 603: 864-870. Aizer, A., Hoynes, H. and Lleras-Muney, A. (2022) “Children and the US Social Safety Net: Balancing Disincentives for Adults and Benefits for Children”. Journal of Economic Perspectives 36(2): 149-174. Albaddawi, M., Boncenne, A., de Hoop, J., Lee, A., Natali, L., McNaughton, M., and Rodriguez Pumarol, M. (2020) “How a Displacement Crisis Helped Jordan Support Its Population during COVID-19”. UNICEF blog (May 20). New York. Alfers, L., Chen, M. and Plagerson, S. (2022) Informal Workers and Social Contracts in the Global South. Edward Elgar. Alfers, L., Ismail, G. and Valdivia, M. (2020) “Informal Workers and the Social Protection Response to COVID-19: Who Got Relief? How? And Did It Make a Difference?” WIEGO, Policy Insights 2. Manchester. Almenfi, M., Breton, M., Dale, P., Gentilini, U., Pick, A. and Richardson, D. (2020) “Where is the Money Coming From? Ten Stylized Facts on Financing Social Protection Responses to Covid-19”. OECD, UNICEF, and World Bank, Social Protec- tion and Jobs Policy and Technical Note 23. Washington DC. Alvarez-Iglesias, A., Hessel, P., Bauer, A. and Evans-Lacko, S. (2021) “Extending COVID-Related Reforms to Conditional Cash Transfers Could Improve the Life Chances of Young People in Colombia”. LSE blog (May 10). London. Aminjonov, U., Bargain, O., and Bernard, T. (2021) “Gimme Shelter: Social Distancing and Income Support in Times of Pandemic”. IFPRI, Discussion Paper 2042. Washington DC. Anderson, S., Rowe, G., Toby, C., Gentilini, U., Paternostro, S. and Zafar, U. (2022) “Review of Emerging Lessons from the Initial Wave of World Bank Social Protection Responses to Covid-19”. FEG and World Bank. Washington DC. Ando, M., Furukawa, C., Nakata, D. and Sumiya, K. (2020) “Fiscal Responses to the COVID-19 Crisis in Japan: The First Six Months”. National Tax Journal 73(3): 901-926. 72 VI. REFERENCES Arezki, R. (2021) “Cash Payments in Africa Could Boost Vaccine Uptake”. Nature 596: 9. Arruda, P., Lazarotto de Andrade, M., Falcao, T., Teixeira Barbosa, D. and Morgandi, M. (2021) “Auxílio Emergencial: Lessons from the Brazilian Experience Responding to COVID-19 – Summary Note”. World Bank and FCDO. Brasilia. Arza, C. (2021) “Argentina’s Social Policy Response to Covid-19: Protecting Income and Employment”. CRC 1342, Covid-19 Social Policy Response Series 17/2021. Bremen. Asfaw, A. (2021) “The Effect of Income Support Programs on Job Search, Workplace Mobility and COVID-19: Interna- tional Evidence”. Economics and Human Biology 41(C): 1-11. Aslanyan, G., Baghdasaryan, V., and Shakhmuradyan, G. (2021) “Armenia’s Social Policy Response to Covid-19: Mitigat- ing Expectations, Financial Stress, and Anxiety”. CRC 1342, Covid-19 Social Policy Response Series 10/2021. Bremen. Auerbach, A. and Thachil, T. (2020) “How Does Covid-19 Affect Urban Slums? Evidence from Settlement Leaders in India”. World Development 140: 1-11. Baker, S., Farrokhnia, R., Meyer, S., Pagel, M. and Yannelis, C. (2020) “Income, Liquidity, and the Consumption Response to the 2020 Economic Stimulus Payments”. NBER, Working Paper 27097. Cambridge. Baldizón, A. (2021) “Nicaragua’s Social Policy Response to Covid-19: Business as Usual at All Costs” CRC 1342, Covid-19 Social Policy Response Series 29/2021. Bremen. Bance, P. and Gentilini, U. (2020) “Suspended Until Further Notice”: Is There a Role for Public Works in COVID-19 (Coro- navirus) Response?”. World Bank, Jobs and Development Blog 7. Washington DC. Bance, P., Bermeo, L. and Kabemba, F. (2021) “Cash and the City: Digital COVID-19 Social Response in Kinshasa”. Brook- ings, Future Development (September 8). Washington DC. Banerjee, A., Faye, M., Krueger, A., Niehaus, P. and Suri, T. (2020) “Effects of a Universal Basic Income during the Pan- demic”. Working paper. Cambridge. Barca, V. (2021) “Useful COVID-19 and Social Protection Materials”. Social Protection Approaches to COVID-19 Expert Advice Service (SPACE). London. Barca, V., Hebbar, M., Cote, A., Schoemaker, E., Enfield, S., Holmes, R. and Wylde, E. (2021) “Inclusive Information Sys- tems for Social Protection: Intentionally Integrating Gender and Disability”. SPACE. London and Bonn. Barnes, H., Espi-Sanchis, G., Leibbrandt, M., McLennan, D., Noble, M., Pirttilä, J., Steyn, W., Van Vrede, B. and Wright, G. (2021) “Analysis of the Distributional Effects of COVID-19 and State-led Remedial Measures in South Africa”. Interna- tional Journal of Microsimulation 14(2): 2-31. Barrett, C. (2020) “Actions Now Can Curb Food Systems Fallout from COVID-19”. Nature Food 1(June): 319-320. Barrientos, A. (2021) “Should We Expect a Post-Covid-19 Social Protection Epiphany in Latin America?” Global Devel- opment Institute Blog (February 18). Barrientos, A. (2020) “COVID-19 and Social Protection in Latin America”. Bulletin of Latin American Research 39(S1): 52-55. Baskaran, G., Bhorat, H. and Köhler, T. (2020) “South Africa’s Special COVID-19 Grant: A Brief Assessment of Coverage and Expenditure Dynamics”. DPRU, Policy Brief 202055. Cape Town. Bassier, I., Budlender, J., Zizzamia, R., Leibbrandt, M., and Ranchhod, V. (2021) “Locked Down and Locked Out: Repur- posing Social Assistance as Emergency Relief to Informal Workers”. World Development 139(C): 1-23. Bastagli, F. and Lowe, C. (2021) “Social Protection Response to Covid-19 and Beyond: Emerging Evidence and Learning for Future Crises”. ODI, Working Paper 614. London. Bauer, A., Garman, E., McDaid, D., Avendano, M., Hessel, P., Díaz, Y., Araya, R., Lund, C., Malvasi, P., Matijasevich, A., A-La Park, Silvestre Paula, C., Ziebold, C., Zimmerman, A., and Evans-Lacko, S. (2021) “Integrating Youth Mental Health into Cash Transfer Programmes in Response to the COVID-19 Crisis in Low-Income and Middle-Income Countries”. The Lancet Psychiatry 8(4): 340-346. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 73 Bauer, L., Pitts, A., Ruffini, K., and Whitmore Schanzenbach, D. (2020) “The Effect of Pandemic EBT on Measures of Food Hardship”. Brookings Institution, The Hamilton Project. Washington DC. Bauer, L., Ruffini, K., and Whitmore Schanzenbach, D. (2021) “An Update on the Effect of Pandemic EBT on Measures of Food Hardship – Technical Appendix”. Brookings Institution, The Hamilton Project. Washington DC. Beazley, R., Marzi, M., and Steller, R. (2021) “Drivers of Timely and Large-Scale Cash Responses to COVID-19: What Does the Data Say?” SPACE. London and Bonn. Behsudi, A. (2022) “Technology-Driven Development: Togo’s Cina Lawson Views Digitalization as Key to Her Country’s Growth and Development”. IMF, Finance and Development (March): 44-45. Beland, D., Cantillon, B., Hick, R., and Moreira, A. (2021) “Social Policy in the Face of a Global Pandemic: Policy Re- sponses to the COVID-19 Crisis”. Social Policy and Administration 55(2): 249-260. Béland, D., Cantillon, B., Hick, R., and Moreira, A. (2021) “Social policy in the face of a global pandemic: Policy respons- es to the COVID‐19 crisis” Social Policy and Administration, 55(2), 249-260. Béland, D., Dinan, S., Rocco, P., and Waddan, A. (2021) “Social Policy Responses to COVID‐19 in Canada and the United States: Explaining Policy Variations between Two Liberal Welfare State Regimes”. Social Policy and Administration 55(2): 280-294. Berkouwer, S., Biscaye, P., Puller, S., Wolfram, C., (2022) “Disbursing Emergency Relief Through Utilities: Evidence from Ghana”. Journal of Development Economics 156(C), in press. Berner, H. and van Hemelryck, T. (2020) “Sistemas de Información Social y Registros de Destinatarios de la Protec- ción Social No Contributiva en América Latina: Avances y Desafíos Frente al COVID‐19”. CEPAL, Santiago. Betliy, O. (2021) “Ukraine’s Social Policy Response to Covid-19: A Quick but Fragmented Approach”. CRC 1342, Covid-19 Social Policy Response Series 24/2021. Bremen. Bhan, G., Chowdhury, A., Margosa, N., Sampat, K. and Sohane, N. (2020) “Lessons for Social Protection from the COVID-19 Lockdowns Report 1 of 2: State Actors”. Indian Institute for Human Settlements. Bangalore and Delhi. Bhattacharya, S. and Sinha Roy, S. (2021) “Intent to Implementation: Tracking India’s Social Protection Response to COVID-19”. World Bank, Social Protection and Jobs Discussion Paper 2107. Washington, DC. Bhorat, H., Oosthuizen, M. and Stanwix, B. (2021) “Social Assistance Amidst the COVID‐19 Epidemic in South Africa: A Policy Assessment”. South African Journal of Economics 89(1): 63-81. Bierbaum, M., Orton, I. and Behrendt, C. (2021) “Towards Solid Social Protection Floors? The Role of Non-Contributory Provision during the COVID-19 Crisis and Beyond”. ILO, Brief. Geneva. Bista, C., Nguyen, Q. and Stern-Plaza, M. (2021) “Extending Unemployment Protection as a Response to the COVID-19 Crisis and as Part of Efforts to Build Back Better”. In Peres et al (eds) What’s Next for Social Protection in Light of COVID-19: Challenges Ahead. Brasilia. Blattman, C., Cerero, D., Duncan, G., Hernandez, S., Lessing, B., Martínez, J., Mesa-Mejía, J., Montoya, H., and Tobón, S. (2020) “Crime in the Time of COVID-19: How Colombian Gangs Responded to the Pandemic”. IPA, University of Chicago, and Universidad EAFIT. Chicago. Blofield, M., Giambruno, C. and Pribble, J. (2020) “Policy Expansion in Compressed Time: Assessing the Speed, Breadth and Sufficiency of Post-COVID-19 Social Protection Measures in 10 Latin American countries”. ECLAC, Social Policy Series 235. Santiago. Blofield, M., Giambruno, C. and Pribble, J. (2021a) “Breadth and Sufficiency of Cash Transfer Responses in Ten Latin American Countries During the First 12 Months of the Covid-19 Pandemic”. CEQ, Working Paper 114. New Orleans. Blofield, M., Lustig, N. and Trasberg, M. (2021b) “Social Protection During the Pandemic: Argentina, Brazil, Colombia, and Mexico”. CGD Note. Washington DC. 74 VI. REFERENCES Borges, F. A. (2021) “Bolivia’s Social Policy Response to Covid-19: Hindered by Political and Economic Crises”. CRC 1342, Covid-19 Social Policy Response Series 19/2021. Bremen. Bottan, N., Hoffmann, B. and Vera-Cossio, D. (2020) “Stepping Up During a Crisis: The Unintended Effects of a Noncon- tributory Pension Program during the Covid-19 Pandemic”. IDB, Working Paper 1153. Washington DC. Bourgault, S. and O’Donnell, M. (2020) “Women’s Access to Cash Transfers in Light of COVID-19: The Case of Pakistan”. CGD, Policy Note. Washington DC. Boushey, H. and Shambaugh, J. (2019) “Introduction”. In Boushey et al (eds) Recession Ready: Fiscal Policies to Stabi- lize the American Economy. Brookings Institution. Washington, DC. Bowen, T., del Ninno, C., Andrews, C., Coll-Black, S., Gentilini, U., Johnson, K., Kawasoe, Y., Kryeziu, A., Maher, B., and Williams, A. (2020) Adaptive Social Protection: Building Resilience to Shocks. World Bank. Brooks, W., Donovan, K., Johnson, T. and Oluoch-Aridide, J. (2022) “Cash Transfers as a Response to COVID-19: Experi- mental Evidence from Kenya”. Journal of Development Economics, in press. Brown, C., Ravallion, M. and van de Walle, D. (2020) “Can the World’s Poor Protect Themselves from the New Corona- virus?” NBER, Working Paper 272. Cambridge. Brubaker, R., Day, A. and Huvé, S. (2021) “COVID-19 and Humanitarian Access: How the Pandemic Should Provoke Sys- temic Change in the Global Humanitarian System”. UNU. New York. Brum, M. and De Rosa, M. (2021) “Too Little but Not Too Late. Nowcasting Poverty and Cash Transfers’ Incidence in Uruguay during COVID-19’s Crisis”. World Development 140: 1-11. Bui, D., Dräger, L., Hayo, B. and Hong Nghiem, G. (2022) “The Marginal Propensity to Consume During the COVID-19 Pandemic: Evidence from Thailand and Vietnam”. University of Hannover, Hannover Economic Papers 695. Han- nover. Busso, M., Camacho, J., Messina, J. and Montenegro, G. (2021) “Social Protection and Informality in Latin America during the COVID-19 Pandemic”. PLoS-ONE 16(11): 1-15. Calcutt, E., Maher, B. and Fitzgibbon, C. (2021) “Emerging Lessons in Financing Adaptive Social Protection”. World Bank. Washington DC. Cantillon, B., Seeleib‐Kaiser, M., and van der Veen, R. (2021) “The COVID‐19 Crisis and Policy Responses by Continental European Welfare States”. Social Policy and Administration 55(2): 326-338. Carranza, E., Farole, T., Gentilini, U., Morgandi, M., Packard, T., Santos, I. and Weber, M. (2020) “Managing the Employ- ment Impacts of the COVID-19 Crisis: Policy Options for Relief and Restructuring”. World Bank, Jobs Working Paper 49. Washington DC. CBPP (2021) “States Are Using Much-Needed Temporary Flexibility in SNAP to Respond to COVID-19 Challenges”. Washington DC. Chipenda, C., and Tom, T. (2021) “Zimbabwe’s Social Policy Response to Covid-19: Temporary Food Relief and Cash Transfers”. CRC 1342, Covid-19 Social Policy Response Series 23/2021. Bremen. Cho, Y. and Johnson, D. (2022) “COVID-19 and Social Assistance in the Philippines: Lessons for Future Resilience”. World Bank. Manila. Cho, Y., Avalos, J., Kawasoe, Y., Johnson, D. and Rodriguez, R. (2021) “The Impact of the COVID-19 Pandemic on Low Income Households in the Philippines: Impending Human Capital Crisis”. World Bank, COVID-19 Low Income HOPE Survey Note 3. Manila. Chodorow-Reich, G. and Coglianese, J. (2019) “Unemployment Insurance and Macroeconomic Stabilization”. In Boushey et al (eds) Recession Ready: Fiscal Policies to Stabilize the American Economy. Brookings Institution. Wash- ington, DC. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 75 Choudhuri, P., Pramanik, S. and Desai, S. (2022) “Urban Exclusion: Rethinking Social Protection in the Wake of the Pandemic in India”. NCAER, Working Paper 135. New Delhi. Chowdhury, A., Lawson, C., Kellison, E., Chia, H., Kharas, H., Fuller, J., Faye, M., Rutkowski, M., Salvado, R., and Dercon, S. (2022) “Accelerating Digital Cash Transfers to the World’s Poorest”. Brookings, Future Development (February 17). Washington DC. Christiano, L., Eichenbaum, M., and Rebelo, S. (2011) “When is the Government Spending Multiplier Large?” Journal of Political Economy 119(1): 78-121. Chung, H., and Thewissen, S. (2011). “Falling Back on Old Habits? A Comparison of the Social and Unemployment Crisis Reactive Policy Strategies in Germany, the UK and Sweden”. Social Policy and Administration 45(4): 354–370. Cichon, M. (2021) “The Vaccine Against Poverty, Inequality and Insecurity Needs a Stronger Prescription”. Develop- ment Pathways blog (February 17). London. Coibion, O., Gorodnichenko, Y. and Weber, M. (2020) “The Cost of the Covid-19 Crisis: Lockdowns, Macroeconomic Expectations, and Consumer Spending”. NBER, Working Paper 27141. Cambridge. Coleman-Jensen, A. and Rabbitt, M. (2021) “Food Pantry Use Increased in 2020 for Most Types of U.S. Households”. USDA, Economic Research Service. Washington, DC. Cook, S. and Ulriksen, M. (2021) “Social Policy Responses to COVID-19: New Issues, Old Solutions?” Global Social Policy 21(3): 381-395. Cooper, H. and Szreter, S. (2021) After the Virus: Lessons from the Past for a Better Future. Cambridge University Press. Crossley, T., Fisher, P., Levell, P. and Low, H. (2022) “Stimulus Payments and Private Transfers”. European University Institute, IFS, ISER, and University of Oxford. London. Dasgupta, S. and Robinson, E. (2022) “Impact of COVID-19 on food insecurity using multiple waves of high frequency household surveys”. Nature (Scientific Reports) 12(1865): 1-15. De La Flor, L., Mujica, I., Fontenez, M., Newhouse, D., Rodriguez Alas, C., Sabharwal, G., and Weber, M. (2021) “Taking Stock of COVID-19 Labor Policy Responses in Developing Countries”. World Bank, Jobs Watch COVID-19. Washington DC. De la Porte, C., and Jensen, M. D. (2021) “The next generation EU: An Analysis of the Dimensions of Conflict Behind the Deal”. Social Policy and Administration 55(2): 388-402. de Leon, F., Malde, B. and Mcquillin, B. (2021) “The Effects of Emergency Government Cash transfers on Beliefs and Behaviours during the COVID Pandemic: Evidence from Brazil”. Working Paper (v. November 5). University of Kent and University of East Anglia. Kent and Norwich. Debenedetti, L. (2021) “Togo’s Novissi Cash Transfer: Designing and Implementing a Fully Digital Social Assistance Program during COVID-19”. IPA, RECOVR Case study. Washington DC. Demuria, V., and Absandze, T. (2021) “Georgia’s Social Policy Response to Covid-19: Targeted Social Assistance”. CRC 1342, Covid-19 Social Policy Response Series 13/2021. Bremen. Deparday, V., Nuala, M., Cowan, C., and Gevaert, I. (2022) “Screens Over Shovels: Using Digital Technology to Modern- ize Public Works”. World Bank, Sustainable Cities blog (February 11). Washington DC. Development Initiatives (2021) Global Humanitarian Assistance Report 2021. Bristol. Devereux, S. (2021) “Social Protection Responses to COVID-19 in Africa”. Global Social Policy 21(3): 421-447. Devereux, S., Béné, C., and Hoddinott, J. (2020) “Conceptualising COVID-19’s Impacts on Household Food Security”. Food Security 12(4): 769-772. Dhemba, J. (2021) “Eswatini’s Social Policy Response to Covid-19: Temporary and Donor-Driven” CRC 1342, Covid-19 Social Policy Response Series 30/2021. Bremen. 76 VI. REFERENCES Duchoslav, J. and Hirvonen, K. (2021) “Delivery of Social Protection Programs to Combat COVID-19 in Africa”. In Ulimwen- gu et al (eds) Building Resilient African Food Systems after COVID-19. ReSAKSS and IFPRI. Kigali and Washington, DC. Duran-Valverde, F., Pacheco-Jiménez, J., Muzaffar, T. and Elizondo-Barboza, H. (2020) “Financing Gaps in Social Pro- tection: Global Estimates and Strategies for Developing Countries in Light of the COVID-19 Crisis and Beyond”. ILO, Working Paper 14. Geneva. Ebbinghaus, B., Lehner, L. and Naumann, E. (2022) “Welfare State Support During the COVID- 19 pandemic: Change and Continuity in Public Attitudes Towards Social Policies in Germany”. European Policy Analysis, in press. ECHO (2022) “Cash Transfers”. Thematic Policy Document 3. Brussels. ESCAP and ILO (2020) “The Protection We Want: Social Outlook for Asia and the Pacific”. Bangkok. Escobar Correa, P., de la Cadena, S., Gentilini, U., Rodriguez, L., and Wiseman, W. (2021) “Strengthening Social Protec- tion”. World Bank, background paper for Colombia’s PFR. Bogota. Eurofound (2021) “Emergency Income: Factsheet for case IT-2020-21/996 (Updated) – Measures in Italy”. Brussels. Farnsworth, K. and Irving, Z. (2011) Social Policy in Challenging Times: Economic Crisis and Systems. Policy Press. Feldman, N. and Heffetz, O. (2020) “A Grant to Every Citizen: Survey Evidence of the Impact of a Direct Government Payment in Israel”. NBER, Working Paper 28312. Cambridge. Fiala, N., Rose, J., Aryemo, F. and Peters, J. (2021) “Cash Transfers and Shocks: Long-Term Effects During a Pandemic in Uganda”. G²LM|LIC Policy Brief 43. Bonn. Furman, J. (2020) “The Fiscal Response to the Great Recession: Steps Taken, Paths Rejected, and Lessons for Next Time”. In Bernanke et al (eds) First Responders: Inside the U.S. Strategy for Fighting the 2007-2009 Global Financial Crisis. Yale University Press. Furman, J. and Summers, L. (2020) “A Reconsideration of Fiscal Policy in the Era of Low Interest Rates”. Brookings Institutions. Washington, DC. Gelb, A. and Mukherjee, A. (2020) “Digital Technology in Social Assistance Transfers for COVID - 19 Relief: Lessons from Selected Cases”. CGD Policy Paper 181. Washington DC. Gelb, A., Giri, A., Mukherjee, A., Rautela, R., Thapliyal, M., and Webster, B. (2022a) “Beyond India’s Lockdown: PMGKY Benefits During the COVID-19 Crisis and the State of Digital Payments”. CGD, Policy Paper 257. Washington, DC. Gelb, A., Mukherjee, A. and Webster, B. (2022b) “Citizens and States Projects on Digital Services and Payments: Recap, Progress and Lessons”. CGD, PPT presentation (May 27). Washington DC. Gelb, A., Mukherjee, A. and Webster, B. (2022c) “Can Digital G2P Transfers Drive Financial Inclusion and Digital Pay- ments? Evidence from India”. CGD, Working Paper 618. Washington DC. Gentilini, U. (2021) “A Game Changer for Social Protection? Six Reflections on COVID-19 and the Future of Cash Trans- fers”. World Bank, Let’s Talk Development Blog (January 11). Gentilini, U., Almenfi, M., TMM Iyengar, H., Okamura, Y., Downes, J. A., Dale, P., Weber, M., Newhouse, D., Rodriguez Alas, C., Kamran, M., Mujica, I., Fontenez, M., Ezzat, M., Asieduah, S., Martinez, V., Hartley, G., Demarco, G., Abels, M., Zafar, U., Urteaga, E., Valleriani, G., Muhindo, J., and Aziz, S. (2022a) “Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures”. World Bank, Version 16 (February 2). Washington DC. Gentilini, U. Almenfi, M., TMM Iyengar, H., Okamura, Y., Urteaga, E., Valleriani, G., Muhindo, J., and Aziz, S. (2022b) “Tracking Global Social Protection Responses to Price Shocks”. World Bank, Social Protection and Jobs Discussion Paper 2208. Washington DC. Gentilini, U., Almenfi, A., Blomquist, J., Dale, P., De La Flor Giuffra, L., Desai, V., Fontenez, M., Galicia Rabadan, G., Lopez, V., Marin Espinosa, A., Natarajan, H., Newhouse, D., Palacios, R., Quiroz, Ana., Rodriguez Alas, C., Sabharwal, G. and Weber, M. (2021a) ““Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures”. World Bank, Version 15 (May 14). Washington DC. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 77 Gentilini, U., Khosla, S. and Almenfi, M. (2021b) “Cash in the City: Emerging Lessons from Implementing Cash Transfers in Urban Africa”. World Bank, Social Protection and Jobs Discussion Paper 2101. Washington DC. Georgarakos, D. and Kenny, G. (2022) “Household Spending and Fiscal Support During the Pandemic – The Role of Public Perceptions”. European Central Bank, Research Bulletin 94 (April 21). Brussels. Gerard, F., Imbert, C. and Orkin, K. (2020) “Social Protection Response to the COVID19 Crisis: Options for Developing Countries”. Oxford Review of Economic Policy (August 29). GIZ, AU and World Bank (2021) “Towards a Shared Understanding of Digital Social Protection: What Are the Lessons from the COvid-19 Crisis and the Implications for the Future? Report from a Regional Workshop in Africa”. Bonn. Goldfayn, O., Lewis, V. and Wehrhöfer, N. (2022) “Spending Effects of Child-Related Fiscal Transfers”. CEPR, Discussion Paper 17058. London. Goolsbee, A. and Syverson, C. (2020) “Fear, Lockdown, and Diversion: Comparing Drivers of Pandemic Economic De- cline 2020”. NBER, Working Paper 27432. Cambridge. GoSA (Government of South Africa) (2022) “Social Assistance Act, 2004: Regulations Relating to Covid-19 Social Relief Distress Issued in Terms of Section 32, Read with Section 13, of the Social Assistance Act, 2004 (Act No.13 of 2004), as Amended and with the Concurrence of the Minister of Finance, Made the Regulation of the Schedule”. Department of Social Development, Government Notices 46271 (April 22). Pretoria. Grand Bargain (2022) “Cash Coordination Caucus: Outcomes and Recommendations”. Geneva. Gray Molina, G. and Ortiz-Juarez, E. (2020) “Temporary Basic Income: Protecting Poor and Vulnerable People in De- veloping Countries”. UNDP. New York. Greve, B., Blomquist, P., Hvinden, B., and van Gerven, M. (2021) “Nordic Welfare States—Still Standing or Changed by the COVID‐19 Crisis?”. Social Policy and Administration 55(2): 295-311. Gronbach, L., Seekings, J. and Megannon, V. (2022) “Social Protection in the COVID-19 Pandemic: Lessons from South Africa”. CGD, Policy Paper 252. Washington DC. Gudmalin, C., Perante-Calina, L., Balbosa, J., Mangahas, J. and Aloha Samoza, M. (2021) “Protecting the Poor and Vul- nerable against the Pandemic”. ADBI, Development Case Study 2021-3 (September). Manila. Guliyev, F. (2021) “Azerbaijan’s Social Policy Response to Covid-19: Swift Response, Modest Benefits”. CRC 1342, Covid-19 Social Policy Response Series 8/2021. Bremen. Gupta, S. and Tovar Jalles, J. (2022) “Can COVID-19 Induce Governments to Implement Tax Reforms in Developing Countries?” Applied Economics 54(20): 2288-2301. Hagen-Zanker, J. and Both, N. (2021) “Social Protection Provisions to Refugees during the Covid-19 Pandemic: Lessons Learned from Government and Humanitarian Responses”. ODI, Working Paper 612. Washington DC. Hale, T., Angrist, N., Goldszmidt, R., Kira, B., Petherick, A., Phillips, T., Webster, S., Cameron-Blake, E., Hallas, L., Majum- dar, S. and Tatlow, H. (2021) “A Global Panel Database of Pandemic Policies (Oxford COVID-19 Government Response Tracker)”. Nature Human Behaviour 5: 529-538. Hammad, M., Bacil, F. and Veras Soares, F. (2021) “Next Practices: Innovations in the COVID-19 Social Protection Re- sponses and Beyond’. IPC-IG. Brasilia. Hassan, S. (2021) “Pakistan’s Social Policy Response to Covid-19: Expansion of the Ehsaas Program” CRC 1342, Covid-19 Social Policy Response Series 34/2021. Bremen. Hattori, T., Komura, N., and Unayama, T. (2021) “Impact of Cash Transfers on Consumption During the COVID-19 Pan- demic: Evidence from Japanese Special Cash Payments”. RIETI Discussion Paper 21-E-043. Tokyo. Henry-Lee, A., and Jones, J. (2021) “Jamaica’s Social Policy Response to Covid-19: Societal Inequalities Laid Bare” CRC 1342, Covid-19 Social Policy Response Series 32/2021. Bremen. 78 VI. REFERENCES Hick, R., and Murphy, M. P. (2021) “Common Shock, Different Paths? Comparing Social Policy Responses to COVID‐19 in the UK and Ireland”. Social Policy and Administration 55(2): 312-325. Hill, R., Patel, D., and Plichtal, M. (2020) “What Have We Learned After Six Months of Tracking Covid-19 Funding? We Don’t Just Need More Money, We Need a Different Approach”. Center for Disaster Protection. London. Holmes, R. (2021) “Can the COVID-19 Crisis be a Catalyst for Gender-Responsive and Inclusive Social Protection?” SPACE blog post (March 6). Hoynes, H. and Whitmore Schanzenbach, D. (2019) “Strengthening SNAP as an Automatic Stabilizer”. In Boushey et al (eds) Recession Ready: Fiscal Policies to Stabilize the American Economy. Brookings Institution. Washington, DC. ILO (2021a) “Assessment of the Social Security: Responses to COVID-19 Lessons from the Western Balkans and Eastern Europe during the First Stage of the Pandemic”. Geneva. ILO (2021b) World Social Protection Report 2020-2022. Geneva. ILO (2020a) “COVID-19 and the World of Work. Country Policy Responses: Thailand”. Geneva ILO (2020b) “Social Protection Policy Responses to Mitigate COVID-19 Impact in Vietnam”. Geneva. IMF (2022) Fiscal Monitor (April): Fiscal Policy from Pandemic to War. Washington DC. Immervoll, H., Fernández. R., Hyee, R., Lee, J., and Pacifico, D. (2022) “De-Facto Gaps in Social Protection for Standard and Non-standard Workers: An Approach for Monitoring the Accessibility and Levels of Income Support”. IZA, Discus- sion Paper 15289. Bonn. IPC-IG, UNICEF, and WFP (2021) “Social Protection and Venezuelan Migration in Latin America and the Caribbean in the Context of COVID-19”. Brasilia. Jain, R., Budlender, J., Zizzamia, R., and Bassier, I. (2020) “The Labor Market and Poverty Impacts of COVID-19 in South Africa”. CSAE, Working Paper 2020-14. Oxford. Jara, H., Montesdeoca, L. and Tasseva, I. (2021) “The Role of Automatic Stabilizers and Emergency Tax–Benefit Policies During the COVID-19 Pandemic in Ecuador”. UNU-WIDER, Working Paper 2021/4. Helsinki. Jaroszewicz, A., Jachimowicz, J., Hauser, O. and Jamison, J. (2022) “How Effective Is (More) Money? Randomizing Uncon- ditional Cash Transfer Amounts in the US”. Working paper. Cambridge and Exeter. Jodar, J., Kondakhchyan, A., McCormack, R., Peachey, K., Phelps, L. and Smith, G. (2020) The State of the World’s Cash Report. London. Johnson, K., Tisei, F., Walker, T. and Wanniarachchi, T. (2022) Responsive by Design: Building Adaptive Social Protection Systems in South Asia. (Eds) World Bank, Washington DC. In press. Kaasch, A., Starke, P. and van Hooren, F. (2010) “The Welfare State as Crisis Manager? Comparing Social Policy Re- sponses to Three Major Economic Crises”. Paper presented at the American Political Science Association’s Annual Meeting (September 2-5, Washington DC). Bremen. Kalikeka, M., Bwalya, M., Adu-Ababio, K., Gasior, K., McLennan, D., and Rattenhuber, P. (2021) “Distributional effects of the COVID-19 Pandemic in Zambia”. UNU-WIDER, Policy Brief 2/2021. Helsinki. Kaneda, M., Kubota, S., and Tanaka, S. (2021) “Who Spent Their COVID-19 Stimulus Payment? Evidence from Personal Finance Software in Japan”. Japanese Economic Review 72(3): 409-437. Karger, E. and Rajan, A. (2020) “Heterogeneity in the Marginal Propensity to Consume: Evidence from Covid-19 Stimu- lus Payments”. Federal Reserve Bank of Chicago, Working Paper 2020–15. Chicago. Khan, A. and Jamy, G. (2020) “COVID-19 G2P Cash-Transfer Payments Country Brief: Pakistan”. World Bank. Washington DC. Kimani, E., Lanthorn, H., Stein, D. and Bergemann, D. (2021) “Cash transfers and COVID-19: Experiences from Kiryan- dongo, Uganda – Round 3 of 3”. IDInsight. London. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 79 Kisanga, E., Leyaro, V., Noble, M., Oliveira, R., Rattenhuber. P. and Wright, G. (2021) “Limited Impact of the Tax-Benefit System in Tanzania during the COVID-19 pandemic in 2020”. UNU-WIDER, Policy Brief 6/2021. Helsinki. Köhler, T. and Bhorat, H. (2021) “Can Cash Transfers Aid Labour Market Recovery? Evidence from South Africa’s Special COVID-19 Grant”. DPRU, Working Paper 202108. Cape Town. Kondylis, F. and Loeser, J. (2021) “Intervention Size and Persistence”. World Bank, Policy Research Working Paper 9769. Washington DC. Kone Consulting (2020) “Pandemic EBT Implementation Documentation Project”. Report prepared for CBPP and FRAC. Washington DC. Kubota, S., Onishi, K. and Toyama, Y. (2021) “Consumption Responses to COVID-19 Payments: Evidence from a Natural Experiment and Bank Account Data”. Journal of Economic Behavior and Organization 188 (August): 1-17oi Kuhfeld, M., Soland, J. and Lewis, K. (2022) “Test Score Patterns Across Three COVID-19-impacted School Years”. Brown University, EdWorking Paper 22-521. Providence. Kühner, S., Nakray, K., and Neff, D. (2021) “India’s Social Policy Response to Covid-19: Temporary Relief in a Rigid Wel- fare Landscape”. CRC 1342, Covid-19 Social Policy Response Series 4/2021. Bremen. Kumar, A., Saroj, S., Mishra, A. and Rashid, S. (2022) “Government Transfers, COVID-19 Shocks, and Food Insecurity: Evidence from Rural Households in India”. IFPRI, Discussion Paper 2110. Washington DC. Kurolova, D., and Schiek, S. (2021) “Uzbekistan’s Social Policy Response to Covid-19: A Test for the Mirziyoyev Regime”. CRC 1342, Covid-19 Social Policy Response Series 22/2021. Bremen. Lampraki C., Hoffman A., Roquet A., and Jopp D. (2022) “Loneliness During COVID-19: Development and influencing Factors”. PLoS ONE 17(3). Lastunen, J., Rattenhuber, P., Adu-Ababio, K., Gasior, K., Jara, X., Jouste, M., McLennan, D., Nichelatti, E., Oliveira, E., Pirttilä, J., Richiardi, M. and Wright, G. (2021) “The Mitigating Role of Tax and Benefit Rescue Packages for poverty and Inequality in Africa amid the COVID-19 Pandemic”. UNU-WIDER, Working Paper 2021/148. Helsinki. Lawson, C., Ekouhoho, K., Koudeka, M., Cárdenas Martínez, A., Iñaki Alberro Encinas, L., and George Karippacheril, T. (2022) “Novissi Togo: A Paradigm Shift for Social Protection Delivery Systems”. World Bank, forthcoming. Lawson-McDowall, J., McCormack, R. and Tholstrup, S. (2021) “The Use of Cash Assistance in the Covid-19 Humanitar- ian Response: Accelerating Trends and Missed Opportunities”. Disasters 45(S1): S216-S239. Legros, M. (2021) “Country Chapter France”. In Spasova et al (eds) “Social Protection of Non-Standard Workers and the Self-Employed During the Pandemic – Country Chapters: Belgium, France, Ireland, Italy, Lithuania, Portugal, Romania, Sweden”. ETUI, Report 2021.05. Brussels. Leisering, L. (2021) “Social Protection Responses by States and International Organisations to the COVID-19 Crisis in the Global South: Stopgap or New Departure?” Global Social Policy 21(3): 396-420 Levine, S., Both, N., Salomon, H., Socpa, A., and Fomekong, F. (2022) “Integrating Assistance to the Displaced into a Social Protection System in Cameroon. An Ideal, but in Whose Interests?” ODI. London. Levy, S. and Filho, N. (2022) “The Impact of the Covid Emergency Aid Transfers on Female Labor Supply in Brazil”. In- sper Institute of Education and Research. Sao Paulo. Lindert, K., George Karippacheril, T., Rodriguez Caillava, I. and Nishikawa Chavez, K. (2020) Sourcebook on the Foun- dations of Social Protection Delivery Systems. (Eds) World Bank, Washington DC. Lindert, P. (2021) Making Social Spending Work. Cambridge University Press. Longhurst, D., Evans, S., Connolly, D., Lung, F., McCord, A., Allan, S. and Plichta, M. (2021) “What Are Future Financing Options for Shock Responsive Social Protection? A Technical Primer”. SPACE. London and Bonn. Longhurst, D. and Smith, G. (2020) “Identifying Practical Options for Linking Humanitarian Assistance and Social Pro- 80 VI. REFERENCES tection in the COVID-19 Response”. SPACE. London and Bonn. Lowe, C., McCord, A. and Beazley, R. (2021) “National Cash Transfer Responses to Covid-19: Operational Lessons Learned for Social Protection System-Strengthening and Future Shocks”. ODI, Working Paper 610. London. Lustig, N., Martinez Pabon, V., Sanz, F. and Younger, S. (2020) “The Impact of COVID-19 Lockdowns and Expanded Social Assistance on Inequality, Poverty and Mobility in Argentina, Brazil, Colombia and Mexico”. CGD, Working Paper 556. Washington DC. Makkar, S., Manivannan, J., Swaminathan, S., Travasso, S., John, A., Webb, P., Kurpad, A. and Thomas, T. (2022) “Role of Cash Transfers in Mitigating Food Insecurity in India during the COVID-19 Pandemic: A Longitudinal Study in the Bihar State”. British Medical Journal 12(e060624): 1-9. Maltseva, E. (2021) “Kazakhstan’s Social Policy Response to Covid-19: Developmentalism and Productive Welfare”. CRC 1342, Covid-19 Social Policy Response Series 20/2021. Bremen. Mansour, H. (2021) “Morocco’s Social Policy Response to Covid-19: A Special Fund and a Structural Reform Proposal” CRC 1342, Covid-19 Social Policy Response Series 33/2021. Bremen. Marcus, M. and Yewell, K. (2022) “The Effect of Free School Meals on Household Food Purchases: Evidence from the Community”. NBER, Working Paper 29395. Cambridge. Markhof, Y. (2020) “Pakistan’s Social Protection Response to the Covid-19 Pandemic: The Adequacy of Ehsaas Emer- gency Cash and Road Ahead”. IPC-IG, Working Paper 188. Brasília. Massard da Fonseca, E., Arantes, L. B., and Portella, B. C. (2020) “Brazil’s Social Policy Response to Covid-19: Health- care and Poverty Alleviation”. CRC 1342, Covid-19 Social Policy Response Series 2/2020. Bremen. Matsaganis, M., Ozdemir, E., Ward, T. and Zavakou, A. (2016) “Non-Standard Employment and Access to Social Security Benefits”. European Commission, Research Note 8/2015. Brussels. Matthews, D. (2021) “The Year that Congress Just Gave People Money”. Vox (Feb 5). Misra, K., Singh, N. and Zhang, Q. (2020) “Impact of Stay-at-home-orders and Cost-of-living on Stimulus Response: Evidence from the CARES Act”. NYU Stern School of Business. New York. Moffitt R. and Ziliak, J. (2020) “Covid-19 and the US Safety Net”. NBER, Working Paper 27911. Cambridge Moreira, A., and Hick, R. (2021) “COVID‐19, the Great Recession and Social Policy: Is this Time Different?” Social Policy and Administration 55(2): 261-279. Moreira, A., Léon, M., Coda Moscarola, F., and Roumpakis, A. (2021) “In the Eye of the Storm… Again! Social Policy Re- sponses to COVID‐19 in Southern Europe”. Social Policy and Administration 55(2): 339-357. Mukherjee, A. (2020) “Digital Cash Transfers for Stranded Migrants: Lessons from Bihar’s COVID-19 Assistance Pro- gram”. CGD, Policy Note. Washington DC. Nanau, G. L., and Labu-Nanau, M. (2021) “The Solomon Islands’ Social Policy Response to Covid-19: Between Wantok and Economic Stimulus”. CRC 1342, Covid-19 Social Policy Response Series 18/2021. Bremen. Narayanan, S., Oldiges, C. and Saha, S. (2022) “Does Workfare Work? India’s Employment Guarantee During COVID-19”. Journal of International Development 34: 82-108. Negură, P., Gașper, L., and Potoroacă, M. (2021) “Moldova’s Social Policy Response to Covid-19: Citizen Support for Government Initiatives” CRC 1342, Covid-19 Social Policy Response Series 35/2021. Bremen. Neri, M. (2021) “Desigualdade de Impactos Trabalhistas na Pandemia”. FGC Social. Rio de Janeiro. Neuberger, Z., FitzSimons, C., Rosenbaum, D. and Philbin, E. (2021) “Lessons from Early Implementation of Pandem- ic-EBT: Opportunities to Strengthen Rollout for School Year 2020-2021”. CBPP. Washington DC. Nishtar, S. (2021) “How Pakistan Tackled Its Largest-Ever Social Protection Crisis”. McKinsey and Company (April 26). New York. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 81 Nishtar, S. (2020) “COVID-19: Using Cash Payments to Protect the Poor in Pakistan”. WEF (May 5). Davos. Noyoo, N. (2021) “South Africa’s Social Policy Response to Covid-19: Relief Measures in an Unequal Society”. CRC 1342, Covid-19 Social Policy Response Series 21/2021. Bremen. Obinger, H., Petersen, K. and Starke, P. (2018) Warfare and Welfare: Military Conflict and Welfare State Development in Western Countries. Oxford University Press. OECD (2022) “Social Protection for the Forcibly Displaced in Low- and Middle-Income Countries: A Pathway for Inclu- sion”. Development Policy Papers 43. Paris. OECD (2020a) “Supporting Livelihoods during the Covid-19 Crisis: Closing the Gaps in Safety Nets”. Paris. OECD (2020b) OECD Employment Outlook 2020: Worker Security and the COVID-19 Crisis. Paris. Ohlenburg, T. (2022) “Social Protection in a Pandemic – Trends, Challenges & Technology”. GIZ and ADB. Bonn. Oktem, K. G. (2021) “Turkey’s Social Policy Response to Covid-19: Labor Market Reforms to Protect Employment”. CRC 1342, Covid-19 Social Policy Response Series 12/2021. Bremen. Olivera, J. (2021) “Peru’s Social Policy Response to Covid-19: Jeopardizing Old-Age Security”. CRC 1342, Covid-19 Social Policy Response Series 11/2021. Bremen. Ouma, M. (2021) “Kenya’s Social Policy Response to Covid-19: Tax Cuts, Cash Transfers and Public Works”. CRC 1342, Covid-19 Social Policy Response Series 27/2021. Bremen. Oxfam (2020) “Shelter from the Storm: The Global Need for Universal Social Protection in Times of COVID-19”. Oxford. Palacio, M. (2021) “Ecuador’s Social Policy Response to Covid-19: Expanding Protection Under High Informality”. CRC 1342, Covid-19 Social Policy Response Series 14/2021. Bremen. Pelly, I. (2021) “Good Practice Review on Cash Assistance in Contexts of High Inflation and Depreciation”. SPACE. Lon- don and Bonn. Pereirinha, J. and Pereira, E. (2021) “Social Resilience and Welfare Systems under COVID-19: A European Comparative Perspective”. Global Social Policy 21(3): 569-594. Peres, A., Brito, R., Bilo, C. and Balboni, M. (2021) “Editorial”. In Peres, A., Brito, R., Bilo, C. and Balboni, M. (eds) “What’s Next for Social Protection in Light of COVID-19: Challenges Ahead”. IPC-IG, Policy in Focus 19(2): 6. Perry, J. (2021) “The Politics of Economic Insecurity in the COVID-19 Era”. UNDESA, Policy Brief 91. New York. Pierson, P. (2004) Politics in Time: History, Institutions, and Social Analysis. Princeton University Press. Pople, A., Hill, R., Dercon, S. and Brunckhorst, B. (2021) “Anticipatory Cash Transfers in Climate Disaster Response”. Centre for Disaster Protection, Working Paper 6. London. Prieto, A. (2022) “A New Generation of Digital G2P Payments in Colombia”. World Bank, powerpoint presentation. Washington DC. Pruce, K. (2021) “Zambia’s Social Policy Response to Covid-19: Protecting the Poorest or Political Survival?”. CRC 1342, Covid-19 Social Policy Response Series 9/2021. Bremen. Rahim, F., Allen, R., Barroy, H., Gores, L. and Kutzin, J. (2020) “COVID-19 Funds in Response to the Pandemic”. IMF, Spe- cial Series on COVID-19 Notes. Washington DC. Ramos, C. G. (2021) “The Philippines’ Social Policy Response to Covid-19: Troubles with Targeted Social Assistance” CRC 1342, Covid-19 Social Policy Response Series 36/2021. Bremen. Ravallion, M. (2020) “On the Virus and Poor People in the World”. Economics & Poverty blog (April 2). Razavi, S. (2021) “The COVID-19 Crisis: A Turning Point or a Tragic Setback?” In Bilo, C., Brito, R., Peres, A. and Balboni, M. (eds) “What’s Next for Social Protection in Light of COVID-19: Country Responses”. IPC, Policy in Focus 19(1): 14-16. 82 VI. REFERENCES Reece, L., Owen, K., Foley, B., Rose, C., Bellew, B. and Bauman, A. (2021) “Understanding the Impact of COVID‐19 on Children’s Physical Activity Levels in NSW, Australia”. Health Promotion Journal of Australia 32(2): 365-366. Rixen, T. and Viola, L. (2015) “Putting Path Dependence in its Place: Toward a Taxonomy of Institutional Change”. Jour- nal of Theoretical Politics 27(2): 301-323. Rodrik, D. (2022) “Development Economics Goes North”. Project Syndicate (April 11). Roelen, K., Archibald, E. and Lowe, C. (2021) “Covid-19: Crisis as Opportunity for Urban Cash Transfers?” ODI, Working Paper 609. London. Roll, S., Constantino, S., Hamilton, L., Miller, S., Bellisle, D. and Despard, M. (2022) “How Would Americans Respond to Direct Cash Transfers? Results from Two Survey Experiments”. Working Paper (June 14). Romer, C. (2021) “The Fiscal Policy Response to the Pandemic”. Brookings Papers on Economic Activity. Washington DC. Romer, C. and Romer, D. (2021) “A Social Insurance Perspective on Pandemic Fiscal Policy: Implications for Unemploy- ment Insurance and Hazard Pay”. NBER, Working Paper 29419. Cambridge. Romer, C. and Romer, D. (2017) “Why Some Times Are Different: Macroeconomic Policy and the Aftermath of Financial Crises”. NBER, Working Paper 23931. Cambridge. Rossel, C., and Gutiérrez, M., (2021) “Uruguay’s Social Policy Response to Covid-19: Strong Fundamentals, Selective Expansion”. CRC 1342, Covid-19 Social Policy Response Series 15/2021. Bremen. Rutkowski, M. (2020) “How Social Protection Can Help Countries Cope with COVID-19”. World Bank, Voices blog (April 15). Sahm, C., Shapiro, M., and Slemrod, J. (2012) “Check in the Mail or More in the Paycheck: Does the Effectiveness of Fiscal Stimulus Depend on How It Is Delivered?” American Economic Journal - Economic Policy 4(3): 216–50. Sandford, J., Rajput, S., Coll-Black, S. and Kargbo, A. (2020) “Safety Nets, Health Crises and Natural Disasters: Lessons from Sierra Leone”. World Bank, Social Protection and Jobs Discussion Paper 2010. Washington DC. Sedik, T. and Xu, R. (2020) “A Vicious Cycle: How Pandemics Lead to Economic Despair and Social Unrest”. IMF, Working Paper 216. Washington DC. Seekings, J. (2020a) “Failure to Feed: State, Civil Society and Feeding Schemes in South Africa in the First Three Months of Covid-19 Lockdown, March to June 2020”. CSSR, Working Paper 455. Cape Town. Seekings, J. (2020b) “Bold Promises, Constrained Capacity, Stumbling Delivery: The Expansion of Social Protection in Response to the Covid-19 Lockdown in South Africa”. CSSR, Working Paper 456. Cape Town. Seekings, J., Gronbach, L. (2021) “Pandemic, Lockdown and the Stalled Urbanization of Welfare Regimes in Southern Africa”. Global Social Policy 21(3): 448-467. Seekings, J., Gronbach, L., and Nattrass, N. (2020) “Covid-19 Grant: We Can We Learn from Namibia”. GroundUp blog (April 29). Seemann, A., Becker, U., He, L., Hohnerlein, E. and Wilman, N. (2021) Protecting Livelihoods in the COVID-19 Crisis: A Comparative Analysis of European Labor Market and Social Policies”. Global Social Policy 21(3): 550-568. Shi, S., and Soon, S. (2020) “Taiwan’s Social Policy Response to Covid-19: Protecting Workers, Reviving the Economy”. CRC 1342, Covid-19 Social Policy Response Series 1/2020. Bremen. Shim, J. (2021) “South Korea’s Social Policy Response to Covid-19: Swift and Bipartisan Attempts”. CRC 1342, Covid-19 Social Policy Response Series 7/2021. Bremen. Shonchoy, A., Rigol, N., Roth, B., Chandra, S., Franco, A. and Hussam, R. (2021) “Safety Net and Pan- demic: The State of Social Benefit Payments during COVID-19”. IPA, Working Paper. Washington DC. Varshney, D., Kumar, A., Mishra, A., Rashid, S. and Joshi, P. (2020) “India’s COVID-19 Social Assistance Package and Its CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 83 Impact on the Agriculture Sector”. Agricultural Systems 189: 1-11. Sitko, N., Knowles, M., Viberti, F. and Bordi, D. (2022) “Assessing the Impacts of the COVID-19 Pandemic on the Liveli- hoods of Rural People – A Review of the Evidence”. FAO. Rome. Slater, R. (2022) “Sustaining Existing Social Protection Programmes During Crises: What Do We Know? How Can We Know More?” IDS, BASIC Research Working Paper 14. Brighton. Slater, R., McCord, A., Devereux, S., Baur, D., Trotter, D., and Bright, R. (2021) “Beyond Covid-19: Lessons for Social Pro- tection from WFP’s Work in East Africa in 2020”. WFP. Nairobi. Smith, G. (2021) “Overcoming Barriers to Coordinating across Social Protection and Humanitarian Assistance – Build- ing on Promising Practices”. SPACE. London and Bonn. Soon, S., Chou, C. C., and Shi, S. J. (2021) “Withstanding the Plague: Institutional Resilience of the East Asian Welfare State”. Social Policy and Administration 55(2): 374-387. Spasova, S., Ghailani, D., Sabato, S., Coster, S., Fronteddu, B. and Vanhercke, B. (2021) “Non-standard Workers and the Self-Employed in the EU: Social Protection During the Covid-19 Pandemic”. ETUI, Report 2021.02. Brussels. Stampini, M., Ibarrarán, P., Rivas, C. and Robles, M. (2021) “Adaptive, but Not by Design: Cash Transfers in Latin Amer- ica and the Caribbean Before, During and After the COVID-19 Pandemic”. IDB, Technical Note 2346. Washington DC. Sumarto, M., and Ferdiansyah, F. (2021) “Indonesia’s Social Policy Response to Covid-19: Targeted Social Protection under udget Constraints” CRC 1342, Covid-19 Social Policy Response Series 28/2021. Bremen. Sumner, A., Hoy, C. and Ortiz-Juarez, E. (2020) “Estimates of the Impact of COVID-19 on Global Poverty”. UNU-WIDER, Working Paper 2020/43. Helsinki. Tallio, V. (2021) “Uganda’s Social Policy Response to Covid-19: Rudimentary Relief Measures”. CRC 1342, Covid-19 Social Policy Response Series 26/2021. Bremen. Tarasenko, A. (2021) “Russia’s Social Policy Response to Covid-19: Temporary and Categorized Benefits”. CRC 1342, Covid-19 Social Policy Response Series 3/2021. Bremen. ten Brink, T., Li, Y., and Liu, T. (2021) “China’s Social Policy Response to Covid-19: Pragmatic Adjustment of Existing Welfare Institutions”. CRC 1342, Covid-19 Social Policy Response Series 25/2021. Bremen. The Economist (2021) “Bounding Back: How to Make a Social Safety Net for the Post-Covid World”. Print edition, March 6-12. Tossou, Y. (2021) “COVID-19 and the Impact of Cash Transfers on Health Care Use in Togo”. BMC Health Services Re- search 21(882): 1-10. UN (2022) “Experts in Social Development Commission Highlight Protection Gaps, Structural Inequities, as Delegates Detail Efforts to Build Back Stronger from COVID-19”. Commission for Social Development, Meeting Coverage (Febru- ary 9). New York. UN (2021) “COVID-19 Social Protection and Economic Policy Responses of Governments: Lessons for Social Protection Readiness and Building Forward Better – Global Policy Brief”. New York. UNHCR (2022) “Socio-Economic Evidence in Practice in Forcibly Displaced Settings: Examples of Uses of Socio-Eco- nomic Data in Advocacy, Policy and Programmes”. Geneva. UNHCR (2021) “Inclusion of Refugees in Government Social Protection Systems in Africa”. Geneva. UNHCR (2020) “Social Protection Responses to COVID-19 for Forcibly Displaced Persons”. Geneva. UNICEF (2020) “UNICEF Social Protection Response to COVID-19”. New York. United Nations Human Rights Council (2022) “Non-Take-Up of Rights in the Context of Social Protection: Report of the Special Rapporteur on Extreme Poverty and Human Rights, Olivier De Schutter”. General Assembly, A/HRC/50/38. New York. 84 VI. REFERENCES US Federal Reserve Board (2022) “Economic Well-Being of U.S. Households in 2021”. Washington DC. Van der Auwera, M., van de Meerendonk, A. and Kumar, A. (2021) “COVID-19 and Social Protection in Asia and the Pa- cific: Projected Costs for 2020–2030”. ADB, Sustainable Development Working Paper 80. Manila. Velázquez Leyer, R. (2021) “Mexico’s Social Policy Response to Covid-19: A Path of Minimal Action”. CRC 1342, Covid-19 Social Policy Response Series 5/2021. Bremen. Vera-Cossio, D., Hoffmann, B., Ibarrarán, P., Stampini, M., Pecha Garzon, C. and Gallego, J. (2020) “The Impacts of the Ingreso Solidario Program to Confront the Covid-19 Crisis in Colombia”. IPA, Study Summary. Washington DC. Verhaeghe, E. (2020) “Yes, Crises Do Happen: A Plea for Feeling More Vulnerable After COVID-19”. UNU-CRIS blog (March 23). Voorend, K., and Alvarado, D. (2021) “Costa Rica’s Social Policy Response to Covid-19: Strengthening Universalism During the Pandemic?” CRC 1342, Covid-19 Social Policy Response Series 6/2021. Bremen. Waxman, E., Gundersen, C. and Fiol, O. (2021a) “How Far Did SNAP Benefits Fall Short of Covering the Cost of a Meal in 2020?” Urban Institute. Washington DC. Waxman, E., Gupta, P., Pratt, E., Lyons, M., and Green, C. (2021b) “Documenting Pandemic EBT for the 2020–21 School Year: State Perspectives on Implementation Challenges and Lessons for the Future” Urban Institute. Washington DC. Weber, M. (2020) “Digital Public Works – A Viable Policy Option for the COVID-19 (Coronavirus) Jobs Crisis?”. World Bank, Jobs and Development Blog 8. Washington DC. WFP (2020a) “Supporting National Social Protection Responses to the Socioeconomic Impact of COVID-19”. Rome. WFP (2020b) “Social Protection and COVID-19 in Asia and the Pacific”. Bangkok. Woo, J. J. (2021) “Singapore’s Social Policy Response to Covid-19: Focusing on Jobs and Employment”. CRC 1342, Covid-19 Social Policy Response Series 16/2021. Bremen. World Bank (2022a) Poverty and Shared Prosperity Report 2022: Toward an Inclusive Recovery. Washington DC. Forth- coming. World Bank (2022b) “Using Big Data and Innovative Approaches for Social Assistance Programs”. Washington DC. Mimeo. World Bank (2021a) “Restructuring Paper on a Proposed Project Restructuring of Covid-19 Emergency Income Support Project Approved on June 9, 2020, to Republic of Maldives”. Report RES47642. Washington DC. World Bank (2021b) “Prioritizing the Poorest and Most Vulnerable in West Africa: Togo’s Novissi Platform for Social Protection Uses Machine Learning, Geospatial Analytics, and Mobile Phone Metadata for the Pandemic Response”. Washington DC. World Bank (2021c) “Digital Works for Urban Resilience: Supporting African Youth. Rapid Pilot Phase”. Washington DC. World Bank (2021d) “Stress Testing Social Protection: A Rapid Appraisal of the Adaptability of Social Protection Sys- tems and Their Readiness to Scale Up. A Guide for Practitioners”. Washington DC. Zeufack, A., Calderon, C., Kabundi, A., Kubota, M., Korman, V., Raju, D., Girma Abreha, K., and Kassa, W. (2022) “Boosting Resilience: The Future of Social Protection in Africa”. World Bank, Africa’s Pulse 25. Washington, DC. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 85 VII. ANNEX CASH TRANSFER PROGRAMS, DESCRIPTIVE STATISTICS (LARGEST COVERAGE PROGRAM ONLY) Benefit Beneficiaries Expenditure amount Country Program name Policy Adaptations Unit Coverage (actual) (USD) (monthly USD) Albania Support for Horizontal expansion 176,000 Individuals 6.20% - 394 employees in (new targeted one-off) accomodations (1st package) Algeria solidarity Vertical expansion (one- 319,983 Household 4.30% 16,717,821 76 transfer off) Angola "Kwenda" Horizontal expansion 5,808 Household 0.10% - 4 Monetary (new targeted multiple Social Transfer payments) Program Argentina Emergency Horizontal expansion 9,000,000 Individuals 19.80% 2,176,278,168 61 Family Income (new targeted multiple (IFE) payments) Australia Economic Horizontal expansion 6,800,000 Individuals 26.50% 8,595,904,286 376 Support (new targeted one-off) Payment Azerbaijan One-off cash Horizontal expansion 555,000 Individuals 5.50% - 112 payment (new targeted one-off) Bangladesh Poor Horizontal expansion 3,200,000 Household 8.70% 147,405,660 29 families cash (new targeted one-off) assistance Barbados Barbados Horizontal expansion 3,453 Household 3.50% 6,350,000 300 Adopt- (new targeted multiple A-Family payments) programme Belgium Creative Horizontal expansion 1,572 Individuals 0.00% - 1,825 sector workers (new targeted one-off) grant Belize COVID-19 Horizontal expansion 21,362 Household 22.00% 12,400,000 150 Cash Transfer (new targeted multiple Program payments) (BCCAT) Bermuda Supplemental Horizontal expansion 127 Individuals 0.20% - 1,520 Unemployment (new targeted multiple Benefit payments) Bhutan Druk Gyalpo’s Horizontal expansion 52,644 Individuals 6.80% 2,038,219 136 Relief Kidu (new targeted multiple – Income payments) support to affected individuals Bolivia Bono Contra Horizontal expansion 3,991,179 Individuals 34.20% 576,000,000 145 el Hambre (new targeted one-off) 86 VII. ANNEX Brazil Auxilio Horizontal expansion 68,300,000 Individuals 32.10% - 117 Emergencial 1 (new targeted multiple (AE1) payments) Bulgaria Family Horizontal expansion 35,700 Household 1.20% 45,406,481 560 targetеd (new targeted multiple monthly payments) allowance Burkina Faso Solidarity Horizontal expansion 69,755 Household 2.00% 9,276,315 Fund (new targeted multiple payments) Cambodia COVID-19 Horizontal expansion 700,000 Household 19.30% 320,000,000 20 Cash Transfer (new targeted multiple Programme payments) for ID Poor Households Cape Verde Emergency Horizontal expansion 21,048 Household 19.70% - 61 Social Income (Existing programs) for Inclusion Cayman Permanent Vertical expansion (one- 1,948 Individuals 3.00% 3,000,000 510 Islands Financial off) Assistance Central Service Horizontal expansion 15,800 Household 1.60% 8,000,000 15 African Delivery and (Existing programs) Republic Support to Communities Affected by Displacement Project Chile Ingreso Horizontal expansion & 12,121,438 Individuals 63.40% 6,728,869,087 135 Familiar de vertical expansion Emergencia China Dibao Horizontal expansion, 83,900,000 Individuals 6.00% - vertical expansion & admin simplification Colombia Solidarity Horizontal expansion 3,084,987 Household 21.40% 2,164,021,661 46 Income (new targeted multiple Program payments) Comoros Social Safety Horizontal expansion 21,600 Household 13.30% 6,500,000 Net Comoros (Existing programs) Congo, Rep. Emergency Horizontal expansion 220,000 Household 17.20% 9,276,315 93 cash transfer (new targeted one-off) (ECT) Cook Island School Closure Horizontal expansion & 4,742 Individuals 27.00% 675,676 135 Support (Cook vertical expansion Island) Costa Rica Protect Grant Horizontal expansion 723,520 Individuals 14.20% 488,293,670 155 (new targeted multiple payments) Côte d'Ivoire Humanitarian Horizontal expansion 125,000 Household 0.50% 24,674,997 46 Emergency (new targeted multiple Solidarity and payments) Support Fund - Cash transfer for poor and vulnerable Cyprus Covid Horizontal expansion 4,060 Individuals 0.30% - 608 unemployed (new targeted one-off) support program CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 87 Czech Tour guides Horizontal expansion 404 Individuals 0.00% 1,109,139 1,996 Republic support (new targeted one-off) Djibouti Programme Horizontal expansion 5,750 Household 5.40% 15,000,000 50 Nationale de (Existing programs) Solidarite Famille Dominican Stay at Home Horizontal expansion & 1,500,000 Household 48.10% - 86 Republic program vertical expansion Ecuador Bono de Horizontal expansion 1,430,000 Household 30.60% 57,000,000 60 Protección (new targeted multiple Familiar por payments) la Emergencia por la presencia de COVID-19 Egypt Exceptional Horizontal expansion 478,000 Household 1.90% 12,783,964 32 cash (new targeted multiple assistance payments) El Salvador Compensation Horizontal expansion 1,230,000 Household 77.20% 450,000,000 300 Bonus (new targeted one-off) Estonia parental Horizontal expansion 1,298 Individuals 0.10% - 967 allowance (new targeted multiple for disabled payments) children Eswatini WFP cash Horizontal expansion 94,000 Household 38.10% 127,467 9 transfer (Existing programs) Ethiopia UPSNP - Direct Advanced payment 220,000 Household 0.90% - Support component Fiji Poverty Vertical expansion (one- 26,000 Household 13.30% - 48 Benefit off) Scheme Finland cash Horizontal expansion 9,500 Individuals 0.20% 66,933,500 7,302 assistance (new targeted one-off) to workers in cultural and creative arts sector Gabon UNDP cash Horizontal expansion 500 Individuals 0.00% - 1,000 support (new targeted one-off) Gambia Nafa Quick Horizontal expansion 78,000 Household 26.60% - 29 program (Existing programs) Georgia One-off Horizontal expansion 371,233 Individuals 10.00% 9,128,530 91 compensation (new targeted one-off) Ghana LEAP Vertical expansion 334,084 Household 3.80% - (multiple payments) Greece Social Vertical expansion (one- 256,562 Household 6.10% - Solidarity off) Income Guatemala Bono Familia Horizontal expansion 2,692,018 Household 76.70% 769,238,166 128 (new targeted multiple payments) Guinea COVID-19 Horizontal expansion 20,000 Household 1.00% - 26 economic (new targeted multiple response, payments) April 6 2020 88 VII. ANNEX Guinea- Temporary Horizontal expansion 1,587 Household 0.70% - 74 Bissau cash transfer (new targeted multiple for vulnerable payments) families Guyana COVID-19 relief Horizontal expansion 153,000 Household 73.80% 33,573,141 120 package (new universal one-off) Haiti Transfert Horizontal expansion 319,959 Individuals 2.80% 12,808,755 39 en espèces (new targeted one-off) unique (subvention de l’État) Honduras Unique Bonus. Horizontal expansion 260,000 Individuals 2.60% 68,000,000 83 Assistance to (new targeted one-off) Independent Workers or Own Account Hong Kong The Cash Horizontal expansion 6,558,910 Individuals 87.70% 9,102,564,103 1,282 Payout (new universal one-off) Scheme India PMJDY Horizontal expansion 200,000,000 Individuals 14.50% - 7 (new targeted multiple payments) Indonesia Village Funds Horizontal expansion 11,000,000 Household 15.50% 7,323,738,941 28 Unconditional (new targeted multiple Cash Transfers payments) Iran, Islamic Corona Horizontal expansion 3,000,000 Individuals 3.60% - 36 Rep. support (new targeted multiple package payments) Iraq Emergency Horizontal expansion 11,580,000 Individuals 28.80% 230,061,350 23 Grant (new targeted multiple payments) Israel Grant to Every Horizontal expansion 8,820,000 Individuals 95.70% 2,001,539,646 231 Citizen (new universal one-off) Italy Relaunch Horizontal expansion 9,840,000 Individuals 16.50% - 1,217 Decree (new targeted one-off) Jamaica COVID Horizontal expansion 378,919 Individuals 12.80% - 70 Allocation of (new targeted one-off) Resources for Employees (CARE) Japan Universal $930 Horizontal expansion 116,520,000 Individuals 92.60% 124,171,201,114 963 cash handouts (new universal one-off) Jersey Spend local Horizontal expansion 105,000 Individuals 97.40% 14,712,629 134 card (new universal one-off) Jordan Temporary Horizontal expansion & 237,000 Household 11.00% 277,000,000 Cash transfer vertical expansion Kazakhstan Monthly Horizontal expansion 4,000,000 Individuals 21.30% - 101 payments of (new targeted multiple US$100 payments) Kenya National Horizontal expansion 1,300,000 Household 8.80% - 18 Safety Net (new targeted multiple Programme payments) (NSNP) Korea, Rep. COVID-19 relief Horizontal expansion 51,606,633 Individuals 99.70% 11,660,000,000 913 funds (new universal one-off) CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 89 Kosovo One-off Horizontal expansion 206,852 Individuals 11.70% - 353 assistance for (new targeted one-off) worker Lebanon Emergency Horizontal expansion 170,000 Household 10.70% 49,751,244 265 National (new targeted multiple Social payments) Solidarity Programme Lesotho Child Grants Horizontal expansion & 50,000 Household 7.80% - Programme vertical expansion (CGP) and Old Age Pension (OAP) Liberia Emergency Horizontal expansion 14,855 Household 1.30% 2,700,000 15 Cash Transfer (Existing programs) (SCT-COVID) Lithuania temporary Horizontal expansion 266,000 Individuals 9.50% 322,497,773 243 jobseeker’s (new targeted multiple allowance payments) Madagascar TOSIKA Horizontal expansion 215,093 Household 3.80% 15,000,000 27 FAMENO (new targeted multiple payments) Malawi Malawi’s Advanced payment 289,628 Household 6.80% - National COVID-19 Preparedness and Response Plan Malaysia Cash transfers Horizontal expansion 1,400,000 Individuals 4.30% 14,792,236 123 for e-hailing (new targeted one-off) drivers Maldives Special Horizontal expansion 22,946 Individuals 4.20% 54,484,737 1,233 allowance, (new targeted multiple Income payments) Support Allowance Malta Parental Horizontal expansion 8,360 Individuals 1.60% 37,361,063 809 Benefit (new targeted multiple payments) Mauritania Covid-19 Horizontal expansion 210,000 Household 30.90% - safety net (new targeted multiple support payments) Mauritius Self-Employed Horizontal expansion 197,000 Individuals 15.60% 60,357,467 128 Assistance (new targeted one-off) Scheme (SEAS) Mexico Pensiones Advanced payment 801,201 Individuals 0.60% 2,124,786,684 131 para personas con discapacidad Micronesia Assistance Horizontal expansion 535 Individuals 0.50% - 1,250 for stranded (new targeted one-off) citizens Mongolia Child Money Horizontal expansion & 1,186,289 Individuals 36.20% 494,440,578 35 program vertical expansion Montenegro Unemployment Horizontal expansion 17,157 Individuals 2.80% 1,043,980 benefit (new targeted one-off) 90 VII. ANNEX Morocco Tadamon Horizontal expansion 5,500,000 Household 78.00% 1,282,297,553 112 (new targeted multiple payments) Mozambique Post Horizontal expansion 623,500 Household 8.70% 79,000,000 10 Emergency (new targeted multiple Direct Cash payments) Transfers Program (PASD-PE Covid) Myanmar COVID-19 Horizontal expansion 5,600,000 Household 43.50% 301,839,228 21 Economic (new targeted multiple Relief Plan payments) (CERP) Namibia One-time Horizontal expansion 867,281 Individuals 34.10% 37,703,444 50 emergency (new targeted one-off) income grant Netherlands Tofa Horizontal expansion 7,500 Individuals 0.00% - 2,008 (new targeted one-off) Niger Strengthening Horizontal expansion 311,500 Household 7.60% 6,848,306 28 the Adaptive (new targeted one-off) and Scalable Safety Net System Nigeria National Horizontal expansion & 3,700,000 Household 8.80% - Social Safety vertical expansion Net Program (NASSP) North GMI Horizontal expansion & 4,739 Household 0.80% - 163 Macedonia vertical expansion Pakistan Ehsaas Horizontal expansion 17,000,629 Household 52.40% 1,197,901,851 75 Emergency (new targeted one-off) Cash Programme Panama Panama Horizontal expansion 442,323 Individuals 10.30% 857,582,100 100 Solidarity Plan (new targeted multiple payments) Paraguay Pytyvõ Horizontal expansion 1,200,000 Individuals 16.80% 275,720,791 36 Program (new targeted multiple payments) Peru Bono Familiar Horizontal expansion 7,749,298 Household 88.20% 1,933,883,341 187 Universal (new targeted one-off) Philippines Social Horizontal expansion 17,597,757 Household 67.90% 4,161,142,557 135 Amelioration (new targeted multiple Program payments) Poland Tourism Horizontal expansion 7,000,000 Individuals 18.40% - 204 voucher for (new targeted one-off) families with children Portugal Proteccion of Horizontal expansion 156,313 Individuals 1.50% - 533 self-employed (new targeted multiple and informal payments) workers Russia Child one-off Horizontal expansion 22,800,000 Individuals 15.80% 2,663,051,111 135 transfer (new targeted one-off) Rwanda VUP, Direct Horizontal expansion & 70,634 Household 2.30% - Support vertical expansion CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 91 Samoa National Horizontal expansion 144,275 Individuals 72.70% 4,750,693 20 ID project (new universal one-off) incentive Serbia Universal one- Horizontal expansion 6,145,529 Individuals 89.00% 86,405,063,636 122 off payment (new universal one-off) Sierra Leone COVID-19 Horizontal expansion 35,816 Household 2.60% 4,000,000 135 Emergency (new targeted one-off) Cash Transfers Singapore Care & Horizontal expansion 4,600,000 Individuals 80.90% - 450 Support (new universal one-off) Package one- off assistance Somalia Baxnano Horizontal expansion 73,478 Household 2.70% 175,000,000 20 program (new targeted multiple payments) South Africa Child Support Vertical expansion 12,337,152 Individuals 20.80% - 64 Grant (multiple payments) Spain Ingreso Horizontal expansion 284,000 Household 1.50% 5,398,491,018 572 Mínimo Vital (new targeted multiple (GMI) payments) Sri Lanka Loss of Horizontal expansion 1,338,442 Household 23.20% - 27 Livelihood (new targeted multiple Program payments) St. Lucia Public Horizontal expansion 850 Household 1.30% - 93 Assistance (Existing programs) Program (PAP) St. Vincent Social safety Horizontal expansion 1,644 Individuals 1.50% 170,370 111 and the net Interim (Existing programs) Grenadines Assistance Benefits Sudan The Sudan Horizontal expansion 2,350,000 Household 30.00% 765,000,000 5 Family (new targeted multiple Support payments) Project Sweden Housing Vertical expansion (one- 138 Household 0.00% 66,982,914 84 allowance off) Taiwan, China Triple stimulus Horizontal expansion 22,970,000 Individuals 96.20% 1,751,988,507 70 voucher (new targeted multiple payments) Tajikistan one-time Horizontal expansion 700,000 Household 44.00% 22,100,000 35 emergency (new targeted one-off) cash support Thailand Rao Chana Horizontal expansion 21,500,000 Individuals 30.80% - 465 (new targeted multiple payments) Timor-Leste Universal one- Horizontal expansion 334,000 Household 100.00% 66,856,800 100 off transfer (new universal one-off) Togo Novissi Horizontal expansion 819,972 Individuals 9.90% 20,407,893 21 (Solidarity) (new targeted multiple UCT. payments) Tonga Secondary Horizontal expansion 1,142 Household 6.20% - 88 school (new targeted one-off) students assistance 92 VII. ANNEX Trinidad and Income Horizontal expansion 25,000 Household 5.90% - 222 Tobago Support (new targeted multiple Grants payments) Tunisia Aide Horizontal expansion 370,000 Household 12.50% - 73 exceptionnelle (new targeted one-off) de 200 dinars Turkey Pandemic Horizontal expansion 6,200,000 Household 29.90% - 129 Social Support (new targeted one-off) Program Tuvalu Universal Cash Horizontal expansion 10,507 Individuals 89.10% - 30 transfer (new universal multiple payments) Uganda Child- Horizontal expansion 56,200 Individuals 0.10% 4,000,000 39 Sensitive (new targeted one-off) Social Protection Ukraine One-off Horizontal expansion 10,600,000 Individuals 24.00% 376,290,578 35 support to (new targeted one-off) vulnerable pensioners United Universal Horizontal expansion & 5,700,000 Household 19.90% - 1,338 Kingdom credit vertical expansion standard allowance. United States CARES Act Horizontal expansion 162,000,000 Individuals 49.20% 271,000,000,000 1,200 stimulus (new targeted one-off) check Uruguay Asignación Vertical expansion 372,231 Individuals 10.70% 7,854,515 Familiar Plan (multiple payments) de Equidad Vietnam Allowance Horizontal expansion 7,953,060 Individuals 8.20% 298,107,664 22 for poor (new targeted one-off) households and near poor households West Bank Emergency Horizontal expansion 33,487 Household 4.10% 13,900,000 222 and Gaza cash transfer (new targeted one-off) Zambia Cash transfers Horizontal expansion & 204,000 Household 5.70% - 10 vertical expansion Zimbabwe Informal Horizontal expansion 309,146 Individuals 2.10% - workers (new targeted multiple support payments) CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 93 VIII. ENDNOTES 1 See https://ec.europa.eu/social/main.jsp?langId=en&catId=89&newsId=10101&furtherNews=yes#navItem-1 2 See for example the COVID-19 Social Policy Response Series launched by the Collaborative Research Centre’s “Global Dynamics of Social Policy” research program (https://www.socialpolicydynamics.de/crc-1342-publications/crc-1342-covid-19-social-policy- response-series), and the UNU-WIDER collection on a range of Covid-related country level analysis (https://www.wider.unu.edu/ article/unu-wider-ongoing-work-effect-covid-19-economies-states-and-societies-global-south). 3 See https://www.econ.uzh.ch/en/people/faculty/pomeranz/conferences.html 4 Keywords used for the search included “conditional cash transfers”, “unconditional cash transfers”, “cash transfers”, “social pensions”, “non-contributory pension”, “child benefits”, “cash grants”, “guaranteed minimum income”, “basic income”, “universal basic income”, and “UBI”. 5 Keith Hansen, Country Director at the World Bank, put it elegantly: “the Covid pandemic has shown that in times of stress a comprehensive social protection system is as important as a central bank. In its own way, each ensures that shocks do not trigger systemic crises or cause irrevocable losses. Together, they buttress individuals and the economy at large against catastrophic damage” (personal communication). 6 See the compilation of 32 webinars under the Social Protection Responses to COVID-19 Webinar Series (https://social- protection.org/connect/communities/social-protection-responses-covid-19-task-force/3-webinar-series-social), a dozen events under the RECOVR initiative (https://www.poverty-action.org/program-area/social-protection/events), and another collection of 6 webinars (https://www.povertyactionlab.org/event/webinar-series-strengthening-safety-nets-and-health- care-during-covid-19-pandemic). Various other events included seminars in April 2020 (https://www.cgdev.org/event/ online-event-cash-transfers-covid-19-low-and-middle-income-countries), June 2020 (https://www.ifpri.org/event/virtual- event-social-safety-nets-and-covid-19-can-we-protect-food-security-and-nutrition and https://mbrg.bsg.ox.ac.uk/event/webi- nar-social-protection-response-covid-19-crisis-low-and-middle-income-countries); September 2020 (https://avpn.asia/event/ avpn-webinar-financing-solutions-for-social-protection-in-covid-19-response-and-recovery/, as well as https://interagencystand- ingcommittee.org/grand-bargain-official-website/webinar-linking-humanitarian-cash-and-social-protection-response); February 2021 (https://www.who.int/news-room/events/detail/2021/02/25/default-calendar/who-webinar---confronting-poverty-and-in- come-insecurity-during-covid-19-through-strengthening-social-protection); March 2021 (https://socialprotection.org/discover/ multimedia/webinar-presentation-taking-stock-one-year-mark-social-protection-during-covid); July 2021 (https://www.adb.org/ news/events/social-safety-nets-beyond-covid-19-crisis and https://sustainabledevelopment.un.org/index.php?page=view&- type=20000&nr=7905&menu=2993); November 2021 (https://www.sparkblue.org/event/undp-webinar-next-practices-innova- tions-covid-19-social-protection-responses-and-beyond and https://europeanevaluation.org/events/webinar-where-next-in- digital-social-protection-lessons-from-the-covid-19-crisis-taken-into-the-future/); and April 2022 (https://www.adb.org/news/ events/think20-associated-webinar-on-rethinking-social-protection-reform-in-the-post-pandemic-recovery). 7 According to Sedik and Xu (2020), pandemic events lead to significantly higher risk of civil disorder after 14 months, with a peak around 24 months. Five years after the pandemic, civil disorder is 7% higher than the mean of the sampled 133 countries over 2001-2018. 8 Romer (2021) discerns three main differences between Covid-19 and past crises: first, while in ordinary recessions governments try to reignite aggregate demand and restore employment in a variety of ways, options during a pandemic are constrained by public health considerations. In fact, governments can “… stimulate only as much production and employment as can happen relatively safely” (ibid, p.6, emphasis added). Second, distributional impacts (in terms of both unemployment and safety) can be even more unequal than other shocks, especially impacting, at least in the first phases, “essential workers” and vulnera- ble, low-income workers in sectors particularly affected by the pandemic. And third, “… [t]he usual knock-on effects behind a traditional Keynesian multiplier—spending in one area flows to spending throughout the economy—fail to operate when part of the economy is shut down” (ibid, p.7). In practice, the speed at which the health shock affected the economy was much higher compared to Great Recession’s: the latter “… took months and years rather than days and weeks to occur” (Moreira and Hick 2021, p.265). 9 The authors find less encouraging evidence on reforming tax administrative structures: “… [i]t appears pandemic events boost the possibility of certain tax policy reforms but not that of revenue administration” (ibid, p.2299). 10 Note that the figure refers to programs with announcement dates. Such information is available for 73% of measures, or 2,796 in total. 11 This is a conservative estimate: by relaxing particular assumptions, the global level of direct beneficiaries of cash transfers coverage may have been 20% higher. See Annex 4 in Gentilini et al (2022) for a discussion on methodology. 12 The Oxford Stringency Index is a composite indicator measuring the strictness of Covid-19 containment measures. It is a com- posite indicator with 9 dimensions measuring school closures, workplace closures, and travel bans. The index’s values range from 0 to 100, with 100 indicating maximum stringency. For full description, see Hale et al (2021) as well as https://github.com/OxCGRT/ covid-policy-tracker/blob/master/documentation/index_methodology.md. For cash transfers, the duration is calculated as the 94 VIII. ENDNOTES average among multiple cash transfer schemes available as Covid-19 responses within a country. The analysis is also conducted for values above 50 as well as by restricting countries where social assistance accounts for more than one-third or at least half of Covid-19 social protection measures. Those scenarios yield results generally consistent with those presented in the text. 13 In absolute terms, monthly transfers were on average $253, ranging from $536 in high income countries to $42 in low-income settings. 14 This is estimated by comparing COVID cash transfer for each country with the 2019 median income or expenditures. Because the latter is estimated on a daily basis, cash transfer programs were also calculated as daily transfers. Data for median income or expenditures is from PovCal. 15 Such country-level perspective should be interpreted in tandem with other supra-national adaptations. For example, Perei- rinha and Pereira (2021) noted that “… the suspension of compliance with the budgetary rules of the Stability and Growth Pact and the massive fiscal solidarity that occurred in the EU are evidence of resilience at the international level” (p.574). 16 Some programs offered amounts equivalent to the cost of the meal, while in most cases cash was a top-up to take-home rations. The state government of Bihar in India provided cash transfers for every child who was receiving a school-based transfer for uniforms. Similarly, back-to-school and other allowances were also provided to ensure continuity of children’s learning (e.g., Bolivia and 16 economies). Malaysia provided cash assistance to purchase devices for families with children belonging to the bottom 40% of the income distribution. Another 10 countries have used student relief funds to support their students studying abroad, such as Cyprus (Gentilini et al 2022a). 17 This model is reminiscent of the US system of “blue and orange food stamps” active in the late 1930s. 18 See https://www.service.nsw.gov.au/campaign/covid-19-help-businesses/grants-loans-and-financial-assistance 19 Malaysia subsidized childcare for working parents with e-vouchers worth about $471 (Hammad et al 2021). 20 An internal review of 2,121 pre-Covid programs included in the ASPIRE database identified 1465 cash-based measures. Out of the latter, 1,180 were unconditional cash transfers, which therefore represented 80.5% of the cash-based portfolio (conditional cash transfers amounted to 96 programs and public works accounted for 189 interventions). Total unconditional transfers, includ- ing both cash and inkind measures, were 78% of the portfolio. Note that the ASPIRE database has limited information for HICs, hence data should be interpreted cautiously. 21 This doesn’t consider the substantial share of fee waivers as they are not “transfers”, but discounts or payment postpone- ments. Those instruments represented between 25-38% of social assistance across country income levels and have been used extensively in Africa. For example, see Berkouwer et al (2022) for Ghana’s experience. 22 Note that universal transfers occurred in 13 economies, but involved 16 programs due to three interventions by Taiwan (April 2020, June 2020, and September 2020) and two by Singapore (February and May 2020). 23 This includes 28 cases where programs identified beneficiaries that were waitlisted in or had graduated from existing program databases. For example, this allowed Iraq to expand coverage of its Social Safety Net Program by 60,000 beneficiaries. Note that data on identification and registration presented by Hammad et al (2021) doesn’t include most high-income countries. 24 A related issue in India’s initial response was communication and beneficiaries’ awareness of transfers. A survey by Gelb et al (2022c) found that in April 2020, all respondents were enrolled in at least one transfer program that utilized DBT, yet only 76% reported not knowing that a payment had been made in their account. 25 Nonparametric correlation of 0.370 significant at the 1% level. Coverage refers to ‘direct’ and ‘actual’ beneficiaries (see Gentilini et al 2022a, Annex 4 for definitions). 26 The extremes include Colombia program “Cash transfers for positive COVID test individuals”, which took only two days, and Singapore’s “One-off GST Voucher”, which took nearly four months. 27 About 83% of funding to local and national actors is directed to national governments ($624 million). 28 The OECD (2022) shows that while COVID-19 provided a window of opportunity to include refugees in national social protec- tion systems, de jure access to programs was not necessarily ensured de facto. This is also reflected in the case of Venezuelan migrants: preexisting cash transfer programs were generally not adapted to include those mobile populations (IPC-IG et al 2021). Yet there are cases where assistance during Covid-19 was provided to displaced populations and host communities, such as Cam- eroon (for 10,000 people) and Djibouti (1,000 urban refugee and asylum seekers, and 200 host families) (UNHCR 2021) 29 Cho and Johnson (2022) also reported that food transfers reached most households within two weeks of the first lockdown; instead, it took between “… few weeks to several months for SAP beneficiaries to receive their cash” (p.22). 30 Such proportion is also reflected in the broader size of the fiscal stimulus packages. Romer (2021) note that the average fiscal package early in the pandemic was 5.2% (with a median of 4.4%), while he average package early in the Great Recession was 1.4% GDP (median of 1.6%). 31 The authors found that since 1980, OECD countries with initial debt ratios one standard deviation below the sample mean decreased their high-employment surpluses by over 3% of GDP in response to significant financial distress. Conversely, countries with initial debt ratios one standard deviation above the sample mean actually increased their high-employment surpluses by 2 to 3% of GDP, i.e., they adopted highly contractionary fiscal policy. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 95 32 This section draws heavily from discussion in Gentilini et al (2022), Annex 1. 33 For a thought-provoking discussion on Africa, see Arezki (2021). 34 See https://www.youtube.com/watch?v=mBn59ij3ICY. 35 There are examples of “lighter” integration. For example, UNHCR (2022) shows that high-frequency phone surveys conducted by national statistical offices and development partners included forcibly displaced persons in 12 countries (i.e., Bangladesh, Burkina Faso, Chad, Costa Rica, Djibouti, Ethiopia, Iraq, Jordan, Kenya, Mexico, Uganda, and Yemen). These provided the “first systematically collected data on refugees and hosts during a global shock” (ibid p.11). 36 See https://www.workandincome.govt.nz/covid-19/leave-support-scheme/index.html 37 Some studies examined sectoral effects. For example, see Varshney et al (2020) for an analysis of the impacts of cash transfers among Indian farmers’ purchases of agricultural inputs. In Kenya, a one-off cash transfer for female microentrepreneurs of $50 (equal to approximately 1 month of average profits) provided at the very early stages of the pandemic doubled weekly business profits, recouping about one-third of initial losses that occurred between January and May 2020 (Brooks et al 2022). Also, benefi- ciaries increased spending on personal protective equipment by 17% and increased an index of mitigation practices such as hand washing and mask wearing by 0.23 standard deviations. 38 A forthcoming report by the World Bank (2022a) presents fresh analysis on the performance of cash transfers on parameters like, among others, poverty reduction, adequacy, targeting, and incidence 39 See Devereux et al (2021) for a conceptual and methodological discussion on the impacts of the pandemic on food security. 40 Of course, such broad-brush categorization withholds considerable variation on the extent to which impacts were mitigated. For example, within such yellow band, in some cases mitigation was relatively substantial (e.g., Colombia) while in others more limited (e.g., Uruguay). Also, a country may appear in different color bands pending on studies’ timing, scope and methodologies (e.g., South Africa). In a narrower number of cases, social assistance mitigated more than half – or even fully countered – the effects of the pandemic. In the case of poverty and food security, this involved instances of generous, large-scale provisions in Brazil and South Africa. With regards to health, particularly striking are cross-country results by Aminjonov et al (2021) and esti- mates on lives saved by Ashaw (2021). In a range of “red” cases, welfare effects of responses were modest. This mostly involves a compilation of tax-benefit microsimulations from Sub-Saharan Africa and as well as the notable case of Mexico. 41 There are at least four main channels through which cash transfers generate multipliers, namely through effects on firms, wages and employment; integration of economies and retail-level spending; indirect effects; and through marginal propensity to consume. 42 This is not to say that all one-off universal transfers are HM. For example, Hong Kong government’s 2020 universal transfers were financed, along with all the measures in the emergency package, by raising money on the bond markets. 43 Unemployment insurance appears to have bolstered a higher MPC: Farrell et al (2020) estimate that the spending of those benefit increased $0.73 for every $1 of additional benefit received by beneficiaries. 44 Put in (recent) historical perspective, results are consistent with 2001 tax rebates (Shapiro and Slemrod 2003; Johnson 2006; Agarwal et al 2003); and they are slightly higher than 30% MPC recorded for the 2008-9 stimulus (Parker et al 2013; Sahm et al 2010). They are also generally line with n line with 40% MPC found for Australia in 2008-09 (Leigh 2012). 45 A multi-arm experiment by Jaroszewicz et al (2022) found that “… cash allowed participants to spend more money, improving objective financial outcomes for the few weeks immediately following the transfer and then dissipating thereafter” (p.4). 46 Another key activation measure, trainings, included 140 programs in 77 countries (Gentilini et al 2022a). For example, Indonesia adapted the design of the Kartu Prakerja (Pre-Employment Card) program and doubled the initial budget. The program targets both informal and formal workers with vocational training, cash and credit. In Singapore, the SGUnited Skills Program aimed to support 30,000 jobseekers with subsidized training courses for industry-relevant skills. The Seychelles Employee Transition Scheme intended to support workers made redundant by encouraging re-entry through retraining, up-skilling, and providing temporary work placement services. And in Belgium, a platform helped facilitate job matching between employers and students working part-time (especially hospitality and non-food sectors). 47 For example, Portugal provided hiring subsidies of around €5000/employee, specifically targeted at young workers and the long-term unemployed. Croatia announced a scheme of $645/employee to support employers organized as sheltered workshops, integrative workshops, and employment units for people with disabilities that experienced a decrease in turnover of at least 50%. 48 See https://www.oecd.org/els/soc/recipients-socr-hf.htm 49 In addition, about 36% of low- and middle-income countries established measures that deliberately encouraged the unem- ployed and other target groups to become self-employed (De La Flor et al 2021). 50 Normally, benefits for a claimant that has been self-employed for a certain amount of time are calculated “… as if they earned the National Minimum Wage for the hours they are expected to work, even if their actual earnings are lower. This was therefore a considerable improvement for the self-employed” (Seemann et al 2021, p.561). 51 TOZO differs from “regular” social assistance as it does not take property and capital into consideration for eligibility, although it includes other forms of income in the household. 96 VIII. ENDNOTES 52 It’s interesting to note hoe the Irish experience with PUP contrasts with the UK’s extension of an existing scheme, the Universal Credit program: while PUP was extended three times (initially set for 3 months, then extended for an additional 2, then for other 7 months), the UC increased benefits (by £20/week, bringing it to £94) and eased access for a full year from the start. 53 Vera-Cossio et al also found that Ingreso transfers increased household expenditures on education, and increased the prob- ability that school-aged children spent at least four hours on virtual schooling activities by 11 percentage points. The program also increased the time students spent on education by 27-46 minutes. Yet it didn’t reduce the probability of domestic violence (a concern during the pandemic with lockdowns). 54 Beyond cash-based measures, see Seekings (2020a) for a discussion on the implications of suspending the national school meal program. 55 Namibia provided a one-off grant (R750) for people aged 18-59 who have lost informal livelihoods or were already unemployed with an application process by phone. People with work, including all people with formal employment, were excluded, as was anyone who already received one or other social grant. Anyone who had been formally employed but had lost their job would be supported through a separate scheme via Namibia’s Social Security Commission. About 739,000 people were estimated to be eligible to the grant: three days after the application process was opened, 350,000 applications had been received. Such wave of interest overwhelmed the SMS applications system which had to be redesigned quickly, leading to minor delay. Applications rose to 579,000 after a week. At that point, about 20% of the target population was reached (Seekings et al 2020). 56 The WhatsApp platform was subsequently hosted by a private company (GovChat) instead of the Department of Health used in the test. 57 They also found that the main profiles of beneficiaries were young, relatively low-education, black males. 58 The Kafaalat program replaced BISP just before the pandemic in February 2020. According to Bourgault and O’Donnell (2020), Kafaalat “… uses an updated door-to-door digital survey to identify recipients, and the payments are dispersed monthly (PRK 2,000/month, or about USD $12) rather than quarterly as under BISP. And though the BISP program did not automatically set up bank accounts for intended payment recipients, the Kafaalat program opens bank accounts for women in a conscious effort to increase their access to financial services” (p.2) 59 For example, Cantillon et al (2021) show that in Germany, “… all families were entitled to an additional €300 for every child. The top 20% of income earners, that is, those with an annual income of more than 67,816 Euros, only benefitted partially or not at all, due to the relevant offsetting with child tax allowances” (p.334). 60 The informal networks and governance of such informal settlements is an important dimension to consider. Based on phone surveys of 321 leaders across 79 settlements in Bhopal and Jaipur, India, Auerbach and Thachil (2020) show that in the early phases of the pandemic, slum leaders help with day-to-day problem-solving; handle request for government relief (especially food assistance); make claims; and collect administrative information. However, they also rely on partisan networks, which can affect and politically skew some of their activities. 61 For example, techniques for conceptualizing and diagnosing urban poverty may underestimate the problem; there could be skeptical urban attitudes and adversarial narratives around social assistance (e.g., transfers not needed because “there are jobs opportunities in cities”); related fears of “dependency” and inducing migration could be amplified among local government au- thorities; and, critically, adapting design of social assistance to dynamic urban livelihoods has received relatively little attention, alongside solving specific data, delivery and institutional issues. 62 See for example World Bank (2022b). 63 In the case of Mauritania, a special fund for social solidarity involved a state contribution of $170 million; similarly, in Uz- bekistan $44.8 million was raised through public-private partnerships, including donations from individuals and businesses to support the “Kindness and Support” program of the government (Almenfi et al 2020). 64 Rutkowski (personal communication, 2022) noted that social insurance was conceived as an “automatic stabilizer”, albeit for idiosyncratic risks and not for systemic shocks. If risks could be properly predicted and estimated, provisions could be made through risk mitigation with no need for last resort risk coping through social assistance. 65 Stress test appraisals may provide a measurable understanding of how much existing systems could be scaled up; they may help identify gaps and areas for strengthening; they may enhance accountability by linking stress test results with investments in crisis preparedness; they may foster coherence and reduce fragmentation across sectors involved in crisis responses (e.g., social protection and disaster management authorities); and finally help ensure that national structures are the default option for so- cial protection, thereby using “parallel systems” only in exceptional circumstances. This may include cases of pure humanitarian action and contexts where scale up efforts may “overheat” domestic systems (World Bank 2021). 66 See https://www.africa-newsroom.com/press/sapm-dr-sania-nishtar-briefs-diplomatic-corps-on-declaration-on-establish- ing-a-global-social-protection-knowledge-platform?lang=en 67 The framework was inspired by Acemoglu (2020) and his broader reflections on the post-pandemic state. Clearly, the current “triple blow” of pandemic-displacement-inflation is redefining the scenarios. As such, future pathways may be interpreted in light of all the concomitant shocks, and not necessarily Covid alone. CASH TRANSFERS IN PANDEMIC TIMES: EVIDENCE, PRACTICES, AND IMPLICATIONS FROM THE LARGEST SCALE UP IN HISTORY 97 “Over the course of 2020-2021, social protection was rapidly elevated as a fundamental component of countries’ response framework. This has energized an intense debate on the future of social protection, including whether a radical disruption like the pandemic could alter the long-term, structural equilibrium of social protection coverage, adequacy, and comprehensiveness.” 98