We thank Elizabeth Pérez Chiques and Sergio Campos for their helpful comments on an earlier version of this manuscript presented at the Public Administration Division seminar at CIDE. We would also like to thank Pablo Ibarrarán and Marco Stampini of the InterAmerican Development Bank for ideas and questions that sparked our research, as well as two anonymous reviewers for their insightful and constructive feedback. Cejudo thanks Columbia University's Institute for Latin American Studies, where he was a visiting scholar while writing this article. He is also grateful to the New York Public Library's Center for Research in the Humanities. 

Introduction

1The COVID-19 pandemic made heavy demands on governments’ capacities. The unprecedented public-health crisis posed formidable challenges for public administrations around the world (Boin et al., 2020). Governments had to reorganize processes swiftly and made use of old and new capacities to cope with a set of complex problems (Peña et al., 2020). This simultaneous deployment of policy responses allows for comparative analyses that shed light on the reasons for commonalities and dissimilarities in countries’ reactions to emerging problems (Liu & Geva-May, 2021), as well as on the capacities required for doing so (Dunlop et al., 2020), particularly in developing countries where those capacities cannot be taken for granted.

2In Latin America and the Caribbean (LAC), in addition to the public health emergency, governments faced an economic crisis and an increase in poverty. In response, they made extensive use of cash transfer programs (CTPs) aimed at compensating for the temporary loss of income, albeit with wide variations. These differences can be explained by political and ideological preferences, fiscal conditions, policy legacies, and even broader institutional factors, such as levels of democracy or decentralization (Baccini & Brodeur, 2020; Béland & Schlager, 2019; Rozell & Wilcox, 2020). Notwithstanding the type of intervention chosen, governments were also constrained by their existing policy capacities.

3In this paper, we study how policy capacities enabled LAC countries to respond to the social and economic consequences of the pandemic, with cash transfer programs to address reductions in employment and income. We build on Capano et al. (2020)’s understanding of policy capacities in the face of the pandemic as a mix of government experience and preparedness to respond to an emergency. We argue that policy capacities are not only context-specific but also task-specific. They are not generic assets valuable for addressing all kinds of policy issues at all moments of the policy process. In the case of the social protection response, the challenge was specific: governments needed to determine the attributes of the households or persons that would benefit from a cash transfer, identify who they were, and be able to deploy the resources to reach them. By analyzing 122 CTPs, implemented in 27 countries from the beginning of the pandemic until March 2021, we explain how different combinations of policy capacities enabled governments to reach their populations (old and new beneficiaries) in response to this emergency by making use of their experience implementing CTPs and the information gathered over the years.

4Of course, we do not state that capacities were the sole reason behind these responses. Neither do we argue that greater capacities resulted in well-designed interventions or even in effective responses to the pandemic: we do not seek to assess the causal link between policy capacities and the design of these interventions or their effectiveness. Our aim is simpler: we show how combinations of policy capacities enabled different types of responses.

5The paper is structured as follows. In the first section, we explore the relevant literature on policy capacities. We then explain our analytical framework and methodological decisions. After operationalizing both dimensions of capacity (experience and preparedness), we group countries by their configuration of capacities and types of response. We then focus on specific CTPs to exemplify the different pathways linking capacities and policy responses. We conclude with a discussion of how these findings interact with the literature on policy capacities, and with ideas for further comparative research.

Policy Capacities in Response to the Pandemic

6Latin America has experienced a significant expansion in social policy in the last two decades (Blofield, 2019). There are three factors commonly identified in the literature as determinants for this expansion. The consolidation of democratic regimes allowed for a diverse array of actors to mobilize in favor of a more extended social policy. The favorable economic conditions resulting from the commodity boom enabled governments to address such demands. And policy legacies shaped the nature of the policy expansion by defining needs and alternative solutions (Arza et al., 2022).

7The COVID-19 pandemic, in contrast to this steady expansionary wave, required simultaneous, effective responses from numerous policy sectors. Yet countries were not all equally equipped for doing so: not all governments had access to the same policy options in the same way when they were required. As a result, social policy responses varied across the region. Yet, these variations are not necessarily explained by the same factors that shaped the gradual expansion developed over two decades. Indeed, Blofield (2021) argues that three factors account for variation in the social policy responses to the pandemic: party legacies, divided government, and fiscal space.

8Of course, political preferences and fiscal resources may explain the intention and the fiscal and political leeway for a response. But any CTP put in place would need to be implemented, and thus require a bureaucracy equipped to identify eligible populations and to deliver the transfer. As Capano et al. (2020) argue from a public policy perspective, the managerial and organizational resources that governments had at their disposal when the pandemic hit influenced their reaction to the crisis. There were countries characterized by their preparedness to face pandemics with some having had similar experiences that resulted in realistic, prudent, and safe public health responses (see Yen et al., 2022). On the other hand, other countries were ill-prepared and, having no relevant experience, lacked the skills to make decisions to mitigate, or effectively handle, the problems associated with the pandemic. The authors identify governments’ preparedness for a pandemic and their experience in coping with similar situations as two dimensions of policy capacity.

9Policy capacity needs to be distinguished from related concepts such as state capacity. These concepts come from different fields, each focusing on different dimensions of such ability to implement a policy, but with overlapping implications. In the literature on comparative politics, the notion of state capacity is associated with the structures that set out the power systems through which a state controls its territory, extracts revenue, allocates resources, and performs essential state functions (Besley & Persson, 2009; Soifer, 2015; DeRouen et al., 2010; Duarte, 2014).

10From the literature on public policy and administration, on the other hand, policy capacity is regarded as the ability of bureaucratic structures to perform the necessary functions throughout the policy process to address public issues. Policy capacity entails a combination of skills (analytical, operational, and political), at different levels (individual, organizational, and systemic) along the policy cycle (Wu et al., 2015). Analytical capacity is the ability of governmental actors to acquire and apply knowledge about the task they have to perform, through a learning process within policy arenas (Dunlop, 2015, p. 2015; Howlett, 2015; Pattyn & Brans, 2015). Operational capacity is the ability of governmental actors to deploy and develop the administrative resources that are key to the task they are to perform, by learning from the experience of governmental and non-governmental actors (Dunlop & Radaelli, 2018). Political capacity is the ability to understand the relevant interests and ideologies through the development of learning relationships between both governmental and non-governmental actors, and the public (Dunlop, 2015; Pal & Clark, 2015).

11Both state and policy capacity enable effective policy implementation, but for different reasons. A government with high state capacity may be able to finance the necessary infrastructure to provide health services efficiently throughout its territory without political constraints, and yet this would be insufficient to address a public health-related issue if they lacked the policy capacity to design a triage system that allows for effective and timely diagnoses, for example, or to translate into practice scientific and technological breakthroughs. Moreover, different levels of policy capacity may exist in different public agencies within the same state, even if these agencies are part of the same policy area (i.e., primary care clinics can be more efficient and effective than medical centers that provide secondary-level services). Finally, policy capacity may vary quickly over time: a country may have a poor public health response in the initial stages of the pandemic, but an effective and properly implemented vaccination strategy months later.

12Thus, policy capacities are dynamic and context-specific, not generic assets valuable for addressing all kinds of policy issues (Dunlop et al., 2020). In the case of the social protection response to the pandemic, the challenge was specific: governments not only needed budgets and political support for deploying CTPs; they needed to determine the attributes of the households or persons that would benefit from a cash transfer, to identify who they were, and to be able to deploy the (bureaucratic, financial, and logistic) resources to reach them. Thus, having socioeconomic information about the population, and governmental agencies with skills to use it, was crucial for conducting these tasks. Moreover, the pertinence of specific capacities depends on the type of response: Expanding an existing program to include additional beneficiaries or increasing the amount given to current recipients is not equivalent to creating a brand-new program.

13To incorporate the dynamic and context-specific nature of policy capacities into the analysis of LAC’s response, we built on Capano et al.'s understanding of policy capacities as “the managerial and organizational resources'' (preparedness) and the “levels of existing skills and competences” (experience) that each government has to face public issues (Capano et al., 2020, p. 298). We rely on this notion of policy capacities since we aim to observe these capacities not in a vacuum, but in a specific crisis. As Toshkov et al. (2022) put it, “the way governments have reacted to the outbreak is a function of the capacities and characteristics of the public administrative system already in place”.

14Beyond the public health response, countries had to cope with the dramatic loss of income due to the suspension of economic activity (Gentilini et al., 2021). LAC countries implemented CTPs and this choice demanded specific policy capacities. By definition, a CTP entails delivering a benefit to individuals. This cannot occur unless there is information about who these individuals are and where they can be reached. Policy capacity for deploying CTP responses means a specific form of preparedness: having the informational resources to reach the target population. It also means a specific form of experience: knowing how to implement CTPs is the product of a bureaucracy familiar with this type of program.

15When the pandemic hit, many countries found that some basic information for policymaking was non-existent, or constantly changing, rendering governments unprepared to act (Boin et al., 2020). For a social policy response, having information is crucial to identify citizens, their location and income, and their potential eligibility. Without this ready-to-use information, governments would not have been able to find the people that they intended to help, or would have had to retort to burdensome practices to obtain this new information (Barca & Hebbar, 2021; Chudnovsky & Peeters, 2022; Lee & Zhang, 2017; Muralidharan et al., 2016). We consider that the informational tools that governments had at their disposal to respond, and the mechanisms for information gathering and use, correspond to this notion of preparedness.

16Still, a government may have the tools and not know how to use them. Capano et al.’s (2020) understanding of policy capacity includes the know-how that comes from the experience of implementing a certain type of program. Indeed, governments and their agencies also require key skills that enable them to make strategic use of available information to attain their desired results. In ordinary circumstances, policymaking requires organizations and officials with the ability to assess existing alternatives, or to support positions already adopted (Landry et al., n.d.) by knowing how to decipher an issue, access information, and apply knowledge (Dunlop & Radaelli, 2018; Howlett, 2015). In times of crisis, having experience is even more important because people and organizations seek to interpret events and scenarios when emergencies arise. To produce policy, public officials resort to the know-how of their “governance partners” to learn about the functioning of a service (Dunlop & Radaelli, 2018; Pierre & Peters, 2006). Without experience, public officials’ skills are more limited and the range of policy solutions they may envision narrower, since people and organizations act based on their (shared) accumulated knowledge and established routines: they rely on choices that are familiar to them because they were actions implemented previously, or processes currently at work (McConnell & Drennan, 2006; Weick, 2001). Thus, we adopt Capano et al.'s (2020) definition of policy capacity as a mix of preparedness and experience to examine CTP responses in LAC, exploring whether there was information on potential beneficiaries in place and if governments had experience of implementing CTPs.

Methods

17To analyze how LAC governments deployed their capacities to implement CTPs in response to the pandemic, we carry out a mixed-methods research design, combining quantitative and qualitative data (Ivankova et al., 2006) while answering complementary but distinct questions (Hendren et al., 2022). To identify LAC countries’ policy response and analyze how policy capacities were used, we combine quantitative information about the number, type, and coverage of the CTPs implemented in each country, with qualitative information about the process by which governments identified beneficiaries and delivered transfers. We integrate this information using a sequential explanatory design (Ivankova et al., 2006) with two phases: a quantitative phase (data about programs), followed by a qualitative phase (short case studies about the use of capacities in the implementation of types of response), so that findings from specific CTPs will help in contextualizing the quantitative data and in providing an analysis of how patterns identified in the first phase came about. Our analysis of these programs, within their national contexts, allows us to “answer ‘how’ questions that go beyond numerical representations of relationships and trends” (Hendren et al., 2022).

19With this information, we make a comparative analysis of how policy capacities enabled LAC governments to launch new, or expand existing, CTPs. To do so, we develop a working definition for observing policy capacities in practice, based on the analytical framework developed by Capano et al. (2020). We then identify patterns of relationship among these dimensions of policy capacity (experience and preparedness), on the one hand, and types of innovation (new program, vertical or horizontal expansion) and their coverage on the other.

20Then, to understand how those capacities were deployed, in the second phase we develop brief analyses of specific CTPs, based on official records, press reports, and, when available, scholarly literature. We selected CTPs to represent the various types of response associated with the availability of policy capacities. By doing this, we demonstrate how each intervention was enabled by the capacities that governments had before the pandemic and, in addition, how some governments compensated for their lack of these capacities. We do not claim a causal relationship (since the determinants of these expansions are beyond the reach of this article), but we show that the existing capacities enabled specific types of response but not others.

Operationalization: How do we look at capacities and governments’ responses?

21We use experience in implementing CTPs and preparedness for adapting them for new purposes as dimensions of policy capacity (Capano et al., 2020). In this section, we develop working definitions of experience, preparedness, and policy response, and explain how we observe these.

Experience

22The relevant experience for deploying CTPs is the knowledge that governments have about how to implement them, which is a product of having a bureaucracy familiarized with this type of program. To operationalize our working definition of experience, we use information about whether a government had a CTP in place before the pandemic began. We take into account the longest-running national CTP with at least 1.5% coverage in the last available year, according to the Non-contributory Social Protection Programs Database by ECLAC (2020b). We exclude social pensions and productive-inclusion programs, because our focus is on interventions to compensate for the loss of income regardless of people’s job status. We use only ongoing CTPs since, although some countries have been implementing them for many years, their termination or substitution usually means losing or diminishing policy capacities (when administrative processes are dismantled and personnel are dismissed).

Preparedness

Policy responses (coverage during the pandemic)

25This indicator assumes that the more people that benefited, the better. However, it does not account for the amount of money being transferred. The latter is clearly relevant, but it would be more indicative of governments’ fiscal capacity than of their policy capacity for reaching people.

Policy responses and capacities

26By early March 2021, 122 CTPs were created or modified in 27 LAC countries to reach, on average, a number of beneficiaries equivalent to 23.8% of their populations. Yet the reach of these programs varied widely. Chile, Bolivia, Perú, and the Dominican Republic reached 40% or more of their population. On the other hand, 10 countries in the region responded with programs reaching less than 10%. In this section, we describe these variations in coverage, and the capacities that LAC governments had for implementing CTPs, and identify patterns resulting from the interaction of capacities and types of response.

27Figure 1 shows the number of people reached by all the CTPs implemented in each country, divided by its population. The bars are made up of the coverage of each program. Bolivia is the country with the largest coverage, with direct beneficiaries equivalent to 106.8% since people could have received more than one program. Indeed, Bono contra el Hambre in Bolivia (one of the largest CTPs in the region) was explicitly designed to reach beneficiaries of a previous CTP in response to the pandemic (Bono Universal), as well as additional beneficiaries.

28These responses not only varied in their coverage; they also took different forms. Some countries increased the amount of money transferred to existing beneficiaries (vertical expansions), others increased the number of beneficiaries of existing programs (horizontal expansions), and others launched new interventions. As Figure 1 shows, the most common social protection response to the pandemic was the creation of new CTPs. A total of 27 countries launched 95 new programs — on average, 3.5 CTPs per country, ranging from eight in Argentina and Chile to one in Mexico. Nine governments opted to expand 18 CTPs vertically, whereas seven countries carried out horizontal expansions in nine of their CTPs.

29As is evident in Figure 1, CTPs were not an unusual response. On the contrary, LAC countries had a long tradition of implementing them (Rawlings, 2005). This experience, as argued before, allowed bureaucracies to develop skills that enabled them to implement emergency responses. Figure 2 shows how most governments already had experience of implementing CTPs before the pandemic began, some with programs launched more than 15 years earlier. The average number of years that a country has been implementing an ongoing CTP is 10.4 (14.1 if only those 20 countries with a CTP are considered). Seven countries without an ongoing CTP in place get a score of 0.

The interaction between capacities and responses

31Policy capacities enable policy responses, they do not determine them. This means that policy capacities set the higher limit of what is possible in the intended response in terms of coverage, but also regarding the type of response (type of innovation). In this subsection, we show the interaction between coverage of the responses, the dimensions of policy capacity (experience and preparedness), and the type of innovation, to identify patterns across countries.

32Based on variations in coverage, experience, and preparedness, we distinguish between three groups of countries (see Figure 4 and Table 1). The first group is composed of governments with experience of implementing an ongoing CTP, whose responses had coverage at least as large as their source of information. These countries appear in Figure 4 with a dotted line and with their names in bold. This group includes 13 countries that already knew how to implement a CTP and had the information to reach people. This could have allowed them to target specific populations, to access banking information or payment mechanisms easily, and so to transfer the benefit expeditiously. Chile’s response reached a number of beneficiaries equivalent to the coverage of its social registry (around 75% of its population). Meanwhile, Argentina, Colombia, Costa Rica, the Dominican Republic, Mexico, and Uruguay reached a number far below what its largest source of information would have allowed.

33The second group includes countries with experience, whose responses reached more people than those included in their largest source of previous information. In this group, seven countries appear in Figure 4 with their names in bold (because they had experience) and a continuous line (because they reached more than the number of registrants in their information source). Countries such as El Salvador, Panama, and Paraguay reached significantly more people than those included in their social registries. Meanwhile, there are three countries (Belize, Guatemala, and Jamaica) that did not have a social registry at all but had experience of implementing an ongoing CTP, which could have enabled them to obtain information about potential beneficiaries. Although Bolivia has a social registry, its coverage was lower than its ongoing CTP. Bolivia’s response was the most outstanding case: it launched four programs with an aggregate number equivalent to 106.8% of its population, almost two times greater than the coverage of its ongoing CTP (54.1%).

34Finally, in the third group, there are seven countries without an ongoing CTP (no experience) and with no available information, and yet they managed to design and implement a new intervention, even if the number of beneficiaries was small. These countries reached a number of direct beneficiaries smaller than 10% of the population (except for Guyana which reached 19.8%). Despite their lack of experience, these governments managed to respond either by using alternative administrative records (whether from programs’ registries or from non-social registries, such as education and labor data) or by relying on individuals providing their own data when making applications.

Table 1. Configuration of experience and preparedness in CTP responses.

	Countries with experience	Countries without experience
CTPs reached fewer people than the number of individuals in their registries	13 countries: Argentina, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, Haiti, Honduras, Mexico, Peru, Trinidad and Tobago, and Uruguay	Not observed
CTPs reached more people than the number of individuals in their registries	7 countries: Belize, Bolivia, El Salvador, Guatemala, Jamaica, Panama, and Paraguay	7 countries: Antigua and Barbuda, Bahamas, Barbados, Dominica, Grenada, Guyana, and Saint Vincent

Source: Authors’ elaboration.

35Not only was coverage of the response different between countries, but also the types of response varied (new programs, vertical or horizontal expansions). Even within the same country, governments implemented different types of innovation. Different combinations of capacities enabled governments to design new interventions, as well as to carry out activities to enroll beneficiaries (such as processing applications and determining eligibility) and to deliver the benefits (locating the target population and choosing the payment mechanism) (see Figures 5 and 6). By definition, all vertical and horizontal expansions required an ongoing CTP and thus occurred in countries with experience of implementing this kind of program.

36All 18 CTPs that were vertically expanded used existing information, since they were targeting people already enrolled. Most horizontal expansions also relied on the available information (seven out of nine to identify and enroll beneficiaries, and six to deliver benefits), but others did not: for instance, Guatemala gathered new information to identify potential beneficiaries, whereas Belize used it to send the money through a new application.

37All 27 countries implemented at least one new program. To do so, 74.7% of these new interventions used previous information to select beneficiaries, and 44.2% for delivering the cash transfer. All but seven countries had experience of implementing a CTP.

38To sum up, despite important variations in their responses, governments relied on existing capacities. Even though all countries launched at least one new program, most used information from existing sources to identify beneficiaries and deliver the transfers. The extent to which information was used depended on the information requirements of these responses: vertical expansions did not require new information, whereas horizontal expansions and new programs did. On the other hand, the most ambitious responses took place in countries with experience. Indeed, even if countries without experience of implementing CTPs managed to launch new programs, these reached less than 10% of their population.

How policy capacities enabled countries’ responses

Table 2. Programs analyzed according to type of innovation and policy capacities.

	Both past experience and preparedness	Only past experience	Only preparedness	Neither past experience nor preparedness
Horizontal expansion	Bolsa Familia in Brazil	Programa de Aporte al Adulto Mayor in Guatemala	Not possible
Vertical expansion	Boost in Belize	Not possible
New program	Quédate en Casa in the Dominican Republic	Bono Familia in Guatemala	Not observed	Covid-19 Relief Fund in Guyana

Source: Authors’ elaboration.

40These six programs are useful examples of how varying configurations of information availability and experience of implementing CTPs depict diverse pathways that governments had to go through to extend existing CTPs or create new ones. These programs exemplify how policy capacities, or the lack thereof, completely modify the process of launching and operating a program. As will be shown in the following sections, when policy capacities are in place, deploying any type of response entails a seamless decision-making process, whereas when policy capacities are (totally or partially) absent, many stumbling blocks have to be dealt with before being it is possible to deploy a policy response. In what follows, we tell the stories of these six programs as a way of capturing these variations in policy capacities, and grouping them according to the different types of innovation that each program involves.

Horizontal expansions

41For horizontal expansions, governments needed policy capacities at various stages of the implementation process — during the identification of potential beneficiaries, during the beneficiaries’ intake assessment and enrollment, and, finally, during delivery of the cash transfer. Governments needed to be prepared — by having information about potential beneficiaries — to identify new individuals to enroll in the ongoing CTP. They also needed experience for assessing the eligibility of potential beneficiaries so as to enroll them. At the same time, they had to be prepared with information and had to have experience of delivering CTPs to be able to provide the cash benefit to the new beneficiaries.

42For most horizontal expansions, governments deployed such skills: 77% of the programs that were horizontally expanded used previous information for identification of new beneficiaries, and 66% used information for delivering the cash benefit. Moreover, all horizontal expansions took place in countries with experience of implementing at least one CTP. For example, besides creating the largest program in the region — Auxílio Emergencial — Brazil increased the coverage of Bolsa Família, a program designed to reduce poverty by promoting school and health-service attendance, which had been running for 17 years. This horizontal expansion was possible because its social registry, Cadastro Único, contained information about potential beneficiaries who were on the waiting list to receive the benefits of the program (6.9 million new direct beneficiaries). The availability of this information meant that Brazil was prepared to deploy a rapid horizontal expansion of an already large program, and that it had the skills to assess and enroll new beneficiaries, as a result of having more than two decades of experience of doing so.

43This experience contrasts with the horizontal expansion of the Programa de Aporte Económico al Adulto Mayor — a non-contributory pension for the elderly — in Guatemala, a country with experience (an ongoing CTP) but lacking a social registry. As a response to the pandemic, the government temporarily increased its budget so that it could be horizontally expanded. However, since it lacked ready-to-use information, it relied on its ability developed over the years. The program already had an established way of processing applications and payments, which required people to apply in person. Interested adults over 75 years old and living in poverty had to apply personally to the Ministry of Labor and Social Security and, if selected, had to collect their pension at a BANRURAL branch. When the pandemic hit, the process of identifying new beneficiaries remained the same: poor, elderly people (the most vulnerable to COVID-19) had to carry out these burdensome procedures in person.

Vertical expansions

44Even though increasing the size of the cash benefit (vertical expansion) of a given program seems straightforward, this is because governments had ready-to-use information about beneficiaries, making it administratively easy to send more money to them. Some long-standing programs in the region used this mechanism. For instance, Belize increased Boost‘s monthly payment from US$268 to US$325, a task that did not require additional information to enroll new beneficiaries or additional skills because, for more than nine years, this program had been reaching approximately 3,000 households, conditioning its benefits to school attendance and regular health check-ups.

45All other countries that expanded their programs vertically did so in a similar way. For governments to be able to expand a CTP vertically, they needed experience in delivering CTPs. In other words, they needed the know-how derived from implementing ongoing programs, and also preparedness (banking or beneficiaries’ location information from the ongoing program’s registry). There cannot be a vertical expansion if a CTP does not exist, and no further information is needed because beneficiaries are already enrolled and receiving the cash transfer. Naturally, all 18 vertical expansions occurred in countries (9) with experience of implementing CTPs.

New programs

46All countries that responded to the pandemic with a CTP created at least one new program. This type of innovation was the most frequent response (95 out of 122). Designing and implementing a new program demanded both the know-how to operate and the analytical skills to make sense of the information and resources garnered. Governments needed analytical skills to select the most appropriate policy instruments to attain their goals, a task that could be facilitated by the experience they had of implementing other ongoing CTPs. Meanwhile, they also required information on potential beneficiaries, and skills to assess their eligibility, to enroll them, and to transfer the cash benefit to them. As mentioned above, most new programs relied on existing information to identify beneficiaries, and to deliver the transfer. A total of 78 out of 95 new programs were launched in countries that had experience of implementing an ongoing CTP. The ways in which these capacities were deployed was diverse. In the following examples, we show how different combinations of capacities shaped the responses: CTPs in countries with experience and preparedness, which reacted as their capacities allowed them to do; CTPs in countries with experience, which benefited more people than they had on social-protection registries; and CTPs in countries without this specific set of policy capacities.

47Quédate en Casa (Stay at Home), a new CTP created in the Dominican Republic to benefit informal workers, is an example of a program that took advantage of its existing skills throughout the design and implementation stages, as well as of existing information —particularly concerning its Single-Beneficiary System (SIUBEN), created in 2004. The program benefited from the targeting that the program Progresando con Solidaridad had already done when first implemented more than eight years earlier. Stay at Home automatically enrolled the beneficiaries of that program, thus taking advantage of experience and information that enabled the Dominican government to act expeditiously. Indeed, since Progresando con Solidaridad already had 800,000 registered families, and a network of more than 2,000 businesses that accepted prepaid bank cards, the government used the existing banking information and infrastructure to target and deliver cash benefits (Visa, 2020). To reach even more households, Stay at Home also selected new people from the SIUBEN where their socioeconomic conditions were already registered. All countries with experience of implementing a CTP and information in their social or beneficiary registries had the capacities needed to deploy new programs if they decided to do so. The Dominican Republic was one of 13 countries in this situation, with previous information and experience that enabled them to benefit their populations.

48Unlike the Dominican experience, some countries launched new programs that did not use information derived from existing social registries although the countries did have experience of implementing an ongoing CTP. Among these, Guatemala stands out for the magnitude of its response. Indeed, Guatemala did not have information about beneficiaries and was thus not prepared to reach new people easily. However, through its new program, Bono Familia (Family Bonus), which sought to benefit those people who could be most affected by the confinement measures, it had coverage of two million individual beneficiaries, equivalent to 12% of its population. Indeed, even though the country had administrative records and information about poverty among the population, these datasets were not compatible (CIEN, 2020). Thus, the government created a new dataset by combining information from energy-distribution companies, the Guatemalan Social Security Institute, and the Ministry of Social Development (MIDES). This ministry merged and systematized the information to identify those households with electricity consumption below 200 kW, which was the criterion for eligibility. The government then let people know that they were pre-approved as beneficiaries through their energy bill. Then the pre-selected beneficiaries had to register their personal identification number and telephone number through a webpage, a phone call, or even an SMS, to apply to the program. After MIDES assessed the eligibility of pre-approved applicants — as it had been doing for Bono Social for the previous eight years — the beneficiaries received a code through their mobile phone, which they could then use to make purchases and cash withdrawals. The ability to deliver the cash benefits, present in the entire financial industry’s infrastructure in the country, was crucial for this response to take place (Visa, 2020). All those countries which, like Guatemala, intended to benefit more people than they had information about, had to come up with new ways of identifying and enrolling beneficiaries. They were able to do so, thanks to the experience they had in implementing similar programs.

49Other countries made unexpected responses. Indeed, seven Caribbean countries responded to the emergency with interventions that reached people beyond their existing capacities, because they had no experience of implementing an ongoing CTP. For instance, Guyana launched the COVID-19 Relief Fund program — a one-time transfer to support the socioeconomic effects of the pandemic — with a number of direct beneficiaries equivalent to 19.8% of their population. However, in a country with no experience of implementing an ongoing CTP, and with low-tech information systems (administrative records are incomplete and some of these are kept only on paper) (Bailey & Ciardi, 2020), the government had to rely on the territorial structure of its local authorities which visited each household to carry out physical verification of the potential recipients’ conditions. With no further assessment process, local officials selected those households eligible to receive the US$120 transfer which was delivered by authorized officials in person. Governments without experience and preparedness had to implement responses that were inevitably time-consuming and demanding of human resources. The Guyanese response was similar to those deployed by the other six countries that lacked both experience and preparedness, and was not as expeditious as that implemented in countries with capacity. Although the coverage reached by Guyana was slightly lower than the regional average (23.8%), the other six countries did not benefit more than 10% of their populations.

50From this analysis of the decisions and processes involved in the design and implementation of these programs, we can draw two key implications. First, while the types of response that governments were able to deploy depended on each country’s departure point, the extent to which experience and preparedness were necessary for responding also depended on the intended type of response that each country wanted to deploy. Second, governments might make different use of their policy capacities to respond to the same policy issue. Indeed, most large countries in the region responded in more than one way: they deployed new programs and, at the same time, carried out vertical or horizontal expansions. However, the use they made of their information or experience was not the same for every response they provided. As argued before, policy capacities are dynamic and contextual.

Unbundling policy capacities

51Observing specific programs from the standpoint of the policy capacities used to deploy them makes it evident, first, that countries employed different configurations of policy capacities, depending on their type of CTP response. This finding supports the idea that policy capacities are not one generic asset equally relevant for every decision involved in the policy process. They are not a unitary asset because they are made up of resources (informational resources, in this case) and a set of organizational practices and structures through which governments have learned over time how to manage them. Belize’s vertical expansion of Boost, for instance, shows that alleviating the economic stress of certain families did not require any implementing agent to learn how to do anything that was different from what they had done for the previous nine years. But policy capacities are not equally relevant for every decision either, because — as shown in the six programs described above — the type of information, its purpose, and the skills required for its analysis are not the same for identifying beneficiaries as for delivering the cash transfer. Bono Familia (Guatemala) showcases this distinction: for the government, identifying new beneficiaries was possible thanks to its experience, which allowed it to create a dataset using existing information from energy-distribution companies. This information made it possible to target the program to certain households, as well as to let them know that they were eligible. However, this information was not useful for delivering the payment. Information for this second phase had to be obtained through other means: each beneficiary had to provide information directly on a webpage created for it; the government would then use it to send a code with which the beneficiary could make cash withdrawals or purchases.

Table 3. Policy capacities throughout the implementation process, by type of response.

	Instrument selection	Identifying potential beneficiaries	Beneficiaries’ intake assessment and enrollment	Cash transfer delivery
Vertical expansion (increasing the size of the payment)	Having experience and being prepared is irrelevant CTP design is already in place	Having experience and being prepared is irrelevant There is no need to identify new beneficiaries	Having experience and being prepared is irrelevant Beneficiaries already enrolled	Experience Operational skills derived from ongoing program and Preparedness Information (banking or beneficiaries’ location from the ongoing program’s registry)
Horizontal expansion (increasing the coverage)	Having experience and being prepared is irrelevant CTP design is already in place	Preparedness Socioeconomic information about potential beneficiaries from social/beneficiary registry	Experience Analytical and operational skills derived from ongoing program to include new beneficiaries	Experience Operational skills derived from ongoing program and Preparedness Information (banking or beneficiaries’ location from social/beneficiary registry)
New program (intervention designed ex profeso in response to the pandemic)	Experience Analytical skills derived from implementing other CTPs	Preparedness Socioeconomic information about potential beneficiaries from social/beneficiary registry	Experience Analytical and operational skills derived from other program to include new beneficiaries	Experience Operational skills derived from another program and Preparedness Information (banking or beneficiaries’ location from social/beneficiary registry)

Source: Authors’ elaboration.

53By definition, unless governments had an ongoing CTP, they would not be able to respond with a vertical or a horizontal expansion. To conduct a vertical expansion, however, it is irrelevant whether governments had experience and were prepared to select policy instruments and to identify and enroll beneficiaries, because this kind of response is limited to providing larger benefits to the population that they were already targeting. However, they would need to be experienced (having the operational skills derived from implementing an ongoing program) and to be prepared (possessing banking or beneficiary location information from the ongoing program’s registry) to be able to deliver a higher-size payment to beneficiaries.

54Also, by definition, for a horizontal expansion, it would be necessary to have an ongoing CTP. Although having experience and being prepared for selecting policy instruments is irrelevant, because the CTPs were already designed, this type of response required governments to have information to identify new beneficiaries and to have had experience of assessing and enrolling them. Indeed, governments would need to have socioeconomic information about potential beneficiaries to expand the program’s coverage. Moreover, a bureaucracy with the analytical and operational skills to assess the eligibility of new beneficiaries, and to enroll them in the ongoing CTP, would be crucial. Delivering this horizontal expansion would also require the agencies responsible for the implementation of the CTP to have the requisite operational skills and banking information concerning the new beneficiaries.

55Finally, responding to the pandemic with the creation of a new CTP would require further policy capacities, since it could not be done unless governments had the analytical skills to design the program, as well as the experience and information needed to implement it. When dealing with complex problems, designing a new policy demands a meticulous articulation of the stated policy goals, the selected policy tools, and the current governance modes (Tosun & Lang, 2017). While having these skills is crucial for developing new programs, having experience of implementing similar responses makes this kind of innovation less dependent on the individual’s and the organizations’ analytical capacities.

Discussion

56Unbundling the different configurations of policy capacities needed for carrying out every task involved in responding with a CTP (Table 3) makes it evident that capacities are specific. This has been explained in the policy instruments literature when empirically demonstrating that the behavioral assumptions that guide the selection of a specific set of policy instruments “are contingent not only on the policy instrument at hand but also the specific policy and local context” (Wegrich, 2021, p. 12). However, as our analysis shows, policy capacities are not only context-specific but also task-specific. This feature is what distinguishes them from broader notions of state capacities that use generic indicators (such as the World Bank Governance Indicators, Gross Domestic Product Per Capita, or taxation) to measure a government’s capacity to implement policies. State capacities vary from sector to sector (Knill, 2005) and thus their empirical analysis requires sector-specific indicators. For instance, for measuring health responses, Toshkov et al. (2021) use the number of nurses per capita or health expenditure per capita. Yet variations in the responses to the pandemic have been attributed to state capacity when it is reasonable to expect that, within the same government, the ability of a health agency to vaccinate its population will differ from the ability of other agencies to implement any social program or to manage remote education. Moreover, recent research on governmental responses to COVID-19 has shown that the capacities required differed from task to task. For example, contact tracing (Cingolani, 2022) required informational capacities different from those required for vaccination strategies (Peeters et al., 2023).

57Our analysis of CTP responses to the pandemic also supports the idea that policy capacities are more than just a set of generic assets or static skills. As the in-depth analysis of the cases shows, having ready-to-use information made the task of responding easier and swifter; however, the lack of this did not impede some other governments in their responses (several Caribbean countries are examples of this). The fact that such governments could compensate for their lack of information by developing new mechanisms to obtain socioeconomic information through very diverse means, demonstrates that they could be agile and resilient. Capacities must be understood and developed as part of a dynamic context where stability and predictability are, in fact, exceptional (Mazzucato & Kattel, 2020).

58Having governments that made CTP responses despite lacking information might give the wrong idea that policy capacities do not explain CTP responses except, perhaps, in terms of their speed. Yet a closer look at the data presented above shows that countries that lacked social-protection information systems for targeting and deploying a response did so with a very limited scope. The coverage of those responses was influenced by the experience of their decision-makers and bureaucracy in obtaining information about potential beneficiaries or new mechanisms for delivering the payments. Moreover, those few countries that created programs from scratch not only reached a small percentage of their populations but placed heavy burdens of time and effort on their beneficiaries and public officials. Indeed, while creating mechanisms to retrieve information may be possible in the context of an emergency, the knowledge needed for designing and creating a program that reaches a specific sector of the population could only result from having experience of doing so. This means that, while experience may compensate for a lack of preparedness, the lost time in developing new skills cannot be made up.

59The different configurations of policy capacities for conducting specific tasks in the delivery of CTPs depict a more complex understanding of this concept, one which goes beyond having resources or even having the skills to make strategic use of resources.

Conclusion

60In this paper, we have extended the discussion of policy capacities during the pandemic, building on Capano et al.’s (2020) conceptualization of capacities as a mix of preparedness and experience, by looking at different configurations of policy capacities during the design and implementation of CTPs in LAC. Extending current work on the deployment of policy capacities during the pandemic, and building on recent research on policy capacities which emphasizes that these are not generic skills or resources but, instead, are context-dependent, we show that there are multiple possible combinations of both dimensions of policy capacities, and that the ways in which they have enabled policy responses vary greatly.

61By observing preparedness and experience in every type of policy response, we have shown that both dimensions of policy capacity should be operationalized in relation to the task that they are supposed to be enabling. Based on our operationalizations, we have provided an analysis of how policy capacities enabled policy responses to a specific issue in the form of CTPs for an emergent, temporal response. We have explained how different combinations of policy capacities enabled governments to reach their populations (old and new beneficiaries) and to provide assistance in response to this emergency by making use of their experience in implementing a CTP and of the information gathered over the years.

62Our findings cannot be automatically generalized to other policy sectors. We have shown the importance of generating sector-specific operationalizations of policy capacities. A promising avenue of research in comparative public administration is to compare how experience is built over time in different sectors, and to show what preparedness means in specific policy areas. There is also a need to explore how capacities might be rapidly developed when time is at stake. This agenda could benefit from engaging with the policy-feedback literature (Béland & Schlager, 2019), policy learning and innovation capacities (Borrás, 2011), and social policy expansion (Arza et al., 2022) to gain a better understanding of how previous decisions on expanding social programs to new populations facilitated the response during the pandemic and could do the same for future expansions. This research would confirm the importance of generating (and preserving) bureaucratic capacities and the key importance of information systems in dealing with future challenges.