Mapping the center of government in Latin America and the Caribbean: a typology

Over two decades ago, Bardach (1998, p. 4) highlighted a core problem of public administration: “Agencies cooperate? Of course not!” One of the solutions identified by scholars and practitioners has been the center of government (CoG), that is, organizations created by elected officials to facilitate inter-ministerial cooperation for the design and implementation of priority policies. Originally developed in European parliamentary systems or at Franklin Roosevelt’s presidential center in the United States, the CoG has recently gained momentum in Latin America and the Caribbean (LAC) after a collective effort between the IDB and the OECD. This paper maps the historical path and most recent institutional design of CoG organizations in 27 LAC countries as of 2020. The outcome of an exhaustive data collection process is a typology with five dimensions and 15 ideal types that help scholars and practitioners to better understand how and why these organizations have now been widely adopted in the region.


INTRODUCTION
In his seminal book "Getting agencies to work together, " Bardach (1998, p. 4) shouted one of the core problems of public administration: "Agencies cooperate? Of course not!" This derives from, among others, Allison's (1969) Bureaucratic Politics Model, which analyzes the conflicts of interests resulting from the interaction between top managers from different public departments. Roughly five decades later, we know that coordination problems are not exclusive to agency heads. They go from politicians in charge of different portfolios (Preston & Hart, 1999) to street-level bureaucrats motivated by conflicting interests (Brower & Abolafia, 1997;Kalkman & Groenewegen, 2019).
How do Latin American governments attempt to improve inter-agency cooperation at the national level? One of the responses to address coordination problems is the development of central agencies, that is, structures responsible for planning, guiding, and monitoring the activities of different portfolios within government (Peters, Dahlström & Pierre, 2011). In this paper, I map the status of centers of government (CoG) in Latin America. Adding up to the existing literature, this exploratory effort includes classifying the historical development of these structures, their main responsibilities, and how they fit the administrative and political structure of different countries. Overall, the goal of this paper is to provide a comprehensive typology of these central agencies that function as a bureaucratic response to coordination problems.
The working definition of CoG used in this paper derives from the seminal study of Alessandro, Lafuente and Santiso (2014, p. xvii) in Latin America. They conceptualize this term as: the organizations and units that provide direct support to the country's chief executive (president or prime minister), generally for the political management of the government's actions, the strategic management of its priorities, the coordination of policy design, the steering of policy implementation, the monitoring of performance and delivery, and the communication of results.
Whereas Alessandro et al. (2014) are dedicated to the study of CoG in Latin America, they highlight that this is neither a new structure nor one restricted to this region. Traditional examples are found in Australia (i.e., Department of the Prime Minister and Cabinet, created in 1911), the United Kingdom (i.e., Cabinet Office, 1916), and the United States (i.e., Executive Office of the President, 1939). All of these organizations had in common the task of guiding the implementation of relevant policies especially when cooperation of different agencies was required.
Interest in this type of government response has been increasing (Inter-American Development Bank [IDB], 2020;OECD, 2020). Different factors explain this surge among scholars and practitioners. Consider the growth of government and recent administrative reforms. First, there has been a substantive expansion in the roles of the state over time (Higgs, 2013). This has been made possible by, among others, technological advances (Cowen, 2009;Kau & Rubin, 1981) and increased political franchise (Husted & Kenny, 1997), thus leading to more complex governmental structures.
In the meanwhile, governments have adopted several reforms toward decentralizing their functions. One of the core aspects of New Public Management (NPM) is the empowerment of lower-level managers (R. Andrews & Walle, 2013), who gained more attributions and discretion (Peters & Pierre, 2000). Whereas this may increase efficiency in certain policy areas, Peters and Savoie (1996, p. 282) highlight that "In practice, […] the fragmentation of control structures may contribute to policy incoherence and result in a manifest failure to exercise effective control over the bureaucracy. " This explains the "whole-of-government" movement that, following NPM reforms, attempted to deemphasize isolationism and make cooperation across portfolios more common (Christensen & Laegreid, 2007).
These temporal changes imply that, if central agencies were already relevant in the early development of democratic governance, their presence in administrations in the 21 st century is indispensable. This led the Inter-American Development Bank (IDB) to establish the regional project Strengthening and Promoting Innovation in Center of Government Institutions in 2013 (Alessandro et al., 2014), now the Institutional Capacity of the State project (IDB, 2020). This is complemented by the Organization for Economic Cooperation and Development's (OECD) LAC Regional Program which, starting in 2016, has also been promoting technical assistance and international dialogue to strengthen central agencies (OECD, 2018).
Motivated by the growing relevance of the topic in Latin America, the remaining parts of this paper are divided as follows. First, I review the existing literature which is mostly interested in the study of central agencies in developed countries. These studies often include the historical development of these structures, their main responsibilities, and how they fit the administrative structure of different countries. In the next section, I use these pillars to describe the state of the art of the CoG in the region and to propose a typology of coordination solutions. This is based on data collection conducted in 27 countries and territories in Latin America and the Caribbean -the most comprehensive study in the region so far. Finally, I discuss the status of these structures in the region in light of the existing literature. This leads to a conclusion where I present inputs on how to move forward in this research agenda, as well as potential reforms in the region.

WHAT DO WE ALREADY KNOW?
Coordination is an essential feature of policy implementation. 1 When studying the failure of a project developed in Washington, D.C. for implementation in Oakland (California), Pressman and Wildavsky (1984, p. 94) list some reasons why different agencies had to work together: Some of these participants […] became involved because they possessed jurisdictional authority over important parts of the project; others […] entered the process when they felt their interests being impinged on; and still others […] were intentionally brought into the program […] to build local support for the projects.
Indeed, horizontal coordination (i.e., among members of the same administrative level) may appear to be less complex than vertical cooperation (e.g., between the federal and a local government). Yet, it is not unlike to see different organizations joining a single implementation process for the same reasons listed above. Furthermore, one should not expect prima facie that these agencies will automatically adhere to a single mission and, leaving aside their differences, agree on all the details of the project and dedicate full efforts to its execution (see Allison, 1969). After all, what is appropriate for one organization (e.g., mission, concepts, way of doing things) will not necessarily be for another (March & Olsen, 1989).
There are at least two additional complexities that harm collective action within governments. First, cooperation is not a one-shot game (Allison, 1969;Bardach, 1980). Many agencies must cooperate on multiple occasions throughout time, thus creating a history of interactions that will affect how they see each other and, ultimately, their willingness to work together. Second, these organizations will not necessarily share the same views as those espoused by elected officials (Downs, 1967) -in fact, this relationship may be even more complex since public officials are replaced from time to time. In this wicked context, one should expect that governments develop institutions to mitigate collective action problems and that these institutional arrangements will be adapted over time as a manner of responding to the history of inter-agency interactions and the reality of the current administration. This explains the rise and evolution of the center of government.

Conceptual framework
The literature on the CoG uses different terms to refer to this concept. This often includes central agencies and the presidential center. Whereas they are comparable, the use of each of these terms often involves certain particularities. First, central agencies have a starker focus on specific structures within the government. For instance, Painter (1981) discusses the creation of new agencies as a tool to alleviate coordination problems. These structures have tasks which "concern aspects of policy and administration that transcend departmental boundaries, whether because a broader 'whole government' perspective is required or because several departments or agencies overlap in parts of their work" (Painter, 1981, p. 274). Thus, the reference to central agencies may occur in the plural, given that their responsibilities need not be concentrated in a single structure -in the U.S. government, these include the Office of Management and Budget, the Office of Personnel Management, and the General Services Administration (Levine, 1986).
The concept of presidential center follows, in many aspects, the definition of central agencies. Yet, their recall to a regime type that is mostly encountered outside of the European scope highlights the relevance of these structures in countries as the United States and most of Latin America and the Caribbean. The first presidential center among developed democracies arises from the Executive Office of the President created in the United States in 1939 aiming "to strengthen the president's ability to coordinate the work of cabinet departments and other executive agencies" (Bonvecchi & Scartascini, 2014, p. 147).
Whereas the presidential center may refer to the creation of new organizations, as discussed by Painter (1981) when analyzing central agencies, that term is often linked to the inner circle of the presidential office (Mejía-Guinand, Botero & Solano, 2018). In this sense, Bonvecchi and Scartascini (2014, p. 147) define the presidential center as the "closer group of advisers with no necessary departmental responsibilities who work under their most direct supervision. " In comparison, the reference to a CoG is broader. Savoie (1999, p. 3) begins his analysis of the topic in Canada by saying that "by center of government, I mean the prime minister, his office, the Cabinet, and central agencies. " Naturally, the starting point is the term's applicability to systems other than presidential -as it also occurs when one refers to central agencies. Yet, it allows for a more abstract approach, as this is not restricted to the formal nature of structure-creation that is implied when 'agencies' are evoked. For that reason, I opt for the broadest conceptualization of all. Besides allowing the inclusion of organizations across different regime types, this also permits studying agencies that perform CoG functions even though they are not often labeled as central agencies.

Historical development
The origin of the CoG is often related to the secretariat of the head of government. One common example in the literature is the Australian Department of the Prime Minister and Cabinet, whose history dates back to the 1910s (Hamburger & Weller, 2012). Yet, it was the excessive growth of government in the post-World War II (Hamburger & Weller, 2012;Hamburger, Stevens & Weller, 2011;Yeend, 1979) and the direct influence of the British Cabinet secretary (Savoie, 1999) that led this structure to gain the traditional roles of central agencies as it has nowadays.
Similarly, the creation of the Cabinet Office in the United Kingdom was also influenced by a world war. In 1916, following the growth of government caused by the war economy (Hayek, 1997), the Cabinet Secretariat gained prominence as the central body in the organization of portfolios (Seldon & Meakin, 2016). The further expansion in the role and expenditure of public authorities during the second war consolidated this agency, which remained central in times of downsizing (Seldon & Meakin, 2016).
In presidential regimes, the American case is among the initial references to the development of a comprehensive CoG. After the 1929 crisis and the election of Franklin D. Roosevelt, there has been a significant expansion of government in terms of political relevance and financial expenditure (Higgs, 2013). Dickinson (1996) explains that, during this administration, the president maintained power over the expanding portfolios of government by institutionalizing his inner circle. This means establishing routines and empowering his closest advisors who were responsible for coordinating the different arms of the executive branch. Ultimately, that strategy resulted in the creation of the Executive Office of the President. The same rationale is promoted nowadays by the Inter-American Development Bank in Latin America and the Caribbean (Alessandro et al., 2014).

Responsibilities and overlap
Painter (1981) highlights that the CoG is responsible for the coordination of both administrative and policy issues. The first requires the enforcement of routines that are going to be valid for the whole government. The second relates to the substantive content of policies that often require multiparty convergence. In both areas, CoGs must also include proper planning to define priorities and reduce overlap and engage in monitoring to make sure that the different agencies of government are following the expectations of the center.
Whereas overlap is a common problem in the list of issues to be addressed by the CoG, Johnson (1974) suggests that this is also a strategy adopted by some elected officials when designing their central agencies. Duplication may lead to competition that increases the efficiency of certain public actors (Niskanen, 1971). Additionally, this is also a tool to assure that policies will be implemented even when existing agencies decide to shirk (Costa, 2008).
In other cases, the overlap is an outcome of specialization. Consider the case of ministries of finance, public service, and planning. All of them have responsibilities that could be attributable to a presidential center or the cabinet office. Yet, most countries have specific agencies for at least one of these sectors. Many authors tend to include these structures in the concept of CoG (Alessandro et al., 2014), while, as already argued, these may be left out by those who exclusively discuss the presidential center (Mejía-Guinand et al., 2018).
Overall, while the presidential center may have overlapping functions with the rest of the CoG, these are implemented through different means. The inner circle of the head of government will help the elected officials to provide guidelines that will be followed by other central agencies. These, in return, will use their expertise in the specific policy area to further develop and implement policies that are in accord with the expectations of the government.

EMPIRICAL STRATEGY
I collected data from the CoG of 27 Latin American and Caribbean (LAC) countries and territories. Here, I adopt a broad definition of LAC. This means that the list is not exclusive to those countries that were colonized by Portugal, Spain, and France. Instead, I include all countries to the south of Rio Grande. This includes former British and Dutch colonies, as Suriname, Belize, and Guyana. Thus, this represents the broadest data compilation in the topic of CoG in the Americas so far.
Further than including countries and territories that are often overlooked in the study of public administration and politics in the region, this strategy adds a relevant complexity to CoGs, i.e., variation in regime type (Box 1). Most often, the study of Latin American countries is focused on presidential systems. Yet, among the developing democracies in the Americas, there are six parliamentary systems and one semi-presidential regime. Furthermore, two autocracies merit comparisons.
Source: Elaborated by the author.
The data collection process follows the longstanding tradition of documentary research. According to McCulloch (2004), this methodological approach allows us to understand how texts fit and help explain complex contexts. In his words, The bureaucracies of the nation-states produced copious records of their development, and of their dealings with different interest groups. They produced large numbers of reports on the problems that they encountered and the policies that they favored, and their consultations were also transcribed in loving details.
[…] Strict, formal record keeping was routini[z]ed and became a discipline in its own right. Such records might well provide insights into the processes and workings of the social structures of the modern age (McCulloch, 2004, p. 11).
For that purpose, McCulloch (2004) suggests the collection and analysis of different types of documents. I am interested in two of them. First, legislation, which does not include interpretations of any sort. These include constitutional texts, laws, and other types of regulatory documents that were designed either by elected officials or bureaucrats as a manner of creating new organizations or institutionalizing their practice. These legal pieces should be complemented by interpretative pieces (e.g., academic literature, government reports, news reports), which help make sense of legislation in their specific context.
Thus, the initial step in this data collection process regarded an assessment of the existing literature, as this is helpful not only to understand the different contexts of Latin America and the Caribbean but also to identify relevant legislation. In most cases, this was absent. The first exception regards regional reports, which mostly derived from the work of Alessandro et al. (2014) under the umbrella of the Inter-American Development Bank. When it comes to country cases, this field of study is most developed in Brazil (e.g., Cavalcante & Batista, 2019), followed by recent studies in Uruguay (Lanzaro, 2016;Milanesi & Gadea, 2017) and Chile (Baraona & Herrera, 2011;Franco, 2012;Palacios & Jofré, 2016).
Second, I collected data from government documents. This includes a review of constitutional texts, laws, executive orders, and other regulations that involve the creation or reform of central agencies. This information is complemented with descriptions obtained through public websitesin many cases, agencies publish a brief narrative of the organization's history. This has been helpful to identify pieces of legislation, as well as to trace back their historical path. Furthermore, selected policy guidelines are often published online. This includes, for instance, the documents published by Jamaica's central agency that shall be used by different agencies to report their main projects and indicators.
Finally, I relied on information from local newspapers. These were often used to complement the existing data especially in countries that recently created or reformed their central agencies. In certain cases, this was an adequate mechanism to detect the influence of international organizations on the development of CoGs -a common trend in several Central American and Caribbean countries.
In all these cases, I opted for a broad understanding of CoG. That is, I focused on the different organizations responsible for at least one of the functions of the CoG, as expressed in the conceptualization of Alessandro et al. (2014). For instance, I included all organizations responsible for supporting inter-agency cooperation whenever this was referred to in the literature, government documents, or local newspapers.
To process the data, I categorized the information based on a series of criteria (e.g., historical path, relevant legislation, head of the center of government, executive structures, collegiate bodies, policy tools, etc.). To a certain extent, this resembles the proposition made by Bardin (2011) regarding content analysis. That is, after reviewing the full corpus, I created the selected labels and went through each institution to label them according to the proposed typology.
As will be further explained in the following section, these ideal types are distributed within five dimensions. Two of them regard the application of the history of the CoG in the region (Historical Path and Change over Time), two are related to agency type (Executive Agencies and Collegiate Bodies), and the last is specific to instruments (Policy Tools). This approach allows a comprehensive comparison of the 27 selected cases, ranging from how these institutions evolved to their relevant mechanisms to facilitate inter-agency cooperation. In the following section, I systematize and present the findings of this exploration.

CENTER OF GOVERNMENT IN LATIN AMERICA
The categorization process led to five dimensions: historical path, change over time, executive agencies, collegiate bodies, and policy tools. Each of these dimensions has either two or three ideal types. Following the Weberian empirical approach, this does not mean that countries or cases fit perfectly well into single ideal types. On the contrary, most cases include a mix of different categories. Yet, the typology is relevant to understand the main elements in the formation and organization of CoG in the region. The ideal types are presented in Box 2. They are explained and illustrated in the following subsections of this paper. Yet, before moving forward, it is important to clarify the five dimensions proposed here. First, two of them are dedicated to the longitudinal nature of the CoG: while the Historical Path captures the plural historical junctures that influenced the creation and evolution of these institutional arrangements, Change over Time regards how the political context led to different problems and solutions in government.
The next two dimensions regard agency type. On the one hand, governments adopted centralized solutions through the creation of agencies responsible for designing, controlling, and overseeing implementation processes. On the other, these tasks were delegated to collegiate bodies where members of different organizations were tasked with coming up with common solutions. Finally, the last dimension is dedicated to exploring those tools used by the different organizations to perform the tasks of the center of government.

Historical path
Regardless of country, the center of government tends to evolve. Even in the absence of specific institutions responsible for inter-ministerial relations, it is possible to identify government officials responsible for similar tasks. For instance, the prime minister, by definition, coordinates the cabinet; additionally, the ministry of finance is responsible for distributing budgets according to the priorities of the head of government. Thus, to argue that some countries only developed a center of government in the last, say, five years, is not entirely accurate.
Yet, if we focus on structures that are specifically created for policy coordination, it is possible to point out three categories of historical paths: post-independence, structuralist, and globalization. Framed as ideal types, countries should not be understood as belonging exclusively to only one of these categories. Nonetheless, they help to perceive the development of CoG in the region.
The first category, the post-independence period, includes those countries which created an institution for policy coordination as soon as they left colonial dominance in the 19 th century. These structures were often created as a secretary in the early presidential circle to support the head of government in handling the different ministerial portfolios. They speak to the need of strengthening the political control over newly created states.
This was the case of Colombia, where Law 3/1898 established the Secretaría General de la Presidencia de la República. Over time, this structure evolved to its form, the Departamento Administrativo de la Presidencia de la República -most recently reformed through the Decree 901/2020, which added the presidential policy priorities in the list of responsibilities of the organization. Even older, the Dominican Republic created the Secretaría de Gobierno through Law 38/1845. In 1927, Law 685 transformed this into the Ministerio de la Presidencia, which still exists nowadays.
Especially in larger countries, the responsibilities of the post-independence CoG included both horizontal and vertical cooperation. That is, the organization coordinated inter-ministerial cooperation (horizontal), as well as political control over the full territory (vertical). Whereas the latter is less relevant for this paper, this was necessary given constant territorial disputes and the under-development of sub-national divisions. While in some cases the CoG still holds the role of vertical coordination, the common trend was separating the institution into two organizations. The one responsible for vertical coordination is often referred to as Ministerio de Gobierno or Ministerio de Interior. Here, the Mexican case serves as an example. After independence in 1821, the newly established government created the Secretaría de Interior y de Relaciones Exteriores, which was dismembered after it became the Ministério de Gobernación in 1867. Since 1917, it evolved into this newer structure for horizontal coordination, which is called the Secretaría de Gobernación.
The second category regards the structuralist period. In the aftermath of World War II, the Economic Commission for Latin America and the Caribbean (ECLAC) pushed for greater intervention of the state in the economy. This structuralist agenda included protectionist policies, subsidies to local industries, as well as the active role of government in the creation and expansion of new firms (Prebisch, 1962). Similar to what happened in the US during the crisis of 1929 (Higgs, 2013) or in Europe in the war period (Hayek, 1997), the growth of government required complex management systems dedicated to economic planning and the implementation of inter-sectoral policies. Given the focus of several governments on the economic agenda, new executive and collegiate bodies dedicated to central planning became the new central agencies.
In The third category includes those countries that have either only recently developed their CoG or promoted substantive reforms in the organization of their central agencies due to interactions with international organizations. There is a vast literature dealing with policy diffusion and policy transfer (Graham, Shipan & Volden, 2013;Marsh & Sharman, 2009;Shipan & Volden, 2012). Focusing on anti-money laundering policies, Sharman (2008) argues that developing countries are significantly more vulnerable to international power-based influences over the creation of new structures to respond to pressing issues.
This has been the case of Guatemala, where the Acuerdo Gubernativo 16/2017 created the Comisión de Gestión Estratégica, the embryonic structure that evolved into the Comisión Presidencial del Centro de Gobierno in 2020 (Acuerdo Gubernativo 31). This collegiate body is composed of the Ministro de Finanzas Públicas, the Secretario de Planificación y Programación de la Presidencia, the Secretario General de la Presidencia, and the Director Ejecutivo de la Comisión Presidencial Contra la Corrupción. The latter speaks directly to Sharman's (2008) work, as it has also been established in response to international influences.

Ideal Type Explanation
Post-Independence Created in the late 1800s and early 1900s to create state capacity and territory control.

Structuralist
Created in the mid-20 th century to facilitate economic control over growing expenditures.

Globalization
Created in response to the pressures of international organizations in the late 1900s and early 2000s.

Source: Elaborated by the author.
This is similar to the case of Haiti. Whereas the cabinet (Conseil des Ministers) functions as the main organization in the CoG, the government created the Unité de le Lutte Contre la Corruption in 2004 as a response to the Interamerican Convention Against Corruption. This inter-ministerial effort is responsible for coordinating policies related to corruption prevention and detection.
Paraguay is part of this list of countries not due to the recent creation of a CoG but because of its recent shifts. In 2018, President Benítez gave ministerial powers to the Unidad de Gestión de la Presidencia de la República and assigned Hugo Cáceres Agüero as its minister. Mr. Agüero is a former high-level employee at the Inter-American Development Bank.

Change over time
The historical path of CoG in Latin America hints at the fact that, within the same country, governments may have different strategies to mitigate coordination problems. The recent literature focused on specific country cases highlights that it does not take a century or a decade to reform the CoG. Whereas incremental changes are expectable, central agencies also tend to be modified depending on the existing coalition or the head of government's trust over political appointees.
Argentina offers an example of incremental changes. Especially since the Ley de Ministérios (Decree 438 of March 12 th , 1992), every Argentine president publishes a new decree containing the distribution of ministries and their responsibilities. Based on that, President Alberto Fernandez enacted the Decree 8/2019 reinforcing the role of the Jefatura de Gabinete de Ministros as the center of government. This structure was created by President Ménem in 1995 following 1994 Constitutional Reform but only gained prominence during the Macri Administration (2015-2019) (Coutinho, 2018).
In Brazil, the attributions of the center of government have been distributed across different agencies, which had a volatile level of influence depending on the presidential administration After the impeachment of Rousseff, PAC has been mostly discontinued. Then, President Temer used the Ministério da Casa Civil to create the Comitê Interministerial de Governança. Composed of representatives of the Ministério da Casa Civil, Ministério da Economia, and Ministério da Controladoria-Geral da União, this committee is responsible assisting the president in the implementation of the federal standards of governance. This includes strategic planning and the preparations of guidelines to support the different ministries of the country.
The case of Chile is not significantly different. Baraona and Herrera (2011) suggest that the role of strategic planning was controlled by the Ministerio Secretaría General de la Presidencia (SEGPRES-Ch) and Ministerio de Hacienda during the Aylwin administration (1990( -1994( ) but shifted under president Frei (1994( -2000 to a steering committee that also included Ministerio de Interior, Ministerio de Defesa and, later, the Secretaría de Gobierno -the so-called Círculo de Hierro. These structures lost power when President Lagos (2000)(2001)(2002)(2003)(2004)(2005)(2006) strengthened the inner circle of the presidency, or "Segundo Piso del Gobierno" (Baraona & Herrera, 2011, p. 146). This has been formalized through the Law 19,882/2003, which created the Sistema de Alta Dirección Pública, allowing the appointment of high-level bureaucrats in the presidential center (Franco, 2012). Since Bachelet (2006Bachelet ( -2010, the roles of the CoG were distributed between the Segundo Piso and the ministries within the Comité Político (Baraona & Herrera, 2011) -a mix of the strategies adopted by the three previous presidents. All authors converge in that the way the CoG has been managed depended on the president's coalition (Baraona & Herrera, 2011;Franco, 2012;Palacios & Jofré, 2016).
In Uruguay, Lanzaro (2016) highlights that each president emphasized structures in different ways. During his first administration, President Tabaré Vázquez (2005-2010) delegated more powers to the Consejo de Ministros if compared to previous governments. According to a press release from the Uruguayan presidency (Presidencia de Uruguay, 2019), this head of government hosted over 70 public meetings of this council in the different subnational units of the country, leading to over 23 proposals -most of which were implemented.

Incremental
Minor changes promoted by new administrations seeking the specific goals of each chief executive.

Coalitional
Re-organization motivated by the balance of power between the government and its allies.
Source: Elaborated by the author.
Vázquez's strategy also included the creation of sectoral cabinets, i.e., councils composed of ministries of related areas. However, they were downplayed in the Mujica administration (2010)(2011)(2012)(2013)(2014)(2015) (Lanzaro, 2016;Milanesi & Gadea, 2017). During his second administration, Vázquez (2015-2020) reduced the power of different ministries by creating secretaries under the umbrella of the presidency (Milanesi & Gadea, 2017). That is, those areas which were strategic to the presidency were transferred to structures that worked closer to the presidential office than structures that could be led by distant political allies.

Executive agencies
In parliamentary systems, the prime minister often is the chair of the cabinet, thus being responsible for the center of government. This is not starkly different in standard presidential or semi-presidential systems. Since the president is responsible for appointing the cabinet, ministers end up being responsive to the head of government anyway. Yet, regardless of regime type, there are a few Latin American countries where the core administration of the CoG is constitutionally delegated to another actor instead of the head of government. Guyana, Suriname, and Cuba exemplify this model. Guyana has a system of government that is significantly different than most countries in the region. It resembles the Swiss parliamentary system, in that the president has a fixed term, yet the position is filled indirectly by the party with the largest representation in the legislature. The president appoints the cabinet, including the prime minister. In systems as such, the center of government tends to be placed in the office of the prime minister. Yet, as defined in article 106 of the Constitution, the Cabinet -the central body of government intersectoral decision-making -is composed of the president, the prime minister, the vice-presidents, and other ministers. This body, coordinated by the president, is tasked with "aid[ing] and advis [ing] the President in the general direction and control of the Government of Guyana and shall be collectively responsible therefor to Parliament" (Constitution of Guyana, article 106, paragraph 2).
In Suriname, following the Constitution of 1987, the vice-president chairs the cabinet (Raad van Ministers) (article 116, paragraph 1). This is the highest body of government (article 119, paragraph 1) and is tasked with policy design, implementation, and oversight, as well as horizontal (within the national government) and vertical (among sub-units) coordination (article 122). Thus, whereas all ministers are responsive to the president (article 123, paragraph 2), the coordination of the CoG lies in the office of the vice president, which includes many policy advisers and theme experts.
As in the case of Suriname, it is the first vice-president of Cuba who is assigned to chair the CoG's collegiate organization (Decree 358/2018, article 4). The chair responds to the president, as well as to the Consejo de Estado, which is composed of the president, the secretary of the legislative house, and the vice-president (Constitution of 2019, article 120).
Despite whom the chair of CoG is, countries often have at least three types of executive agencies to mitigate inter-ministerial coordination problems. I refer to them as inner circle, independent, and auxiliary. The first category is aligned with studies that discuss the presidential center (Bonvecchi & Scartascini, 2014;Mejía-Guinand et al., 2018). These organizations differ from those which I call independent since they do not have their portfolio. In the inner circle category, the central agencies are located within the presidential office (or the office of the prime minister). The last type, auxiliary, are those structures that are created within or above certain agencies, being complementary to the main executive bodies.
There are diverse examples to represent the first ideal type: Argentina, Colombia, the Dominican Republic, El Salvador, Honduras, Nicaragua, Panama, Peru, and Puerto Rico. Even though the main executive agency of these countries' CoG is located in the inner circle of the presidency, they are not structured in the same fashion. While some of these have a single agency, others have a fragmented institutional setting. Consider Nicaragua. Whereas the Secretaria de la Presidencia holds the main tasks of the CoG, the presidential center also includes 10 other secretaries with responsibilities that could be attributable to the CoG, like planning, internal communication, and multi-sectoral coordination.
The case of Costa Rica exemplifies independent executive agencies. As earlier mentioned, the Ministerio de Planificación Nacional y Política (Mideplan) was dismembered from the presidential center in 1974. Besides additional political power, this structure also gained the role of coordinating certain public policies, especially when an inter-ministerial effort was required.
In some regards, the Mideplan has overlapping attributions with another independent portfolio: the Ministerio de Hacienda. One of their mechanisms of control is the Lineamiento de Planificación y Presupuestración, which is an instrument to link to the long-term Plan Nacional de Desarrollo to the budget of the different ministries. Based on a collegiate body created in 2011 (the Comisión Hacienda-Mideplan), these ministries were responsible for developing the annual Informe de Avance, which is an oversight instrument to monitor the advancement of the different policies being implemented by the national government. This project is directly influenced by the Inter-American Development Bank through the proposition of a Gestión para Resultados en el Desarrollo.
Brazil has a mix of independent and auxiliary agencies. In 2019, Law 13,844 enacted the formulation of a new ministerial composition of the country, considerably reducing the number of federal departments. It reinforced the role of the independent Ministério da Casa Civil as the center of government. Yet, it also tasked the auxiliary Secretaria Geral da Presidência da República and the Secretaria de Governo with complementary tasks, especially when it comes to the main projects of the federal administration. While the former is also responsible for the planning of strategic policies (an overlap with Ministério da Casa Civil), the latter is responsible for coordination with different levels of the federation as well as non-public sector players. In some cases, this overlap is intentional as it helps the president to distribute responsibilities to actors that may have a greater capacity for specific tasks or closer political or personal ties .

Ideal Type Explanation
Inner circle Located within the chief executive's portfolio, responding directly to the president or prime minister.

Independent
Responsive to the chief executive yet having their portfolio; most often enjoying ministerial status.

Auxiliary
Hybrid organizations responsible for specific policy areas and the coordination of selected agencies only.
Source: Elaborated by the author. Now, consider the auxiliary agencies of Venezuela and Honduras. Below the autocratic leader Nicolás Maduro, the Consejo de Ministros is a council composed of the president, a vice-president and a secretary general, a public prosecutor, as well as seven Vicepresidentes Sectoriales. Each of these vice-presidents coordinates a set of ministries which are organized based on the strategic priorities of the central government. They are (1) Economía, (2) Planificación, (3) Socialismo en lo Social y Territorial, (4) Soberanía Política, Seguridad y Paz, (5) Comunicación y Cultura, and (6) Obras Públicas y Servicios.
The same structure is used in Honduras. The Ley General de la Administración Pública (Decree 146/1986) allows the creation of Gabinetes Sectoriales, which are interagency portfolios in which different secretaries work together in pursuit of broader themes (article 15). In 2014, the presidency established seven Gabinetes Sectoriales through the Decree PCM-001. To illustrate, the Gabinete Sectorial de Desarrollo e Inclusión Social, is composed of 10 different structures. It include secretaries, national institutes, public funds, and policy programs.
Most of these agencies share at least some of the "six Ps" identified by Barber (2016) as essential elements to allow chief executives to govern. That is,

Collegiate bodies
The main structure in the CoG of parliamentary systems is the cabinet. This comprises a collegiate body where all ministers have a seat and discuss the matters of government. Yet, this is not exclusive to these regime types. Most presidential countries also have their council of ministers. These differ from other auxiliary collegiate bodies that are composed of specific ministers or technocrats to put forward specific policy priorities. I differentiate these two types of collegiate bodies based on the labels cabinet and auxiliary council.
Belize, for instance, became independent from the United Kingdom in 1981. Following the inherited Westminster model, the prime minister acts as the chair of the Cabinet. All ministers but the Ministers of State have a guaranteed seat in this collegiate body. According to the Constitution of 1981, the latter are junior ministers who may be invited to the Cabinet, but who do not necessarily hold a seat in the center of government. This is like Dominica, whose independence from the UK occurred in 1978.
In presidential systems, the chair of cabinet may be the president (e.g., El Salvador), his vice president (e.g., Cuba), or an executive agency (e.g., the Ministerio de la Presidencia in Bolivia). The case of El Salvador exemplifies a common trend in the region. The president is the chair of the Consejo de Ministros, but the coordination is supported by two inner circle executive agencies: the Secretaría para Asuntos Legislativos y Jurídicos and the Secretaría de la Presidencia.
In Suriname, where the vice-president heads the cabinet (Raad van Ministers), two types of auxiliary councils support the center of government. They are the mid-level committees (midden commissies) and the subcommittee (onderraden). The former is composed of specific ministers to discuss general inter-ministerial policies, while, in the latter, technocratic personnel discusses elaborate policy operationalizations.
In Peru, the auxiliary councils are hosted by the Secretaría de Coordinación -the inner circle executive agency responsible for inter-ministerial coordination. It oversees a collegiate body composed of vice-ministers, besides other auxiliary councils that facilitate multi-sectoral policies. There are the policy-centered commissions Comisiónes Multisectoriales and policy instruments, e.g., the Sistema Informático de Comisiones Multisectoriales. This digital tool allows mapping and monitoring the policies that are relevant to each different sectors of government.
Another type of auxiliary council is the consejo nacional. This is a broader organization that includes state and non-state actors interested in a specific policy priority of the government. In Panama, there are three national councils hosted by the presidential center: the Consejo Nacional para el Desarrollo Sostenible, the Consejo de Concertación Nacional para el Desarrollo, and the Consejo de Seguridad Nacional. This is like the Ecuadorian structure, where there are five national councils related to different aspects of equality -a core priority of the government.

Ideal Type Explanation
Cabinet Located in the cabinet to coordinate all ministries; often called the Council of Ministers.
Auxiliary Council Hybrid organizations responsible for specific policy areas and the coordination of selected agencies only.
Source: Elaborated by the author.
To a certain extent, these solutions reflect one of M. Andrews, Pritchett and Woolcock's (2013) propositions to improve state capacity and facilitate policy implementation. According to these authors, "change primarily takes root when it involves broad sets of agents engaged together in designing and implementing locally relevant solutions to locally nominated and prioritized problems" (M. Andrews et al., 2013, p. 240). That is, the interaction of the different players engaged in a certain solution may create bridges that will not only contribute to the current implementation process but also the development of other solutions.

Policy tools
The last dimension of CoG in Latin America encompasses policy tools. These are instruments developed by central agencies to map the activities of government (mapping), monitor and improve the performance of implementation processes (monitoring), and prevent and punish those responsible for corruption and other types of wrongdoings (oversight). In the previous subsections, I described some of these instruments. Here, I illustrate this three-category typology of coordination policy tools.
Two of these mechanisms are identified in Jamaica. Within the Office of the Cabinet, the Cabinet Support and Policy is responsible for issues related to policy design and implementation. One of its subunits is the Policy Analysis and Review Unit, which provides policy advice to the different ministries within the cabinet. This also includes the effort of mapping the different national policies that are in place. Periodically, the central government publishes at its website a report called the Government of Jamaica Policy Register, where all ongoing policies are listed.
In terms of monitoring, the Performance Management and Evaluation Branch (PMEB) is the main structure within the Office of the Cabinet. One of its tools is the Performance Monitoring and Evaluation System. Annually, every ministry is expected to submit its Operational Plan to the Office of the Cabinet, the Ministry of Finance, and the Ministry of Public Service. This document includes information on the major initiatives of the portfolio, its intended results, performance measurements tools, targets, and expenditures per quarter. These data are then used by the PMEB to assess the performance of each ministry and by the Ministry of Finance to identify the budgetary needs of each sector of government. This is complemented by the three-year Whole-of-Government Business Plan, which includes the overall goals of government in terms of functional and fiscal issues.
Trinidad and Tobago have developed a complementary tool to improve performance in implementation processes. The Ministry of Public Administration and Digital Transformation is dedicated to the organizational matters of government, ranging from public service training to structural management. The latter is guided by the Public Management Consulting division, which provides advice to the different government agencies when they request support to increase efficiency. Interestingly, this is not a proactive body as in other countries, performing its tasks only when invited by different organizations.

Ideal Type Explanation
Mapping Identify the activities of government, improve internal communication, and avoid redundancy.
Monitoring Control the quality of government action and whether this complies with the chief executive goals.
Oversight Reduce corruption and other types of wrongdoings.
Source: Elaborated by the author.
The creation of monitoring organizations is commonly seen as a result of new public management reforms (Fleischmann, 2019;Paes de Paula, 2005). However, there is a new trend in the region: guided by the international community, many countries strengthened their oversight institutions (Sharman, 2008). In Honduras, these tasks are coordinated by two inner circle auxiliary agencies: the Dirección Presidencial de Gestión por Resultados (DPGR) and the Dirección Presidencial de Transparencia y Modernización del Estado (DPTME). DPGR is the monitoring agency of the national government. It oversees whether the structures within government are pursuing the goals defined by the Dirección Presidencial de Planificación Estratégica. This national plan includes the Planificación Estratégica Nacional y Sectorial (as seen in the Marco de la Visión de País) and the Plan de Nación (embedded in the Plan Estratégico Anual and the Plan Estratégico Plurianual). DPTME is responsible for oversight. Its functioning is like the Cuban Sistema Nacional de Auditoria. Created in 2001, this system hosts the different auditing subunits responsible for controlling public portfolios, firms, and civil society organizations.

DISCUSSION
The literature highlights the role of the growth of government in the strengthening of central agencies (Hamburger et al., 2011;Hamburger & Weller, 2012;Yeend, 1979) formerly a powerful coordinating agency, is the most obvious marker of the change. The PSB's departure left the Secretary of PM&C with a role that is now often described as 'head of the public service' . More broadly, the 1987 changes to the machinery of government both formalised and enabled a sea-change in PM&C's role. Before 1987 a large policy initiation and development project would usually have been considered as beyond PM&C's scope. Since then, extensive and direct policy development work by PM&C has become common. The continuing debates have been over whether PM&C actually delivers in these roles (an empirical question. This occurred especially as a response to the events that drastically changed the world economy in the first half of the 20 th century (Hayek, 1997;Higgs, 2013).
In Latin America and the Caribbean, this is only part of the story. As I propose based on the Historical Path dimension of the typology, indeed, the structuralist movement led by ECLAC introduced central planning agencies in several countries in the region. Yet, the formation of CoG institutions was also motivated by an initial push toward territorial control and government efficiency in the post-independence period in the 19 th century. Beyond, there has also been the influence exercised by international organizations, especially the IDB, for better coordination structures.
This reinforces the relevance of a broader conceptualization of which institutions should be included in the CoG, as some creative solutions were specifically adopted to respond to challenges faced by the region in different historical contexts. This is in also line with the studies on policy diffusion guided by foreign actors (Sharman, 2008) and part of the story behind the development of the Australian CoG (Savoie, 1999), as discussed earlier in this paper. That is, while it was the UK that functioned as a motivating factor for the development of the CoG in Australia, international organizations as the IDB and the OECD have been successful at convincing countries in LAC to adopt this type of solution to coordination problems.
This dimension also contributes to debates on state capacity (Cowen, 2009). The CoG works as a feasible strategy for new countries aiming to establish and strengthen control over their territory and government action. This is in line with those authors who, as Huntington (2006), suggest that order is a central element of development. Overall, it appears that these institutions are focused either on territorial control (post-independence), economic growth (post-structuralist), and efficient policy implementation (globalization).
As discussed, CoGs did not necessarily emerge through a single path. Most likely, different historical stages influenced how institutions evolved, centralized powers, and were fragmented. If such incremental changes are likely to exist in the long run (Cohen, March & Olsen, 1972), the coalitional nature of LAC politics also forces heads of government to reform central agencies in the short term. As highlighted through the Change over Time dimension, these are reflected in the shift of powers between collegiate bodies and executive agencies, or between independent agencies and those hosted in the inner circle of the presidency. The proposed typology shows that the study of the center of government is relevant not only to those interested in public administration but also to the students of political coalitions. In brief, the CoG is a dynamic and plural institution (or set of institutions) that requires continuous analysis over time.
Furthermore, it is a cue for the challenges and priorities of elected officials. An assessment of the distribution of powers made by the head of government allows identifying the most trusted appointees and agencies, as well as the main projects that the administration seeks to implement. If that is correct, the students of public policy should consider investigating the projects assigned to central agencies as a cue for political priority.
Another relevant element in the study of CoG in LAC is the plurality of institutions that may be used by elected officials -a trend that was already expected, following the initial discussion on the different concepts often used in this field of study. They range from executive agencies to collegiate bodies, some of them within the presidential portfolio, others independent, or even hosted as auxiliary subunits of certain structures. This offers a broader perspective than those central agencies that are often located in or closely related to parliamentary cabinets. Furthermore, this plurality shows that these organizations are not exclusive to a single regime type. Similar models of CoG institutions were identified in parliamentary and presidential systems, regardless of their level of democracy.
One possible gap in the present study is an assessment of the pros and cons of each type of central agency. That is, when do executive agencies outperform collegiate bodies? Or, is it better to have them in the presidential center or as independent portfolios? Most probably, the accurate answer is 'It depends. ' The specific literature assessing the cases of Brazil Cavalcante & Batista, 2019;Fleischmann, 2019), Chile (Baraona & Herrera, 2011;Palacios & Jofré, 2016), and Uruguay (Lanzaro, 2016;Milanesi & Gadea, 2017) shows -as highlighted before -how the needs of different presidents shape the design of these agencies. Yet, it is still unclear how successful each strategy was.
Finally, there are multiple Policy Tools developed to exercise the functions of the CoG. This is explored in the analysis of the last dimension. Some of these tools are structured as institutions, while others are instruments of governance. They can mostly be divided into three categories: mapping, monitoring, and oversight. The two last ideal types are especially relevant in the regional context. The literature on New Public Management reforms emphasizes the pursuit of outcome-based performance management in public administration -a similar approach to that in the private sector (R. Andrews & Walle, 2013). Yet, this generates the need for better control given the higher degree of discretion that it involves (Peters & Pierre, 2000;Peters & Savoie, 1996). In LAC, as well as in other regions, this means the development of monitoring tools.
These tools are ultimately connected to the betterment of policy implementation, the delivery of public goods and services, and corruption prevention. Even though these policy tools are often created with different goals, the task of monitoring becomes oversight when it is guided toward the identification of wrongdoings. Recent corruption scandals in the region are the result of the empowerment of such organizations that are not necessarily seen as central agencies (Lagunes & Svejnar, 2020). They have the power to increase compliance with existing legislation, improving government output, and, finally, maximizing the transparency and accountability of the public administration.
These goals and instruments reflect an international trend in policy diffusion (Sharman, 2008) and the need for political control that emerged with growing governments (Higgs, 2013) and decentralized policymaking (R. Andrews & Walle, 2013). Again, the next step in the study of policy tools regards assessing their efficiency. Indeed, this recommendation is valid for all other structures -as already affirmed for the case of executive agencies versus collegiate bodies. This dichotomy reflects only a small portion of the multiple types of CoG organizations in the region, thus, reaffirming the need for a broad concept that captures the creativity of LAC governments in dealing with coordination problems. Moving forward from this paper, which contains a description and classification of CoG across the region, future efforts should be dedicated to assessing the quality of coordination policies in the plural contexts of LAC.