Policy Diffusion in Latin America
Studies about public policy diffusion address questions about political change. Among these we can find: Why different governments adopt the same type of policy? Why others do not do it? Why some innovations in policy are scattered and concentrated in a given territory? What explains the speed of a policy diffusion process? (Coêlho 2016). To answer these questions, it is assumed that the information about new policies or institutions is communicated by the members of a social system in an uncoordinated way, and where the first adoptions of an innovation affects the chances of doing so by those who have not applied it (Rogers 2004).
From the 1960s onwards, the contributions of innovation diffusion studies – such as those of Rogers (2004) – began to be applied to the field of public policy. The objective is to explain policy-making processes where several different actors formulate and design similar policies, assuming that this process is not explained exclusively by domestic variables. Therefore, it is aimed at determining the independent variables that explain an adoption (which becomes the dependent variable), as well as the mechanism that operates in this process. A dual view is adopted, which links the domestic level where the policy is applied with the regional/global where it has been disseminated.
The Diffusion Model and Its Main Components
One of the first works on diffusion came from Jack Walker in 1969 (Berry and Berry 2007), whose objective was to develop several propositions that could be used as guides for the diffusion of innovations and applied, also, for the analysis of decision-making (Walker 1969). Its object of study was the innovation capacity in different areas by the federal states of the United States, and its questions were similar to those still carried out in the studies with this model. Among the main results, Walter outlines, indirectly, that some of the mechanisms that influence diffusion are competition and emulation among the states (Walker 1969).
Subsequently, the number of papers about policy diffusion has been increasing, the same as with the sectors studied. These make reference to the diffusion of economic, health, education, retirement, poverty, agriculture, institutions, regulators, and environment policies among others. In general, several models are applied that have been designed to explain the diffusion (Berry and Berry 2007), starting from the innovation theories and taking a step forward when considering how diffusion occurs between different units and which are the variables that explain it. Along with the variety of areas, the diffusion model has been applied at different levels, from the municipal, regional, federal to the national level.
Since the diffusion of policies arise from the theories of diffusion of innovations, some authors work with the definition proposed by Rogers (2004) who points out that “it is the process through which an innovation, defined as an idea perceived as new, spreads via certain communication channels over time among the members of a social system” (Rogers 2004: 13). Others apply proposals such as the one given by David Strang (1991), understanding diffusion as the process in which the adoption of a practice alters the possibility of doing it for those who have not done it (Elkins and Simmons 2005). Including aspects of the above, Levi-Faur defines diffusion “as the process by which the adoption of innovation by member (s) of a social system is communicated through certain channels and over time and triggers mechanisms that increase the probability of its adoption by other members who have not yet adopted it” (Levi-Faur 2005: 23). Or, in other words, understand diffusion as a process whereby the information about new policies or institutions is communicated through certain channels in time among members of a social system in an uncoordinated way, and where the first adoptions of the innovation affect the probabilities of adoption by those who have not adopted yet (Jordana et al. 2009).
All the definitions of diffusion assume that is a process that is not due exclusively to reasons specific to the internal political process of each state, because they consider the possibility that “policy choices in one country affect the policy choices in other countries, so causing policies to converge” (Meseguer and Gilardi 2009: 528). This model then discards the possibility that countries respond independently to similar domestic conditions and that this could explain the coincidence in decision-making. It also seems unlikely that convergence will occur exclusively due to coordination among actors, through international cooperation processes or due to pressures from more powerful agents. In a sense, it is defined as a mixed position in that there is interdependence between countries over time or with a central actor who acts as a transmitter. Countries would not act in a completely independent way, but neither in a coordinated way. The diffusion process answers then to an uncoordinated interdependence (Elkins and Simmons 2005). “Under this conception, governments are independent in the sense that they make their own decisions without cooperation or coercion but interdependent in the sense that they factor in the choices of other governments” (Elkins and Simmons 2005: 35). Therefore, the studies on diffusion seek to rule out the null hypothesis that assures that national socioeconomical and political factors are the exclusive cause of change in policies.
The diffusion process can be characterized by at least three components; first, they develop in waves, starting slowly in terms of the numbers of actors adopting the initiative, increasing significantly to then decrease. This implies that, in graphic terms, the adoption rate over time will take the form of an “S” curve (Weyland 2006). Secondly, the diffusion usually presents a significant geographic convergence, in the same region or neighboring countries (Weyland 2006). Finally, generally literature assumes that a third element in diffusion is the similarity between the central characteristics of the policies but with divergence in other components (Weyland 2006).
Finally, in terms of the object of diffusion, some researchers have suggested distinguishing whether it is a model or a principle (Weyland 2006). The first refers to a diffusion of a specific policy or program that is replicated, such as the Chilean model of privatization of pensions or the Grameen Bank microcredit system. The second refers to the diffusion of a principle, a guideline that leads the decisions towards specific polices, such as central bank autonomy or the universal access to primary health services (Weyland 2006). Therefore, a new route guide is disseminated, a key idea, but it can be imbued with institutional characteristics and diverse components. The result of this diffusion is not a pattern of change as profound and uniform as in the case of the diffusion of a model (Weyland 2006).
The Mechanism of Diffusion
Diffusion studies also seek to understand the objective behind the adoption of a given policy and also determine how that process worked. The mechanisms of diffusion allow us to elucidate the causal relation of that decision, and if, for example, it is due to a desire for competition among states, if it seeks to emulate decisions that confer some status in the international arena, or was the product of a learning process that enabled identification of such policy as the most relevant and functional (Elkins and Simmons 2005). The literature also argues that the type of mechanism could have consequences in the quality of the policies. For example, a policy adopted by emulation may not be the most appropriate alternative to solve a specific problem (Meseguer and Gilardi 2008).
Simmons et al. (2008) distinguish four mechanisms of diffusion – coercion, emulation, competition, and learning – on which there is relative consensus, although the denominations and certain components may partially vary.
The coercion mechanism refers to the imposition of the policies of the national governments by international organizations or of more powerful and influential countries. Therefore, the convergence of policies is promoted by the dominant actors and it is a mechanism that operates vertically.
The other three mechanisms operate horizontally among the different actors and would have voluntary character. Emulations occur when “the policies are adopted by their social value, instead of its own value, looking for loyalty and status” (Osorio 2015: 34). It differs from the other mechanisms because it is based on the logic of appropriateness, where the action of adopting a policy implies the evocation of an identity or role in a specific situation rather than in the rational analysis of the domestic variables to be considered in the decision-making process (Gilardi 2012). On the other hand, the policymakers are not interested in effective policy solutions, rather in facilitating the public acceptance or viability of a policy change. So emulation, in terms of results, may not lead to an improvement or even lead to a deterioration in a given situation. It is due to the above that it is argued that emulation usually implies copying of the policy (Bender et al. 2014).
Third, there is a competition “when countries that dispute the same resources adopt the policy of their competitors through fear of economic loss” (Osorio 2015: 34). It is usually assumed that it refers exclusively to the economic sphere; however, it is possible to identify political and social competition. The first occurs when the states try to adopt the role of international leaders or early followers of a policy in order to influence in the evolution of the policy and minimize the adaptation costs. On the other hand, the second denotes a relationship of two entities towards a third one. In order to stabilize that relationship these two entities observe each other and, if necessary, they imitate each other (Bender et al. 2014).
Finally, in the mechanism of learning there is a process of knowledge of the policy in question, and as a consequence of this arises the decision for adopting it (Simmons et al. 2008). “Policies adopted in some countries turn out to be natural experiments from which others can learn” (Meseguer 2008: 323). Now, literature has made an important distinction between full or rational and limited learning. In its first version, learning improves knowledge since it implies the review of the causal links between policies and the results in the light of the observed experience and contributes to the understanding of the consequences of the policies (Meseguer 2005). Therefore, this definition contributes to distinguishing between the rational learning and the emulation or the symbolic imitation (Weyland 2004); since the latter seeks credibility, the status, or the simple conformity with the international tendencies (Meseguer 2005). On the other hand, limited learning resorts to heuristic shortcuts. “Rather than scanning all information, governments look at relevant information” (Meseguer 2005: 72).
The debate about the mechanisms is constant and alternatives to the aforementioned emerge. For example, Simmons has regrouped the mechanisms distinguishing between those where the adoption by one actor alters the practice of the other (these are the conditions by which he makes his decision) and where the adoption by some gives information by which the decision in made (for example, about the advantages and disadvantages) (Simmons and Zachary 2005). On the other hand, Meseguer and Gilardi (2008, 2009) exclude coercion from the list of possible mechanisms. The verticality and the pressure exerted would condition the voluntariness in the decision.
The Critical View of the Policy Diffusion Models and Their Mechanisms
As already noted, the policy diffusion model provides a perspective to understand the convergence among countries that adopt a policy, helping to identify the mechanism(s) that explain this process. However, some aspects in literature that can be discussed to enrich its application can be observed.
One of them is the overlap that exists between the different mechanisms of policy diffusion. The boundaries between the definition of one mechanism or another are diffuse and tend to have aspects in common that make it difficult to distinguish among them. This is a key point to consider during the interpretation and empirical testing (Meseguer and Gilardi 2009). For example, in the case of emulation, the main distinction with learning has to do with the motives, since the adoption of the policy would be due to the aspiration to imitate a country with high prestige or adopt a policy that is in fashion. Now, in spite of this, it is possible that there is in the motivation an expectation about the results of the policy that provokes certain expected consequences (Meseguer 2004); because otherwise, the decision would be completely irrational. Hence, a difficulty has to do with being able to ensure that the motivation was exclusively to emulate a particular policy. Likewise, a mechanism can be classified as learning if deep knowledge of the initiative is identified before its adoption. However, it is not possible to rule out completely that the motivation to emulate has not been present. Therefore, the idea of a 100% pure mechanism does not capture the complexity of the process and in mythological terms, it is hard to apply.
The other mechanisms, competition and coercion, are easier to identify and apply. In the case of the former, it is understood that the policies adopted would directly or indirectly affect other countries in terms of economic aspects, so the motivation would be not to be at a disadvantage. Finally, in reference to coercion, the pressure exerted by an international institution or by an influential country is possible to identify, for example, through the conditionality of a loan. However, more subtle ways of coercion could overlap with learning or emulation, which requires the definition of precise indicators.
In that approach, a tension present in the mechanisms is that they involve – although not always defined properly – two components: the means and the motivations (Osorio 2015). The attention given to one or the other varies in the diffusion studies and also according to the research methods applied. The identification of the motives and means of each one will contribute to a more precise definition of its characteristics and their indicators for their empirical observation (Osorio 2015).
Also, it is necessary to consider that different mechanisms can operate during a diffusion process and, in addition, a variety of them can be observed depending on the countries involved or according to the sector where the diffusion takes place (Jordana et al. 2009). This should avoid a homogenizing assumption, since in a process of diffusion among a group of countries, it may be the case that some adopt new policies because they learn from others, and that there are countries that do it by simple emulation or by competition (Meseguer and Gilardi 2009). In this sense, the consideration of the domestic aspects of each country opens the alternatives to observe variables of political type, such as the proximity of an election or the impact of specific institutional spheres (Meseguer and Gilardi 2009). Finally, it is a matter of continuing to clarify how national politics condition the diffusion.
Specifically regarding learning, one of the challenges is to move forward in identifying the components of the process that influence the final outcome. In that sense, what is the object of study could be distinguished; that is to say what is learned. Here Weyland’s already commented distinction between the adoptions of the model or principle (Weyland 2006) is useful. Then it must also be considered that learning a principle or a model could also be the result of this mechanism and the decision made in its development; that is, it could be a consequence of how it was learned. Another aspect that must continue to be worked on is to define who the learners are. This has not been approached in detail until now by the literature and its inclusion is assumed in the same group as policymakers, technocrats, and politicians. The difference regarding their motivations and attributions could be important, differing, for example, between substantial concerns about the nature of the policy or the obtaining of electoral revenues.
Finally, when analyzing this mechanism at the level of policymakers, the question whether the learning process has been individual or collective arises. As already pointed out, the learning process of a policy can begin at an individual level and eventually could be collective learning, but not necessarily at the moment of decision-making for adopting the initiative. However, this is a point to take into account, since it has to do directly with the identification of who learns at the moment of adopting a policy.
Studies About Cases of Diffusion in Latin America
Studies in Latin America have addressed global and regional policy diffusion processes using a variety of methodologies. Among the research are the studies carried out by Jacint Jordana, who analyzes the creation and reform period of regulatory agencies in 19 countries of the region, questioning the traditional approach and considering sectorial areas, as well as the national variations in the diffusion of regulatory authorities in Latin America (Jordana and Levi-Faur 2005). In this sense, they propose two approaches to explore the institutional transfer channels: the national approach and the sectorial policy approach (Jordana et al. 2011). This contribution considers how the transfer channels vary during the stages of the diffusion process and, at the same time, complicates the approach in the mechanisms of the transfer, considering that they could operate not only among states but also between policies sectors.
Sarah Brooks (2009) conducted an extensive study to understand the diffusion wave of Pension Fund Administrators (PFA). Her work Social Protection and the Market in Latin America (2009) analyzes the wave of reforms in the social welfare sphere, this is, the shift from the delivery system to the individual capitalization one. The study tries to answer the arguments that prevail at the time to privatize the system, despite the consequences that this brings and the strong opposition that is generated from the different actors.
On the other hand, it is worth mentioning the numerous studies carried out by Kurt Weyland (including from 2002, 2004, 2005, 2006 and 2014) that contribute significantly to the research about the diffusion processes, as well as the understanding of waves of diffusion of specific policies in Latin America. In Learning from Foreign Models in Latin America Policy Reform (2004), a series of articles are collected that address the process of diffusion of reforms in the security and welfare sector, unemployment insurance, and the health sphere. In Bounded Rationality and Policy Diffusion: Social Sector Reform in Latin America (2006), Weyland develops a psychological approach to decision-making, deepening and developing some of its guidelines and analyzing cases compared qualitatively. His research found that a clearly delimited form of rationality prevails in the transnational diffusion of political innovations.
Others works continue analizing specific social policies in the region. For instance, in The diffusion of the Conditional Cash Transfer programs (CCT) in Latin America 1990–2010 (Osorio 2014) seeks to resolve which variables explain this diffusion process and the mechanisms of diffusion that are observed in the processes of adoption at a national level. The hypothesis developed is that diffusion is explained by the existence of a consensus about the CCT generated by a regional epistemic community that is also linked with international organizations presents in Latin America. It ratifies the small role of these domestic variables as the degree of inequity and develops the role of the regional epistemic community, identifying the experts, their beliefs about the programs, their ways of influence and linking with the national processes of policy formulation.
The diffusion of CCT at a global level is studied by Brooks (2015) through a quantitative analysis in Social Protection for the Poorest: the adoption of antipoverty Cash Transfer Programs in the Global South (2015). The findings point out that there are more possibilities for implementing the programs with the presence of the CCT between countries of the same region (the so-called neighbor effect in the literature) and among democratic countries. In the domestic arena, democratization increases the probability of adopting, along with the existence of governments with ideological differences between executive and legislative branches (Brooks 2015).
Then, at national level, Natasha Borges Sugiyama, in Diffusion of Good Government: Social sector reforms in Brazil (2013), examines the policies behind the programs of good government and researches the forces behind its dissemination. Specifically, it focuses on local Brazilian governments and the diffusion of two social programs: Family Grant and Family Health Program (Programa Saúde da Familia) at a municipal level. The results show that the electoral incentive offers a weak explanation for the diffusion of innovative social policies (Sugiyama 2013). Therefore, the alternative explanation refers to ideology and the socialized norms at a professional level.
The work of Denilson Coêlho (2016) focuses on how the Family Health Program initiative became a model to be followed by other federal states, starting with a diffusion process inside Brazil. It focuses its attention on the adoption process of this CCT factor of a political, institutional nature and how the regional influence conditions the adoption of a program.
De Faria, in Ideas, Knowledge and Public Policies. A succinct inventory of recent major analytical strands (2003), studies the importance of ideas and knowledge in the public policy process. Along with it, in the work The diffusion of social policies as a strategy of international insertion: Brazil and Venezuela compared (2012) comparatively analyzes the way that Lula’s Brazil (2003–2010) and the Chavez’s Venezuela (1999–2013) tried to provide the processes of international social policy diffusion with tools that allowed them to obtain benefits in their strategies of international insertion, taking into account that both countries had the intention to exercise leadership in a regional context.
Finally, Porto de Oliveira (2016) studies the process of international diffusion of participatory financing policies. This policy was conceived in Porto Alegre in 1989 and now has innumerable iterations at international level, defining the importance of a group of actors – participation ambassadors – in the process of advocating for the definition of participatory financing policies in the global agenda.
Studies regarding the diffusion of public policies constitute a consolidated analysis perspective that contributes to understanding the diffusion processes of policies that may occur within a country or at regional or global level. At domestic level, for example, it is considered that variables such as the type of political regime, institutions and their characteristics, the gross domestic product, and government ideologies could explain the decision as to whether to adopt a particular policy or not. At an external level, the presence of the so-called neighbor effect, the link to loans of the multilateral banks, the incidence of economic crises, the expert influence of epistemic communities, and advocacy coalitions or international organizations are usually assumed.
These variables are tested through diverse methodologies. There is some consensus about the advantages of complementing quantitative analyzes at macro level – either state, regional, or global – that allows us to test the variables and deepen their incidence. Then, the detailed look at how these variables affect and operate in the decision-making process should be studied by case or comparative case analysis, which will enable us to advance with the questions concerning how and what is adopted.
- Bender K, Keller S, Willing H (2014) The role of international policy transfer and diffusion for policy change in social protection – a review of the state of the art. IZNE Social Protection Working Paper 14/1, International Policy Learning and Policy Change: Scientific Inputs for the Dialogue on Social Protection with Global PartnersGoogle Scholar
- Berry FS, Berry W (2007) Innovation and diffusion models in policy research. In: Sabatier PA (ed) Theories of the policy process. Westview, Boulder, pp 223–260Google Scholar
- Brooks S (2009) Social protection and the market in Latin America: the transformation of social security institutions. Cambridge University Press, New YorkGoogle Scholar
- De Faria C (2003) Idéias, conhecimento e políticas públicas. Um inventário sucinto das principais vertentes analíticas recentes. in Revista Brasileira de Ciências Sociais 18(51):21–29Google Scholar
- De Faria C (2012) A difusão de políticas sociais como estratégia de inserção internacional: Brasil e Venezuela comparado. Interseções 14(2):335–371Google Scholar
- Gilardi F (2012) Transnational diffusion: norms, ideas, and policies. In: Carlsnaes W, Risse T, Simmons B (eds) Handbook of international relations. SAGE Publications, Thousand Oaks, pp 453–477Google Scholar
- Jordana J, Levi-Faur D, Fernández i Marín X (2009) The global diffusion of regulatory agencies: channels of transfer and stages of diffusion. IBEI Working Papers 2009/28Google Scholar
- Meseguer C, Gilardi F (2008) Reflexiones sobre el debate de la difusión de políticas. en Política y Gobierno Vol. XV(2):315–351Google Scholar
- Osorio C (2014) La difusión de los Programas de Transferencia Condicionada en América Latina 1990–2010, Doctorado en Ciencias Políticas y Sociales. Departamento de Ciencias Políticas y Sociales, Universitat Pompeu Fabra, BarcelonaGoogle Scholar
- Osorio C (2015) ¿Cómo se difunden y adoptan los Programas de Transferencia Condicionada en América Latina? Propuesta de operacionalización de los mecanismos de difusión aplicado al caso chileno. In FLACSO – Ecuador (ed) Revista Íconos. Número 53. pp 31–48. ISSN: 1390-1249Google Scholar
- Porto de Oliveira O (2016) International policy diffusion and participatory budgeting: Ambassadors of participation, international institutions and transnational networks. Palgrave McMillanGoogle Scholar
- Simmons B, Dobbins F, Garrett G (eds) (2008) The global diffusion of markets and democracy. Cambridge University Press, Cambridge, MAGoogle Scholar
- Sugiyama NB (2013) Diffusion of good government. University of Notre Dame Press, Notre DameGoogle Scholar
- Weyland K (2002) The politics of market reform in fragile democracies. Princeton University Press, PrincetonGoogle Scholar
- Weyland K (ed) (2004) Learning from foreign models in Latin American policy reform. Woodrow Wilson Center Press/Johns Hopkins University Press, Washington, DC/BaltimoreGoogle Scholar
- Weyland K (2006) Bounded rationality and policy diffusion: social sector reform in Latin America. Princeton University Press, PrincetonGoogle Scholar