By Samantha
Within international policy circles, a big push is underway to expand social protection in developing countries. A Global Partnership for Universal Social Protection was launched at the most recent UN General Assembly by ILO and the World Bank, and social protection targets are included in the Sustainable Development Goals. Central to the efforts of international organizations is policy diffusion, or the active spreading of social policy reforms between countries. These efforts are assiduously rational, providing evidence on experience, costs and benefits on social protection schemes between countries of the global South. International organizations assume that governments will compare global best practices – and in turn use these to craft social policies that serve domestic contexts and needs.
However, theories on policy diffusion indicate that the process is often anything but rational; a range of factors and mechanisms come into play when policies diffuse between governments. For social protection, the 800 pound gorilla in the room is ideology. Universal social protection, that will make a real dent in reducing poverty and inequality, requires redistribution of income from wealthier groups to the least well-off, and as such is typically associated with left-leaning governments and political forces. The question posed by this precis is to what extent ideology promotes or inhibits the diffusion of universal social protection? Will universal social protection models spread more rapidly between left-leaning governments, or is ideology irrelevant to the spread of innovative social welfare policies?
In his study of Latin American pension reform, Weyland (2005) does not identify ideology as an immediate impediment to policy diffusion. Focusing on Chile’s market-oriented pension privatization, he shows that the model spread fairly rapidly across Latin America countries, including to those with substantively different political orientations. Moreover, it was more influential than pension innovations in Europe that were potentially more attractive to left-leaning governments. Weyland attributes this to cognitive heuristics, which is based on ‘bounded rationality’. Rather than processing information in a thorough and balanced way, humans resort to shortcuts, which are subject to distortions and biases. One bias is availability, with a salient innovation in a neighboring country attracting disproportionate attention. Another bias is representativeness, in which decision-makers jump the gun in assessing innovations, giving greater value to early success than long-term evidence. Finally, the anchoring heuristic posits that the adaptation of policy models will be limited, with the core attributes of the model retained, explaining why the same model will diffuse to widely different national contexts.
Assuming that Weyland’s model of policy diffusion is correct then, the introduction of universal social protection will be driven by the most prominent models, which early on demonstrate an ability to solve pressing problems, rather than by government’s ideological stance. In this regards, the global evangelization of Brazil’s Bolsa Familia model, which emphasizes targeted, conditional cash transfers to the most vulnerable families, makes it front-running model. However, for sub-Saharan Africa, the region with the lowest social protection coverage, neighboring early starters like South Africa may be more influential. South Africa’s Child Support Grant provides unconditional but targeted cash transfers to families, and at 3.5 percent of GDP[1] is more redistributive than Bolsa Familia, which is equivalent to just 0.5 percent of Brazil’s GDP.[2] Ultimately, according to Weyland, it is not the level of redistribution and poverty reduction that matters, but which model gains prominence and traction with other countries in the region first.
By contrast, in Ideology and Learning in Policy Diffusion (2004), Grossback, Nicholson-Crotty and Peterson argue that ideology plays a determining role in policy diffusion in the United States. The dominance of the two-party system and division between liberal and conservative ideas means that policymakers assess new policies along this dimension. State policy-makers will more readily adopt policies from ideologically similar states, and disregard information about policies from dissimilar states. In this regards, their theory is also informed by ‘bounded rationality’, with ideological cues the main shortcut employed by decision-makers. Using event-history analyses of data sets from 49 states, they find evidence of a strong relationship between ideology and policy diffusion for lottery adoptions, academic bankruptcy laws and sentencing guidelines.
Applying this model to the diffusion of universal social protection, policy-makers will look for ideological cues from other polities as a way to determine if the policy will pass muster with their elites and electorate. However, an immediate challenge appears in applying Grossback’s model outside of the US; its dependence on a competitive and polarized two-party system, and the ability of states to readily use ideological cues. These conditions are not a given in other countries. Multi-party systems reduce the distinction between liberal and conservative, and in many parts of the world, particularly Asia and Africa, this dimension does not hold sway. In these regions differences between and within countries may be determined more by ethnicity, religion, colonial legacies and authoritarian versus democratic regimes. This may complicate countries abilities to read ideological cues from one another. The underlying dynamic of using these cues as a short-cut for policy uptake may still be relevant, but without replication of the study in different contexts, it is difficult to say if this model applies globally.
Likewise for the model developed by Weyland, the lack of replication for other regions and for policy areas other than pension privatization, limits the strength of his claims. It may be that pension privatization spread so quickly and widely in Latin America, because there were no competing policy options within regional pension reform circles. As such the privatization model was able to dominate reform efforts. For other sectors, such as education, which have multiple policy options readily available, ideological similarity could be a more prominent variable in policy diffusion. In the current social protection environment, there are a number of policy options to choose from, and ideology could serve as one cue among others.
Turning back to the initial questions, there is no clear answer. When ideological differences between countries are clear, then it could potentially be used as a cue for the suitability of redistributive universal social protection measures. This need not be the ideological differences between liberal and conservatives, but could instead be cues based on other political differences. However, the evidence also suggestions that the same cognitive short-cuts could also lead to decision-makers to disregard ideological differences, and instead adopt social protection systems based on the most prominent examples in the region. Whichever model gains momentum and is recognized as providing a solution to countries’ pressing problems will likely be reproduced with minimal modifications to different contexts. In this case, ideology and politics will have little influence on the level of redistribution achieved by social protection policies; instead it will be driven by cognitive heuristics of policy diffusion.
[1] ILO (2016) Social protection floor in South Africa, Child Benefits http://www.social-protection.org/gimi/gess/ShowWiki.action?wiki.wikiId=853
[2] Wikipedia (2016) Bolsa Familia https://en.wikipedia.org/wiki/Bolsa_Fam%C3%ADlia
Within international policy circles, a big push is underway to expand social protection in developing countries. A Global Partnership for Universal Social Protection was launched at the most recent UN General Assembly by ILO and the World Bank, and social protection targets are included in the Sustainable Development Goals. Central to the efforts of international organizations is policy diffusion, or the active spreading of social policy reforms between countries. These efforts are assiduously rational, providing evidence on experience, costs and benefits on social protection schemes between countries of the global South. International organizations assume that governments will compare global best practices – and in turn use these to craft social policies that serve domestic contexts and needs.
However, theories on policy diffusion indicate that the process is often anything but rational; a range of factors and mechanisms come into play when policies diffuse between governments. For social protection, the 800 pound gorilla in the room is ideology. Universal social protection, that will make a real dent in reducing poverty and inequality, requires redistribution of income from wealthier groups to the least well-off, and as such is typically associated with left-leaning governments and political forces. The question posed by this precis is to what extent ideology promotes or inhibits the diffusion of universal social protection? Will universal social protection models spread more rapidly between left-leaning governments, or is ideology irrelevant to the spread of innovative social welfare policies?
In his study of Latin American pension reform, Weyland (2005) does not identify ideology as an immediate impediment to policy diffusion. Focusing on Chile’s market-oriented pension privatization, he shows that the model spread fairly rapidly across Latin America countries, including to those with substantively different political orientations. Moreover, it was more influential than pension innovations in Europe that were potentially more attractive to left-leaning governments. Weyland attributes this to cognitive heuristics, which is based on ‘bounded rationality’. Rather than processing information in a thorough and balanced way, humans resort to shortcuts, which are subject to distortions and biases. One bias is availability, with a salient innovation in a neighboring country attracting disproportionate attention. Another bias is representativeness, in which decision-makers jump the gun in assessing innovations, giving greater value to early success than long-term evidence. Finally, the anchoring heuristic posits that the adaptation of policy models will be limited, with the core attributes of the model retained, explaining why the same model will diffuse to widely different national contexts.
Assuming that Weyland’s model of policy diffusion is correct then, the introduction of universal social protection will be driven by the most prominent models, which early on demonstrate an ability to solve pressing problems, rather than by government’s ideological stance. In this regards, the global evangelization of Brazil’s Bolsa Familia model, which emphasizes targeted, conditional cash transfers to the most vulnerable families, makes it front-running model. However, for sub-Saharan Africa, the region with the lowest social protection coverage, neighboring early starters like South Africa may be more influential. South Africa’s Child Support Grant provides unconditional but targeted cash transfers to families, and at 3.5 percent of GDP[1] is more redistributive than Bolsa Familia, which is equivalent to just 0.5 percent of Brazil’s GDP.[2] Ultimately, according to Weyland, it is not the level of redistribution and poverty reduction that matters, but which model gains prominence and traction with other countries in the region first.
By contrast, in Ideology and Learning in Policy Diffusion (2004), Grossback, Nicholson-Crotty and Peterson argue that ideology plays a determining role in policy diffusion in the United States. The dominance of the two-party system and division between liberal and conservative ideas means that policymakers assess new policies along this dimension. State policy-makers will more readily adopt policies from ideologically similar states, and disregard information about policies from dissimilar states. In this regards, their theory is also informed by ‘bounded rationality’, with ideological cues the main shortcut employed by decision-makers. Using event-history analyses of data sets from 49 states, they find evidence of a strong relationship between ideology and policy diffusion for lottery adoptions, academic bankruptcy laws and sentencing guidelines.
Applying this model to the diffusion of universal social protection, policy-makers will look for ideological cues from other polities as a way to determine if the policy will pass muster with their elites and electorate. However, an immediate challenge appears in applying Grossback’s model outside of the US; its dependence on a competitive and polarized two-party system, and the ability of states to readily use ideological cues. These conditions are not a given in other countries. Multi-party systems reduce the distinction between liberal and conservative, and in many parts of the world, particularly Asia and Africa, this dimension does not hold sway. In these regions differences between and within countries may be determined more by ethnicity, religion, colonial legacies and authoritarian versus democratic regimes. This may complicate countries abilities to read ideological cues from one another. The underlying dynamic of using these cues as a short-cut for policy uptake may still be relevant, but without replication of the study in different contexts, it is difficult to say if this model applies globally.
Likewise for the model developed by Weyland, the lack of replication for other regions and for policy areas other than pension privatization, limits the strength of his claims. It may be that pension privatization spread so quickly and widely in Latin America, because there were no competing policy options within regional pension reform circles. As such the privatization model was able to dominate reform efforts. For other sectors, such as education, which have multiple policy options readily available, ideological similarity could be a more prominent variable in policy diffusion. In the current social protection environment, there are a number of policy options to choose from, and ideology could serve as one cue among others.
Turning back to the initial questions, there is no clear answer. When ideological differences between countries are clear, then it could potentially be used as a cue for the suitability of redistributive universal social protection measures. This need not be the ideological differences between liberal and conservatives, but could instead be cues based on other political differences. However, the evidence also suggestions that the same cognitive short-cuts could also lead to decision-makers to disregard ideological differences, and instead adopt social protection systems based on the most prominent examples in the region. Whichever model gains momentum and is recognized as providing a solution to countries’ pressing problems will likely be reproduced with minimal modifications to different contexts. In this case, ideology and politics will have little influence on the level of redistribution achieved by social protection policies; instead it will be driven by cognitive heuristics of policy diffusion.
[1] ILO (2016) Social protection floor in South Africa, Child Benefits http://www.social-protection.org/gimi/gess/ShowWiki.action?wiki.wikiId=853
[2] Wikipedia (2016) Bolsa Familia https://en.wikipedia.org/wiki/Bolsa_Fam%C3%ADlia