Why do similar types of democracies respond differently to health epidemics, such as AIDS? In his book, Evan Lieberman is startled by the fact that although Brazil and South Africa share similarities in levels of political and economic development, they responded differently to AIDS. This is surprising if one considers the recent transition to democracy and proclaimed political elite commitments to human rights and elections in both cases. How and why did this happen?

In explaining this puzzle, Lieberman provides a unique explanation for differences in policy response—measured in terms of bureaucratic development and prevention and treatment policy, as well as human rights. In contrast to other approaches focusing on the impact of democratic institutions, state capacity, and civic mobilization, Lieberman's emphasizes the impact of boundary politics: that is, the informal and formal categorization of ethnic groups that shape policy preferences and the government's perceptions of the wider systemic threat of AIDS. These boundaries also shape civic perceptions and incentives to approach the state for reform. As seen in South Africa, when boundaries are thick, measured in terms of formal measures of ethnicity such as census data, or active pubic discussion such as newspapers and/or legislative debates, policymakers have few incentives to respond aggressively. The risk of AIDS is relegated to isolated pockets of communities behaving in a devious manner. At the same time, thick boundaries in society generate few incentives to approach the state for resources to combat AIDS, as communities fear disclosing their health status due to the stigma of having the disease.

Conversely, when boundaries are weak, as seen in Brazil, AIDS is not described in boundary terms. Politicians view the risk of AIDS as a grave national concern, not to be associated with any particular group. Politicians are thus more likely to aggressively implement policy. Likewise, civic organizations, such as nongovernmental organizations (NGOs), are more likely to work with the state for a timely and effective response.

Lieberman evaluates his theory by using qualitative and quantitative evidence. He combines case study analysis not only of Brazil and South Africa but also of India in order to test for the wider applicability of his theory. This is followed up with multivariate regression analysis, using multiple sources of data, as well as brief case studies supporting his statistical results. Altogether, the book provides a very impressive package of multimethodological techniques.

As with any bold enterprise, there are strengths and weaknesses. My discussion of the strengths of this book focuses on the application of political science theory and diversity in methodology. The book's weaknesses emerge in the unsubstantiated dismissal of rival claims, lack of methodological clarity, and the evidence.

In my view, the most important contribution of this book is the application of political science theory to a new and emerging topic in political science: public health policy. Lieberman applies theories from comparative politics, focusing on ethnic conflict and rent seeking not only to reveal their limitations in explaining AIDS policy but also to proffer an alternative theory: boundary politics. In addition, he devotes an entire chapter to the issue of global health governance in order to reinforce the fact that political science has not theorized enough about the interaction of international and domestic politics for AIDS policy reform. He fills in this lacuna by positing that international policy discourse and organizational coercion influence domestic policy. I wholeheartedly agree with his position that in the field of political science, much more work needs to be done on this topic.

Methodologically, the book is very impressive. Lieberman submits a combination of qualitative and quantitative evidence in a clear and persuasive manner. When discussing Brazil and South Africa, he explains his reason for comparing both nations, while providing qualitative evidence through interviews and myriad reports and news articles. To further test his argument, he follows up with an even more detailed case study of India, drawing on several primary sources. And lastly, a statistical chapter is provided to further test for the generalizability of his claim. In contrast to his case studies, he randomly selects cases for statistical analysis. He then specifies the data that he used to measure the dependent variables and boundary politics, which were extensive and drew from a host of sources, including his own. He also introduced alternative ways of measuring data, such as the usage of content analysis for estimates of budgetary commitments. And finally, Lieberman was careful to clarify the scope of his argument, stating that it only applies to initial government responses, as elites are eventually motivated by rising AIDS cases and international and domestic pressures.

The author does not stop there, however. In a fashion similar to his nested approach to theory development, after discussing his broader statistical findings Lieberman breaks down the analysis by region and discusses the impact of his theory in Africa, Latin America, and Asia. In addition to once again testing his theory by providing limited statistical analysis for each of these regions, he briefly describes how his theory accounts for policy reforms across several different types of political regimes.

Yet there are some limitations to Lieberman's work. The first has to do with his methodological intentions, which are somewhat unclear. For example, what I initially thought to be in-depth case studies testing his theory—Brazil and South Africa—actually appeared to be case illustrations of the efficacy of his theoretical approach, something akin to what Theda Skocpol and Margaret Somers call the parallel demonstration of theory in historical analysis (in “The Use of Comparative History in Macrosocial Analysis,” Comparative Studies in Society and History 22 [1980]: 156–73). He eventually states that the Brazilian and South African cases were not intended to test his theory but that India was (p. 173). But then at one point he refers to the Brazil/South African comparison as an example of a “model-building” analysis … an “inductive, theory driven” comparative analysis, which confuses me about their methodological purpose (p. 173). Were these cases introduced for inductive theory building, or were they used for illustrating the strength of his theory?

We are thus left with only one crucial case study: India. And even then, Lieberman admits that this study did not provide the best test, due to a lack of evidence. As he states: “But additional evidence is needed to support the link from cause to effect, especially because it has been rare for Indian policy-makers or for other analysts to discuss AIDS in terms of ethnic politics” (p. 193).

Next, the author seems to readily dismiss rival theoretical claims without good reason. While he provides evidence showing that state capacity and the presence of NGOs is insufficient for predicting policy responses, it is hard to say the same for political leadership. The dismissal of AIDS leadership fell on President Nelson Mandela's avoidance of the issue in South Africa, Thabo Mbeki's interest in AIDS prior to election, and then his avoidance of it once in office, notwithstanding ongoing political support. In Brazil, Lieberman claims that aggressive policies predated Presidents Fernando Henrique Cardoso's and Luiz Inácio Lula da Silva's political leadership. Yet this is factually incorrect. Prior to Cardoso's and the World Bank's loans in 1993, there was no aggressive AIDS program. Thus, leadership under Cardoso, at the presidential and bureaucratic level, was important for reform.

Finally, there are problems with his evidence. Essentially all of the evidence used to substantiate Lieberman's claims—with the exception of India—is based on secondary sources. Especially with regards to Brazil, there are no interviews with key AIDS officials, NGOs, and civil society, nor are any Brazilian news and scholarly sources cited. The same can be said for South Africa, notwithstanding his discussion of a brief visit to the city of Mtubatuba in 2003.

Despite these limitations, scholars contemplating adopting Lieberman's framework may wish to consider supplementing it with an analysis that incorporates the role of international organizations, pressures, and resources as causal forces working with boundary institutions. Researchers may also wish to consider several in-depth qualitative case studies in order to further test his theory. And finally, comparing middle-income nations to more advanced democracies, such as the United States, may provide a fruitful analysis. Notwithstanding differences in levels of development, AIDS policy in the United States has also been troubled by boundary institutions; public health institutions and policy, especially at the urban level, are relatively weak, while the federal government is often unresponsive to local needs. Moreover, incorporating the United States and other advanced democracies can provide yet another example of the usefulness of Lieberman's framework.

In closing, as a political scientist conducting research on global health policy, I learned a great deal from this book. Boundaries of Contagion provides a good example of the potential that the discipline of political science has for illuminating the challenges of AIDS policy success in developing nations. At the same time, it widens our thinking about the health policy consequences of ethnic divisiveness. This is surely an important contribution to the field of comparative politics and public policy.