Livio Silva-Muller

This article shows how the legal architecture of state finance enables some policy funds to survive periods of state dismantling. It identifies a pattern of jurisdictional hedging, in which lawyers and bureaucrats insulate policy by intentionally distributing funding rules (how funds are raised), governance rules (how decisions are made), and disbursement rules (how funds are paid) across public–private and national–international jurisdictions. Drawing on a novel environmental finance dataset, 70 in-depth interviews, and archival research, the article demonstrates jurisdictional hedging through a mixed-methods analysis of the Brazilian environmental state since 1990. First, it shows how and why bureaucrats and lawyers opted for different legal configurations across policy funds. Second, leveraging data that differentiate between resource commitment, allocation, and disbursement, it shows that some resource flows survived Bolsonaro’s dismantling attempts while others did not. It ties survival to specific strategies of legal dispersion that make policy more resistant to political intervention. For scholars of democratic backsliding, the anatomy of state finance highlights how dismantling is uneven and constrained when state capacities are dispersed beyond the national state. For sociologists of law and globalization, it shows how legal coding and relational design can shape public policy with sometimes protective effects.

While elite surveys have a long tradition in the social sciences, scholars have questioned their feasibility and analytical value, citing problems of sampling, access, and the limits of structured questionnaires. Drawing on metadata from over 300 elite samples across more than 60 countries, as well as our own mixed-method research with political, economic, and social elites in Brazil and South Africa, we reassess three recurrent critiques: that probabilistic sampling is ill-suited for elites, that elites are uniformly hard to survey, and that standardized instruments fail to capture elite attitudes. We show that probabilistic sampling can produce high-quality elite samples and defensible population-level estimates, and that structured questionnaires do not merely constrain expression but discipline elite respondents by predefining analytical categories and limiting rhetorical maneuvering. By compelling elites to position themselves within a shared classificatory space, surveys make variation and contradiction within elite worldviews empirically visible. We conclude that methodological choices in elite research should be guided less by assumptions about elites as a special population worthy of special methods than by clearly specified research questions, inferential aims, and substantive interests.