Beyond decentralization:
intergovernmental relations in Brazil and South Africa
1
Under what conditions can central governments control or reverse processes of
decentralization? Why do some central governments actually push for certain
decentralizing moves, when it would seem this concedes power to subnational
governments? In this paper I explore developments in intergovernmental relations in
Brazil and South Africa since 1994. I conclude that processes of decentralization over
the last decade have not always enhanced, but have sometimes actually reduced the
autonomy of subnational governments. In both cases, central governments have
increasingly worked through bureaucratic channels to earmark the money that flows to
the states (estados) and provinces. Central governments in both countries have also
recently placed restrictions on the ability of states and provinces to borrow money. The
comparison shows that similar changes in intergovernmental relations can occur across
countries with different histories and institutions. In limiting or reducing subnational
autonomy, strong presidential leverage within governing coalitions was crucial; this
leverage can be wielded either in the legislative arena, or within the executive. At the
same time, variations across the cases point to an additional causal factor affecting the
outcomes in these countries: economic crisis. The Brazilian president’s successful
reckoning with a generalized economic crisis gave him leverage that South Africa’s
president did not have, and this enabled the more dynamic set of (re)centralizing changes.
A proposed measure: subnational fiscal autonomy
Political studies of federalism and decentralization have advanced considerably in recent
years, with a shift from positive economic models and assessments of administration to
closer analyses of the underpinnings of intergovernmental politics.
2
Further
improvements can be made, however, by looking not only at fiscal decentralization, but a
wider range of outcomes in intergovernmental fiscal relations (IGFR).
One important step will be to avoid potential problems arising from selection on
the dependent variable. Studies of decentralization may examine only processes where
the central government concedes or devolves authority to subnational governments.
3
Of
course, many studies are attentive to variations in decentralization outcomes by
examining cases where decentralization does not occur, or where the extent or timing of
1
Acknowledgements: I would like to thank several institutions and individuals for support
during the research phase of this project: The Center for International Studies, the Center for
Regional Studies, the Program in Latin American Studies at Princeton University, and the
Woodrow Wilson International Center for Scholars. Special thanks are also due to Peter Spink at
the Fundação Getúlio Vargas (São Paulo) and Annette Seegers at the University of Cape Town
for their assistance. I also owe thanks to Kent Eaton, Jeffrey Herbst, and Deborah Yashar for
numerous conversations, and to Eduardo Gómez, Conor Healy, and Evan Lieberman for
insightful comments on this and related work; any errors are mine, of course.
2
Classic references in federalism and decentralization include Oates 1972 and Cheema and
Rondinelli 1983, inter alia.
3
I would like to thank Marta Arretche for bringing this point my attention.
1
1
Convention