by Janet I. Lewis
International donors have heavily invested in decentralization programs in developing countries, to the tune of over 7.4 billion dollars since 1990.1 They did so amidst a global shift in the 1980s and 1990s away from centralized political and economic regimes, and in the wake of widely-held expectations that decentralization reforms would bring about numerous benefits—from consolidating democracy to improving local service delivery and mitigating ethnic conflict. However, recent research shows that decentralization reforms’ record in attaining those goals is mixed, at best.2
In recent research, I argue that decentralization reforms can, in fact, contribute to an unintended recentralization of power. In a project co-authored with Guy Grossman, assistant professor of political science at the University of Pennsylvania, we examine a specific, rarely-noted process through which decentralization reforms may lead to diminished local power relative to the center: the proliferation of local administrative units.3 Since the mid-1990s, following or alongside decentralization reforms that were promoted by international donors, almost half of Sub-Saharan African countries increased their number of sub-national administrative units by over twenty percent. Beyond Africa, several countries that underwent decentralization reforms significantly increased their number of local governments. For example, as part of the post-communist decentralization reform process, Czechoslovakia and Hungary increased their number of municipalities by about fifty percent between 1989 and 1993. Similarly, after devolving power to the district level in 1998 and 1999, in the wake of Suharto’s fall, Indonesia increased its number of districts from 292 to 497 in less than a decade. Vietnam also increased its number of provinces from forty to sixty-four between 1996 and 2003, following liberalization reforms. In Brazil, from the return to civilian rule in the mid-1980s to 2000, the number of municipalities increased over fifty percent.
But even amidst decentralization reforms that are characterized by far-reaching de jure devolution of power, administrative unit proliferation can lead to a de-facto recentralization of power. This is the case because the rapid creation of a large number of new local governments, fragments existing ones into smaller units with lower intergovernmental bargaining power. This contributes to an increased dependence of local governments on the resources and administrative capacity of the central government.
Why does administrative unit proliferation occur? First, the introduction of decentralization reforms that devolve power to localities imbues those localities with new importance and value to their residents, generating demand for them. Second, to explain when and where within a country new administrative units are created, one should consider the incentives of both local political actors and national leaders. We argue that administrative unit proliferation is largely driven by a convergence of interests between citizens in marginalized areas of localities, which seek more direct access to local government resources and who view the creation of new top-tier units as an avenue to such access; elites in marginalized localities who seek job opportunities and greater control of public resources; and the central government, which seeks to implement popular policies and thus attain increased electoral support.
We think of localities that could potentially split as being comprised of two areas: a core area that controls the local government and is home to its headquarters, and a more rural area that lacks control of local resources and where residents and elites are, or perceive themselves to be, politically and economically marginalized. This conceptualization reflects the emergence of small, semi-urban areas in developing states’ peripheries. For example, in 2005, over half of the urban population in Sub-Saharan Africa lived in small cities of less than 200,000 people.4 The creation of a new district in a marginalized area brings it public jobs and resources, such as hospitals, as well as access to previously remote local decision makers. Especially in developing countries, where travel in rural areas is often quite costly due to poor road conditions, citizens prize proximity to public services and decision makers. Elites in such areas also tend to favor the creation of new districts because small districts mean more limited pools of competitors for local office. Demand for new districts is most likely to come about, we argue, in areas that are the most economically, politically, and ethnically marginalized—meaning that they are less well developed, less well-represented politically on influential local committees, and less likely to have the same majority ethnic group as the core area.
While local demand thus plays an important role in generating new localities, ultimately only the central government can formally approve a new local government, thus it is important to consider the central government’s incentives. We argue that a proliferation of local administrative units will be broadly desirable to central government incumbents. This is the case because given the widespread popularity of creating new administrative units in marginalized areas, meeting demands for new units can provide a significant electoral boost. The primary costs to a central government of creating a new local administrative unit will be a budgetary burden—but this is often offset, at least in part, by international donors who often view the creation of new units as a positive step towards decentralization.
Administrative Unit Proliferation and Recentralization in Uganda
To examine the determinants of administrative unit proliferation, we analyze how it unfolded in Uganda, whose decentralization process has been heralded as “one of the most far-reaching local government reform programs in the developing world.”5 Another observer noted that: “Within a very short time, Uganda has achieved one of the most decentralized and stable systems of subnational government in the entire Sub-Saharan region.”6 Alongside Uganda’s extensive decentralization reforms, beginning in the mid-1990s the number of districts increased dramatically: from thirty-four in 1990 to 112 in 2011. The maps that follow show the process of district splitting unfolding over time in Uganda. The maps show Uganda’s county boundaries—counties are a governance unit one tier below districts, and with very few exceptions, new districts that formed in Uganda were comprised of one or more counties, which split away from their “mother” district. The maps show that the creation of new districts occurred steadily since the mid-1990s, and affected the majority of Ugandan counties.
We use multilevel regression analysis to analyze the determinants of district formation. We find that less-developed counties, and those that are underrepresented in the leadership of a district’s committee governing intra-district resource allocation, are more likely to form a new district. This finding is robust to various model specifications, including to the incorporation of spatial dependencies. Additionally, we find that counties in which the largest ethnic group is different than the district’s largest group are more likely to split from that district, suggesting that being a concentrated ethnic minority worsens perceptions of marginalization.
We also test the supply-side of our argument by examining how district splits in the five years leading up to an election affect the vote share received by the incumbent president, Yoweri Museveni. Using a battery of statistical models, we find that the incumbent president receives an electoral bonus of approximately 2.5 to three percent in counties that were elevated to the status of a district prior to an election and is not penalized by “mother” areas that have recently lost territory due to a split. These findings support our claim that the electoral benefits to the central government of providing new districts are significant.
Finally, we find substantial evidence from interviews with local officials, secondary accounts, and national and local government budget data that is consistent with our argument that the proliferation of districts that followed Uganda’s decentralization reforms has contributed to a substantial recentralization of power. Among several other indications of recentralization, we find that the portion of the national budget that is apportioned to districts has declined since the early 2000s, limiting the resources available to districts. Additionally, Uganda’s graduated tax—which was once the districts’ primary source of locally derived revenue—was ended by the central government in 2005—making the districts almost entirely dependent on the central government for budgetary support. The central government has also recently made key local government positions—especially that of the chief administrative officer, who oversees the entire technocratic arm of each district—a centrally appointed position, whereas it was formerly a locally-elected position. As a result, instead of answering to local politicians, the entire technocratic arm of each district answers to the central government. We interpret these changes as evidence of a growing de-facto recentralization of power.
Implications for Research and Policy
What are the policy implications of our findings? They suggest that proponents of decentralization reforms should be more aware of the relationship between decentralization reforms and the pressure to create new administrative units. Furthermore, there are some reasons for concern. Several commentators in Uganda have argued that the creation of new districts generates unnecessarily burdensome administrative costs and destabilizes local interethnic relations. On the other hand, the recentralization of fiscal authority may play a fundamental role in state-building in countries that have suffered from instability.7 More research will be needed in examining the effect of rapid district creation, in particular, its effect on economic development and on the provision of local public goods and social services.
This work also contributes to our broader understanding of the effects of Africa’s political liberalization beginning in the 1990s. Our findings about the center’s electoral incentives to supply districts to marginalized rural localities contributes to a growing body of work that shows that voting in African states is not simply an ethnic census as prior work had suggested. Rather, it provides further support to the idea that African voters respond to policy initiatives, and that elections—even in hybrid regimes like Uganda—generate incentives for national elites to implement highly visible policies that are perceived as redistributive.
- The World Bank, Independent Evaluation Group. 2008. “Decentralization in Client Countries: An Evaluation of World Bank Support, 1990-2007.” Washington, DC: The World Bank.
- See, for example, Eaton, Kent, Kai Kaiser and Paul J. Smoke. 2010. The Political Eonomy of Decentralization Reform: Implications for Aid Effectiveness. Washington, DC: The World Bank.
- Grossman, Guy and Janet I. Lewis. 2013. “When Decentralization Begets Recentralization? The Intergovernmental Politics of Adminstrative Unit Proliferation.” Working Paper.
- Kessides, Christine. 2005. “The Urban Transition in Sub-Saharan Africa: Implications for Economic Growth and Poverty Reduction.” World Bank Africa Region Working Paper Series No. 97. Washington, DC: The World Bank.
- Francis, Paul and Robert James. 2003. “Balancing Rural Poverty Reduction and Citizen Participation: The Contradictions of Uganda’s Decentralization Program.” World Development 31(2):325–337.
- Obwona, Marios, Jesper Steffensen, Svend Trollegaad, Yeko Mwanga, Francis Luwangwa, Ben Twodo, Abel Ojoo and Fred Seguya. 2000. “Fiscal Decentralisation and Sub-National Government Finance in Relation to Infrastructure and Service Provision in Uganda.” Washington, DC: The World Bank.
- Diaz-Cayeros, Alberto. 2006. Federalism, Fiscal Authority, and Centralization in Latin America. New York: Cambridge University Press.
Janet I. Lewis is a postdoctoral fellow in the Department of Government and a former WCFIA Graduate Student Associate from 2007–2011.