Decentralization and local government in Bolivia : an overview from the bottom up

This paper examines the remarkable case of Bolivia to explore decentralization's effects on government responsiveness and poverty-orientation. The author first summarizes econometric results on the effects of decentralization nationally, and then turns to qualitative research - the focus of the paper - that digs deep into local government processes to understand how decentralization did this. In Bolivia, decentralization made government more responsive by re-directing public investment to areas of greatest need, for example, investment in education and water and sanitation rose after 1994 where illiteracy rates were higher and water and sewerage connection rates lower, respectively. Investment shifted from economic production and infrastructure to social services and human capital formation, and resources were rebalanced in favour of poorer districts. The author explains these results as the aggregate of discrete local institutional and political dynamics. A conceptual model is presented which construes local government as the nexus of two political markets and one organizational dynamic, where votes, money, influence and information are freely exchanged. In order for local government to be effective, these three relationships must counterbalance each other and none dominate the other. Such a stable tension leads to a self-limiting dynamic where pressures from various interest groups are contained within the bounds of political competition. Breaking this tension can hobble government, leaving it undemocratic, insensitive to economic conditions, or uninformed and unaccountable.

In an appendix detailed econometric models on responsiveness to local need are presented for education, water and sanitation, and watershed management.

The model for water and sanitation finds that investment rises under decentralization where more people have no sewerage. It also rises where the share of population without access to drinking water increases, though this finding is sensitive to specification and drops out when other variables are included in the model. Thus local governments invest more where need is greatest. Participative planning methodologies are significant and negative, thus decreasing investment, and the private sector and civil institutions are both insignificant. The latter is surprising given the positive effect of civil institutions on investment in education.

The model for watershed management (related to water and sanitation but broader in scope as it includes reservoirs and wastewater treatment lagoons, levees, and storm drainage works) finds that investment is lowest where the share of population with no access to water is highest. The model also finds that investment rises as more people have access to public and private standpipes, and falls again as internal plumbing becomes widespread. Investment is also highest where few people have access to sewerage, or access to rudimentary sewerage, and decreases as municipal sewerage systems become widespread. These results point to investment that increases in need at intermediate and high levels of provision. But at the lowest levels of provision, local government fails to respond to need and central government is superior. This would make sense if initial investments in water were so great (e.g. from developing water sources, laying water mains and building treatment plants) that local governments cannot undertake them alone, but once these initial investments are made the marginal costs of extending the system are manageable. Perhaps surprisingly, civic and institutional variables appear to have no effect on investment - only variables of need matter.