Weak Foundations: Street-Level Infrastructure and Abandoned Policy Implementation in Brazil

Abstract

States physically transform communities--for better or for worse--through investments in street-level infrastructure: public buildings and other facilities, including libraries, clinics, schools, and community centers that occupy physical, visible spaces within neighborhoods and that serve as loci for state-society interactions. This dissertation seeks to explain unevenness both in the pursuit of infrastructure development as well as the outcomes of those efforts. Conditioning on existing capital stock, why do some governments make investments in street-level infrastructure while others do not? And what factors distinguish successful cases of implementation from unsuccessful ones? In line with research that ties political competition in developing contexts to normatively undesirable outcomes like clientelism, corruption, and inefficient policymaking, the findings presented in this dissertation indicate that competition has perverse effects on the delivery of high-visibility public goods with long time horizons. I expand upon previous scholarship by emphasizing the noninstrumental nature of this problem. Specifically, I argue that political competition increases the perceived cost of delaying credit claiming, leading vulnerable incumbents to accelerate policy planning to the detriment of its quality. This intertemporal trade-off is riskiest when a policy's objectives are irreducible and when its intermediary outputs are irreversible; in these cases, policies become "locked-in" after plans are finalized, with deviations accommodated only at high cost. My work further highlights the importance of autonomous bureaucracies and technically skilled bureaucrats at both the local and federal levels to plan policies and limit political discretion over the pace of policy implementation. The empirical foundation for these arguments comes from fixed effects regressions performed on a time series cross-section panel of Brazilian municipalities spanning three mayoral administrations, and survival analyses of primary health care clinics built or renovated with grants from the Ministry of Health's Requalifica UBS program. The proposed mechanisms are evaluated against alternative explanations with qualitative evidence collected over eleven months of fieldwork in Brazil.