Low quality infrastructure is a major barrier to economic advancement in developing countries. This paper develops an empirical framework to explain the persistence of this problem as the result of a targeted program of utility subsidies. I estimate a structural model of household demand for electricity, using customer billing data from Colombia matched to household characteristics and network outage data. I then predict the change in consumption from upgrading households with low quality connections. Combining this with cost and regulatory data, I calculate the change in the utility firm’s profits from the upgrade. I demonstrate that the existing program of targeted subsidies in Colombia deters investments to modernize infrastructure in areas with unreliable electricity supply. Households in these areas receive low quality service for which they do not pay, firms receive transfers from the government to tolerate areas with non-payment, and the government provides these transfers to prevent mass disconnections of non-payers. Based on the model estimates, I analyze less costly subsidy programs that provide stronger incentives for firms to upgrade neighborhoods with low quality connections. This paper closes with a discussion of how these results can be applied to the design of upgrade and subsidy programs in other countries.
