Can Big Push Infrastructure Unlock Development? Evidence from Ethiopia (Job Market paper)

Roads are instrumental to market access. Electricity is a key technology for modern production. Both have been widely studied in isolation as potential drivers of development. In reality, infrastructure investments are commonly bundled. How such big push infrastructure investments interact in causing economic development, however, is not well understood. To this end, I first develop a spatial general equilibrium model to understand how big push infrastructure investments may differ from isolated investments. Second, I track the large-scale road and electricity network expansions in Ethiopia over the last two decades and present causal reduced-form evidence confirming markedly different patterns: access to an all-weather road alone increases services employment, at the expense of manufacturing. In contrast, additionally electrified locations see large reversals in the manufacturing employment shares. Third, I leverage the model to structurally estimate the implied welfare effects of big push infrastructure investments. I find welfare in Ethiopia increased by at least 11% compared to no investments, while isolated counterfactual road (electrification) investments would have increased welfare by only 2% (0.7%). Both structural and reduced-form findings highlight that interactions of infrastructure investments are in fact material to economic development. Such interactions may be especially crucial to understand the contested effects of electrification.

The Heterogeneous Effects of Rural Electrification: Evidence from Zambia
(joint with T. Figueiredo Walter)

Electrification is crucial for economic development. However, where andhow it may have transformative effects remains unclear. In light of recent micro-economic evidence on electrification’s disappointing effects, we study the potential sources of heterogeneity underlying these results in the context of Zambia, where a large, country-wide rural electrification program is being rolled out. We develop a novel measure of electrification with high temporal and spatial coverage derived from administrative records on the universe of Zambian primary schools and health centres. We can confirm, first, that hundreds of localities have gained access to the electric grid since 2006. Second, our best estimate of a local average treatment effect that exploits inconsequentially electrified localities is indeed virtually zero across a wide range of outcomes. Third, exploiting two full rounds of the Zambian Population and Housing Census, we document a new, key determinant of heterogeneity: any beneficial effects of electrification are concentrated on localities that have at least one non-residential building in commercial use before electrification, even conditional on locality size. In line with classic theories of agglomeration, we conclude that indivisible investments, such as pre-existing productive building capacity, can determine if rural electrification will cause economic development.

Seasonality and Development
(joint with T. Figueiredo Walter)

Complementarities in Public Infrastructure Provision: Evidence from School Electrification in Zambia
(joint with T. Figueiredo Walter)

Distracted by Dreamliners: Noise Pollution and Pupil Performance under the Heathrow Flightpath

This work provides the first causal estimate of the adverse productivity effects of noise pollution emitted by airplanes flying over densely populated areas. Combining high frequency geo-located data of each incoming plane to London Heathrow international airport between 2012 and 2016 with geo-identified, individual-level National Pupil Database test scores, I make use of plausibly exogenous variation in the level of noise pollution experienced by pupils while they sat universal, high-stakes English ‘key stage’ exams in schools on the ground. Scheduled runway alternations generate large and sudden variation in the number and tact of overflying planes. Therefore, individual schools’ exposure to noise pollution while pupils sit centrally scheduled exams varies from hour to hour on exam days, with nearby schools experiencing widely different exposure levels.