Research

Job Market Paper

The Price of Power: Costs of Political Corruption in Indian Electricity

Political capture of public electricity provision may benefit targeted consumers through informal subsidies. However, this causes leakages in utility revenues, inhibiting their ability to reliably supply electricity to the broader consumer base. Using a close-election regression discontinuity design, and a confidential dataset on the universe of geo-coded electricity bills from a large state in India, I show that billed electricity consumption is lower for constituencies of the winning party after an election. However, actual consumption, as measured by satellite nighttime lights, is higher for these regions. I find new evidence to explain this discrepancy -- politicians illicitly subsidize their constituents by systematically allowing the manipulation of electricity bills. To address this corruption through policy, it is important to measure the size of the welfare losses, and compute demand elasticities. I develop a method to estimate elasticities in the presence of data manipulation by leveraging exogenous variation from policy-led price changes and predictive analytics. The net deadweight loss I estimate is large enough to power 3.7 million rural households over an electoral term.

Media mentions: Hindustan Times, Live Mint, World Bank Impact Blog, Ideas for India



Working Papers

Unbundling the Load: How Electricity Restructuring Affects the Manufacturing Sector

The Indian power sector was largely comprised of state-owned, vertically-integrated firms that, for years, accrued massive losses. Between 1998 and 2003, a number of electricity reforms, under the umbrella of the Electricity Act, were phased in, which sought to make the electricity sector more competitive and efficient. I use an event-study framework to study the effect of these reforms on the manufacturing sector, one of the biggest consumers of electricity in India. The reforms affected manufacturing firms via two main channels: the price per unit of electricity and the quality of electricity. Surprisingly, while many expected states with multiple distributors to benefit consumers, I find that instead states that ended up with a single distributor supplied more reliable access to energy. Single distributors found it easier to raise prices, reflecting political realities. In such states, the reforms increased the amount of electricity consumed by manufacturing firms. Consistent with moving to a higher price tier, the average price paid by firms increases, and firms re-optimize their production decisions. Despite a price increase, firms respond to decreased blackouts (greater reliability) by increasing purchased electricity, worker hours and worker productivity. I find corroborating evidence of greater electrification using luminosity data (satellite night lights) that shows an increase in light density in these particular states.



Behavioral Dynamics in Transitions from College to the Workforce

(Under Review) with Catalina Franco, Universidad del Rosario, Bogotá

We show the first estimates of changes in preferences and economic decision-making over major, expected life transitions. We follow Colombian students on the job market and those in lower years across three stages: job search, receiving a job offer, and receiving a paycheck. Using a difference-in-differences setup, we find that job-market students become more altruistic, less impatient and perceive greater liquidity after receiving a job offer, despite this being an expected outcome. These effects dissipate by the paycheck stage, creating "windows of clarity", which may improve long-term decisions about healthcare, pensions or insurance. Furthermore, the results are driven by poorer students.


Works in Progress

Predicting Electricity Consumption with Satellite Nighttime Light Density

Several studies in data-poor contexts use satellite-measured nighttime light density as a proxy for electricity consumption. This project uses high-frequency geo-located confidential data on electricity consumption for over 72 million people in a large state in India to estimate the elasticity between satellite lights data and actual electricity consumption. With these estimates we can more accurately interpret results from satellite data and understand what such estimates imply for daytime electricity consumption, for instance.

Reducing Roof-top Solar Power Supply Delivery Costs in India (with Robyn Meeks)

The Effects of Extreme Weather Events on Electricity Consumption

Does Time-of-day Pricing in Electricity Markets Work?


Other Work

Seeing is Believing: Poverty in the Palestinian Territories, World Bank (2014), with Tara Vishwanath, Brian Blankespoor, Faythe Calandra, Nandini Krishnan and Nobuo Yoshida

Poverty Mapping in the Kyrgyz Republic: Methodology and Key Findings, Working Paper 76690, World Bank (2013), with Nobuo Yoshida and Larisa Praslova

Understanding Seasonal Extreme Poverty in the Northwestern Region of Bangladesh, Vol. 2: The Extent of Seasonal Deprivation and Informal/Formal Coping Mechanisms, World Bank (2012), with Shinya Takamatsu and Nobuo Yoshida