Editor’s note: This story is one in a series, #PolicyForward, that spotlights how faculty, students and alumni at the Harris School of Public Policy are driving impact for the next generation. Leading up to the May 3 grand opening of Harris' new home at the Keller Center, these stories will examine three of the most critical issues facing our world: strengthening democracy, fighting poverty and inequality, and confronting the global energy challenge.

Fiona Burlig uses machine learning methods to study environmental economics and energy.

Can you talk a little about the cost of energy efficiency upgrades and its role in climate change mitigation?

If you look at the way in which countries have committed to greenhouse gas reduction around the world, about 70 percent of the promised reductions in almost every country’s plan are supposed to come through energy efficiency upgrades. This is because, in theory, we all have access to new technologies that increase the efficiency with which we use energy so we can get the same outputs with fewer inputs. There have been many engineering studies supporting this idea and suggesting that, while achieving important climate benefits, consumers also benefit from savings as the energy efficiency upgrades pay for themselves with things like reduced utility bills or increased miles per gallon in cars. From an economist’s perspective, however, we question why we are not seeing more energy efficiency technologies being adopted if these energy efficiency upgrades are actually negative costs. One prominent hypothesis is that these engineering models are over-promising savings.

Can you talk a little bit about how your research explored that hypothesis?

My team worked in the California school system because the state legislature passed a bill that ended up funneling a billion dollars toward energy efficiency upgrades in their schools. The timing presented us with a perfect opportunity to go in and use these schools as our lab. As you can imagine, however, measuring the effectiveness of upgrades is super challenging for a host of reasons. But one thing we were able to in this project was to build on machine learning techniques from the computer science literature to measure hourly data consumption at all the schools. The challenge here was that we then had millions of potential models that we could construct to try and explain energy consumption and it’s very hard to think about the right way is to choose among those models. But we were able to use decision learning architecture to do that, which was very exciting.

What were your findings?

We found that the energy efficiency upgrades were under-delivering to some extent. Our main estimate found that average energy efficiency upgrades provided between 70-78 percent of expected savings. This may not be as bad as what some other studies have found—which is closer to 50 percent of expected savings—but it still resulted in a glass-half-full scenario. When you look 50 years down the road at what we are trying to achieve in terms of reduced GHG, we now have to consider that we are only getting three-quarters of the expected savings from efficiency upgrades. This has serious implications for global investments and the overall focus of climate plans and strategies in jurisdictions around the world.

You’ve done a lot of work around electricity use in developing countries like India. Can you talk a little bit about your research?

Yes. The story here is that around a billion people around the world still lack access to modern electricity and, up until recently, 400 million of those people are in India. We have a lot evidence to suggest strong correlation in the global cross-section between energy use and economic growth, so wealthy countries tend to use more electricity than poor countries. As a result, we’ve seen billions, if not trillions, of dollars around the world being poured into policies to extend energy access to rural areas and pursue these last-mile-grid connections where we bring electricity to people who don’t currently have it. My team wanted to understand what that grid access actually achieved and was it worth the enormous cost to get it.

What did you find?

Let’s start with an amazing statistic. When India achieved its independence, 1,000 villages had access to electricity; now, all of the country’s 597,464 villages are connected to the electrical grid. This is the result of a large series of investments by various governments. Our project was able to take advantage of one of the more recent investments designed to bring electricity infrastructure to approximately four hundred thousand villages with the expressed intention of enabling people to run business and pursue commercial activity off that electricity. Using satellite data, my team was able to see, from space, that villages using the program did have an increase energy access and consumption. What we were interested in looking at is what affect this increased energy access had on well-being. So we gathered a bunch of different metrics, ranging from things like how long kids stayed in school to use of commercial goods like phones and bikes to access to bus lines to impact on job sectors, and more. What we found is that, across the board, electrification led to a precisely measured zero effect on those outcomes. We found no evidence—and actually strong evidence against—this electrification program as having been widely transformative in the way it was promised.

What are the implications of these findings?

Well, when we were able to go and visit some of these villages in person we found that people on the ground corroborated our empirical evidence that electrification wasn’t transforming their opportunities to earn income or economic prosperity at all. We did find, however, that they did like having electricity and it generally made them happier. So there are two main implications of these results for policy makers. One is that the best available evidence suggests that these last-mile rural electrification programs aren’t delivering on the quantitative economic indicators they promised. But we did find that people were happier with electricity so we can’t say these investments are not completely worthwhile. What we can do now, though, is hand policy makers information they can use to weigh the trade-offs between spending for electrification—which is significant—versus other investments that may benefit society more.

About Fiona Burlig

Fiona Burlig is an Assistant Professor at the Harris School of Public Policy. She studies energy and environmental economics, with a focus on the developing world. Her recent research examines the impacts of rural electrification in India, uses machine learning methods to quantify the effectiveness of energy efficiency upgrades, and proposes tools for designing randomized controlled trials. Prior to joining Harris, Fiona was a Postdoctoral Scholar in the Department of Economics and Energy Policy Institute (EPIC) at the University of Chicago. She holds a PhD in agricultural and resource economics from the University of California, Berkeley, and a BA in economics, political science, and German from Williams College.

Read more #PolicyForward stories that spotlight how faculty, staff, students and alumni at the Harris School of Public Policy are driving impact for the next generation.