Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Michael K. Price

Second Advisor

James C. Cox

Third Advisor

Garth Heutel

Fourth Advisor

John A. List


This dissertation consists of three essays that evaluate the effectiveness of interventions promoting residential energy conservation. I study three programs that were implemented as natural field experiments by energy utilities in the United States. In the experiments, households were randomly assigned to a control group that was not exposed to the program or to treatment groups that participated in the program(s). The main objective of this research agenda is to generate insights that help design effective energy policy.

In the first chapter, I study a program that provided weekly emails containing information about a customer's natural gas expenditure to more than 150,000 households in California. I find that exposure to the emails caused significant reductions in natural gas use. Additional analyses suggest that households changed their heating behavior in response to the program. Furthermore, reductions in usage persisted over a 20-month treatment period.

The second chapter focuses on the interaction between a voluntary financial rewards program for reductions in energy use and the Home Energy Report (HER), a repeated mailer that compares a recipient's consumption to that of similar households. We test whether the introduction of financial incentives increased energy savings by households that also received the HER. Our findings show that the two programs complemented each other in several important ways and promoted larger energy savings by a broader set of customers than the HER alone.

Finally, the third chapter investigates whether social nudges, such as the HER, are subject to crowd out effects. Crowd out occurs when the effectiveness of a new social nudge is negatively affected by previous exposure to an existing social nudge. We use two distinct social nudges in an experiment in which households are exposed to no nudge, one nudge in isolation, or both nudges in combination. To determine the extent of crowd out, we compare the effect of the combined intervention to the summation of the effects of each nudge in isolation. We do not find evidence of crowd out. In contrast, the combined effect was marginally larger than the sum of the individual effects, although the difference is not statistically significant.