Energy Efficiency

Lecture 1

Read Allcott & Greenstone (JEP 2012).
Come to class prepared to discuss the following questions:
1. What is the energy efficiency gap?
2. What are some “rational” explanations?

Do the problems from this old quiz. Ignore question 6.
- You do not need to submit to Canvas, but please complete this before class on Thursday! We are going to go over the answers in class, and you will get a lot more out of it if you attempt them on your own first.

We present a simple model of the adoption of energy saving capital. Basic idea is that a rational individual wants to minimize the total costs of energy services.
We then discuss what an energy efficiency “gap” looks like in this model. Know why energy efficient technology adoption might be lower than socially optimal, and even lower than what is privately optimal.
Understand how to use the different MB curves: social, private and decision to compute outcomes and welfare.

Do this homework problem on the welfare effects of a mandate for energy efficient appliances. Submit your handwritten answers to Canvas.

Discuss another possible reason people might not want to adopt energy efficient technologies: efficient goods are somehow less desirable in non-energy attributes (e.g., aesthetics, performance).
Introduce the concept of nudges and libertarian paternalism.
Discuss different policy instruments, including mandates and subsidies.

Do this homework problem, which is a compuational exercise mirroring the home energy audit example discussed in class.
- This problem set gives you a numerical example of the issue of “selection” bias that we discussed in class. It uses a fake dataset where we know the “truth” to illustrate how selection bias can lead to incorrect estimates of energy savings when individuals opt in to energy policy.
- To do the problem, download the data here (this is a .csv file).
- You can do this exercise in Stata, R, Python, or even Excel if you prefer. I will post solutions in R and Stata if you want to check your work.

We often don’t know how beneficial some policy intervention is (ie how much money people will save if they adopt energy efficient measures). Policymakers must then attempt to measure the benefits of these interventions using data.
This often leads to selection bias, where the individuals who choose to adopt energy efficient measures may not be representative of the general population.
Understand the simple framework comparing outcomes for different groups of individuals who do or don’t adopt energy efficient measures under their own volition.
Randomized assignment of individuals to treatment and control groups can help mitigate selection bias and provide more accurate estimates of policy impacts. But experiements can only estimate the causal effect for the portion of the population induced to change behavior.

In our simple model of energy efficient appliance adoption, we assume that we know how much energy each option uses. In reality, this is often not clear.
Recently, there was a very public debate of the benefits of weatherization. Some prominent economists conducted a study which called into question the wisdom of a popular energy efficiency program.

Read this summary of their critique here, and a blog post from one of the authors here
Then read the published article on the study here (Fowlie, Greenstone and Wolfram QJE 2018)
- Read the intro, try to understand the experiment setup, and read the results.
(Optional) You can read a rebuttal in the New York Times here