Empirical Methods
Slides [link]Key Points
- Lecture 1: Translating policy problems into empirical questions
- Lecture 2: Correlation vs causation
- Correlated explanatory variables and omitted variables
- Lecture 3: Experiments
- Example: Time varying electricity prices.
- Lectures 4/5: Natural experiments
- Theory
- Difference-in-Differences and panel regression
- Example: Nuclear plant privatization
Required reading and response questions
Lecture 1
Required reading
- Read Borenstein, Severin, and James Bushnell. 2015. “The US Electricity Industry After 20 Years of Restructuring.” Annual Review of Economics 7 (1): 437–63. link. Alternative BC link here.
- Focus on section 3, “ELECTRICITY MARKET PERFORMANCE SINCE RESTRUCTURING BEGAN” (you can skim the rest of the article).
Response questions:
- The authors look for empirical evidence that electricity deregulation reduced costs. Describe their statistical approach to answering this question (in your own words).
- Name one complication with this analysis described by the authors.
Lecture 2
No required readings. Blog post 1 due.
Optional reading: “Quasi-Experimental and Experimental Approaches to Environmental Economics” by Greenstone and Gayer (ungated version here)
Lecture 3
Required reading
- Read Jessoe and Rapson (AER 2014)
- Optional: There are actually dozens of experiments like this. See this review for a synthesis.
Response questions: Explain the experimental setup in your own words (1-2 sentences each)
- What is the research question?
- Why do the authors need to use an experiment to answer it?
- What are the different experimental groups?
Lecture 4
Required reading
- Read the 2021 Nobel Prize Award Summary
Response questions:
- What is a “natural experiment”?
- How did Card and Krueger estimate the effect of the minimum wage on employment?
Lecture 5
Required reading
- Read “Deregulation, Consolidation, and Efficiency: Evidence from US Nuclear Power,” Lucas W. Davis and Catherine Wolfram (ungated version here)
Response questions:
- The authors are using using “difference-in-differences” here. What are the two differences? Explain the groups being compared in your own words.
- Do you find this empirical approach believable? Why or why not?
- Optional: “Environmental Health Risks and Housing Values: Evidence from 1600 Toxic Plant Openings and Closings”, by Currie et al (ungated version here)