Research

This paper provides causal estimates on how inland oil spills, one major type of environmental disaster, affect local labor markets. By exploiting severe inland petroleum oil spills and their news coverage status, I find that spills negatively affect county-level labor markets, but only when a spill is in the news. In the five years after a severe inland spill that receives news coverage, employment, wages, the number of establishments, and the labor force all decrease significantly. Severe inland spills not reported in the news yield no such effects. Exposure to spill information induces composition changes in county-level gross migration, weakening labor market conditions in low-tradability industries, such as retail and food services. Information on environmental disasters triggers sorting, which has distributional effects and degrades labor markets in counties that experienced spills with news coverage. Back-of-the-envelope calculations suggest that compared to the control group, counties with spills that receive news coverage lost 407,000 jobs, the monetary value of which is equivalent to \$41.4B, and \$27.1B in foregone wages in aggregate during the post period.

Exploiting county-level variation in exposure to severe inland oil spills and their news coverage status, I estimate that oil spills raise ambient air pollution levels and mortality rates, but only when a spill is not reported in the news. The increases in mortality rates are caused by the elevated air pollution and are concentrated in the most susceptible group: elderly adults. When a spill is covered in the news, there are not only no changes in ambient air quality but also persistent decreases in county-level mortality rates. By exploring heterogeneous effects, I show that the decreases in mortality rates are due to out-migration. The differential effects on air pollution and mortality imply that information on environmental disasters is beneficial to the environment and human health.

This paper evaluates the impacts of expanding green jobs on non-green employment, energy production and consumption, and CO₂ emissions. By employing a shift-share instrumental variable approach, I find that one additional green job generates about four non-green jobs at the state level in the United States. Most of the increase comes from low-tradability industries. The cleaner an industry in terms of air pollution emissions, the larger the multiplier effect of green jobs. Furthermore, having more green jobs also significantly boosts the production and consumption of renewable energy, and lowers CO₂ emissions.

Collective Choice May Tell Nothing About Anyone’s Individual Preferences

Muye Chen (first author), Michel Regenwetter, and Clintin P. Davis-Stober.
Decision Analysis. 2021.

Beyond Theory and Data in Preference Modeling: Bringing Humans into the Loop

Thomas E. Allen, Muye Chen, Judy Goldsmith, Nicholas Mattei, Anna Popova, Michel Regenwetter, Francesca Rossi, Christopher Zwilling.
In: Walsh T. (eds) Algorithmic Decision Theory, New York, NY: Springer. 2015.