Three Essays on U.S. Electricity Policy

There continues to be an ongoing debate about the best design of electricity markets in the United States and this debate is complicated by important discussions about how the population should interact with those markets. At the same time our energy grid is moving towards larger, multi-state systems, technology improvements have allowed power production and storage to be provided at the household level. The interactions between generators, the population, the states, and the markets will grow in complexity, making it all the more important that we understand which policy mechanisms optimally achieve desired outcomes. This dissertation examines some of the facets of U.S. energy policy across each of these four domains.

In Chapter 1, I focus on the strategic behavior of generators participating in U.S. wholesale electricity auctions. Specifically, I examine one particular market clearing mechanism, the Uniform Price Auction (UPA), which has seen near ubiquitous adoption as the price setting rule for U.S. real-time markets. Theoretical literature on these types of auctions predicts that participants will engage in bid shading and that this can lead to inefficient auction outcomes. Yet, this equilibrium outcome seems to be overlooked in much of the literature studying power markets. The convention that has been adopted, instead, is to assume generators are offering their supply at marginal costs. Any deviations from marginal cost bidding behavior has been explained away as a consequence of a congested grid, this congestion fracturing the total market into smaller oligopolies and rendering market power to the participants. I study the real-time electricity market in the Electric Reliability Council of Texas (ERCOT) from May through July 2021, narrowing my lens to only those periods where the grid is uncongested and at its most competitive levels. In observing the bids of suppliers and using three different methods to estimate marginal costs, I find widespread rent seeking behavior among generators and empirical evidence of inefficient auction outcomes. Under my most conservative marginal cost estimates, I predict the bid shading - inherent to UPAs - creates at least $21 million of annual rents to generators compared to an efficient auction mechanism. Furthermore, this research determines that the UPA could as much as double the cost to serve load and, all else equal, increase CO2 emissions to meet demand by between 1.5% and 5%. Despite these issues, the UPA may still be the best of all available market clearing options. It is easy to understand, easy to implement, and deploys a larger fleet of generators in any given time period. If the grid reliability products in ERCOT are not fully functional, then the UPA induces a more reliable grid than an efficient auction could support.

In Chapter 2, I explore the distributional impacts created by rooftop solar adoption. There is remarkable, well-documented heterogeneity in pollution damages avoided by rooftop solar capacity in the United States. Furthermore, there is concern that the \textit{costs} of solar policy are regressive. Yet, there is no research that examines the characteristics of households receiving the environmental benefits, even as a growing environmental justice and distributional equity literature highlights its importance. Using data on the locations of nearly 1.5 million rooftop solar panels, demographic data from the U.S. Census Bureau's American Community Survey, and an empirical model that estimates the avoided damages of rooftop solar generation, I examine the distributional effects of rooftop solar adoption. I find that households of color receive greater per capita benefits on average, perhaps alleviating some of the environmental justice concerns. On the other hand, benefits are increasing in income, suggesting that the distribution of benefits from currently installed systems is not sufficient to overcome the regressivity of costs. Moreover, I show that there exists essentially no efficiency-equity tradeoff in this context: capacity allocations that maximize total environmental benefits are nearly identical to allocations that maximize benefits received by people of color and low-income households. Thus, existing solar capacity not only foregoes up to $1 billion annually in total environmental benefits but also fails to achieve substantial improvements in distributional outcomes.

Finally, in Chapter 3, I consider the relationships between U.S. states and organized electricity markets. Using a simulation tool, I investigate the effects created by a state defecting from the wholesale electricity market in PJM, an organized electric grid in the eastern United States, on the states that remain in the coalition. I find, generally, that if a net-importing state defects from the wholesale energy market, the remaining states’ producers are worse off and the remaining states’ consumers are better off. The opposite effect takes hold if the defecting state is a net-exporter. Furthermore, I find evidence that defection impacts the remaining states’ climate initiatives. The effectiveness of electric vehicle and solar photovoltaic policies are conditional on the number and characteristics of defecting states. My simulations suggest that, for state legislatures pursuing these climate goals, the best strategy to adopt is to pass laws that are both geographically targeted and adaptable to shifting consortiums.