UC Berkeley

Energy production is associated with a number of significant environmental externalities. For instance, coal-fired power plants emit both local pollutants (such as particulate matter and sulphur dioxide) and the global pollutant carbon dioxide. This creates a need for government intervention: left to their own devices, energy producers will do more environmental damage than is socially optimal. The choices faced by policy-makers in regulating the energy industry are, however, rarely clear. Government regulators must trade off the externalities caused by different types of energy production. While nuclear power generation does not emit carbon dioxide, there is the risk of significant environmental damage in the event of a nuclear meltdown. While many proponents of biofuels hoped that replacing fossil fuels with biofuels would decrease carbon dioxide emissions, land-use changes associated with biofuels production can cause environmental damage. These trade-offs motivate the chapters of this dissertation.

In the first chapter, I study changes to nuclear power safety following major regulatory changes in electricity markets. Following electricity market restructuring, approximately half of all commercial U.S. nuclear power reactors were sold by price-regulated public utilities to independent power producers. At the time of the sales, some policy-makers raised concerns that these corporations would ignore safety. Others claimed that the sales would bring improved reactor management, with positive effects on safety. Using data on various safety measures and a difference-in-difference estimation strategy, I find that safety improved following ownership transfers and the removal of price regulations. Generation increased, and this does not appear to have come at the cost of public safety.

This paper contributes to several strands of the energy literature. First, it fits in with the literature on electricity deregulation. While this literature has considered a broad set of outcomes, my paper is the first to look closely at safety, an outcome of particular interest for nuclear energy. In line with that, it also contributes to the literature on nuclear safety, which has been of particular interest given accidents at Three Mile Island, Chernobyl, and Fukushima. Finally, my work is germane to the literature on the consequences of deregulation for outcomes beyond private efficiency gains. While there is now some consensus that deregulation can lead to the alignment of private costs and thus to efficiency gains, less is known about the effect on external costs. Papers in this literature are necessarily industry-specific: the interaction of private cost reductions with changes to quality or changes to external costs is highly context-dependent. However, this paper provides intuition for the mechanisms at work, some of which are generalizable beyond the nuclear power industry.

In the second and third chapters, I study land-use changes relating to biofuels production. Transportation in the U.S. accounts for a significant portion of greenhouse gas emissions. Motor gasoline, excluding ethanol, accounts for around 20 percent of U.S. greenhouse gas emissions, or over 1 billion metric tons each year. Biofuels have been promoted as an alternative to petroleum products that bypasses some of the fundamental problems with the oil market: supporters claim that it is renewable (whereas conventional oil is exhaustible), produced in the U.S. (as opposed to regimes in some cases unfriendly to the U.S.), and carbon-friendly. As the acreage devoted to biofuels crop production expands, however, it can compete with cropland used for food or with natural ecosystems.

In the second chapter, joint with Maximilian Auffhammer and Peter Berck, I examine price impacts of biofuels production. The last ten years have seen tremendous expansion in biofuels production, particularly in corn ethanol in the United States, at the same time that commodity prices (e.g., corn) have experienced significant spikes. While supporters claim that biofuels are renewable and carbon-friendly, concerns have been raised about their impacts on land use and food prices. This paper analyzes how U.S. crop prices have responded to shocks in acreage supply; these shocks can be thought of as a shock to the residual supply of corn for food. Using a structural vector auto-regression framework, we examine shocks to a crop's own acreage and to total cropland. This allows us to estimate the effect of dedicating cropland or non-crop farmlands to biofuels feedstock production. A negative shock in own acreage leads to an increase in price for soybeans and corn. Our calculations show that increased corn ethanol production during the boom production year 2006/2007 explains approximately 27 percent of the experienced corn price rise.

In the final chapter, I study land-use change in Brazil arising from biofuels production. Scientists and economists are increasingly worried that biofuels production is leading to deforestation, and hence loss of habitats and increased carbon dioxide emissions. I estimate land use changes in response to shocks in sugarcane (a biofuels feedstock) and soybean (thought to be affected by United States corn ethanol production) prices in Brazil at a national and regional level. Using county-level data from 1973 to 2005, I consider a dynamic panel data model of input demand for land, conditioning on price changes of other commodities. Unlike the existing literature, I apply a dynamic panel data estimator that is unbiased (unlike OLS with fixed effects) and more precise than GMM. The short-run price elasticity of sugarcane acreage in Brazil is estimated to be approximately zero, whereas the elasticity of soybean acreage is 0.9 when both spot and futures prices change. The regional estimates for soybeans show considerable variation, and are highest in areas of ecological importance, such as the cerrado. Sugarcane estimates are more homogeneous. These results should be taken into account in impact assessments of biofuels.