Electricity Markets Price Risk, Pollution, and Policies
Publication or External Link
The electricity sector is a significant contributor to the economic and environmental health of the United States, with annual revenues well over $300 billion and responsibility for approximately one third of all carbon emissions. The last several decades have brought significant changes to economic and environmental policies applicable to the electricity sector, including market restructuring and a variety of air quality improvement policies. This thesis builds on previous research of these issues through three related essays on energy economics and policy.
The first essay examines the local environmental impacts that can be attributed to renewable portfolio standards. Renewable portfolio standards (RPS) have been adopted by the majority of states in the U.S. to encourage electricity from renewable sources. Previous studies omit an analysis of local and regional pollutants, so this paper provides an empirical investigation into pollutant reductions from RPS while accounting for policy heterogeneity across states. Using a nation-wide panel of pollution monitoring stations in conjunction with local and national economic data, I find that adopting RPS results in significant sulfur dioxide reductions and modest nitrogen oxide reductions. I find no evidence of particulate matter reductions. Lastly, the analysis shows that pollution reductions are driven by groups of states whose neighbors also adopt RPS, which is likely because of pollution spillover effects.
The second essay examines the importance of ramping cost to electricity price volatility. High price volatility has plagued electricity market participants for decades and is increasingly important in the context of growing intermittent renewables. Although electricity market price behavior generally has been well studied in the last decade, the literature is sparse when discussing the importance of generator ramping costs to price volatility. This paper contributes to the literature by first formalizing the intuitive link between ramping costs and price volatility in a multi-period competitive equilibrium. The fundamental result of the model shows how price volatility rises with ramping costs. This notion is tested empirically using a pooled event study regression, a two-stage least squares (2SLS) specification, and a generalized autoregressive conditional heteroskedasticity (GARCH) model. The econometric results all confirm that price volatility is significantly decreased by additional natural gas capacity, which has comparatively low ramping costs. This marks the first rigorous study to quantify the pecuniary externalities within the New England market's generating profile. A simulation also explores how annualized volatility changes over time during a shifting generation profile, noting that natural gas generators can offset the volatility increases from increasing wind generation. Lastly, there is no evidence that natural gas capacity additions reduce the forward premium.
The third essay examines price convergence in the wholesale electricity markets in the context of transaction costs on virtual bids. Virtual bidding has been introduced in most restructured electricity markets in the United States with the intent to manage price risk, increase financial liquidity, and minimize deviations between forward prices and spot prices. Previous literature argues that even without virtual bids, generators can attempt to exploit the forward premium through altering bids related to physical scheduling, which is a costly way to induce price convergence. While previous literature has shown that the introduction of virtual bids does lead towards price convergence, it is also a relatively large market shock that potentially introduces new market participants with different risk preferences. This paper is the first to explicitly test the effect of increasing virtual bid transaction costs on forward price premiums using a natural experiment in a market where virtual bidding is already established. Using high-frequency price data with an event study approach, I find that increasing transaction costs on virtual bids leads to significant increases in forward premiums and significant decreases in the total number of cleared virtual bids. Additionally, my analysis supports recent literature arguing that the day-ahead prices have converged to become an unbiased predictor of real-time prices, which is an important condition for efficient markets. Lastly, I find no evidence that increasing transaction costs on virtual bids translates to increases in intra-day price volatility.