Three essays on strategic behavior, information revelation and restructuring in the U.S. electricity markets

The first chapter of this dissertation focuses on one phenomenon in restructured electricity markets, called price spikes. There are two divergent explanations for occurrence of price spikes: ordinary but tight supply and demand conditions, or abnormal behavior by suppliers. In the latter case, if suppliers alter their behavior by refusing to bid some of their otherwise-profitable output into the market with the purpose of raising price sufficiently so that their remaining output earns considerably more than the loss on the withheld quantity, this represents a perverse supply shift that can cause an abnormal price spike. This research develops an operational empirical approach for distinguishing between these two sources of price spikes. The data used in the analysis is hourly bidding data from the New York Day-Ahead electricity market in the summer of 2001. We use an OLS and non-linear exponential models to estimate highly non-linear supply curves in the electricity markets. The results indicate that strategic behavior of several large bidders, as well as two particular generators have contributed to or produced the price spikes observed in the New York Day-Ahead electricity market in the summer of 2001. These results establish the legitimacy of concern over strategic withholding and focus policy attention on periods of superpeak prices and on especially large bidders.The second chapter of this dissertation analyzes the effect of exchange of and access to rivals' private information on the pricing strategies of influential suppliers (so-called marginal bidders) in the electricity markets and on the probability of those bidders being marginal. In the majority of competitive markets exchange of cost and demand information between market participants is not a common practice since it can create opportunities for anticompetitive behavior and cause distortion of market equilibrium. Rules and regulations of some restructured electricity markets in the U.S., however, allow their participants to get access to rivals' private information by allowing short-term contractual agreements between bidders of electricity generation and multiple generating units simultaneously or over time. We provide a simple theoretical model that demonstrates how access to rivals' cost information may alter bidding behavior of a marginal, or likely to be marginal, bidder, and then test this model using our two unique measures of information revelation. This research uses hourly bidding data from the New York Day-Ahead electricity market in 2006-2008. We use fixed effects panel data model to investigate the effect of information revelation on price bids and Heckman two-step model to estimate the probability of bidders being marginal. Estimation results reveal differences between groups of marginal and non-marginal bidders and indicate that access to rivals' private cost information may cause higher marginal price bids by marginal bidders and, therefore, higher market equilibrium price.The third chapter of this dissertation analyzes the drivers behind states' choice to implement restructuring and the differences in performance of restructured and non-restructured states. Since implementation of electricity restructuring has not been mandatory in the U.S., there are more than half of U.S. states that still operate their regulated electricity markets. Although 23 states have been restructured up to date, there has been no conclusion reached yet on whether electricity restructuring achieved its goals of reduced costs and prices to consumers and greater market efficiency. We use annual state level data for all 51 U.S. states from 1990-2008. The analysis involves multiple estimations, including GLS model, two-stage least squares estimation and simultaneous equations model with endogenous switching to account for the heterogeneity in the states' decision to implement restructuring or not, and for unobservable states' characteristics. We find that significant heterogeneity is indeed present in the sample between restructured and non-restructured states and that there are substantial differences in determinants of retail electricity retail rates between the states that implemented restructuring and those that did not. Our results indicate that as state's restructuring choice is not exogenously defined, results from simple OLS or GLS models without correction for endogeneity bias, should be interpreted with care. These results are particularly important in developing states' decisions about electricity restructuring and further restructuring reforms that keep in line with the potential impact of restructuring on electricity prices. (Abstract shortened by UMI.)...