"Virtual bid" and "Virtual offer" are purely financial products offered in certain electricity markets. Theoretically, virtual bids and offers can change the electricity price as the bids and offers are stacked along with the demand and supply, respectively. This dissertation discusses how virtuals can be used to hedge and speculate in the electricity market.

A statistical simulation model is developed based on the day-ahead (DA) demand and real time (RT) load data from Midwest Independent Transmission System Operator's (MISO) footprint and DA and RT price observed at Cinergy hub. The simulation models are intended to mimic the load and price processes, taking the cyclical and correlation patterns in the market data into account as well as to provide a mechanism to incorporate stochastic variations that impact the processes. This model can then be utilized to study how the various trading strategies perform under deferent scenarios and thus provide better decision making tools to a trader. The DA Demand and RT Load are simulated using a combination of unobserved component models (UCM) and a set of regression variables. The DA Price and RT Price processes are replicated with GARCH based regression models. The regressor variables include principal components of different weather variables to capture the weather variation across MISO footprint and a set of dummy variables to model key patterns observed in the electricity market. The simulation models are used to generate test data sets which are then used to analyze different strategies involving virtuals. The simulation models also help to understand the relationship between DA and RT clearing prices. This research finds no evidence of DA/RT price convergence purely based on the virtuals trading at M1SO. Based on the simulation results, the virtual bids appear to be most profitable during summer and winter and virtual offers appears to be most successful during shoulder months.