Three Essays on Commodity Risk Management

114 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2007.This dissertation consists of three papers on risk management with empirical applications for commodity markets. The first two papers analyze selective hedging, where risk managers have views on future market conditions and sometimes hedge selectively based on these views. I develop two Bayesian optimal hedging models based on the Bayesian portfolio optimization framework. The Bayesian approach is chosen because it jointly considers subjective views and parameter estimation risk. The first paper considers only subjective views and estimation risk regarding the expectation vector of asset returns, while the second paper extends the framework to the covariance matrix of asset returns. Numerical examples in these studies show that subjective views can have a substantial impact on risk managers' hedging decisions and that the impact is most evident when the hedger speculates on market price direction and/or is pessimistic about the effectiveness of hedging, i.e., a breakdown in the correlation among different markets. Overall, the Bayesian optimal hedging models not only help explain the large cross-sectional and time-series variation in hedging positions often observed in practice, but also provide risk managers with a theoretically intuitive yet quantitatively rigorous tool to blend their views on market conditions with a "market-wide" or "firm-wide" consensus in determining optimal hedging positions. The third paper estimates the cost (i.e., risk premium) of pre-harvest forward contracting for wheat in Illinois and Kansas. Given the similarities between forward and futures markets, regression models used for testing the risk premium hypothesis in futures markets are applied to forward markets. Cost is estimated for both unconditional and conditional levels. The empirical results show that the average cost of forward contracting is higher than that of futures hedging in both states and cost varies systematically in relation to the level and volatility of forward prices.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD