Three Essays on Energy and Agricultural Price Analysis

This dissertation consists of three essays on energy and agricultural commodity price analysis: 1) Natural Gas Price Forecasting in a Changing World; 2) The Effect of EIA Storage Announcement on Natural Gas Returns: A Comprehensive Analysis; and 3) Forecasting the U.S. Season-Average Farm Price of Corn: Derivation of an Alternative Futures based Forecasting Model. The first essay evaluates the performances of various individual and composite forecasting models when predicting natural gas prices in the United States. The empirical results show that forecast generated by the Energy Information Administration Short-Term Energy Outlook provides a more accurate price prediction at longer forecasting horizons (6- and 12-month ahead) while futures-based forecasts perform better in the short-run (1- and 3-month ahead). Projections based on time-series models perform well at longer forecast horizons when price volatility is relatively low. Further, the Hotelling model performs well for 1- and 3-month ahead forecast horizons. Our findings further support the additional benefit of composite forecasts based on individual methods for more accurate predictions; however, the performance is not uniform at different forecasting horizons. The second essay examines how natural gas prices react to inventory surprises contained in Energy Information Administration’s weekly inventory report. Results indicate that natural gas prices are more responsive to 1) negative (more-than-expected) surprise storage news as compared to positive (less-than-expected) surprises, 2) news released during the injection season as compared to the withdrawal season, and 3) inventory surprises occurring in periods of tight supply in withdrawal season compared to when the market has an abundant supply. Finally, we find that EIA’s inventory report has exerted a smaller impact on natural gas prices over time. Possible contributing factors to this declining impact include the increasing availability of alternative information providers in the market, the relatively over-supply of natural gas during the period of analysis since the rise of unconventional production, and a more integrated regional market that can transport natural gas from production to consumption regions more efficiently. The third essay investigates an alternative futures-based procedure to forecast the season-average farm price (SAFP) for U.S. corn, an under-researched price forecast. With the exceptionally volatile conditions experienced in the corn market since 2006, the need for price forecasting has become more critical. The new model developed in this essay performs better than two widely watched season-average price forecasts (World Agricultural Supply and Demand Estimates and the Hoffman futures-based forecasts) at the beginning of the post-harvest season, and just as well as those forecasts at the beginning of the forecast cycle and in the later post-harvest season. We attribute the performance of the proposed model’s forecasts to its assignment of heterogeneous weights to both futures and cash prices depending on the underlying market conditions. Improved performance of the proposed model’s forecasts is especially noticeable when the market is more volatile