My puzzlement over the close connection between natural gas and crude oil prices continues. In the last year or so, gas prices have moved up and down (mostly up) together with oil prices. Gas prices have averaging about $2/MMBtu lower. When the prices of gas and oil are nearly equal they are connected by the fact that some industrial consumers can switch from one to the other. But with a $2 difference and no fuel switching potential, why the connection? My curiosity was further aroused by last week's Energy Commission hearing on the Integrated Energy Policy Report (IEPR) at which staff reported the results of their price forecast model. This model did a lousy job of forecasting the last ten years' gas prices, so why should it be relied upon to predict the next ten years? I decided that any forecasting model worth its salt should at least fit the historical data fairly well. That's what we demand of climate change models, for example. So I rummaged around in the data files that accumulate in my computer and came up with a surprising (to me anyway) result. Over the last eight years, if you had predicted that the natural gas price would be equal to 75 percent of the crude oil price, you would have been much more accurate than any other analyst on the planet. You would have missed the sharp spikes in the winters of 2000-01 and 2002-03 when gas storage levels sank to alarming levels. Gas supplies relative to consumption really do matter, but apparently they don't matter much. Will the price of gas continue to follow oil? If so, there are some counter-intuitive consequences. For example, importing liquefied natural gas would lower U.S. gas prices only slightly, if at all. A gas pipe from Alaska wouldn't reduce prices either. More frightening is the possibility that oil prices will skyrocket if crude oil production begins to decline as some analysts believe. If gas prices follow, they would also reach stratospheric levels. It's much too soon to declare that the price of gas will continue to track oil prices into the future. However, my gas price forecasting model now assumes that it will. And my model fits the historical data a whole lot better than any other I've seen. Now, if I only knew what oil prices were going to do.
