قیمت های نفت : نقش بهره برداری از پالایشگاه، بازارهای آتی و غیر خطی

We test the hypothesis that real oil prices are determined in part by refinery capacity, non-linearities in supply conditions, and/or expectations and that observed changes in these variables can account for the rise in prices between 2004 and 2006. Results indicate that the refining sector plays an important role in the recent price increase, but not in the way described by many analysts. The relationship is negative such that higher refinery utilization rates reduce crude oil prices. This effect is associated with shifts in the production of heavy and light grades of crude oil and price spreads between them. Non-linear relationships between OPEC capacity and oil prices as well as conditions on the futures markets also account for changes in real oil prices. Together, these factors allow the model to generate a one-step ahead out-of-sample forecast that performs as well as forecasts implied by far-month contracts on the New York Mercantile Exchange and is able to account for much of the $27 rise in crude oil prices between 2004 and 2006.

Causes for the rapid rise in the price of crude oil between 2004 and the summer of 2006 are the subject of debate. Some of the debate focuses on changes in the so-called downstream sector especially the refining sector. The number of refineries in the United States has not increased since 1981 (Annual Energy Review, 2006), and in the spring of 2007, a significant fraction of refining capacity was closed due to unscheduled maintenance (New York Times, 2007). Under these conditions, a lack of spare refining capacity is seen as one cause for the on-going rise in the price of crude oil and refined petroleum products. Other hypotheses for the sharp rise in oil prices include the lack of spare production capacity, a non-linear relationship between oil prices and supply, and changed perceptions of the balance between supply and demand. Although a linear relationship can be a reasonable approximation under normal circumstances, extreme events may shift the market equilibrium between supply and demand towards different types of market functioning in which prices are much more sensitive to shocks than under normal conditions. On the supply side, non-linearities may be caused by lags associated with building additional extraction and refining capacity (Kaufmann and Cleveland, 2001 and Kaufmann, 2007). Given these constraints, oil prices would be more sensitive to supply as production approaches capacity. Finally, expectations about the supply/demand balance, as reflected by conditions in the futures market, may affect current prices. Hypotheses that refining capacity, non-linearities, and expectations, have an important effect on oil prices are consistent with the performance of models that exclude their effect. For instance, the model by Dees et al. (2007), which specifies crude oil prices as a function of OPEC capacity, OECD crude oil stocks, OPEC quotas and cheating by OPEC on those quotas, performs well in-sample (1986–2003), but consistently under-predicts real oil prices out-of-sample, 2004–2006 (Fig. 1). This bias indicates that the model omits variables that are largely responsible for the increase in oil prices between 2004 and 2006. Full-size image (28 K) Fig. 1. The observed value of the near month contract on the NYMEX (solid line). The forecast for the average prices for US crude oil imports generated by a model that omits the effects of refinery utilization, non-linearities, and market conditions in the NYMEX (dotted line). The one-step ahead out of sample forecast generated by the econometric model (Eqs. (1) and (2)) is given by open circles (root mean square error = 4.07), the forecast implied by the near month contract on the NYMEX is given by the open squares (root mean square error = 3.54), a random walk, as given by the lagged value of the near month contract on the NYMEX (mean square error = 3.08). Open diamonds represent the price simulated by the econometric model with information about the exogenous variables only (root mean square error = 6.87). Figure options In this paper, we test the hypothesis that real oil prices are influencedby by refinery capacity, non-linearities in supply conditions, and/or expectations about supply/demand balances and that observed changes in these variables can account for the rise in prices between 2004 and 2006. To do so, we expand the equation described by Kaufmann et al. (2004) to include observations for US refining utilization rates and price differences between the far month and near month contract for crude oil on the New York Mercantile Exchange (NYMEX). To test for non-linearities, the linear specification for capacity utilization by OPEC in Kaufmann et al. (2004) is replaced with a cubic function. Results indicate that the refining sector plays an important role in the recent price increase, but not in the way described by most analysts. The relationship is negative such that higher refinery utilization rates reduce crude oil prices. This effect is associated with shifts in the production of heavy and light grades of crude oil and price spreads between them. Together with a non-linear relationship between OPEC capacity utilization and oil prices as well as conditions on the futures market, the expanded equation generates a one step ahead out-of-sample forecast that performs as well as the forecasts implied by far-month contracts on the New York Mercantile Exchange and is able to account for much of the $27 rise in the real price of crude oil between 2004 and 2006. These results and the methods used to obtain them are described in five sections. Section 2 describes the data and econometric techniques used to estimate a cointegrating relationship for crude oil prices. Results are described in Section 3. Section 4 discusses the effect of refinery utilization rates on crude oil prices, the importance of non-linearities in marginal supply, and expectations, as measured by conditions in the futures market. It also presents the ability of this econometric equation to forecast oil prices. Section 5 concludes.

The rapid rise in the price of crude oil between 2004 and the summer of 2006 has been difficult to explain with the usual fundamentals related to the supply/demand balance. This paper investigates additional factors that might have contributed to the oil price increase. Most of the increase can be explained by concerns about future oil market onditions, as represented by the shift of the futures market from backwardation to contango, as well as changes in the refining sector, with a drop in the refinery utilization rate. Factors related to crude oil supply continue to be important when we account for nonlinear relationships between OPEC supply behavior and oil prices. Interestingly, results of this analysis indicate that there is little vidence that increasing refining capacity will lower crude oil prices. Of the variables identified by this paper to effect prices, only stocks of crude oil could effectively lower prices—each day of forward consumption reduces real oil prices by about $2 in the long run. Nonetheless, despite a recent upturn, days of forward consumption have generally declined over the last 20 years, from about 90–95 days of forward consumption to 78–82 days of forward consumption. Interestingly, this reduction is not due to a reduction in stock levels, but is due to the fact that the increase in storage capacity has been considerably slower than the increase in demand. This implies that market conditions may not provide the economic incentives needed to expand storage facilities with demand. Against this background, as long as demand remains robust, there are very few reasons to expect oil prices to return to levels observed before 2006.