|
Article Excerpt
Gasoline prices increased dramatically in the U.S. Midwest in the summer of 2000, generating allegations of collusion among gasoline marketers. We examine the causes of the price increase, and find no evidence to support the collusion story. Instead, a combination of industry characteristics and unanticipated problems in switching to a new, federally-mandated gasoline specification caused the spike. Once prices rose, firms responded as quickly as possible to get additional supplies to affected markets.
INTRODUCTION
When gasoline prices rise people notice. Many believe that price increases are the result of anticompetitive practices; even in the wake of an impending war and the collapse of Venezuelan production there were charges of "gouging" in winter 2003. We examine the spring 2000 Midwest gasoline price spike and discuss the market forces that caused it. In particular, we focus on two categories of causes: structural problems and supply shocks. We find that structural problems in the gasoline refining and distribution system, including limited refining capacity, and problems in the logistics of distribution, set the stage for volatile prices. Supply shocks, such as refinery outages or pipeline ruptures, triggered rapid price increases.
In May and June 2000, retail gasoline prices in the Midwest increased sharply, hitting $2.50 per gallon in parts of Chicago. (1) The peak wholesale price rose from $0.93 at the end of March to a high of $1.60 per gallon in Chicago. The fall was just as dramatic; wholesale prices were back to $0.93 by July 11. (2)
The dramatic increase in Midwest gasoline prices in the first few weeks of May 2000 prompted both Congress and President Clinton to request an FTC investigation into the causes of the price spike (Federal Trade Commission, 2001). The obvious, immediate culprit for the price increase was the implementation on May 1 of the new EPA standard for reformulated gasoline (RFG), known as Phase II.
Beyond the new gasoline standard, several institutional factors created an environment conducive to price spikes. We describe these factors in Section 3, "Setting the Stage for a Price Increase." We describe other contributing factors in Section 4, most notably several pipeline disruptions and longer-than-anticipated refinery maintenance.
Although the price increase was severe, it was of brief duration. As soon as prices in the Upper Midwest increased above those in the Gulf Coast by more than normal levels, refiners took steps to increase supplies to the affected areas. However, this process took several weeks because refiners first had to switch production in other regions over to the special gasoline used in Chicago and Milwaukee, and then ship the product on a slow-moving barge or pipeline. Section 5 looks at the varied responses of area refiners. Their behavior was consistent with normal competitive behavior in the face of a supply shock. We conclude there is no credible evidence that the price increase was due to coordinated behavior by the oil companies. Observed prices are fully explained by non-collusive market forces. Finally, we argue that criticisms of the refiners for not better anticipating the shortages that developed are unjustified.
1. AN OVERVIEW OF MIDWESTERN REFINING
The Energy Information Administration reports a large amount of petroleum industry data by regions, which are known as Petroleum Administration for Defense Districts ("PADDs"). The Upper Midwest, which we consider to include Illinois, Indiana, Wisconsin, Michigan, and Minnesota, is part of PADD II. (3) PADD II has substantial refining capacity, although this capacity has decreased slightly between 1999 and 2001 as two Upper Midwest refineries have closed. (4)
PADD II as a whole and the Upper Midwest in particular are not self-sufficient in petroleum products production. The PADD and the Upper Midwest must "import" substantial quantities of products from outside, primarily from refineries located along the Gulf Coast in Texas and Louisiana. The principal means of delivery of imports, as well as distribution of locally produced products, is by pipeline. However, pipeline capacity into the region has been used to its limits in recent years, particularly during the summer driving season, so that at peak times additional supplies are barged into the region along the Mississippi and Ohio Rivers. This means that the market price in major consumption centers of the region tends toward the purchase cost of product in the Gulf Coast states plus the cost of the most expensive mode of transportation necessary to meet the prevailing demand, i.e., pipeline tariffs in the winter and barge fees in the summer. However, if these facilities are at or near maximum capacity, the abil ity of the system to respond to short run supply disruptions is limited. Prices may then tend to rise until demand is rationed to the available supplies.
Reformulated gasoline must meet the pollution reduction targets predicted by the EPA's "Complex Model," which estimates the degree of pollution reduction attainable through the specific properties of a particular blend of gasoline. Beginning on January 1, 2000, refiners were required to be distributing Phase II RFG in those areas in which it was legally required, and, effective May 1, 2000, they had to be distributing Summer-grade Phase 11 RFG. Production of Phase II RFG, particularly the Summer grade, apparently requires several higher quality blend stocks, such as raffinate, alkylate, and tolulene, that are often used in premium gasoline.
By law, reformulated gasoline requires the use of an oxygenate. The two most commonly used oxygenates are MTBE, which is a chemical compound created in the refining process, and ethanol. (5) Because general product pipelines cannot transport gasoline containing ethanol, ethanol-oxygenated RFG is actually created at the terminal, when up to ten percent ethanol is blended with a base stock (known as RBOB), usually as the product is being placed in tank trucks for delivery to local service stations. During the summer months, the volatility levels of RFG are closely controlled to reduce the smog-creating potential. Producing the RBOB for Summer-grade, ethanol-oxygenated RFG requires particular care because of the synergistic effect between ethanol and the blend stock in raising volatility levels (often referred to as Reid Vapor Pressure, or RVP, for the commonly used measurement of this property). In Section 4.3 below we discuss in more detail why Midwest refiners chose ethanol as an oxygenate and the implicatio ns of that decision.
Within limits, a refinery can vary the proportions of the types of gasoline it blends from the various blend stocks on a one-for-one substitution. However, if a blend stock that is required for producing minimum standard conventional gasoline is used more intensively in the production of reformulated gasoline, then beyond a point an extra gallon of RFG may come at the cost of more than a gallon of regular. This may have happened at one or more Midwest refineries as they tried to maximize the amount of RFG produced in summer 2000. The distinction between the average versus the marginal cost of refining RFG II is particularly important when looking at transitory supply shortages. The EPA estimated the additional cost of RFG II at about 1-2 cents per gallon (cpg) over the cost of RFG I and 4-8 cpg over the cost of conventional gasoline. As a result, senior EPA officials expressed some mystification as to the magnitude of the Midwest price increases, (6) saying that "EPA believes the cost of producing RFG II does not account for the extreme prices being paid by Midwest consumers." However, these officials clearly applied estimates of the average cost of making REQ during "normal" conditions to a period of multiple supply disruptions, as we discuss below, where the marginal barrel of RFG was very expensive.
2. PRICES ROSE, THEN FELL QUICKLY
At the beginning of May 1999, the average retail price for regular unleaded RFG in PADD II (including Chicago, St. Louis, Louisville, and Covington) was $1.11 per gallon, including taxes. (7) By May 1, 2000, the average price was $1.48/gallon and starting to rise sharply. On June 5 the average was $1.84 and on June 19 it peaked at $2.00. By July 24, the price had fallen to $1.43/gallon.
Rising crude oil prices played an important role in the increase over the previous year, as Figure 1 shows. On May 3, 1999 the reported price for West Texas Intermediate (WTI) crude oil was 44.8 cents per gallon. (8) One year later, on May 3, 2000, the price was 63.7 cpg. The price of crude oil was near its peak on June 15, 2000 -- just about the time that gasoline prices peaked -- at 78.5 cpg, or an additional 15 cpg from the May 3, 2000 price. (9) At this point, the price of WTI was more than triple the price in December 1998. These higher costs, combined with the supply and transportation shocks, are what pushed Chicago retail prices over $2.00 a gallon.
Although the retail price increases were the ones that made headlines, the widespread nature of the price increases suggested that any anticompetitive behavior occurred further upstream, at the refining or marketing level. Gasoline retailers are generally independent businessmen and operate at a less concentrated level of the market than do marketers. The focus in this paper will be on wholesale prices.
The average RFG branded average rack price for regular unleaded at the Chicago terminal on May 2, 2000 was $1.04/gallon, compared with $0.67 on May 3, 1999. (10) At the peak, on June 15, 2000, the price had risen to $1.65/gallon, but by July 24, 2000 the price had fallen to $0.88/gallon. During this period, conventional gasoline prices also rose in the Midwest. The price of regular unleaded conventional gasoline on May 2, 2000 was $0.83/gallon, peaking at $1.51/gallon on June 15, and falling back to $0.83/gallon by July 22.
A more useful way to assess Midwest price increases is to look at the relative price between Chicago and a market unaffected by the problems in the Midwest, such as Dallas. In May and June 1998 the average difference between RFG prices in Chicago and Dallas was 11.2 cents per gallon; in May and June 1999 the difference averaged 5.7 cpg. (11) Throughout much of March 2000 the difference was actually negative, by as much as 6.7 cpg on March 16. The difference rose through April, reaching 8.4 cpg by May 1. The difference remained in the 10-11 cent range until the middle of May, then began rising sharply, reaching 41.1 cents by the end of the month, and peaking at 49.1 cents on June 9. At this point, the price differential fell rapidly, to 10.0 cents on June 24 and to 4.3 cents on June 29. By July 6 the difference had turned negative, and remained negative, at times by as much as 10 cents, throughout the rest of July. (12) The conventional gasoline differential between Chicago and Dallas is generally smaller, but exhibits a similar pattern.
Figure 2 shows RFG and...
|
