The effects of recent volatility in international petroleum markets on Canadian wholesale and retail gasoline prices

Wholesale prices

There are various sets of wholesale prices. Furthest from retail prices are those determined where many buyers and sellers trade in large quantities. This is true of the commodities exchanges where futures contracts are sold. The same applies to the spot prices arrived at in centres such as New York, Chicago and Rotterdam where large volume sellers and buyers quote asking and offering prices. There are no equivalent markets in Canada, so refiners here (as in the U.S.) look to these markets (as well as to crude prices) for information on the near term course of the next level of wholesale prices, which is product sold at the rack at refineries or terminals.

The refiners (or other sellers) quote or post these rack prices. These prices often hold for two or three days, but the changes could be more frequent. Unlike spot and future prices, in Canada rack prices are rarely transaction prices since the purchases are generally made as part of a contract that includes discounts off the rack price based on considerations such as the volume of sales over specified periods and the credit worthiness of the buyers. Wholesale buyers are responsible for transportation from the rack to retail outlets and ordinarily are purchasing generic gasoline; i.e., without the additives in the gasoline marketed by the refiners. In the U.S. there is a parallel set of rack prices for branded gasoline. There, refiners have surrendered the recruiting of dealers and supply to them that has been handled internally by Canadian refiners, to jobbers. Recently, Imperial has entered into a jobber arrangement with an Ontario firm.

The final set of wholesale prices is that charged to branded retailers. In the case of one refiner, these prices are tied to rack prices unless retail prices charged in an outlet controlled by the refiner in the vicinity of the retailer fall below rack prices. In that event prices are adjusted. Some other refiners base the wholesale price on prevailing retail prices.

Geographical Relationships between Spot Prices

Petroleum products, like crude oil, are internationally traded commodities. Over time, when adjustments to regional changes are allowed for, there is pattern of price relationships. Over a six year period starting in 1998, there were the following average daily spot prices for regular gasoline, in U.S. cents per gallon: Rotterdam ? 71.02; U.S. Gulf ? 74.33; New York ? 75.54; Chicago ? 78.21; and Los Angeles ? 85.94. The price in New York cannot long exceed the cost of transportation from the U.S. Gulf, the major source of supply, or from Rotterdam, one of several sources of imports. There is a distinct upward east to west drift in prices. The west coast, does not have close supply points alternative to regional production. Thus although the average price in Singapore over the same period was 69.02 cents/gal, which was lower than Rotterdam's, the great distances do not allow it to have a dampening effect on prices on the west coast.

There is a similar, but broken east to west pattern for rack prices in Canada. Several points arising from Table 2 and information outside it are worth noting. Information for a number of centres has been omitted from the table in the interest of manageability. In the case of St. John's, its regulated price is a cent per litre higher than that in Saint John and somewhat more than that in Halifax. The difference in transaction prices might be even higher if the regulated environment does encourage the negotiated discounts that occur in othe parts of the country.³⁶

As seen in Table 2, what has been identified as the east to west pricing pattern did not hold for the entire period. This is explored more fully later.

Prices in Montreal and Quebec City are virtually identical. This suggests that Ultramar is the key price leader in both centres. It would be very surprising if it were not responsible for the price in Quebec City, where it is the only refiner. Thus it is likely that prices in both centres are set by Ultramar at the same time.

Toronto and Sarnia, two refining centres connected by a pipeline, also have virtually identical prices. Based on similar reasoning to that applied to Montreal and Quebec City, it is likely that Imperial is the price leader in both locations, since it is the only company with refineries in both centres.

Finally, as discussed below, the east to west pattern is broken by the fact that pricing from Thunder Bay to Kamloops is determined by the price in Edmonton plus a differential. Thus Thunder Bay and all of the other points east of Edmonton have higher prices than Edmonton itself.

Table 2 Selected Canadian Rack Prices

Period	Saint John	Montreal	Toronto	Edmonton	Vancouver
1998-2004	31.00	30.51	31.72	31.91	32.99
Jan-May/04	42.47	42.87	43.49	43.47	43.92

Source: JC Ervin & Associates

Pricing in Western Canada

Edmonton is the refining center in Western Canada with each of the three integrated major oil companies operating a relatively large refinery. The only other multi-product refinery operating between Sarnia and the east coast is the Consumers' Coop in Regina. The output from Edmonton is shipped east and south by a number of product pipelines that reach Calgary, Regina, Saskatoon and Winnipeg. Product to Thunder Bay is transshipped from Winnipeg by train. Kamloops and Vancouver are supplied by a dual-purpose pipeline that accommodates crude oil as well as refined products. The price set in Edmonton, along with fairly stable location differentials' determines the price throughout a wide area, stretching from Kamloops at the Western extreme to Thunder Bay in the East. As the series of charts numbering from two to six demonstrate, the difference between prices in Edmonton, on the one hand, and Calgary, Regina, and Winnipeg on the other, have for the most part either been constant, or have undergone discrete changes, as in the case of the Calgary-Edmonton price differential.

Apart from occasional variations, there have been very stable relationships between the rack prices in Edmonton and Calgary. In the earliest period the differential was 0.40 cents³⁷/L. This was subsequently changed to 0.60 cents/L and then to 0.80 cents/L. The amount of the Calgary-Edmonton differential is larger than the cost of transporting gasoline from Edmonton via the pipeline which is presently 0.455 cents/L, well below the differential of 0.80 cents/L. We were informed that the capacity of the pipeline was sometimes fully utilized and that product has to be shipped by truck, thus adding to the average cost. It is difficult to know how important this factor is in explaining the differential, but it is doubtful that the stability of the differential and the shifts in it can be explained by stable transportation cost. It is more likely that there is a strong pattern of price leadership. The same thing can be said with regard to the other differentials. One of the most stable differentials has been the steady 0.60 cents/L between Regina and Edmonton. In 1998 the transportation cost was approximately 0.24 cents/L. It increased to about 0.43 cents/L in 2001 and was about 0.52 cents/L in 2003. Clearly the differential was established and maintained independently of the cost of transportation. In the case of Winnipeg-Edmonton the differential has been fairly steady around a dollar, which is close to the current cost of transportation, but does not correspond to earlier costs.³⁸

Chart 2: Rack Price Differentials: Calgary

Source: MJ Ervin & Associates

Chart 3: Rack Price Differential: Winnipeg-Edmonton

Chart 4: Rack Price Differentials: Kamloops-Edmonton, Vancouver

Chart 5: Rack Price Differentials: Regina

Because of the possibility of arbitrage between Kamloops and Vancouver, the relation between prices in Kamloops and Edmonton is the most interesting. With a road distance of roughly 350 kilometers between Edmonton and Kamloops it is feasible to transport product by truck in either direction. It will be noted that the chart showing the differential between Kamloops and Edmonton also shows the differential between Vancouver and Kamloops. The purpose is to see whether large differences in the latter differential triggered changes in the Kamloops-Edmonton differential as presumed arbitrage causes the price in Kamloops to rise or fall thus causing a respective widening or narrowing of the Kamloops-Edmonton differential.³⁹ Although several deviations in the Kamloops-Edmonton differential can be traced to large changes in Vancouver-Kamloops, other factors also must have been in play.

Chart 6: Rack Price Differentials: Thunder Bay-Edmonton

Chart 7: Rack Price Differentials: Thunder Bay-Sarnia

Source: MJ Ervin & Associates

Pricing in Thunder Bay is only tied to Edmonton in the second half of 2001. Before then there is no discernible relationship. Since product is shipped to Thunder Bay from both Sarnia (by ship) and Edmonton the possibility that Sarnia was used as a benchmark for Thunder Bay prices in the earlier period was explored. Although the Thunder Bay-Sarnia differential is not as regular as that between Thunder Bay and Edmonton, prices in Thunder Bay do appear to be tied to those in Sarnia until June 2001.

Given the fairly stable relationships between prices in Edmonton and those in centres ranging from Kamloops to Thunder Bay,⁴⁰ it is clear that the key to understanding the level of prices in Western Canada is an understanding of how prices in Edmonton are determined. Unlike cities in the east, Edmonton itself is fairly isolated relative to U.S. cities that might serve as supply points to independent wholesalers. Thus it is unlikely that a concern about arbitrage would be a motivating factor in pricing. According to information supplied by companies, prices in Minnesota and, to a lesser extent, Chicago are used as benchmarks in setting prices in Edmonton. Regardless of whether refiners are facing continuing pressure from potential arbitrage, it must be helpful to have benchmark prices to use as a guide when setting their own prices.

Like many cities in the East, Vancouver is easily accessible to imports, either by sea for large volumes, or overland by truck from nearby Seattle. There is no easily discernible relationship between prices in Edmonton and Vancouver. Unlike in the rest of Western Canada, the Edmonton refiners are almost certainly price takers rather than price setters. While the overall average difference in prices of 1.07 cents/L is close to the current pipeline tariff of approximately 1.13 cents/L, this is a coincidence. During the first part of the period covered, from January 1998 until January 2001, the average difference was 1.64 cents/L and there were many months when the differential was far in excess of the pipeline tariff. The reverse is the case for the succeeding period, from January 2001 until May 2004, when the average differential is 0.57 cents/L, clearly well below the cost of transportation. These numbers suggest that it is highly unlikely that prices in Edmonton are set at a level to meet Vancouver prices and cover transportation costs. It is more likely that the benchmarks used in setting prices in Edmonton bear some relationship to the forces acting on Vancouver prices. And of course one common factor is the price of crude oil.

Pricing in Eastern Canada

Most regional centres in Eastern Canada are easily accessible by sea or the Great Lakes. In addition, many cities in Southern Ontario are within easy driving distance of either Buffalo or Detroit. Thus successive investigations of the pricing of wholesale gasoline have concluded that domestic refiners have limited pricing freedom.

But it is also important to look beyond locations where there are a number of domestic refiners within close proximity to each other that have a large volume of capacity to utilize. To this end the differential between Saint John and Halifax rack prices, respectively, were compared with those in Montreal. Montreal was chosen as providing a competitive Canadian benchmark since wholesale prices there have consistently and closely reflected the level of prices in the Atlantic Basin. The differentials for Saint John are shown in Chart 8. Differentials of the order of two cents/l in the early and middle 1990s represented very high refiner margins. But for reasons that are not immediately apparent, there has been a gradual shift towards more competitive price levels in both Halifax and Saint John for both gasoline and diesel, to the point that prices in both centres are, presently, somewhat below those in Montreal. It is noteworthy that there is a single refiner in both locations, Irving in Saint John and Imperial in Dartmouth. The existence and possibility of purchase/sale agreements did not provide a sufficient presence of other refiners to create a competitive outcome.

Chart 8: Rack Price Differential: Saint-John-Montreal

Conclusion: Relationship Between Canadian Rack Prices and U.S. Spot Prices

As is evident in Chart 9, rack prices in Canadian centres are aligned with spot prices in New York. Much of this close relationship is undoubtedly due to a strong common factor affecting all prices, namely changes in crude oil prices. Thus of greater interest than the timing of price changes is the differential between spot prices in New York and rack prices in Montreal and Toronto. These are more clearly seen in Chart 10, that traces these differentials.

The key consideration in both comparisons is that there was no widening in the differentials during 2004 generally, and none as well when prices peaked in May compared to average differentials since the beginning of 2002. In fact it is somewhat surprising that the differentials did not widen in 2004, since they are not immune to general market conditions as indicated by narrower differentials in 2002 when market conditions were less favourable for refiners than in the preceding and succeeding years. It is doubtful that the differential covered discounts off of rack prices and the terminal costs. In 2003 there was probably some cushion. The average differential in cents per litre for Montreal⁴¹ in 2002, 2003 and 2004, respectively, was: 2.36; 3.46 and 3.16. The pattern is similar for Toronto: 3.55; 4.73; and 3.78 cents per litre in 2002, 2003 and 2004, respectively. The pricing experience in these two cities is similar across the country. There was no spike in wholesale prices in 2004 that cannot be traced to increased crude costs and widening margins in the U.S. When the latter declined, so did wholesale prices in Canada to the extent of more than offsetting increasing prices for crude oil.

These data support the well-established conclusion in the industry that prices in Canada tend to closely track those in the U.S. This is not to say that divergences are not possible, but price movements of the magnitude that occurred in 2004 in Canada are inconceivable unless they are reflective of similar movements in the U.S. and supply sources in Europe and the Caribbean. In general price leaders in Canada anticipate the results that would occur under arbitrage so the product movements are not necessary for the tracking of prices to occur.

Chart 9: Relation of Canadian Rack Prices to NY Spot Spice

Chart 10: Difference in Toronto, Montreal Rack and NY Spot

³⁶ Maximum wholesale and retail prices are set by the Island Petroleum Regulatory Commission and Appeals Commission and the Petroleum Products Pricing Commission, respectively, in Prince Edward Island and Newfoundland and Labrador. The latter body was integrated with the Public Utilities Board at the end of May 2004. Because of the large distances in Newfoundland and Labrador maximum prices are established for a number of geographic zones.

³⁷ One of the major breaks in the pattern occurred between April and October of 2001 as average monthly prices underwent several large changes. The average monthly rack prices per litre in Edmonton during those months were: 36.53, 42.8, 41.72, 33.37, 40.33, 45.08, and 30.63 cents, respectively. As prices increased sharply between April and May in Edmonton, they increased by a larger amount in Calgary and the reverse happened as prices rose in August and September. What is noteworthy about these departures from a strong pattern is that the same departures occurred in the more or less constant price differentials between Edmonton and the other centres during those months. Similarly, when another sharp break in the Calgary-Edmonton differential occurred in March 2003, there were similar changes in the price differentials between Edmonton and the other centres. Once again it was during a period of price volatility due to changes in the price of crude oil. However, the recent rapid run up of crude oil prices and the accompanying increases in wholesale prices have not resulted in similar departures in established differentials.

---

³⁸ There are two legs to the route: Edmonton to Gretna, the longer haul, and Gretna to Winnipeg. Rates for the former were easier to obtain since the tolls are controlled by the N.E.B. In 1998 the rate was 0.43 cents/L and reached 0.75 cents/L in 2003. I am grateful to Craig Rubie of the N.E.B. for his help in obtaining the rates for earlier years.

³⁹ The pipeline cost of product shipped from Edmonton to Kamloops was 0.81 cents/L in 2000. (The next relevant figure available is 0.67 cents per litre in 1996.) Thus the rack price differential is closer to the pipeline costs than is the case for the other centres.

⁴⁰ Some appreciation of the degree to which prices in Western Canada are related can be obtained by comparing the coefficient of variation for the differential in prices between points for which there is no product movement and the existence of price leadership is highly doubtful. Over a similar period to the price comparisons in Western Canada the coefficient of variation of the Saint John-Montreal price differential was 2.49, due to considerable variability and a relatively small average difference.

⁴¹ Montreal has been used rather then Ottawa, even though the latter is a focus of interest, because Montreal is most directly affected by imports and the threat of imports. The rack price in Ottawa has been 0.52 cents/L higher than in Montreal over the last six years and declined somewhat in 2004.