Abstract
This paper examines the impact of variation in transportation options – what I denote the “supply network” – on observed price differences between locations for a specific good, retail gasoline. I use a unique data set of weekly gasoline prices across 44 Canadian cities to analyze how the existence of variation in the available modes of transportation for gasoline between cities (via pipeline, marine tanker, rail or truck) accounts for observed price differences across locations. I find that the supply network is significant – cities connected by lower cost-per-unit methods like pipelines or seaports exhibit smaller mean- and weekly-price differences than those connected only by road or rail, after controlling for distance, regional effects and market size. A pipeline connection results in a reduction in weekly price dispersion equivalent to a 53% reduction in distance between cities, while a maritime connection has the equivalent effect of a 38% reduction in distance between cities.
Appendix A
Retail Price Correlations vs. Mean Price Differences: Regional.
Coefficient of Variation: Regional.
Mean Price Differences Regression: Excluding Que/ATL.
| Regression estimation: dependent variable | ||
|---|---|---|
| Variable | Reg. 1 | |
| Constant | −0.0188 | |
| (0.0197) | ||
| Distance | 0.0520*** | |
| (0.0176) | ||
| Region | 0.0751*** | |
| (0.0200) | ||
| Pipeline | −0.0405*** | |
| (0.0146) | ||
| Seaport | 0.0358 | |
| (0.0299) | ||
| Terminal distance | −0.0003 | |
| (0.0026) | ||
| Total population | 0.0240*** | |
| (0.0069) | ||
| Population density | 0.0042 | |
| (0.0049) | ||
| R2 | 0.3966 | |
| Adj. R2 | 0.3835 | |
| N (obs.) | 378 | |
***-significant at 99% CL, **-sig. at 95% CL, *-sig. at 90% CL.
Weekly relative price regression: Excluding Que/ATL.
| Regression estimation: dependent variable | ||
|---|---|---|
| Variable | Reg. 1 | |
| Constant | 0.0148*** | |
| (0.0054) | ||
| Distance | 0.0544*** | |
| (0.0022) | ||
| Region | 0.0660*** | |
| (0.0029) | ||
| Pipeline | −0.0298*** | |
| (0.0017) | ||
| Seaport | 0.0433*** | |
| (0.0033) | ||
| Terminal distance | −0.0004 | |
| (0.0004) | ||
| Total population | 0.0209*** | |
| (0.0008) | ||
| Population density | 0.0054*** | |
| (0.0005) | ||
| R2 | 0.2288 | |
| Adj. R2 | 0.2288 | |
| N (obs.) | 342,090 | |
***-significant at 99% CL, **-sig. at 95% CL, *-sig. at 90% CL.
Mean Price Differences Regression: Excluding YK/NWT.
| Regression estimation: dependent variable | ||
|---|---|---|
| Variable | Reg. 1 | |
| Constant | −0.0299*** | |
| (0.0037) | ||
| Distance | 0.0044** | |
| (0.0023) | ||
| Region | 0.0076* | |
| (0.0045) | ||
| Pipeline | −0.0157*** | |
| (0.0038) | ||
| Seaport | −0.0095** | |
| (0.0044) | ||
| Terminal distance | −0.0000 | |
| (0.0007) | ||
| Total population | 0.0019 | |
| (0.0014) | ||
| Population density | 0.0030* | |
| (0.0018) | ||
| R2 | 0.0791 | |
| Adj. R2 | 0.0704 | |
| N (obs.) | 861 | |
***-significant at 99% CL, **-sig. at 95% CL, *-sig. at 90% CL.
Weekly Relative Price Regression: Excluding YK/NWT.
| Regression estimation: dependent variable | ||
|---|---|---|
| Variable | Reg. 1 | |
| Constant | 0.0556*** | |
| (0.0024) | ||
| Distance | 0.0114*** | |
| (0.0008) | ||
| Region | 0.0054*** | |
| (0.0014) | ||
| Pipeline | −0.0040*** | |
| (0.0007) | ||
| Seaport | −0.0130*** | |
| (0.0009) | ||
| Terminal distance | −0.0001* | |
| (0.0003) | ||
| Total population | 0.0014*** | |
| (0.0004) | ||
| Population density | 0.0035*** | |
| (0.0003) | ||
| R2 | 0.0381 | |
| Adj. R2 | 0.0381 | |
| N (obs.) | 779205 | |
***-significant at 99% CL, **-sig. at 95% CL, *-sig. at 90% CL.
Data Sources and Implementation
The gasoline price data comes from the Kent Group Ltd. website, publicly available at http://charting.kentgroupltd.com/. I use weekly data on retail prices, excluding taxes, for regular gasoline, from the 44 cities listed, and compile it over the years 2001–2017. These prices represent a city-wide weekly average of gasoline prices in each city, as sampled by the Kent Group from a wide selection of branded and independent gasoline retailers every Tuesday morning at 10:00 A.M. local time.
The distance between cities is calculated as the shortest driving distance as suggested by online navigation system Mapquest.[30] While many papers use variations of the Great-circle or Euclidean distances between locations to measure distance, this paper focuses on the arbitrage condition that governs price dispersion, which is a function of the costs associated with physically transporting gasoline products between locations. The default alternative for transporting gasoline is by truck, as it is the only method accessible to all locations. I therefore use the shortest highway route provided by Mapquest, measured in kilometers, as the measure of distance between any given cities.
The region dummies are calculated to correspond with the natural supply orbits suggested by Natural Resources Canada. Cities in British Columbia, Alberta, Saskatchewan and Manitoba, as well as the Yukon and Northwest territories fall into the West region; cities in Ontario and Quebec fall into their respectively named regions; and cities in New Brunswick, Nova Scotia, Prince Edward Island and Newfoundland and Labrador fall into the Atlantic region.
The pipeline dummies are calculated according to the current pipeline infrastructure in Canada, available from the Canadian Association of Petroleum Producers (CAPP). Cities are considered to be connected by pipeline if they are directly connected by an existing pipeline, or if they are both connected to a common third city by pipeline. For example, while Edmonton and Regina may not share a direct pipeline link, they are both connected to Calgary, and are thus considered to be linked via pipeline.
The seaport dummies are set to one if there exists a plausible maritime connection between the two cities, whether or not gasoline products are currently shipped via marine tanker between the two cities. This reflects the fact that the price dispersion between cities is defined by the arbitrage condition governed by the available transportation options, whether they have been employed in the past or not. For example, Saint John, N.B. and St. John’s, NFLD are considered to be connected via seaport, where existing shipping routes exist; Thunder Bay, ON and Sault Ste. Marie, ON are considered to be connected by seaport, even though they are not currently serviced by marine tanker; however Thunder Bay and St. John’s are not considered to be linked via seaport, since it is not feasible for marine tankers to navigate the sea route between the cities, due to their size and scale, even though both cities have accessible ports.
The distances to the nearest terminals are calculated by the shortest highway distance provided by Mapquest, as well, using the terminal locations provided by Petroleum industry suppliers, like Esso.[31]
References
Adams, A. F. (1997) “Search Costs and Price Dispersion in a Localized, Homogeneous Product Market: Some Empirical Evidence,” Review of Industrial Organization, 12(5–6):801–808.10.1023/A:1007779422525Search in Google Scholar
Burdett, K. and K. L. Judd (1983) “Equilibrium Price Dispersion,” Econometrica, 51(4):955–969.10.2307/1912045Search in Google Scholar
Canadian Energy Pipeline Association. Available at: http://www.cepa.com/.Search in Google Scholar
Ceglowski, J. (2003) “The Law of One Price: Intranational Evidence for Canada,” The Canadian Journal of Economics, 36(2):373–400.10.1111/1540-5982.t01-1-00005Search in Google Scholar
Chandra, A. and M. Tappata (2011) “Consumer Search and Dynamic Price Dispersion: An Application to Gasoline Markets,” The RAND Journal of Economics, 42(4):681–704.10.1111/j.1756-2171.2011.00150.xSearch in Google Scholar
Crucini, M. and H. Yilmazkuday (2014) “Understanding Long-Run Price Dispersion,” Journal of Monetary Economics, 66:226–240.10.1016/j.jmoneco.2014.04.012Search in Google Scholar
Crucini, M., C. Telmer and M. Zachariadis (2003) “Price Dispersion: The Role of Borders, Distance and Location,” Tepper School of Business, Paper 490.Search in Google Scholar
Crucini, M., M. Shintani and T. Tsurugu (2015) “Noisy Information, Distance and the Law of One Price Dynamics Across US Cities,” Journal of Monetary Economics, 74(C):52–66.10.1016/j.jmoneco.2015.06.003Search in Google Scholar
Eckert, A. (2013) “Empirical Studies of Gasoline Retailing: A Guide to the Literature,” Journal of Economic Surveys, 27(1):140–166.10.1111/j.1467-6419.2011.00698.xSearch in Google Scholar
Engel, Charles M. and J. Rogers (1996) “How Wide Is the Border?” American Economics Review, 86(5):1112–1125.10.3386/w4829Search in Google Scholar
Feyrer, J., E. Mansur and B. Sacerdote (2017) “Geographic Dispersion of Economic Shocks: Evidence from the Fracking Revolution,” The American Economic Review, 107(4):1313–1334.10.1257/aer.20151326Search in Google Scholar
Kano, K., T. Kano and K. Takechi (2013) “Exagerrated Death of Distance: Revisiting Distance Effects on Regional Price Dispersions,” Journal of International Economics, 90(2):403–413.10.1016/j.jinteco.2013.02.002Search in Google Scholar
Kent Marketing Services, The Kent Group. “Petroleum Price Data,” Available at: http://www.kentmarketingservices.com/dnn/PetroleumPriceData.aspx.Search in Google Scholar
Lewis, M. (2011) “Asymmetric Price Adjustment and Consumer Search: An Examination of the Retail Gasoline Market,” Journal of Economics & Management Strategy, 20(2):409–449.10.1111/j.1530-9134.2011.00293.xSearch in Google Scholar
Marvel, H. (1976) “The Economics of Information and Retail Gasoline Price Behavior: An Empirical Analysis,” Journal of Political Economy, 84(5):1033–1060.10.1086/260495Search in Google Scholar
Mills, J., J. Schmitz and G. Frizelle (2004) “A Strategic Review of Supply Networks,” International Journal of Operations & Production Management, 24(10):1012–1036.10.1108/01443570410558058Search in Google Scholar
Nagurney, A., J. Dong and D. Zhang (2002). “A Supply Chain Network Equilibrium Model,” Transportation Research Part E: Logistics and Transportation Review, 38(5):281–303.10.1016/S1366-5545(01)00020-5Search in Google Scholar
Natural Resources Canada. (2019a). “Infrastructure for Crude Oil and Petroleum Products,” Available at: http://www.nrcan.gc.ca/energy/sources/infrastructure/1390.Search in Google Scholar
Natural Resources Canada. (2019b). “Trans Mountain Expansion Project,” Available at: http://www.nrcan.gc.ca/energy/resources/19142.Search in Google Scholar
Parsley, D. C. and S. J. Wei (1996). “Convergence to the law of one price without trade barriers or currency fluctuations,” The Quarterly Journal of Economics, 111(4):1211–1236.10.2307/2946713Search in Google Scholar
Pennerstorfer, D., P. Schmidt-Dengler, N. Schutz, C. Weiss and B. Yontcheva (2015) “Information and Price Dispersion: Theory and Evidence,” CEPR Discussion Paper No. DP10771.Search in Google Scholar
Shen, Z. J. M. (2006) “A Profit-Maximizing Supply Chain Network Design Model with Demand Choice Flexibility,” Operations Research Letters, 34(6):673–682.10.1016/j.orl.2005.10.006Search in Google Scholar
Stahl, K. (1982). “Differentiated Products, Consumer Search, and Locational Oligopoly,” The Journal of Industrial Economics, 31(1/2):97–113.10.2307/2098007Search in Google Scholar
Statistics Canada. (2017a). “Household Spending,” Table 11-10-0222-01 Canada, regions and provinces.Search in Google Scholar
Statistics Canada. (2017b). “Inter-provincial transfers, Imports, Exports, Supply and Disposition of Refined Petroleum Products,” Table 134-0004.Search in Google Scholar
Stigler, G. (1961) “The Economics of Information,” Journal of Political Economy, 69(3):213–225.10.1086/258464Search in Google Scholar
Yilmazkuday, H. (2014) “Price Dispersion Across U.S. Districts of Entry,” Economic Letters, 123(3):361–365.10.1016/j.econlet.2014.03.018Search in Google Scholar
Yilmazkuday, H. (2017) “Geographical Dispersion of Consumer Search Behaviour.” Applied Economics, Taylor & Francis Journals, 49(57):5740–5752, December.10.1080/00036846.2017.1340575Search in Google Scholar
Yilmazkuday, H. (2018) “Spatial Dispersion of Retail Margins: Evidence from Turkish Agricultural Prices,” Agricultural Economics, 49(4):455–462.10.1111/agec.12428Search in Google Scholar
Yilmazkuday, D. and H. Yilmazkuday (2016) “Understanding Gasoline Price Dispersion,” The Annals of Regional Science, 57(1):223–252.10.1007/s00168-016-0775-4Search in Google Scholar
Yilmazkuday, D. and H. Yilmazkuday (2019) “Redistributive Effects of Gasoline Prices,” Networks and Spatial Economics, forthcoming.10.1007/s11067-018-9435-9Search in Google Scholar
©2018 Walter de Gruyter GmbH, Berlin/Boston
