Trends in Freight Railroad Rates and Competition (GAO-07-292SP), an E-supplement to GAO-07-291R

  • Table of Contents
  • Scope and Methodology

    We used the Surface Transportation Board’s (STB) Carload Waybill Sample to identify railroad rates from 1985 through 2005 (the latest year for which data was available at the time of this review). The Carload Waybill Sample is a sample of railroad waybills (in general, documents prepared from bills of lading that authorize railroads to move shipments and collect freight charges); the sample contains information on rail rates. This STB database includes information on rail rates, tonnage, federal regulation, and other statistics but disguises some revenues to avoid disclosing confidential business information to the public. We obtained a version of the Carload Waybill Sample that did not disguise revenues. We used these data to obtain information on rail rates, car ownership, and other data across the industry, for certain commodities, by state, and for certain routes, by shipment size and length of haul. According to STB officials, data from the Carload Waybill Sample are not adjusted for such factors as year-end rebates and refunds that railroads may provide to shippers.

    We aggregated the rail rate and other data in this report that were derived from this database at a level sufficient to protect confidentiality. Since much of the information contained in the Carload Waybill Sample is confidential, STB disguises the revenues associated with these movements before making this information available to the public. Consistent with GAO’s statutory authority to obtain agency records, we obtained a version of the Carload Waybill Sample that did not disguise revenues associated with railroad movements made under contract. Therefore, the rate analysis presented in this report presents a more complete picture of rail rate trends than analyses that may be based solely on publicly available information.

    We determined that the data used in this report were sufficiently reliable for the purpose of our review. However, during our work we noted anomalous tonnage data estimated by one carrier in 2005 for one commodity (miscellaneous mixed shipments, including intermodal shipments). This carrier reported a significant number of waybill records with a single tonnage value rather than the range of values reported by other carriers and by this carrier in years prior to 2005. This lack of variation caused us to question the reliability of these records and to work with STB officials to investigate further. STB officials stated that the anomaly resulted from a change this carrier instituted in its methodology for estimating the tonnage of certain rail cars and that despite the lack of variation it had no basis for believing these data were in error. STB provided us additional information on this new methodology, however, we remained concerned about the lack of conformity with the reporting practices of other railroads for similar movements. As such, we explored with STB various options for handling these 2005 data, including excluding it from our analysis. However, we determined that because a significant number of waybill records would be excluded, excluding these data would dramatically understate the amount of overall industry tonnage. We also discussed with STB not reporting 2005 data on selected routes and for selected states where the records in question exceeded 50 percent of the tonnage for the particular route or state we examined. However, STB believed that any such action to exclude selective data in this manner represented a significant distortion of the data. We therefore decided to include this 2005 data because despite the lack of variation in the reporting, the average tonnage transported is relatively consistent with prior year data and data reported by other carriers, and thus we believe provides a more accurate estimate than the alternatives of excluding or modifying these data would.

    We used rate indexes and average rates to measure rate changes over time. A rate index attempts to measure price changes over time by holding constant the underlying collection of items that are consumed (in the context of this report, items shipped). This approach differs from comparing average rates in each year because, over time, higher- or lower-priced items can constitute different shares of the items consumed. Comparing average rates can confuse changes in prices with changes in the composition of the goods consumed.

    We placed each corridor in one of three distance-related categories: 0 to 500 miles, 501 to 1,000 miles, and more than 1,000 miles. In the context of railroad transportation, rail rates and revenues per ton-mile are influenced by, among other things, the average length of haul. Therefore, comparisons of average rates over time can be influenced by changes in the mix of long-and short-haul traffic. We judgmentally selected these distance categories to represent reasonable proxies for short-, medium-, and long-distance shipments by rail. In general we reported on the top 25 corridors for the three distance-related categories by commodity tonnages. An exception to this was that we did not report on corridors that did not meet proprietary or sampling requirements for 10 or more years in the 20-year period that we reviewed.

    To examine the rate trends on specific traffic corridors, we first chose a level of geographic aggregation for corridor end points. We defined start and end points as the regional economic areas defined by the Department of Commerce’s Bureau of Economic Analysis. An economic area is a collection of counties in and about a metropolitan area (or other center of economic activity); there are 179 economic areas in the United States, and each of the nation’s 3,141 counties is included in an economic area.

    To determine revenue-to-variable-cost (R/VC) ratios, which when over 180 percent STB regards the traffic as potentially captive, we used data from the Carload Waybill Sample to identify the specific revenues and variable costs and to compute R/VC ratios for the commodities and markets we examined. Using this information, we then identified those commodities and areas whose R/VC ratios were above or below two levels—180 percent (the statutory threshold for challenging a railroad’s rates as unreasonable and considering actions to bring rate relief to potentially captive shippers) and 300 percent (a rate substantially over the statutory threshold for rate relief).

    Although STB regards traffic traveling at rates equal to or above 180 percent R/VC as potentially captive, this must be interpreted with caution because use of R/VC levels as a proxy measure for captivity can understate or overstate captivity. For example, it is possible for the R/VC ratio to increase while the rate paid by a shipper is declining. Assume that in Year 1, a shipper is paying a rate of $20 and the railroad’s variable cost is $12; the R/VC ratio—calculated by dividing the rate by the variable cost—would be 167 percent. If, in Year 2, the variable cost declines by $2 from $12 to $10 and the railroad passes this cost savings directly on to the shipper in the form of a reduced rate, the shipper would pay $18 instead of $20. However, because both the revenue and the variable cost have declined, the R/VC ratio would increase to 180 percent.





    Table of Contents