The committee was asked to review “the performance of the nation’s major railroads regarding service levels, service quality, and rates” and “the projected demand for freight transportation over the next two decades and the constraints limiting the railroads’ ability to meet that demand.” These three issues—recent trends in rail rates, service quality issues, and concerns about future capacity constraints—are examined in this chapter. Sampled railroad waybill data are analyzed, the complaints and concerns expressed by shippers about service problems are surveyed, and railroad freight demand forecasts and projections of long-range capacity constraints are reviewed.
Shortly after the original congressional request for this study in 2005,1 the Surface Transportation Board (STB) sponsored an independent economic study of the postderegulation freight railroad industry by Laurits R. Christensen Associates, Inc. (2009a; 2009b; 2010). The Christensen Associates’ reports examine railroad rates, service levels, and capacity issues from the 1980s through 2008. In view of their relevance to the topics reviewed in this chapter, results from the Christensen Associates reports are discussed in several places.
This section examines recent rail rate trends and patterns on an industrywide basis with regard to a selection of commodities and to shipments moved in common and contract carriage. The analysis period for the
1 Public Law 109-59 §9007. Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users.
most part is 2000 to 2012 or 2013.2 Rates are measured in average revenue per ton-mile (RPTM) on the basis of railroad revenues from both common (tariff) and contract carriage, as derived from shipment waybills sampled by STB. STB’s annual Carload Waybill Sample (CWS) program is explained in Box 2-1. The analyses that follow consist mostly of simple indices and cross-tabulations and are intended for background. The results yield insights relevant to the review of regulatory policies in subsequent chapters, such as trends in commodities moved in unregulated contract and regulated common carriage.
Historical Trends in Industrywide Average Rates
Figure 2-1 shows several indices presented in inflation-adjusted (real dollar) terms. Two of the indices show rate levels for 1989 to 2007–2008. They were originally constructed by STB and Laurits R. Christensen Associates, and both use revenue and ton-mile data obtained from the confidential version of the CWS containing actual contract revenue.3 Although they were constructed in slightly different ways, the two indices use a chain-weighting technique to adjust for annual variability in the mix of freight. Changes in average RPTM are measured and proportionately weighted for subgroups of traffic that share characteristics, such as commodity type, length of haul, shipment size (number of carloads), and rail car ownership.4 The third rate index (labeled “NAS”) was developed by the study committee for the more recent period 2002 to 2013. It too
2 Some of the analyses end in 2012 and others end in 2013 because of the timing of the committee’s receipt of 2013 Carload Waybill Sample data late in the study period.
3 Laurits R. Christensen Associates (2010, Chapter 2, Table 2-1) provides two indices for the period 1989 to 2008, one constructed from freight revenue and another constructed from freight and miscellaneous revenues combined. For simplicity, only the former index is shown in Figure 2-1. STB’s rate index (STB 2009) used 1985 as the base year, which the committee rebaselined to 1989 for comparison with the Christensen Associates index. STB’s rate study and documentation on the development of its index can be found at http://www.stb.dot.gov/stb/industry/1985-2007RailroadRateStudy.pdf.
4 The indices were developed by using a chain-weighting technique known as the Törnqvist method. An explanation of the advantages of chain-weighted indices over fixed-weighted indices is given by Diewert (1976).
STB’s CWS PROGRAM
A freight waybill is a document issued by a carrier giving details and instructions relating to the shipment. The document normally contains the names of the consignor and consignee and the shipment’s commodity, origin, destination, and route. Most freight waybills are maintained in electronic form. In this study, each railroad waybill is referred to as a “shipment.” A single waybill shipment can consist of one carload or an entire trainload (100 or more carloads). Because of the need for railroads to interchange traffic, the Association of American Railroads (AAR) has developed standard waybill data elements and forms, including rules for electronic data interchange.
STB requires all railroads that terminate 4,500 or more carloads to sample their waybills and report the sample on a monthly or quarterly basis, depending on traffic activity. Sampling rates vary between 2.5 and 50 percent, depending on the number of carloads in the shipment. Shipments consisting of one or two carloads are sampled at the lower rate, and shipments of 101 or more carloads are sampled at the higher rate. Other sampling rates apply to shipments with 3 to 15, 16 to 60, and 61 to 100 carloads (8.3, 25, and 33.3 percent, respectively). The sampled waybills are submitted in electronic form to a private contractor, Railinc Corporation, which processes and corrects errors in the records under contract with STB and the Federal Railroad Administration (FRA).
During processing, additional information is paired with the sampled record such as details on the rail car (e.g., capacity, dimensions, and mechanical characteristics) and location identifiers (e.g., census region, station zip code, standard production location code). The processed records, typically numbering more than 500,000 for a year, thus contain a range of information on the shipment, including routing, billed tons, miles traversed, revenue, origin, destination, interchange points, railroads traversed, car type, car ownership (e.g., railroad or private), and commodity. Commodity type is recorded by using the U.S. Department of Commerce’s Standard Transportation Commodity Codes (STCCs). STTCs are two- to seven-digit codes, with the first two digits corresponding to major commodity groups and each additional digit a refinement (e.g., 01 = farm products, 011 = field crops, 0113 = grain,
STB’s CWS PROGRAM
Expansion factors are applied to each record to estimate the total population of similar shipments. The expansion factor is the inverse of the sampling rates (e.g., each shipment consisting of one or two cars is multiplied by 40). Other data added to CWS records include STB’s estimate of each railroad’s “variable cost” for transporting each shipment. The costing system used for these calculations, the Uniform Railroad Costing System (URCS), is discussed in more detail in Chapter 3. URCS calculates a variable cost for each shipment on the basis of railroad accounting and operating data and cost apportionment methods that take into consideration characteristics of the shipment such as commodity, number of carloads, number of railroads involved, and rail car type.
Because of the law’s contract confidentiality restrictions, railroads do not submit the actual revenue data for waybills involving contract shipments. They submit encrypted data that only STB can decode. The encrypted records are marked with a flag indicating that revenue data are “calculated.” When the processed waybills are delivered by Railinc to STB, the agency replaces the encrypted revenue data with the actual revenues but restricts access to the databases according to federal regulations (49 CFR 1244.9). Some nonrevenue fields are treated as confidential as well, including the origin and termination freight stations, junction points, and rail carrier identifications. The CWS is therefore released in versions that contain different masked and unmasked fields, with the most restricted version having masked contract revenues. STB also releases the CWS in a public version stripped of all confidential fields. In this version, and in other confidential versions that mask contract revenue, STB replaces the actual contract revenues with revenues that would have been generated by the public tariff rate.
STB uses the CWS for various purposes, including special studies and creation of the Railroad Cost Adjustment Factor (see text). Federal and state agencies use the CWS for transportation planning, as do transportation practitioners, consultants, and law firms with formal proceedings before STB. They must apply for access to the confidential versions of the CWS. All of the analyses using the CWS in this study were conducted with the confidential version containing unmasked contract revenue data.
FIGURE 2-1 Historical and recent trends in industrywide real average rate and unit input costs, 1989–2013. [Source: NAS rate data from CWS 2002–2013; STB rate index from STB 2009; Christensen rate index from Laurits R. Christensen Associates 2010; RCAF-A from Laurits R. Christensen Associates 2010 and AAR (various years). Adjusted for inflation by using U.S. Bureau of Economic Analysis (BEA) gross domestic product (GDP) chained price deflator (http://www.bea.gov/national/index.htm).]
was based on the confidential CWS and chain-weighting according to the method used by STB.5
Accompanying the three rate indices is STB’s productivity-adjusted Railroad Cost Adjustment Factor (RCAF-A) index,6 which summarizes changes in the price of major railroad operating inputs, including labor,
5 The updated rate trend was developed by using the same chain-weighting method that STB used. In this analysis and all others in this chapter, no CWS records were filtered or excluded, but revenue reported as “miscellaneous” during 2003–2007 was added to the freight revenue field to account for misreporting of fuel surcharge revenues.
6 RCAF-A summarizes railroad unit-of-output costs, representing the net effect of input price changes and productivity changes. A description of the components and construction of the RCAF-A, which is developed according to STB rules, can be found at the websites of STB (http://www.stb.dot.gov/stb/industry/rcaf.html) and the Association of American Railroads (https://www.aar.org/data-center/rail-cost-indexes). The base year 1989 for the RCAF-A index shown in Figure 2-1 was constructed by Laurits R. Christensen Associates (2010), and subsequent years were updated and made consistent across periodic changes in base years.
fuel, materials, equipment rents, and interest on debt. In recognition that railroads do not use inputs in fixed amounts and ratios from year to year, the RCAF-A adjusts for variations in the mix and quantity of inputs per unit of output because of productivity changes, such as reductions in fuel, materials, and labor per ton-mile.
Figure 2-1 shows that even when inflation is taken into account, rate levels declined during the 1990s. Real rates fell on average by 30 percent during the decade, driven by a steep decline in input costs per unit of output. This development, as indicated by the RCAF-A, was largely the result of productivity gains, such as the conversion to larger cars and consolidation of traffic in multicar shipments and unit trains of 100 or more cars. The productivity improvements after deregulation, including the shedding of low-volume branch lines and the restructuring of labor agreements, were discussed in Chapter 1.
However, as can be seen in Figure 2-1, starting in the period 2001–2003 rates began to rise 1 to 3 percent per year in nominal terms and then in real terms before jumping markedly from 2005 to 2008. During the early part of the decade, input costs, as summarized by the RCAF-A, began to stabilize after their long secular decline, only to become more volatile after 2004.
Recent Trends in Industrywide Average Rates
Because the focus of this study is on recent developments, the remainder of the section focuses on post-2000 rate trends and patterns. As noted above, from 2001 to 2003 there was a break in the downward movement in real rates that had commenced in the 1980s. Therefore, Figure 2-2 shows trends since that breakpoint through a rebaselining to 2002 and the addition of a trend line for ton-miles to indicate changes in traffic volumes.
The rebaselined rate index shows that real rates rose 27 percent from 2002 to 2013, but with periods of volatility. Two of the many potential reasons for the rate increase are the growth in ton-miles in an industry that had been shedding excess capacity for years and the slight growth in input costs. The post-2006 volatility in rates, input costs, and demand complicates the assessment of secular trends. The freight railroad industry experienced a sharp decline in traffic after 2006 as
FIGURE 2-2 Recent trends in ton-miles and real industrywide average rate and unit input costs, 2002–2013. [Source: NAS rate data and ton-miles are from the CWS 2002–2013; RCAF-A data are from AAR (various years). Adjusted for inflation by using BEA’s GDP chained price deflator (http://www.bea.gov/national/index.htm).]
the national economic recession took hold from 2007 to 2009. Sharp swings in fuel prices have also occurred since 2006; for example, the average price of a gallon of diesel fuel dropped from $3.12 in 2008 to $1.77 in 2009 and then jumped back to more than $3 by 2011 [AAR various years (2014, 63)]. Fluctuations in rates tended to parallel the fluctuations in input costs and in demand from 2006 to 2011. Since 2011, the industry has experienced relative stability in input costs and demand, which may have contributed to steady rates.
Rates by Commodity and Selected Shipment Characteristics
As shown in Table 2-1, more than three-fourths of shipments tendered in 2012 involved commodities and car types ruled exempt from
SOURCE: CWS 2012.
common carrier regulation.7 Examples of exempt shipments are intermodal containers, general merchandise, and fresh fruits and vegetables carried in refrigerated boxcars. These shipments are deregulated mainly because they can be transported competitively by truck. They are often made in single carloads, which explains their large percentage share of railroad shipments. In comparison, coal, grain, minerals, and other bulk commodities are usually shipped in multiple carloads. The latter commodities are not often suited to long-haul movement by truck and remain subject to regulation except when they are carried by contract; thus, they are referred to in this report as “nonexempt.” Shipments of nonexempt commodities, including those moved by contract, account for a much larger share of railroad traffic than shipments of exempt commodities in terms of ton-miles. In 2012, nonexempt commodities accounted for more than two-thirds of ton-miles.
Figure 2-3 shows recent trends in the average rate (RPTM) for several major commodities, including exempt intermodal containers and nonexempt coal, chemicals, grain (e.g., corn, wheat, oats), and oilseeds
7 The discussion refers to traffic as having been ruled exempt from common carrier regulation as distinct from traffic that is temporarily exempt by virtue of contracting. The latter traffic is not ruled exempt because it reverts to common carrier regulatory status on expiration of the relevant contract. While the law does not state explicitly that truck-competitive traffic should be ruled exempt, the ability to be moved competitively by truck is the practical reason for most exemption rulings by commodity and car type. The reasoning is that if a commodity can be moved competitively by truck, it has effective competition in all markets, because trucks are ubiquitous.
FIGURE 2-3 Trends in nominal rates by commodity, 2002-2013.
(Source: CWS 2002–2013.)
(e.g., soybeans, sunflower seeds). Coal, grain, and oilseed shipments have traditionally had the lowest average rates among commodity groups because of their high density, large shipment volumes, long-haul movements, and ease of loading and unloading. Nonexempt commodities moved by contract are included in the nonexempt category because they are not ruled exempt and can move in and out of common carriage over time.
The average rate for all commodities, converted to an index and adjusted for inflation, is shown in Figure 2-4. As noted earlier, it rose by more than 25 percent from 2002 to 2013. Rates for coal and grain and oilseeds grew fastest, up by nearly 50 and 40 percent, respectively.
Examination of Coal
Laurits R. Christensen Associates (2010, 6-2) surmised that a contributor to the faster growth in coal rates after 2005 was the expiration
FIGURE 2-4 Index of trends in real rates by commodity, 2002–2013. [Source: CWS 2002–2013. Adjusted for inflation by using BEA’s GDP chained price deflator (http://www.bea.gov/national/index.htm).]
of many long-term, or “legacy,” contracts and their renegotiation at higher rates as volatile diesel fuel prices were pushing up rates generally. The Christensen Associates report also presented data showing that fewer opportunities existed for productivity gains to offset higher fuel prices—for example, through further consolidation of coal shipments in dedicated coal trains. By 2000, more than 95 percent of coal ton-miles were moving in shipments of 50 carloads or more. This high percentage could increase only marginally, as it did. It rose to 99 percent by 2013.8
Table 2-2 shows changes in coal traffic from 2000 to 2012 on the basis of shipments and carloads as well as ton-miles. The tabulations distinguish between coal shipments originating in the East and the
8 On the basis of the committee’s review of the 2002 and 2013 CWS.
|Origin||Shipments||Carloads||Ton-Miles (millions)||Revenue (current $, thousands)||Average Distance per Shipment (miles)||Carloads per Shipment||Tons per Carload||Percentage of Carloads by Private Car||Average RPTM (current $)|
|Percentage Change, 2000–2012|
NOTE: “West” includes all coal shipment origins in states west of the Mississippi River. Rates and revenues are not adjusted for inflation.
SOURCE: CWS 2000 and 2012.
West, because they differ substantially in character. The data show a continuing decline in coal volumes in the East, where average rates remained substantially higher than in the West, partly because of shorter average hauls, less private car use, and smaller shipment sizes. Coal rates in both regions increased during the period, but more in the East. The tabulations also show how tens of thousands of smaller shipments, each averaging relatively few ton-miles, were consolidated into larger shipments after 2000. This pattern is exhibited in both regions, but most dramatically in the East, where smaller shipments had been the norm. As discussed in more detail below, the consolidation of coal traffic occurred as most coal shippers were leaving common carrier service for contract service, which likely accelerated the pace of shipment consolidation.
Examination of Grain and Oilseeds
In contrast to the situation for coal, substantial room remained for consolidation of grain and oilseeds traffic in 2002, because shipments of 50 or more carloads accounted for only half of ton-miles. This share increased to 65 percent in 2013, perhaps prompted by the growing gap in rates between smaller- and larger-carload shipments, as shown in Figure 2-5. For grain and oilseeds shipments that continued to move in smaller-carload shipments (less than 50 carloads), the average rate had become 35 percent higher than the average rate paid by larger-carload shippers (Figure 2-5 and Figure 2-6).
Consistent with the trend toward traffic consolidation, railroads tried to use their grain car fleets more efficiently and to encourage larger shipments through volume incentives, the auctioning of rail cars in large allocations (e.g., 24- or 40-car deliveries), and the promotion of 75- to 110-car train shuttle services (Wilson and Dahl 2005; Prader et al. 2013).9 As a result of these efforts, railroad-owned grain cars expanded their share of ton-miles relative to private cars from 65 to 80 percent between 2002 and 2013 (Figure 2-7). Accompanying this development,
9 Grain train shuttle services typically involve a dedicated set of 75 to 110 cars that move as a unit from a single origin to a single destination. The shuttle operator and the railroad enter into a contract to move the train on a continuous basis for a specific time, generally 1 year. The shuttle operator is provided incentives for the commitment, and grain elevators are provided incentives to accommodate the trains and to have the capability for fast loading.
FIGURE 2-5 Trends in ton-mile shares and rate differentials for small shipments of grain and oilseeds, 2002–2013. The black line is the average RPTM for traffic moved in shipments of less than 50 cars divided by the average RPTM for shipments of 50 or more cars, expressed as a percentage minus 100 percent (e.g., in 2013, shipments of less than 50 cars had an average RPTM 135 percent of that of shipments involving 50 more cars; this is shown as a 35 percent positive rate differential). (Source: CWS 2002–2013.)
FIGURE 2-6 Average rates (nominal) paid for smaller and larger shipments of grain and oilseeds, 2002–2013. (Source: CWS 2002–2013.)
FIGURE 2-7 Share of ton-miles by private- and railroad (RR)-owned cars and their rate differentials, grain and oilseeds, 2002–2013. (Source: CWS 2002–2013.)
and likely related to it, was the erosion of the long-standing rate premium for shipments using railroad-owned grain cars (Figure 2-7).
The declining premium for railroad-owned grain cars had been observed earlier by STB (2009).10 In 2009, the average rate for service in railroad-owned cars actually fell below that of private cars. The reason for this change is not apparent from the aggregated data. It could be the result of many factors, such as differences in the propensity of private- and railroad-owned cars to be used in contract versus common carriage, for larger versus smaller shipments, and for longer versus shorter hauls. These other factors, more than differences in equipment costs, may have contributed to the declining gap in average rates for service in private- and railroad-owned cars.
10 According to STB (2009), average rates for service using railroad-owned grain cars were consistently 10 to 20 percent higher than average rates for service using private grain cars during the 1990s; however, the premium had nearly disappeared by 2000.
None of the recent developments, all of which would seemingly favor lower rates (i.e., further shipment consolidation, dedicated trains), explains why rates for shipments of grain and oilseeds rose faster than rates for shipments of other commodities except coal (see Figure 2-4). For example, rates for small shipments and for larger shipments of grain and oilseeds increased by more than 80 percent and by nearly 70 percent, respectively, from 2002 to 2013. Furthermore, grain shippers—unlike coal shippers in the early 2000s—were not major users of contract carriage. Increases in grain shipping rates, therefore, are less likely to have been exacerbated by the expiration of low-rate legacy contracts, as may have occurred for shippers of coal.
The data in Table 2-3 show how grain and oilseeds traffic has been consolidating over the period. Total shipments went down relative to carloads. With the exception of oilseeds, the total volume shipped was lower in 2012 than in 2000, which likely reflects yearly fluctuations in harvests and grain export demand. The level of grain consolidation does not approach that of the more concentrated coal segment. Grain shippers include many elevators dispersed across a farming region, whereas coal mines deliver large and predictable volumes of coal to electric utilities and ports. Thus, most coal shippers regularly transport large volumes over fixed traffic lanes, but many grain shippers do not.
As discussed in the next section, coal shippers have converted almost exclusively to contract carriage during the past 10 to 15 years. This trend has been accompanied by even more traffic consolidation. Grain shippers have not converted to contracts in large numbers, perhaps because their shipment characteristics and volume fluctuations require more flexibility than contract commitments would allow. In addition, some new features of tariff service emulate contract features. For example, grain shippers can bid for future allotments of rail cars and locked-in tariff rates via auctions.
Contract and Common Carriage Rates
Prevalence of Contract Carriage
More than 30 years after the Staggers Rail Act permitted railroads to negotiate confidential contracts with shippers, contract carriage
|Commodity||Shipments||Carloads||Ton-Miles (millions)||Revenue (current $, thousands)||Average Distance per Shipment (miles)||Carloads per Shipment||Tons per Carload||Percent of Carloads by Private Car||RPTM (current $)|
|Percentage Change, 2000–2012|
NOTE: Small amounts of corn, wheat, and oilseeds traffic that may move in exempt rail cars are not included in the table. Rates and revenues are not adjusted for inflation.
SOURCE: CWS 2002 and 2012.
SOURCE: CWS 2000 and 2012.
has become commonplace among shippers of many nonexempt commodities such as coal, ores, and chemicals.11 In 2012, contract service accounted for more than two-thirds of nonexempt traffic, whether measured in shipments, carloads, ton-miles, or revenues (Table 2-4). Between 2000 and 2012, total ton-miles by common carriage declined by nearly 50 percent. Having started the century as the predominant means by which shippers of nonexempt commodities procured rail service, common carriage had become the minority means by 2012.
11 The CWS records do not state definitively whether a shipment was moved by contract or tariff. A flag in the record indicates whether the revenue data are confidential. Only contract traffic has this confidentiality privilege. However, a railroad is not required to report its contract revenues as confidential. Thus, some contract shipments could be included in the tariff shipments in these tabulations.
The large reduction in common carriage since 2000—part of a general industry trend that commenced years earlier—was largely the result of a rapid migration of coal to contract carriage. As shown in Table 2-5, coal ton-miles were split evenly between common and contract carriage in 2000. By 2012, only 5 percent of coal ton-miles were moved in common carriage. During this period many chemical shippers also switched to contract service. Contract carriage accounted for only 35 percent of the chemical (non–hazardous materials) ton-miles in 2000 but for 77 percent in 2012. Shippers of several other nonexempt commodities having shipment characteristics similar to those of chemicals or coal, including hazardous materials, petroleum products, ores, and stone, also became majority users of contract service by 2012.
Despite the general shift toward contracting over the past decade, shippers of some commodities, particularly agricultural commodities, have remained users of common carriage. Contract carriage grew dramatically among coal and chemical shippers but barely changed among shippers of corn and wheat. As a consequence, grain has become the largest commodity grouping in common carrier service. It has far surpassed coal to account (along with oilseeds) for about one-third of all ton-miles by tariff (Figure 2-8).
Contract Versus Common Carrier Tariff Rates
Contract and tariff rates for the same commodities can be compared by using the revenue-unmasked version of STB’s CWS. However, all contract terms, including service periods, incentives, and performance requirements, are not recorded in the CWS. Thus, the comparisons may not be valid. For example, the sampled waybills do not indicate whether a contracting railroad agreed to a lower rate in return for a traffic guarantee or whether a contracting shipper agreed to a higher rate in return for a level of service standard. Furthermore, a contract rate may reflect the demand and supply conditions that existed years before a particular shipment was tendered, while a tariff rate is more likely to reflect momentary conditions, similar to a spot rate.
The preceding qualifications should be kept in mind in reviewing Table 2-6, in which average rates and traffic characteristics for tariff and contract shipments among the major nonexempt commodities during 2012 are compared. Except for the slightly lower contract rates
|2000||2012||Percent Change, 2000–2012|
|Ton-Miles (billions)||Percent Contract Carriage||Ton-Miles (billions)||Percent Contract Carriage||In Ton-Miles||In Contract Ton-Miles|
|Chemicals (excluding hazardous materials)||103||35||106||77||3||126|
|Food and kindred products||60||48||86||39||43||17|
|Lumber, wood products||44||52||36||45||−18||−29|
|Stone, clay, glass||30||49||26||63||−13||11|
|Petroleum and coal products||18||19||15||70||−17||207|
NOTE: Ton-miles are measured for movements made in nonexempt rail cars only; for example, ton-miles of hazardous materials transported in boxcars are not included in the tabulations.
SOURCE: CWS 2002 and 2012.
|Commodity||Shipments||Carloads||Ton-Miles (millions)||Revenue (current $, thousands)||Average Distance per Shipment (miles)||Carloads per Shipment||Tons per Carload||Percent of Carloads by Private Car||RPTM (current $)|
|Chemicals (non–hazardous materials)||135,482||198,625||17,800||890,836||852||1.5||96||79||0.050|
|Commodity||Shipments||Carloads||Ton-Miles (millions)||Revenue (current $, thousands)||Average Distance per Shipment (miles)||Carloads per Shipment||Tons per Carload||Percent of Carloads by Private Car||RPTM (current $)|
|Chemicals (non–hazardous materials)||906,915||1,089,801||87,826||3,953,489||828||1.2||98||93||0.045|
NOTE: CWS 2012.
for most commodities, the comparison does not indicate any consistent patterns. For example, in the case of coal, tariff shipments tend to move shorter distances than contract shipments, but the reverse is true for most other commodities. Grain is moved in larger shipments in common carriage than in contract carriage, while the opposite is true for coal. Reasons for these differences have already been discussed. Among them are the addition by railroads of features resembling those found in contracts to common carrier grain service, such as the use of multicar discounts and rate locks. The substantially higher tariff rate for coal may be an artifact of the small number of coal shippers who have remained in common carriage; many of their shipments may move intermittently or have special transportation demands that preclude contractual commitments or are more costly to provide.
Table 2-7 shows average rates for tariff and contract traffic by commodity in 2000 and 2012. With the major exceptions of coal and chemicals, contract rates for most nonexempt commodities were 5 to 20 percent lower than tariff rates in 2000. However, contract rates rose faster than tariff rates. By 2012, the gap between tariff and contract rates had closed substantially for most commodities, with the exception of coal.
Summary of Recent Rate Trends
During the 1990s, both railroad rates and input costs experienced a secular decline, which reflected dramatic improvements in railroad productivity. Although the railroad industry has been characterized by volatility in rates, input costs, and demand in recent years, real rates rose by more than 25 percent from 2002 (when real rates reached their low point) to 2013. Rates grew nearly twice as fast as ton-miles and far exceeded growth in input costs, which exhibited periods of volatility but rose in real terms by 2 percent. The observed volatility in input prices may have led to higher renewed contract rates to account for uncertainty, particularly in fuel prices.
Real rates increased for all commodities from 2002 to 2013, with most increases between 15 and 25 percent. Among major commodities, coal rates grew the fastest (up nearly 50 percent) followed by grain rates (up nearly 40 percent). Consolidation of both coal and grain into
|Name||2000||2012||Percent Change, 2000–2012|
|Chemicals (non–hazardous materials)||0.033||0.033||0.033||0.047||0.051||0.048||42||55||45|
|Food and kindred products||0.023||0.026||0.025||0.038||0.041||0.040||66||57||62|
|Lumber, wood products||0.026||0.027||0.026||0.042||0.039||0.041||61||44||55|
|Stone, glass, clay||0.032||0.036||0.034||0.058||0.054||0.057||81||48||65|
|Petroleum and coal products||0.032||0.036||0.035||0.060||0.052||0.058||87||44||63|
NOTE: Rates are not adjusted for inflation.
SOURCE: CWS 2000 and 2012.
larger shipments has continued. In the past, consolidation had been a major contributor to growth in industry productivity and downward pressure on rates, but this effect is not evident in recent rate trends.
During the 2000s, shippers who had previously used common carriage continued to shift to contracting. In 2012, 75 percent of all nonexempt ton-miles were moved in contract carriage, compared with 44 percent in 2000. Coal shippers have turned almost exclusively to contract carriage. Shippers of grain remain the most committed to common carriage, with only 1 in 5 ton-miles being moved by contract. Characteristics of grain shipments, including irregularity in volumes and routings, may make this traffic less suited to contractual commitments than commodities with regular routings such as coal. However, railroads have added features to grain common carrier service that are characteristic of contracts, including the ability to reserve cars through auctions.
The continued reliance by grain shippers on common carriage, which is the only form of rail transportation that remains subject to direct regulation, has implications for STB rate and service oversight responsibilities that are discussed in the next two chapters.
As the preceding section makes clear, there is substantial information on railroad traffic and revenues, much of it derived from STB’s CWS. However, shipment-level data for evaluating or benchmarking railroad service quality do not exist.
Apart from requiring railroads to report and publish aggregated statistics on train operations and car fleet status, STB does not sample or require the reporting of shipment-specific data pertaining to aspects of service performance such as delivery times or speeds. Hence, service trends and patterns cannot be examined with as much precision as rates, as Congress requested of this study. All that proved to be practical is a survey of shipper commentary about service quality. It was assembled mainly from past STB hearings, reports in the Christensen Associates study, and the study committee’s meeting with shipper organizations (see Preface). That survey is given next and is followed by a summary of explanations offered by railroads for episodic
Because information on service quality is largely anecdotal, the record must be considered with caution. There is a consistency to shipper complaints and there are commonalities in their timing and location, which suggest some problematic service levels and particular time periods when problems have been exacerbated. However, regulatory hearings about service quality are bound to attract dissatisfied shippers, and there will always be some who are dissatisfied. STB hearings do not necessarily gauge the satisfaction levels of other shippers who may be content or who can only express their dissatisfaction with service through the enforcement of contracts in the courts.
Shippers remark that without explicit standards for common carrier service and means of monitoring railroad performance, they can do no more than use STB as a sounding board for service complaints. A review of comments in service-related hearings by STB and Congress, trade publication survey results, and statements by shippers invited to brief the committee indicates shipper desires that regulators do more to compel railroads to improve common carrier service.12
In late 2013 and early 2014, STB received many complaints from shippers, especially coal, grain, and automobile shippers, who reported widespread disturbances in rail service, including an inability to obtain service, lengthy delays in transit, and unusually long rail car cycle times.13 As a result of these problems, shippers reported significant inventory backlogs and shortages of materials, including fertilizer for crops and coal for electricity generation. Some also expressed concerns
12 See STB Ex Parte No. 677 and No. 724 for many shipper comments on service. As explained in the Preface, the committee asked shipper groups who had previously submitted comments in STB service-related hearings to brief members on their concerns. The public briefings were held in conjunction with the committee’s second meeting on March 14, 2014 (slides are available at http://www.trb.org/PolicyStudies/RailTransReg.aspx).
13 A rail car “cycle” is the time required for loading a rail car with product at a shipper’s origin, transporting it to the consignee, unloading it, and transporting it back to the shipper for reloading.
about a lack of communications and timely information by railroads about service status.14
Shippers cited problems in several regions, but disturbances were particularly severe in the Upper Midwest. For example, grain shippers described their experiences to STB and to the study committee as follows:
The sheer gravity, magnitude, and scope of rail service disruptions now being experienced are unprecedented and have rippled through all sectors of grain-based agriculture. … Another fallout is illustrated in the values paid in the secondary rail car freight market. … The majority of secondary freight has traded at values of approximately $4,000 per car, equating to $1 per bushel.15
The late summer/fall and winter of 2013–2014—has proven to be one of the worst rail service meltdowns in modern history—affecting all classes of traffic but especially the northern plains movements.16
Later in 2014, automobile manufacturers described rail service to a congressional committee:
The greatest logistics problem faced by auto manufacturers is the carriers’ failure to provide a sufficient supply of empty railcars to transport finished vehicles. Automakers have also incurred significant delays in the movement of railcars loaded with finished vehicles. In this regard, it appears that the priority of auto shipping has become less than that of other shippers. … These vehicles should have been transported much sooner via contracted rail services to dealerships. … [E]xtreme weather merely exacerbated underlying problems stemming from a lack of capacity—in cars, as well as crews and locomotive power.17
14 See shipper comments to STB Ex Parte No. 724.
15 National Grain and Feed Association, Statement to STB, April 17, 2014 (http://www.ngfa.org/2014/04/21/ngfa-urges-stb-to-require-reporting-of-service-related-metrics-by-class-icarriers-at-public-hearing-on-rail-disruptions/).
16 T. Whiteside, Alliance for Rail Competition (Montana-based shipper organization involving many wheat shippers), presentation to the committee, March 14, 2014 (http://www.trb.org/PolicyStudies/RailTransReg.aspx).
17 S. Karr, Statement of the Alliance of Automobile Manufacturers Before the Committee on Commerce, Science, and Transportation, September 10, 2014.
The Western Coal Traffic League (WCTL), an organization of coal-burning electric utilities, reported concerns to the study committee expressed by the organization’s members that slow deliveries of coal will jeopardize the reliability of electricity generation.18 In a petition to STB, it stated the following:
The limited coal deliveries and the uncertainty of adequate future deliveries have caused most of the WCTL membership to curtail coal-fired production. These curtailments have forced the utilities to seek alternative generation at significantly higher costs, which in turn has cost electric consumers and ratepayers hundreds of millions of dollars.19
In October 2014, WCTL petitioned STB to require one railroad (BNSF) “to submit a coal-specific service recovery plan, which the Board should then review, approve or revise, and, most importantly, enforce.” The recovery plan would specify coal delivery standards to be met, and STB enforcement could include fines for failure to meet deliveries.20 STB declared the continuation of rail service performance throughout the national system to be a priority and ordered the railroad to provide a detailed description of the contingency plans it would use to mitigate an acute coal inventory shortage at key generating stations in each region.21
The periodic surveys conducted by research firms and trade publications are supplemental sources of information on rail shipper satisfaction. Their industrywide representativeness is difficult to gauge, but their coverage may be broader than the record of complaints from industry groups to STB and Congress. Survey results reported in the trade press indicate a low level of shipper satisfaction with rail service during 2013 and 2014 and a general perception of deteriorating service over time.22
18 D. Jaffe, WCTL, presentation to the committee, March 14, 2014 (http://onlinepubs.trb.org/onlinepubs/railtransreg/Jaffe031414.pdf).
19 WCTL, STB Ex Parte No. 724—United States Rail Service Issues. Petition of the Western Coal Traffic League for an Order Requiring BNSF Railway Company to Submit a Coal Service Recovery Plan, October 22, 2014.
20 WCTL, STB Ex Parte No. 724.
21 STB Ex Parte No. 724, December 3, 2014.
22 The survey, conducted by Wolfe Research, found that 70 percent of respondent shippers had experienced what they believed were capacity-related service problems during 2013–2014 (Szakonyi 2014).
- In 1997, while the operations of the merged Union Pacific (UP) and Southern Pacific Railroads were being integrated, western rail shippers experienced extraordinary service delays as congestion at certain terminals spread into a systemwide problem. STB intervened by ordering UP to release certain shippers from contracts and to cooperate with other railroads in relieving congestion (GAO 1999, 67).
- In 1999, while Norfolk Southern and CSX were merging the operations of the disbanded Conrail, shippers experienced delays in obtaining service and in transit times (GAO 1999).
- In 2004, during a period of rapid growth in container and other rail freight traffic, the Southern California seaports experienced severe congestion that was attributed to lack of rail capacity for the transportation of arriving containers as well as to port capacity constraints. Rail shippers complained of degraded service in other regions at the same time (CBO 2005, 1–3; Lavigne 2014).
Shipper expressions of concern in 2013–2014 have precedent in the earlier periods of tight capacity. Some of the circumstances in 2004 resembled those of 2013–2014 as growth in traffic resumed after the recession. Although capacity was strained in 2004, the Congressional Budget Office (CBO), asked by Congress to examine the causes of service disturbances, concluded that “the feared ‘meltdown’ of service had not materialized” because the railroads were able to take action to mitigate the effect of the surge in traffic on service (CBO 2005, 2). CBO found that the railroads were able to expand capacity by hiring new workers and adding equipment, changing routings to reduce congestion, and managing traffic demand patterns by selectively raising rates during the year. CBO noted that the large share of rail traffic moving by long-term contract rates presumably constrains the railroads’ ability to adjust rates in response to sudden increases in traffic but concluded that “the BLS [freight rail industry] price index data suggest that railroads have been able to raise rates within the terms of their contracts” (CBO 2005, 12).
STB commissioned Christensen Associates, as part of its freight rail competition study, to interview shippers in various sectors of the rail freight market to solicit opinions on railroad capacity, rates, competition, and service quality. The interviews were conducted from November 2007 through August 2008 (Laurits R. Christensen Associates 2009a, 5-4–5-7). According to the authors’ summary, service-related themes that emerged from the interviews included the following (Laurits R. Christensen Associates 2009a, 5-12–5-13, 18-30):
- A sense that service quality was declining but somewhat improved in 2008 compared with the 2004 congestion episode;
- A belief that the variability in delivery times had increased, leading to larger shipper inventories, the need for more rail cars, and the need for more shipper personnel to manage shipping;
- Claims that railroads having market dominance lacked motivation to provide good service and that shippers could not negotiate contracts with standards for service accountability; and
- A view that tight capacity was a primary contributor to service problems and that disturbances arising at local chokepoints were a main cause of the disruptions propagating through railroad networks lacking slack capacity.
The railroads have responded to the public complaints made by shippers concerning service quality and to STB inquiries. They have generally maintained that they cannot fully predict changing economic conditions that lead to abrupt changes in demand, nor can they predict and fully prepare for exogenous factors such as extreme weather.23 They contend that adjustments are made to networks and operations as quickly as possible when unusual events arise. The railroads are less definitive in explaining when circumstances make it more beneficial for them to maintain some slack capacity to avoid service disruptions, but they emphasize the impracticality and high cost of maintaining extra capacity as a standard means of quickly recovering from rare events.
23 AAR comments to STB Ex Parte No. 724, April 17, 2014.
a confluence of events that have affected rail service in particular regions of the country. These included a historically harsh winter that forced railroads to dramatically shorten train lengths and limit crew exposure to the elements; a record grain harvest and unexpected surge in grain exports; and higher coal volumes as utilities sought to replenish stockpiles consumed when generating additional electricity that this winter demanded.24
In another statement, AAR disputed reports that the traffic delays during that period were caused “by increased demand to move any one commodity or product,” apparently a reference to the rapid growth in petroleum tank car movements. It emphasized instead “a surge in demand to transport a mix of more and more commodities and products, … something that neither railroads nor their customers anticipated.”25
Traffic data in 2013 and 2014 are consistent with the railroad industry’s description of unanticipated market conditions, especially surging demand. Railroad traffic growth, measured in originating carloads, was flat during 2013, and volumes of coal and grain experienced declines (Table 2-8). However, traffic rebounded in 2014. Grain traffic increased as a result of larger harvests and rising exports in 2014, and petroleum traffic continued its rapid growth. Both commodities account for modest shares of total U.S. rail freight traffic but are important in the Upper Midwest, where shippers described especially severe service disturbances.
In responding to complaints that railroads are slow to invest in the capacity needed to forestall service disruptions, AAR presented data to the study committee showing how trends in rail carloads, excluding coal and grain, have aligned closely with trends in output from the U.S. manufacturing sector.26 Railroads maintain that they are like most other businesses in being required to react to broader, economywide
24 AAR Weekly Rail Traffic Report: Week 52. January 2, 2014.
25 As the U.S. Economy Roars Back to Life, Freight Rail Provides the Ride. E. Hamberger, AAR, advertisement in the Washington Post, November 19, 2014, p. A15.
26 J. Gray, AAR, presentation to the committee, March 14, 2014, Slide 5. http://onlinepubs.trb.org/onlinepubs/railtransreg/Gray031414.pdf.
TABLE 2-8 Total Carload Origins and Origins for Selected Commodities in 2014, with Percentage Change from 2012 and 2013, Class I Railroads
2012 to 2013
2013 to 2014
|Petroleum and products||
SOURCE: AAR Weekly Rail Traffic Report: Week 52, Jan. 2, 2014, and Week 52, Jan. 2, 2013.
It is beyond question that the railroad industry is committed to making investments in the network designed to meet the demand for rail service now and in the future. The nation’s freight railroads project that they will spend approximately $26 billion this year to build, maintain, and upgrade their nationwide rail network. Railroads also expect to hire more than 12,000 people in 2014.27
Railroads have emphasized the need to keep making such capacity investments to ensure adequate service availability and performance and have consistently claimed that regulatory interventions can make matters worse by suppressing capital spending. For example, in its comments to the April 2014 STB hearing on service quality, AAR stated:28
In order for railroads to continue to invest at levels necessary to meet increasing demand for rail service, a necessary predicate is a regulatory
27 AAR comments to STB Ex Parte No. 724, April 17, 2014, p. 2.
28 AAR comments to STB Ex Parte No. 724, April 17, 2014, p. 3.
In briefing the committee, AAR presented 2014 data indicating that railroads had increased their expenditures on roadway, structures, and equipment by 28 percent since the low point of the 2007–2009 recession.29
Whether the service expectations of shippers and railroads are the same is unclear. As discussed later in this report, the development of a common set of expectations about service quality, particularly for common carrier service, must start with improvements in data for assessing service levels and monitoring performance. That information base is limited, as is evident from the review of service-related data given next.
Aggregate Data on Service Quality
Shippers and the railroads have recognized a need for better data on freight railroad service performance that can be collected and published in a timely manner. Better data could aid shippers in planning for and coping with transportation conditions, reinforce the railroads’ accountability, and help regulators evaluate shipper complaints. In response to the merger-related service disturbances of 1997–1999, AAR has published a weekly series of railroad performance measures (RPM) for each Class I railroad that includes the number of cars on line by car type and by owner (the railroad on which the car is located, another railroad, or a nonrailroad), average train speeds (for five train categories: intermodal, manifest, multilevel, coal unit, and grain unit), and terminal dwell times.30
29 J. Gray, AAR, presentation to the committee, March 14, 2014, Slide 38. http://onlinepubs.trb.org/onlinepubs/railtransreg/Gray031414.pdf.
30 Train speed measures the line-haul movement between terminals. The average speed is calculated by dividing train-miles by total hours operated, excluding yard and local trains, passenger trains, maintenance of way trains, and terminal time. Terminal dwell time is the average number of hours a car resides at the specified terminal location expressed in hours. The measurement begins with a customer release, received interchange, or train arrival event and ends with a customer placement (actual or constructive), delivered or offered in interchange, or train departure event. Cars that move through a terminal on a run-through train are excluded, as are stored, bad ordered, and maintenance of way cars.
FIGURE 2-9 Average train speeds (mph), BNSF Railway, October 2013 to September 2014. (Source: AAR, Railroad Performance Measures, http://www.railroadpm.org.)
As an example of the potential application of the RPM data in gauging service performance, Figure 2-9 shows average BNSF train speeds from October 2013 to September 2014. BNSF data were chosen because many customer complaints about service disturbances during 2013–2014 were concentrated in the Upper Midwest, where BNSF provides much of the service. An inspection of the 2013–2014 data does not immediately suggest service disturbances as described by shippers for the period and largely acknowledged by railroads to have been problematic.31 Average train speeds fluctuated by a few miles per hour from month to month, with a downward drift for all train types. The data are too coarse to make meaningful determinations about service quality.
Dwell time data are more indicative of service disturbances, as shown in Figure 2-10. A prominent feature in the 2013–2014 series is a spike in average dwell times in the Northtown, Minnesota, terminal during late winter 2014, when dwell times averaged 60 to 75 hours com-
31 BNSF comments to STB Ex Parte No. 724, September 13, 2014.
FIGURE 2-10 Average terminal dwell times for BNSF trains, October 2013 to September 2014. (Source: AAR, Railroad Performance Measures, http://www.railroadpm.org.)
pared with a system average of 30 to 35 hours during the overall period. However, for the most part the extent of aggregation of the RPM data obscures any meaningful insight into the types and degree of service quality problems experienced by shippers. Furthermore, an estimate of how long shipments took to move between any two particular points cannot be derived from the two data series, and neither sheds any light on how long shippers at various locations had to wait for rail cars.
In its evaluation of the RPM data, Laurits R. Christensen Associates (2009a, 17–19; 2009b, 2-31–2-34) reached the same conclusion: average train speed and dwell time data are too gross to offer more than a rough indication of service performance. For example, it calculated correlations of changes in real GDP with changes in dwell time, cars on line, and train speed by railroad during the period 2006–2008 and found that the measures did not consistently change in the expected direction. Christensen Associates also pointed out that the performance features of greatest concern to shippers, such as route-specific or corridor-specific information on on-time performance and the variability of performance, are not part of the measurement system.
In response to shipper service complaints from the winter of 2013–2014, in October 2014 STB issued a temporary order requiring all Class I railroads to report weekly performance data.32 The required data include the standard RPM measures of train speed, terminal dwell times, and cars on line, as well as measures of the following:
- Unit train origin dwell time, by train type;
- Trains held short of destination, by train type and cause;
- Loaded and empty cars in service that have not moved in more than 48 hours, by car type;
- Grain cars loaded, by state and by commodity;
- Past due car orders;
- Grain shuttle round-trips;
- Coal unit train loadings; and
- Car counts at Chicago yards and numbers of trains held for delivery to Chicago (for railroads operating at Chicago).
The order states that the reporting “will give the agency and stakeholders access to data needed for real-time understanding of regional and national service issues” and cites shipper contentions that “performance metrics are important for rail users to plan logistics, minimize economic harm to operations and revenues, assist with business planning, and to better serve their own customers during the service recovery period.”33
STB issued a notice of proposed rulemaking in December 2014 that would make such weekly reporting requirements permanent and that would modify the data specified in the October temporary order. The notice describes the value of the reporting as follows:34
32 STB Ex Parte No. 724-3: United States Rail Service Issues—Data Collection, October 8, 2014; STB Ex Parte 724-4.
33 STB Ex Parte No. 724-3: United States Rail Service Issues—Data Collection, October 8.
34 STB Ex Parte No. 724-4: United States Rail Service Issues—Performance Data Reporting, December 30, 2014.
The permanent collection of performance data on a weekly basis would … improve the Board’s ability to identify and help resolve future regional or national service disruptions more quickly, should they occur. Transparency would also benefit rail shippers and other stakeholders, by helping them to better plan operations and make informed decisions based on publicly available, near real-time data, and their own analysis of performance trends over time.
The committee does not know how these efforts will proceed, because the data collection proposal was introduced during the course of the study. However, the data to be collected are not specific with regard to shipment or even to origin and destination (with the exception of unit train data) in the same manner as are the on-time arrival data collected for many years by the U.S. Department of Transportation (USDOT) for airlines. Furthermore, the proposed collection effort appears to be an ad hoc response to the disturbances of the previous winter; it does not appear to have been strategically devised in the sense of there being a plan for routine use of the information in monitoring performance.
Summary of Service Quality Issues
STB maintains a waybill sampling program that allows the monitoring of railroad traffic and rates at the shipment level. However, it does not collect comparable shipment-specific records for monitoring the performance of railroads in carrying out their common carrier duty of providing adequate service. Trends in service reliability, speed, and other aspects of performance must be identified from a largely anecdotal record of shipper complaints. The record suggests that railroad service is interrupted at intervals by disturbances that arise when traffic volumes escalate unexpectedly and outpace the railroads’ deployment of capacity. Service during the winter of 2013–2014 was particularly problematic for this reason and was made worse by severe weather.
Whether service problems during such episodes are more severe or whether reliability is routinely inferior for common carrier traffic cannot be ascertained from the complaint records or by assessing the aggregated service-related data collected by STB. The complaint record is naturally
skewed toward shippers of common carriage because only their service is regulated. Better service-related data at the shipment level, for both common and contract carriage, would allow more objective analysis of common carrier service quality, particularly to evaluate whether this service is chronically substandard and how it changes relative to that of contract carriage when capacity is tight.
The committee was asked to examine “the projected demand for freight transportation over the next two decades and the constraints limiting the railroads’ ability to meet that demand.” This section begins with a review of two long-range freight rail volume forecasts that are made regularly by the federal government and notes three studies undertaken in recent years to assess the potential effects of forecast freight growth on railroad capacity needs. Consideration is then given to economic factors tending to motivate railroads to supply capacity, which are typically neglected in studies that project freight volumes and predict capacity shortages. The section concludes with a brief review of federal programs intended to make investments in railroad capacity more attractive.
Forecasts of Rail Freight Traffic and Transportation Capacity Needs
Federal Freight Demand Forecasts
Commonly cited freight traffic forecasting series are USDOT’s Freight Analysis Framework (FAF)35 and the Energy Information Administration’s (EIA’s) National Energy Modeling System (NEMS). FAF is intended to aid transportation investment planning and policy analysis, while NEMS is used by EIA to produce the Annual Energy Outlook (AEO) series, which supports energy program planning and policy making (EIA 2014a). Both project traffic for all freight transportation
FIGURE 2-11 FAF: freight ton-miles by mode and selected rail commodities (actual for 1997–2012 and forecast for 2012–2035). Truck and railroad ton-miles do not include shipments that move by multiple modes. [Source: Freight Analysis Framework Data Tabulation Tool (http://faf.ornl.gov/fafweb/Extraction1.aspx), accessed February 5, 2015.]
modes on the basis of the assumption that transportation capacity does not constrain growth.
The FAF freight projections are derived from a proprietary economic model that produces regional projections of growth by industry and projections of the resulting freight traffic on the basis of a matrix of historical interregional flows by commodity and mode. The projections are made under the assumption that all of the needed transportation capacity will be available and deployed regardless of cost (i.e., no capacity constraints) (FHWA 2012, 4–8). The most recent FAF projections indicate an average annual growth rate from 2012 to 2035 of 1.3 percent for rail freight ton-miles, compared with 2.6 percent for truck ton-miles and 2.9 percent for multiple-mode shipment ton-miles, most of which are on railroads but include truck and water movements (Figure 2-11).36 FAF also projects railroad ton-miles by commodity; for
36 USDOT states that ton-miles cannot be separated by mode for multiple-mode shipments in the historical data source it uses (the U.S. Census Bureau’s Commodity Flow Survey) (FHWA 2012, 6).
example, coal ton-miles are projected to decline at an average annual rate of 1.2 percent, whereas grain ton-miles are forecast to grow at 5.7 percent annually over the forecast period (Figure 2-11).
EIA’s NEMS model projects rail freight demand by multiplying projected industrial output for the individual commodities in each Census division by a set of constant ton-mile-per-dollar coefficients (EIA 2014b, 80). EIA publishes NEMS projections in its AEO series according to high-, mid-, and low-range economic growth assumptions (Figure 2-12). The 2014 mid-range scenario, used for most purposes, shows railroad freight ton-miles in 2025 unchanged from 2012, after a recovery from depressed levels from 2013 to 2016, a short-term trend that now appears improbable (Figure 2-12). Ton-miles in 2025 are 9 percent higher in the high-range scenario than in the low-range one.
Freight ton-mile projections can vary widely from one year’s AEO edition to the next. Projections are sensitive to near-term economic conditions and often result in depressed or exaggerated extrapolations, depending on when they were made in the business cycle. The 2014 edition’s high-range projection is lower in all future years than the 2013 edition’s mid-range projection (Figure 2-12). The 2014 AEO edition’s ton-mile mid-range (reference) projection for 2025 is 23 percent
FIGURE 2-12 AEO: rail ton-miles, historical and projected.
(Source: EIA 2013, Table A7; EIA 2014a, Table A7.)
FIGURE 2-13 U.S. rail ton-miles projected for 2015 and 2025 in AEO reference cases by publication year of projection. (Source: AEO, various years.)
below the mid-range projection for 2025 made in the 2007 AEO edition, which was published just as the economy was entering a recession (Figure 2-13). In view of these discrepancies, the forecasts for future years will probably be even farther from the actual values. In Table 2-9, the 10-year (2015–2025) NEMS and FAF forecasts are contrasted with the actual ton-mile growth from 1990 to 2000 and 2000 to 2010. The results suggest that forecasts are heavily influenced by recent trends in traffic growth.
Projections of Long-Range Capacity Shortages
Freight output forecasts such as NEMS and FAF are sometimes used to inform studies of long-range transportation investment needs. Such studies are more common during business cycle peaks, when there are perceptions of tightening capacity and freight volume projections tend to produce exaggerated trend extrapolations, as noted above. The last round of prominent studies of freight capacity needs coincided with the peak in railroad ton-miles (2006), and they were released at
|Source||Period||Actual or Forecast||Percent Change in Ton-Miles|
|Class I railroadsa||1990–2000||Actual||42|
|Class I railroadsa||2000–2010||Actual||15|
|FAF (excluding multimodal)b||2015–2025||Forecast||15|
|AEO (NEMS) reference casec||2015–2025||Forecast||10|
NOTE: The increases in real GDP from 1990 to 2000 and from 2000 to 2010 were 40 and 18 percent, respectively, which closely parallels the Class I railroad ton-mile increases (http://www.bea.gov/national/index.htm).
a AAR various years (2014).
c EIA 2014a.
the commencement of a recession that would quickly quiet concerns about capacity shortages (as ton-miles fell by more than 10 percent). Three prominent studies all forecast significant railroad investment and capacity gaps by 2020 to 2035:
- Cambridge Systematics (2007), whose study was sponsored by AAR, estimated a $39 billion gap between railroads’ capabilities and capital spending for the infrastructure required to accommodate traffic growth and maintain service between 2007 and 2035.
- The National Surface Transportation Policy and Revenue Study Commission (NSTPRSC 2007) estimated a $1 billion to $3 billion annual gap between sustainable capital spending by the railroad industry and investment required to improve performance for 2008 to 2020 (the report utilized elements of the AAR-sponsored study cited above).
- The American Association of State Highway and Transportation Officials (AASHTO 2007) estimated an annual gap of $3 billion to
- $4 billion between railroads’ investment capabilities and economically justified capital spending from 2007 to 2027.
These studies imply that shippers would be unwilling to pay for or railroads would be incapable of financing all the railroad capacity that would be economically and socially beneficial to provide. The source and size of the purported gaps are computed in different ways. In some cases, assumptions are made about the level of capacity needed to meet a specific quality of service target under different freight growth scenarios, and the investment shortage is calculated on the basis of projected railroad revenue and a fixed ratio of investment to earnings (NSTPRSC 2007, Vol. 1, 5–6; Vol. 2, 4–17). Candidate causes for the underinvestment are alluded to but seldom defined. Among them are the influence of regulations specific to railroads (including economic regulation and railroad labor laws), government subsidies to trucking and barge transportation (e.g., public provision of highways and waterways), and external benefits (e.g., pollution and congestion reduction) of shifting freight from highways to rail.
Such capacity-needs projections have substantial weaknesses. One is the assumption that railroads invest in fixed proportion to their earnings. The studies tend naively to treat the demand for freight transportation and the supply of capacity as largely exogenous. They discount or neglect the incremental profits that railroads can generate from capacity investments and fail to explain in a convincing manner why railroads would sacrifice profits by letting large capacity gaps persist. The predictions of capacity gaps are often accompanied by policy proposals to make rail investments more attractive. However, the proposals fail to compare alternatives for correcting or compensating for the supposed causes of rail underinvestment, such as improved pricing of the public facilities used by the competing modes; pollution charges; and cost-reducing truck, rail, barge, and pipeline regulatory reforms.
Rail Capacity Supply Incentives
A common shortcoming of studies assessing future rail capacity needs, as exemplified by those cited above, is that they seldom define what constitutes capacity. The capacity of a transportation network can be
difficult to define and even more difficult to measure. It cannot be characterized simply in terms of a maximum rate of throughput of some aggregate measure such as trips or ton-miles traveled. The ability of a freight network to carry any specified quantity of traffic will depend on the distribution of origins and destinations of the shipments, the temporal pattern of shipments, and shipper preferences with regard to speed and reliability. As traffic on a rail system grows, congestion is likely to begin to appear at local chokepoints, which may be in terminals or heavily used segments of main line. As traffic continues to grow, localized congestion may spread until systemwide service problems arise. The optimum level of congestion depends on the value that each shipper places on speed and reliability and on the cost to the railroad of mitigating congestion through physical expansion, asset redeployments, and refinements in operating practices. The response may be to add more infrastructure, equipment, and workers. However, railroads may also respond by changing routings and schedules; increasing productivity through technological improvements in infrastructure, equipment, and operations; and rationing demand through pricing.
As discussed in Chapter 1, the ability of railroads to discriminate on the basis of price through contracting allows them to set rates that do not price profitable traffic out of the market and thus to avoid systematic underinvestment. Both the railroad and the shipper have an economic interest in reaching agreements ensuring that no profitable traffic goes unserved.37 In this regard, a qualitative definition of a rail capacity shortage might be a circumstance in which a shipper or group of shippers are paying, or willing to pay, a rate that generates revenue sufficient to cover the cost to the railroad of improving speed and reliability, but improvements, for whatever reason, are not forthcoming over some protracted period.
One reason for a protracted capacity shortage might be that the railroad lacks access to credit markets. For example, a railroad that is financially weak may not be able to raise investment capital generally and thus not be able to add capacity even in individual markets in which
37 The incentive to add capacity to accommodate all profitable traffic is apparent as railroads respond to the increasing demand for transportation of crude oil in the Upper Midwest. This demand did not exist less than a decade ago.
additions would be profitable. Another possibility is that some shippers may not be able to commit to a contract, perhaps because they lack a sufficiently regular service need. Normally, a railroad that posts a high tariff rate would negotiate contracts with shippers having a lower willingness to pay and thus ensure that all profitable traffic is served. If a shipper is unable to contract in these cases, it may not be served; however, both the railroad and the shipper will have a strong incentive to bargain, so that the scenario of an unmet demand ought to be relatively rare.
In light of these profit incentives, protracted capacity shortages would not be expected. However, capacity provision may not be smooth or well targeted in the short term. The observed pattern of increasing shipper complaints at the start of economic expansions suggests that there can be lags in capacity deployment and investment during upticks in demand, especially if the new demand is viewed as short-lived or simply outpaces the physical ability of a railroad to respond. Investments involving the addition of fixed infrastructure can be “lumpy” and difficult to target precisely and quickly. Railroads must also make choices that minimize opportunity costs when temporary constraints arise. Accordingly, a railroad may price some traffic that is normally profitable out of the market; however, service for that traffic should resume over the longer run when capacity adjustments can be made. In addition, there is no guarantee that a shipper will have its traffic transported from its preferred location. The reason is that railroads make pricing and investment decisions in a network environment in which interdependent demands affect where railroads add capacity. Shippers with a high willingness to pay may be compelled to ship from alternative locations where traffic is concentrated; the trend toward consolidation of grain-loading facilities to serve shuttle trains (as noted earlier) is one example.
Thus, the price-discriminating capabilities of railroads should preclude prolonged and significant underinvestments in capacity. However, the profit-maximizing calculus of railroads and shippers will not necessarily lead to resource allocations that are desirable from a societal perspective when externalities are factored in. Studies of future capacity needs are often accompanied by policy recommendations that would make investments in rail capacity more attractive to reduce
externalities such as highway congestion and emissions from freight moved by truck. As discussed next, such externalities are a common justification for government programs to attract more capital to the railroad industry.
Public Policy and Railroad Capacity
Congress has authorized a number of government programs that can be used by the freight railroads and that are intended to increase the attractiveness of investing in rail capacity. They include two credit assistance programs (the Railroad Rehabilitation and Improvement Financing and Transportation Infrastructure Finance and Innovation Act programs) and a discretionary grant program (the Transportation Investment Generating Economic Recovery program) administered by FRA and USDOT. In addition, Congress funds occasional projects specifically to aid railroads with capacity investments. An example is the Heartland Corridor project, which was administered by the Federal Highway Administration. That project increased tunnel clearances for trains moving double-stacked intermodal containers between Chicago and Norfolk, Virginia. Its purpose was to reduce the number of containers moved by truck on the public highways. While all of these programs play a minor role in overall rail freight capital funding, they are examples of public-sector efforts intended to guide rail freight investments toward perceived public interests.38
FRA, whose primary responsibility is to regulate railroad safety, also regards itself as responsible for promoting socially beneficial investments in rail freight transportation. For example, the FRA website states the following: “To meet the needs of the current and future freight rail industry and to maximize the benefits of public investments, FRA is committed to supporting current freight rail market share and growth and developing strategies to attract 50 percent of all shipments 500 miles or greater to intermodal rail.”39 FRA’s 2010 National Rail Plan
38 For example, the Obama administration’s Fiscal Year 2015 federal surface transportation program reauthorization proposal called for $10 billion in spending over 4 years on road, rail, and port projects to relieve freight bottlenecks (http://www.whitehouse.gov/sites/default/files/omb/budget/fy2015/assets/transportation.pdf).
Progress Report identifies two goals related to freight capacity: “support the current freight rail market share and growth” and “develop strategies to attract 50 percent of all shipments 500 miles or greater to intermodal rail” (FRA 2010, 14). As a rationale for its pursuit of these goals, FRA states that greater use of rail freight will bring about lower casualty rates, shipper cost savings, reductions in energy consumption and pollutant emissions, and reduced highway congestion and infrastructure costs.40
Summary of Long-Run Capacity Issues
Railroads maintain that service disturbances do not indicate chronic underinvestment in capacity. Instead, they are a temporary phenomenon arising from a short-run inability to adjust supply, which can cause traffic to move slowly and some normally profitable traffic to go unserved. Nevertheless, concerns about railroads falling short of the investments required to handle future growth in freight traffic were prevalent before the recent recession. At that time, the railroad industry’s networks had been made lean, traffic had been steadily growing, and forecasts of rapid traffic growth had become exaggerated by the postderegulation volume peak. Predictions of capacity gaps were often dire but were seldom accompanied by explanations of why the profit motive of railroads would allow such a suboptimal outcome to persist over time periods in which adjustments can be made. A profit-maximizing railroad that can access credit markets (i.e., that is revenue adequate) and can price according to its customers’ willingness to pay should generally have the ability and incentive to deploy and invest in the capacity required to move all profit-generating traffic. The profit incentive should oppose any large and protracted capacity shortfalls.
However, the profit motive by itself may not produce an equilibrium rail output that maximizes public welfare when externalities are considered. Forecasts of long-run capacity shortages seldom distinguish between valid concerns about railroads underinvesting in the capacity needed to handle socially optimal traffic and more questionable
40 S. Greene, FRA, presentation to the study committee, May 29, 2014 (http://www.trb.org/PolicyStudies/RailTransReg.aspx).
concerns about railroads underinvesting in the capacity required to handle all profitable traffic. Shifting freight from truck to rail may create positive externalities, such as reductions in air pollution or highway congestion, that neither carriers nor shippers will take into account. In this sense, railroads may fall short in supplying welfare-maximizing levels of rail capacity, and policy interventions may be warranted to fill the gap. However, that possibility was not examined in this study. It concerns issues and requires analyses that are outside the study charge and that are better suited to a multimodal study of national freight policy.
The main points from this chapter that are discussed in the summary assessment of the final chapter are given in the following paragraphs.
Freight rail rates declined for more than two decades after the railroad industry was largely deregulated, but real rates increased over the past decade and gains in productivity slowed. Since 2007, the railroad industry has been characterized by volatility in rates, input costs, and demand. During the past decade, average rates for coal and grain have grown the fastest, for reasons that could not be established in this chapter. Contract carriage is now predominant for many bulk commodities, including coal and chemicals. Grain shippers continue to rely mostly on common carriage and represent the largest user group of this service.
Shippers have repeatedly raised concerns about the reliability and general quality of freight rail service, particularly common carrier service. Complaint levels tend to be highest during periods of sharply rising demand and have been exacerbated by bouts of severe weather. Because STB only regulates common carrier service, it largely receives complaints only from this segment of traffic. Thus, whether common carrier service is chronically inferior to contract service or whether it suffers more when capacity is tight is difficult to ascertain. Data pertaining to service quality that are collected by STB are anecdotal and do not allow objective evaluation of service quality trends and the responsiveness of railroads to their common carrier service obligations.
Long-range forecasting of freight rail capacity levels tends to be unreliable. It is affected by the near-term business cycle and focuses
more on general factors influencing freight demand and less on those influencing service supply. While short-term capacity shortages can be expected, the reasoning offered for anticipated long-term shortages is often vague. Profit-maximizing railroads should be expected to supply all the capacity needed to transport profitable traffic over the longer term. Whether the resulting rail freight volumes are welfare-maximizing when the external costs of freight transportation are considered is another matter. Government programs exist to help make more freight profitable for railroads to move and thus to shift traffic away from trucks out of concern over highway congestion, safety, and emissions. Whether these programs are effective and justified was deemed to be outside this study’s scope.
|AAR||Association of American Railroads|
|AASHTO||American Association of State Highway and Transportation Officials|
|CBO||Congressional Budget Office|
|EIA||Energy Information Administration|
|FHWA||Federal Highway Administration|
|FRA||Federal Railroad Administration|
|GAO||General Accounting Office|
|NSTPRSC||National Surface Transportation Policy and Revenue Study Commission|
|STB||Surface Transportation Board|
|TRB||Transportation Research Board|
AAR. Various years. Railroad Facts. Washington, D.C.
AASHTO. 2007. Transportation: Invest in America: Freight-Rail Bottom Line Report. Washington, D.C.
Cambridge Systematics. 2007. National Rail Freight Infrastructure Capacity and Investment Study. Prepared for AAR, Sept. Cambridge, Mass.
CBO. 2005. Freight Rail Transportation: A Review of the 2004 Experience. Washington, D.C.
Diewert, W. E. 1976. Exact and Superlative Index Numbers. Journal of Econometrics, Vol. 4, pp. 115–145.
EIA. 2014a. Annual Energy Outlook 2014. U.S. Department of Energy. May.
EIA. 2014b. Assumptions to the Annual Energy Outlook 2014. U.S. Department of Energy.
FHWA. 2012. Freight Analysis Framework: User Guide. U.S. Department of Transportation, Washington, D.C., June.
FRA. 2010. National Rail Plan: Moving Forward: A Progress Report. Sept.
GAO. 1999. Railroad Regulation: Changes in Rates and Service Quality Since 1990.
Laurits R. Christensen Associates. 2009a. A Study of Competition in the U.S. Freight Railroad Industry and Analysis of Proposals That Might Enhance Competition. Madison, Wis., Nov.
Laurits R. Christensen Associates. 2009b. Supplemental Report to the U.S. Surface Transportation Board on Capacity and Infrastructure Investment. Madison, Wis., March.
Laurits R. Christensen Associates. 2010. An Update to the Study of Competition in the U.S. Freight Railroad Industry. Madison, Wis., Jan.
Lavigne, G. 2014. Port of Long Beach Says Congestion Has Been Worse. JOC.com, Oct. 23.
NSTPRSC. 2007. Report of the National Surface Transportation Policy and Revenue Study Commission: Transportation for Tomorrow. Dec.
Prader, M., A. Sparger, and D. O’Neil. 2013. The Effects of Increased Shuttle-Train Movements of Grain and Oil Seeds. U.S. Department of Agriculture, Agricultural Marketing Service, Aug.
STB. 2009. Study of Railroad Rates: 1985–2007. Office of Economics, Environmental Analysis, and Administration, Section of Economics. http://www.stb.dot.gov/stb/industry/1985-2007RailroadRateStudy.pdf.
Szakonyi, M. 2014. Poor U.S. Rail Service Limiting Potential Intermodal Volume Growth. JOC.com, Oct. 6. http://www.joc.com/rail-intermodal/intermodalshipping/poor-us-rail-service-limiting-potential-intermodal-volume-growth_20141006.html.
TRB. 2003. Special Report 271: Freight Capacity for the 21st Century. Transportation Research Board of the National Academies, Washington, D.C.
TRB. 2009. Special Report 297: Funding Options for Freight Transportation Projects. Transportation Research Board of the National Academies, Washington, D.C.
Wilson, W. W., and B. L. Dahl. 2005. Railcar Auctions for Grain Shipments: A Strategic Analysis. Journal of Agriculture and Food Industrial Organization, Vol. 3, No. 2, pp. 1–27.