Freight Data Cost Elements (2013)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4Freight cost analyses and the factors that influence them are discussed in this chapter, which seeks to identify the spe- cific types of direct freight transportation cost data elements required for public investment, policy, and regulatory decision- making, and describe and assess different strategies for iden- tifying and obtaining the needed cost data elements. The goal of this analysis is to ensure that the recommended approaches are conceptually valid and in accordance with current best practices and the latest research. The complexities involved in defining “cost” are explored further in Appendix A. Depending on the type of decision being made, definitions of cost may vary. In this context, dis- cussing the various cost definitions ensures a common basis for discussion. 2.1 Freight Costs and External Factors that Impact Them Freight costs are influenced by a number of external factors that impact the resources used in the activity. These external factors include: • Business Models: The freight industry is very heteroge- neous and complex, and encompasses a wide range of modes, users, and business relationships. As such, a multi- tude of business models have been developed that reflect the needs of the various industry segments. Just within the trucking industry, examples include: – Specialized (e.g., auto carriers, hazardous materials, liq- uid cargo, bulk) – Full truckload (FTL) vs. less than truckload (LTL) – Express couriers Some models are found in all freight modes, while others are specific to a given mode. Different models emphasize different aspects of the business and employ various strat- egies to succeed in the marketplace. For example, express carriers emphasize fast deliveries, while small FTL carriers compete on the basis of price. • Private vs. For-Hire Operations: The market orientation of a company could play a role in its freight costs. In the case of private carriers—whose primary mission is to pro- vide ancillary transportation services to a parent or related company—some of the cost components could be shared with other units of the company. This leads to internal cross- subsidization that, in addition to being extremely difficult to ascertain, could change the cost structure. In contrast, for-hire carriers, which provide services to an open market, need to cost out all the key expenses. As a result of these differences, private and for-hire carriers are likely to have dramatically different cost structures that in turn, may lead to different behavioral responses to public-sector policies. • Units: When evaluating across modes, the analysis gener- ally must make the costs consistent. The models usually convert costs per vehicle-mile to costs per ton-mile, which requires data on the amount of a commodity that can fit on a given vehicle type. Given that the volume-to-weight ratio varies across commodities, the tonnage per vehicle type usually varies across commodities. • Line Haul vs. Trip: In comparing the cost of freight trips across modes, the analysis is complicated by the inclusion of these important costs: loading, drayage, transloading, and unloading costs. For all freight modes except truck, a trip often involves multiple modes, so cost estimates must reflect this multi-dimensionality. • Vehicle Type: Costs vary for each configuration of a specific mode. Numerous truck sizes and configurations exist, each with different costs and cargo capacities. Barge sizes and drafts also vary, and trains may carry containers in single or double stacks. • Vehicle Configurations: Multiple configurations of vehi- cles are used in freight movements. Truck movements may be truckload (TL) or less than truckload (LTL); rail freight may be in carloads or unit trains; and barges move in vari- ous tow sizes, carrying commodities in containers, bulk, or break-bulk. C h a p t e r 2 Fundamentals of Freight Cost Estimation and Its Use

5 • Service Area: Vehicle cost will often vary according to ser- vice area. For example, truck costs will be different for short haul versus long haul, or urban versus rural, due to labor costs, speed and fuel usage, and vehicle age. Barge costs will vary depending on up-bound versus down-bound travel, the speed of the current, and the presence of locks. Rail costs vary as well, often dropping with the length of haul. • Overhead: Overhead items include administrative fees, management, advertising, and marketing. Numerous other factors associated with public-sector inter- ventions and market-related factors can impact freight costs. Some examples include: • Fuel Price Volatility: Sudden changes in fuel costs can neg- atively impact freight costs and profits, as most contracts only allow for price revisions once the fuel cost increase reaches a minimum threshold. Increases below that thresh- old are absorbed by the carrier. • Labor Shortages: Since labor is one of the industry’s largest cost components, across all modes, labor shortages have a significant impact on costs. In some cases, regulations that limit labor supplies have the unintended consequence of increasing labor costs. • Congestion Costs and Parking Availability in Urban Areas: It is widely known that urban congestion costs are significant. Cost estimates produced by the research team and corroborated with industry input indicate that, in equal conditions, delivering at night in New York City is 30% cheaper than delivering in the congested daytime hours (Holguín-Veras 2006). The same research reveals that parking fines—an obvious byproduct of lack of suit- able parking—average between $500 and $1,000 per truck per month. • Urban Sprawl: As urban areas grow, land prices in the sur- rounding areas increase, which forces freight transporta- tion companies to locate farther away. An analysis of the business relocation patterns in New Jersey from1990–1999 indicates that the location decisions of freight compa- nies is negatively correlated with the distance to both New York City and Philadelphia (Holguín-Veras et al. 2005). In other words, freight companies tend to locate farther away from these cities. The net effect of this is to increase even further transportation costs to the urban areas. However, while freight companies located far from city centers have increased variable costs, their fixed costs are lower because of lower land prices. • Other Restrictions: The research confirms that restric- tions such as delivery time curfews, truck routes, large truck bans, etc., tend to increase freight costs. While the use of such restrictions may be justified for other reasons, well- documented cases suggest that some of these restrictions are clearly counterproductive. 2.1.1 Classification of Cost Components The main emphasis of this project is on operational costs. Other costs, such as value of time and logistic costs, are not discussed. Operational costs include all cost components that are the direct result of the provision of the service, such as fuel, tires, crew wages, indirect costs, and fringe benefits. 2.1.1.1 Operational Costs “Operational costs” (also called “operating expenses”) are those expenses incurred in the daily running of a business. Operational costs are internal to the carriers and include both fixed and variable costs. From an accounting perspective, “variable costs” are incremental costs that can go up or down based on the amount of business activity or consumption. (By contrast, “fixed” costs do not change depending on the level of activity or consumption.) “Marginal” costs are variable costs, indicating the amount of cost that goes up or down if produc- tion or consumption increases (or decreases) by one unit. Variable and fixed costs may be either “direct” or “indirect.” A direct cost is one that can be directly associated with a spe- cific cost item (e.g., a task, service, or material). Indirect costs cannot be directly tied to a specific cost item. In this context, “indirect” costs refer to the portion of the fixed costs that are adjudicated to the direct cost (typi- cally in the form of a multiplier) so that the resulting pric- ing structure recovers both variable and fixed costs. Unlike direct costs, indirect costs are not affected by consumption. Typical freight operational cost components include the following: • Variable Costs – Fuel, fuel taxes, oil, tires, and depreciation – Maintenance and repair – Crew wages – Travel time – Paid parking and tolls • Fixed Costs – Capital investment (buildings, equipment, land, etc.) – Obsolescence – Insurance – Registration fees 2.2 Freight Cost Estimation Techniques Because the movement of goods is largely handled by private-sector businesses for private-sector customers, cost and rate information can be difficult to obtain. However, estimating the cost of producing a good or service is one of the most important subjects in microeconomics, and cru- cial to freight cost analysis. For that reason, current research

6emphasizes the importance of developing increasingly sophis- ticated estimation techniques. Four factors make freight cost estimation particularly complex: 1. Freight costs depend on the amount of cargo being trans- ported. 2. In transportation, operators frequently provide service to different markets—each being a different output—as part of a multi-output process in which the concept of average cost ceases to have meaning. 3. Especially in regulatory analyses, there may be a need to assess the degree of scope and scale economies, which requires the use of functional forms that do not impose any constraints on these parameters. 4. All of these data may be hard to obtain. This section briefly summarizes the wide range of tech- niques specifically designed to overcome these challenges. Freight cost estimation can be approached from two differ- ent perspectives: accounting and statistical (Pels and Rietveld 2000). Readers interested in more information should consult Pels and Rietveld (2000), Jara-Díaz (2012), and Small (1992). 2.2.1 Activity-Based Costing (ABC) and Other Accounting Techniques Activity-based costing (ABC) methods identify the data cost elements that influence costs, estimate the amounts in which the cost elements enter into the provision of a unit of output, and compute their contribution to the cost. These methods implicitly assume a linear relation between costs and output. In the case of freight, for example, one can identify the vari- able cost components that are related to distance traveled and time traveled, as well as additional fixed costs. Then, by estimating the amounts in which they enter into the trans- portation of a unit of cargo, one can aggregate their contri- butions and obtain the parameters that quantify the cost per unit distance, per unit time, and at a fixed cost. The limi- tations of such approaches include that no consideration is given to substitution effects. This prevents the consideration, for instance, that a carrier who faces increasing fuel costs may purchase a more efficient vehicle. These techniques also assume constant returns to scale, given that they cannot con- sider how costs could decrease or increase depending on the amount of cargo transported. As a result, the use of account- ing techniques is only appropriate for short-term analyses. 2.2.2 Statistical Techniques The second approach to cost estimation involves the use of statistical techniques, of which there are two subgroups: econometric models, in which statistical estimation is com- bined with economic theory to estimate the parameters of cost models, and statistical models that do not use economic theory in the estimation process. In econometric models, the functional form used in the analyses may restrict the analyses and the results that can be obtained (Pels and Rietveld 2000). Recognizing this, econo- metricians make a distinction between “basic” and “flexible” functional forms. The basic functional forms (e.g., Cobb- Douglas, Leontief, and Constant Elasticity of Substitution) impose one or more restrictions on the cost functions. Flex- ible forms impose no restrictions and allow the data to deter- mine the nature of the cost function. This family of models includes quadratic, trans-log, and generalized Leontief mod- els, among others. A drawback of statistical techniques is that they require significantly more data than do accounting techniques. This is particularly true with flexible forms, which have a large number of parameters. This challenge is an important con- sideration in freight cost estimation, as such data may be hard to obtain. Statistical approaches are also needed when cost components are not measurable, either because doing so is too difficult (e.g., because of proprietary data), or because direct observation is impossible. In the latter context, statistical approaches are usually applied to (1) transport and economic modeling, to assess preferences of decisionmakers in the sector to perform freight transportation demand analysis; and (2) non-market valuation in cost-benefit analysis, where not all values are expressed in market prices. Costs measured using statistical techniques include environmental costs (to assess the envi- ronmental impacts of industry decisions) and the value of transport time (to assess the value of services to shippers). 2.3 State of the Practice This section summarizes the state of the practice for freight cost estimation for the various modes of transportation, indicating those approaches currently in use. Professional freight cost estimation techniques range from extremely simple to fairly advanced techniques. Because the state of the practice evolves over time—in most cases toward increased sophistication—the research team included methodologies that represent the highest standard of practice, even when used by a relatively small number of practitioners. Not surprisingly, the literature reflects the varying levels of practice. Table 2.1 and Table 2.2 summarize the key pub- lications on freight cost estimation in the United States and Europe. The tables classify the literature in terms of intended use, mode, type of costs considered, geography of interest, needs area, freight cost estimation technique used, data, and main focus. The publications studied also are broken down by mode. If a publication covers two or more modes, it has been counted

7 Study Pu b cil Pr iv eta R oa d R ai l ri A W at er anoitarep O l L go is it ac l E x ret n eitila s la V ue of itm e ano Z U/l r ab n R eg i ano l N a it ano l l G abo l R eg ul a ito n Po ycil Pl nna in g B A C lacitsitatS C os t tne mele s tagergg A e B e vah i aro l C o detcell ? C o ts s tekra M etar s ATRI (2008) Y Holguín-Veras and Brom (2008) Y Vanegas et al. (2005) Y Litman (2002) N Musso (2001) Colombian Government (2000) Y Litman (1996) N Waters et al. (1995) Y Archondo-Callao et al. (1994) Y Harrison (1990) N Chesher and Harrison (1987) Y Resor and Blaze (2004) Y Resor and Smith (1993) N Meyer and Kraft (1961) N Márquez-Ramos et al. (2010) Y Micco and Pérez (2002) Y Johnsson and Gaier (1998) N Lall et al. (2009) Y Data used Focus area Use Mode Costs considered Geography Needs area FCE Note: FCE refers to freight cost estimation technique; ABC is activity-based costing. Table 2.1. Summary of key literature on freight cost estimation—United States. Study cilbuP etavirP dao R lia R ri A lado mretnI reta W lanoitarep O lacitsigo L seitilanretx E e mit fo eula V nabr U/lano Z lanoige R lanoita N labol G noitaluge R yciloP gninnalP C B A lacitsitatS stne mele tso C etagergg A laroivahe B ?detcello C stso C setar tekra M Ballis and Golias (2002) N Cantos et al. (1999) N Combers and Lafourcade (2005) N NEA (2009) Y NEA (2010) Y European Commission (2002) N Rothengatter et al. (2002) N Boerkamps and van Binsbergen (1999) Y Combes (2010) N Fischer et al. (2005) N Friedrich (2010) Y Jin (2005) N Klaus et al. (2009) Y Mauer (2008) N Tavasszy et al. (1998) Y Data used Focus area Use Mode Costs considered Geography Needs area FCE Note: FCE refers to freight cost estimation technique; ABC is activity-based costing. Table 2.2. Summary of key literature on freight cost estimation—Europe.

8once per mode, and the total counts for the research reflect this adjustment. The publications studied are: NEA(2010), Cantos et al. (1999), NEA (2009), NEA (2004), European Com- mission (2002), Ballis and Golias (2002), Holguín-Veras and Cetin (2008), Ruta (2002), Litman (2002), Litman (1997), Litman (1996), Archondo-Callao et al. (1994), Harrison (1990), Chesher and Harrison (1987), Meyer and Kraft (1961), Short et al. (2010), ATRI (2008), Lall et al. (2009), Holguín-Veras and Brom (2008), Vanegas et al. (2005), Rizet (2003), Holguín-Veras (2003), Musso (2001), Ministry of Transportation of Colom- bia (2000), Combers and Lafourcade (2005), Márquez-Ramos et al. (2010), Micco and Pérez (2002), Resor et al. (2012), Resor and Smith (1993), Waters et al. (1995), and Klaus et al. (2009). Most publications are primarily intended for public-sector use. The predominance of research on highway-related modes is apparent. Out of the 33 unique publications, 22 focus on highway users, 11 on freight rail, 5 on waterways, 4 on air and intermodal transport, and 9 on logistic costs. To complement the information gathered from the published literature, and to gain insight into industry cost estimation practices, the research team conducted a number of in-depth interviews with industry representatives who are familiar with freight costs. Given that most research occurs at the national level, there is a lack of research publications with a local focus that could support local decisionmaking. The majority of these publica- tions seem to have been motivated by planning and policy needs. In terms of freight cost estimation techniques, activity- based costing (ABC) is the preferred choice. Statistical mod- eling was most often used to explain the relationship between freight rates and a set of independent variables.