Logistics Costs and U.S. Gross
Domestic Product

Prepared for:

Federal Highway Administration
Department of Transportation

Prepared by:

MacroSys Research and Technology
Washington, DC

August 25, 2005

Table of Contents

Executive Summary
I. Introduction
II. Some Conceptual Issues
III. CASS Methodology for Calculating Transportation Costs
IV. ENO Methodology for Calculating Transportation Costs
V. Estimates for Recent Years
VI. Assessments and Recommendations
Improved Measure of Logistics Costs
GDP-Comparable Measures of Logistics

---

Logistics Costs and U.S. Gross
Domestic Product

Executive Summary

This report addresses two major issues. They are:

1. Do current metrics used to quantify logistics costs (specifically, the CASS estimate of logistics costs) accurately measure the costs associated with business logistics activities within the United States?
2. Are the frequent comparisons of logistics costs to GDP (namely the practice of stating as a percentage the share of GDP attributable to logistics costs) misrepresentations of what the figures in question represent?

This report concludes that the CASS estimate of logistics costs is generally valid as a measure of business freight logistics costs. The CASS estimate does not include passenger logistics costs associated with business activities. While valid, the CASS methodology could be improved by making four adjustments. First, business related passenger transportation should be included in the CASS estimate. Second, inventory should be measured in constant dollars. Third, inventory levels should be smoothed to account for economic cycles. And finally, interest rates used to estimate inventory capital costs should be held constant.

This report further concludes that the most common comparison between logistics and GDP is misleading. In fact, the most commonly cited estimate of logistics costs is not inherently related to GDP because it is neither a measure of how much of GDP is consumed by logistics nor is it a measure of how much logistics contributes to GDP.

The difficulty in comparing logistics (or any industry) with GDP is in the way "logistics" is classified and calculated. Which elements are included in the estimation of "logistics" can significantly change the meaning of any comparison. In any comparison, GDP is the constant - it always has the same definition. Simply stated, it is the total value of final goods and services produced for consumption within a country's borders in a given time period (usually a year.)

GDP = private consumption + government consumption + investment + net exports
Where: net exports = exports - imports

GDP does not include intermediate goods and services, only final goods and services.

One way to evaluate logistics relative to GDP is to determine how much of GDP is consumed by the logistics activities or the total final demand for logistics by all user categories (consumer, government, business, and net export.) This comparison says nothing of the contribution of the industry to GDP other than to determine the level of final demand that logistics requirements generate. This calculation of logistics must only include purchases that are not consumed by the production of services that ultimately contribute to GDP. This measure of logistics is not currently available. Furthermore, the usefulness of such a measure is not readily apparent.

Another way to evaluate logistics relative to GDP is to determine how much the logistics industry contributes to GDP. To make this comparison one must determine the level of added value generated by the industry. Such a metric would be desirable because it would be a statement of logistics' contribution to GDP. Importantly, intermediate goods are not included in GDP. As a result, this is a difficult comparison to make. One must determine which of the many goods and services purchased by logistics firms are consumed in the production of the goods and services produced by the firms, the difference between the two equals the industry's contribution to GDP. The required calculation is a two-step process. The first is to develop a clear boundary separating logistics activities from other business activities. Within the established boundary (as drawn in this report or as employed in the CASS methodology) the secondary challenge is to estimate value-added from both outsourced logistics activities (i.e., transportation and warehousing industries) and from in-house logistics activities. The transportation element of logistics is measured in this manner by the TSA. However, there is currently no comparable measure of other logistics activities. As such, there is currently no means of calculating the contribution of logistics activities to GDP from a value-added perspective.

Finally, one can calculate fairly easily the amount firms spend on logistics activities (as CASS does) and compare that number to GDP. The problem is that this cost calculation contains 1) intermediate goods and services, and 2) internal business operating costs unrelated to logistics. When firms outsource logistics activities they are purchasing not only the services produced by the logistics providers, but also the intermediate inputs used in the production of the services. When firms run in-house logistics operations, their "logistics costs" also include internal business activities and purchases that are not strictly logistics functions. Furthermore, logistics costs include inventory-carrying costs, which includes opportunity cost of capital, which is not a component of GDP.

Therefore, assuming the most commonly cited estimate of logistics costs:
It is incorrect to say that logistics costs account for X % of GDP.
It is incorrect to say that logistics contributes X % to GDP.
To say that logistics costs are equal to X % of GDP is acceptable, but it is simply a statement of their relative sizes, not a statement of how much one is dependent on the other.

I. Introduction

The Council of Supply Chain Management Professionals (CSCMP), previously known as the Council of Logistics Management, defines logistics as "that part of supply chain management that plans, implements, and controls the efficient, effective forward and reverse flow and storage of goods, services and related information between the point of origin and the point of consumption in order to meet customers' requirements."¹ With certain clarifications and qualifications discussed in the next section, this definition can serve as a general starting point for logistics costs measurement. Broadly consistent with this definition, CASS Information Systems Inc. (formerly CASS Logistics Inc.) has produced annual estimates of national logistics costs for the United States since 1973. The sponsorship of this effort has recently moved from CASS to CSCMP, but the CASS methodology is unlikely to be changed.²

James L. Heskett originally put forward the concept of macroeconomic logistics costs and developed a methodology to measure them.³ However, CASS' estimates, published in its "Annual State of Logistics Report", have been most frequently cited by government agencies and trade associations. As one United Nation's document states, the CASS report "has taken on oracle status in the profession and statistics in it are often cited in federal government reports."⁴ Since the change in sponsorship is unlikely to change the CASS methodology, and since the statistics from applying that methodology will likely continue to be used in highly visible government reports and presentations to justify large public expenditures on freight transportation, the research done in preparing this report sought to give the methodology a rigorous review to ensure that the government is using credible statistics. Furthermore, the CASS logistics statistics are often presented alongside, and compared to the U.S. gross domestic product (GDP). As discussed in detail below, it is imperative when making such comparisons to accurately describe the nature of the relationship between the two measures in order to avoid misinterpretation by the reader and/or audience. Therefore, this report addresses the issue of comparability between the CASS logistics statistics and GDP.

The remainder of the report is structured as follows: Section II raises and examines several basic conceptual issues concerning the measurement of logistics activity and the appropriateness of comparisons between logistics costs and GDP. Section III describes the CASS methodology for calculating logistics costs as reflected in its definition of logistics, methods of cost measurement, and source data used for the measurement. Section IV describes the Eno Transportation Foundation methodology for calculating transportation costs, as the CASS methodology relies entirely on the Eno methodology for transportation cost estimates, which comprise the largest component of the total logistics costs. Section V presents some recent estimates from the CASS report. And finally, Section VI provides an assessment of the CASS and Eno methodologies in light of the conceptual discussion in section II, and, based on this assessment, suggests potential improvements to the current procedures used to calculate logistics costs. This last section also provides suggestions about how to develop GDP-compatible measures of logistics.

II. Some Conceptual Issues

The CSCMP definition of logistics cited in the introduction includes all activities concerning the movement and storage of goods between the point of origin and the point of consumption of the goods.⁵ According to this definition, logistics includes freight movement and excludes people movement. The measure of logistics costs as discussed in this report is accordingly limited to freight logistics, although it must be recognized that the overall logistics does involve passenger movement. The remainder of this section discusses several important conceptual issues related to the definition of logistics, and logistics cost measurement and use. Through the discussion, some basic criteria are established. They are later used to assess the CASS methodology in the last section of this report.

Point of Consumption (Business versus Households)

The term "point of consumption" must be defined. Should the point of consumption be extended to households? In other words, should the movement of goods such as that from a grocery store to a private kitchen be part of the logistics chain? Obviously, such movements require the same kind of inputs, individually at a much smaller scale of course, as those between wholesalers and the grocery store. Both require loading, transporting, and unloading the goods. However, since much of the goods movement to households is carried out by households, and household activities are not treated as part of the national production process in official economic accounting systems such as the U.S. National Income and Products Accounts and the U.S. Input-Output Accounts, treating household logistics activities as part of the overall logistics chain expands the production boundary and causes problems of consistency and comparability with official economic statistics.

One way to avoid this difficulty is to limit the measurement of logistics to business logistics only. And in fact, CASS' methodology includes only the logistics costs for the U.S. business system. Furthermore, the CSCMP definition may be interpreted to cover only business logistics because it refers to the supply chain, which normally does not include household activities. Clearly, the final stop of goods movement in business logistics is not always consumption. The inbound movement of goods to retailers is often the last leg of goods movement for business logistics, yet retailers are not the ultimate consumer of the goods. Although the movement of goods from retailers to the final point of consumption by households creates real value to consumers, this kind of value creation is not counted in current economic accounting methodologies. Instead, it is treated as consumption. In other words, value-addition in current economic accounting only happens in the business sector.⁶ Therefore, to achieve clarity, the definition for business freight logistics should refer to the last point of value-addition rather than the point of consumption. Whereas consumer transportation of goods from market to household is properly excluded from logistics activities, merchandise home-delivery on behalf of sellers is part of the production and logistics chain. To exclude the former and include the latter, business logistics may be better defined on the basis of who performs certain activities rather than where those activities are performed. To the extent that supply chain management refers to business activities only, the CSCMP definition of logistics on the basis of supply chain management is sufficiently clear for delimitating where business logistics start and end.

Included Cost Items

According to the CSCMP, logistics management activities "typically include:

• inbound and outbound transportation management;
• fleet management;
• warehousing;
• materials handling;
• order fulfillment;
• logistics network design;
• inventory management;
• supply/demand planning, and
• management of third party logistics services providers.

To varying degrees, the logistics function also includes sourcing and procurement, production planning and scheduling, packaging and assembly, and customer service. It is involved in all levels of planning and execution — strategic, operational, and tactical. Logistics management is an integrating function which coordinates and optimizes all logistics activities, as well as integrates logistics activities with other functions, including marketing, sales, manufacturing, finance, and information technology."⁷

Goods movement requires transportation and storage requires warehousing. There does not seem to be any controversy over including costs of transportation and warehousing in logistics costs. Measuring them, however, is not straightforward because both involve substantial amount of in-house operations. As discussed above, in-house operations refer to those business operations that a company conducts to provide services for its own use. In contrast to the services from for-hire operations that are bought and sold in market-places, in-house operations are provided and consumed internally without market mediation. Some internal operations should be included because they are either the same kind of activities as external logistics or the natural and immediate extensions of external logistics.

It is a well-known fact that much of the trucking and warehousing operations in the U.S. are in-house operations. Therefore, measures of transportation and warehousing costs should cover services from both for-hire and in-house operations. It is important to note that the for-hire cost should be the total price charged to the service users, not the cost to the carriers providing the services. The full cost of in-house services should include an imputed return to capital as well as the costs of intermediate inputs, labor compensation, and capital consumption.

While the cost of transportation measures the cost of goods movement, the cost of warehousing does not fully capture the cost of goods storage. Clearly, warehoused goods tie up capital and capital is costly. Moreover, businesses may suffer losses as materials in storage may experience obsolescence, physical deterioration, and loss in value. Also, businesses have to spend money to insure goods in storage against accidental losses such as those due to natural disasters and fire hazards. Based on these considerations, an inventory-carrying cost is often measured to include costs of interest, taxes, obsolescence, depreciation, insurance, and warehousing.

While how much of a certain material a business decides to keep in warehouses may heavily depend on transportation system performance, inventory-carrying costs may change independently for reasons completely unrelated to transportation. Costs of interest change in response to fluctuations in interest rate. Insurance costs depend on the level of insurance premium. Even the level of inventory itself may be determined by factors outside transportation and logistics management such as business cycles. A business may experience an unexpected increase in inventory due to an economic slowdown, no matter how well it manages its logistics operations. Factors such as these do not constitute a valid argument against the inclusion of inventory-carrying costs in total logistics costs, but they do indicate that certain price and business cycle variables need to be controlled if the resulting logistics cost measure is utilized for trend analyses. This point is further discussed in the final section of the report.

There are internal business operations that are immediate extensions of transportation and warehousing and therefore should also be counted. These include industrial traffic management, loading and unloading by shippers, inventory planning and analysis, and support by central distribution staff.

Inter-plant versus Intra-plant

Yet another problem in defining logistics is the question of how much intra-plant activity should be included as logistics activity. For example, goods are moved around within the business establishment where the goods are to be retailed. This is relevant because extensive intra-plant movement of materials may cause a significant increase in the estimate of total logistics costs depending if the activities are included as logistics. The current literature does not provide clear guidance on the issue. In some of the literature on business logistics management, production scheduling, materials handling, purchasing, order processing, and market forecasting are included as business logistics activities, where materials handling is basically intra-plant material movement.⁸ Although these activities affect the movement of goods and level of inventory, they are not themselves intrinsic to goods movement and inventory management. The literature and practice of macroeconomic costing of logistics seems to indicate that the costs associated with these functions should not be counted as logistics costs. For example, the Heskett approach and the CASS methodology both exclude many of those internal business activities. For Heskett, such exclusion may be because his approach was originally developed to measure the macroeconomic costs of physical distribution.

The reason for the exclusion of these cost categories from the calculation of logistics costs relates to the purpose for which logistics costs are being defined, measured, and analyzed. Within an individual firm, for the purpose of planning and managing business operations, logistics may well be defined to include everything that is involved in physical supply, physical distribution, and intra-plant materials movement. For a manufacturing company for example, anything that is not general administration and direct manufacturing operations may be counted as part of the company's logistics operations. It is certainly useful and necessary for the firm's management to know how much it costs the company to move things in and out and get them to the right spot in right quantity at right time. However, this cost information is less useful outside the firm (other than to its competitors.) Specifically, the cost information based on this broad concept of logistics is of little consequence to public decision-makers because it contains many elements on which public decisions have no affect. For instance, an inefficient plant layout that hampers intra-plant material movements will cause the firm to incur costs no matter how efficient the highway system or trucking operations are.Therefore, information on the internal logistics costs of a business enterprise is largely irrelevant to public decision-makers.

Of course, cost measurement or any other measurement efforts by the government sector are not necessarily all for the purpose of public decision-making. A logistics cost measure that includes all intra-plant material movement may be useful for some purposes, but the logistics cost measure exclusive of intra-plant material movement has its own value, and is perhaps more valuable to decision-makers in transportation.

The questions and discussion above define in theory the proper boundary of business freight logistics for the purpose of calculating national logistics costs. As determined, many internal business activities should be excluded from the calculation. However, the boundary must be cast to include specific cost items. Unfortunately, the exclusion of certain internal business operations is easier said than done. A certain level of arbitrariness in drawing the cut-off line is unavoidable. However, the consequence of this arbitrariness to the final estimates of overall logistics costs is likely to be inconsequential so long as the list of internal operations included does not get too long.

Logistics and GDP

Many economic statistics including logistics costs are often cited in relation to Gross Domestic Product (GDP,) implying an underlying relationship between the two. Though a relationship does indeed exist, the relationship is a multi-faceted one prone to misinterpretation.⁹

GDP is the total value of all goods and services produced during a certain period of time, such as a year, that has not been consumed by the production itself. In physical terms, GDP is a huge basket of goods and services that is available for final uses including consumer use, government use, investment use, and foreign use (export). Therefore, the total value of GDP is equal to the total value of final use or final demand. When the popular press states that two thirds of the US GDP is accounted for by consumer purchase, it means that consumers purchase two thirds of the goods and services in the GDP basket. In other words, consumers consume two thirds of the GDP.

Final use or final demand can be classified into different categories according to their general purposes. For example, consumers purchase cars, tires, gasoline, auto insurance, etc. to serve their transportation needs. Governments purchase steel, concrete, asphalt, etc. to build highways for transportation needs. Therefore, consumer expenditures on cars, tires, gasoline, and auto insurance and government expenditures on highway construction have one thing in common: they are all for the purpose of transportation. All purchases for transportation purposes by consumers, governments, and businesses (investment), and foreign users (export) can be put together and be called total transportation final demand. When the transportation literature states that transportation accounts for 11% of the US GDP, it means that all the final users purchase or consume 11% of the goods and services in the GDP basket to serve their transportation needs.¹⁰ Similar final demand measures can be developed for other broad social functions such as education, health, housing, and so on. One may treat logistics as a broad social function that includes transportation. In that case, total logistics final demand as a percentage of GDP means that logistics consumes that percentage of goods and services in the GDP basket. When a particular sector or activity is related to GDP as a basket of goods and services for final uses, the resulting comparison is a measure of how much of the GDP that sector or activity consumes. It is not a measure of how much the sector in question produces or contributes to GDP.¹¹

Value-Added

Each industry purchases all different kinds of goods and services to run its operations. Through different operational procedures equipped with capital and staffed with labor, each industry consumes what it purchased and produces products and services to sell to its customers. The industry's purchase of goods and services for its internal operations constitutes its intermediate purchase. The total value of the industry's products and services constitutes its gross output. The gross output minus the intermediate purchase is the industry's total value-added. All industries' combined value-added constitutes GDP. Therefore, GDP as total industry value-added can be divided into different parts according to its origin. Unlike GDP as a basket of goods and services for final use to which some industries may have very low contributions (e.g., iron ores and pig iron,) GDP expressed as the total industry value-added is produced by each and every industry in an economy.

An industry's contribution to GDP is normally measured by that industry's total value-added. It is indeed in that sense that Triplett and Bosworth state that the transportation sector represented 3.3% of GDP in 2001.¹² It is also in this sense that one can talk about logistics' contribution to GDP. The difficulty is that all logistics activities are not neatly classified into a unique industry. To gauge the true contribution of logistics to GDP, one must make extensive reclassifications, which greatly complicate measurement effort.¹³

As discussed above, GDP has different economic meanings. However, GDP is also a well-known economic variable, the level of which can be used without any of its underlying meanings attached. For example, often people may find it convenient to compare some economic measures with GDP, simply because GDP is a good yardstick. In those cases, relating a particular economic measure to GDP conveys no more intrinsic meaning than relating the size of the moon to the size of the earth. Anything that is not comparable to GDP but is so compared should be interpreted this way. The level of logistics costs is such a measure because it includes intermediate expenses by businesses and is therefore non-comparable with GDP. Eno Transportation Foundation's transportation bill, including the passenger bill and the freight bill, is another such measure. The result of comparing such measures to GDP only indicates the relative sizes of the objects measured. It reveals nothing about the role of the entities' contribution to or consumption of GDP. Therefore, phrases such as "contribute to" and "account for" should be avoided in these comparisons. Instead, "GDP equivalent" may be used.

III. CASS Methodology for Calculating Logistics Costs

The CASS measure of logistics costs includes three broad cost components comprising the business logistics system. They are inventory-carrying costs, transportation costs, and logistics administration costs. The inventory-carrying costs and transportation costs both contain subcomponents. Below each of these cost categories is described in accordance with a publication by the United Nations.¹⁴

Inventory-Carrying Costs

Inventory carrying costs include the cost of money (opportunity or interest), ad valorem taxes, insurance and shrinkage. Inventory carrying costs vary with the level of inventory stored. They can be categorized into the following four groups: (1) capital costs, (2) inventory service costs, (3) storage space costs, and (4) inventory risk costs.

Capital Costs for Inventory Investment: Holding inventory ties up money that could be used for other types of investments. Consequently, a company's opportunity cost of capital should be used to reflect accurately the true cost involved. All inventory carrying cost components must be stated in before-tax numbers, since all the other costs in the trade-off analysis, such as transportation and warehousing, are reported in before-tax dollars.

Inventory Service Costs: Inventory service costs consist of taxes and insurance paid as a result of holding inventory. In general, taxes vary directly with inventory levels. Insurance rates are not strictly proportional to inventory levels, but are related to the value of inventory over a specified time period.

Storage Space Costs: Storage Space Costs can be incurred at four types of facilities:

• plant warehouses;
• public warehouses;
• rented (leased) warehouses;
• company-owned (private) warehouses.

Inventory Risk Costs: Although inventory risk costs vary depending on the company, in general, they include charges for: (1) obsolescence, (2) damages, (3) pilferage, and (4) relocation.

Inventory carrying costs, the cost of taxes, and obsolescence, depreciation and insurance are estimated according to the Alford-Bangs Production Handbook formula.¹⁵ In this formula, obsolescence accounts for nearly 40 per cent of total inventory carrying costs, thus demonstrating the challenges facing inventory managers in the world of fast cycles and just-in-time procurement. Total warehousing cost estimates encompass both public warehouses and private warehouses operated by manufacturing and distribution companies. Public warehousing costs are obtained from the public warehousing services data reported by the Commerce Department's Census Bureau. Private warehousing costs are independently obtained by CASS. Relocation costs are incurred at the transshipment of inventory from one warehouse location to another to avoid obsolescence.

Transportation Costs

Total transportation costs include costs for both primary and secondary transportation. Primary transportation is the movement of finished goods from plants and vendors to warehouses. Primary transportation costs include costs for replenishment movement from plants or distribution centers to other plants or distribution centers, and inbound freight on purchased finished goods movement to plants or distribution centers for resale. Secondary transportation is the delivery of finished goods to customers. Secondary transportation costs include payments to carriers, pickup allowances, truck or rail equipment and operations costs, and freight allowed. Freight may originate in plants, distribution centers or terminals.

Transportation costs include carriers' charges for all modes, including trucking, rail transport, water and oil pipeline, and both international and domestic airfreight transport, as well as freight forwarding and shipper-related costs. The freight transportation costs in the CASS report account for the largest portion of logistics costs. These estimates are based on the annual Transportation in America report published by the Eno Transportation Foundation. Of total transportation costs, trucking costs dominate the United States business logistics system, accounting for more than 80 percent of the nation's freight bill. Shipper-related costs include the loading and unloading of transportation equipment, as well as traffic department operations.

Logistics Administration Costs

Logistics administration costs include indirect management and support staff, which includes central distribution staff, planning and analysis staff, and the traffic department. Computer software and hardware cost allocations are another important distribution expense. Such costs are included in the appropriate cost categories, with any remainder considered part of administration costs.

Logistics administration costs are set at four per cent of the sum of the inventory-carrying costs and transportation costs, in line with the methodology that has been consistently employed since the data series was first published in 1973.

The details of these cost components and the CASS methods of measuring them are included in Table 1, below.

Table 1: Cost Components and CASS Methods of Measurement

Table 1 description: Table 1 provides a summary of the information in the text on the cost components and measurement methods according to CASS method. All information in the table is available in the discussion in Section III, presented above the table.

IV. ENO Methodology for Calculating Transportation Costs

The Eno Transportation Foundation estimates the total cost of freight transportation in the United States and publishes this estimate with modal details as nation's freight bill in its annual series Transportation in America. As stated above, CASS relies on the Eno freight transportation cost estimate (freight bill,) which is the largest component of the CASS estimate of logistics costs. Therefore, the Eno methodology for calculating the freight bill should be considered part of the overall CASS methodology. The following description of the freight bill estimate is based on Transportation in America, 19th Edition.¹⁶

Intercity Truck

Intercity truck includes trucking operations for freight, mail, and express services. Recent year estimates for ICC-authorized trucks are based on Census For-Hire Truck Revenue surveys.¹⁷ Estimates for non-ICC-authorized trucks are derived as follows. First, the ICC-authorized truck vehicle-miles are subtracted from total freight-carrying truck vehicle-miles, using large truck vehicle-mile data for non-urban highways from the Federal Highway Administration's Highway Statistics. Second, the resulting truck-vehicle miles are multiplied by average revenue per vehicle-mile estimated using data for Class I and Class II ICC-regulated carriers.

Local Truck

Costs for local trucking are estimated as follows: The total truck vehicle-miles for urban travel (other than interstate urban travel) are taken from FHWA's Highway Statistics Table VM-1. An estimated percentage representing use of small urban trucks (pickups, vans, etc.) for passenger travel is subtracted from the FHWA total.¹⁸ Vehicle miles for trucks carrying tools-of-the-trade (plumbers, repairmen, etc.), obtained from U.S. Census data, are added to the remaining truck vehicle miles. The estimated per-mile costs were determined using data showing the cost of different classes of single-unit trucks (truck driver costs added) from the Federal Motor Vehicle Fleet Report, as published by the General Services Administration.

Railroads

Figures for 1980 to present are computed by using Class I rail freight revenues shown in Railroad Facts, published by the Association of American Railroads (AAR), and adding revenues from non-Class I railroads, plus outlays for rail freight service assistance and for rail employee pension benefits under the Railroad Retirement Act paid by the U.S. Government. Data for earlier years includes freight, express, baggage, mail, switching and demurrage, and is taken from the Transport Statistics in the U.S., Part I, published by the ICC.

Water

International: Figures for 1980 to present represent outlays for ocean transportation for U.S. imports (paid by shippers), including shipping, port expenditures and charter, but not military freight, taken from Department of Commerce (DOC) analysis entitled "International Transportation Transactions." U.S. export figures are from footnote 'c' of the same source. Earlier figures were taken from special reports in the DOC's Survey of Current Business.

Coastal/intercoastal: Figures are estimated by multiplying ton-miles from "Waterborne Commerce of the U.S., Part 5, National Summaries," as published by the U.S. Corps of Engineers (Corps), by estimated revenue per ton-mile.

Inland waterways: Figures are calculated using the same methodology used to calculate coastal/intercoastal costs. Figures include both domestic and foreign traffic moving in U.S. inland waterways.

Great Lakes: Figures are calculated using the same methodology used to calculate coastal/intercoastal and inland waterways costs.

Locks and channels: Figures include actual outlays made by the Corps for construction, operation and maintenance of channels and harbors, locks and dams, and navigation, but exclude multi-purpose projects. Figures are taken from the Budget of the United States.

Oil Pipeline

Regulated and Non-regulated: Current figures are based on Federal Energy Regulatory Commission data as reported by the Oil Pipeline Research Institute. Earlier figures were from ICC data for regulated pipelines, and estimated to be 16% of total for non-regulated pipelines (estimates provided by the ICC and the Association of Oil Pipelines.)

Air

Domestic and international: All scheduled and non-scheduled carrier figures are from the Department of Transportation's Air Carrier Financial Statistics. Section 418 All-Cargo carrier figures are from separate DOT reports, plus air-freight waybill taxes.

Freight Forwarders

Figures through 1975 are from ICC annual reports, and include only remaining operating revenues after deducting transportation purchased from rail, truck, and water carriers to avoid duplication of revenues from forwarders paid to carriers. The figures from 1980 to present are estimates based on annual changes in freight forwarder employment (Department of Labor data) and total average compensation (DOC data.)

Shipper Related Costs

Loading and unloading freight cars: Figures are derived by multiplying the sum of total rail freight carloads by estimated cost per car from Railroad Facts, published by AAR. Yearly average cost is adjusted to reflect changes in average weekly earnings in manufacturing industries shown in "Economic Indicators," published by the President’s Council of Economic Advisors (CEA.)

Operation of traffic departments: Estimates are based on published surveys of total costs of salaries of industrial traffic executives, administrators, supervisors and traffic clerks, handlers, and secretaries.  Average salary levels are adjusted annually based on average weekly earnings of manufacturing workers, from "Economic Indicators," published monthly by the CEA.

V. Estimates for Recent Years

The latest available estimates of logistics costs from CASS are those for the year 2002. Table 2, based on data from the 14th edition of the CASS Annual State of Logistic Reports, shows that the total logistics costs were $910 billion in 2002, equivalent to 8.7% of the U.S. gross domestic product in the same year.

Table 2 - U.S. Business Logistics Costs, 2002

Carrying Costs - $1.444 Trillion all business inventory		Billions of Dollars
	Interest	23
	Taxes, Obsolescence, Depreciation, Insurance	197
	Warehousing	78
Transportation Costs
	Motor Carriers:
	Truck - Intercity	300
	Truck - Local	162
	Other Carriers:
	Railroads	37
	Water	27
	Oil Pipelines	9
	Air	27
	Forwarders	9
	Shipper Related Costs	6
Logistics Administration		35
Total Logistics Cost		910

The following two charts present graphically the information contained in table 2 above. Transportation accounted for 63% of the total logistics costs, while inventory-carrying costs accounted for 33%, of which interest costs account for 8%.

Chart 1 description: Chart 1 presents the shares of the three main components of the logistics costs such as inventory-carrying costs (33%), transportation costs (63%), and logistics administration costs (4%).

Chart 2 description: Chart 2 shows the subcomponent shares of the inventory-carrying costs such as interest (8%), warehousing (26%), and taxes, obsolescence, depreciation and insurance (66%).

Transportation costs, the largest component of logistics costs, are largely composed of trucking costs. Intercity and local trucking make up a combined total of nearly 80%, which is more than 10 times as large as the second largest mode, railroads.

Chart 3 description: Chart 3 shows the subcomponent shares of the transportation costs such as intercity truck (50%), local truck (27%), railroad (6%), logistics administration (6%), water (4%), air (4%), oil pipelines (1%), forwarders (1%), and shipper-related costs (1%).

Table 3, based on data from the same CASS publication, provides time-series data. Both logistics costs and its largest component, transportation cost, have experienced a steady decline as compared to GDP, while the inventory cost shows some cyclical fluctuations over time. Along its decline compared to GDP, transportation as a share of the total logistics costs has trended upward, ending more than 18 percentage points higher in 2002 than in 1981.

Chart 4, based on data in table 3, more clearly shows the different growth patterns of the three major components of the logistics costs.

Chart 4 description: Chart 4 shows the different growth patterns of the three major components of the logistics costs such as inventory-carrying costs, transportation costs, and logistics administration costs. While administrative cost component is the lowest in level and almost flat over time, transportation component is the highest in level and fastest in growth. Inventory-carrying cost is in the middle and fluctuates over time.

VI. Assessment and Recommendations

The CASS methodology was designed and implemented to measure U.S. business logistics costs. What are actually measured are U.S. business freight logistics costs. Business logistics encompasses not only freight logistics, as businesses must get the right people to the right place in right time no less so than they must ensure the delivery of their supplies or outputs. A full measure of business logistics costs should thus include both freight and passenger logistics costs. The CASS measure of business logistics costs is incomplete because it excludes business passenger logistics costs.

For measuring business freight logistics costs, the CASS approach is generally valid. The CASS measure captures key components of the freight logistics without including unrelated internal business functions. In fact, among many of the business operations outside transportation and warehousing that are often incorrectly classified as logistics functions, the CASS approach only includes logistics administration. As was discussed in section II, a more inclusive measure can be useful for some purposes but the narrower measure based on the CASS approach is more useful for public policy analyses and public decision-making.

The measurement of logistics administration costs under the CASS approach is undoubtedly very rough. However, this element of the overall logistics costs is very small in comparison to transportation costs and inventory-carrying costs, which together account for over 90% of the total. 
The CASS methodology relies on Eno estimates of transportation costs. Section IV details the data sources and methods used to calculate the costs. Because input data come from so many different sources, and because the data were originally collected or compiled for different purposes, whether they are internally consistent is not clear. The Eno approach is generally defensible under the very real condition that not many alternative source data exist.

Eno estimates of trucking costs do include the extensive in-house trucking operations since they are derived on the basis of truck vehicle miles, and vehicle mile data cover private trucks. There are also in-house operations for other modes such as railroads that are not counted in Eno estimates because the estimates are revenue-based and in-house operations usually do not have identifiable revenue records. The U.S. Transportation Satellite Account (TSA) was designed to produce estimates of all transportation operations for all modes. The Bureau of Transportation Statistics of the U.S. Department of Transportation is currently exploring ways to capture in-house railroad operations, corporate jets, and private barges and ferries as well as in-house trucking. Future TSA estimates of transportation output may provide a better choice for transportation costs. Before then, Eno provides more readily usable data than the national accounts because the national accounts data are often industry-based without detailed enough distinction between passenger and freight transportation. Also the national accounts do not cover any in-house operations.

The third component of the CASS logistics costs is inventory-carrying costs. The CASS estimates of these costs are well based, using inventory data from the national accounts and warehousing data from the Census. The annualized commercial paper rate is a good approximation of the rate of interest cost for capital tied-up in the form of inventory. The parameters from the Alford-Bangs Handbook that the CASS Information System uses are well within the range of the industry standards.

Improved Measure of Logistics Costs

While the CASS methodology is generally sound, it may be improved. Several changes, if implemented, would allow the CASS estimates to better reflect real changes in the logistics system.

First, the prices of the goods in inventory should be held constant to allow inventory levels to be estimated in constant dollars. This is a standard practice applicable to all other logistics cost items. Without controlling price effects, inventory level fluctuates even if the real inventory level does not change.

Second, the level of inventory can be smoothed over time to lessen the effect of cyclical changes. As discussed in section II, an unexpected economic slowdown usually pushes up business inventory causing an increase in inventory carrying cost, other things being equal. Likewise, an unexpected economic upturn causes inventory to go down. While the resulting level of inventory-carrying costs can still be usefully measured, its changes are not good indicators of whether the underlying logistics system is working better or worse. A moving-average or some other time-series processes may be applied to the inventory data so that a more persistent trend can be identified.

Third, the interest rate used to estimate the inventory capital costs should be held constant. While the tax rates and the insurance premiums can both change, the CASS estimation does not individually utilize tax rate and insurance premium data. Interest rates are also relatively more volatile. Fluctuations in interest rates directly result in changes in the inventory-carrying cost even if the underlying logistics system stays unchanged. For a trend analysis, interest rates should be held constant.

With these adjustments for price effects, a constant-dollar time series or a quantity index could be developed to complement the available current-dollar cost estimates.

GDP-Comparable Measures of Logistics

Two measures of logistics can be developed that are components of GDP, one from the final demand perspective and one from the value-added perspective. The distinction between the two is important, and critical to understanding the relationship between logistics and GDP. A comprehensive measure of logistics from the demand-side captures the total final demand for logistics, which includes purchases by all final user categories (consumer, government, business, and net export.) Items included in this form of business logistics estimates should include private logistics investments such as those for vehicle purchases and warehouse construction. Importantly, estimation of those items requires extensive data extraction and regrouping from the national accounts because final demand data provided in the national accounts are not functionally grouped. Furthermore, there are important final demand items that are clearly logistics-related but are not classified as business logistics items, such as public investment in transportation. While techniques to estimate demand-side logistics expenses may, with substantial effort, be practically developed, the usefulness of such a measure is not readily apparent.

A more useful measure would be that from the value-added perspective, because the logistics value-added is a correct measure of logistics' contribution to GDP. Since value-added is often measured on the basis of its origination, the first step in measuring logistics value-added is to develop a clear boundary between logistics and other business activities. This may be approached in one of two ways.

One way is to treat logistics as a group of pre-defined industries. Based on current industry classification -- the North American Industry Classification System (NAICS) -- industries that clearly reside under the logistics umbrella are Transportation and Warehousing (NAICS Sectors 48-49.) For the transportation and warehousing industries, calculating total value-added is straightforward because the national industry accounts provide the data as needed.

The problem with using the existing industry classification, however, is that many logistics activities occur in sectors that would not fall entirely under logistics. Logistics operations carried out by other than transportation and warehousing industries are called in-house logistics. For example, a retail store may own and run its own warehouses as a part of its regular retail business. Although such warehouse operations are logistics functions, NAICS may categorize these functions as part of retail business operations if they are not operated as separate establishments. All business activities that facilitate the movement and storage of materials are freight logistics activities, regardless of who performs the operations. Calculating the added value of logistics activities within this boundary poses serious practical challenges, as estimating in-house logistics activities/costs/expenses requires a careful redefining and regrouping of economic activities on a functional basis.

For the transportation portion of such an estimate, there is a readily available alternative. Specifically, the U.S. Transportation Satellite Accounts provide estimates of value-added from both for-hire and in-house transportation. While in-house transportation is likely to be a large component of in-house logistics, it is certainly not all-inclusive. For example, shipper-related logistics activities and logistics administration are two other broad categories of in-house logistics functions. Estimating the value-added from these types of in-house logistics activities is something that is not currently done, largely because of the substantial effort that would be required. As such, there is currently no means of calculating the contribution of logistics activities to GDP from a value-added perspective.

As a result, the common comparison between logistics expenses (as calculated by CASS) and GDP is ripe for misinterpretation. The problem is that this logistics cost calculation contains 1) intermediate goods and services, and 2) internal business operating costs unrelated to logistics. When firms outsource logistics activities they are purchasing not only the services produced by the logistics providers, but also the intermediate inputs used in the production of the services. When firms run in-house logistics operations, their "logistics costs" also include internal business activities and purchases that are not strictly logistics functions. Furthermore, logistics costs include inventory-carrying costs, which includes opportunity cost of capital, which is not a component of GDP.

Therefore, assuming the most commonly cited estimate of logistics costs:
It is incorrect to say that logistics costs account for X % of GDP.
It is incorrect to say that logistics contributes X % to GDP.
To say that logistics costs are equal to X % of GDP is acceptable, but it is simply a statement of their relative sizes, not a statement of how much one is dependent on the other.