Identification and Evaluation of Freight Demand Factors (2012)

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.

12 2. Measuring and Estimating Transportation Demand In Task 1, the research team reviewed existing freight demand models and related literature to identify and evaluate a wide variety of freight demand measurements and the potential independent factors that could be used to predict these demands. The following principles were applied to this review of potential measures of demand and of the independent, predictive variables: • The independent variables that influence transportation demand should be easy and inexpensive to gather and should be recognized as important and regularly updated economic or demographic variables; • The dependent variables that express transportation demand should provide insights into the overall activity of the nation’s freight system and be useful in generating forecasted needs for new investment in operational improvements and infrastructure; • The data representing all of the variables in the analysis, both the independent factors and the dependent measurements of transportation demand, should be publicly available at minimal cost. • The mathematical form of the model should not be restricted or in any way pre- determined; many possible formulations should be considered. Measures of Transportation Demand The research team posited that the contributing factors to total freight volumes might be significantly different by mode. Most demands for freight transportation are predisposed to use a particular mode, based on commodity type, value, weight or size, origin and destination location, and urgency. This assumption is further supported when one notes that many carriers provide several “tiers” of service tied to price, speed and service guarantees to ensure their freight customers don’t find the need to choose other services. Of course there are important exceptions – “zones” of contestable traffic where modal qualities compete - although in terms of overall volume, they are a minority. Depending on traffic lane and shipment value or urgency, the movement of medium-value, containerized freight over a “medium-haul” of 750-1,500 miles can occur via either intermodal rail, with drayage moves on each end, or all-highway. The long-term decrease in waterborne freight tonnage and increased exportation via Columbia River ports suggests there may be some shifting of freight to railroads. Notwithstanding these exceptions, the preponderance of freight typically moves via a given mode based on its value, weight, origin-destination, and perishability. The research team originally identified nine selected measures of transportation demand (dependent variables) that are described below, along with their data sources and frequency. It should be noted that all these datasets offered national-level data – with limited exceptions, data is generally not available on a regional or state-wide level. R ailroad measurements are well represented, primarily because their data is better and their operations are plainly defined. Almost all United States rail freight moves over at least one of the seven Class I railroads which

13 means that their activity is measured weekly, monthly and annually by both the Association of American Railroads (AAR) and Surface Transportation Board (STB). National measurements of truck freight are complex due to the challenge that “Counting Truckers Can be Complicated3 ” and the risk inherent in aggregate tonnage volumes that goods may be counted more than once as they are staged through the system (a problem that other modes may face as well, but with substantially lower incidence than trucking). T ruck ton-mile and vehicle-mile measures avoid this problem, and are obtainable for long time periods. Waterborne traffic is available via annual summaries only from the Institute of Water Resources (IWR), a department of the Army Corps of Engineers. Per the research team’s communications and agreements, the data described below are freely available through the Research and Innovative Technology Administration (RITA), which coordinates the U.S. Department of Transportation's research programs, including the Bureau of Transportation Statistics (BTS)4 Rail Tonnage , the STB, or the IWR. This annual statistic, gathered and made public by the Association of American Railroads (AAR), provides a basic measure of the weight of goods transported, regardless of distance between their extraction, production, or importation supply points and their utilization, consumption, or exportation demand locations. As railcars continue to improve – the new 286,000-lb gross limit means that railcars can now carry 110-115 net tons of freight – railroads provide a very efficient method of moving heavy loads of goods long distances within North America. Rail Ton-Miles With a combination of higher terminal and train building costs, the rail mode tends to concentrate on longer hauls of freight than trucking. Railroads use net ton-miles (net weight of freight, vs. gross which includes the weight of the rail equipment itself) as the primary tally of their productive operations. These data are similar to rail revenue ton-miles (discussed below), with the relatively small exception of railroad-owned freight, such as maintenance-of-way material. The dataset is estimated on an annual basis by the BTS based on A AR data. Moreover, ton-miles are an important statistic for railroad operations as they are used for pricing or rate estimates. Ton-miles correlate well with railroads’ consumption of fuel and their medium- and long-term requirements for locomotives and track maintenance. Rail Train-Miles Since railroad freight, as well as the equipment used to carry it, moves between stations and terminals in trains, it is also useful to measure train-miles. Annual estimates of train-miles and the next variable, rail car-miles, are both published by BTS as part of the National Transportation Statistics. While train-miles increase with traffic, efficiencies can slow train-mile growth even as rail transportation demand increases. For instance, railroads can haul more freight per train 3 Transportation Research Circular E-C146, “Trucking 101 An Industry Primer,” Transportation Research Board, December 2010, 4 See http://www.bts.gov/publications/national_transportation_statistics/.

14 using double-stack intermodal cars sized for international container lengths and low-tare aluminum gondolas that carry 20 more tons of coal than steel gondolas. Railroads offer shippers discounts when they ship more railcars at a time because 90-car trains are not much more expensive to operate than 87-car trains. Rail Car-Miles Measuring the movement of railroad cars, as opposed to tons, provides a key insight when it is noted that for the purpose of this statistic, a car-mile refers to a mile run by a freight car or an intermodal trailer or container, with or without a load.5 Rail Revenue Ton-Miles (Quarterly) Besides the efficiencies of loading more tonnage into non-intermodal railcars, the increasing volume of intermodal “units” as a percentage of rail carriers’ loadings affects this annual measure. This is the only dependent variable included in the analysis gathered on a quarterly basis for two reasons. First, as explained above, freight transportation data is limited for the truck and waterway freight. Second is that Revenue Ton-Miles is the best metric of the railroad industry’s productive efforts and used regularly in industry financials to explain revenues (with adjustments for traffic “mix”). The dataset is estimated by the Surface Transportation Board from AAR data. The smaller time frame allows for better identifying “lag” effects from some of the independent factors as predictors of increases or decreases in freight transportation demand. These data were utilized for a s pecialized “back-cast” to test the accuracy of the predictive model on f reight demand. Truck Ton-Miles Surface freight moves primarily by truck. With intermodal rail capturing a moderately higher percentage of longer-distance hauls, most truck moves are less than 200 miles. Ton-miles is a good measure of how most, but not all, trucking companies price their services, as well as a proxy for the useful work trucks provide (and correlates well with the fuel they use and the emissions they generate), although ton-miles aren’t particularly relevant to less-than-truckload (LTL), package and courier freight. Truck Vehicle-Miles Besides moving up to 25-28 tons of freight (depending on axle and tire configuration, applicable road restrictions, and tare of the cab and trailer equipment), trucks of various sizes may be utilized for shorter-distance deliveries and the essential “last-miles” for transporting air freight, intermodal and even small packages. LTL networks move small packages throughout the country, while courier services interface with air transport to connect nearly every office and home via “Next Day” or “2nd Afternoon” services. Indeed, the activity of truck movement, irrespective of the total tonnage they carry, is a key indicator of economic vitality as tracked by the regularly published National Transportation Statistics. 5 Per various Bureau of Transportation Statistics’ publications on annual transportation statistics, for example: https://www.bts.gov/publications/transportation_statistics_annual_report/2008/html/appendices/glossary.html

15 Domestic Waterway Tons In general, different kinds of freight moves on barges over domestic waterways than by railroad (the notable exception, in some cases, is grain). Barge freight is typically heavy, low-value, and high-volume where timeliness is not important. It also typically does not move far overland to or from its waterway origination or termination because these transfers would obviate the advantages of water transportation. The Army Corps of Engineers’ Institute of Waterway Research (IWR) measures how much revenue tonnage is loaded each year onto barges and ships moving domestically on United States river systems (mostly Ohio-Mississippi-Missouri) and various maintained waterways (mostly the Atlantic and Gulf Intracoastal Waterways (GIWW); see Figure 1). 6 Figure 1. Composition of Internal Tonnage by Waterway Source: Waterborne Commerce of the United States, Calendar Year 2005, Part 5-National Summaries, U.S. Army Corps of Engineers Domestic Waterway Ton-Miles Also available from the IWR is a summary of annual loadings in short tons times distance traveled over the domestic waterway system which is a good approximation of the value created by the waterway carriers and how much shippers and consignees paid for the services. 6 Waterborne Commerce of the United States, CY 2005, Part 5 – National Summaries, U.S. Army Corps of Engineers Other 9% Ohio River 35% Columbia/Snake 2% GIWW 16%Mississippi Main Stern 38%

16 Waterways attract freight that is extremely heavy and usually has low value relative to its weight. The origin-to-destination transportation can involve, but does not require, a long length-of-haul to be appropriate to the water mode. The Task 2/3 report describes in more detail the evaluation and selection of dependent measures of freight demand (see Chapter 3).7 Independent Factors Affecting Transportation Demand The National Cooperative Highway Research Program (NCHRP) Report #388, “A Guidebook for Forecasting Freight Transportation Demand,” provides a helpful overview of the various exogenous economic and societal factors that contribute to the demand for freight transportation. The Guidebook offered a wide variety of factors along with some guidance on how they affect a) either the total demand of goods transported, or b) the distances and origin-destination markets in which these goods were actually moved. As first described in the Task 1 r eport for this research,8 these factors can be divided into two categories: • “Pure” or direct factors affect transportation by creating or demanding “more stuff.” Crop production, mineral extraction, population growth, housing starts, retail sales, and exports all impact the total volume or tonnage of goods that gets moved. • “Network” or indirect factors affect transportation just as much by influencing where, when, and how goods are moved. W here people live and how and where they buy interact with new developments in logistics and sourcing strategies (e.g., an interconnected global village that comes to the big-box mall, vs. “buying locally” via markets and in downtowns). P ublic policies, infrastructure investments, and logistics technologies influence freight demand as much as the demographics of where people work and live. In addition to the five general categories of factors described in the Scope of Work for this research effort – demographic, economic, environmental, technology, and public policy – the research team found that a sixth category, operations strategy (or logistics) was helpful in conceptualizing the research effort. Demographics is a Direct Factor How many people and where they live is one of the factors most basic to macroeconomic activity and hence, transportation demand for food, fuel, consumer goods, waste – in short, everything. People form households and utilize housing – another factor in transportation demand. A 7 Project No. NCFRP-11, Identification and Evaluation of Freight Demand Factors, Report #2, Tasks 2-3, June 9, 2010. 8 Project No. NCFRP-11, Identification and Evaluation of Freight Demand Factors, Report #1, Task 1, Candidate Demand Factors, Sept. 4, 2009; Chapter 1.

17 percentage of them become the labor force that provides the means of production, cultivation and extraction of other freight that must be transported. Economic Activity is a Direct Factor GDP, particularly the goods production component, is a fundamental measure of the national output of freight that gets that produced, consumed and at some level, transported. Production of durable goods tends to be more cyclical and correlates well with the growth in freight demand (as measured by ton-miles), while consumable goods correlate more closely with population size and average wealth. Levels of production for goods that have a high transportation component – typically either because of their weight or the distance they are moved – were considered as candidate independent variables or factors in freight demand. Specific examples are coal and grain (including oil seeds), which typically move by railroad or water, and imported goods, which often move long distances over inland surface transportation networks after they are unloaded from container ships. Globalization and the increased economic efficiency and prosperity of people around the world has resulted in increased demand for goods, as well as the ability to source supply from a much broader set of locations. Capital investment and effective labor mean that goods can be produced all over the world. The resulting economic wealth means that goods are now demanded and consumed by a greater diversity of the world. Not only does the Walt Disney Company source the toys for its Florida and California resorts from China and other locations in Asia, but as the Chinese economy grows it becomes wealthy enough to demand the attractions, entertainment and even food exported from the United States to the newest Disney parks in Hong Kong and (soon) Shanghai. Fuel is both an Indirect and Direct Factor Fuel constitutes a significant and relatively volatile component of cost for all freight modes, as well as a significant percentage of most consumers’ cost of living. During the early 1990s, fuel accounted for 7.1% of total operating expenses for Class I railroads; fuel, oil, lubricants and coolants accounts for about 13.5% of operating expenses for truck-load carriers and about 6% of operating expenses for LTL carriers; and 30-40% for air carriers.9 Because of an approximately three-fold increase in price that has outpaced most other expenses, fuel has recently accounted for approximately 10~15% of railroad operating expenses and 15~25% of long-haul trucking. Higher fuel prices make long-haul trucking services less price-competitive against railroad alternatives, shifting some long-haul freight to intermodal rail. The volatile increases in gasoline prices also tend to have a dampening effect on consumer economic activity, as much personal automobile travel is relatively price elastic over the short- medium term. A s consumers pay more for gasoline, they have less money to spend on ot her things thereby dampening consumption and the economy. 9 A Guidebook for Forecasting Freight Transportation Demand, NCHRP Report 388, Transportation Research Board, 1997.

18 Environmental Policy is an Indirect Factor Federal and state environmental policies may place particular restrictions on por ts and waterways. Port expansion is limited by the Clean Water Act, requirements for “no net loss” of wetlands, or limitations and permitting processes for channel deepening or harbor dredging. The Oil Pollution Act requires oil tankers that have single hulls to be replaced with double hull tankers and tankers with single hulls are required to meet other criteria. With trucks, emissions controls and clean fuel requirements have also affected costs, and environmental policy towards carbon emissions is an unknown future factor in fuel prices. Air transport is often subject to noise reduction acts. By increasing transportation costs, all these factors can affect the demand for freight. In addition, environmental policies such as air and water quality regulations influence the locations at which raw materials are produced and those at which industrial plants are located. Recycling and overall incentives to reuse solid waste may have an important effect on some transportation demands. Recycling plants, as well as waste-to-energy incinerators, are typically located near urban centers or the source of their raw materials, while industrial manufacturing plants that use virgin materials are usually located farther away with the need to ship their products much longer distances. Technology is an Indirect Factor The use of computers and telecommunication equipment has had an important effect on t he freight industry. Air carriers and many leading manufacturers and retailers have implemented sophisticated systems for tracking shipments, as well as computers for sorting shipments and optimizing the use of freight vehicles such as aircraft and trucks. By improving efficiency, level of service, and reliability, these factors have indirectly promoted modern just-in-time (JIT) delivery methods. JIT systems focus on keeping inventories at minimum levels through coordination of input deliveries with production schedules. The effect on freight demand may be to increase the number of individual shipments, decrease their length of haul, or increase the importance of on-time delivery. Public Policy is an Indirect Factor The deregulation of various transportation industry sectors during the early 1980’s allowed private operators to serve their freight customers with innovative and remunerative business offerings. These changes were phased in over several years during this period. Almost all highway and waterway freight movement is dependent on publicly constructed, maintained and, in many cases, operated infrastructure. The 50 State Departments of Transportation, along with dozens of associated highway authorities are primarily responsible for building and maintaining intercity roads. They in turn are guided and heavily funded by the U.S. DOT Federal Highway Administration (FHWA). The Army Corps of Engineers maintains harbors and waterway channels. The FHWA works with state departments of transportation to maintain and improve the national highway system. While transportation infrastructure tends to be expanded and improved more slowly than many users would like, measures of public infrastructure investment in dollars or in the physical characteristics of lane-miles, miles of

19 dredged channels, or new feet of runway could all be considered as independent variables. However, these measurements might actually be a result of increased freight demand, rather than a cause, so they were not considered as “influences.” User charges, directed tax revenues, and in many cases, tolls are used to fund U.S. infrastructure, including airports, ports and highways. Besides direct user charges such as tolls and facility charges, complementary taxes are levied on gasoline, highway diesel, and fuel sold along inland waterways. These taxes add to transportation costs and can make some modes more attractive relative to others. While freight railroads largely operate on their own infrastructure, they too are impacted by tax, regulatory and other public policies. A rail infrastructure tax incentive provides economic encouragement for private railroad companies to invest in new roadbed and track that will ultimately provide more capacity and support greater demand. International trade agreements reduce barriers to trade, encourage consistency in government regulations, and generally promote trade and cross-border freight traffic. NAFTA was a big boost to Mexico-United States trade and, over time, has led to hundreds of cross-border investments and greater use of transportation to access the competitive advantages of each country. The Contestability of Freight is an Indirect Factor Some kinds of freight shipments, depending on their weight and size per shipment, its value and per ton, and its perishability, can move on alternate modes. For example, most of the heavy, typically low-value/ton, freight that moves via waterways can also move via rail. In many cases, grain, fertilizer, and gravel/stone/sand also moves via truck, although typically these are short- haul, last-mile moves to and from waterway or rail terminal locations. It is a combination of new and changing services offered by the major freight transportation modes and changing factors in logistics and freight distribution that has affected the portion of freight that is contestable among mode. There are two major shifts of freight between modes that have occurred over the past 30 years that are worth mentioning. One is the strong growth of intermodal movement of truck trailers, now superseded by the use of ISO containers, moving via railroad instead of over the highway for medium-long-hauls between North American metropolitan areas and/or ports. The second is that waterway traffic has lost a s ignificant share of their long-haul freight, possibly because railroads are better able to connect the Midwest with Asia via Pacific Northwest ports than through the waterway modes’ traditional connections via Gulf transloading locations. Since the mid-1980’s, practically all intermodal rail traffic is a shift from trucking. For railroads it’s a big deal -- 14~15 million boxes of 10~16 net tons each, moving 750~2,000 miles (150~200 million tons and 300~350 billion ton-miles of freight) is now close to 1/4th of their freight revenues. For trucking, where the entire industry moves an estimated 10 billion tons of overall tonnage each year, that 1.5% share isn’t so significant. Actually, trucking may not have lost any tonnage handled since all intermodal shipments move by truck – often twice!

20 Intermodal’s effect on overall truck volumes is tiny, with the exception of long-haul, dry-van and reefer markets where it is a b it more significant. R ITA data indicates trucking carries approximately 1.3 trillion ton-miles of freight, so intermodal rail accounts for a small, but still countable 2.3% of surface trailer traffic, mostly due to the skewing of intermodal competitiveness towards longer haul moves. After another season of floods, we have seen more challenges to the movement of waterborne freight. As reported in the Wall St. Journal10 Operations Strategy (Logistics) is both a Direct and an Indirect Factor , increased siltation and delays in repairs to locks and channels by the Army Corps of Engineers have made it more difficult and expensive to move freight by waterways. Waterborne Commerce of the United States data indicates that while total tonnage has experienced a modest decline during the past 20 years, ton-miles have decreased by over a third indicating the loss of longer-haul freight. Since there weren’t good explanations from independent economic variables including the Inland Waterway Fuel tax, and the volumes of underlying demand for goods movement haven’t decreased, other factors such as the deterioration of the physical waterway network or the competitiveness of a rejuvenated railroad industry may have played a part in shifting this freight and accounting for the substantial decrease in waterway freight movements. Intermodal operations developed not only from technological innovation, but also from a change in operations by international shipping lines interested in extending their services between Asia and numerous U.S. inland origins and destinations. Interoperable equipment including standard- sized containers, twist-lock lifting mechanisms, and double-stack rail cars were technologies that allowed operational innovations. At the same time, the trend has been towards using fewer, more efficient warehouses. Third- party logistics operators and innovative warehouse developers with advanced forklift, stacking and pallet-picking technologies have helped this trend by supporting advanced logistics management while improving the inventory/sales ratio. Appropriate measures beyond this ratio might be the number of firms using modern, computerized warehouses, or simply a count of square feet of warehouse space built or refurbished after the marriage of inexpensive PC-based information systems with modern storage technologies. Trends in the location of distribution and manufacturing centers have influenced transportation mode choice. As new facilities have developed in rural, exurban locations along interstate highways, dependency on rail transportation has decreased, except for the heaviest, low-value freight. Summary All of the above factors were considered in varying degrees by choosing a set of independent variables that would represent most of these influences on f reight transportation demand. For example Housing Starts (demographics, consumption, interest rates) and Industrial Production Index (manufacturing, mining and extraction, and goods processing) measure two distinct areas 10 Wall St. Journal, “Old Locks Jam River Traffic,” January 6, 2011, and “Silt Buildup Muddies Trade on River,” July 5, 2011”

21 of the economy while Exports and Imports summarize trading volumes, which generate significant demands in long-distance transportation. Candidates for Factors that Affect Transportation Demand Based on the guidance provided by the Transportation Research Board and discussed above, as well as on past and expanded review of literature and existing freight models, the research team identified potential candidates to be used as the independent factors affecting freight transportation demand. This was one of the key efforts of Task 1, and Appendix B presents the detailed results of the team’s literature review. The research team used a “case” approach to review and analyze a variety of recent and relevant studies on w hat factors affect the demand for transportation. First, a significant number of documents, including the bibliography of NCFRP-01, the “Review and Analysis of Freight Transportation Markets and Relationships,” were scanned to identify relevant documents. A thorough review of these documents yielded a representative collection of freight demand models that have been used in the past. During the second step of the literature review process, the research team performed an in-depth analysis of the various models described in the studies summarized in the Appendix B – Literature Review. In reviewing the literature, it was important to distinguish between general demand factors and the specific datasets actually utilized to represent these factors. In some cases, datasets employed to represent certain factors may actually be outputs of other models or summarizations (e.g., the output from an external regional economic growth model). For each of the complex and varied datasets in the model survey literature, the research team endeavored to tease out the underlying factors for consideration as independent variables. In subsequent efforts as part of Tasks 2 a nd 3, these candidates were narrowed based on availability, differentiation from one another and the ability to predict transportation demand. Finally, the addition of new dependent variables measuring freight demand via inland waterways required several additional independent factors to be added. Determining the Specific Independent Factors In this step, the research team identified and evaluated specific datasets that could be used as independent variables in modeling freight demand. To maximize the benefit of using a limited set of independent factors that affect transportation demand, the research team had to determine which of the potential factors were most useful to model freight demand. As analyzed below, two important criteria in choosing the independent factors were a) their ability to explain key changes in the economy that affected transportation demand; and b) their independence or uniqueness vs. other potential independent factors. This work is described in greater detail in the Task 2/3 report.11 11 Project No. NCFRP-11, Identification and Evaluation of Freight Demand Factors, Report #2, Tasks 2-3, June 9, 2010.

22 The initial goal was to consider data only if it w as publicly available and collectable on a monthly, quarterly and annual basis. However since all but one of the freight demand measures were publicly available only on an annual frequency, the independent factors were summarized on an annual basis as well. The candidate list includes most of the “pure” demand factors that were identified in the preceding section. The specific factors were next grouped into broad categories to identify and discuss their relevance to economic and demographic factors as well as to help evaluate the possibility that many of the factors measure the same things and therefore some might be unnecessary. These analyses were broadly broken down into the following categories of demand drivers: • Income – as an indicator of the size of the economy and its strength; • Housing Starts – a significant statistic which ties into other closely-followed macro- economic activities while corresponding with other goods acquisition and transportation activities; • Production – various macroeconomic measurements of manufacturing and extraction as well as general business sentiment around imminent investment and business activity; • Trade and Foreign Exchange – as international trade and the long supply chain of imported and exported goods movement are large consumers of demand, measuring the value of the dollar vs. the currencies of its major trading partners was expected to be helpful in predicting transportation demand to and from U.S. ports; • Inventory/Sales Ratio – provides a counter-cyclical measure of business activity during times of increasing and decreasing economic activity; • Retail Sales – offers the most direct measure of the quantity of goods consumed and pulled through the production-import-warehouse-store sequence of transport and logistics activity; • Fuel Prices – changes can make many consumers feel less or more well off, affecting their demand for consumables and other goods. Also affects costs for trucking more than other freight modes, making rail transportation relatively more attractive. Income Included in the candidate list of demand factors are various measures of income. Real GDP and real GDP per capita are among the most widely used measures of economic activity. Figure 2 shows growth rates of per capita income-related factors including real GDP, real personal consumption and real disposable income based on da ta from the U.S. Bureau of Economic Analysis (BEA). Freight demand typically follows growth in GDP.12 However, during the current recession, freight traffic has declined more than the decline in GDP (for example, rail carloads and intermodal units declined more than 20% in May 2009 compared to May 2008.13 12 Economic Impact of Freight. Bureau of Transportation Statistics (BTS). www.bts.gov/programs/freight_transportation/html/freight_and_growth.html. ). Similarly, 13 Levinson, Marc. “Freight Pain: The Rise and Fall of Globalization.” Foreign Affairs, Vol. 87, No. 6, November/December 2008

23 growth in GDP – a broad measure of economic activity – translates into consumer (i.e., household) expenditures that are ultimately spent on pur chasing goods and services, either domestically produced or imported, can be captured by real personal consumption. Historically personal consumption accounts for about two thirds of domestic final expenditures.14 Real disposable income reflects what people receive from land, labor and capital after paying taxes and inflation. These output and income measures reflect how much of household earnings go into purchasing goods and services and implicitly the demand for freight services. -4.0% -2.0% 0.0% 2.0% 4.0% 6.0% 8.0% 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Real Per Capita Disposable Personal Income $ 05 Real Per Capita Pers. Cons Exp. 2005 $ Real GDP Per Capita, Chained 2005 $ Figure 2. Per Capita Growth in GDP, Personal Consumption and Disposable Income Source: Bureau of Economic Analysis 14 NIPA Handbook. Bureau of Economic Analysis, October 2009, Chapter 5.

24 Housing Housing starts are counted and released monthly by the Census Bureau and provide insight on construction trends in the U.S. They tend to follow closely-monitored, large macro-economic developments in the U.S. However, as the recent housing-led economic recession has proven, the slowdown in housing started much sooner than in the rest of the economy. After reaching a peak in 2005, total housing starts in the United States started to decline from 2006 onwards (see Figure 3). Trends in housing have matched trends in freight. Rail carloads and intermodal units, for example, reached a peak in 2006 a nd have declined consistently throughout 2007, with large year on year declines (in excess of 5%) starting in the latter half of 2008 and increasing to more than 20% year on year in 2009.15 300 500 700 900 1,100 1,300 1,500 1,700 1,900 2,100 2,300 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 in T ho us an ds US Housing Starts, Total Figure 3. Total Housing Starts in the United States (Seasonally Adjusted at Annual Rates) Source: U.S. Census Bureau 15 Ibid

25 Production The Federal Reserve Board measures production activities in the United States. These include goods produced domestically for consumption or exported, as well as imports of semi-finished goods that may be further processed in the U.S. before their final use. The Industrial Production Index includes activities of the manufacturing sector, including mining and utilities, while the Industrial Manufacturing Index excludes mining and utilities. In 2007 the Industrial Manufacturing Index accounted for 78.7% of the Industrial Production Index, vs. 85% in 1999.16 Total capacity utilization measures output in the manufacturing, mining and utilities industries against total capacity. This measure can be used to understand how well the resources in the economy are employed in production activities. Total capacity utilization measures are also maintained by the Federal Reserve Board. Figure 4 shows capacity utilization and industrial production for the historical period 1980 to 2007 based on data from the Federal Reserve Board. The Purchasing Managers’ Index (PMI) from the Institute of Supply Management is an important supplemental indicator for the U.S. economy. It is based on a survey of purchasing managers that weights: 1) new orders; 2) production; 3) employment; 4) supplier deliveries; and 5) new inventories. Conducted monthly, the PMI is considered a leading indicator of business sentiment, with values above 50 i ndicating that that the manufacturing economy is generally expanding while a reading below 50 shows that it is generally declining.17 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 110.0 120.0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Purchasing Managers Index Total Capacity Utilization Industrial Production Index Figure 4 includes recent movement of the PMI along with other measures of production. Figure 4. Industrial Production, Total Capacity Utilization and Purchasing Manager Index Source: Federal Reserve Board 16 Industrial Production and Capacity Utilization: The 2008 Annual Revision. Federal Reserve Bulletin, Federal Reserve Board, August 2008, p. A58 17 Overview of the Manufacturing Report on Business. Institute of Supply Management, www.ism.ws/ISMReport

                                                                   26  Trade and Foreign Exchange In a world where supply chains have become increasingly globalized, trade as measured in both volume and value plays an extremely important role in determining freight demand. Imports to the U.S. have grown rapidly, partly as a result of increased containerization. The demand for freight transport is a derived demand; international trade and domestic consumption are the main drivers. The reduction in freight transport costs from deregulation and resulting efficiency gains helped bolster trade and enable the creation of long-distance supply chains. Trade itself is partly a function of the exchange rate between the U.S. dollar and the currencies of its trading partners, and therefore such exchange rates also affect freight transport demand. For example, the Chinese government has been under considerable pressure to allow the Yuan to appreciate, since the low value of the Yuan keeps the price of Chinese goods artificially low, thereby stimulating U.S. imports from China. Figure 5 presents recent U.S. trade value data for the past 17 years. Similarly, weakness in the U.S. dollar against Asian currencies has helped U.S. export growth, especially during the most current recession, as many of the healthiest segments of the American economy – agriculture, defense, and high-tech electronics – have been supported by a strong export market accompanied by dollar weakness.18 The Trade Weighted Foreign Exchange Index, a measure of the U.S. dollar vs. the currencies of our country’s primary trading partners, factors currency strength (or weakness) into observations on changes in imports and exports. 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 3,500,000 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 R ea l T ra d e V al u e (M il li on s U SD , 20 05 ) Figure 5. Real U.S. Exports & Imports Source: US Census http://www.census.gov/foreign-trade/statistics/historical/ 18 Levinson, Marc. “Freight Pain The Rise and Fall of Globalization.” Foreign Affairs, Vol. 87, No. 6, November/December 2008

27 Inventory/Sales Ratios Inventory/sales ratios are a counter-cyclical measurement in that during an initial period of business slowdown, sales decline and inventory piles up. A s business improves, sales increase while inventory initially is depleted (before orders can catch up). Inventory/sales ratios offer better insight than simply looking at sales or inventories; for example, during 2005 inventories increased while the inventory/sales ratio fell, because sales were increasing even faster. Figure 6 shows quarterly data on manufacturing and trade inventories held by firms in real U.S. dollars and a chained measure for inventory/sales ratio from the BEA. Generally, as can be observed in Figure 6, inventory-sales ratios increase rapidly during recessions and show a declining trend during periods of economic expansion. 1,100,000 1,150,000 1,200,000 1,250,000 1,300,000 1,350,000 1,400,000 1,450,000 2000.1 2001.1 2002.1 2003.1 2004.1 2005.1 2006.1 2007.1 2008.1 2009.1 R ea l I nv en to ri es (2 00 5 $, M ill io ns ) 1.20 1.25 1.30 1.35 1.40 1.45 1.50 In ve nt or y S al es R at io s Real Inventories Inventory Sales Ratio (BEA) Figure 6. Recent Real Manufacturing & Trade Inventories and Inventory/Sales Ratios Source: US Census of Manufacturing and Trade Inventories & Sales and BEA

28 Retail Sales Retail sales are an important potential factor affecting transportation demand that was identified as part of the literature review.19 320,000 330,000 340,000 350,000 360,000 370,000 380,000 390,000 2000.1 2000.4 2001.3 2002.2 2003.1 2003.4 2004.3 2005.2 2006.1 2006.4 2007.3 2008.2 2009.1 M ill io ns o f 2 00 9 $ (S ea so na lly A dj us te d) Retail Sales in Real $ Increased demand for goods by households due to economic growth leads to increased retail activity. Increased retail activity is an important generator of truck trips particularly in urban areas, as well as long-haul moves between ports and manufacturing regions to local markets. Figure 7 shows recent retail sales data from the Census Bureau in real 2009 dollars (deflated using the urban consumer price index from the Bureau of Labor Statistics). Real retail sales declined rapidly during the recession, declining to 2000 levels after experiencing robust growth during 2003 to 2006. Figure 7. Recent Retail Sales in Real U.S. dollars Source: U.S. Census Bureau 19 Beagan, D., M. Fischer, and A. Kuppan, Cambridge Systematics, Inc. Quick Response Freight Manual II. Federal Highway Administration, 2007

29 Fuel Prices Even though trucks and trains use diesel fuel, gasoline prices are a good economic indicator that may affect the demand for truck transportation. First, the price of gasoline is highly correlated with that of diesel. Higher fuel prices, which account for 35~40% of operating costs (including driver pay), directly makes trucking more expensive, also contributing to reduced demand. The impact of gas and diesel prices on rail freight is less because fuel is only 20-25% of railroad operating costs. Second, because the purchase of gasoline is essentially a staple requirement for many consumers, higher gas prices reduce disposable income and thus consumer demand for other goods, reducing demand for trucking those goods. Figure 8 shows both nominal and real urban average gas prices that have been deflated using the consumer price index (CPI) based on data from the Bureau of Labor Statistics. 0 50 100 150 200 250 300 350 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 C en ts / G al lo n Real Urban Gas Price (2009 $) in Cents/Gallon Nominal Urban Gas Price in Cents / Gallon Figure 8. Real and Nominal Urban Gas Prices in U.S. Cents per Gallon Summary Based on these analyses, a set of independent factors representing a diverse set of economic activities influencing freight transportation demand were ultimately chosen. Consistent with the requirements that the data needed to be available, non-confidential and low-cost or free, the independent factors were chosen from highly accessible public sources such as the Census Bureau, Bureau of Labor Statistics, Bureau of Economic Analysis and the Federal Reserve Board. The Table 1 below summarizes the list of candidate variables that, based on the literature review, are believed to drive total demand for goods transportation, which in turn drives infrastructure needs. The table lists the sources of each dataset and the specifics used.

30 TABLE 1 - Chosen Independent Factors Variables Sources Notes Real GDP 2005 Chained $ Bureau of Economic Analysis GDP in 2005 dollars, chain price deflated Real GDP 2005 Chained $/ Capita Bureau of Economic Analysis & Census Bureau Mid-year Population Estimates Real Personal Consumption 2005 Chained $ Bureau of Economic Analysis Real Income / Capita, Chained 2005 $ Bureau of Economic Analysis & Census Bureau Mid-year Population Estimates Total Housing Starts Census Bureau Avg of Monthly Data Industrial Production Index Federal Reserve Board Avg of Monthly Data Industrial Manufacturing Index Federal Reserve Board Avg of Monthly Data Purchasing Managers Index Institute of Supply Management Avg of Monthly Data Trade Weighted Foreign Exchange Index (Broad Trading Partners) Federal Reserve Board Avg of Daily Data Trade Weighted Foreign Exchange Index (Major Trading Partners) Federal Reserve Board Avg of Daily Data Employment Total Bureau of Labor Statistics Total Non-Farm Payrolls, Avg of Monthly Data Employment in Wholesale Industry Bureau of Labor Statistics NAICS code 42, Avg of Monthly data Real Exports in Goods (in $) Census Bureau Census Basis, deflated with CPI, avg. of monthly Real Imports in Goods (in $) Census Bureau Census Basis, deflated with CPI, avg. of monthly Total Capacity Utilization (% of Total) Federal Reserve Board Inventory Sales Ratio, Chained $ Bureau of Economic Analysis Manufacturing & Trade, SIC Basis (1980-96) chained 1996 $, NAICS Basis (1996-09) Chained 2005 $, Avg. of Monthly data Inventory Sales Ratio Census Bureau Total Business, SIC Basis (1980-91), NAICS Basis (1992-09),Avg. of Monthly data, actuals adjusted Real Gas Price Bureau of Labor Statistics U.S. Urban Avg, CPI Deflated Retail Sales in Real $ Census Bureau U.S. Retail Sales, SIC Basis (1980-91), NAICS Basis (1992-09), actuals adjusted, CPI Deflated Inland Waterway Trust Fund Fuel Tax/Gal (Lagged 1 year) Inland Waterways User Board Historic adjustments to the inland waterway user’s excise tax on fuel Grain and oil seed production (tons) US Dept of Agriculture Historic summaries of crop production Coal production (tons) Energy Information Admin. Annual Coal Report Total grain & coal production (tons) US DoA and EIA Sum of coal & grain, in short tons SOURCE: Developed by the Research team