Chapter 3 - State of Transportation Statistics

Transportation statistics arise from an array of data systems each constructed for specific, sometimes narrow, purposes. These systems exist much like a collection of pieces from different jigsaw puzzles of the same picture. The pieces answer some questions well but leave many others unanswered or partly or poorly answered. The pieces do not constitute a whole because of a number of factors, such as incompatible definitions, diverse collection methods, data overlaps and omissions, coverage and timeliness of data collection and releases, and apparent inconsistencies. In addition, transportation statistics need to satisfy a variety of users (see table 1).

This chapter focuses on five core transportation areas: freight, passenger travel, air transportation, economic, and geospatial data. Each section provides an analysis of why these data are important, a review of existing data, and possible options for filling crucial data gaps. The challenge for the Bureau of Transportation Statistics (BTS) will be to solve, where possible, some of the more pressing data needs in these areas.

Freight Data

Changes in freight transportation reflect the dynamic nature of national and global economies and the continuing improvements and innovations in technology. Alterations in the mix of manufactured products, shifts in global production and trade patterns, and growing domestic demands from industry and consumers all affect freight transportation and related data needs.

The consensus on collected freight data is that they are often too out of date to capture current developments and, despite progress, there are many missing pieces to the freight picture. Furthermore, data are often not comparable across modes. Table 2 compares features of several freight data sources.

Some freight data needs come up consistently among data users. Major categories of data where important gaps remain include:

• freight flows (modal combinations used; tonnage, weight, ton-miles, volume, and value of shipments) at the local, state, national, and international levels;
• true origins and destinations of shipments (not necessarily the same as the place of record);
• commodities (type; value; whether hazardous, containerized, or bulk);
• transit times (how long it takes to ship commodities);
• shipment costs;
• intermodal connections (seaports, airports, railyards) through which freight passes; and
• infrastructure, vehicles, vessels, airplanes, and pipelines used.

In many of these categories, some of the data may be available but not on a comprehensive basis.

Existing Freight Data and Data Programs

Commodity Flow Survey (CFS)

The CFS, conducted by the U.S. Census Bureau for BTS, is the broadest, most comprehensive national survey available of multimodal freight trends. Every five years, the U.S. Census Bureau surveys shippers for the CFS. Release of 2002 survey data will begin in late 2003; previously available data cover 1993 and 1997. For a sample of shipments, shippers are asked to provide information on the shipment's origin and destination, its value and weight, the mode or modes of transportation, and the commodity type (using codes from the Standard Classification of Transportable Goods—SCTG).

Uses of CFS data include:

• determination of the magnitude of freight moving over the national transportation system,
• projection of growth in freight transportation demand,
• calculation of ton-miles using origin and destination points,
• estimation of traffic routing, and
• calculation of the modal distribution of shipments by value, tonnage, and ton-miles.

While the CFS is the most extensive survey available of domestic freight activity, it does not cover all sectors of the economy. The shipments of farms, logging and fishing operations, governments, construction firms, service establishments, transportation firms, and most retail firms are not within the scope of the CFS. The distribution center or warehouse of a retail firm will be in scope, but its retail store will not. For instance, farm produce is not captured when it is transported from the farm to a grain elevator or processor but is when reshipped by the elevator or processor. Pipeline shipments of crude petroleum are also outside the scope of the CFS.

The CFS excludes imports unless the shipment is received by an importer at its establishment within the United States and then reshipped; the reshipment portion would be counted. The CFS includes the value and tonnage of exports. However, only the domestic portion of the shipment mileage is counted; the international portion is not. The CFS excludes "landbridge" cargo carried across the United States (i.e., cargo that does not originate or terminate in the United States—truck traffic from Canada to Mexico, or rail traffic carrying Pacific Rim freight from West Coast seaports to East Coast seaports for shipment to Europe).

The relatively small sample size of the CFS limits its geographic and commodity detail. The 1993 CFS had an establishment sample size of 200,000; the 1997 sample was reduced to 100,000 (covering 5 million shipments); and the 2002 sample was halved again to 50,000 (covering 2.7 million shipments). The Census Bureau designs the CFS to survey national and state-to-state flows, so data on flows to and from metropolitan areas are often not available. Planners and policymakers have indicated a strong need for greater geographic detail—at the county level—for analysis and forecasting.

Carload Waybill Sample

This specialized database provides detailed information on freight movements by rail. The Surface Transportation Board collects this stratified sample of rail waybills, with the sample rate varying from 2.5 percent to 50 percent depending on the size of the shipment. About 577,000 shipments were sampled in 2001.

Data collected include rail system origin and destination, weight, commodity (by 7-digit Standard Transportation Commodity Classification (STCC) code), and mileage. No value data are collected. The main weakness of the Carload Waybill Sample is the lack of true origin and destination data. The sample gathers no information beyond the scope of the rail system, so that the origin and destination are treated as occurring when the shipment is tendered to or delivered by the system. The true origin and destination, if it requires a connection to the rail system by truck, is not shown. Moreover, shipments interlined between two railroads are often shown as two movements, with the interchange point being listed as the destination of the first shipment and the origin of the second. Export traffic is also excluded unless it is rebilled at the border.

Waterborne Commerce Data

The Waterborne Commerce Data collected by the U.S. Army Corps of Engineers provide details on maritime freight. It is a complete census of all waterborne domestic and foreign movements and gathers data on the weight of the shipment, the origin, destination, and routing of the waterborne portion of the shipment, the waterborne distance traveled, and a code for the commodity carried (using the Corps of Engineers' own commodity coding system). No value data are collected for domestic shipments, but the Corps collects customs data on the value of foreign shipments. The foreign data lack information on routing.

Air Freight Data

BTS gathers data on air freight directly from the carriers. For each carrier, these data show the enplaned weight of freight at each airport, the weight of freight carried over airport-to-airport segments, available ton-miles of capacity, and actual revenue ton-miles flown. The data do not include information on the commodity classification of the freight carried or its value. The data also generally do not report the true origin and destination of cargo, that is, the point from which it is shipped by truck to the airport and to which it is delivered from the airport. Moreover, when the cargo is interchanged at an airport, even by the same airline, it is treated as a new shipment, so even the airborne origin and destination points are lost. The database also does not count cargo moved by the smallest carriers (i.e., air taxis).

International Data

The U.S. Census Bureau provides surface freight import data collected from the Bureau of Customs and Border Protection¹ to BTS. These data are then reported as part of BTS's Transborder Surface Freight Data. BTS also gathers data on exports from Canada and Mexico to the United States to complete the database. The data measure the weight of imports (but not exports), the value of shipments, the modes of transportation, the commodity classifications (from the 2-digit Harmonized Schedule), the ports of entry or exit, the freight charges, and codes to indicate if the shipments were containerized. Data on origin and destination show only the state or province. Similar data are not available for air and water exports and imports to and from Canada and Mexico, and the data available on exports and imports from other countries are even more limited.

The federal government plans to substantially improve the quality of international freight data using an International Trade Data System (ITDS). The ITDS, a multiple agency effort spearheaded by the Bureau of Customs and Border Protection, promises to provide comprehensive data on imports and exports, including origin and destination, mode, commodity, weight, and value.

Motor Carrier Data

Data on the number of vehicles passing through the transportation network are needed for many purposes. For instance, the number of vehicles is important for calculating the level of congestion in the transportation system. Many freight vehicles travel empty, contributing to highway congestion but not showing up in freight statistics.

The U.S. Census Bureau conducts the Vehicle Inventory and Use Survey (VIUS) every five years. The VIUS provides data on the number of trucks (including data on the size of truck fleets owned by carriers), their types, vehicle weights and lengths, fuel economy, the industries they serve, the kinds of commodities they carry, their range of operation (e.g., local, regional, or long distance) and the percentage of miles operated out-of-state, weeks operated, model years, how they are acquired, and the nature of the carriers (e.g., for-hire, private, or contract). These data do not provide any information about freight flows themselves, such as origins and destinations, but are useful in conjunction with other databases for understanding truck movements.

Options for Filling Freight Data Gaps

As the above discussion suggests, there are many important freight data gaps. Two pressing examples are the lack of good origin and destination data for truck shipments and data on intermodal shipments. While the CFS is, in principle, a comprehensive freight database, it in fact omits many sectors of the economy and reports limited details on a relatively small sample of shippers. Moreover, the sample may contain significant errors in the reported mode of transportation when the shipper often has incomplete information (e.g., the shipper that hires a trucking company to transport the shipment may not know that the trucking company contracts with a railroad to move the trailers by rail flatcar for a considerable portion of the trip). Databases on rail, air, and maritime shipments fill some of these gaps, but little data are available for truck and intermodal shipments. Moreover, other databases cover only the portion of the trip handled by the mode on which they focus and do not cover intermodal connections.

The Committee on National Statistics of the National Academy of Science's Transportation Research Board (TRB) released a study on the CFS in March 2003 [8]. The committee recommended that the five-year survey be continued until an improved alternative could be established. The committee reasoned that the existing CFS provided unique and important data on domestic freight movements, despite its limitations and shortcomings. Among the limitations cited by the committee were gaps in shipment and industry coverage and a lack of geographic and commodity detail. These limitations, according to the committee, are compounded by the survey's shrinking sample size. (Recommendations from another report on freight data options—by the TRB Committee on Freight Transportation Data—A Framework for Development—were not available in time for inclusion in this report.)

A new American Freight Data Program would aim to fill in key data gaps. It could have many elements, but the extent to which gaps could be eliminated would depend on resources available. Three specific components of a new program could be:

• filling the most important gaps and making freight data more timely by changing the Commodity Flow Survey to an annual or continuous collection of data that includes specialized sample frames of construction, agricultural, retail, and services shippers;
• providing complete statistical coverage of transportation of imports and exports through the ITDS; and
• developing a mail-out survey of nonpassenger vehicles to collect data on vehicle movements and vehicle and driver characteristics.

Such approaches would build on existing data efforts. Over the longer term, new innovations and approaches may emerge as practical alternatives for gathering data. Two of these—electronic transmission of operational data from vehicles and infrastructure and possible adoption of transportation data elements as part of a universal bill of lading—are discussed in box 1.

CFS Enhancement

For the foreseeable future, a survey such as the CFS is likely to remain a central component of a national freight data program. A top priority for CFS enhancement would be to gather information from shippers currently outside the scope of the survey. The focus for sample frames for the shippers could be those that generate the most vehicle-miles of traffic (e.g., construction, services, retail trade, and agriculture).

Two other key weaknesses in the CFS are lack of timely data and a sample size that is too small for local planning. It may never be possible in a national freight flow survey to generate sufficiently detailed data to meet the needs of all local planners, but if data were collected annually they could be aggregated to provide larger sample sizes. This would allow the most recent data to be analyzed if the focus is on currency at the expense of geographical detail; it would allow data for the last five years to be analyzed if the focus is on maximum geographic or commodity detail.

BTS lacks independent authority to require responses to a survey. Previously, BTS collected freight data under an interagency agreement with the Census Bureau, which has mandatory authority under its five-year economic census. It has been demonstrated that voluntary responses are less timely, complete, and accurate than responses collected under a mandatory survey, reducing the quality of the data and increasing the cost of the survey.

Another way to provide data for local planning might be through partnerships with state and local governments willing to provide resources to increase sampling rates in geographical areas in which they have a special interest. This approach would add data to the CFS while allowing state and local governments to obtain information on their own local traffic flows, probably at a lower cost than conducting their own surveys. This is similar to the approach taken with the 2001 National Household Travel Survey.

While the CFS lacks import data, some data are available from the Bureau of Customs and Border Protection. The ITDS, while still in the planning stage, will eventually remedy this existing data gap in the CFS.

Vehicle Movement Data

Even if all the gaps in the CFS are filled, certain kinds of data will still not be available. As indicated above, the CFS only tracks movements of freight, whereas for many transportation planners and policymakers, movement of all nonpassenger vehicles is of interest, whether those vehicles carry freight or not.

There are also other important data needs that cannot be met through a freight survey alone. Safety exposure data describe the drivers operating vehicles (e.g., their age, level of experience, number of hours on duty), but more descriptive vehicle data than that provided by the VIUS are needed (e.g., tare weights, axle loadings, safety equipment), as are data on where vehicles are operating.

A number of approaches already exist for surveying vehicles. The United Kingdom uses a mail-out vehicle survey in which truck operators fill in weekly diaries on the operations of a particular vehicle for a week's period. These diaries report, for each trip during the week, information on the origin and destination of the trip, the number of miles operated full and empty, the weight of the shipment, and the type of goods carried (with a special code to note hazardous materials). While these diaries do not provide data on transit times or the vehicle's driver, this approach could be used to gather this information.

Canada uses a roadside intercept survey, in which data collectors stop trucks at 238 roadside points and ask the driver questions about origin, destination, routing, and transit time of the trip (including information on stops), the weight and commodity classification of the cargo, the nature of the truck, and information about the driver and the carrier. Data are collected on handheld computers.

The California Air Resources Board (CARB), in conjunction with the Federal Highway Administration (FHWA), conducted a one-time study in 1999 using global positioning system (GPS) technology to track vehicles in Southern California. CARB obtained usable data from 140 trucks and tracked them for 4 months. Data were collected on truck routes and stops, travel times and speeds, idling times, and times of starts and stops. This approach gives the most detailed data on the actual route used by the truck, the actual times when the truck was in operation, and the amount of time required to make a trip at particular times of the day, week, and year.

Passenger Travel Statistics

A much-valued feature of American life is the ability to travel from place to place with relative ease, at a reasonable expense, and relatively quickly, whether it is across town, across country, or to a foreign destination. Americans average 1,500 daily trips annually, covering an average of 14,500 miles per person [19]. Many kinds of data are needed to evaluate this demand for passenger travel and how well the supply meets the demand. For instance, data are needed for the different modes of transportation and at various levels of detail, including geographic scale. Questions that help evaluate the needs of current and future travelers include why people travel, how and when they travel, what their origins and destinations are, how long travel takes, and how much it costs. Travel data, in combination with other types of data, can also be used to assess the costs and benefits of travel, including transportation safety and its environmental effects.

Government and private organizations need passenger travel data. Policymakers, planners, and forecasters use these data to determine infrastructure investment requirements and to measure and design strategies to influence transportation demand, particularly in the face of growing congestion. Data can also be used to identify geographic areas and groups of individuals who may be underserved by public transportation. Passenger travel data, including the number, type, and use of motor vehicles, can provide measures of exposure for safety analyses, energy efficiency, and the environmental impacts of travel. State-by-state estimates of vehicle travel are the foundation of decisions about funding, cost, and fee allocation.

Businesses and other private sector entities (e.g., consumer organizations and public interest groups) also want passenger travel data for a wide variety of reasons, and despite the cost, they frequently collect data for their own purposes. These data are often of use to others and may be available to the public. Privately collected data may be supplemented with government data; however, in many cases, government data are the only source of information available for business planning and other analyses.

Existing Passenger Travel Data and Data Programs

There are three main types of passenger travel data: survey, regulatory/administrative, and operations/industry data. Each type provides different levels of detail in terms of coverage, periodicity, and geography; and each possesses different strengths and weaknesses.

Passenger Transportation Surveys

Since the 1960s, the federal government has periodically conducted nationwide surveys of passenger travel in the United States. The most recent, the 2001 National Household Travel Survey (NHTS), combines two previous surveys: the Nationwide Personal Transportation Survey (NPTS—a survey primarily of daily travel), and the American Travel Survey (ATS—a survey of long-distance travel). BTS and FHWA conduct the NHTS.² The 2001 NHTS national sample surveyed 26,000 households nationwide on their daily travel and asked about the number and purpose of their trips, the mode(s) of transportation used, the distance traveled, as well as the social and demographic characteristics of the household. In addition, the survey asked household respondents to report all long-distance trips (trips of 50 miles or more one way) during a four-week period. Additional detail on purpose, mode, duration, overnight stops made to and from the farthest destination, and so on, were collected for each of these trips. In addition to the national sample, another 40,000 households were surveyed from 9 additional jurisdictions: 5 local, mostly urban, areas and 4 states [19].

The NHTS travel data along with the demographic data collected on households and individuals provide one of the few tools to analyze the travel patterns of sociodemographic groups. For instance, a recent study by TRB used data on the travel of children from the 1995 NPTS to assess their safety going to and from school [9].

Several improvements were introduced in the 2001 NHTS, including data collection on the travel of children under the age of five and better data on walking and medium-range trips (from 50 to 100 miles). These changes mean that the most recent survey data cannot always be easily compared with data from prior year surveys. An important supplement to this kind of travel survey is the upcoming time-use survey being developed by the U.S. Department of Labor [13].

Commuting behavior, which affects congestion and the scheduling of work, is captured most fully by the long form of the decennial census. Mode, origin and destination, start time, and travel time can be linked to the full range of social, economic, and demographic characteristics of the population at a very fine geographic scale. Information from the decennial census is relied on for federal policymaking and program development; for metropolitan transportation planning, including air quality analyses under the Clean Air Act; and for transit planning. The U.S. Census Bureau plans to eliminate the long form from the 2010 census and replace it with an annual American Community Survey to provide up-to-date demographic and housing data on a continuing basis. National-level journey-to-work data are also collected every two years in the American Housing Survey, conducted by the U.S. Census Bureau for the Department of Housing and Urban Development.

Some other, more specialized surveys also exist. For instance, the Office of Travel and Tourism Industries in the Department of Commerce surveys between 69,000 and 95,000 passengers per year through a Survey of International Air Travelers (In-Flight) Program, providing information on the origin and destination of foreign visitors coming to the United States by air and the foreign travel of Americans by air [12]. This information is supplemented by arrivals and departures data from the Bureau of Customs and Border Protection. Additional information on the travel of Canadians to the United States and Americans to Canada is provided by Statistics Canada. The Federal Aviation Administration conducts the General Aviation and Air Taxi Activity Survey annually. The VIUS, discussed earlier in the Freight section, collects information on private and commercial trucks, including passenger-oriented light trucks, their use, and a host of other related variables at the state level [11]. Examples of private surveys include the Travel Industry Association of America's survey of tourist travel [10] and the American Bus Association's survey of the intercity bus industry [3].

Regulatory Data and Administrative Records

Regulatory programs that require service providers to report information to the government generate passenger travel data, as do administrative records and other data collections by states and localities. Regulatory program data include information on airlines, intercity bus travel, and transit. Administrative records and other data collected by state and local governments include the Highway Performance Monitoring System (HPMS) and fuel tax revenue data. Since deregulation of the transportation industry, beginning in the late 1970s, availability of regulatory data has declined.

The National Transit Database, the most comprehensive information on transit, contains data collected from major urban transit agencies receiving federal aid and is maintained by the Federal Transit Administration (FTA). Data cover services supplied and consumed, finances, safety, and security.

Airline data, collected by BTS's Office of Airline Information, include flight on-time data, other service quality indicators (mishandled baggage and oversales), enplanements by airport and airline, airplane- and passenger-miles, and financial data. Other airline data include the Passenger Origin and Destination Survey, a sample of 10 percent of all airline tickets used in scheduled service, as discussed later in the Air Transportation Statistics section.

FHWA, through the HPMS, collects data from states on vehicle-miles traveled by vehicle and road type, as well as information on the extent and condition of the road system. This information is available for states and urbanized areas. Among other things, HPMS data are used for federal funding apportionment and for a major annual congestion study by the Texas Transportation Institute [7]. States also report other types of information to FHWA derived from administrative records, such as fuel consumption from state fuel tax revenue records and motor vehicle registrations.

Operations and Industry Data

Public operators of the transportation system, particularly state and local departments of transportation and industry groups, collect a vast amount of data. Industry sources releasing operations data include the American Public Transportation Association (APTA) (transit), Amtrak (intercity train travel), and airline associations. APTA's transit database adds to the National Transit Database by collecting data that their members do not report to the federal government. Amtrak operations data include number of passengers served, miles traveled, and on-time performance.

Real-time monitoring of vehicle travel by roadside sensors can provide data for a number of purposes such as evaluating the impacts of intelligent transportation systems, computer model calibration, congestion monitoring, and transportation planning. For instance, an FHWA-sponsored project is using roadside sensor data to develop indicators of congestion in 10 urban areas on a monthly basis [6]. The Texas Transportation Institute is working on a similar project [7]. Ensuring confidentiality and privacy of this type of data collection is an area that is and will continue to be important.

Operations data also exist in air travel. FAA's Air Traffic Control System Command Center monitors, in real-time, air traffic over the continental United States and provides real-time airport status information to the public.

The Bureau of Customs and Border Protection captures inbound border crossing information by vehicle type (passenger car, bus, and train) and of pedestrians [17]. The Bureau does not collect comparable data on outbound vehicles. Other sources of information include individual bridge operators, border state governments, and the Mexican and Canadian governments.

Options for Filling Passenger Travel Data Gaps

Data exist for all modes of passenger travel and related factors on a wide range of variables, but significant gaps exist. A data gap can result from deficiencies in data existence, completeness, detail, quality, timeliness, integration, and accessibility.

Timeliness and quality are virtually always issues. Options for making passenger travel surveys more timely are limited, because they tend to be conducted five or more years apart and the data usually take more than a year to publish after collection. Continuous measurement may help solve the problem of timeliness, although it has its own set of challenges [15]. Operational and administrative data also takes time to produce, compile, and publish. For instance, it takes about a year to assemble and publish a national set of annual data for the National Transit Database.

Variability in definitions, populations surveyed, reference periods and timeframes, sampling question formats, and so on also affect comparability of data collected. Sometimes a good deal of data exist on a topic from various sources, but the collection method has not been standardized so that integrating the data in a statistically valid way is difficult or impossible. Much data collected and analyzed pertains to only one mode of transportation. Only a few data sources, such as the NHTS, identify the ways in which different modes are used in combination to make a single trip. Consequently, there is little information on subjects such as wait times between different trip segments.

Data integration is also a problem for state and local users who need system data, such as infrastructure extent and condition, integrated with data on system use and surrounding characteristics such as land use. Indeed, data to assess the link between transportation and land use is typically a top concern for analysts at the local level.

Several types of important passenger travel data are not available in usable form, because, among other things, they are not publicly available or cannot be aggregated. Data gaps exist for travel cost and time. Some cost information is available for commercial modes and from the Bureau of Labor Statistics's (BLS's) Consumer Expenditure Survey, but these sources generally do not provide information on the prices consumers paid nor do they relate costs to specific trips or types of trips [16]. Using the BTS Passenger Origin and Destination Survey data, BTS is working with BLS to develop an airline price index. Travel time or on-time performance is available for air carriers and Amtrak, and the decennial census collects data on time spent commuting. In addition, the BTS/FHWA NHTS includes travel times for all trips. However, NHTS data provide national averages, making it difficult to assess highway congestion or transit reliability trends in specific areas.

Publicly available information on intercity bus and rail are limited even on a national level, not to mention corridor-specific data or use of these modes by population subgroups. Furthermore, data on these long-distance modes in the 2001 NHTS are more limited than the data that were collected in the 1995 ATS. And, despite work by FTA and the Community Transportation Association of America, rural transit data are also quite poor [20].

Publicly available information on passenger travel using commercial water transportation, such as cruise vessels, are limited although a recent congressionally mandated ferry survey by FHWA provides important information for that segment of the industry. Historically, data on the use of recreational boats (e.g., motorboats, jet skis, and sailboats) have not been collected, limiting safety analysis. However, a new National Boating Survey conducted by the U.S. Coast Guard may begin to correct this deficiency. Exposure measures for general aviation aircraft are also limited.

The U.S. Census Bureau initiated the Longitudinal Employer-Household Dynamics (LEHD) pilot program to demonstrate the potential for linking existing economic and demographic administrative and survey data. BTS is participating in this pilot to conduct research on developing detailed origin and destination tables based on residence and employment information of U.S. workers. This research effort will combine information about workers (place of residence, employment, and income) and employers (location, type of business, number of employees, and payroll). Once available, these data will illustrate the travel flow of workers to places of employment and will be made available to the transportation community, subject to confidentiality requirements. It is anticipated that the LEHD data will provide greater geographic specificity than has been available in the past and will lead to the ability to derive enhanced travel demand patterns for use by transportation researchers, planners, and engineers.

In many instances, few or no data exist on populations with special transportation needs; for example, people with disabilities and special segments of the elderly population such as the housebound, those living in special care facilities, and those living in rural areas. BTS sponsored questions about the travel of people with disabilities in the 1995 National Health Interview Survey on Disability conducted by the National Center for Health Statistics, but the survey provided very limited information. In 2002, BTS conducted its own small-scale³ national survey about how persons with disabilities use transportation, the barriers they face, and their overall satisfaction with the transportation system [18]. A preliminary findings report will be available in late 2003.

Detail and completeness are problems with most travel datasets in even the major travel surveys. Despite its utility, the NHTS does not include origin and destination information for local trips. Data for state and local analyses are limited to the few "add-on" areas where states or locations have supported collection of additional data. The decennial census, conducted every 10 years, is limited to commuting, trips that account for only a small percentage of all trips made. To improve pedestrian data, BTS includes related questions in its bimonthly Omnibus Survey. In addition, BTS and the National Highway Traffic Safety Administration conducted a small national survey in 2002 on bicycling and walking to ascertain the scope and magnitude of these activities and the public's behavior and attitudes regarding them [21].

More comprehensive solutions to passenger data issues may be appropriate. In late 2001, BTS asked TRB and the Committee on National Statistics to convene a panel of experts to review the key BTS survey programs. The Expert Panel to Review BTS Survey Programs recommended the following options in their June 2002 letter report:⁴

• An enhanced NHTS. This could involve: 1) improving the sample size and representativeness, 2) increasing the overall precision of resulting estimates, 3) providing more timely data to users by changing the frequency of the data collection, 4) adding more statistical rigor in data processing and final products, and 5) expanding the use of existing and emerging technologies.
• A new, targeted data-collection program. The program would develop and implement ways to address known data gaps but be flexible enough to handle new gaps that arise.

Air Transportation Statistics

Airline traffic and financial statistics were first collected by the federal government in the 1930s for use in monitoring and promoting the fledgling air transport industry. Today, the U.S. Department of Transportation (DOT) collects traffic, operational, and financial statistics from more than 240 domestic and foreign airlines serving the United States.

Prior to deregulation of commercial airlines in 1978, traffic and financial statistics supported the federal government's responsibility to ensure universal air service through regulation of routes, airfares, and freight and mail rates. After deregulation, the federal government's responsibilities were reduced, although key data series in such areas as safety, financial, and operating statistics were retained. Air carriers have also increased their use of traffic and financial information as they monitor and adjust their competitive position and determine new route opportunities.

While scheduled service enplanements grew by 47 percent from 452 million to 665 million from 1991 to 2000, by 2001 these enplanements decreased to 622 million [14]. The decrease was due primarily to the economic downturn and to the September 11, 2001, terrorist attacks on the United States, after which the aviation system was shut down entirely for one day and then partially for several days, costing the airline industry an estimated $330 million per day [1]. Since that time, the airline industry has faced a fundamental shift in travel patterns and increased security and safety measures. Higher operational costs and narrow profit margins create economic challenges for the airlines. Financial, operational, air traffic, pricing, and safety airline statistics aid in addressing these and other issues and provide policymakers and industry stakeholders with information to understand airline travel trends.

Air transport data supports policy initiatives and international air service negotiations, monitoring of air carrier fitness, allocating airport improvement funds, ensuring the provision of essential air services, setting international and intra-Alaska mail rates, and safety and security analyses. For instance, the Department of Labor uses aviation data in their computation of productivity and consumer price indices. The Department of Justice uses data to monitor the collection of customs service fees and for anti-trust cases.

Examples of other uses of air transport statistics include airport planning, traffic forecasting, and development of tourism initiatives by state and local governments; travel planning by the general public; planning and marketing by the travel and tourism industry; and forecasting and analysis by airlines.

Existing Air Transport Data and Data Programs

The four categories of airline statistics—financial, operational and traffic, pricing and fees, and safety—provide different levels of detail in terms of coverage, periodicity, and focus. The federal government collects the majority of publicly available aggregated airline statistics directly from air carriers. The Air Transport Association (ATA) reports member data on a quarterly and monthly basis on airfares, a cost index, and passenger and cargo traffic.⁵ The International Civil Aviation Organization (ICAO) collects international air data covering 188 countries.

Financial Data

BTS, through its Office of Airline Information (OAI), collects air carrier financial data, including income data and limited carrier operating expense data. Fuel constitutes the industry's second-largest operating expense. Air carriers report to OAI monthly on the gallons and costs of fuel consumed, which allows OAI to calculate and publish the average price per gallon. BTS also collects, on a monthly basis, the weighted average number of full time employees per labor category and the maintenance costs for flight equipment for carriers by aircraft type.

ICAO reports annual data on the fleets and personnel of both international and domestic scheduled and nonscheduled carriers. The statistics cover the number and types of aircraft operated, their capacity and utilization, and the numbers of airline personnel by job category and the annual expenditures for these personnel. ICAO also collects financial data from international scheduled airlines on annual revenues and expenditures, year-end assets and liabilities, retained earnings, and summary traffic data. [5]

Operational and Traffic Data

Operational and traffic airline statistics cover information such as aircraft departures performed, delays, fleets, aircraft revenue-miles, and passenger and cargo revenue-hours. Additional data cover information, such as revenue passenger-miles, available seat-miles, revenue ton-miles (for passengers, freight, and mail), segment and market traffic and capacity by airport-pair, and enplanements by airport.

Particular OAI data collections include flight on-time data, other service quality indicators (mishandled baggage and oversales), enplanements by airport and airline, airplane- and passenger-miles, and financial data. Monthly data submitted to OAI include arrivals, departures, and aircraft revenue-miles and hours. Twelve of the largest U.S. air carriers report to OAI on scheduled and actual arrival/departure times and cancellations by flight number and day of month. Finally, information on aircraft operating expenses by aircraft type and changes in aircraft fleet inventory are also reported to OAI.

The Federal Aviation Administration (FAA) collects data in a variety of areas of aviation including the Operations Network database, which captures flight operations and delays; the Air Traffic Control System Command Center, which monitors air traffic in real-time over the continental United States as well as providing airport status information to the public; and the Enhanced Traffic Management System (ETMS), which contains detailed information on flights for which a flight plan has been filed. FAA uses the ETMS to strategically manage traffic flow in the National Airspace System, balancing demand and capacity to avoid congestion that eventually translates to delays.

Every year, FAA conducts the General Aviation and Air Taxi Activity Survey, which provides estimates of the number of active aircraft, hours flown, primary use, and many other characteristics by aircraft type. Additional sources of information are ATA reports on passenger and cargo traffic, as well as ICAO collections of monthly traffic data for major international airports that include international aircraft movements, number of passengers embarked and disembarked, and tons of freight and mail loaded and unloaded.

Pricing and Fees Data

OAI's Passenger Origin and Destination (O&D) Survey samples 10 percent of all airline tickets used in scheduled service and reports on the full ticket itinerary, operating and ticketing carrier on each coupon, fare basis codes, and total dollar value of the ticket. BLS publishes both a Consumer Price Index (CPI) and a Producer Price Index (PPI) for airfares. The CPI measures changes in the prices paid by consumers for domestic and international airline trips, including taxes and any distribution costs not received by the carriers (e.g., travel agents' fees). The PPI measures changes in revenues received by producers of domestic airline trips only. Monthly prices for the two programs are gathered from different data sources: CPI prices come from the SABRE system,⁶ while PPI prices are gathered directly from airline pricing departments. Additional sources of this category of information are available from ATA, which reports on member passenger prices.

Safety

The National Transportation Safety Board (NTSB) investigates the causes of accidents and uses the information gathered to make safety recommendations to the transportation industry.

The National Airspace Incident Monitoring System is a repository that includes the FAA Accident/Incident Data System, the Near Mid-Air Collision System, the Pilot Deviation System, and other aviation safety-related databases. The Accident/Incident Data System contains NTSB's recommendations to FAA and FAA's responses.

The Aviation Safety Reporting System (ASRS) is a voluntary, confidential, and anonymous incident reporting system that collects information used to identify hazards and safety discrepancies in the National Airspace System. Data from the ASRS are also used to formulate policy and to strengthen the foundation of aviation human factors safety research.

Options for Filling Air Transportation Data Gaps

The discussion above illustrates the variety of air transport data currently available. However, many gaps exist. One important gap is the multimodal nature of passenger and freight air travel, that is, movement to and from airports. Passengers arrive at airports in cars (their own or taxis), by transit, by shuttle and intercity buses, and, to some extent, by rail. Nearly all air freight arrives by truck, but the extent of prior movement is not known. Some of these data may be captured but in data systems with different formats, definitions, and data elements, increasing the challenge to integrate the data into single trips.

One major challenge in the collection of airline statistics is air carrier concern about revealing proprietary information that could put them at a competitive disadvantage. A second challenge is that the immense volume of data that could be collected by carriers is expensive to house and analyze.

In collaboration with BLS, BTS is investigating a new method of computing price indices for air travel. The research aims to produce an airfare index series based on actual transaction prices, because the current BLS price indices for air travel do not capture the effects of special discounts (e.g., Internet specials or frequent flyer awards) on price trends. The BTS series will also provide more geographic detail; for example, estimate indices for particular cities of itinerary origin to facilitate local area economic analysis. However, greater detail is needed for the O&D data to be able to distinguish different products or services and their associated prices.

Options for improving air transportation data include:

• Traffic data on air taxis (on-demand air carriers) and corporate jets, when combined with information from NTSB, could be used to conduct exposure/risk analyses. The information would also be useful in assessing the degree to which these forms of air transportation substitute for traditional carriers.
• On-time statistics for smaller carriers and international flights could be used to calculate delays and travel time for an important segment of the U.S. aviation industry and would benefit consumers of these services.
• Flight-specific data would allow more accurate estimates of the performance of the airline industry, including the cost to passengers of flying, the time required to travel, the reliability of travel times, and the number of passengers affected by airline delays. Currently, however, BTS is unable to collect these data because of a legislative prohibition on collecting passenger data by flight number.

Transportation Economic Data

Transportation economics refers to industry performance on key economic measures such as prices, quantities, productivity, and externalities. It looks at not only how the industry performs directly in meeting the needs of its customers, but also how it affects the economy as a whole, based on, for example, measures of employment, output, and international competitiveness.

Existing Economic Data

Prices

Many customers focus on the price they pay for transportation as the key indicator of how well the transportation system is meeting their needs. The costs of passenger transportation rise and fall in line with prices in the overall economy but are particularly affected by energy prices and the productivity of individual modes. Good data exist on the component costs of automobile transportation, which represents the bulk of passenger movement. Reasonably good data are available on the price of passenger rail transportation and transit.

On the freight side, there are some data available on all modes of transportation, but these data often do not show the actual prices paid on particular routes or for particular commodities. Without this information it is difficult to determine which customers benefit from lower prices, even when average prices go down. Price data are also often not adjusted for changes in quality, such as delivery time and reliability.

See box 2 for Transportation Services Output Indices

Quantities

Quantities are the physical amount of transportation produced and thus measure the level of mobility that the transportation system enables. Quantities (e.g., passenger-miles or ton-miles) are essential to understanding the impacts of transportation on the economy and to making investment and other economic policy decisions.

Reasonably good data on physical movement, except for trucking, are publicly available, but details on specific routes, particular times, and vehicle types are not. Data on highway transportation of freight and passengers, which account for a substantial portion of all movement, have gaps in completeness, accuracy, and timeliness. Better data are needed for truck ton-miles and highway passenger-miles, as well as those for transit, intercity bus, and Amtrak (discussed earlier under Freight Data and Passenger Travel Statistics).

Investment

Transportation investment includes public and private infrastructure and vehicles and affects both the capacity and condition of the transportation system. Capital stock is the accumulated stock of these investments over a period of years, reflecting new investments and depreciation of existing investments. Information on capital stock is essential for measuring the productivity of the transportation system and its effect on economic growth.

Reasonably good data are available on most components of private capital stock: railroads, privately owned automobiles, trucking, airlines, waterway operators, pipelines, privately owned local transit, transportation services (e.g., travel agents and freight brokers), and the transportation-related capital stock of in-house transportation companies, as well as public highways and streets. All of these forms of transportation capital stock have increased with the growth of the economy over the last 50 years, with the exception of railroad capital stock, which declined (in inflation-adjusted terms), and privately owned local transit, which declined (in inflation-adjusted terms) until recently.

Productivity

Productivity measures how effectively economic inputs are converted into output. Labor productivity measures the output per hour of labor. Multifactor productivity, which is a more comprehensive measure, relates changes in output to changes in all inputs, including capital, labor, energy, materials, and services. If the transportation system becomes more productive, then fewer resources are needed and prices can decline.

Labor productivity data are widely available for a large number of industries. Multifactor productivity data are not as widely available for the economy as a whole and are only available for the railroad industry in the transportation sector. Railroad multifactor productivity increased more rapidly than that of either manufacturing or business over the last decade.

Externalities and Regulation

Transportation often generates negative effects such as pollution, collisions, congestion, and noise, which are known as externalities and are the driving force behind regulation and many other public policies affecting transportation. Measures of the costs of external effects of transportation (e.g., on safety, congestion, and the environment) are important in determining whether various kinds of regulation of transportation are appropriate.

The physical quantity data described earlier can be used to determine the amount of total pollution and noise emissions, but not necessarily the exact locations of these emissions or their costs. The physical quantity data also can be used to analyze the growth of congestion, although questions remain about the impacts on travel patterns of urban growth, system and demand management policies, and changes in technology. In addition, congestion tends to be very site- and time-specific, so estimates or policy decisions may need to be based on specific local data.

Historically, transportation has also been subject to economic regulation. Reduced economic regulation over the last 20 years has led to increased competition and lower prices for consumers. However, some carriers have found it difficult to stay in business in the face of increased competition. There is also considerable debate about whether transportation mergers have reversed some of the benefits of deregulation. Data on prices, quantities, investment, and productivity are crucial to understanding these trends.

Impact on the Economy

The transportation system links producers and consumers to markets. Public and private investment and other improvements to the transportation system increase capacity, reduce cost, and improve service. Improvements in freight transportation allow economies of scale in production, diversity in products available to consumers and sources available to producers, and increased competition in markets. All of these factors lower the cost of goods Americans buy and broaden the markets for the goods produced.

Similarly, passenger transportation affects several areas of the economy. Improved local commuter transportation supports broader labor markets, increasing the choice of jobs and economic efficiency. Better intercity transportation allows greater opportunity for personal travel and makes possible better management of geographically decentralized business organizations. Transportation also represents a major source of demand for many industries, both for suppliers to the transportation industry and producers of complementary services, such as the hotel and restaurant industries. Transportation thus creates efficiencies throughout the general economy, increasing incomes, providing higher paying jobs, and enhancing the international competitiveness of the economy.

Considerable debate continues, however, about appropriate means of measuring these effects. While data are available on several aspects of the overall efficiency of the transportation system (e.g., the productivity and physical flow data mentioned earlier), data are lacking on other important aspects, such as the efficiency of specific port and intermodal yard facilities and their related surface transportation links.

BTS's Transportation Satellite Accounts (TSAs) provide a broad measure of the value of transportation services to the economy, including both for-hire transportation and a large portion of in-house transportation, such as trucks owned by manufacturing companies. BTS continues to develop the TSAs as a way of measuring the more direct impacts of transportation on the economy.

Options for Filling Transportation Economic Data Gaps

Research is underway at BTS to fill some of the known gaps in economic data. One project will improve the quality of data on airline transportation prices with the development of an Airfare Index, covering both domestic and international travel. This research could eventually be extended to other modes.

Other key projects include the following:

• expansion of Transportation Satellite Accounts data to include additional modes, such as private barge, rail, and air, and updating of the household production of transportation services (e.g., local errands in a private automobile and people commuting to work) component;
• augmentation of available privately owned capital stock data by developing data on publicly owned capital stock, such as airports, waterways, and transit systems; and
• development of multifactor productivity measures for modes other than railroads to better measure transportation industry productivity in relation to the rest of the industry sector of the economy.

Geospatial Data

Geospatial information technologies have become increasingly useful decisionmaking tools for the transportation industry and agencies responsible for transportation planning. Previously used only by expert operators on specialized mainframe systems, they are now available for desktop systems and distributed computing services that give nontechnical users access to spatial analytical tools. One example of this technology, geographic information systems (GIS), provides desktop computer applications, Internet mapping services, and other information technology systems designed to store, display, and analyze geographically referenced information (geospatial data). This technology enables the presentation of masses of data in more useable forms (see map).

Geospatial data describe both manmade and natural features on the earth's surface, such as roads, rivers, and political boundaries. These features reference the geographic coordinates or location on the earth and also include associated characteristics that describe the features. These characteristics may define the type of road, river depth, or political boundaries. These data for use in a GIS are collected through a number of methods including the use of Global Positioning System (GPS) satellites and receivers to collect accurate geographic locations, remote sensing technology to capture earth images, and conversion of tabular data or paper maps through digitization.

As the use of geographic information technologies spreads, the range of applications expands. Today, GIS is used to track urban crime patterns, coordinate emergency response efforts, visualize data critical to national security, and support city zoning decisions. From urban planning support systems to on-board global positioning systems (GPS) and mapping technologies for public transport vehicles and automobiles, GIS tools provide a means for displaying and analyzing transportation data in real-world physical contexts. Public transportation agencies use these tools to plan system services and expansion, as well as maintain maps and analyze service areas. Geographic information technology enables freight and passenger carriers to warehouse data on operating resources and model the flow of goods and passengers across the transportation system.

Existing Geospatial Data and Data Programs

BTS creates, maintains, and distributes geospatial data through the National Transportation Atlas Database (NTAD) Program. These data are obtained from multiple sources and include the National Highway Planning Network, a national rail network, public-use airports and runways, and Amtrak stations. In addition, the NTAD includes state, county, congressional district, and metropolitan statistical area boundary files to provide a geographic reference for transportation features.

Geospatial information depicting transportation infrastructure can be stored and managed in a GIS for development and maintenance planning. For instance, the National Bridge Inventory maintained by FHWA contains information on structurally deficient bridges. Information describing the location and bridge conditions can be displayed cartographically and analyzed. Geographically accurate maps can be produced using a variety of data tables or layers placed one on top of the other to show geographic relationships (see map).

BTS in partnership with the DOT Office of Intermodalism and FHWA developed GeoFreight: The Intermodal Freight Display Tool. GeoFreight calculates and measures the intensity of infrastructure use for intermodal facilities (e.g., airports, seaports, and truck-rail interchanges). The information is displayed on regional or national maps. A CD-ROM version of GeoFreight is expected to be issued in late 2003.

New transportation applications for GIS continue to be developed, such as emergency response planning for crisis management. New applications place a premium not only on timely development of geospatial data but also on commonly agreed-on standards for data quality, content, and technology. As GIS becomes a more mainstream information technology tool, issues of open access to geographic data and software interoperability will grow in importance.

Organizations such as the Federal Geographic Data Committee (FGDC) and the Open GIS Consortium work to facilitate the wide availability and implementation of standardized geospatial data and technologies. FGDC is an interagency committee of representatives from the Executive Office of the President and cabinet-level and independent agencies. FGDC, in cooperation with state, local, and tribal governments, the academic community, and the private sector, is developing the National Spatial Data Infrastructure (NSDI). The NSDI is composed of geospatial data themes of national significance; documentation describing the data, standards (metadata), and partnerships; and the National Spatial Data Clearinghouse, which provides access to documented geospatial data and metadata from a variety of publicly available data sources [4].

The Geospatial One-Stop, a federal e-government initiative that extends the goals of the NSDI, facilitates public and private access to geographic data, with particular reference to the framework themes of the NSDI. In keeping with the trend toward development of data standards, Internet GIS applications, and online data warehousing, the Geospatial One-Stop effort will result in the creation of an online portal through which users can freely access geospatial data from the federal government and other sources, as well as gather information on spatial analytical tools and data standards. The Departments of the Interior, Commerce, and Transportation are the federal agencies with lead responsibilities for coordinating the national coverage and stewardship of specific NSDI framework data themes. The Bureau of Transportation Statistics (BTS) is the operating administration within DOT responsible for coordinating the development of the NSDI transportation standards.

Options for Filling Geospatial Data Gaps

Geospatial information technologies and geospatial data have become permanent fixtures in the information technology landscape, and the variety of Internet GIS applications and data available to the public and private sector will continue to develop and expand. Key areas for future geospatial data and standards development critical for transportation analysis include land-use planning, employee-based travel pattern analysis, and fine-grained data on infrastructure and operations critical to transportation safety and security analysis. Specific areas include:

• integration of bridge, tunnel, and transit data into the National Transportation Atlas Database;
• development of a North American Transportation Atlas Database covering the United States, Mexico, and Canada;
• expansion of the current web-based mapping center to enable customers to generate interactive maps and spatial analysis to support their projects and requirements; and
• provision of technical assistance/training workshops on geospatial data to state DOTs and metropolitan planning organizations.

¹In 2003, functions of the Immigration and Naturalization Service, the U.S. Customs Service, the Animal and Plant Inspection Service, and the Border Patrol were combined into the U.S. Department of Homeland Security's Bureau of Customs and Border Protection.

²Prior to the 2001 NHTS, the NPTS was conducted in 1969, 1977, 1983, 1990, and 1995, and the ATS (or similar surveys) was conducted in 1963, 1967, 1972, 1977, and 1995.

³The completed survey sample consisted of 2,698 nondisabled respondents and 2,321 disabled respondents.

⁴A comprehensive final report by the Expert Panel is expected in late fall 2003. As of August 2003, three letter reports had been issued by the Committee, covering the NHTS, Omnibus, and CFS.

⁵ATA members are common carriers in air transportation of passengers and/or cargo that operate a minimum of 20 million revenue ton-miles annually and have done so for 1 year preceding the date of application, and operate under a valid certificate issued by the Federal Aviation Administration [2].

⁶A commercial system that provides users with airline schedules, availability, pricing, and other information, as well as reservations and ticketing capability.

Conclusion

A focused yet comprehensive transportation statistics program can be concentrated in five key areas: freight, passenger travel, air transportation, transportation economics, and geospatial information. The preceding review shows that, while a wealth of data exist to inform stakeholders about the state of transportation, much work remains to be done. Data need to be collected or collected differently, relevant linkages among datasets need to be established, and data need to be analyzed and offered in ways useful for stakeholders at all levels of government and the private sector.

Sources

1. Air Transport Association, Airlines in Crisis: The Perfect Economic Storm, available at http://www.airlines.org/public/industry/binAirlinesInCrisis.pdf, as of July 2003.

2. ______, data available at http://www.airlines.org/public/home/default1.asp, as of March 2003.

3. American Bus Association, Motorcoach Census 2000 (Washington, DC: 2000), available at http://www,buses.org/industry/ABA-RLBanksReport.pdf, as of February 2003.

4. Federal Geographic Data Committee, available at http://www.fgdc.gov, as of March 2003.

5. International Civil Aviation Organization, data available at http://www.icao.org, as of March 2003.

6. Jung, S., J. Larkin, A. Toppen, and K. Wunderlich, "Urban Congestion Reporting," draft report prepared by Mitretek for the U.S. Department of Transportation, Federal Highway Administration, January 2003.

7. Texas Transportation Institute, 2002 Urban Mobility Study (College Station, TX: 2002).

8. Transportation Research Board, "Letter Report to the Bureau of Transportation Statistics," Mar. 20, 2003.

9. ______, The Relative Risks of School Travel (Washington, DC: 2002).

10. Travel Industry Association of America, TravelScope Survey (Washington, DC: 2003), also available at http://www.tia.org/Travel/, as of February 2003.

11. U.S. Department of Commerce, U.S. Census Bureau, 1997 Economic Census: Vehicle Inventory and Use Survey (Washington, DC: 1999).

12. U.S. Department of Commerce, Office of Travel and Tourism Industries, Survey of International Air Travelers, available at http://tinet.ita.doc.gov/, as of February 2003.

13. U.S. Department of Labor, American Time Use Survey, available at http://www.bls.gov/tus/#docs, as of March 2003.

14. U.S. Department of Transportation, Bureau of Transportation Statistics, Annual Historical Air Traffic Statistics, 1981-2001, available at http://www.bts.gov/oai/, as of March 2003.

15. ______, Implications of Continuous Measurement for the Uses of Census Data in Transportation Planning (Washington, DC: 1996).

16. ______, Transportation Statistics Beyond ISTEA: Critical Gaps and Strategic Responses (Washington, DC: January 1998).

17. ______, Border Crossing Data, available at http://www.bts.gov/itt/cross/, as of February 2003.

18. ______, National Transportation Availability and Use Survey for Persons with Disabilities, 2002, information on the survey available at http://www.bts.gov/omnibus/, as of April 2003.

19. U.S. Department of Transportation, Federal Highway Administration and Bureau of Transportation Statistics, 2001 National Household Travel Survey, Preliminary Data Release Version 1, available at http://nhts.ornl.gov/2001/index.shtml, as of February 2003.

20. U.S. Department of Transportation, Federal Transit Administration and Community Transportation Association of America, Status of Rural Public Transportation: 2000 (Washington, DC: April 2001).

21. U.S. Department of Transportation, National Highway Traffic Safety Administration and Bureau of Transportation Statistics, National Survey of Pedestrian & Bicyclist Attitudes and Behaviors, Highlights Report (Washington, DC: nd).