You are here

Rail Freight Carloads and Intermodal Traffic

Rail Freight Carloads and Intermodal Traffic

Source: Weekly Railroad Traffic, compiled by Association of American Railroads (private, proprietary).


Transportation Services Index (TSI) data for rail freight are compiled from the Association of American Railroads publication, Weekly Railroad Traffic (WRT). The data presented in WRT show weekly carload and intermodal traffic originating in the United States and Canada and are not a direct measure of ton-miles. Canadian railroad traffic reflects Canadian and U.S. operations and the operations of U.S. subsidiaries. U.S. traffic reflects the U.S. and Canadian operations of U.S. railroads (this applies to both carload and intermodal traffic).


BTS converts WRT weekly data to monthly numbers by deriving an average daily number from the weekly data and then multiplying that number by the number of days in the month after adjusting for the number of weeks and days in that particular month. There are two separate indexes: (1) carload traffic and (2) number of intermodal units. The two indexes present two somewhat different aspects of monthly freight activity (due to different commodity mixes, traffic, etc.).

Why not exclude intermodal units from the rail freight index, because carloads make up the majority of rail freight traffic? The reason is that carload traffic is comprised mostly of bulk commodities, such as coal (46% of rail tonnage in 2001), agricultural products, and nonmetallic minerals and products, while rail intermodal transports a huge range of goodsfrom bicycles to automotive parts, lawnmowers to glassware, greeting cards to bottled water, and toys to computers. Through globalization manufacturers supply chains have become widely distributed geographically, and rail intermodal plays a critical role in making supply chains more efficient for retailers and other firms and industries. If intermodal units are excluded, rail freight economic activity may not be properly tracked.

Data Quality Questions


Coverage is sufficient even though not all U. S. railroads report to Weekly Railroad Traffic (WRT). All U.S. Class I railroads report, but data are submitted by only 8 out of almost 500 non-Class I railroads. U.S. Class 1 Railroads are those earning revenues of at least $289.4 million in 2004, as defined by the Surface Transportation Board. Although, this is a voluntary response survey, the WRT covers 87 percent of U.S. carloads and 96 percent of U.S. intermodal units. Inclusion of Canadian railroad U.S. operations accounts for 96 percent of U.S. carloads and 100 percent of U.S. intermodal units. This is the only available source of this data.

  • Is this data source a frame or sample?
    WRT is a sample of all U.S railroads. The accounting departments for the reporting railroads report all revenue movements during the week to the Association of American Railroads (AAR).
  • Does the sample cover the entire frame? Or is some group missing or underrepresented in the sample? For the excluded group, what percentage of the total do they make up?
    Not all U.S. railroads report to AAR. All U.S. Class I railroads submit data. But there are almost 500 non-Class I railroads in the United States, and only 8 of them submit data to AAR.
    Reporting to the WRT is voluntary. Using the Surface Transportation Boards Rail Waybill Sample to estimate how much of total U.S. traffic is covered in the WRT, AAR concluded that 87 percent of carloads and 96 percent of intermodal loads are covered in the WRT. However, this does not include the U.S. operations of Canadian railroads. Traffic for the former Illinois Central, Grand Trunk Western, and Soo line (among others) is now reported by their parent companies, Canadian National (CN) and Canadian Pacific (CP). All of these cars originating in the United States are now reported in the Canadian railroad lines of the WRT. CN and CP do not separate out the carloads originating in the United States from the rest of their traffic in the WRT. When U.S.-originated CN and CP traffic is added to the other U.S. traffic, the WRT covers 96 percent of carloads and 100 percent of intermodal originated in the United States, although it is spread over both the U.S. and Canadian sections.


  • Are the data available monthly?
    Weekly data are converted to monthly numbers by BTS.
  • How soon are the data available after the month is over?
    There is a 2-week lag time after the end of the period for the weekly data.
  • Are the data easy to access and use?
     Yes, the report comes via email through paid subscriptions.
  • Are the microdata available for use?
    They may be available with subscription/purchase.


  • Are there duplicate records?
    There do not appear to be duplicate records for weekly carload and intermodal traffic.
  • Are there outliers in the data?
  • Are data missing for individual records? If so, how are they identified?
    AAR says there are no missing data for individual records in the WRT. All reporting carriers send a full set of data every week.
  • How accurate are the key data fields?
    For the time they are reported, the data are accurate. Each car origination is counted once. Two railroads frequently sent minor revisions to the current week 2002 data during 2003. Other railroads send revisions to their data when errors are discovered, which happens infrequently. Revisions to the traffic data may be submitted by reporting railroads through the end of the following calendar year. When a railroad is unable to submit its traffic figures for the current week, the AAR repeats figures from the last period. These figures are replaced with carrier-reported figures as soon as possible, and cumulative figures and weekly data incorporate these changes.
  • Are variances available for this data source? If so, what method was used to calculate variances?


  • Are the data comparable over time within the data source? If not, can data be made to be comparable (e.g., combining two data series)?
    Yes, they are comparable. Individual week comparisons to the previous year in the WRT are made to the week which ended 52 weeks earlier.

Other Questions and Important Information

  • Is sufficient documentation available for the data source?
    All documentation included is in the report, and more is on AARs website ( (link is external)).
    • Data dictionary:
      Information is not available to BTS.
    • Detailed description of the methodology:
      AAR has an explanation along with the explanation of WRT.
    • Estimation methods:
      Information is not available to BTS.
  • Are other sources available for the same data?
  • Who is the contact for the data source?
    Association of American Railroads, 202/639-2323, (link is external).
Updated: Thursday, March 22, 2018