USA Banner

Official US Government Icon

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure Site Icon

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

U.S. Department of Transportation U.S. Department of Transportation Icon United States Department of Transportation United States Department of Transportation

Development of FAF4 Database

Friday, September 23, 2016

2.1 OVERVIEW

The FAF is built on the CFS data, while integrating additional data to estimate volumes of shipments from many industries that were not covered by the CFS.  Because it is CFS-based, 2012 CFS definitions of the 132 domestic areas, the 43 commodity codes (SCTG 2-digit), and the modes of transportation were adopted by the FAF4.  The only exception is a “no domestic” mode, which was added to represent imported crude petroleum that is processed at the “dock” thus not involving any mode transport domestically.  Similar to previous releases, the 2012 CFS captured shipments that accounted for approximately 70% of the FAF4-estimated total volumes by dollar value.  As shown in Figure 2–1 , the FAF base-year matrix consists of shipment flows from two major groups: domestic shipments captured by the CFS and out-of-scope (OOS) shipments from many industry sectors or trade (both foreign and domestic).  In addition to the 2012 CFS data, major data sources utilized in estimating flows of OOS shipments are also presented in Figure 2–1.  These data were analyzed and, in most cases, coupled with sector-specific models and algorithms to generate freight flow details as required in producing the FAF database.

If you are a user with a disability and cannot view this image, please call 800-853-1351 for further assistance.

Figure 2–1.  Components of the FAF4 database.

2.2 CFS DOMESTIC

Being a U.S. shipper-based survey, the CFS does not include imports.  While the CFS does capture exports, there is a data quality concern on its exported shipment flows, mainly due to sample size limitations.  As a result, only domestic shipments from the CFS were used in constructing the FAF. 

Because of a combination of data suppression for confidentiality reasons, limited sample size, potential large standard errors in estimates, as well as limitations to the scope of the CFS (across industrial sectors), many cells in the CFS matrix where flows may exist were suppressed (i.e., missing).  A prior study of the complete set of CFS data indicated that a good set of available data matrices could be used to estimate data gaps within the CFS matrix.  This includes the most detailed information of the published matrices (Tables 25), which reports annual tons and dollar values shipped by origin, destination, mode, and 2-digit SCTG.  Other CFS tables also provide various dimensional (e.g., 2- or 3-dimention) data, including flows broken down to the CFS-domestic geographic regions of interest.  To separate domestic and exported shipments from the 2012 CFS data, the Census Bureau provided a set of special CFS tables that broke out domestic shipments from exports.

Detailed discussions on methodology developed for filling missing cells in the domestic CFS flow matrix, under the creation of FAF4 database effort, are further described in Section 4 of this document. 

2.3 CFS OUT OF SCOPE AREAS

As pointed out previously, FAF4 takes into account shipments from establishments not covered by the 2012 CFS.  This includes shipments from industries classified in transportation, construction, most retail and service industries, farms, fisheries, foreign establishments (imports), petroleum and natural gas extraction, municipal solid waste, logging, as well as household and business moves.  

The OOS-component databases for FAF4 were constructed based on various data sources, including administrative records (e.g., Foreign Trade Data) and other industry-based data.  For missing or suppressed data elements–either due to disclosure concerns or geographic limitations of the source data files–models and imputation techniques were employed to estimate those data components.  Specifically, modeling approaches such as iterative proportional fitting (IPF), log-linear, and spatial interaction models were often used during the FAF4 estimation process.  Several improvements, over the FAF3 process, were applied in estimating flows of commodities for OOS sectors in FAF4. 

As seen in Figure 2–1 , like its predecessors in 1993, 1997, 2002, and 2007, the 2012 CFS does not include shipments originating from several OOS business sectors, including:

  • Farm-based Agriculture
  • Fisheries
  • Logging
  • Construction and Demolition Debris
  • Municipal Solid Waste (MSW)
  • Services
  • Retail
  • Household and Business Moves
  • Crude Petroleum
  • Natural Gas

In addition to industry sectors listed above, foreign trade (i.e., imported and exported shipments) are also included as an OOS area in the FAF process.  Data sources and flow estimation procedures for each of the OOS components are covered in several sections following the discussions of the domestic CFS component in this document.