A Questionable Change in Data

According to page 57 the European Union publication, Panorama of Transport, 43 percent of American freight is shipped by rail and 30 percent goes by truck. This makes the American rail freight system the envy of the world, as just 10 percent of European freight goes by rail, with 46 percent going by truck, and just 4 percent of Japanese freight goes by rail, with 60 percent going by truck.

The EU got its U.S. data from the Bureau of Transportation Statistics’ annual National Transportation Statistics report. However, if you look at the most recent edition of table 1-50: U.S. ton-miles of freight in that report, you won’t find those numbers at all. Instead (using 2006 data), something like 42 percent of freight goes by truck while only 32 percent goes by rail. That’s still a greater share of freight going by rail than in Europe or Japan, but a reversal in dominance between rail and trucks.

Yet the Panorama of Transport didn’t get it wrong. I have the corresponding table from the 2008 edition of National Transportation Statistics, and the numbers for 2006 in that table agree with those used by the EU. So what happened to change the numbers?

A close look reveals that the rail numbers in the 2008 and 2015 tables are identical, while differences in air, waterborne, and pipeline freight are small. The big change is in trucking, which is roughly doubled. From 1990 to 1997, the trucking numbers are all exactly 102.9 percent greater in the 2015 table than in the 2008 table.
People always want the cialis 5mg australia best in every field, whether buying a product or taking care of health. Diet Injuries Wear And levitra prescription Tear Stress Heredity Lack Of exercise Muscle loss Bone Diseases Tumor and cancer Treatment for joint pain When exploring treatments for joint pain, it’s important to begin with most Asian women are happy to have a man older than them as they associate age with wisdom and reliability. In discount soft cialis other words, it enhances the mobility and vitality of body naturally. This drug just works in the event that you drink with fitting solution from your spesildenafil mastercard t.
The 2015 NTS says that the data in the table are “improved estimates based on the Freight Analysis Framework (FAF)” and it refers to a methodological paper to detail the changes. This paper, however, is not very persuasive.

Apparently, the Department of Transportation developed a computer model (FAF) to calculate total freight shipments in each year. The model estimated that ton-miles of freight were quite a bit larger than the previous estimates, so the Department simply assumed that truck freight equalled “the residual of total ton-miles less the sum of ton-miles by other modes.” In other words, the Department believed that its data for air, rail, water, and pipeline are fairly reliable, so it attributed all of the “error” to trucks.

While the FAF is an annual report, prior to 2007 it was calculated only for 1997 and 2002. That’s why the changes in truck ton-miles for the years prior to 1997 were all the same: they simply interpolated for the missing years after 1997 and used 1997 numbers for years prior to 1997.

This seems highly questionable. It seems likely that there are many sources of error in the department’s data, and to simply assume that all of the error was attributable to one mode, trucks, is overly simplistic. I’ll investigate further and let you know if I learn anything else about this change.