One thing I didn’t see mentioned so far is that nowadays in the US TOFC (trailer) loadings is only a smallish fraction of COFC (container) loadings.
The latest figures I could get are from an AAR 2016 PDF (which apparently I cannot hot-link to)
Class I railroads, in million units, 2015 preliminary:
2013: COFC - 11.35; TOFC - 1.48
2014: COFC - 11.97; TOFC - 1.53
2015: COFC - 12.23; TOFC - 1.48
So over the past few years, TOFC has been roughly a tenth of COFC loadings in the US.
Reasons for this ratio have been debated on many railroad forums, but in general:
– Containers haulage involves less dead weight - no chassis or wheels. This is kind of the reason Road-Railer service (Trailers with train wheels) never really took off, and in fact has been reduced to only a few traffic lanes.
– Double stack - 6 containers on freight cars of less length (and less tare weight) than the equivalent for 5 trailers (trailers have been mostly hauled using spine cars over the past decade or so)
– Finally, not all trailers are designed for the stresses of loading/unloading via lifting (the older roll-on/roll-off circus-style loading is incompatible with spine cars, to say nothing of the few trailer capable well cars around). A lot of the 89ft TOFC flats from the 1980s have been repurposed for hauling pipe, beams, trusses etc.
Maybe one or another of the various TOFC loading schemes being trialed in Europe will resolve enough issues that North American railroads will adopt it, and and the percentage of TOFC traffic will increase again.