In response to this post on 19th-century U.S. tariffs and economic growth , the ever-insightful Tim Worstall  sends me an e-mail with this important point:
[Consider this additional point, one found in] Power and Plenty , [by Ronald] Findlay and [Kevin] O’Rourke.
Tariffs themselves are not the determinant of trade barriers. The cost of transport should be included as well. We tend not to think of this these days as transport is so cheap (as Brad DeLong has ponted out, if you’re on the container network you are $5,000 per 40 tonnes of goods from anywhere else on the container network. These sorts of transport rates have implications for bulk materials, ores, base metals etc. But for just about anything manufactured they are near irrelevant. In fact, I’m told that 80% of the US consumption of iPods actually comes in air freight! And we certainly move the more expensive metals by air freight….$2,000, $3,000 per tonne from Russia to Taiwan say, as against $145,000 a tonne in value).
But while those tariffs were indeed rising to the McKinley peak transport costs were falling precipitately over the previous 50 years. The move from (for transatlantic) sail to steam lowered rates so much that, even despite tariff rises total trade barriers were falling. The evidence they use is that commodity prices converged over this time period, despite the rises in tariffs.
I always think it slightly strange that many people fail to consider transport costs: when much of the modern literature on inter-African trade makes much of that very expense of inter-African transport.