Analysing trade by mode of transportation requires some background data work as a comprehensive global database on trade by mode of transportation does not exist yet. If the UN Comtrade database – a widely used dataset on international trade – contains a variable for modes of transportation, not all countries report this variable to the UN Statistical Division. For example, in 2022, around a third of all countries filled the mode of transportation variable for at least one trade flow at the HS 6-digits level. In this report, the UN Comtrade data were used for Australia, Brazil, Canada, and Türkiye. For the European Union, Japan and the United States, national and community sources were used, as they are not reported in the UN Comtrade dataset. However, some major economies, such as the United Kingdom or China, could not be included in the analysis, as trade data by mode of transportation is not publicly available at the moment. The countries included in the analysis database cover 29.5% of total value of exports and 33.4% of total import value in 2023.
Apart from gathering the data, some post-processing was applied to create a complete and comparable dataset. For the United States, two datasets were used: the Census data, provided by the US Census Bureau of the US Department of Commerce and the TransBorder dataset from the Bureau of Transportation Statistics of the US Department of Transportation. The Census data provides information on trade between the United States and all its trade partners by sea, air, and total mode of transportation. It does not contain specific information on rail, road and other modes of transportation. The TransBorder dataset compiles data on trade between the United States and Mexico and Canada by all modes of transportation. Since trade by rail and road to and from the United States occurs exclusively with Canada and Mexico, we merged those two databases to obtain dataset covering air, sea, rail, road and other modes of transportation.
However, these two datasets do not have the same coverage in terms of commodities. While the Census data provides trade flows at the HS 6-digits level, the TransBorder dataset reports trade at the HS 2-digits only. Since we need HS 6-digits to group trade by main industries in our analysis, we estimated trade by modes of transportation for HS 6-digits. To do so, HS 6-digits trade flows for which the mode of transportation was unknown (i.e. not air or sea), were distributed into rail, road and other modes proportionally using the shares of trade by modes for their parent codes at the HS 2-digits level. For example, in 2023, imported value to the United States with unknown modes of transportation for HS 050690 was USD 3 540 473. In 2023, for its HS 2-digits parent code (05), trade flows which did not occur by air or sea were shipped at 87.4% by road, 4.2% by rail, 0.2% by other modes and 8.1% by unknown modes of transportation. Thus, we split USD 3,540,473 according to those shares: USD 3 094 621 was shipped by road; USD 149 579 by rail; USD 8 280 by other modes; and USD 287 993 by unknown modes.
European data were also adjusted. Eurostat provides extra-EU trade by mode of transportation for the EU aggregate at the time being: it means that data before 2020 corresponds to EU-28 while data from 2020 corresponds to EU-27. To obtain a comparable time series, the trade between the UK and the rest of the world prior to 2020 is excluded and the United Kingdom as a partner is excluded for 2020 and after. This way, a consistent dataset is obtained across the periods, corresponding to EU-27 trade with the rest of the world except the United Kingdom. Ideally, UK trade to the rest of the world and added UK trade with the European Union before 2020 would have been subtracted to obtain EU-27 data at each time period, but the UK trade with the European Union is not available for each the period. This correction was done only when comparing across years that include dates prior to 2021. Figure 2.20 and Figure 3.1 are based on 2023 data as provided by Eurostat. Finally, Japanese and European data were converted to US dollar, using the OECD Exchanges Rates database.