One is not enough! Understanding world trade collapses

In order to put the mainstream narrative for the recent world trade collapse into a consistent economic-historic framework, this paper builds a comparable data set for the analysis of world trade collapses consisting of 72 periods of import decline (27 in the 1930s; 45 in 2008-10). The empirical analysis explains about three quarters of cross-country variance and supports the emerging professional consensus that identifies the decrease in domestic demand and the share of manufacturing trade as key determinants of the severity of a world trade collapse, but the paper also reveals significant differences between the 1930s and the 2000s. Both the demand shock and the composition effect are comparatively speaking less important in the recent trade collapse. The paper identifies country-specific determinants (level of development, political system and openness) that have not (yet) been considered in the mainstream narrative for the recent world trade collapse. In line with the theory of trade uncertainty (van Marrewijk and van Bergeijk 1993, van Bergeijk 2010), I find that more democratic, more open, and wealthier countries reduced their imports to a smaller extent.