Grubel and Lloyd (1975) developed an index, now known as the Grubel-Lloyd index, for measuring the fraction of trade flows that take place within industries as opposed to across sectors. Using this decomposition they showed that in many countries the majority of trade is intra-industry rather than intersectoral." H. Helpman, 2011, Understanding Global Trade, page 102.
As I explained in the previous post, the finding that "the majority of trade is intra-industry rather than intersectoral" explains why models of the New Trade Theory have known such a success since they typically analyze this trade of varieties between similar nations. When I teach this story to my students I always ask, as every teacher, "Have you got any questions?", they don't have, I beg "please, try to criticize", they don't. It's a pity because many questions can be ask, such as "how do you define exactly an industry?", this is an important question because if you compute the Grubel and Lloyd indicator at different levels of the international classification of goods, you obviously don't have the same results (see below the comparison between 3-digit and 5-digit from Brulhart, 2009).
When you see this number, you realize there is "only" 43.7% of intra-industrial trade at the world level (at the 3-digit level) which means that there is still a large amount of trade that is intersectoral.
Furthermore, in the sentence of Helpman ("in many countries the majority of trade is intra-industry"), after questionning "the majority of trade" we also need to question who are these "many countries" between which intra-industry trade occurs.
Intrasectoral trade concerns mainly rich and intermediate countries (see the Figure below from Brulhart (2009), line Hi-Hi and Med-Hi), representing respectively around 40% and 30% of their trade. In contrast, one can observe that there is very few intra-sectoral trade between similar countries that are poors or between countries with intermediate level of incomes (Low-Low, Med-Med).
You also have to be skeptical about what are "similar products", indeed since the 80s, the literature has improved its description of intra-industry trade and has observed that a bulk of that trade concerns goods that are vertically differentiated. So don't mix up the theoretical models of the 80s that analyze trade of horizontally differenciated varieties and the reality. On the Figure below, from Fontagné, Freudenbeg and Gaulier (2006) you can see that vertical intra-EU12 trade has increased between 1980-99 while trade of varieties differenciated horizontally has been constant.
With Cabral et al. (2006) we can conclude that "the distinction between trade in horizontally and vertically differentiated products [...] seems to be as important as the distinction between inter-and intra-industry trade” (pg. 562).*
Another important point (quite related): many of the models that we use in the classroom speak about trade of final products. In fact, a large share of trade concerns trade in intermediate goods, and this is also true to explain intra-sectoral trade, see below the share of intra-sectoral goods by product group.
As a soundtract of this post, because he just won a well deserved Nobel Prize in literature, Bob Dylan, "Don't Think Twice, It's All Right".
References and sources: Figure 1, 2 and 4 comes from Brülhart (2009), An Account of Global Intra-industry Trade, 1962-2006, The World Economy, vol. 32(3), pages 401-459, 03. Figure 3 comes from Fontagné, Freudenberg and Gaulier, (2006) A Systematic Decomposition of World Trade into Horizontal and Vertical IIT. Review of World Economics.
*I have borrowed this quotation of Cabral et al. to Azhar and Elliott (2011) that proposes an interesting measure of trade induced adjustment in volume and quality space.