Specialization : Pro - and Anti - Globalizing , 1990 - 2002 ∗ †

Specialization alters the incidence of trade costs to buyers and sellers, with pro-and anti-globalizing effects on 76 countries from 1990-2002. The structural gravity model yields measures of Constructed Home Bias and the Total Factor Productivity effect of changing incidence. A bit more than half the world’s countries experience declining constructed home bias and rising real output while the remainder of countries experience rising home bias and falling real output. The effects are big for the outliers. A novel test of the structural gravity model restrictions shows it comes very close in an economic sense. JEL Classification: F10, F15, R10, R40. ∗To be presented at the NBER ITI meetings, Spring 2010. †Contact information: James E. Anderson, Department of Economics, Boston College, Chestnut Hill, MA 02467, USA; Yoto V. Yotov, Department of Economics, Drexel University, Philadelphia, PA 19104, USA. Specialization is a powerful force churning the world’s economies (76 countries) from 1990 to 2002, demonstrated in this paper by applying the structural gravity model. Increasing specialization is revealed by a fall in the correlation between national shares of world manufacturing output and expenditure. Trade costs inferred from gravity are large and vary over distance, so the shifts in the location of production and consumption must be changing ‘average’ trade costs. But what is ‘average’, and isn’t incidence what matters? Our measures of theoretically consistent average incidence are derived from the structural gravity model. They reveal both proand anti-globalizing effects of specialization — some countries become more open to trade while others become less open, all despite unchanging bilateral trade costs. The changing incidence has significant real output effects — some countries gain while others lose. The changing correlation between output and expenditure shares in total world manufacturing and for three 3-digit ISIC categories is illustrated in Table 1. For all goods and Table 1: Correlations: Output and Expenditure Shares (1) (2) (3) (4) Year All Manufacturing Apparel (322) Leather (323) Food (311) 1990 0.97 0.94 0.82 1.0