A puzzle about intra-industry trade: A reply

Abstract

regrouping U.S. trade data for 1929-60 into standard industrial classification (SIC) categories, Frederic Pryor identifies a behaviour pattern for intra-industry trade (IIT) which is seemingly inconsistent with predictions derived from our findings for the post-1970 period.1 Specifically, he finds that IIT is roughly unchanged between 1929-50, in contrast to its well-documented increase in the 1960s and 1970s and subsequent decline in the 1980s. He argues that our model would have predicted a decline in U.S. IIT, since (i) income, trade and discretionary income decreased; (i"i) bilateral trade increased; (wi) barriers to trade increased. We suggested in our article that reduced discretionary income should reduce consumers' taste for variety and thus reduce IIT; however, we were unable to provide empirical confirmation of the hypothesis. While better specified measures of income growth might have yielded statistically significant results, it is not clear from our study how much weight, if any, should be placed on the linkage running from lower incomes in trading partner countries to changes in IIT. Our approach also suggested that trade liberalization should increase IIT, ceteris paribus. By implication, the erection of trade barriers should reduce IIT. We also argued that the higher the level of bilateral trade, the smaller the increase in a country's IIT, when trade levels increase further. The reasoning was that with an already high level of bilateral trade, any significant increases in overall trade would draw in countries with heterogeneous tastes, exerting a downward influence on IIT. While we did not hypothesize and test a relationship between the change in bilateral trade and the change in IIT, by extension we would predict a positive rather than a negative relationship. If in fact the increase in barriers to trade was accompanied by an increase in bilateral trade, the overall impact on IIT might well be neutral or even positive.