Desktop version

Home arrow Business & Finance arrow Globalisation, comparative advantage and the changing dynamics of trade

Stylized facts and anomalies

Relationship between diffusion and downstreaming in cross-sections and time series

Figure 8.1 presents a scatter plot of the relationship between HHI (diffusion) and EXPRELY (relative income level of economy with revealed comparative advantage) in 2006. For ease of interpretation, the names of the 15 economies are placed on the horizontal axis approximately at the level of their relative per capita income in 1997, as used to construct the index. A fifth-order polynomial is fitted to the data (Figure 8.1). The overall pattern is U-shaped. On the right, exports are concentrated in the highest-income economies, the United States and Japan. In the middle, exports are relatively diffused among all the economies, and associated on average with economies in the middle of the income distribution, e.g. Italy and Chinese Taipei. On the left, exports are concentrated in the lowest- income economies. While there are several of these, the left tail is accounted for primarily by concentration in China. For each of the 201 products with HHI > .25 and EXPRELY < .4 in 2006, China accounts for the largest market share. Of the outliers, some are clustered in upward-reaching “fingers” from the main U. These correspond to products that are concentrated in particular middle-income economies.

Figure 8.1. Scatter of HHI and EXPRELY, 2006

If taken from right to left, this pattern suggests something like the traditional Vernon product cycle (diffusion followed by downstreaming), followed by a final phase in which exporting is concentrated in China. This impression may be misleading, as figure 1 represents a cross-section and not a time series. Time-series behaviour may not be the same as crosssection behaviour.13 Thus, we approximate the typical dynamic behaviour of HHI and EXPRELY between 1997 and 2006 using flexible second-order polynomial regressions with dHHI and dEXPRELY as the dependent variables (Figure 8.2 and Table 8.A13).

The resulting dynamics are superimposed over the stylized U in Figure 8.2. The results suggest on average that during the period in question, exports of many products became both more concentrated and more extreme in terms of the level of relative income they were associated with. Products that in 1997 were associated with a level of EXPRELY above .8 became more concentrated and moved upstream toward either the United States or Japan. Products associated with an upper-middle level of income (France, Germany, Hong Kong, Singapore, United Kingdom) remained about where they were. At somewhat lower incomes (Italy, Chinese Taipei) the typical product downstreamed but remained diffuse, while products associated with income levels equal to that of Korea or lower experienced both downstreaming and concentration (in China). While there are many special cases among the products in question, the overall pattern is one of agglomeration of exports in one of the three largest economies - China, Japan, or the United States - for the products in question.

Figure 8.2. Fitted relationship between HHI and EXPRELY in 2006, showing dynamics from 1997 to 2006

< Prev   CONTENTS   Source   Next >

Related topics