"پیچیدگی" صادرات : اندازه گیری تجاری جدید

Trade data are often classified by product characteristics. We propose a new classification “sophistication” as a means of distinguishing between products. We construct a sophistication index based on the income levels of exporting economies. Sophistication captures a range of factors including technology, ease of product fragmentation, natural resource availability, and marketing. We calculate sophistication scores at the 3- and 4-digit levels and test how far the index relates to existing technological classifications of products. We use the index to examine trade patterns and illustrate how it can be applied in the analysis of export performance of individual economies.

Researchers seeking to explain trade flows or to assess the competitiveness of individual economies have long sought ways of classifying and categorizing the vast and ever increasing range of products that enter into world trade. Classification by factor intensity and technological intensity are two popular ways of addressing this issue that has yielded useful insights. A basic problem with both approaches however stems from the relatively high level of aggregation at which both can be applied. Factor input data will normally come from input–output tables or industrial censuses, typically at a 2-digit level, while technology classifications are normally based on R&D expenditure by the manufacturing sub-sector. Trade data, on the other hand, are available at highly disaggregated levels: for instance, the Standard International Trade Classification (SITC) Revision 2 has 236 items at the 3-digit level and 778 items at the 4-digit level. In the 2002 version of the Harmonized System, the new trade data recording system, there are over 1 200 items at the 4-digit level and over 5 000 items at the 6-digit level. In contrast, in relation to R&D expenditure, US data are available for only 38 manufacturing industries (NSB, 2002), while the OECD Science, Technology, and Industry Scoreboard provides cross-country R&D data for 19 industries. Given the extensive and disaggregate information on products that enter international trade, it is useful to have a means of distinguishing between similar products produced by different countries and to have a broad indication of product differences at a highly disaggregate level. With this aim we put forward a new method of classifying exports that does not require industry data, but only information on exports of each product and per capita incomes of exporting countries. We term this a “sophistication index” and calculate sophistication scores for 1990 and 2000 for 237 products at the 3-digit level (of SITC Revision 2), and for 766 products at the 4-digit level.1 We show how the index can be used in comparisons of export structure and in discussions of individual country competitiveness. Section 2 explains the calculation of our index. Section 3 discusses individual product scores at the 3- and 4-digit levels and relates the index to one well-known set of technology categories. Section 4 discusses the results when product data are aggregated for individual countries and the potential use of the index in discussions of export promotion and competitiveness strategies. Finally, we draw some conclusions.

‘Sophistication’’ provides a new way of ana- lyzing trade and location patterns and tracking export composition and competitiveness in developing countries. Its main advantage is that it can be calculated quickly at any level of detail and for any period. It can provide unique scores for products at disaggregated levels. Its main disadvantage is that it is not a specific technology measure: it captures many other factors affecting export location, and care is needed in interpreting the results. —Sophistication correlates relatively poorly with technology and in particular its impact is diluted by fragmentation, which allows technology-intensive activities to locate exports in countries that received theory would not predict. However, all ‘‘fragment- able’’ activities (with discrete processes) do not fragment to similar extents: sophistica- tion provides a useful tool to map this and identify activities with ‘‘location inertia.’’ This can lead to further research on the causes of inertia—economies of agglomera- tion, links with innovation systems, special skill needs, government policy, and soon—and provide insights to countries that wish to attract or upgrade those industries. —Resource-based exports have the obvious pattern that has little to do with income levels (that is countries export products for which they have the resources). However, there are exceptions caused by technological factors (accumulated skills), marketing and brands, and government protection and sub- sidization. —In the aggregate sophistication does not have a strong relationship with growth rates. Exports by richer countries do not grow rel- atively fast: industrial catch-up means that exports by poorer countries are likely to grow more rapidly, aided by relocation of activities within global value chains. Thus, there is widespread (but not universal) ‘‘de-sophistication’’ of manufactured prod- ucts. —Exports at the bottom of the sophistication scale largely from poor countries do not grow rapidly. Most products in this category have a low income elasticity of demand and may be suffering from declining prices. The poor- est countries lack the industrial capabilities to move into more attractive products or attract hi-tech production networks. Low wages per se are not the driver of relocation but low wages for technically proficient workers, backed by modern infrastructure, suppliers, and other capability and institu- tional factors needed for modern industry. Bythesamereasoning,havinghighpercapita incomes is not a guarantee of a sophisticated export structure. Countries may become rich without building advanced industrial skills and capabilities; doing this requires specific strategies. This is clearly illustrated by the contrast between the low sophistication export structure of Hong Kong, China, one of the richest countries in East Asia, and the more advanced export structures of other mature Tigers in the region. —The ‘‘de-sophistication’’ process is thus highly skewed. It is largely a shift from high- to middle-income countries, not to the poorest ones (a differently constructed sophistication index could illustrate this). There is a bulge in the middle of the global sophistication scale, where the largest and most dynamic exports are concentrated. —Aggregate sophistication scores by coun- try conform broadly to expectations, with South Asian economies near the bottom of the scale.While this paper is only a first cut at using the sophistication index, it suggests that the tech- nique can be useful in several ways. First, the scores are product specific and can be as de- tailed as needed. This allows the index to be applied as a continuous variable for econo- metric analysis. Second, in conjunction with technology and value chain information, sophistication data can be used to analyze fragmentation and location inertia. Third, our sophistication index can be used as a prelimin- ary step for country competitiveness analysis, allowing a rapid mapping of location shifts inexports of interest to a country. More broadly, it can show if a country’s sophistication struc- ture of its exports is in line with its income level. Although the normative implications of this data are limited, they provide a starting point for more detailed analysis. We argue that the index provides a useful, new way of organizing very detailed trade data. More work is needed to refine the index and construct other indices (geared to middle and low income countries or to finer levels of prod- uct detail). It is hoped that this preliminary exploration will stimulate such work