China’s export profile is a fascinating topic – hence the frequent debate and discussion. The latest contribution is by the FT‘s Guy de Jonquieres:

Now China’s manufacturing miracle is entering a second phase, as producers start to drive aggressively up-market. Exports of aircraft parts, ships, microchips and cars all grew by about 70 per cent last year, more than four times faster than traditional exports such as shoes and clothing. At the same time, it is set on becoming more than a processing economy that just screws together components made elsewhere. With official backing, industries such as steel, vehicles and electronics are steadily raising the value added locally by performing more advanced processes and making more parts in China.

That description echoes the message the FT‘s Richard McGregor has pushed repeatedly, but how much does the growth rate of sophisticated exports tell us about China’s development? China exports a lot of labor-intensive manufactures, so we would be shocked if their growth rate was anywhere near 70 percent. Reporting the level as well as the growth rate would substantially aid our understanding of China’s move up the value-added chain.

Newspapers are unlikely to print stories proclaiming “China’s exports not as sophisticated as some claim,” so we have to turn to the academic literature for contrary assessments. In a working paper titled “Measuring the Technology Content of China’s Exports,” Bin Xu argues that China’s export profile has not been so remarkable:

First, we reexamine the finding of Rodrik (2006) that China is an outlier in terms of technology sophistication of exports. Without adjusting product quality in the measurement, our data confirms Rodrik’s finding: China is an outlier in every year from 1989 to 2000. However, using our quality-adjusted measure of TCE [technology content of exports], China disappeared to be an outlier after the mid 1990s. This result highlights the importance of considering the product quality dimension in evaluating the technology content of China’s exports. By ignoring the low quality nature of Chinese exports, Rodrik (2006) overestimated the technology sophistication of Chinese exports.

Second, we find that the product-level TCE of China is significantly lower than that of a country at the same development level in every year from 1991 to 2001. Moreover, the gap had been widening steadily during the period. Low TCE at the product level is driven by low product quality. This result is a mirror image of Schott’s (2006) finding of increasing price discounts of Chinese exports.

In the paper we identify several new features that help to explain the pattern and trend of the technology content of Chinese exports. First, at both the industry and product levels, we find a positive correlation between the growth rate of an industry/product’s export share (in China’s total exports) and its TCE level. On average, industries/products with higher TCE levels expanded more in exports.

Tragically, econometric analyses will always lag behind anecdotes – Xu’s work covers 1989-2001. So the recent FT reports may be right, but it’d be nice if their coverage were more informative — please describe not only the trend but also the current state of play! Similarly, Brad Setser casts his assessment in terms such as China “increasingly has the export and industrial production profile of a middle income country” and “China no longer just imports parts… for final assembly” rather than commenting on the levels.

When will China actually have the export profile of a middle-income country? When will less than X percent of China’s exports be labor-intensive low-quality manufactures? Obviously it is difficult to construct, measure, and obtain data needed to answer such questions. That’s one reason why the discussion often centers upon more easily accessed numbers, such as those used by the FT reporters. But we would learn a lot by contextualizing them with a few more pieces of information.

[Update: Emmanuel of IPE Zone points to this World Bank update. It also describes the increased importance of non-processing exports without reference to level.]