In previous posts I’ve discussed some of the estimates of aggregate trade elasticities. Some new work presented at a recent IMF conference on Chinese trade suggests that we may need to revise some of our views on the efficacy of yuan appreciation for inducing expenditure switching.

To place matters in context, recall in this post on the enigmatic yuan, I discussed the estimates provided by Marquez and Schindler [pdf]. The point estimate that they obtain is perverse for Chinese imports — that is imports fall in response to a yuan appreciation.

On a bilateral (US-China) basis, similarly odd results are obtained. Estimates by Mann and Plueck [pdf], Thorbecke [pdf] and Cheung, Chinn and Fujii are discussed in this post.

Some of the low price elasticities — especially for Chinese imports — were rationalized on the basis of the idea that domestic value added in Chinese exports was quite small. Brad Setser ([1], [2], and [3]) has taken issue with this view, characterizing it as outdated. The debate goes on, partly because empirical work on the issue is hampered by data limitations.

In one of the papers entitled “Measuring the Vertical Specialization in Chinese Trade” [pdf], presented at the conference, Judith Dean, K.C. Fung (a former colleague) and Zhi Wang write:

In recent years, two important related developments have transformed the nature of world trade: the explosive growth of Chinese trade, and the growth of vertically specialized trade due to international production fragmentation. The literature in each of these two separate topics is large and growing. However, very few papers quantitatively assess these two trends together. In this paper, we measure the degree to which Chinese trade has become vertically specialized, using a new measure adapted from Hummels, Ishii and Yi (2001). By making use of the latest Chinese input-output tables, and a new detailed Chinese trade data set which distinguishes
processing trade from other forms of trade, we develop a new method of identifying intermediate goods imported into China. With this new method, we measure Chinese vertical specialization over time, and by sector, export destination, and input source. We find about 35 percent of the value of China’s exports to the world is attributable to imported inputs. This vertical specialization exceeds 50% in some sectors, and is growing over time.

The authors conclude:

While trade fragmentation and China’s rapidly growing trade have been recognized as important economic phenomena, the importance of fragmentation in China’s trade growth has been
left unexamined until recently. In this paper, we provide the most up-to-date and comprehensive measures of the degree of vertical specialization in China’s trade, using a new detailed Chinese
dataset which allows us to distinguish processing imports and exports from ordinary trade. We further utilize the United Nations Broad Economic Categories (BEC) system to identify nonprocessed
imported intermediates from imports of final goods These data are incorporated into the Hummels, et al. (2001) measure of vertical specialization, using both the 1997 and 2002 benchmark
Chinese IO tables. We then quantify vertical specialization over time, by source, by trading partner, sector, and type of firm.

Our results show that the vertical specialization in China’s exports to the world was more than 30% in 2002. In the sectors with the most fragmented trade–plastic products, steel processing, communication equipment, industrial machinery, metal products and computers—vertical
specialization exceeded 50%. Not surprisingly, the firms with the most fragmented trade are the foreign-invested enterprises. There is strong evidence of the importance and persistence of an Asian
supplier network to China. About 58% China’s aggregate vertical specialization in 2002 was attributable to imports from Japan, Taiwan, South Korea and other Southeast Asian economies. Among bilateral partners, China’s vertical specialization was high (about 30%) with not only the US, the EU 15, and Canada, but also with Taiwan and Singapore. Our evidence also suggests that the
fragmentation in China’s global trade is growing, particularly its trade with the US and the EU.

Our results are generally robust to the use of alternative methods of measurement. However, for some specific sectors, there can be some variations across methods. For example, in 1997, for motor vehicles, the extent of fragmentation was 34% using our own DFW method, but it was 31.9% using the BEC method and 35% using the IO approach. These variations could become important for our future work since we intend to econometrically determine the how the VS shares can vary with
different variables as suggested in the theoretically literature. The alternative measures give us an opportunity to test the robustness of our future estimates.

Figure 5 from the paper exhibits the degree of vertical specialization by industry. No surpise that the highest degree of vertical specialization is in plastic products, communication equipment and electronic computers. Still, even in that last category, domestic value added is almost one-half.

The authors provide broad trends up to 2005; it is interesting that they assert that vertical specialization is increasing, although at a decelerating rate. This is shown in Figure 7 from the paper.

What are the key implications, from the perspective of global adjustment? There are two. First, the relatively small price elasticities estimated for Chinese exports are consistent with a high degree of vertical specialization. Second, the considerable extent of Chinese value added in many of the categories of Chinese exports provides some scope for expenditure switching by yuan appreciation (although the impact on US-China bilateral trade would likely be fairly minor, for the reasons I’ve given before.)

Another paper presented at the conference had a slightly different message. From the abstract of Li Cui and Murtaza Syed, 2007, “Is China Changing It’s Stripes? The Shifting Structure of China’s External Trade and its Implications”
[pdf]:

Using disaggregated trade data, this paper assesses how the expansion of China’s production capacity and its changing production structure may be affecting its trade linkages with other countries. It argues that although processing trade still accounts for an important share of its total trade, China has moved away from traditional assembly operations and its exports have started to rely more on domestically sourced components, particularly in less sophisticated product categories. As a result, China’s imports and exports have to a large extent delinked in recent times, with imports of parts and components increasingly driven by domestic production needs rather than final demand abroad. This increased domestic sourcing may be one of the important factors behind the recent increase in China’s trade balance. In addition, as China moves up the value chain, both its imports and exports have become more sophisticated than in the past. This shift in the product composition of China’s trade bundle
could imply that its trade balance has become more sensitive to demand changes in foreign markets and to relative price changes. Indeed, the empirical results suggest that China is becoming more exposed to fluctuations in the strength of the global economy, and that changes in its exchange rate could have a bigger impact on the trade balance and the domestic economy than commonly believed. The paper also discusses some potential implications that this shift in China’s trade and production structure is likely to have for other
economies in the region, as well as for the world.

Since the posted paper is preliminary, I won’t quote further. I will note that the results pertain to only a subset of Chinese manufactures, in so far as the methodology requires matching specific imported inputs to specific outputs in the SITC system.

The bottom line is great for academics. More research needs to be done!

Technorati Tags: China, imports,
exports,
US-China trade deficit,
renminbi,
exchange rate,
vertical specialization, value added, and
expenditure switching,