Foreign-Invested Enterprises

Evidence presented in the previous section attests to the importance of trade in intermediate goods to the PRC's overall relationship with the US and Japan. An equally important aspect of the PRC's trade is its reliance on FIEs as the source of import demand and export supply. Fueled by reforms that legalized various types of non-state-owned enterprises but retained limitations on domestic credit, FIEs became the main vehicle for the PRC's integration into the global economy. Currently, FIEs consist of fully-funded foreign enterprises (FFEs), Sinoforeign contractual joint ventures (CJVs), and Sino-foreign equity joint ventures (EJVs). American and Japanese firms have invested heavily in foreign affiliates in the PRC, both to gain access to the PRC's labor for processing, and as a platform to serve the growing PRC consumer market. From 1998 to 2007, American FDI in the PRC averaged US$3.9 billion per year. Over the same period, Japanese FDI in the PRC averaged somewhat more at US$4.3 billion per year. Though far outweighed by FDI from Hong Kong, China, these investments are large, persistent, and indicative of the rapid integration of production between the PRC, Japan, and the US.

Why firms choose to operate foreign affiliates rather than service a market through exports from the home country is the subject of a large theoretical literature. Markusen (1984) introduced the proximity/concentration trade-off as an explanation for why firms choose a particular mode of entry into a foreign market. Serving a foreign market through a local affiliate replicates the home production process abroad but saves on transport costs. However, if there are economies of scale in production, exporting may be more profitable than serving the local market through FDI because production remains concentrated in a single location. This “horizontal” approach has been extended and tested by Brainard (1997), Markusen and Venables (1998, 2000), and Helpman, Melitz, and Yeaple (2004), among others. An alternative “vertical” approach follows the work of Helpman (1984), who shows that if there are increasing returns to scale in “headquarter services” and crosscountry differences in factor prices, a firm may split the production of headquarter services and manufacturing across countries. Helpman's model predicts that the extent of multinational activity will be increasing in relative factor endowment differences across countries. Empirical evidence on the vertical explanation for multinational activity is provided by Yeaple (2003a) and Hanson, Mataloni, and Slaughter (2005). Some empirical research, such as that done by Yeaple (2003b), combines both the vertical and horizontal motive for foreign investment and attempts to capture complementarities between the two forms of activity.

Because this literature focuses on the mode of entry into a foreign market, empirical studies typically rely on detailed information on foreign affiliate sales to the host country domestic market. As noted by Greaney and Li (forthcoming), there are significant differences in the extent to which American and Japanese firms use their PRC subsidiaries to serve the local market. Their analysis reveals that in 2003, 70% of sales by US majority-owned non-bank manufacturing affiliates in the PRC were to the local market. Less than 8% of their sales were exports to the US. In contrast, in 2003 only 46% of sales by Japanese majority-owned non-bank manufacturing affiliates in the PRC were to the local market, while 34% of sales were exports to Japan. These data are consistent with Markusen's (1984) framework, in that transportation costs are larger for American firms than for Japanese firms and, thus, may be an explanation for the greater intensity with which American firms use local affiliates to serve the PRC market rather than to act as export platforms.

Foreign firms play an important role in the PRC's position in the global supply chain. Table 4 [ PDF 13.6KB | 1 page ] shows the share of the PRC's exports to various destinations that is classified as processing trade, the total share of exports carried out by FIEs, and the share of processing trade that is performed by FIEs. While PRC exports to the US are somewhat more likely to be processing exports than are exports to Japan (62.5% versus 56.6%), for both destinations a remarkably high share of total exports is carried out by FIEs: 68.3% for exports to the US and 67.7% for exports to Japan in 2007. Secondly, and perhaps less surprisingly, the majority of processing trade is performed through FIEs, as 86.2% of processing exports come from FIEs to the US and 86.9% from FIEs to Japan in 2007. Less obvious, however, is the extent to which total FIE trade is processing trade. Detailed examination of the data shows that, for many industries, FIEs do very little trade other than processing importing and exporting.

When we consider how transport costs might drive differences across industries, we note that Krugman (1980) identifies a “home-market effect” for industries producing differentiated products. While transport costs lead firms to locate production as close as possible to final consumers, fixed costs associated with differentiating products in response to consumer tastes encourage the concentration of production in a single location. As a result, there is a tendency for a differentiated-products industry to concentrate production in the country with the larger market for home varieties, making the home country the net exporter of differentiated goods.

Hanson and Xiang (2004) tested the home-market effect with international trade data organized into two groups of industries: those with high transport costs and more differentiated products, and those with low transport costs and less differentiated products. In Table 5 [ PDF 15.4KB | 1 page ], we provide characteristics of PRC imports from Japan and the US for industries divided into the two groups identified by Hanson and Xiang. Because transport costs to the PRC are much larger for American firms than for Japanese ones, we might expect to see the influence of firm-level fixed costs more clearly in Japan-PRC trade and, thus, the influence of the home-market effect in these flows. The left panel provides the share of PRC imports from Japan that are ordinary imports and the share that are processing imports, for the three-digit SITC industries classified as having low transport cost and low differentiation (top panel) and high transport cost and high differentiation (bottom panel).¹² Interestingly, we see that the PRC's imports from Japan of the highly differentiated products are much more likely to be ordinary trade, and less likely to be processing trade, than the low-differentiation products. This evidence suggests that, for Japan-PRC trade, the home-market effect may indeed explain some of the differences in the ordinary trade shares across industries.

In contrast, the same comparison using US values in Table 5 shows very little difference between the two industry groups. The PRC's imports from the US in both these categories are more likely to be ordinary trade, and less likely to be processing intermediates than are the PRC's imports from Japan. Moreover, for imports from the US, the share of trade that is ordinary trade is nearly identical across the two industry groups. That no difference appears between high-differentiation and low-differentiation industries suggests that transport costs between the US and the PRC dominate sourcing decisions.

Some caveats should be kept in mind. First, the differential results for the US and Japan across the product groups may reflect their different positions in the global supply chains. Second, as emphasized by Ferrantino et al. (2008), there is considerable heterogeneity within these sectors and the Hanson and Xiang dichotomy may not reveal true differences in transportation costs across goods.

Download this Paper [ PDF 171.8KB| 30 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.