Contractual Frictions and Organizational Form

While the importance of processing trade is evidence that firms have broken the production process into several fragments, some component production is done within the firm while some involves arm's length transactions. Recent theories of organization and trade, drawing upon models of contractual frictions, seek to understand which activities take place within the firm's boundaries. Thus, they address a narrower aspect of the data: whether production of intermediates takes place inside the firm (in-sourcing) or outside the firm (outsourcing). They presume that production is fragmented, rather than trying to explain how fragmented it is.

Simply observing the extent to which trade is mediated by FIEs does not capture the extent to which an American or Japanese firm controls production decisions. However, some insight into firm boundaries can be obtained through recognition of a unique feature of PRC customs data. Processing imports are subdivided by PRC customs into two categories: process and assembly (P&A), which is conducted mostly by SOEs and FIEs; and processing with imported materials (PWIM), which is largely conducted by FIEs. The key distinction between these two customs regimes is control over inputs: with P&A, the PRC firm receives materials and processes them according to orders taken from the foreign firm, while with PWIM, the PRC firm has full control of decisions related to input sourcing, production, trading, and financing. The PRC firm involved in these relationships may be an SOE, an FIE, or a private domestic enterprise.

Use of this PRC customs distinction to study firm boundaries was exploited by Feenstra and Hanson (2005), who developed their approach from recent advances in the study of imperfect contracts. Since production of final goods may require highly customized and specialized intermediate inputs to be produced by input suppliers, its quality may not be verifiable by a third party. In such cases, the final good producer and the input supplier may find it impossible to write a complete contract specifying the price-quality relationship. Moreover, even if such a contract could be written, it may not be enforced by the judicial system. Thus, the division of the economic surplus from production and use of the input may be subject to ex post facto bargaining between the final good producer and the input supplier or producer. This can result in what, in the literature, is called a “hold up” problem: distortions in the incentives for investment and effort in input production because the producer is able to get only a fraction of the returns to his or her investment or effort. If a hold up problem occurs, the input supplier will provide less investment or effort than is optimal for the maximization of the joint production surplus. Such a situation may also characterize resources provided by the final good producer.

Helpman (2006) argues that intermediate inputs supplied by the final good producer suffer less from agency problems than intermediate inputs that require the engagement of suppliers. Also, the effective bargaining power of the final good supplier is higher under integration than under outsourcing as, under the former, the final good supplier has some control over inputs and can recover some of the value of the final good if bargaining fails. This is good from the point of view of getting the resources provided by the final good producer as close to the optimum level (i.e., that which maximizes the joint surplus from production) as possible, but it adversely affects the level of activity of the input producer. Thus, incentives are closer to optimal under integration when goods require intensive use of headquarter services (provided by the final good producer) in their production, while outsourcing is better when goods require intensive use of specialized inputs (provided by the intermediate good producer). If (i) headquarter services are capital intensive and input production is labor intensive, and if (ii) the two have to be combined in the same country to produce a specific tradable intermediate input used in the production of a given nontradable final good (and there are many different intermediate and final goods), then we should see, ceteris paribus, a positive correlation between the share of intra-firm imports of a country and the capital abundance of the exporting country (Helpman 2006; Antràs 2003).

Note that intra-firm imports are correlated with vertical FDI, while interfirm imports are correlated with offshore outsourcing. With vertical FDI, inputs are being produced in the same multinational firm as the final output. This is not so with offshore outsourcing. Thus, according to this theory, in less capital-abundant countries such as the PRC and India, we should see relatively more offshore outsourcing than offshoring through FDI. Of course, legal institutions are weaker in these countries than in the more capital-abundant countries, and that will be an offsetting force as it influences contract enforcement.

How well these theories fare with actual experience in the PRC is the subject of Feenstra and Hanson's (2005) exploration of ownership and control in PRC processing trade. The authors built a simple model of international outsourcing and applied it to the PRC. They considered a multinational firm that had decided to set up an export-processing plant in a low-wage country. In this arrangement, the firm sends intermediate inputs to a processing factory, which converts the inputs into finished goods and then exports the final output. The decisions facing the multinational firm include who should own the processing factory and who should control the input-purchase decisions the factory makes. Feenstra and Hanson posit that parties use control rights over productive assets to ameliorate hold up problems created by incomplete contracts. Their model predicts that the joint surplus generated by the partnership depends on model parameters, including the specificity of investments and contracting costs, which they estimate.

Feenstra and Hanson did not observe the value of surplus from outsourcing activities directly. Rather, they used the share of processing trade accounted for by each contractual type to represent the probability that a particular contractual arrangement will be chosen.¹³ Comparing these shares in the PRC's total processing exports over the period 1997–2002, they found that multinational firms tend to split factory ownership and input control with local managers. The most common form, as evidenced by trade shares, is to have foreign factory ownership but PRC control over input purchases.

Following Feenstra and Hanson (2005), we calculated the shares of processing trade by contractual type for 1996 and 2007. In 1996, as shown in Table 6 [ PDF 13.4KB | 1 page ], multinational firms engaged in export processing tended to split factory ownership and input control with local managers. A little over one quarter of processing exports to the US came from PRC owned factories operating under P&A arrangements (foreign control of inputs), while nearly two thirds came from foreign owned factories operating under PWIM arrangements (local control of inputs). A similarly small share of processing exports to Japan (17.8%) come from PRC owned factories operating under P&A arrangements, while the largest share (56.3%) come from foreign owned factories operating under PWIM arrangements. These results indicate similar patterns across the two bilateral relationships and are close to the results found by Feenstra and Hanson for all trade in the 1997–2002 period: 27% of processing exports were produced in PRC factories under P&A arrangements and 49.6% were produced in PRC factories under PWIM arrangements.

Looking at data from 2007 in Table 6, however, we do see some changes in contractual arrangements over the decade. The dominant form of processing trade continues to be foreign-owned factories with PRC managers controlling input decisions. This type of processing trade accounts for more than three quarters of processing exports to the US and nearly as much to Japan. However, the second largest form of processing trade—processing and assembly by PRC firms—dwindles in share. Instead, there is a growing share of processing exports by foreign-owned firms with foreign control over inputs. This pattern appears with respect to both destination countries, but is more dramatic with respect to the US.¹⁴

Further differences between the US and Japan emerge when we dig a bit deeper in the data and distinguish between foreign firms by type. As noted above, not all FIEs operate under the same organizational forms. In FFEs, foreign control over production decisions is complete. With EJVs or CJVs, control is shared between the foreign investor and the PRC partner. In Table 7 [ PDF 12.4KB | 1 page ], it can be seen that PRC processing exports by foreign-owned firms destined for Japan are more likely to come from firms with foreign control over inputs than those destined for the US, regardless of firm type. But there is also variation across firm types. When the exporter is an FFE (i.e., a completely foreign-owned firm), the share of processing exports to the US under processing and assembly is 15% (8.9/59.7), in contrast to about 20% for exports to Japan. When foreign ownership is shared, as under an EJV, about 7% of processing exports to the US are from firms with foreign control over inputs, in contrast to 27% of exports to Japan. Finally, when joint ownership is stipulated only contractually (i.e., in CJVs), the shares of processing exports to the US and to Japan under processing and assembly are 23% and 57%, respectively. This decomposition suggests that foreign firms undertaking processing exports to Japan are progressively more likely to retain control over inputs as their control over the ownership of the firm declines.

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

Comment(s)

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