Endnotes

¹This exercise cannot be conducted with regular trade data because imports are not necessarily used as inputs in the production of exports, but can also be consumed locally.

²In this paper, “province” encompasses all of the PRC’s first-tier administrative divisions: provinces, municipalities, and autonomous regions. We have excluded the Xizang Autonomous Region (also known as Tibet Autonomous Region) and Ningxia from our analysis because they had no processing trade in at least one year of our data sample. Furthermore, we treat Hong Kong, China; Macau, China; and Taipei,China as foreign economies.

³The sample correlation coefficient of –0.3274 is statistically significant at the 10% level. Note that the data used for this figure are adjusted for transshipments through Hong Kong, China. See Section 3 for further details.

⁴Yeaple (2003) uses the term “complex integration strategy.”

⁵The structure of our model closely follows that of Helpman, Melitz, and Yeaple (2004).

⁶We conduct our analysis for a single industry, but it is straightforward to embed our model in a general equilibrium framework with many industries.

⁷As is well known, where nⁱ is the number (measure) of varieties available in economy i and p^t(v) is the price of variety v.

⁸A debate has recently emerged on whether trade costs are truly a linear function of distance (e.g., Brun et al. 2005; Coe, Subramanian, and Tamirisa 2007). In this paper, we follow the standard approach in gravity regressions of using distance as a proxy for trade costs.

⁹We rely on this specific functional form because it ensures that the derivation is symmetric for advanced economies East and West. The key theoretical predictions of our model do not hinge on this specific functional form.

¹⁰As will become clear below, assumption 1 ensures that the unit cost of a type-P firm is sufficiently lower than that of a type-X firm so that aⁱ_x > aⁱ_p.

¹¹(t²w)^ε-1 > 1 implies that the unit cost of conducting both production stages at home for domestic sales is lower than offshoring final assembly to the PRC for re-import. While this assumption is not realistic, it significantly reduces the complexity of our analysis. Furthermore, relaxing this assumption is unlikely to change the two key theoretical predictions of our model.

¹²This graphical approach of presenting the results is adopted from Helpman, Melitz, and Yeaple (2004).

¹³This estimate of the import content share of the PRC’s processing exports is slightly lower than that of Koopman, Wang, and Wei (2008) because they also took into account indirect import content in their estimation.

¹⁴A positive coefficient on lnWage is in line with previous studies. Using data from 1996 to 1999 to examine the effect of labor costs on FDI location, Gao (2002) also obtained a positive coefficient on the OLS estimation of real wages.

Download this Paper [ PDF 340.3KB| 28 pages ].

Comment(s)

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