Foreign outsourcing and the product cycle: evidence from micro data

Abstract

This article provides empirical evidence on the product cycle and the firm's make-or-buy decision by using a firm-level data set with a direct measure of foreign outsourcing. Across industries, foreign outsourcing tends to be inactive in R&D-intensive industries. Within each industry, products exported from the home country are on average significantly more R&D intensive than those outsourced to independent foreign firms. Products manufactured within subsidiaries at South tend to have medium R&D intensity. This ordering in R&D intensity is consistent with the theoretical prediction.

Acknowledgements

The Ministry of General Affairs of Japan allowed the author to access the micro-data files by issuing official approvals. Kei Nara and Mutsuharu Takahashi were helpful for the access. This research was partly supported by Grant-in-Aid for Scientific Research C16530152. Remaining errors are mine.

Notes

¹Though confidential firm-level data cannot be publicly disclosed, anyone can access to the same micro data as long as one obtains individual official permission from the government in advance.

²While all the firms with no less than 50 employees are surveyed, those with less than 50 are sampled with probability of less than one. However, this article will control for the difference in firm size.

³Tomiura (2005) explains the survey in detail and compares it with other outsourcing measures.

⁴Though the survey does not separate Southern and Northern firms among foreign suppliers, the outsourcing to high-wage North is supposed exceptional.

⁵To include a large number of firms owning no patents, this article adds the value one to the number of patents before taking the ratio. One cannot adjust differences in economic value or depreciation across patents due to lack of data. R&D expenditure is recorded in million yen.

⁶The firms are excluded from exporters if they record zero both in R&D expenditure and in patents because the product cycle is not appropriately defined for these exporters.

⁷Only majority-owned subsidiaries are included to concentrate on manufacturing affiliate as opposed to sales offices or portfolio-motivated FDI. The survey does not distinguish 100% among majority ownership.

⁸The model by Helpman et al. (2004) implies that firms producing within Northern subsidiaries must be on average more (not less) productive than exporters.

⁹The capital-labour ratio is defined as tangible fixed assets divided by the number of regular employees. The firms without data on capital are excluded from the columns (2) and (4).

¹⁰This result is also in line with other previous work. Acemoglu et al. (2004) report that outsourcing is more likely when the downstream industry is less R&D-intensive, though they do not discuss the issue of location.