Foreign direct investment in industrial transition: the experience of Vietnam

Abstract

This paper examines the role of foreign direct investment (FDI) in the process of industrial transition in Vietnam under market-oriented reforms. The findings are consistent with the conventional wisdom that concomitant liberalisation of trade and investment regimes, accompanied by creating a congenial environment for market-based decision making by the private agents, is vital for reaping developmental gains from FDI. During the 1990s growth of employment in firms with foreign capital participation lagged behind output growth, reflecting the capital-intensity bias in domestic-market-oriented production in a partially liberalised economy. This pattern has changed notably in recent years as the product mix gradually became export oriented in response to further liberalisation of trade and investment policy regimes.

Keywords:

• Vietnam
• foreign direct investment
• economic transition
• multinational enterprises

Keywords:

• JEL classifications: F23
• O53
• P23

Acknowledgement

We are indebted to the anonymous referee for excellent comments and suggestions.

Notes

1. In Vietnamese policy documents, the term ‘FIE’ is used to refer to all firms with foreign capital participation, regardless of the size of the foreign equity-capital share. Of the total number of FIEs approved during 1989–2009, 83% were fully foreign owned (Statistical Yearbook of Vietnam 2010, Hanoi: General Statistical Office, Table 62).

2. Key references, which also provide useful listing of related works, include Huang 2003, Naughton 2007, Lankes and Stern 1997 and MacBean (ed.) 2000.

3. A business cooperation contract is an agreement between a foreign business partner and a local firm in Vietnam (in most cases a state-owned company) under which the former operates the given business venture on a profit sharing basis while the ownership of the venture rests with the latter.

4. The impact of this law was swift and remarkable; there was a four-fold increase in the number of private enterprises over the period 2000–2006 compared to 1991–1999 (CIEM 2008, Ngo 2006).

5. The estimated value of withdrawn foreign investment licenses in 1996 was US$2875 million, by far the highest value of withdrawn licenses in a single year during the period (1988–1998) covered in this study.

6. Investment figures in this paragraph are from UNCTAD, World Investment Report database (www.UNCTAD.org).

7. Figures given here were computed from the investment approval records (unpublished) of the Ministry of Planning and Investment, Hanoi.

8. Figures reported in this paper, unless otherwise stated, are compiled from the Statistical Yearbook of Vietnam, Hanoi: General Statistical Office (various issues).

9. Hitachi plant in Ho Chi Minh City commenced operation in 2000. It currently employs about 4000 workers.

10. International buying groups are world-wide purchasing organisations of large retail chains in developed countries, which specialise in world-wide purchase of consumer goods such as apparel, toys and footwear.

11. Disaggregated data are not reported here because of the space constraint. They are available on request from the authors.

12. See Jenkins and Sen (2006) and Harrison and Rodriguez-Clare (2010) for surveys of this literature. We believe that our approach is preferable to working with an optimising model derived from first principles assuming a ‘representative’ firm. This approach, notwithstanding its analytical elegance, cannot adequately address issues that arise from imperfect information and heterogeneity relating to industry characteristics and government policies (Kirman 1992).

13. Note that, as the model was estimated using all variables expressed in natural logarithm, the percentage equivalent for any dummy coefficient is [exp(dummy coefficient)−1] × 100.

14. See Harrison and Rodriguez-Clare (2010), Section 5.1 and the works cited therein.

15. In both equations, the coefficient of the labour variable (L) is positive but not statistically significant. As the referee suggested, we examined (using the Cook test (Cook 1977)) whether the estimate is influenced by the presence of some outlier observations in the data series and found no evidence of such bias. Thus, this seemingly counterintuitive result may presumably reflect the fact that the number of workers is not a good measure of the actual contribution of labour in the production process. However, we believe that the measurement error of this variable is unlikely to bias the coefficient estimates of the other variables because the estimation method allows for industry specific effects.