Abstract
This paper analyses regional Foreign Direct Investment (FDI) in Thailand. Its determinants are identified using a unique assembled panel dataset comprising all 76 Thai provinces during the period 1985–2005. The work is premised on the view that foreign investors consider various determinants of location choices in each province (including labour costs, Gross Provincial Products (GPP) per head, areas of industrial estates, communication and transportation issues, population characteristics, educational attainment, population density and distance from the centre of town to the main ports of Thailand) as well as government incentives, before deciding to undertake FDI in a given region. Our econometric model estimates suggest that government regional policy, and the effect of zoning in particular, however, has a significant and positive effect on regional FDI, drawing FDI to those zones where the greatest incentives are on offer. The other possible determinants of FDI amongst regions are largely shown to be insignificant.
Notes
1. One province is omitted due to it affecting the high variation between the values and the number of FDI projects.
2. To which amendments were made in 1991, 1992 and 1993.
3. Office of Industrial Economics. Ministry of Industry and Faculty of Economics, Thammasat University (Citation1996, p. 19).
4. Regarding the Business Act of the Kingdom of Thailand, all firms are enforced to register with the Business Promotion Department, Ministry of Commerce; however, foreign investors may invest through Thai partners as nominees.
5. For further details on the Thailand LFS background and methodology see Lo‐Utai (Citation2003).
6. Before 2003, the study reports were conducted separately between the private and public university and institution. The recent number of lecturers in higher education has been available online since 2004. Available at: http://www.mua.go.th
7. There are three major ports in Thailand including Bangkok, Laem Chabang (Chonburi) and Maptaphut (Rayong) Port. To avoid multicollinearity problems due to the relative closeness of the ports, the analysis employed in this study only uses distances to Laem Chabang Port.
8. In modelling terms, a fixed effect pooled model will not permit the inclusion of variables which are time‐invariant within groups as the fixed group effects are already incorporated in the constant term for each group.
9. We should record that tests on the equality of the dummies for Zones 2 and 3b could not be rejected at the 5% level, showing some inconsistency in the tests. However, given the findings of the other two comparative tests, we would suggest this is most likely an anomaly.