Japan-Vietnam Economist Club
The 10th WORKSHOP
29 May 2004 (Sat), GRIPS, Tokyo 14:00-16:00
Locational Determinants of FDI: The Case of Vietnam
by Mr. Le Viet Anh,
The tenth workshop of JVEC was held on May 29, 2004 at GRIPS with Mr. Le Viet Anh's presentation, "Locational Determinants of FDI: the Case of Vietnam." It is a summary of his master thesis which had been submitted to Nagoya University last year.
Alternative approaches to FDI were reviewed such as capital theory, international trade approach, market imperfection and industrial organization, Dunning's eclectic paradigm, and agglomeration economies. Mr. Viet Anh chose Dunning's eclecticism and agglomeration economies as his framework.
Vietnam's FDI statistics during 1988-2002 was presented, classified by source countries, sectors, forms of investment, and regional distribution within the country. Vietnam's classification of foreign investment (30% foreign owned or more) is different from that of IMF (10% foreign owned or more).
A model of FDI into different regions of Vietnam was set up using the following explanatory variables: market demand and market size (proxied by GRP per capita), infrastructure (proxied by telephone sets per capita), industrialization (regional industrial output/GRP), stock of foreign investment (cumulative FDI/POP), labor cost (wage in the state sector), labor quality (upper secondary school graduates), openness (import/GRP), and policy incentives (numbers of IZ/EPZs).
The panel data covered eight regions from 1991-2001. The model was estimated first by OLS with White correction for heteroschedasticity, and then GLS regression with fixed effects, common intercepts and differenced data. Regressions with sub-sample periods (1991-96) and (1997-2001), without Red River Delta and Southeast Regions, or without cumulative FDI, were also tried.
Agglomeration, labor cost and openness were found to influence FDI. Meanwhile, regional demand, education, and IZ/EPZs had little explanatory power. The author suggested the possibilities of uniform labor quality throughout country and the investors' tendency to avoid developed regions, which might explain the insignificance of some of these variables. Policy implications were also mentioned.
During the discussion, Ms. Diep's comments were delivered from Hanoi (she ran similar regressions in her master thesis while at Hitotsubashi). One of her points was that the simultaneity and causality problems between the dependent and independent variables might be serious. Other participants added a few more technical problems that might arise in the model. Some questioned the choice of data. By aggregating provinces into as few as eight regions, the important determinants of FDI, such as transportation access, vicinity to urban areas, ports and airports, attitudes of provincial governments, etc. might be lost. It was also pointed out that some proxy variables were crude and difficult to interpret. For example, the number of telephones was hard to justify as a variable representing the quality of "infrastructure". The author recognized the problem in data quality and availability. The merger of provinces during the period also complicated the data compilation.
After the closing of the workshop, participants continued another hour of free and friendly discussion over oolong tea.