Determinants of foreign direct investment across China

Qian Sun, Wilson H.s. Tong, Yu QiaoJournal of International Money and Finance 2002 · DOI

TL;DR

This paper investigates the factors influencing foreign direct investment (FDI) across China's provinces from 1986 to 1998. It reveals that the importance of these determinants has shifted over time, particularly after 1991, with market size and openness becoming more crucial. A surprising finding is that a higher ratio of cumulative FDI to domestic investment negatively impacts new FDI, suggesting that simply having more FDI doesn't automatically attract more.

What Problem Does This Paper Address?

While China has attracted a massive amount of foreign direct investment (FDI), understanding the specific provincial-level factors that drive this investment has been less explored. Previous studies often used cross-country data, making it difficult to disentangle the complex interplay of social, cultural, economic, and political factors. This paper aims to provide a more granular analysis by focusing on China's provinces over time, specifically examining how the importance of FDI determinants has evolved, especially in light of China's opening up to the global economy and its eventual WTO accession. The authors are motivated by the observation that the nature of FDI in China has changed, with a shift from export-oriented assembly to investments targeting the domestic market.

Methods

The study employs panel data analysis, a statistical technique that combines cross-sectional data (across provinces) with time-series data (over years). This allows for the examination of both spatial and temporal variations in FDI. The authors use a fixed-effects regression model, which accounts for province-specific unobserved characteristics that might influence FDI. They also explore common intercept models and first-difference regressions to analyze growth rates. To address potential multicollinearity among explanatory variables, they carefully select a set of key determinants and test their robustness. The analysis is conducted on the full sample of 30 provinces and also on a subsample excluding Guangdong and Fujian, two historically dominant FDI recipients.

Datasets

The primary data source is China's Statistical Yearbook, providing provincial-level data on FDI, GDP, domestic investment, and other economic indicators. Data on firms issuing foreign equity shares were obtained from Datastream, and political risk data were sourced from Political Risk Yearbook. The study covers the period from 1986 to 1998.

Key Insights and Ideas

  • Shifting Determinants: The factors attracting FDI to China have changed significantly, especially after 1991. Market size (GDP) and openness to the outside world have become increasingly important, while the role of labor cost has also evolved.
  • Negative Agglomeration Effect? A counterintuitive finding is that a higher ratio of cumulative FDI to cumulative domestic investment (CFDI/CINV) negatively impacts new FDI. This suggests that while agglomeration effects are often cited, simply having more FDI might not guarantee further inflows, possibly due to increased competition or diminishing returns.
  • Labor Quality Matters: High labor quality, proxied by the number of engineers, scientists, and technicians, consistently attracts FDI across different periods.
  • Infrastructure is Key: Good infrastructure, particularly railway networks, is a significant draw for foreign investors.
  • Political Risk Deters: Higher political risk is a strong deterrent to FDI, leading to significant drops in investment.
  • Provincial Differences: While the model explains FDI distribution better in the more developed Eastern provinces, it is less precise for the Central and Western regions, suggesting that these regions may have less mature FDI attraction mechanisms.

Key Results

  • Full Sample (1987-1998):
    • GDP (market size) has a significant positive impact on FDI.
  • RSET (labor quality) shows a positive and significant impact.
  • RAILWAY (infrastructure) is a strong attraction for FDI.
  • RISK (political risk) has a significant negative impact.
  • CFDI/CINV (relative cumulative FDI) has a significant negative impact, suggesting that more accumulated FDI might deter new inflows.
  • Sub-sample 1 (1986-1991):
    • WAGE (labor cost) had a positive relationship with FDI, likely reflecting the demand for skilled labor from Hong Kong manufacturers.
  • GDP was not significant, as FDI was primarily export-oriented.
  • Sub-sample 2 (1992-1998):
    • GDP became highly significant and positive, reflecting the shift towards targeting the domestic market.
  • OPENNESS (degree of openness) significantly attracts FDI.
  • WAGE (labor cost) became significantly negative, indicating its role as a deterrent due to rising costs.
  • FPI (foreign portfolio investment) showed only a marginal impact, possibly acting as a substitute for FDI.

Limitations and Future Work

The authors acknowledge several limitations:

  • Nature of FDI: The study does not break down FDI by industry, and different industries might be attracted by different factors.
  • Tax Effects: Data limitations prevented the inclusion of tax effects, which have been shown to influence investment location decisions in other contexts.
  • Competition with SOEs: The paper suggests further investigation into the trade-off between opening domestic markets to foreign firms and the impact on state-owned enterprises.
  • Information Asymmetry: The role of private information of investing firms and the impact of corruption and red tape (as suggested by other research) warrant further exploration.
  • Model Fit: While the model has a good overall fit, it explains FDI distribution better in coastal provinces than in Central and Western provinces, indicating room for improvement in understanding the latter regions.

Summary updated

BibTeX
@article{sun2002determinants,
  author = {Qian Sun and Wilson H.S. Tong and Yu Qiao},
  title = {{Determinants of foreign direct investment across China}},
  journal = {Journal of International Money and Finance},
  year = {2002},
  doi = {10.1016/s0261-5606(01)00032-8},
}
Figure