Foreign business presence and local entrepreneurship development: evidence from China

The paper asks whether and how the presence of foreign-invested enterprises (FIEs) affects domestic entrepreneurship in host economies, focusing on China. Given China’s simultaneous, rapid growth of FDI and private business since the late 1970s, the interaction between foreign firms and local entrepreneurs is both theoretically important and policy-relevant. Theoretically, FIEs can generate positive spillovers (demonstration, knowledge transfer, market-linkage effects) but may also crowd out local entrepreneurs through competition for resources and technological barriers. The study aims to empirically identify the net effect of FIEs on provincial-level entrepreneurship in China and to assess how this effect varies over time and with changes in the institutional environment, offering lessons for other developing and transition economies.

Prior research links both FDI and entrepreneurship to improved living standards and growth, but evidence on FDI’s impact on entrepreneurship is mixed. Case studies such as Macedonia (Apostolov, 2017) and Nigeria (Onwuka and Chigozie, 2014) find positive FDI–entrepreneurship links. In China, institutional quality—property rights and contract enforcement—has been shown to affect entrepreneurship (Lu and Tao, 2010), and FDI can catalyze institutional reforms that foster entrepreneurship. More recent cross-country work (Fang et al., 2020) suggests foreign ventures may reduce necessity entrepreneurship via employment effects and increase opportunity entrepreneurship via knowledge accumulation, though empirically distinguishing these types is difficult. The literature also documents both positive spillovers (vertical linkages, innovation spillovers: Tang et al., 2019; Luo, 2016; Chen and Zhou, 2023) and potential negative crowding-out via resource competition and technology barriers (Meyer, 2004; De Backer and Sleuwaegen, 2003; Ayyagari and Kosova, 2010). Cross-country evidence indicates greenfield FDI can depress entrepreneurial activity (Ha et al., 2021). In China, political incentives tied to local economic performance may temper crowding-out. This study contributes by providing provincial-level evidence from China, splitting periods pre- and post-WTO entry, and examining how constitutional changes that improve the legal environment modulate the FDI effect.

Design: Two-way fixed-effects (TWFE) panel regressions on 31 Chinese provinces, 1992–2019. The baseline model is ln(entrepreneurship_it) = β ln(FDI_it) + Controls_it + province FE + year FE + ε_it, estimated with Driscoll–Kraay standard errors to address cross-sectional and temporal dependence when T is large. Outcome: Entrepreneurship measured as newly established private firms per 10,000 people (logged). This captures business creation, though it does not reflect firm size or industry. Key regressor: Foreign presence measured primarily by accumulated FDI stock per capita, constructed from annual FDI flows with a 5% depreciation rate: K_it = I_it + δ K_i,t−1. FDI stocks are converted to CNY using annual USD–CNY exchange rates and deflated by China’s CPI (1980=100), then divided by provincial population. Alternative specifications use FDI flow per capita and different depreciation rates (0–10%) and deflators (e.g., U.S. CPI) as robustness checks. Controls: Demographics (elderly dependency ratio; birth rate), investment intensity (fixed asset investment/GDP), human capital (college enrollments per 10,000 population), infrastructure (road kilometers per 10,000 km²), financial depth (loans/GDP), government size (government expenditures/GDP). Institutional environment captured via constitutional amendment dummies (constitution99; constitution04) and interactions with ln(FDI stock pc) to test whether improved legal protections alter the FDI–entrepreneurship link. Subsamples and robustness: Analyses split into 1992–2001 and 2002–2019 to compare pre- and post-WTO periods; pooled regressions with institutional dummies and interaction terms; alternative key regressor (FDI flow pc). All models include province and year fixed effects; Driscoll–Kraay SEs are reported. Endogeneity considerations: Potential omitted variables (e.g., institutional improvements) and reverse causality (FIEs choosing more entrepreneurial regions). The study mitigates omitted-variable bias via extensive controls and institutional proxies (marketization via proxies, constitutional dummies), but acknowledges the absence of a valid external instrument for FDI and suggests future IV-based research.

• Foreign presence promotes entrepreneurship: Across specifications, ln(FDI stock per capita) has a positive, statistically significant effect on ln(new private firms per 10,000 people).
• Magnitude by period (Table 4):
  • 1992–2001: Elasticity ≈ 0.20–0.21 (SE ≈ 0.03), significant at 1% (e.g., Model (3): 0.21). Interpretation: a 1% increase in FDI stock pc is associated with ≈0.21% higher new firm creation.
  • 2002–2019: Elasticity declines to ≈ 0.11–0.16 (SE ≈ 0.02–0.03), significant at 1%.
• Institutional interactions (Table 5): In pooled 1992–2019 models, FDI stock pc remains positive and significant. Constitutional amendments (1999, 2004) are associated with higher entrepreneurship levels (coefficients ≈ 0.36 and 0.45, both 1%). However, interactions show reduced marginal impact of FDI post-amendments: Interaction99 ≈ −0.07***; Interaction04 ≈ −0.09***. Example: baseline FDI elasticity 0.24 falls to ≈0.15 after 2004 (0.24 − 0.09).
• Alternative FDI measure: Using ln(FDI flow per capita), coefficients ≈ 0.04*** (Models (9)–(10)), smaller than stock-based estimates but still positive and significant.
• Other covariates: Fixed investment/GDP consistently positive and significant; education (college enrollment) generally positive and significant; elderly dependency ratio positive and significant (consistent with intergenerational childcare support enabling entrepreneurship); birth rate effects are mixed/negative in later period; infrastructure often positive; government size sometimes negative; loan/GDP generally insignificant.
• Model fit: High within R-squared (≈0.88–0.97 across subsamples; ≈0.97 in pooled models). Results are robust to alternative controls, sample splits, depreciation rates, and deflators.

The findings indicate that, on net, foreign-invested enterprises bolster local entrepreneurial entry in China’s provinces, addressing the central research question in favor of a positive spillover/demonstration effect outweighing potential crowding-out. The stronger effect in the 1992–2001 period aligns with a larger technology and management gap, greater learning opportunities, and stronger demonstration effects from FIEs. The diminished elasticity in 2002–2019 is consistent with narrowing gaps, possible intensification of competition, and macro/institutional changes reducing reliance on foreign demonstration. Constitutional amendments in 1999 and 2004, which elevated the legal status of private businesses and strengthened private property protection, raised overall entrepreneurship but reduced the marginal role of FDI in catalyzing new entry. This suggests that as domestic institutions improve, local entrepreneurship becomes less dependent on foreign spillovers. In China’s context, local officials’ incentives tied to economic performance may limit crowding-out, allowing complementarities between FIEs and domestic firms (e.g., supplier linkages, talent spillovers) to dominate. For policymakers in developing/transition economies, results imply that attracting FDI can stimulate entrepreneurial entry, particularly where gaps are large and institutional environments are still maturing; over time, strengthening domestic legal and market institutions may substitute for reliance on foreign demonstration effects.

Using a provincial panel (1992–2019) and TWFE models, the study shows that FIE presence—measured by accumulated FDI per capita—significantly and robustly increases new private firm creation in China. The elasticity is larger before WTO entry (1992–2001) than afterward (2002–2019). Constitutional reforms that improved the legal environment boosted entrepreneurship overall but reduced the marginal effect of FDI on new entry. Contributions include identifying a positive net FDI–entrepreneurship link in China, documenting time-variation in the effect, and showing that stronger legal protections diminish the reliance on FDI spillovers. Future research should (1) employ valid instruments to address potential reverse causality in FDI location choices, and (2) further examine evolving mechanisms and interactions between FIEs and domestic firms over time and across sectors.

• Causal identification: Despite TWFE controls and institutional proxies, endogeneity remains a concern (omitted variables; reverse causality in FDI location choices). No external instrument is employed.
• Entrepreneurship measure: New registrations per capita do not capture firm size, survival, industry composition, or quality of ventures.
• Mechanism identification: While suggestive channels (spillovers vs. crowding out, institutional change) are discussed, mechanisms are not directly identified or tested.
• Temporal heterogeneity: Although subperiods and constitutional interactions are analyzed, the full dynamics of how FDI’s role changes over time and across contexts/sectors are not fully resolved.