Efficiency of China's Outward Foreign Direct Investment (OFDI) in Belt and Road Countries

Since President Xi Jinping proposed the Silk Road Economic Belt and the 21st Century Maritime Silk Road in 2013, the initiative has garnered enthusiastic responses. By 2020, China’s OFDI in Belt and Road countries/regions reached $153.712 billion, up 12.3% year-on-year, ranking first globally and contributing to world economic growth. Despite notable achievements, the BRI faces multifaceted challenges, including traditional economic and host-country political/institutional factors affecting China’s OFDI, and the business environment as a critical component of institutional context. Higher host-country institutional quality attracts more Chinese OFDI. Meanwhile, global cancellations of MIAs, RIAs, and BITs have increased institutional complexity. Belt and Road countries vary widely in development and internal/external environments, raising questions about how differing business environments affect the efficiency of China’s OFDI, what that efficiency currently is, and how China should optimize its investment layout. This paper empirically evaluates the overall efficiency of China’s OFDI using panel data for 47 Belt and Road countries from 2013 to 2019, clarifies the relationship between host-country business environment and China’s OFDI efficiency, and offers policy suggestions to optimize business environments and improve OFDI efficiency. The paper is structured with literature review, theoretical model, econometric model and data, empirical analysis, and conclusions/suggestions.

OFDI motives and influencing factors: Building on Dunning’s eclectic paradigm integrating ownership, internalization, and location advantages, and gravity-model insights from Tinbergen, research identifies host market size, geographic distance, and other factors as closely related to China’s OFDI. Expanded gravity models find negative associations with geographic distance, positive with host GDP, labor market efficiency, and technology readiness. Recent perspectives emphasize host macro institutions and micro business-environment factors. Research on business environment: Narrowly, it encompasses institutional rules relevant to market entry, operation, and exit (e.g., World Bank Doing Business indicators). Broadly, it includes the overall investment climate, including macro, noninstitutional factors. The paper uses six business-environment indicators from the World Bank governance index as inefficiency variables. Impact of business environment on OFDI: Studies are mixed on whether the host business environment facilitates or hinders OFDI. Some find positive effects of infrastructure, financial services, customs efficiency, and government management on exports and firms’ international engagement, with intra-country business-environment differences also mattering. The literature remains inconclusive, motivating the paper’s empirical assessment.

Data and sample: Panel data for 47 Belt and Road countries over 2013–2019 (listed across Africa, Asia, Europe, Oceania, and South America). Model: A stochastic frontier gravity-type framework for China’s OFDI, with a composed error structure following Aigner et al. (1977), Meeusen and van den Broeck (1977), and a panel-data inefficiency specification per Battese and Coelli (1995). Frontier (deterministic) variables include bilateral economic development (China’s GDP and host GDP), geographic distance, host trade dependence on China, host technological development, host economic freedom, common language, bilateral investment agreements, among others. Technical inefficiency term u captures investment inefficiency; larger u indicates greater investment losses and lower efficiency. Inefficiency determinants are primarily six World Bank Global Governance Indicators reflecting host-country business environment. Estimation and tests: Estimated using Frontier 4.1 with time-varying inefficiency. Likelihood ratio (LR) tests reject the absence of an inefficiency term (LR=10.4129 > 5% critical value 5.99) and reject no time variation in inefficiency (LR=10.42406 > 7.81). Variance parameters indicate sigma-squared ≈ 74.0465 and gamma ≈ 0.9991 (SE ≈ 0.000284), implying most variation is due to inefficiency. Coefficient estimates and t-ratios are reported (Table 7), and significance assessed at the 5% level. Data sources include World Bank governance indicators and official databases referenced in the dataset.

- Model validity: LR tests indicate a significant inefficiency component and time-varying inefficiency in China’s OFDI efficiency across Belt and Road countries (rejecting both nulls of no-inefficiency and no-time-variation). Variance decomposition shows gamma ≈ 0.9991, suggesting inefficiency dominates the composite error variance. - Variable significance (frontier side): Host-country technology level and whether a bilateral investment agreement (BIT) with China exists are not statistically significant at 5%. Significant drivers, in descending order of influence, are: China’s GDP (positive, largest effect), host trade dependence (positive), geographic distance (negative), common language (positive), host GDP (positive, smaller effect), and host economic freedom (negative, small effect). - Interpretations: Larger Chinese economic size increases practical OFDI, consistent with Dunning’s framework. Greater bilateral trade dependence reflects stronger trade-investment complementarity, promoting OFDI. Greater distance raises information and transaction costs, discouraging OFDI, though some prior studies suggest the distance effect may be declining. Shared language reduces communication/negotiation costs, facilitating OFDI. While host GDP is positively linked to China’s OFDI, China may prefer relatively less-developed hosts once development surpasses certain levels. Lower host economic freedom is associated with relatively greater Chinese OFDI under the BRI, implying the initiative weakens the conventional positive role of economic freedom in attracting OFDI and enhances investment toward regions with lower economic freedom. - Inefficiency interpretation: A larger value of an inefficiency determinant increases u and thus investment losses; inefficiency determinants are drawn from World Bank governance indicators representing business environment. - Selected estimates: sigma-squared ≈ 74.0465 (t≈13.63), gamma ≈ 0.99909 (t≈3517.64).

The findings indicate that both traditional gravity factors and institutional/business-environment conditions shape China’s OFDI efficiency under the BRI. Stronger Chinese economic capacity and higher bilateral trade dependence meaningfully enhance OFDI, while geographic distance remains a deterrent due to higher search, information, and coordination costs. Common language eases information frictions, improving efficiency. The limited role of host technology likely reflects many Belt and Road countries’ relatively lower technological and economic development, which leaves scope for Chinese firms to transfer technology and exploit comparative advantages. The insignificance of BITs may stem from widespread existing agreements among Belt and Road countries, diminishing marginal effects on facilitation. Notably, the negative association between host economic freedom and Chinese OFDI under the BRI suggests the initiative channels investment toward less liberalized environments, potentially due to policy support, state-to-state frameworks, and strategic motivations. The stochastic frontier results, with high gamma and significant time-variation, underscore that business-environment-related inefficiency is a key constraint. Policy implications include: for China, further leveraging innovation collaboration and resource accumulation with host countries to elevate OFDI efficiency; for host countries, proactively improving governance and business climates to reduce inefficiency, attract higher-quality Chinese investment, and strengthen trade facilitation. These patterns address the research questions by mapping how business-environment dimensions and classic gravity variables jointly influence OFDI efficiency and where policy levers can improve outcomes.

This study evaluates China’s OFDI efficiency in 47 Belt and Road countries (2013–2019) using a stochastic frontier gravity framework with time-varying inefficiency driven by business-environment indicators. It documents that China’s GDP and bilateral trade dependence are strong positive drivers, distance is a negative constraint, common language facilitates investment, host GDP is positively associated albeit modestly, and host economic freedom is negatively associated with Chinese OFDI under the BRI. Technology level and BITs are not significant in this sample. The results highlight the central role of host-country business environment in shaping inefficiency and thus OFDI effectiveness. Policy suggestions include: enhancing China–host innovation collaboration and research outputs to raise OFDI efficiency; and for host countries, improving governance, upgrading business environments, expanding financing, and increasing R&D to build core competitiveness and better utilize resources. Future research should broaden variable coverage for both frontier and inefficiency terms, expand country coverage as data availability improves, and incorporate post-2020 dynamics to capture COVID-19-era shocks and beyond, as well as micro-level firm heterogeneity to inform enterprise-level strategies.

- Variable coverage: While frontier variables include economic development, distance, trade dependence, technology, economic freedom, language, and bilateral agreements, other relevant determinants of both frontier and inefficiency were not included due to data and estimation constraints; inefficiency measures focus on World Bank Global Governance Indicators. - Data scope and sample: Data limitations led to excluding some Belt and Road countries; results therefore do not represent all BRI countries. - Time coverage: To avoid COVID-19 distortions, analysis is limited to 2013–2019, introducing potential lag relative to current conditions. - Level of analysis: The study adopts a macro perspective and does not incorporate firm-level (micro) factors, which limits the applicability of recommendations to individual enterprises.