The impact of international logistics performance on import and export trade: an empirical case of the "Belt and Road" initiative countries

The paper examines how international logistics performance influences China’s import and export trade with Belt and Road Initiative (BRI) countries amid accelerating globalization and deepening regional cooperation. Trade between China and BRI partners has expanded, reaching 34.7% of China’s total foreign trade in 2022, with concentration in Southeast Asia and notable imbalances. The World Bank’s Logistics Performance Index (LPI) provides a benchmark for logistics capacity that underpins supply chains and trade facilitation. To avoid confounding from the 2008 financial crisis, the authors study 2011–2022 for 61 BRI countries, assessing how LPI and its sub-indicators affect China’s bilateral trade flows and whether effects differ by country size. The study aims to clarify the interaction mechanism between logistics performance and trade, provide evidence using a fixed-effects gravity model with robustness checks, and offer policy guidance for improving logistics and trade within the BRI.

Prior research generally finds that better logistics performance promotes international trade by reducing costs, improving efficiency, and enhancing trade openness (e.g., Hausman et al., Çelebi, Barakat et al., Jayathilaka et al.). Trade growth, in turn, can bolster logistics industry development, with heterogeneous regional effects within China. Studies also report heterogeneity across countries by income level, trade facilitation, and population size; logistics improvements may yield larger benefits for low-income economies and have varying impacts by country size (Kumari and Bharti). However, evidence on how logistics performance affects trade across different country sizes remains limited. This paper addresses this gap by classifying BRI countries by population size and assessing differential effects of LPI and its sub-indicators on China’s import and export trade.

Data: Panel data for 61 BRI countries (40 in Asia, 20 in Europe, 1 in Africa) over 2011–2022, primarily from the World Bank WDI and related LPI datasets; trade data from Prospective Database. Missing biennial LPI values are linearly interpolated using Stata 15. All continuous variables are standardized. Countries are classified by average 2011–2022 population into large (top 25%), medium (25–75%), and small (bottom 25%). Variables: Explained variable is China’s bilateral trade volume (imports+exports) with each BRI country. Explanatory variables include LPI (overall) and its six sub-indicators: TRACE (tracking), SERVICE (logistics services), SHIPMENTS (international shipments/price competitiveness), CLEARANCE (customs efficiency), TIME (timeliness), INFRASTRUCTURE (infrastructure quality). Controls: DIS (distance), modeled as geographic distance multiplied by Brent crude oil price to allow time variation; GDPJ (partner GDP); GDPC (China’s GDP); OPEN (trade/GDP of partner); BORDER (contiguous dummy); WTO (WTO membership dummy). Model: An extended gravity model is estimated in log form with country and year fixed effects. Baseline specification: lnTRADE = β0 + β1 lnLPI + β2 lnDIS + β3 lnGDPJ + β4 lnGDPC + β5 OPEN + β6 BORDER + β7 WTO + ε. Separate regressions replace LPI with each sub-indicator. Subsample analyses are run by country size (large, medium, small). Diagnostics: VIF < 10 indicates no multicollinearity; Hausman test (p=0.0017) supports fixed effects over random effects. Robustness checks include removing outliers (four smallest-trade countries), shortening sample period to 2011–2020, tail trimming (5–95%), and lagging the dependent variable by one period.

- Baseline fixed-effects regression (2011–2022): lnLPI has a positive effect on lnTRADE (coefficient ≈ 0.062, significant at 10%). Distance is negative and significant (DIS ≈ −0.116, 1% level). Partner GDP (GDPJ ≈ 0.803, 1% level), China’s GDP (GDPC ≈ 0.100, 1% level), and openness (OPEN ≈ 0.223, 1% level) are positive. BORDER is negative and marginally significant (≈ −0.048, 10%); WTO is not significant. - Heterogeneity by country size (Table 5): • Large countries: LPI strongly positive (≈ 0.382, 1%), DIS negative (≈ −0.275, 1%), OPEN strongly positive (≈ 1.263, 1%), BORDER positive (≈ 0.206, 1%), WTO not significant. • Medium countries: LPI not significant (≈ −0.010), DIS negative (≈ −0.060, 10%), OPEN not significant. • Small countries: LPI negative (≈ −0.023, 10%), DIS negative (≈ −0.020, 5%), OPEN negative (≈ −0.077, 1%), BORDER positive (≈ 0.025, 1%). - LPI sub-indicators (full sample, Table 6): SERVICE positive (≈ 0.055, 10%); INFRASTRUCTURE positive and significant at 1%. TRACE, SHIPMENTS, CLEARANCE, TIME are not significant overall. - LPI sub-indicators for large countries (Table 7): SHIPMENTS ≈ 0.254 (5%), SERVICE ≈ 0.531 (1%), CLEARANCE ≈ 0.485 (1%), INFRASTRUCTURE ≈ 0.449 (1%); TRACE and TIME not significant. - Robustness checks: After removing outliers, LPI ≈ 0.081 (5%); with 2011–2020 window, LPI ≈ 0.049 (10%); with 5–95% tail trimming, LPI ≈ 0.051 (5%); with lagged dependent variable, LPI ≈ 0.058 (10%). Core conclusions remain stable.

The findings support the hypothesis that improvements in international logistics performance among BRI countries promote China’s bilateral trade, consistent with gravity theory and trade facilitation literature. The strong positive effects for large-population partners suggest that logistics upgrades in sizable markets translate more readily into increased trade volumes, likely due to higher aggregate demand and more extensive supply chain networks. In contrast, for small and some medium-sized countries, LPI improvements do not significantly raise trade with China—and may even correlate negatively—possibly because improved domestic logistics enables better internal resource allocation and import substitution, or because aggregate demand constraints limit external trade expansion. Distance remains a salient trade cost, underscoring the need for infrastructural and procedural improvements to mitigate frictions. Among logistics dimensions, service capability and infrastructure quality most consistently underpin trade growth, especially in large countries, while tracking, timeliness, and customs efficiency matter less overall but are important for large countries (customs) when considered by size. The robustness checks bolster confidence that these relationships are not driven by outliers, pandemic-era turbulence, distribution tails, or reverse causality concerns.

The study demonstrates that higher international logistics performance in BRI partner countries is associated with greater China–partner trade, with the strongest positive effects in large countries. Overall LPI is positively related to trade, and within LPI, logistics service capacity and infrastructure quality are key drivers; for large countries, international shipment competitiveness and customs efficiency also significantly promote trade. Policy implications include: improving logistics performance to lower trade costs; strengthening infrastructure and logistics services to counter distance-related frictions; and prioritizing logistics investments in large BRI countries to maximize trade gains and catalyze broader regional development.

The sample is limited to 61 BRI countries over 2011–2022 and relies on publicly available databases with some interpolated LPI values, which may constrain generalizability. Country classification is based solely on population size; other relevant heterogeneities (income level, geography, development stage) are not explicitly modeled. The empirical strategy centers on a fixed-effects gravity framework with a limited set of controls and sub-indicators; alternative methods and additional variables could refine causal inference. Future research should expand to non-BRI countries, explore alternative stratifications (income, location, development), and apply complementary econometric designs.