An analysis of bilateral trade of belt and road initiative countries using the gravity model

The Belt and Road Initiative (BRI), officially launched in 2013, aims to fill the power vacuum created in Central Asia after the dissolution of the Soviet Union by revitalizing the historic Silk Road with a modern vision. This initiative seeks to more effectively connect China's rapidly growing economy with markets in Asia, Europe, and Africa. The project, shaped under the leadership of Chinese President Xi Jinping and enhanced by land and sea routes, promotes regional trade and economic development. In 2013, announcements made in Kazakhstan and Indonesia introduced the 'Silk Road Economic Belt' and '21st Century Maritime Silk Road' as the two main components of the BRI, presenting a new model of economic cooperation worldwide. The China-Mongolia-Russia, New Eurasia, China-Central Asia-West Asia, China-Indochina Peninsula, China-Pakistan, and Bangladesh-China-India-Myanmar economic corridors, along with maritime routes through the South China Sea, South Pacific, and Indian Ocean to the Mediterranean, have the potential to transform global trade. This extensive network has fundamentally altered the trade dynamics and infrastructure connections among BRI countries, creating new commercial opportunities for member constituents and deepening global economic integration. From this standpoint, our study aims to identify the primary determinants of bilateral trade flows among the countries involved in the project from 2002 to 2020, aiming to guide policymakers and decision-makers in achieving the project's objectives. Drawing on the structural gravity model, frequently utilized in the international economics literature, we first analyzed the expanded gravity model in logarithmic linear form using panel estimators and then conducted parameter estimations with the Poisson pseudo maximum likelihood estimator, which is robust against zero trade flows and heteroscedasticity in error terms. Furthermore, to test the robustness of the findings, we utilized observations spaced at two-, three-, and four-year intervals to accommodate potential changes in trade policies during the study period. Overall, the analysis results demonstrate that the explanatory variables have collectively significant effects on bilateral trade flows. The goodness of fit measures demonstrates that the expanded gravity model and the structural gravity model employed in the study explain a substantial portion of the variance in the dependent variable. Analyses conducted with both panel estimators and continuous and interval observations using the Poisson pseudo maximum likelihood estimator show that there are no model construction errors in the specifications that proxy multilateral trade resistance terms with exporter-time and importer-time fixed effects and also include country-pair fixed effects to control for potential endogeneity in trade policy variables. Findings from various estimators unravel that the economic sizes of the exporting and importing countries bear a positive and statistically significant impact on bilateral trade flows. The effect of the distance variable on bilateral trade flows is consistently negative and statistically significant, albeit the measured elasticity is sensitive to estimator choice. The real effective exchange rate index consistently negatively affects bilateral trade flows, although its impact on the dependent variable is often statistically insignificant. The similarity in demand structures of trade partners negatively and significantly affects bilateral trade flows between BRI countries, supporting the theory of preference similarity. Observable trade costs, represented by the contiguity relationship and the presence of a common language, consistently have a positive effect on bilateral trade flows. The impact of colonial ties on bilateral trade flows is statistically insignificant, positive, and relatively less significant compared to other dummy variables. Religious proximity has a generally statistically significant and positive impact on bilateral trade flows. Ultimately, the presence of free trade agreements has a consistent, significant, and strong positive effect on the dependent variable, irrespective of estimator choice.