Exploring Intra-Industry Trade Dynamics: Spatial Analysis of Türkiye's Provinces

This study investigates intra-industry trade (IIT) among the 81 provinces (NUTS-3) of Türkiye from 2004 to 2021 using Grubel-Lloyd’s methodology with ISIC rev.4 - 4 digit data. Spatial autocorrelation analysis reveals significant clustering, with the western region demonstrating high-high IIT characteristics, whereas the eastern region exhibits low-low patterns, indicating regional disparities. Different spatial estimators, such as the spatial Durbin model (SDM), spatial autoregressive model (SAR), spatial error model (SEM), spatial autoregressive and autoregressive conditional heteroskedasticity model (SAC), and econometric models, are employed to explore spillover effects among cities. However, no statistically significant spillover effects are detected, suggesting IIT is linked to production fragmentation among Türkiye’s cities. Hausman test results show that the most efficient estimators in the analysis are SAR, SEM, and SAC under fixed effects models. The time effects model yields the most significant results, indicating a positive relationship between GDP, GDP per capita, government spending, patent registrations, metropolitan status, and IIT. Conversely, a negative association is found with brand registrations. Trade openness, coastal access, border proximity, and landlocked status do not significantly impact IIT.