Logistics and Sustainability: Predictive analysis of the LPI on <em>CO₂</em> emissions, HDI and GDP Per Capita

The Logistics Performance Index (LPI), developed by the World Bank, is a global benchmark for assessing national logistics efficiency. However, most studies have treated the LPI as a dependent or descriptive variable, overlooking its potential as a predictive indicator of sustainable development. This study reformulates the LPI as a multivariable explanatory construct to evaluate the predictive capacity of its six dimensions—customs, infrastructure, international shipments, logistics competence, tracking and tracing, and timeliness—on three key sustainability indicators: GDP per capita, the Human Develop- ment Index (HDI), and CO2 emissions. Using 2023 data for approximately 120 countries from the World Bank database, thirteen statistical and machine learning models were applied, including linear regression, penalized regressions, support vector regression (SVR), k-nearest neighbors (KNN), and ensemble methods such as ExtraTrees, Random Forest, Gradient Boosting, CatBoost, and XGBoost. Model performance was evaluated using Spearman’s correlation, mean absolute error (MAE), root mean squared error (RMSE), and SHAP interpretability analysis. Among all models, ensemble algorithms—particularly ExtraTrees—achieved the highest predictive accuracy (ρ ≥ 0.79), identifying infrastructure and tracking as the most influential predictors. K-means clustering revealed three distinct logistic–environmental profiles (low, medium, and high emissions), reflecting structural heterogeneity among countries. The findings demonstrate that the LPI can function as a robust and explainable predictive tool for anticipating economic, social, and environmental outcomes, offering a data-driven framework for designing sustainability-oriented logistics policies.