Analyzing the Relationship Between Country Risk Factors and Logistics Efficiency Using Bayesian Networks

The aim of the article is to quantitatively investigate the relationships between country risk factors and the Logistics Performance Index (LPI) indicators. The article analyzes studies that examine the connection between logistics efficiency and socioeconomic factors at the country level, including country risks. It is defined that it remains unexplored how different country risk factors may affect various logistics performance indicators. The LPI assessment is presented as a weighted average based on six indicators: customs, infrastructure, international shipments, logistics competence and quality, timeliness, and tracking. The following factors have been chosen to represent country risk: investment, economic, financial, political, environmental, and corruption risks. The conducted correlation analysis has made it possible to determine that all country risk factors may not be equally important concerning different LPI indicators. To determine the relative importance of country risk factors concerning various LPI indicators, it is proposed to use Bayesian Belief Networks (BBN). This choice of method is due to the fact that other methods are unable to effectively model and analyze various «what-if» scenarios within a probabilistic network. BBN surpasses them due to its ability to analyze direct and reverse conditional relationships. The structure of the Bayesian network has been defined based on the built-in algorithms of the Hugin software environment. The structure learning has been carried out using various algorithms: PC, NPC, Greedy search-and-score, Chow-Liu tree, Naive Bayes, and others. It is found that this dataset describes the Bayes algorithm with an augmented tree (TAN) with the highest accuracy. According to this algorithm, the probability distributions of individual risks affecting LPI and critically important LPI indicators have been determined. The constructed Bayesian network was analyzed for low and high performance of individual LPI indicators to gain insights into the risk profile of countries. In particular, the reverse propagation of the impact of individual risks on LPI indicators in the case of risk reduction was examined, as well as the reverse propagation of the influence of LPI and logistical competence and quality on specific types of country risks. The results of this study may be useful in allocating limited resources to critical risk factors based on the specific risk profile of the country.