The gravitational model is one of the most effective empirical methods in the field of social sciences [14]. In particular, Isard and Tinbergen pioneered the application of the gravitational model to describe patterns of bilateral aggregated trade flows between countries [23, 24]. Their work has produced a vast empirical literature that seems to work well in modeling trade flows and exploring the factors that influence them [14, 25, 26], as flow variation could be captured by the adapted relationship [27]. It is instructive to compare our statement on incentives to set tariffs with the tariffs that would result from compliance with WTO rules. Article XX of the WTO stipulates that members of a preferential trading region do not increase their trade barriers, obliging all members of a customs union to align their external trade policies with the most favourable trade policies of their members. However, the increased incentive to protect trade for members of a customs union complicates the next step towards multilateralism. In short, WTO rules encourage customs unions to take a bigger step towards multilateralism than free trade agreements. But in terms of incentives for further trade liberalisation, customs unions are becoming a stumbling block on the road to multilateralism, while free trade agreements facilitate the next step and are therefore a cornerstone of multilateralism. Preferential trade agreements (EPAs) refer to unilateral trade privileges such as the Generalized System of Preferences (GSP) and non-reciprocal preferential programmes that some WTO Members implement for goods from developing and least developed countries. Information on preferential trade agreements notified to the WTO is available in the PTA database. Second, according to empirical results, the density of the network has decreased during these ten years, which means that the countries of the global commercial network are weaker connected (the cluster coefficient has increased from 0.1370 to 0.1006). In addition, international trade is a system involving many countries and trade relations, and modeling complex networks has the advantage of analyzing a number of complex entities and relationships [19-21]. In addition, network theory can also facilitate the study of local and global properties [22], which is consistent with the purpose of our work.
However, trade flows are the direct result of trade openness, and related studies generally use trade flows to weight the network [9, 19]. Since trade flows could be influenced by a country`s economic volume, geographical factors and artificial barriers, we prefer trade resistance that eliminates the impact of the economy to reflect the purity of trade relations between countries. In addition, the Communities are represented in the international trade network by groups of countries in which trade relations between the countries of the same Community are closer than those of the different Communities [9]. Therefore, comparing union members and typical business communities in the global trade network could facilitate research on the effectiveness of RTAs. In this document, we use trade data from the United Nations Comtrade database, which includes 198 countries/districts. Here we select the product type «Goods» and use the annual sum of all HS (Harmonized Commodity Description and Coding Systems) goods. Due to differences in time and statistics, the flow data reported by the importer and exporter are not always the same. Here we use the importer`s report, with an addition of exporters if the data is missing. The EM algorithm attempts to obtain the marginal probability MLE (Maximum Likelihood Estimate) by iteratively applying the wait step (step E) and the maximization probability step (step M), representing the number of iterations. The detailed process is as follows:(1)E-step: Calculate by step, based on the last parameter estimate, the expected value of the probability of belonging to a certain category. Calculate separately the observation probabilities belonging to category I and category II: then normalize them as follows: The unobserved latent variables, where represents the probability of resistance to trading of category I. (2) Step of probability of maximization (step M): Based on the value obtained from step E, we find the estimate of the parameter that maximizes this probability. The probability function of occurrence is multiplied by the expected probability of all trade resistances as follows: The optimal value of is based on and can be calculated from this function: Regional trade agreements (RTAs) today cover more than half of international trade and operate in parallel with global multilateral agreements within the framework of the World Trade Organization (WTO).
In recent years, many countries have actively sought new bilateral and regional trade agreements – often more modern and progressive – aimed at boosting trade and economic growth. The current prevalence of RTAs partly reflects the call for deeper integration than has been achieved through older multilateral agreements. The innovation of this article is to examine and describe the business unit relationships between countries, separating other typical factors such as economic volume, geographical distance, overall increase in transportation and labor costs. .
