Global e-waste trade network analysis

In this paper, a study for Waste Electrical and Electronic Equipment (WEEE) trade is conducted by using graph theory. In specific, exports and imports for UN COMTRADE data code 854810 which corresponds to waste and scrap of prim cell are collected for 175 countries around the world, spanning the period from 2002 to 2014. WEEE trade networks are generated for each year and communities are produced applying spinglass community detection algorithm. Communities are compared with groups of countries produced by applying detection community algorithms on networks based on common currency, differences in CO₂ levels, geographical distances, common language, colonial ties, and regional trade agreements (RTA). An estimation of the factors that affect key network metrics has also been conducted, using a random effect linear regression. The model assesses the effect that economic, environmental, geographical, and social, as well as intra-country commercial agreements have on degree of nodes, betweenness score, and clustering coefficient. The results indicate that communities of WEEE trade network are very similar with groups produced by clustering countries regarding CO₂ emissions and distance. Distance, contiguity, common currency, colonial ties, common language, and differences in CO₂ levels tend to affect significantly the degree of countries engaged in WEEE trade network. Betweenness score is affected only by common currency while clustering coefficient by common language and CO₂ levels between countries. A statistical validation of WEEE network, with Erdos – Renyi, Small – World and Scale – Free networks, was conducted. The results reveal that in, cycle, and middle clustering coefficients of Erdos – Renyi and Small – World networks were statistically equal to the corresponding of WEEE network for the period 2004–2008, while Scale – Free's out and clustering coefficients coincided with WEEE's across all years.