Constructing a Supplier-Buyer Network: Imported Goods Flow Tracking in the Tax Transaction Records in Armenia

Abstract

Taxpayer networks are among main topics of research given the evolving applicability of modern technological solutions and raising importance for tax authorities to reveal hidden relationships or patterns in the taxpayer behavior. In this study, taxpayer networks are visualized and analyzed with the application of network analysis techniques and a generalizable methodology for product traceability over the supply chain is suggested. In essence, taxpayer networks depict the web of connections among taxpayers, highlighting the flow of goods, services, and financial transactions between them. Previous studies address the issue of product classification from transactional tax documents. Modern NLP techniques were applied for product categorization given only product description in Armenian language and with limited information. The developed algorithm was used as a starting point to identify which taxpayers were involved in the trade of a specific good and thus for network construction. A set of products was chosen for construction of the taxpayer network forming the supply chain from importers to final consumers. Adjusted by a markup, the amount of rice imported to Armenia was found and identified in the last nodes of the supply chain sold to final consumer and documented in tax receipts. Factors like local production are also considered during the analysis and solutions are suggested including price filters on the products subject to trade to filter out local production.