Using Network Analysis to Detect Corruption in Public Procurement: The Bulgarian Case as a Testing Ground

The topic of political corruption remains at the center of the public discourse. With the growing role of mass and social media and the shortening of information routes as well as the growing volume of public information illicit practices are pinpointed by civil society and investigative journalists more often on all possible levels. The academia has recognized, that clientelism, corruption, and patronage are in the basis of ineffective public institutions (Fukuyama, 2017). However, due to the hidden character of corruption, its research remains a challenge. The known methods include - surveys of perception (e.g Transparency CPI), and victimization (Eurobarometer), analysis of the trials of corruption(Shukla 2013), attempts for objectivization of corruption (e.g through discrepancies, of public spending and quality of provided services (Picci and Escresa 2017), (Putnam 1994). This paper will adopt a novel and emerging approach, which is risk assessment through network analysis in public procurement. It will try to assess the virtue of this method through an analysis of the public procurements in Bulgaria for the past five years, where a few political powers and caretaker governments went into power. If significant alterations in the networks of contractors that coincide with changes in political leadership are observed, this will strongly suggest the presence of favoritism and potential corruption within tender procedures. This hypothesis is based on the premise that political shifts should not inherently cause major changes in contractor networks if the procurement processes are transparent and merit-based and thus will act as a proxy indicator. The reasons the public procurement sector is from a specific interest are - regardless, of the growing regulation of this public sector, alleged unlawful activities still occur, especially in developing democracies. Between 15 and 30 percent of the world's GDP is distributed through the mechanism of public procurement, according to Transparency International, and from 15 to 45 percent (in the most extreme cases) sinks into corruption (Wiehen, 2006). A similar approach to examining corruption has been undertaken by Fountoukidis, Antoniou, and Varsakelis (2023) who conducted exhaustive research on the market competition in the EU, through research on the market entropy and clusterization, and found indications for preferential treatment. Wachs, Fazekas, Kertész (2021) found a correlation between the centralization of public procurement markets and corruption risks. De Domenico et al. (2019) claims that network analysis for examining corruption has few benefits - (1) Complex systems are usually composed of multiple interacting elements, each having different types of interactions. (2) The structure of complex systems can be represented as a network of interactions and relationships (3) Complex systems often exhibit collective and nonlinear dynamics, which makes them unpredictable. These characteristics highlight the need for comprehensive approaches to studying and understanding their dynamics and behavior. The Bulgarian case of recent years represents fertile grounds to test the described methodology and hypothesis, which can then be applied universally throughout systems with sufficient open data about procurement.