Using Network Analysis to Map Public Procurement Activities in Ceará Municipalities (Brazil)

Illegal activities in public treasuries worldwide pose a significant threat to the financial stability of cities, states, and countries, resulting in monetary losses. Despite their widespread occurrence, empirical research has not furnished proper analytical tools for objective investigations. More practical insight remains scarce, limiting the effectiveness of public-private networks in efficient spending interactions. Public tenders, the primary mechanism for government procurement, are highly susceptible to fraud. Control institutions make efforts, albeit with limited success, to identify fraudulent activities between governments and firms during the procurement process.
This thesis aims to enhance the understanding of fraud in public procurement and explore the benefits of statistics and network analysis in detecting signs of fraud. Using procurement data from 184 municipalities of Ceará, Brazil, the research spans two empirical studies. The thesis begins with a literature review on data-driven methods for detecting fraud, corruption, and collusion in public procurement activities.
In practical application, it starts with a characterization of the entire procurement network to map the data's structure and complexity. Network science methods examined firms co-bidding relations, identifying communities susceptible to cartel formation. A composite corruption risk indicator (CRI) is adapted based on local empirical data. The CRI focused on binary relations between tenders and winners, using red flags at the contract level. Firms and municipalities were classified into four corruption risk bands based on their scores, allowing the identification of groups more susceptible to fraud.
Overall, this doctoral thesis advances our understanding of irregular activities in public procurement through statistics and network analysis methodologies, providing insights into fraud detection, risk assessment, and the identification of vulnerable groups within the public procurement process.