Detecting corruption in the digital age: big data issues and reform opportunities

The rise of big data offering real-time access to large volumes of highly diverse micro-level electronic data is the most visible face of the third generation of corruption measures. It has raised expectations for innovative methods of detecting corruption and supporting both bottom-up and top-down anti-corruption efforts. The datafied society has indeed opened avenues for new quantitative analyses, most of them related to corruption in public procurement and its political connections. This has emerged mainly in the form of analyzing party and campaign financing, bid-rigging and beneficial ownership frauds, and network analysis based on information available in the news and court documents. However, challenges remain. Academic research using big data is still scarce as it requires, at the very least, access to good-quality data and specialized data processing and analytical tools. This chapter asks whether corruption studies have reached a plateau requiring new connections and measurements using big data, and if so, why. The chapter argues for new measures, in particular, to address “legal corruption” by establishing links between corruption and practices such as lobbying, revolving doors, and to refine our notions of, and tests for, state and media capture. It also highlights the need for reforms to tackle what is referred to here as “big data ISSUES” (Inaccessible, Scattered, Structureless, Unreliable, Erratic, and Sizable data), a mix of existing previous problems with new ones in terms of collecting and processing large volumes of data to measure and better understand corruption. This is crucial to prevent naive uses or superficial big data analytics reducing the value of AI-based anti-corruption technologies, which, if misused, could conceal corruption and mislead analysts and law enforcement.